Journal of Economic and Social Studies Volume 3 Number 2 JECOSS Print ISSN: 1986 - 8499 Online ISSN: 1986 - 8502 Fall 2013 PUBLISHER International Burch University Address: Francuske revolucije bb. Ilidža 71210 Bosnia and Herzegovina Phone : + 387(0) 33 944 400 Phone : + 387(0) 33 782 130 Fax : + 387(0) 33 782 131 Editor Associate Editor Teoman DUMAN Kursad OZLEN tduman@ibu.edu.ba kozlen@ibu.edu.ba Editorial Assistants Mersid POTURAK Merdžana OBRALIĆ mpoturak@ibu.edu.ba mobralic@ibu.edu.ba Journal Manager Recep ZİHNİ rzihni@ibu.edu.ba ISSN (Print) 1986 - 8499 ISSN (Electronic) 1986 - 8502 Frequency Biannually Current Volume 3/2013 Indexing and Abstracting : Econlit (Journal of Economic Literature) Ebscohost Academic Search Complete Pro Quest LLC ASOS Index Sosyal Bilimler İndeksi Index Copernicus Journal Master List STATEMENT OF PURPOSE Journal of Economic and Social Studies (JECOSS) aims to develop scientific knowledge in the areas that include, and are related to Economics, Business Administration, Public Administration, Political Studies, International Relations, Labor Economics and Industrial Relations, Econometry, Sociology and Psychology. As an international social sciences journal with interdisciplinary feature, it will set a ground to bring social science communities across disciplines identified above with a view for sharing information and debate. The journal publishes refereed articles and technical research notes that build on theory and contemporary scientific knowledge. Articles submitted to JECOSS will be peer-reviewed and expected to report previously unpublished scientific work. Submitted manuscripts should follow journal guidelines and should not be under consideration elsewhere. EDITORIAL BOARD AKIN, Zafer BAYRAKTAROGLU, Serkan CIVAN, Abdulkadir TOBB University of Economics and Technology, TURKEY Suleyman Sah University, TURKEY Gediz University, TURKEY COSKUN, Ali Fatih University, TURKEY DEMIROZ, Fatih Florida International University, USA DHAOUI, Abderrazak University of Sousse, TUNISIA DREHER, Sabine York University, CANADA DUMENCI, Levent Virginia Commonwealth University, USA LEE, Chew Ging The University of Nottingham, Malaysia Campus, MALAYSIA GULER, Bulent Indiana University at Bloomington, USA HAFALIR, Isa Carnegie Mellon University, USA HARCAR, Talha Pennsylvania State University, USA IKRAMOV, Nuriddin Sacramento State University, USA IZZI DIEN, Mawil Trinity Saint Davis University, UK KANIBIR, Huseyin Bursa Orhangazi University, TURKEY KAPUCU, Naim University of Central Florida, USA KARAMUJIC, Muharem The University of Melbourne, AUSTRALIA KURTOVIC, Emir Sarajevo University, BOSNIA AND HERZEGOVINA LONCAR, Dragan University of Belgrade, SERBIA MARTINOVIC, Maja Zagreb School of Economics and Management, CROATIA OMERBEGOVIC ARAPOVIC, Adisa Sarajevo Graduate School of Business, BOSNIA AND HERZEGOVINA OZDEVECIOGLU, Mahmut Meliksah University, TURKEY OZTURK, Ahmet Bulent University of Central Florida, USA PASIC, Mugdim Sarajevo University, BOSNIA AND HERZEGOVINA PINNINGTON, Ashly H. The British University in Dubai, UAE RAJIC, Vesna University of Belgrade, SERBIA RAMIC, Sukrija Zenica University, BOSNIA AND HERZEGOVINA ROSE Andrew K. University of California, Berkeley, USA SARI, Ramazan Middle East Technical University, TURKEY SAVEVSKA, Maja University of Warwick, UK STOCKER, John University of Delaware, USA SUVANKULOV, Farrukh Zirve University, TURKEY TOURK, Khairy A. Illionis Institute of Technology, USA UC, Mustafa Epoka University, ALBANIA ULUBASOGLU, Mehmet Ali Deakin University, AUSTRALIA WITKOWSKI, Jaroslaw University of Economics, POLAND YILDIRIM, Yildiray Syracuse University, USA JECOSS Journal of Economic and Social Studies Volume 3 Number 2 Fall 2013 CONTENTS Refereed Articles 5 Do Private Savings Offset Public Savings in Turkey? Hasan Göcen, Hüseyin Kalyoncu And Muhittin Kaplan, 15 Store Personality: Perceptions Towards Consumer Electronics Chain Stores in Turkey:A Case of University Students Keti Ventura, Ipek Kazancoglu, Elif Ustundagli and Rezan Tatlidil 39 The Role of Twin Deficit Problem in Sustainable Growth: An Econometric Analysis for Turkey Halil UCAL and Mehmet BOLUKBAS 53 Cox Regression Models with Time-Varying Covariates Applied to Survival Success of Young Firms Aygül Anavatan and Murat Karaöz 69 Unit Root Properties of Energy Consumption and Production in Turkey Özgür Polat, Enes E. Uslu, Hüseyin Kalyoncu, 87 Measuring and Reporting Cost of Quality in a Turkish Manufacturing Company: A Case Study in Electric Industry Hilmi Kirlioglu and Zülküf Çevik, 101 Social Anxiety and Usage of Online Technological Communication Tools among Adolescents Bilal Sisman, Sinan Yoruk and Ali Eleren 115 State as the Source of Wealth: In Ottoman Economic Thought: A different approach to reflections in the aftermath of the global crisis Birol Çetin 131 Impact of Military Expenditure and Economic Growth on External Debt: New Evidence from a Panel of SAARC Countries Khalid Zaman, Iqtidar Ali Shah, Muhammad Mushtaq Khan and Mehboob Ahmad, 149 Government Expenditure on Nomadic Education in Nigeria: Implications for Achieving the Millennium Development Goals Akighir David Terfa, Okpe. I 166 List of Reviewers for This Issue Journal of Economic and Social Studies Do Private Savings Offset Public Savings in Turkey? Hasan GÖCEN Department of Economics, Faculty of Economics and Administrative Sciences Meliksah University, 38280 Kayseri, Turkey hgocen@meliksah.edu.tr Hüseyin KALYONCU Department of International Trade and Business, Faculty of Economics and Administrative Sciences Meliksah University, 38280 Kayseri, Turkey hkalyoncu@meliksah.edu.tr Muhittin KAPLAN Department of Economics, Faculty of Economics and Administrative Sciences Meliksah University, 38280 Kayseri, Turkey mkaplan@meliksah.edu.tr AbstrAct The issue of whether public savings offset private savings, and visa vice, has important implications for the effectiveness of fiscal policy. This study examines long-run relationship between public and private savings rates using annual Turkish data for the period 1975-2005. The result of Engle-Granger cointegration test has shown that there is no long-run relationship between private and public savings ratios. However, once endogenously determined structural break is allowed, the test results confirm the existence of the cointegration relationship between private and public savings. Econometric estimation of the offset coefficients using both FMOLS and DOLS yields values of between -0.11 and -0.82. The results also indicate that the potency of fiscal policy significantly reduced with the liberalization of financial markets. KEYWOrDs Savings, Offset Coefficient, Ricardian Equivalence, DOLS, FMOLS. ArtIcLE HIstOrY Submitted: 29 Jun 2012 Resubmitted: 13 December 2012 Accepted: 25 December2012 JEL codes: E6, H6, E21. Volume 3 Number 2 Fall 2013 5 Hasan GÖCEN / Hüseyin KALYONCU / Muhittin KAPLAN Introduction The relationship between private and public savings has been central issue in both the theoretical and the empirical literature. The importance of the subject stems from the fact that the effectiveness of fiscal policy is closely related to the responsiveness of private saving to changes in fiscal stance. The relationship between lower public deficits and national savings, however, remains controversial both theoretically and empirically. Theoretically, while Keynes (1936) assumes no relationship between private and public savings, Friedman (1957) and Modigliani (1946) develop models showing full substitution between private and public savings. Barro (1974) also introduced the notion of perfect substitutability between private and public savings, which is called Ricardian Equivalence Proposition (REP). Although there area number of opposing views in the theoretical literature, ultimately, it is an empirical issue to determine the extent to which private savings offset public savings. In the empirical literature, the relationship between private and public savings is investigated for different countries using different econometric methodologies. However, there is no consensus over the size offset coefficient (for a survey see Seater, 1993, Holmes 2006 and Ricciuti 2007). Studies on advanced economies have shown that about half of the change in public savings is offset by an opposite change in private saving (Masson et. al. (1998); Hemming et. al. (2002); Holmes (2006); Mandal and Payne (2007); Seater and Mariano (1985); Leiderman and Razin (1988); Makin and Narayan (2009); De Castro andFernandez (2009)). Although empirical studies are limited in number, offset coefficients were found to be higher for developing countries than for developed countries (Loayza et. al. (2000); Lopez et. al. (2000); De Mello et. al.(2004); Edwards (1996); Masson et. al. 1998; Bulir and Swiston (2009)). This study provides evidence on the validity of the REP by applying powerful econometric techniques of DOLS and FMOLS to time series data of a developing country, Turkey. This paper is organized as follows. Following section sets out the econometric methodology and the data employed in this study. Then, we presented the findings of the study in the empirical section. Last section concludes. 6 Journal of Economic and Social Studies Do Private Savings Offset Public Savings in Turkey? Methodology and Data Empirical studies on testing the REP estimate the following model: (1) where refers to private sector savings as a proportion of GDP, is public sector savings as a ratio to GDP; is the long-run public-private offset (substitution) coefficient is the intercept term and represents usual error term. takes value between 0 (no offset) and -1 (full offset). If , then a decrease in public sector savings is fully offset by an increase in private sector savings. The data employed in our empirical analysis is an annual private and public sector as a percentage of GDP obtained from State Planning Organization (SPO) publications for the years 1975 and 2005. Before estimating the long-run offset function given in equation (1), we first need to investigate the time series properties of the private and public sector saving ratios. Results obtained from unit root tests which are performed to determine whether savings variables have a unit root are presented in Table 1a and Table 1b. While Table 1a presents the results obtained from the ADF, DF-GSL, PP, KPSS and ERS unit root tests, Table 1b shows the Ng-Perron unit root test results. Examination of the Tables show that the null hypothesis of unit root could not be rejected for both private and public sector savings ratios. Table 1.a. Unit Root Test Results Constant -1.432876 -1.367547 -1.454917 0.538798 8.002194 ADF DF-GLS PP KPSS ERS PSR Constant and Trend -1.133958 -1.595668 -1.253357 0.110454 13.83224 Constant -1.473065 -1.384922 -1.479741 9.029962 8.084297 GSR Constant and Trend -2.322051 -1.798766 -1.480789 0.380299 12.96383 Note: ADF, DF-GSL, PP, KPSS and ERS stand for Augmented Dickey-Fuller (1979), Phillips Perron (1988), Elliot, Rothenberg, and Stock (1996), Kwiatkowski, Phillips, Schmidt and Shin (1992), Elliot, Rothenberg, and Stock point optimal (ERS, 1996) unit root tests. Volume 3 Number 2 Fall 2013 7 Hasan GÖCEN / Hüseyin KALYONCU / Muhittin KAPLAN Table 1.b. Ng-Perron Unit Root test Results PSR GSR Asymptotic critical values*: 1% 5% 10% MZa -3.24375 -3.23349 MZt -1.25975 -1.27022 MSB 0.38836 0.39283 MPT 7.53622 7.57531 -13.8000 -8.10000 -5.70000 -2.58000 -1.98000 -1.62000 0.17400 0.23300 0.27500 1.78000 3.17000 4.45000 Note:The number of lags used in Ng-Perron (2001) unit root test is determined by Schwarz Information Criteria(SIC) and turned out to be zero for all specifications. Having established that private and public savings ratios are I(1) variables, we need to test for cointegration between private and public savings to avoid spurious regression. To determine whether there is long-run relationship among these variables, the Engle-Granger (1987) methodology is employed. Testing for cointegration within this methodology involves extracting the residuals from equation (1) and testing for unit root in residuals. The Engle-Granger bivariate cointegration equation and the ADF tests applied to residuals are reported in Table 2. The optimal lag determined by using Schwarz and Akaike information criteria turned out to be zero. The cointegration test statistic is -2.086 with a probability value of 0.251 implying non-rejection of the null of unit root in residuals. Hence, there appears to there is no long-run relationship between private and public sectors savings ratios. Table2. Engle- Granger Cointegration Test Dependent Variable PSR Constant 20.157 (0.531)* ADF test statistics (probability): -2.086 (0.251) Test Critical values: 1% level 5% level 10% level GSR -1.009 (0.101)* -3.671 -2.964 -2.621 Note: The values in parenthesis are standard errors. * indicate significant at 1% level. Furthermore, we employed Johansen multivariate cointegration tests to explore if there is a long run relationship between private and public savings. The number of 8 Journal of Economic and Social Studies Do Private Savings Offset Public Savings in Turkey? lags used in the underlying the vector auto regression (VAR) model were determined as one for the model according to both the Schwarz Bayesian Criterion (SBC) and Akaike’s Information Criterion (AIC). The corresponding values of SBC and AIC criterion are 9.160 and 8.867 respectively. The cointegration test results obtained from the Johansen and Juselius (1990) method for the model (1) is presented in Table 3. The examination of the table indicates that the null hypothesis of no cointegration cannot be rejected by both the maximum eigenvalue and the trace statistic for the model implying that there is no long-run relationship between private and public savings. Table 3. Johansen-Juselius Maximum Likelihood Co integration Tests Trace Test Null Alternative r=0 r≥1 r≤1 r≥2 Statistic 11.946 1.881 Critical Values 15.494 3.841 Maximum Eigenvalue Test Null Alternative Statistic r=0 r=1 10.065 r≤1 r=2 1.8810 Critical Values 14.264 3.8415 Notes: Asterisks (*) denotes statistical significance at 5%. R stands for the number of cointegrating vectors. However, the cointegration tests have a low power in the presence of a structural break (Gregory and Hansen, 1996). For this reason, we applied Gregory-Hansen cointegration procedure to test whether there is long-run relationship among private and public savings. Specifically, Gregory and Hansen (1996) provide the following three structural break alternatives given by equations (2a-2c): (2a) (2b) (2c) where D represents a dummy variable equal to 0 if is less than or equal to unknown timing of change , otherwise it is equal to one; is time trend; other variables are defined as before. The first co integration regression (2a) is allowed to have a level break, the second model includes level shift and time trend and third model includes regime shift variable. Volume 3 Number 2 Fall 2013 9 Hasan GÖCEN / Hüseyin KALYONCU / Muhittin KAPLAN Given that the structural break point is unknown, Gregory-Hansen procedure involves computing the cointegration test statistics for each possible break and taking the minimum test statistics (ADF test) across all possible break points. That is, the break point is unknown and determined by finding the minimum value for the ADF statistic. The Akaike Information criterion (AIC) is used to determine the number of lags of the change in the residual used in computing the ADF statistic and turned out to be zero for all three models. The results of the GregoryHansen Cointegration procedure for all specifications indicate that the null of no cointegration is rejected with an endogenous break year of 1989. The ADF statistics for equations (2a-2c) are -5.082, -5.34836 and -5.15361 respectively and they are statistically significant at 5 percent level. Empirical Results Having found evidence of co integration and having established that private and public saving are I(1), the equations (2a-2c) are estimated using the Dynamic OLS (DOLS) proposed by Stock and Watson (1993) and the FMOLS proposed by Phillips and Hansen (1990).In the estimation of equations (2a-2c) with the Dynamic OLS (DOLS), we used two lead and lag terms. The number of lead and lag-terms are determined by using AIC and SBC criterion. The results obtained from FMOLS and DOLS estimators are presented in Tables 4a-4c. Examination of the Tables indicates that while the FMOLS coefficients of offset (betas) ranges between -0.82 and -0.46, the DOLS coefficients of betas ranges from -0.74 to -0.11 yielding a partial offset. For models (2a) and (2b), coefficient on government savings is statistically significant at 1% level. However, the offset coefficient is insignificant in the model (2c). The long-run offset coefficient estimated by FMOLS (DOLS) is -0.458 (-0.11) but they are both statistically insignificant. However, there was statistically significant (at 5% level) change in the slope coefficient, , after 1989 for DOLS estimates. Thus allowing for the slope change in the regime shift specification in the DOLS case, the long-run coefficient is -0.72 (. The structural break dummy, D, is significant across alternative estimates implying the presence of structural break in the data. Taken together, the results show that a structural break did occur in the long-run relationship between private and public saving in 1989. 10 Journal of Economic and Social Studies Do Private Savings Offset Public Savings in Turkey? Table 4.a. FMOLS and DOLS Estimates for Level Shift Model, 1975-2005 FMOLS 16.129 (1.002)* -0.709 (0.129)* 5.112 (1.268)* Constant GSR D DOLS 15.734 (0.682)* -0.741 (0.0967)* 5.377 (0.891)* Note: *, **, *** indicate significance at 1%, 5% and 10% level of significance respectively. The values in parenthesis are standard errors. Table 4.b. FMOLS and DOLS Estimates for Level Shift with trend Model, 1975-2005 FMOLS 18.263 (1.310)* -0.819 (0.124)* 7.320 (1.503)* -0.193 (0.084)** Constant GSR D TREND DOLS 13.892 (1.393)* -0.577 (0.148)* 4.693 (1.049)* 0.137 (0.088) Note:Seethenote in Table3.a. Table 4.c. FMOLS and DOLS Estimated for Regime Shift Model, 1975-2005 FMOLS 14.571 (2.977)* -0.458 (0.462) 6.627 (3.032)** -0.268 (0.483) Constant GSR D DGSR DOLS 11.685 (2.263)* -0.109 (0.349) 9.355 (2.318)* -0.613 (0.322)** Note:Seethenote in Table3a. Volume 3 Number 2 Fall 2013 11 Hasan GÖCEN / Hüseyin KALYONCU / Muhittin KAPLAN Concluding Comments This study examines the long-run relationship between private and public sector saving ratios using FMOLS and DOLS methodologies. Empirical findings of this study can be summarized as follows: First, there is no long-run relationship between private and public savings unless endogenous structural break in the cointegration relationship is allowed in Turkish case. Secondly, the extent of offset coefficients ranges from -0.82 to -0.11 supporting weak form of Ricardian equivalence. Statistically significant change in the slope coefficient in DOLS case also shows that the substitution (offset) between private and public savings are stronger after 1989. This point is particularly worth mentioning because financial repression in Turkish economy was fully removed at this date. Thirdly, the results of the paper suggest that the effectiveness of fiscal policy implementations by the government has decreased significantly after achieving financial liberalization in 1989.The statistically significant and relatively large coefficient () on regime shift variable can be taken as an evidence for this argument. References Makin, A. J. &Narayan,P. K. (2011), How potent is fiscalpolicy inAustralia?, EconomicPapers, 30, 377-385. Barro, R. (1974), Are government bonds net wealth?,Journal of Political Economy, 81, 1095-1117. Bulir, A. &Swiston,A. (2009),Emergingmarket countriesdon’tbelieve in fiscalstimuli: should web blameRicardo?,CzechJournal of Economicsand Finance, 59, 153-164. De Castro, F. &Fernandez,J. L. (2009), Therelationshipbetweenpublicandprivatesaving in Spain: DoesRicardianEquivalenceHold, Banco De EspanaWorkingPaperSeriesno: 0923. De Mello, L.,KongsrudP. M.,&Price,R. (2004), Savingbehaviourandtheeffectiveness of fiscalpolicy, EconomicsDepartmentWorkingPaper, no. 397. Dickey, D. A. &Fuller,W. A. (1979), Distribution of theestimatorsforautoregressive time serieswith a unitroot, Journal of theAmerican Statistical Association, 74, 427-431. Edwards, S. (1996),Why are Latin American’s savings so low? An international comparative analysis, Journal of Development Economics, 51, 5-44. Elliot, G., Rothenberg,T. J. & Stock,J. H. (1996), Efficient tests for an autoregressive unit root, Econometrica, 64, 813-836. 12 Journal of Economic and Social Studies Do Private Savings Offset Public Savings in Turkey? Engle, R. F. & Granger,C. W. J. (1987), Co-integration and error correction: representation, estimating and testing, Econometrica, 55, 251-276. Friedman, M. (1957), A Theory of the consumption function, Princeton University Press, New Jersey. Gregory, A. W. & Hansen,B. E. (1996), Residual based tests of cointegration in models with regime shifts, Journal of Econometrics, 70, 99-126. Hemming, R., Kell,M. & Mahfouz,S. (2002),The effectiveness of fiscal policy in stimulating economic activity, IMF Working Paper, WP/02/208, IMF. Holmes, M. J. (2006), To what extent are public savings offset by private savings in the OECD, Journal of Economics and Finance, 30, 285-296. Johensen, S. &Juselius, K.(1990), Maximum likelihood estimation and inference on cointegration with applications to the demand for money, Oxford Bulletin of Economics and Statistics, 52, 169210. Keynes, J. M. (1936),The general theory of employment, interest and money, Macmillan, Houndsmills, UK. Kwiatkowski, D. P. C. B. Phillips, Schmidt,P. &Shin,Y. (1992), Testingthenullhypothesis of stationaryagainstthealternative of a unitroot, Journal of Econometrics, 54, 159-178. Leiderman, L.,&Razin,A. (1988), TestingRicardianneutralitywith an intertemporalstochastic model, Journal of Money, CreditandBanking, 20, 1-21. Loayza, N., Schmidt-Hebbel,K., &Serven,L. (2000), What drives private saving across the World?,Review of Economics and Statistics, 82, 165-181. Lopez, J. H., Schmidt-Hebbel, K. &Serven,L. (2000), How effective is fiscal policy in raising national saving?,Review of Economics and Statistics, 82, 226-238. Mandal, A. & Payne,J. E. (2007), The long-run relationship between private and public savings: an empirical note, Journal of Economics and Finance, 31, 99-103. Masson, P.,Bayoumi,T. &Samici,H. (1998), International evidence on the determinants of private saving, World Bank Economic Review, 12, 483-501. Modigliani, F. (1946),Life cycle, individual thrift and the wealth of nations, American Economic Review, 76, 297-313. Ng, S. &Perron,P. (2001), Lag length selection and the construction of unit root tests with good size and power, Econometrica, 69(9), 1519-1554. Phillips, P. C. B. & Hansen,E. E. (1990), Statistical inference in instrumental variable regression with I(1) processes, Review of Economic Studies, 57, 99-125. Phillips P. C. B. &Perron,P. (1988), Testing for a unit root in time series regression, Biometrica, 75, 335-346. Ricciutti, R. (2003), Assessing Ricardian Equivalence, Journal of Economic Surveys, 17, 55-78. Volume 3 Number 2 Fall 2013 13 Hasan GÖCEN / Hüseyin KALYONCU / Muhittin KAPLAN Seater, J. (1993), Ricardian Equivalence, Journal of Economic Literature, 31, 142-90. Seater, J.&Mariano,R. S. (1985), New tests of the life cycleandtaxdiscountinghypotheses, Journal of MonetaryEconomics, 15, 195-215. Stock, J. H. & Watson,M. W. (1993),A simple estimator of cointegrating vectors in higher order integrated systems, Econometrica, 6, 783-820. 14 Journal of Economic and Social Studies Journal of Economic and Social Studies Store Personality: Perceptions Towards Consumer Electronics Chain Stores in Turkey A Case of University Students Keti VENTURA Ege University, Faculty of Economics and Administrative Sciences, Business Administration Ege University Campus, 35100 Bornova İzmir keti.ventura@ege.edu.tr Ipek KAZANCOGLU Ege University Elif USTUNDAGLI Ege University Rezan TATLIDIL Ege University AbstrAct The purpose of this study is to identify, develop and compare the determinants of store personality of the most preferred consumer electronics chain stores, as perceived by young consumers in Turkey. A questionnaire survey including a 22-item store personality scale was conducted among 855 students using a convenience sampling method. Exploratory factor analysis (EFA) and confirmatory factor analysis (CFA) was performed. Findings suggest that greater accuracy of information is needed in the purchasing decision related to high involvement products such as consumer electronics. Also it was found that younger consumers prefer reliable stores that give accurate information, value for money, and provides price-quality fit. This study addresses the neglected area of store personality development and validation for consumer electronics relates through an understanding of young consumers perceptions towards store personality determinants. KEYWOrDs Store personality, Consumer Electronic Chain Stores, Confirmatory Factor Analysis, Turkey ArtIcLE HIstOrY Submitted: 31 July 2012 Resubmitted: 05 November 2012 Resubmitted: 18 December 2012 Accepted: 21 December 2012 JEL codes: M31, M39 Volume 3 Number 2 Fall 2013 15 Ahlem-Selma MESSAI / Fathi JOUINI Introduction In recent years, the growing importance of technology in our daily life has increased enthusiasm for consumer electronics consumption. In particular, young consumers, who are leading the adoption of new technologies, have become more ambitious in their purchases (Accenture, 2012, p. 6). In Turkey, consumer electronics is one of the fastest growing markets, with sales of 3,648 million TL in 2011 (GfK, 2011), with several new entrants establishing themselves the market. However, many of the products and brands sold in consumer electronics retail stores are perceived as being very similar (Yozgat and Deniz, 2008, p. 121). Pursuing sustainable competitive advantage in the current market environment necessitates some degree of differentiation in the products offered by consumer electronics retailers. Therefore, as individual customers tend to attribute humanized characteristics to brands or products, building a particular store personality with appropriate characteristics (Blankson and Crawford, 2012, pp. 311315) can be an effective way to differentiate a particular store from competitors, and position itself through these characteristics, which include product availability, service quality, value for money, and store atmosphere. The concept of “brand as a person”, or in other words “brand personality”, as the set of human characteristics associated with a brand (Aaker, 1997, p. 347) is an important concern for consumers, and in turn, for marketers and academicians (Grohmann, 2003, p. 224; Aaker and Fournier, 1995, p. 392; Wang and Yang, 2008, p. 460; Ambroise, et al., 2005, p. 32). Customers interaction with brands is often similar to their interaction with other people, especially when the brands are attached to such meaningful objects as clothes or cars (Aaker, 1996, p. 142). As individuals tend to associate the brand with their life situations and roles (Ligas and Cotte, 1999, p. 611), interaction with brands as if they were people necessitates defining personality traits that affect the relationship between attributes of people and their behavior in various situations (Hurley, 1998, p. 116), such as purchasing. Furthermore, in order to be useful in a social context, a brand’s meaning should be formed based on three components: its physical make up, functional characteristics, and characterizationi.e., personality (Ligas and Cotte, 1999, p. 612). As human or brand personalities are related in the context of retail stores (Brengman and Willems, 2008, p. 27), consumers are able to identify a particular personality related to a store based on both its tangible (design, environment, etc.) and intangible (attitude of sales personnel, service quality, etc.) components. Despite approximate similarity in price, service, merchandising and marketing activities, some stores are overwhelmingly preferred to others (Martineau, 1958, p. 47). Thus, it could be concluded that there should be other factors that 16 Journal of Economic and Social Studies Predicting Banking Distress in European Countries directly or indirectly affect consumers, and in turn, their preferences. Store personality is considered to be one factor responsible for differentiating and positioning a store, because when a store is humanized, it relates to self-meaning that has congruent personalities to oneself (Chan, et al., 2003, p. 302), this being consistent with the “self-congruity” theory (Sirgy and Samli, 1985; Sirgy, et al., 2000). As Turkish consumer electronics chain stores are all highly competitive and implement similar marketing strategies, some degree of differentiation is necessary within their positioning strategies, especially in regard to younger consumers. In this respect, the purpose of this study is to identify and compare the determinants of store personality of the most preferred consumer electronics chain stores, as perceived by young consumers in Turkey. Additionally, it aims to explore specific store personality determinants, and also to develop a scale in a consumer electronics chain store context. Accordingly, exploratory factor analysis (EFA) and confirmatory factor analysis (CFA) were applied to the most preferred consumer electronics chain stores, in order to validate the scale. The conceptual framework for the empirical research was based on store personality in retailing services. In the first section of the study, the concept of store personality is discussed. Following this, the research design and the results of the study are presented. The study is concluded by outlining the implications and recommendations for the practitioners, the limitations of the study, and opportunities for further research. Store personality has been investigated by relatively few researchers including Martineau (1958), d’Astous and Levesque (2003), Brengman and Willems (2008). As the scales developed in these studies are too general, and based on a specific context, they cannot be appropriately applied to consumer electronics chain stores. Accordingly, the studies of these authors are utilized to develop a new scale to increase understanding of the perceptions of university students’ towards consumer electronics chain stores. The findings of the present study have the potential to contribute to store personality literature by developing and validating a store personality scale for the use of consumer electronics retailers to position themselves in young consumer markets. Literature Review The concept of store personality was introduced by Martineau (1958, p. 47), who defined it as “the way in which the store is defined in the shopper’s mind, partly Volume 3 Number 2 Fall 2013 17 Ahlem-Selma MESSAI / Fathi JOUINI by its functional qualities and partly by an aura of psychological attributes”. Store personality is commonly related to store image, but there are important differences in meaning. Although store image is associated with all the dimensions related to the store, store personality is restricted to those dimensions that correspond to human traits (d’Astous and Levesque, 2003, pp. 455-456). As the concept of human and brand personality are both directly transferable within the context of retail stores (Brengman and Willems, 2008, p. 27; Khan, 2010, pp. 9-10) these traits are relevant to the concept of “brand personality”. Aaker (1997) proposed the commonly accepted brand personality dimensions (sincerity, excitement, competence, sophistication, and ruggedness), and defined the concept, mainly by focusing on its positive attributes. The brand personality concept can help brand strategists enrich their understanding of people’s perceptions and attitudes toward the brand. It contributes to a differentiated brand identity, guiding the communication effort and creating brand equity (Aaker, 1996, p. 150). Customers often interact with brands in the same way they interact with other people (Aaker, 1996, p. 142). When consumers view it as having human characteristics, the brand is said to have a personality. Examples of brands found to have strong personalities include Harley Davidson (ruggedness) and Nike (excitement). Madrigal and Boush (2008) stated that trait inferences can also be made about store personality from the combination of marketing mix elements in which the retailers engage. For the stores themselves, other attributes, such as ambience, design and social components are also of concern (d’Astous and Levesque, 2003, p. 457) as these could directly affect the perception of store personality. Other aspects that affect store personality are store name, store environment, service quality, store personnel, merchandise quality, and carried brand names (Brengman and Willems, 2009, p. 347). A number of studies on store personality were conducted by Martineau (1958), who proposed the main influencing factors could be classified into three main areas: layout and architecture, symbols and colors, and advertising and sales personnel. However, d’Astous and Levesque (2003) stressed that the dimensions used by Martineau (1958) were related with image rather than personality, and they developed a 5 dimensional scale, consisting of “sophistication”, “enthusiasm”, “genuineness”, “solidity” and “pleasantness”, which included a total of 34 sub items. In the same study, they also prepared a reduced scale with the highest factor loadings: “enthusiasm” (welcomeness, enthusiasm, liveliness and dynamism); “sophistication” (chicness, being high class, elegance and style); “genuineness” (honesty, sincerity, reliability and honesty/truth); “solidity” (hardiness, solidity, reputation and prosperity) and “unpleasantness” 18 Journal of Economic and Social Studies Predicting Banking Distress in European Countries (annoyance, irritation, loudness and superficiality). These dimensions were those which had been most commonly accepted and used in previous studies (Brengman and Willems, 2009). In Brengman and Willems’ (2008) study, in which a fashion store was evaluated in the terms of store personality, 5 major classifications were proposed. These were “store atmosphere” related with design, ambience and social factors; “merchandise” related with price, quality, style and assortment; “retailer’s reputation” related with “word-of-mouth”, advertising, communication and social responsibility; “service” related with direct interactivity along with “format and location”. In a store environment, factors as such as attachment, experience and structure are also important. These formations are evaluated through defining mechanisms, triggering memories linked to affect, which is either positive (Orth, Limon and Rose, 2010, p. 1207) or negative. Thus consumers develop a perception about a particular store that will affect shopping decisions. For this reasons, the design and structure of the store, and the qualities of the sales personnel represent important indicators of perceived store personality. Where the environment is formal, expressive, symbolic and has a professionally planned setting that interacts with the product, (Fiore and Ogle, 2000, p. 34) it influences the customer’s perception of store personality (Harrell and Hurt, 1976). As consumers repeatedly see, touch, hear and smell the commodities, as well as experience the environment, they derive value from formal, expressive and symbolic qualities, and this process influences their beliefs about the store (Fiore and Ogle, 2000, p. 34; Darden and Babin, 1994, p. 101). Subsequently, they associate the brand with their lifestyles and roles, thus there is a strong connection with their view of “self ” (Ligas and Cotte, 1999, p. 613). Thus, they use brand personality to define their sense of “self ” and tend to seek brands or stores with congruent personalities (Chan, et al., 2003, p. 304), consistent with the “selfcongruity” theory (Sirgy and Samli, 1985; Sirgy, et al., 2000). However, the way a store is perceived depends on the personality of each customer (Wesley, et al., 2006, pp. 167–168). The impressions in consumers’ minds will result in particular store preferences (Martineau, 1958:55; Brengman and Willems, 2009, p. 352). A good impression can be achieved by building a store personality and finding a suitable “location” in the minds of a group of consumers or a market segment (Keller, 2003, p. 119), a strategy known as “positioning”. Most importantly, brands’ objective personality traits do not exist independent of consumer perception (Zentes, Morschett and Schramm-Klein, 2008, p. 169). The process by which consumers use a product category and attribute information from memory has implications for selecting the appropriate positioning strategy to achieve Volume 3 Number 2 Fall 2013 19 Ahlem-Selma MESSAI / Fathi JOUINI brand association (Punja and Moon, 2002, p. 276). The brand, with its symbolic and human attributes, customizes its marketing mix according to store personality. However the personality that managers aim to implement may be different to that perceived by consumers. If there is a major gap between managers’ implementations and consumers’ perceptions of these, then it will be impossible for managers to communicate with consumers in a satisfactory manner. As store personality is directly related with “the perceptions of consumers”, it becomes important to evaluate store personality in the context of store positioning. Perceptions are mostly formed according to the factors representing the store’s “intangible assets”, which, according to Hooley, et al. (2001), are most difficult for competitors to imitate. Therefore, if a firm can succeed in positioning itself through personality characteristics, it will make an important contribution to its performance in terms of competition and long-term survival. While, the literature contains a number of studies on the determinants of brand personality, there is a need for further studies on store personality and its determinants, as these have received less attention. In this field, Martineau (1958) mostly focused on store image attributes to define store personality, d’Astous and Levesque (2003) utilized Aaker’s (1997) brand personality dimensions to develop a general store personality scale; however, this does not fit all retail stores. One example of the development of a new scale is Brengman and Willems’ (2008), which used qualitative methods to develop a scale for fashion stores based on d’Astous and Levesque (2003). In this study, we draw on this previous research (Martineau, 1958; d’Astous and Levesque, 2003; Brengman and Willems’, 2008) to understand the perceptions of young consumers towards the determinants of store personality. Using survey methodology, a new store personality scale is developed for consumer electronics chain stores. Field Study According to the GfK Retail and Technology 2011 Report, consumer electronics is one of the fastest growing sub-sectors in Turkey, with sales of 3,648 million TL. The consumer electronics sub-sector includes of color televisions, audio appliances, video players, cash registers, audio-video cassettes, television satellite receivers and antennas, electronic scales and electronic calculators. There are various domestic and foreign actors in the Turkish market, notably Teknosa, Vatan Computer, Bimeks, 20 Journal of Economic and Social Studies Predicting Banking Distress in European Countries Darty, MediaMarket, BestBuy, Electroworld and Gold. These highly competitive market players need to be differentiated through their positioning strategies. A consumer electronics product is not only a tool with which the user performs a task, but also a decorative item for the home, or means to express one’s personality and lifestyle (Han, et al., 2001, p. 145). For this reason, the purchase of electronics products entails high involvement, in which consumers seek specific qualities in the products themselves. This very personalized approach makes the consumer electronics sector particularly suitable for store personality research. In this competitive environment, building store personality can be a significant factor in positioning a store. Within this context, the aim of this study is to explore and compare the determinants of store personality, as perceived by young consumers, of the most preferred consumer electronics chain stores in the Turkish market. In March 2011 a questionnaire survey was conducted among 855 students from the Faculty of Economics and Administrative Sciences, Ege University in Izmir, using a convenience sampling method. The questionnaires, which took 10-15 minutes to complete, were distributed in the classes and campus area by three research assistants, who had been informed of the content. This sample was chosen due to the fact that young consumers are more familiar with consumer electronics and often influence family purchasing decisions (Hafstrom, et al., 1992). In addition, the enthusiasm of young people for consumer electronics makes them a potentially highly profitable segment of the market (Accenture, 2012, p. 15). The questionnaire consisted of three sections. The respondents were asked to state their preferred consumer electronics store in the first section. The second section gave 22 statements relating to the determinants of store personality for the consumer electronics context. These statements, generated by the author, were based on the following: Martineau’s (1958) store personality determinants (architecture, colour schemes, advertising, salespeople and others), d’Astous and Levesque’s (2003) determinants for store personality dimensions; and Brengman and Willems, (2008, 2009)’s five major fashion store personality determinants (store atmosphere, merchandise, reputation, service and format). Brengman and Willems, (2009) determined four categories containing different items as follows : (i) “store atmosphere”: interior design, music, layout, window display (Berman and Evans, 2010, pp. 508-509); (ii) “merchandise”: price, quality, product range (Berry, 1969); (iii) “service”: sales people’s attitudes (O’Cass, A. and Grace, D., 2008); (iv) “format”: premium branded products, highincome level target (Brengman and Willems, 2009). A 5-point ‘Likert’ scale ranging from “strongly disagree (1)” to “strongly agree (5)” was utilized. Finally, in the last section, there were some demographic questions related to the respondent’s gender, Volume 3 Number 2 Fall 2013 21 Ahlem-Selma MESSAI / Fathi JOUINI age, income level. The questionnaire was originally prepared in Turkish, and the statements were translated into English for the tables. The findings of the study are presented in two parts. In the first part, EFA was conducted to the determinants (Martineau, 1958; d’Astous and Levesque, 2003; Brengman and Willems, 2008) of store personality in order to identify specific factors of store personality and obtain important insights about the data structure. In the second part of the study, ANOVA is performed to compare the perceptions of respondents to the specific determinants of store personality. Following this, first and second-order CFA was performed for each most preferred store in order to test the validity of scale, that is to assess its fit to the proposed theoretical basis. Findings The sample comprised 53.1 % female and 46.9 % male students. TeknoSa (52.4 %), MediaMarkt (27.9 %) and Vatan Computer (21.1 %) were chosen as the most preferred consumer electronics chain stores, comprising 87.9 % of the sample. The profile of the sample appears in Table 1. Table 1. Profile of the sample Variables Sex Female Male Missing N Valid Percent 379 335 38 53.1 46.9 -- Total 752 100 Variables Monthly Income 650-1379 TL 1380-2109 TL 2210-2839 TL 2840-3569 TL 3570-4299 TL 4300-5029 TL 5030-5759 TL 5760-6489 TL 6490 TL + Missing Total N Valid Percent 236 174 113 49 44 18 12 9 28 69 752 34.6 25.5 16.5 7.2 6.4 2.6 1.8 1.3 4.1 100 Mostly Preferred Electronic Store Age TeknoSA Mediamarkt Vatan Total 18-24 25-31 607 145 80.7 19.3 Total 752 100 22 384 210 158 752 51.0 27.9 21.1 100 Journal of Economic and Social Studies Predicting Banking Distress in European Countries In order to establish the evaluations related to the statements determining store personality, one sample t-test was conducted. As shown Table 2, the findings of one sample t-test about the respondents’ perceptions of store personality determinants revealed that young respondents (a) can easily find the products they are looking for in these stores (t=31.622 df:739 p=0.000), (b) visit these stores for its premium branded products (t=23.362 df:736 p=0.000), (c) get accurate information (t=28.401 df:747 p=0.000) and (d) receive value for their money (t=20.535 df:743 p=0.000). Table 2. One Sample T-Test of the Respondents’ Evaluations Statements Mean I visit this store for the premium branded products. 3.88 I think stores like this sell high-priced products. 3.09 I appreciate the interior design of this store. 3.60 I think this store serves high-income customers. 2.81 I appreciate the window display of this store. 3.32 The music played in this store enables me to take pleasure from 3.31 shopping I appreciate the young and dynamic sales staff of this store 3.15 I find more suitable consumer electronics products for young 3.54 people in this store compared to other stores Only the latest technology products are sold in this store 3.65 The low turnover of salespeople in this store gives me confidence. 3.21 The salespeople in this store are very knowledgeable. 3.56 I can easily find what I am looking for in this store. 3.96 This store determines a price which is appropriate for the quality 3.71 of the product. Information given related to the products is correct in this store. 3.86 I think the advertisements of this store give accurate information. 3.66 I think this store always gives me value for money. 3.70 I do not think that this store makes false claims to increase sales.* 2.36 I do not like the aggressive attitude of the salespeople towards 3.52 customers.* Whenever I want to buy a product from this store, they tell me 3.69 that the product is out of stock * The confusing layout of this store makes it difficult to find what I 3.74 am looking for.* This store’s interior colour irritates me. * 3.77 The constant crowding in this store irritates me. * 3.35 Std. dev. 0.991 1.129 0.982 1.124 1.029 t df Sig. 25.822 2.389 17.865 -4.976 8.852 854 854 854 854 854 0.000 0.017 0.000 0.000 0.000 1.072 4.007 854 0.000 1.029 15.279 854 0.000 0.945 19.963 854 0.000 1.080 1.069 0.961 0.820 5.674 5.628 16.980 34.351 854 854 854 854 0.000 0.000 0.000 0.000 0.899 22.905 854 0.000 0.815 30.762 0.900 21.403 0.895 23.006 0.982 -18.943 854 854 854 841 0.000 0.000 0.000 0.000 1.130 13.310 836 0.000 1.030 19.489 842 0.000 1.009 21.286 838 0.000 0.996 1.108 22.269 835 9.307 846 0.000 0.000 Test value is 3 (neither agrees nor disagrees) *recoded Volume 3 Number 2 Fall 2013 23 Ahlem-Selma MESSAI / Fathi JOUINI Common Factors Factor Interpretation Factor (CumulaLoadtive % of ing variance explained) Statements The confusing layout of this store makes it difficult to find what I am looking for.* This store’s interior colour irritates me. * Pleasantness Whenever I want to buy a product from this store, they tell me that the product is out of stock.* The constant crowding in this store irritates me.* I do not like the aggressive attitude of the salespeople towards customers.* Information given related to the products is accurate in this store. I think this store always gives me value for money. I think the advertisements of this store give accurate Reliability information. The salespeople in this store are very knowledgeable. This store determines a price which is appropriate for the quality of the product. I appreciate the interior design of this store. I appreciate the window display of this store. Welcomeness The music played in this store enables me to take pleasure from shopping. KMO Bartlett Test Cronbach Alpha 0.799 0.791 0.762 21.775 0.80 41.989 0.77 56.032 0.67 0.694 0.640 0.751 0.735 0.727 0.670 0.626 0.812 0.752 0.704 0.834 2940.500; (df: 78; p<0.05) An exploratory factor analyses was perfomed on the total sample in order to identify the specific determinants of the store personality with Varimax rotation. The KaiserMeyer-Olkin (KMO) test and Bartlett’s test of sphericity were computed to assess the appropriateness of factor analyses to the data. The Kaiser-Meyer-Olkin Measure (KMO) of Sampling Adequacy (0.834) also showed that sample adequacy was satisfactory and the Bartlett’s test of sphericity was significant (p<0.000). The factor analysis of the 22 item scale yielded a ‘three principal components’ solution. Each of these three factors had an eigen value greater than one. The statements with factor loadings lower than 0.50 were removed from the study. During the factor analysis, eight variables that were unrelated to any factor were excluded, and a total of three factors were extracted from the remaining 13 items. The analysis produced three store personality factors, and explained 56.3% of the variance. Also, Cronbach’s 24 Journal of Economic and Social Studies Predicting Banking Distress in European Countries alpha coefficients were calculated in order to check the internal consistency and reliability of each factor. Factor loadings ranged from 0.63 to 0.81 and alpha coefficients ranged from 0.67 to 0.80, indicating a high internal consistency and reliability for the factors. The factors were declared as “Pleasantness”, “Reliability” and “Welcomeness” in accordance with their respective factor loadings (Table 3). Table 3. Results of Exploratory Factor Analysis Common Factors Factor Loading Statements The confusing layout of this store makes it difficult to find what I am looking for.* This store’s interior colour irritates me. * Pleasantness Whenever I want to buy a product from this store, they tell me that the product is out of stock.* The constant crowding in this store irritates me.* I do not like the aggressive attitude of the salespeople towards customers.* Information given related to the products is accurate in this store. I think this store always gives me value for money. I think the advertisements of this store give accuReliability rate information. The salespeople in this store are very knowledgeable. This store determines a price which is appropriate for the quality of the product. I appreciate the interior design of this store. I appreciate the window display of this store. Welcomeness The music played in this store enables me to take pleasure from shopping. KMO Bartlett Test Factor Interpretation Cronbach (Cumulative Alpha % of variance explained) 0.799 0.791 0.762 21.775 0.80 41.989 0.77 56.032 0.67 0.694 0.640 0.751 0.735 0.727 0.670 0.626 0.812 0.752 0.704 0.834 2940.500; (df: 78; p<0.05) In order to compare the respondents’ evaluations of “Pleasantness”, “Reliability” and “Welcomness” factors, and explore the differences between the most preferred stores, both a sample t-test and ANOVA were conducted. H1 is generated as follows: Volume 3 Number 2 Fall 2013 25 Ahlem-Selma MESSAI / Fathi JOUINI H1a/b/c/d/e/f/g/h/i/k/l/m/n: All three stores are statistically different according to (a) layout, (b) interior colour, (c) product availability, (d) crowdedness, (e) attitude of salespeople, (f )accuracy of product information, (g) value for money, (h) accuracy of information in ads, (i) knowledge of salespeople, (k) price-quality fit, (l) interior design, (m) window display and (n) in store music. Findings of the sample t-tests (Table 4) revealed that the majority of the mean values for each item were very close among the three stores. Considering the findings, it can be proposed that, although these three are the most preferred ones, they were able to differentiate themselves to a limited extent in relation to store personality determinants. It can be said that, although young people perceive these stores as being similar in all services, Vatan Computer and Teknosa were considered to give better value for money (F=7.847 df: 2/749 p=0.000) compared with MediaMarkt. Vatan Computer sets more reasonable prices according to the quality of their products (F=7.335 df: 2/749 p=0.001) than the other two. Sales personnel in MediaMarkt are perceived to have more knowledge (F=4.662 df: 2/749 p=0.010) than those in Teknosa. Also, the respondents who preferred MediaMarkt and Teknosa reported that the music in these stores gave a more pleasurable shopping experience (F=6.601 df:2/749 p=0.001) than was stated by those who preferred Vatan Computer. Considering these findings, H1g; H1k; H1i and H1n are supported (p<0.05). 26 Journal of Economic and Social Studies Volume 3 Table 4. Comparison of One Sample T-Test of Three Stores MediaMarkt Teknosa Common Factors Number 2 Pleasantness Fall 2013 Mean t** df Mean t** 13.812 386 3.71 9.873 206 3.78 10.093 142 3.74 14.750 385 3.74 10.548 200 3.77 9.972 145 Whenever I want to buy a product from this store, they tell 3.69 me that the product is out of stock.* 14.161 388 3.61 7.932 205 3.69 8.004 144 The constant crowding in this store irritates me.* 5.590 390 3.25 3.038 206 3.49 5.283 145 7.368 987 3.62 8.107 204 3.56 5.600 144 22.667 393 3.81 13.311 209 3.93 12.473 147 Statements Mean The confusing layout of this store makes it difficult to find 3.70 what I am looking for.* This store’s interior color irritates me. * 3.30 I do not like the aggressive attitude of the salespeople to3.41 wards customers.* Information given related to the products is accurate in this 3.83 store. Reliability Welcomeness df df 3.56 12.901 393 3.80 12.442 209 3.84 11.135 147 I think the advertisements of this store give accurate infor3.67 mation. 16.056 393 3.65 9.364 209 3.62 7.941 147 The salespeople in this store are very knowledgeable. 9.607 147 393 3.68 11.022 209 3.63 7.233 This store determines a price which is appropriate for the 3.56 quality of the product. 12.811 393 3.77 11.947 209 3.86 10.594 147 I appreciate the interior design of this store. 3.63 13.560 393 3.54 7.619 209 3.54 6.314 147 I appreciate the window display of this store. 3.35 7.489 393 3.37 5.150 209 3.14 4.369 147 The music played in this store enables me to take pleasure 3.12 from shopping. 2.280 393 3.29 3.909 209 2.87 3.422 147 Test value is 3 (neither agrees nor disagrees) *recoded **p<0.05 3.45 27 Predicting Banking Distress in European Countries I think this store always gives me value for money. t** Vatan Ahlem-Selma MESSAI / Fathi JOUINI EFA is essential in determining the underlying constructs for a set of measured variables. In order to test how well the measured variables represent the number of constructs (Brown, 2006), CFA was conducted with Maximum Likelihood Estimation in LISREL 8.80 (Jöreskog and Sörbom, 1997). CFA verifies the factor structure of a set of observed variables and statistically tests the validity and reliability of a hypothesized factor model (Suhr, 2006, p. 1). The hypotheses tested are as follows: H2: The correlated (first-order) store personality factors explain an important proportion of the variance in the data. H3: The set of store personality determinants comprises the latent variable, overall store personality. First-order CFA was performed to examine the construct more closely and establish construct validity for each of the most preferred consumer electronics chain stores. The estimation of the first-order CFA model revealed that the data defined three factors with regression coefficients varying among the three stores. All the coefficients were significant at the 5 % level (Table 5). Therefore, the hypothesis H2 was supported. Vatan Computer store had the highest correlations between “pleasantness and reliability” (r=0.68), “reliability and welcomeness” (r=0.74), “pleasantness and welcomeness” (r=0.52) respectively. In order to obtain as much understanding of the data as possible, second-order factor analysis was performed (McClain, 1996, p. 131; Correia, et al., 2008, p. 167). In comparison to first-order models with correlated factors, second-order factor models can provide a more interpretable model (Chen, et. al., 2005, p. 472). Second-order CFA was conducted on the three-dimensional model of perceived store personality and compared each store independently, with the same items. Figure 1 presents the hierarchical construct of perceived store personality, consisting of several correlated first-order factors and a single second-order factor. In this model, the latent variables were “perceived store personality”, “pleasantness”, “reliability”, “welcomeness”, and the observed variables are store personality determinants. “Pleasantness” and “reliability” were represented by five items each, and “welcomeness” by three items. 28 Journal of Economic and Social Studies Predicting Banking Distress in European Countries Figure 1. Second-Order Confirmatory Factor Analysis of Perceived Store Personality LAYOUT INTERIOR COLOR PLEASANTNESS PRODUCT AVALIABILITY CROWDED AFFECT ATTITUDE OF SALESPEOPLE ACCURATE PRODUCT INFORMATION VALUE FOR MONEY PERCEIVED STORE PERSONALITYY RELIABILITY ACCURATE INFORMATION IN ADS KNOWLEDGEABLE SALESPEOPLE PRICE-QUALITY FIT INTERIOR DESIGN WELCOMING WINDOW DISPLAY IN STORE MUSIC Volume 3 Number 2 Fall 2013 29 Ahlem-Selma MESSAI / Fathi JOUINI The standardized regression coefficients of the models and associated t-values for the 13 items are shown in Table 5. The output revealed that every standardized regression coefficient and the associated t-values are significant (p ≤ 0.05). The resulting fit indices (GFI, CFI, AGFI, RMSEA, χ2 /df ) for all three store personality models displays satisfactory measure fit, which supported the construct validity of the measures. In order to assess the convergent validity of “pleasantness”, “reliability” and “welcomeness” constructs, construct reliability was computed. Construct reliability values ranged from 0.617 to 0.893, which shows an acceptable convergent validity. Discriminant validity is computed by latent variable correlation matrix (Table 6). As seen from the Table 6, although some of the AVE (Average Variance Exracted) values are at moderate level (AVE<0.5) (Paswan, 2009), the square root values of AVE for all three stores are larger than the correlations, which demonstrates an acceptable discriminant validity. 30 Journal of Economic and Social Studies Volume 3 Table 5. Results of the CFA of Store Personality Scale Among the Most Preferred Consumer Electronics Chain Stores TEKNOSA (N=384) Predictive Parameter Estimates Number 2 Fall 2013 Overall Model Fit: *Significant at p ≤ 0.05 VATAN (N=158) 31 Predicting Banking Distress in European Countries PERCEIVED STORE PERSONALITY → PLEASANTNESS PERCEIVED STORE PERSONALITY → RELIABILITY PERCEIVED STORE PERSONALITY → WELCOMENESS PLEASANTNESS →LAYOUT PLEASANTNESS →INTERIOR COLOR PLEASANTNESS →PRODUCT AVAILABALITY PLEASANTNESS →CROWDEDNESS PLEASANTNESS →ATTITUDE OF SALESPEOPLE RELIABILITY →ACCURACY OF PRODUCT INFORMATION RELIABILITY → VALUE FOR MONEY RELIABILITY → ACCURACY OF INFORMATION IN ADS RELIABILITY → KNOWLEDGABLE SALESPEOPLE RELIABILITY → PRICE-QUALITY FIT WELCOMENESS →INTERIOR DESIGN WELCOMENESS → WINDOW DISPLAY WELCOMENESS → IN STORE MUSIC MEDIAMARKT (N=210) Standardized Standardized Standardized t-value Regression R2 t-value Regression R2 t-value Regression R2 Coefficients Coefficients Coefficients 2.77* 0.37 0.13 5.19* 0.66 0.44 5.53* 0.69 0.48 3.22* 0.83 0.68 4.44* 0.86 0.74 10.04* 0.98 0.97 2.64* 0.50 0.25 4.88* 0.50 0.25 8.22* 0.82 0.56 7.60* 0.76 0.58 7.73* 0.77 0.60 6.33* 0.82 0.67 7.49* 0.74 0.55 7.58 * 0.74 0.55 6.35* 0.83 0.69 7.34* 0.74 0.51 6.59* 0.60 0.36 5.52* 0.58 0.34 6.09* 0.54 0.30 5.57* 0.48 0.23 4.77* 0.45 0.20 0.50 0.36 0.61 0.37 0.46 0.21 6.25* 0.54 0.41 5.08* 0.69 0.48 10.68* 0.76 0.58 6.65* 0.72 0.52 5.07* 0.69 0.48 11.88* 0.85 0.72 6.66* 0.73 0.53 5.14* 0.73 0.53 10.99* 0.78 0.61 0.55 0.30 0.42 0.17 0.73 0.53 5.55* 0.53 0.28 4.30* 0.47 0.22 11.61* 0.83 0.68 0.55 0.30 0.82 0.67 0.74 0.55 3.43* 0.67 0.45 7.30* 0.72 0.52 10.44* 0.83 0.69 2.86* 0.30 0.088 6.34* 0.54 0.30 9.20* 0.70 0.49 GFI=0.93, AGFI=0.90,CFI= 0.94,GFI=0.95, AGFI=0.90,CFI=0.98,GFI=0.90, AGFI=0.87,CFI= 0.90, RMSEA=0.033, χ262=96.34,χ2 /RMSEA=0.031, χ262=73.67, χ2/RMSEA=0.074, χ262=259.78, χ2/ df=1.55 df=1.18 df=4.19 Ahlem-Selma MESSAI / Fathi JOUINI Table 6. The Latent Variable Correlation Matrix: Discriminant Validity Reliability Teknosa (n=384) Pleasantness Welcomeness Reliability MediaMarkt (n=210) Pleasantness Welcomeness Reliability Vatan Computer (n=158) Pleasantness Welcomeness Reliability 0.62 0.30 0.41 0.61 0.57 0.43 0.79 0.68 0.74 Pleasantness Welcoming 0.67 0.18 0.53 0.65 0.33 0.70 0.75 0.52 0.76 Note: Square root of AVE is on the diagonal. The values of some other goodness-of-fit indices are the following: GFITEKNOSA = 0.93, AGFITEKNOSA = 0.90, CFITEKNOSA = 0.94, RMSEATEKNOSA = 0.033, χ2/ dfTEKNOSA=1.55; GFIMEDIAMARKT=0.95, AGFIMEDIAMARKT=0.90, CFIMEDIAMARKT=0.98, RMSEAMEDIAMARKT=0.031, χ2/dfMEDIAMARKT=1.18; GFIVATAN=0.90, AGFIVATAN =0.87, CFIVATAN= 0.90, RMSEAVATAN=0.074, χ2/dfVATAN=4.19. Therefore, the three store personality factors determined perceived store personality, thus H3 is supported. The factor “reliability” had the highest standardized regression coefficients among the three stores. This means that when “reliability” increases by one, the perceived store personality increases by 0.83; 0.86 and 0.98. In all three stores, the most important determinants for the “pleasantness” factor are “layout” (βTeknosa=0.76; βMediaMarkt=0.77; βVatan=0.82) and “interior color” (βTeknosa=0.74; βMediaMarkt=0.74; βVatan=0.83). The reliability factor derives mainly from the “accurate information in ads” (βTeknosa=0.73; βMediaMarkt=0.73) and “value for money” (βTeknosa=0.72; βMediaMarkt=0.69) in Teknosa and MediaMarkt; “value for money” (βVatan=0.85) and “price-quality fit” (βVatan=0.83) in Vatan Computer. The highest coefficients for the “Welcomeness” factor were “window display” (βTeknosa=0.67; βVatan=0.83) in Teknosa and Vatan Computer; “interior design” (βMediaMarkt=0.82) in MediaMarkt. Conclusion and Implications The consumer electronics market is one of the fastest growing in Turkey. In recent years, the entry of global chain stores into the Turkish market and the extension policies of local chain stores have increased competition in this sector. As 34.6 % of 32 Journal of Economic and Social Studies Predicting Banking Distress in European Countries the population in Turkey comprises young people (TSI, 2011, p. 13), the interest in consumer electronics is growing rapidly, thus forcing consumer electronics retailers to be more innovative in the diversification of their marketing and positioning strategies. However, it can be seen that there is insufficient diversity in the retailing mix policies and strategies among the retailers with regard to factors such as goods and services offered, store location, operating procedures, pricing tactics, store atmosphere, customer services and promotional methods. Within this context, the purpose of this study is to identify and compare the determinants of the most preferred consumer electronics chain stores’ personalities, as perceived by a representative sample of Turkish university students. It aims to explore the factors that make up the specific store personality determinants in a consumer electronics chain store context. The reason for this is the critical need for the most popular electronics product chain stores to differentiate and gain competitive advantage, especially for young consumers such as university students, who represent an important section of this market (Accenture, 2012, p. 15). In the literature, there are few studies relating to store personality. This concept was firstly introduced by Martineau (1958), who mainly focused on store image attributes in defining store personality, while d’Astous and Levesque (2003) developed a “general” store personality scale based on Aaker’s (1997) brand personality dimensions. Additionally, Brengman and Willems (2008) developed a scale using retailing mix, aimed specifically at “fashion stores”. As the scales developed in these studies are generalized and context based, they are not appropriate for consumer electronics chain stores. Accordingly, these studies have been utilized as the basis for a new scale specifically designed to gain an understanding of university student perceptions of chain stores in the electronics sector. The findings of the present study contributes to the literature by developing and validating a store personality scale specifically for consumer electronics retailers based on university students’ perceptions of store personality determinants. In the Turkish market, Teknosa, MediaMarkt and Vatan Computer were selected as the most preferred chain stores. In the study, in order to identify specific factors of store personality, EFA was conducted and three factors were identified: “reliability”, “pleasantness” and “welcomeness”. Following this, CFA was performed on these factors and a confirmatory factor analytic model was generated. First-order models with correlated factors were performed to examine the construct more closely, while second-order factor analysis was used to define the most important factors and determine the relative importance of each factor to the overall, for each Volume 3 Number 2 Fall 2013 33 Ahlem-Selma MESSAI / Fathi JOUINI chain store (Correia, et al., 2008, p. 173). The results indicate that “reliability” is perceived to be the most important factor in all three stores. In contrast, factors that are perceived to be least important by university students were “pleasantness” for Vatan Computer and Teknosa, “welcomeness” for MediaMarkt. One of the most important items perceived was “Value for money”, which explains “reliability” factor in all three store. This is due to the fact that purchasing electronic products requires high involvement which means that consumers put a greater emphasis on finding information about brands and products when making purchases. Moreover, in Teknosa and MediaMarkt, “accurate information in the advertisements” of the stores and in Vatan Computer “price-quality fit” best explains the “reliability” factor. The second store personality factor, “pleasantness” is found to be best explained by “layout” and “interior design” in all three stores. “Window display” is very much related with the “welcomeness’ factor in Teknosa and Vatan, while “interior design” is the most important item in explaining this factor in MediaMarkt. These findings emphasize the need for accurate information in the purchasing decision making process for high-involvement products, such as consumer electronics (Ahmed, et al., 2004). Young consumers prefer a reliable store that gives accurate information, value for money and has price-quality fit. In addition, the store atmosphere and merchandise are perceived as the most important determinants of store personality by university students. This may be due to the need for young consumers e.g. university students, to be able to touch, hear and even smell products, as well as see them, which contributes to a pleasant shopping experience and a feeling of being welcome in the store. Considering these findings, in order to increase the reliability of such stores it may be appropriate to recommend that managers ensure that sales personnel are provided with information relative to the products, as well as the price. This is considered to promote sales. In addition, stores should avoid misleading advertising and sales promotions as this can cause store avoidance. In order to prevent this problem and assist in monitoring students’ buying habits, loyalty cards can be given to contracted universities. Also, detailed explanations regarding the usage of products can be given on their web sites in the form of videos. Another vital measure is to improve after-sales service, such as maintenance, installation, and repair and even offer replacement with another product where necessary. Considering of the importance of the university market, it would be very useful for chain stores to promote their most popular products at university festivals, organize a variety of events and give special incentives to students, in order to capture these potential long term customers of the future. 34 Journal of Economic and Social Studies Predicting Banking Distress in European Countries As argued by Brengman and Willems (2009), it may be appropriate to propose that the managers of store chains should reduce the perception of overcrowding by rearranging the layout to facilitate shopping and in store traffic flow, to create a more pleasant store atmosphere. Managers may be encouraged to develop strategies for relieving congestion in crowded areas such as entrance, halls, cash register locations, customer services and shopping areas. This study shows the potential for a store-personality scale developed for consumer electronics chain stores in helping retailers understand the perception of young customers. This understanding can enable stores to reposition themselves in a competitive market by implementing an integrated communication strategy. The main academic contribution of this study is that it can be seen as a preliminary effort to identify the most distinctive personality factors and determinants of consumer electronics chain stores perceived by university students in Turkey. As consumer electronics stores and their personality characteristics have received a limited amount of research attention in the marketing literature, this study is expected to be useful for potential domestic and foreign consumer electronics retailers investing in this sector. Limitations and Future Research This study has a number of limitations. In sampling design in particular, due to budget limitations, the use of convenience sampling method in only one metropolitan city, Izmir, and one university may not represent the general perceptions of Turkish university students. However, despite its limitations, this study can be seen as a preliminary effort in analyzing the perceptions of Turkish university students, a highly profitable segment of the market (Accenture, 2012, p. 15). In his study, Martineau (1958) mostly focused on store image attributes to define store personality, while d’Astous and Levesque (2003) utilized Aaker’s (1997) brand personality dimensions to develop a “general” store personality scale. Additionally, Brengman and Willems’ (2008) developed a scale only for “fashion stores”, which is therefore unsuited to consumer electronics. The scale developed in this study can only be applied to consumer electronics chain stores, because it includes specific determinants. Additionally, this scale can be used to measure the effect of store personality determinants on store loyalty, store image and alternative methodologies, Volume 3 Number 2 Fall 2013 35 Ahlem-Selma MESSAI / Fathi JOUINI including the usage of both qualitative and quantitative techniques. This process would be able to provide a deeper understanding of consumer perceptions and attitudes. Further studies can be made into different types of retailers, as well as customer segments from other countries in order to explore the effects of cultural differences. In addition, further research conducted on a wider sample size involving other universities in Izmir and other metropolitan cities would allow a stronger representative view. References Aaker, J. & Fournier, S. (1995). A brand as a character, a partner and a person: three perspectives on the question of brand personality. Advances in Consumer Research, 22, 391–395. Aaker, D. A. (1996). Building Strong Brands. New York: The Free Press. Aaker, J. L. (1997). Dimensions of brand personality. Journal of Marketing Research, 34(3), 347–356. Accenture (2012), The 2012 Accenture Consumer Electronics Products and Services Usage Report. http://www.accenture.com/SiteCollectionDocuments/PDF/Accenture_EHT_Research_2012_ Consumer_Technology_Report.pdf , Access Date: 22.07.2012 Ahmed, Z. U., Johnson, J. P., Yang, X., Fatt, C. K., Teng, H. S., & Boon, L. C. (2004). Does country of origin matter for low-involvement products?. International Marketing Review, 21(1), 102-120 doi: 10.1108/02651330410522925 Ambroise, L., Ferrandi, J. M., Merunka, D., Vallette-Florence, P. & De Barnier, V. (2005). How well does brand personality predict brand choice? A measurement scale and analysis using binary regression models. Asia Pacific Advances in Consumer Research, 6, 30-38. Berman, B. & Evans, J. R. (2010). Retail Management: A Strategic Approach, (11th ed.). Prentice Hall. Berry, L.L. (1969). The components of department store image: a theoretical and empirical analysis. Journal of Retailing, 45(1), 3-20. Blankson, C. & Crawford, J. C. (2012). Impact of positioning strategies on service firm performance. Journal of Business Research, 65, 311 – 316. doi: 10.1016/j.jbusres.2011.03.013 Brengman, M. & Willems, K. (2008). Determining fashion store personality dimensions: an exploratory study based on repertory grid data and grounded theory. European Institute of Retailing and Services Studies (EIRASS), 15. Brengman, M. & Willems, K. (2009). Determinants of fashion store personality: a consumer perspective. Journal of Product & Brand Management, 18(5), 346–355. doi: 10.1108/10610420910981828 Brown, T. A. (2006). Confirmatory Factor Analysis for Applied Research, New York: Guilford. Chan, P. Y. L., Saunders, J., Taylor, G. & Souchon, A. (2003). Brand personality perception regional or country specific?. European Advances in Consumer Research, 6, 300–307. 36 Journal of Economic and Social Studies Predicting Banking Distress in European Countries Chen, F. F., Sousa K. H., & West S. G. (2005). Testing Measurement Invariance of Second-Order Factor Models. Structural Equation Modeling, 12(3), 471–492, Lawrence Erlbaum Associates, Inc. doi: 10.1207/s15328007sem1203_7 Correia, A., Moital, M., Da Costa, C. F. & Peres, R. (2008). The determinants of gastronomic tourists’ satisfaction:a second-order factor analysis. Journal of Foodservice, 19, 164–176. doi: 10.1111/j.17454506.2008.00097.x Darden, W. R. & Babin, B. J. (1994). Exploring the concept of affective quality: expanding the concept of retail personality. Journal of Business Research, 29, 101–109. doi: 10.1016/0148-2963(94)90014-0 d’Astous, A. & Levesque, M. (2003). A scale for measuring store personality, Psychology & Marketing, 20(5), 455-469. doi: 10.1002/mar.10081 Fiore, A. M. & Paff Ogle, J. (2000). Facilitating students’ integration of textiles and clothing subject matter part one: dimensions of a model and a taxonomy. Clothing and Textiles Research Journal, 18(1), 31–45. doi: 10.1177/0887302X0001800104 GfK Retail and Technology (2011). Press Release: http://www.gfkrt.com/imperia/md/content/ gfkrtturkey/2011-q4_gfk_temax_press_release_turkey_tr.pdf, Access Date: 04.06.2012. Grohmann, B. (2003). Masculine and feminine dimensions of brand personality: initial steps in scale development and validation. European Advances in Consumer Research, 6, 224. Hafstrom, J. L., Chae, J. S. & Chung, Y. S. (1992). Consumer decision-making styles: comparison between United States and Korean young consumers. Journal of Consumer Affairs, 26. doi: 10.1111/ j.1745-6606.1992.tb00020.x Han, S. H., Yun, M. H., Kwahk, J. & Hong, S. W. (2001). Usability of consumer electronic products. International Journal of Industrial Ergonomics, 28, 143–151. doi: 10.1016/S0169-8141(01)00025-7 Harrell, G. D. & Hurt, M. D. (1976). Buyer behavior under conditions of crowding: an initial framework. Advances in Consumer Research, 3, 36–39. Hooley, G., Greenley, G., Fahy, J. & Cadogan, J. (2001). Market-focused resources, competitive positioning and firm performance. Journal of Marketing Management, 17(5-6), 503 – 520. doi: 10.1362/026725701323366908 Hurley, R. F. (1998). Customer service behavior in retail settings: a study of the effect of service provider personality. Journal of the Academy of Marketing Science, 26(2), 115–127. doi: 10.1177/0092070398262003 Jöreskog, K. G., & Sörbom, D. (1997). LISREL 8: User’s reference guide (5th ed.). Chicago: Scientific Software. Keller, K. L. (2003). Strategic Brand Management: Building, Measuring and Managing Brand Equity (2nd ed.). USA: Pearson Education. Khan, B. M. (2010). Brand personality and consumer congruity: Implication for advertising strategy. Journal of Brand Management, 7(1-2), 7 -24. Ligas, M. & Cotte, J. (1999). The process of negotiating brand meaning: a symbolic interactionist perspective. Advances in Consumer Research, Brand personality perception regional or country specific?. European Advances in Consumer Research, 26, 609–614. Madrigal, R. & Boush, D.M. (2008). Social responsibility as a unique dimension of brand personality and Volume 3 Number 2 Fall 2013 37 Ahlem-Selma MESSAI / Fathi JOUINI the consumers’ willingness to reward. Psychology and Marketing, 25(6), 521-37. doi: 10.1002/mar.20224 Martineau, P. (1958). The personality of the retail store. Harvard Business Review, 36(1), 47–55. McClain, A. J. (1996). Hierarchical analytic methods that yield different perspectives on dynamics: aids to interpretation. In B. Thompson (Ed.), Advances in social science methodology, Greenwich, CT: JAI Press, 4, 229-240. O’Cass, A. & Grace, D. (2008). Understanding the role of retail store service in light of self-image – store image congruence. Psychology & Marketing, 25(6), 521-537. doi: 10.1002/mar.20223 Orth, U. R., Limon, Y. & Rose, G. (2010). Store-evoked affect, personalities, and consumer emotional attachments to brands. Journal of Business Research, 63, 1202 – 1208. doi: 10.1016/j. jbusres.2009.10.018 Paswan, Audhesh (2009), “Confirmatory Factor Analysis and Structural Equations Modeling: An Introduction” Seminer Notes, http://www.cob.unt.edu/slides/paswan/BUSI6280/CFASEM%20-%20Intro-May%2018%202009.ppt, Access Date: 5.10.2012 Punja, G. & Moon, J. (2002). Positioning options for achieving brand association a psychological categorization framework. Journal of Business Research, 55, 275– 283. doi: 10.1016/S01482963(00)00165-X Sirgy, M.J. & Samli, A.C. (1985). A path analytic model of store loyalty involving self-concept, store image, socioeconomic status, and geographic loyalty. The Journal of the Academy of Marketing Science, 13(3), 265-291. doi: 10.1007/BF02729950 Sirgy, M.J., Grewal, D. & Mangleburg, T. (2000). Retail environment, self-congruity, and retail patronage: an integrative model and a research agenda. Journal of Business Research, 49(2), 127-138. doi: 10.1016/S0148-2963(99)00009-0 Suhr D. (2006). Exploratory or Confirmatory Factor Analysis?. SAS Usuers Group International Conference. 31, SanFrancisco: SAS User’s Group:Paper #200. Turkish Statistical Institute (TSI) (2011), Turkey in Statistics 2011, Pub. No. 3592, Ankara. Wang, X. & Yang, Z. (2008). Does country-of-origin matter in the relationship between brand personality and purchase intention in emerging economies?. International Marketing Review, 25(4), 458–474. doi: 10.1108/02651330810887495 Wesley, S.C., Fowler, D.C., Vazquez, M.E. (2006) “Retail personality and the Hispanic consumer”, Managing Service Quality, 16 (2), 167-84 Yozgat, U. & Deniz, R. B. (2008). Specialty retailers in consumer electronics technology markets as a business model. Journal of Global Strategic Management, 3, 114-116. Zentes, J., Morschett, D & Schramm-Klein, H. (2008). Brand personality of retailers – an analysis of its applicability and its effect on store loyalty. The International Review of Retail, Distribution and Consumer Research, 1. doi: 10.1080/09593960701868282 38 Journal of Economic and Social Studies Journal of Economic and Social Studies The Role of Twin Deficit Problem in Sustainable Growth: An Econometric Analysis for Turkey Halil UCAL Adnan Menderes University Faculty of Economics and Administrative Sciences Nazilli, Aydin, Turkey hucal@adu.edu.tr Mehmet BOLUKBAS Adnan Menderes University Faculty of Economics and Administrative Sciences Nazilli, Aydin, Turkey mbolukbas@adu.edu.tr AbstrAct In economics literature the relationship between budget deficit and current account deficit is known as twin deficit hypothesis. The Keynesian Approach accepts a relationship between two deficits. In contrast to this, Ricardian Equivalence Hypothesis defends there is no relationship between these two deficits. Twin deficit has become the subject of several studies to test which of these hypotheses are reliable but no consensus has been achieved. Some studies found a relationship from budget deficit to current account deficit but some of them had the opposite result. Especially after 1980 it is known that many developed and developing countries encountered with this twin deficit problem. Turkey also has the problem of twin deficit. Therefore, it is important to find whether there is causality between them and the direction of this causality. In this study the relationship between budget deficit and current account deficit is examined by using Johansen Cointegration Analysis. This study is based on period 1996:Q1-2011:Q4. According to results of cointegration; variable coefficients are statistically significant and consistent with what we expected in hypotheses. Current account deficit (CAD) has a significant negative effect on budget deficit (BD). When there is a 1% increase in CAD, BD decreases 0,12%. This finding is consistent with economic theory because according to Keynesian Approach two deficits have relationship with each other. However, in contrast to this approach, the direction is from CAD to BD and also coefficient is negative. KEYWOrDs Budget Deficit, Current Account Deficit, Sustainable Growth, Econometric Modeling, Turkey ArtIcLE HIstOrY Submitted: 5 October 2012 Resubmitted: 18 October 201 Accepted: 09 January 2013 JEL codes: H62, F32, O4 Volume 3 Number 2 Fall 2013 39 Halil UCAL / Mehmet BOLUKBAS Introduction The Twin Deficit is referred to a situation where an economy is running both Current Account Deficit (CAD) and Budget Deficit (BD). According to Ricardian Equivalence CAD and BD are not correlated. Budget deficit is a result of tax cut which reduces public revenues and public saving (Alkswani, 2000). Decrease in public savings will be compensated by an increase in private saving. Therefore national saving will not be affected and the budget deficit will have no effect on the current trade deficit (Alkswani, 2000). On the other hand, according to Keynesian proposition the two deficits are linked and the direction is from BD to CAD. If there is a budget deficit, government has to borrow more and as a result the interest rates will rise. The rise of interest rates leads inflow of money from abroad and then the local currency appreciates. The appreciation of currency results with increase in import and decrease in export. As a result, trade deficit increase and current account balance distorted. The twin deficit has started to become a problem with the beginning of the 1980’s in USA. Increase in military expenditures and decrease in income tax raised budget deficit. The increase in budget deficit caused increase in debt of US to the rest of the world and therefore caused distortion in balance of payments. After the global crisis in 2008, it is seen that not only in USA also in other developed and developing countries have the same macroeconomic problems. Especially in developed countries such as European countries faced with serious problems in their economies. Growth in developing economies such as China and India has become a danger for developed countries. Foreign trade worsened and caused decrease in balance of payments in western countries. Also high borrowing of governments deepened crisis in European countries. In recent years, CAD has become the most discussed issue for Turkey’s Economy. According to Peker (2009) macroeconomic policies such as inflation targeting generally cause appreciation of local currency and thus stimulate import. Turkey has lack of savings like other developing countries. Because of this, growth in economy depends on import oriented production and consumption. Although Turkish Economy performs high level of growth, the trade balance is worsening. In the last decade Turkish foreign trade has showed a large increase. However, increase in trade volume has become more than increase in export. Also increase of gas and oil prices in the world has increased Turkey’s energy expenditure. Therefore trade balance and also current account balance worsened. 40 Journal of Economic and Social Studies The Role of Twin Deficit Problem in Sustainable Growth: An Econometric Analysis for Turkey After the 1999 earthquake and 2001 crisis, fiscal policies tightened and to increase the revenues new tax policies have been implied. Especially new taxes such as Private Consumption Tax (PCT) on import oriented goods have been implied to help improving budget balance. Especially PCT revenues on petroleum products, almost totally import oriented, helped to finance the budget deficit. Tax burden is 20% in 2011 which was 13% in 1998. Also share of value-added taxes (VAT) from import in total value-added tax revenues raised to 17% which was 11% in 1999. The gap between domestic VAT and VAT from import is closed as of 2011. The growth in economy and tightened fiscal policies reduced the vulnerability to crisis of Turkish economy. However, good performance of budget balance had no positive effect on balance of payments. Export-import ratio was under 70% except 2001 and 2009. After 2001 Trade deficit increased continuously and in period 19972004 CAD/GDP ratio was 1,1% but in period 2005-2010 the ratio raised to 5,1%. Graph 1 shows the relation of BD and CAD in the last 15 years. Figure 1. Budget Deficit and Current Account Deficit in Turkey, 1996-2011(millions of $) Source: Electronic Data Delivery System (EDDS), CBRT, 2012. Volume 3 Number 2 Fall 2013 41 Halil UCAL / Mehmet BOLUKBAS As seen in the figure, especially after the 2001, Current Account Balance continuously worsens. However, in this period Turkish economy experienced high growth rates. With the global financial crises in 2009 CAD decreases sharply. After that it increases sharply too. In this period BD moves in the opposite direction. According to graph, BD did not rise over 30 billion dollars except 2009. Shrink in economy and decrease in foreign trade decreased budget revenues in 2009. However in the last decade BD/ GDP ratio decreased continuously and become -1.4% as of 2011. This ratio is less than 3% which is the reference value in Maastricht Criteria. As of 2011 most of the EU member countries do not meet this criterion. In this paper it is discussed whether CAD and BD has a correlation with each other and if there is, in which direction is this relationship. According to hypothesis of this paper there is a correlation between these two deficits and it is negatively correlated. Because the increase in trade deficit increases the budget tax revenues and this help to decrease budget deficit. Literature Review In economic literature, there are many empirical researches that focused on twin deficit problem. In 1980’s United States faced with increase in federal trade deficit (TD) and federal budget deficit together. After that the relationship between trade deficit and budget deficit has become an important subject for researchers. Darrat (1988) tried to find the linkage between TD and BD by using data period 1960:I to 1984:IV for United States. He found the evidence of causality from budget deficit to trade deficit and stronger causality from trade to budget deficit by using multivariate Granger Causality Test. In the other study for the U.S., Enders and Lee (1990) searched the relationship between BD and CAD in period 1947 to 1987 by using VAR analysis. They found that government spending innovation generates a persistent current account deficit. Also, Abelln (1990) examined the relationship between federal budget deficits and merchandise trade deficit for U.S. He used multivariate time series within autoregressive model for period 1979:02 through 1985:02. He found that indirectly budget deficits affect trade deficits. 42 Journal of Economic and Social Studies The Role of Twin Deficit Problem in Sustainable Growth: An Econometric Analysis for Turkey Another work on US budget deficit and current account deficit linkage is study of Bahmani-Oskooee (1989). He examined the linkage in period 1973-1985 and concluded that the budget deficit contribute to current account deficit. Not only U.S. but also other countries are faced with this twin deficit issue. Therefore, there are studies about other countries too. Islam (1998) examined the casual relationship between budget deficits and trade deficits of Brazil from 1973:1Q through 1991:Q4. The results suggested that there is a bilateral causality between them. Vamvoukas (1999) used annual data in period between 1948 and 1994 for Greece. He used error correction model for the analysis and found that budget deficit has short and long run positive and significant causal effects on trade deficit. Alkswani (2000) studied on twin deficit problem in petroleum economy by using Saudi Arabia annual data from 1970 to 1999. In his empirical analysis he used ECM, Johansen cointegration and Granger bivariate causality tests and as a result found that trade deficit causes budget deficit. Puah et al. (2006) analyzed the twin deficits debate in Malaysia and Johansen-Juselius co-integration test results show that budget deficit and current account deficit do not contain common stohastic trend in long run. In addition unidirectional causality running from current account to budgetary variable where the deterioration in current account deficit could worsen the budgetary position in the case of Malaysia. Merza et al. (2012) examined twin deficits hypothesis for Kuwait for the quarterly period (1993:4-2010:4). To analyze the relationship between variables they applied the VAR model and tested for existence and the direction of causality and the results show that the direction of causality current account to budget balance that is an increase in current account causes a decrease in the government budget surplus or an increase in budget deficit. Also in Turkey, there are many studies focusing on Turkey’s twin deficit problem. Some of these studies are Ay, et al.(2004), Aksu and Başar (2005), Utkulu (2003), Yücel and Ata (2003), Kutlar and Şimşek (2001), Zengin (2000), Sever and Demir (2007), Akbostancı and Tunç (2002). Some of them used current account deficit variable and some used trade deficit variable in their empirical studies. Most of them used quarterly data for Turkey. Volume 3 Number 2 Fall 2013 43 Halil UCAL / Mehmet BOLUKBAS Akbostancı and Tunç (2002) used quarterly variables between 1987:Q1 and 2001:Q3. They used Budget balance and trade balance as a percentage of GDP. By using ECM and Cointegration analysis the empirical results show that there is a long run relationship between two and in the short run worsening of budget balance worsens trade balance. Sever and Demir (2007) used quarterly data between the years 1987 and 2006 to examine the relationship of budget deficit with current account deficit. By using stationarity test, granger causality test and VAR analysis they found that budget deficit influence current account deficit indirectly. Kutlar and Şimşek (2001) used budget deficit and trade balance seasonally adjusted data in log form in period 1984(4) through 2000(2). In the analysis stationarity test, granger causality test, misspecification test, cointegration test and ECM used and found that there is a positive relationship between two variables and trade deficit increase budget deficit. Zengin (2000) used seasonally adjusted quarterly data for period 1987:I through 1998:I. The main variables are trade deficit and consolidated budget deficit as ratios to GNP. In the analysis VAR, Variance decompositions and impulse response function used. The result of the empirical analysis is that budget deficit influence trade balance. Yücel and Ata (2003) used yearly data from 1975 to 2002. The variables are current account deficit and budget deficit both in log form. The result of the empirical analysis is that there is a cointegration between CA and BD and there is a long run positive relationship. Granger causality test results say that causality is from BD to CA in lag(1) and causality is from CA to BD in lag (3,4 and 7). Utkulu (2003) used budget deficit and trade deficit variables as yearly data in period between 1950 and 2000. By using cointegration analysis and ECM, he found that there is a two sided long run causality between budget and trade deficits. Ay et al.(2004) used monthly data between 1992 and 2003 for the empirical analysis to find the linkage between BD and CAD. The variables used in the empirical analysis were in percentage of GDP. They used Granger Causality test and regression analysis. According to the empirical analysis there is reciprocal relationship between two variables. According to two regression analysis the coefficients are positive. 44 Journal of Economic and Social Studies The Role of Twin Deficit Problem in Sustainable Growth: An Econometric Analysis for Turkey Model, Method and Data Set In this section, a multivariate model has established to investigate twin deficit problem in Turkey. (1) where BD and CAD are budget deficit and current account deficit respectively. Budget deficit (BD) is generally defined as an amount by which some measure of government expenditure and some measure of government revenue. BD is dependent variable, whereas, current account is independent variable in this model. And current account deficit (CAD) Current account deficit includes foreign trade in goods, services and transfers. Current account occurs when a country’s total import of goods, services and transfers is greater than total export of goods services and transfers. Many studies in the literature use BD as an independent variable. But in this study BD is used as dependent variable unlike other studies. This paper adopted the method of co-integration first found by Engle-Granger (1987), developed by Johansen (1988) and applied by Johansen and Juselius (1990). This method depends on direct investigation of co-integration in the vector autoregressive (VAR) representation and produces maximum likelihood estimators of the unconstrained co-integration vector, but it allows one to explicitly test for number of co-integration vectors. Johansen’s methodology takes its starting points in the vector auto regression (VAR) of order p given by; (2) Where yt is a k vector of non-stationary variables I(1), xt is a d vector of deterministic variable; and Ɛt indicates an innovation vector. This VAR can be written as; Volume 3 Number 2 Fall 2013 45 Halil UCAL / Mehmet BOLUKBAS (3) where (4) Where cointegration hypothesis defined as a reduced rank of the matrix π is stated in the form of π = αβ. α and β represent the two matrix which have (kxr)-dimensional and r rank. r is the number of co-integration (rank), β is a co-integration vector showing long-term effects of variables in the equilibrium relations and α indicates speed of adjustment in error correction model. Accordingly an matrix π is estimated from an unrestricted VAR in Johansen method and tested that specified conditions with reduced rank of π rejected or not. And determined by the help of Johansen method’s test statistics (λtrace and λmak) how many rank of the matrix π has.In this context, the data set of the variables used to determine the twin deficit problem in Turkey belong to1996:Q1-2011:Q4 period. All data were taken from Electronic Data Delivery System (EDDS) published by the Central Bank of the Republic of Turkey (CBRT). And Econometric Views (Eviews 5.1) software program was used for all tests and estimates. Empirical Results and Discussion Before constructing the Johansen method, it is important to make some process and pre-tests. Firstly the independent variables were transformed into logarithmic form and variables are understood to have seasonal effect deseasonalized by using moving average method. Then univariate time series of variables are checked by using Augmented Dickey Fuller (ADF) (1979) unit root test. ADF unit root test results can be seen in Table 1. Variables were initially tested with first-level values and then tested with the levels of receipt of the first differences. Accordingly determined that all variables are integrated in the same order I(1). Therefore the necessary precondition for co-integration is provided. 46 Journal of Economic and Social Studies The Role of Twin Deficit Problem in Sustainable Growth: An Econometric Analysis for Turkey Table 1. ADF Unit Root Test Variables BD CAD ΔBD ΔCAD Critical Values ADF Test -2.318258 [3] -3.353061 [1] -9.694507 [2] -4.617754 [6] 1% -4.1118 -4.1104 -3.5440 -3.5526 Note: Trend and intercept term is used as test type for BD and CAD variables, but only intercept term is used for the first differences of variables (Δ). The values in square brackets indicates appropriate length of delay according to AIC. It is necessary to determine an appropriate number of delay to apply Johansen method. There are many measurements in the literature to determine the length of delay; Akaike Info Criterion, Schwarz Info Criterion, Hannan-Quin Criterion and Recent Forecast Error Criterion are the most commonly used (Johansen, 1995; Enders, 1995). But these criterions are not enough on their own. Also there should not be econometric problems in the length of delay selected with info criterions. According to this, in this model the length of delay is determined as two. In this context the model presented in Table 2 shows forecasting of diagnostic test is successful. Table 2. Diagnostic Test Results White Heteroskedasticity Chi-sq 14.897 Normality Test Jarque-Bera 0.203535 0.633672 Df 18 Prop 0.669 Df 2 2 Prop 0.9032 0.7285 After checking univariate of all time series variables the relation between BD and CAD variables can be tested by co-integration test. The purpose of the co-integration test is to determine whether a group of non-stationary series is co-integrated or not. According to Table 1, all variables are I(1), that means co-integration relation between BD and CAD can be investigated by using Johansen Co-integration Method.The results of λtrace and λmak statistics are presented in Table 3. λtrace and λmak statistics helps to find existence of co-integration and number of vectors. Volume 3 Number 2 Fall 2013 47 Halil UCAL / Mehmet BOLUKBAS According to the statistics; the null hypothesis (there is no co-integration relation between variables), is rejected against to alternative hypothesis (there is at least one co-integration relationship between the variables). In this case, there has to be at least one co-integration relationship at 5 % critical value. Table 3. Co-integration Test Null Hypothesis (H0) Alternative (H1) Hypothesisis Trace and Mak Sta5% Critical Value tistics Eigenvalue λTrace λTrace statistic r=0 r>0 0.309956 23.57399 15.49471 r≤1 r>1 0.009183 0.571998 3.841466 λMak λMak statistic r=0 r=1 0.309956 23.00199 14.26460 r=1 r=2 0.009183 0.571998 3.841466 The co-integration equation is presented in Table 4. According to results of cointegration; variable coefficients are statistically significant and consistent with what we expected in hypotheses. CAD has a significant negative effect on BD. When there is a 1% increase in CAD, BD decreases 0,12%. This finding is consistent with economic theory because according to Keynesian Approach two deficits have relationship with each other. However, in contrast to this approach, the direction is from CAD to BD and also coefficient is negative. Table 4. Co-integration Equation BD CAD Normalized Co-integration coefficient (β′) 1.000 0.122535 (0.08580) Adaptation rates coefficient (α) -0.000427 (7.68E-05) -5.23E-05 (0.00019) Co-integration Equation BD= 5001.857 - 0.122535CAD 48 Journal of Economic and Social Studies The Role of Twin Deficit Problem in Sustainable Growth: An Econometric Analysis for Turkey If there is a co-integration relationship among non-stationary variables, there has to be an error correction representation (Engle and Granger, 1987) which illustrates the dynamic convergence of the system to the long-run equilibrium. A precondition for the existence of co-integration is that all the variables are integrated of the same order. If this is fulfilled, then the residuals from the long-run estimates can be used as the error correction term (ECT) to explain the short run dynamic.The error correction term in short run indicates that when the deviations in the short run will be adjusted in the next period (Cholifihani, 2008). Error correction model (vector error correction: VEC) was established in order to investigate the short-run dynamics of variables acting together in the long-run and the results are presented in Table 5. As seen in Table 5; coefficient of error correction term (ECt-1) is statistically significant and negative. If the error correction term is negative, that means deviations in the short-run will be eliminated and series converges to the long-run equilibrium value again among the series moving together in the long-run. Namely error correction term is good working. According to the result approximately 87 % of deviations from the long-run equilibrium value eliminate in each period. Table 5. Error Correction Model Estimation Results DBAt= β0 + β1DCAt-1 + αECt-1+ ut Variables DBAt-1 DCAt-1 ECt-1 Invariable term Coefficient -0.049275 -0.247044 -0.874470 -138.1142 R 2 = 0.44 R2 = 0.46 t-statistic -0.33937 -1.45152 -4.69637 -0.31490 F = 17.12 Conclusion In this paper we tested whether there is a relationship between BD and CAD in Turkey with the framework of growth. In the last decade, Turkey’s economy performed well. After the 2001 crisis new economic policies strengthened the economy against Volume 3 Number 2 Fall 2013 49 Halil UCAL / Mehmet BOLUKBAS crises. With the help of tight fiscal policies, government did not compromise on the budget. However increase in consumption, appreciated currency, lack of savings and rise in price of energy products caused an increase in trade deficit. As a result current account deficit rose. According to empirical resultsthere is a significant negative correlation between BD - CAD and the direction is from CAD to BD. When there is a 1% increase in CAD, BD decreases 0,12%. Many studies on Turkey do not cover last decade’s data. But in this study we reflect the effects of structural changes in Turkish Economy after the period 2001 in terms of BD and CAD. In this regard, empirical results of this study are differentiated from others. Because many studies in literature show that incease in CAD results from incease in BD. Unlike studies in literature, the results of this paper indicate that increase in CAD decrease in BD. It is possible to say that Turkey’s fiscal, tax and growth policies in the last periods provide this conclusion. That is to say, an increase in CAD helps to fix the budget balance. 2/3 of tax revenues come from indirect taxes which means most of tax revenues in Turkey come from consumption tax. It seems that economic growth in Turkey bases on consumption and this case result with CAD.This is not a sustainable situation. Because, a period of slowdown in the economy causes not only a decrease in CAD but also a deterioration of budget balance. This situation reduces the credibility of the government and the economy. Therefore Turkish economy has to cope with CAD not with tax policies but with increasing production facilities. If not, the economy may face with both deficits at the same time. References Abell, J. D. (1990) Twin Deficits During the 1980’s: An Emprical Investigation, Journal of Macroeconomics, 12, 81-96. Akbostancı, E. and G.İ. Tunç (2002) Turkish Twin Deficit: An Error Correction Model of Trade Balance, Middle East Technical Universty, ERC working papers, 1–17. Aksu, H and S. Başar (2005) İkiz Açıklar Hipotezi’nin Türkiye Açısından Araştırılması, İktisat, İşletme ve Finansa Dergisi, (20), 109–114. 50 Journal of Economic and Social Studies The Role of Twin Deficit Problem in Sustainable Growth: An Econometric Analysis for Turkey Alkswani, A.M. (2000) Twin Deficit Phenomenon in Petroleum Economy: Evidence from Saudi Arabia, Seventh Annual Conference, Economic Research Forum, Amman. Ay, A., Z. Karaçor, M. Mucuk, S. Erdoğan (2004) Bütçe Açığı - Cari İşlemler Açığı Arasındaki İlişki: Türkiye Örneği (1992-2003), Selçuk Üniversitesi, Sosyal Bilimler Enstitüsü Dergisi, 12, 75–82. Bahmani-Oskooee, Mohsen (1989) Effects of the US Government Budget on its Current Account: An Empirical Inquiry, Quarterly Review of Economics and Business, 29, 76-91. CBRT (2012) Electronic Data Delivery System (EDDS), www.tcmb.gov.tr Cholifihani, M. (2008) A Cointegration Analysis of Public Debt Service and GDP in Indonesia, Journal of Management and Social Sciences, 4(2): 68-81. Darrat, D.A. (1988) Have Large Budget Deficit Caused Rising Trade Deficit?, Southern Economic Journal, 54, 879–887. Dickey, D. and Fuller, W. A. (1979) Distribution of the Estimates for Autoregressive Time Series with a Unit Root, Journal of the American Statistical Association, 74, 427-431. Engle, R. F. and Granger, C. W. (1987) Co-integration and Error Correction Representation, Estimation and Testing, Econometrica, 55, 251-276. Enders, W. (1995) Applied Econometric Time Series First Edition, Wiley New York. Enders, W. and B. S. Lee (1990) Current Account and Budget Deficits : Twins or Distant Cousing, The review of Economics and Statistics, 72(3), 373- 381. Islam, M.F. (1998) Brazil’s Twin Deficits: An Empirical Examination, Atlantic Economic Journal, 26, 2, 121-128. Johansen, S. (1988) Statistical Analysis of Cointegration Vectors, Journal of Economic Dynamic and Control, 12, 1988: 231-254. Johansen, S. and Juselius, K. (1990) Maximum Likelihood Estimation and Inference on Cointegration with Application to the Deman for Money, Oxford Bulletin of Economic and Statistics, 52, 169-210. Volume 3 Number 2 Fall 2013 51 Johansen, S. (1995) Likelihood Basic Inference in Cointegration Vector Autoregressive Models, Oxford University Press, New York. Kutlar, A. and M. Şimşek (2001) Türkiye’de Bütçe Açıklarının Dış Ticaret Açıklarına Etkileri, Ekonometrik Bir Yaklaşım: 1984–2000, Dokuz Eylül Üniversitesi, İİBF Dergisi, 16 (1), 1–13. Merza, E., Alawin, M. and Bashayreh, A. (2012) The Relationship Between Current Account and Government Budget Balance: The Case of Kuwait, International Journal of Humanities and Social Science, 2(7), 168-177. Peker, O. (2009) Türkiye’deki Cari Açık Sürdürülebilir mi? Ekonometrik Bir Analiz, Kocaeli Üniversitesi Sosyal Bilimler Enstitüsü Dergisi, 17(1), 164-174. Puah, C., Lau, E. and Tan, K. L. (2006) Budget-Current Account Deficits Nexus in Malaysia, Munich Personal RePEc Archieves, 37677(27), 1-27. Utkulu U. (2003) Türkiye’de Bütçe Açıkları ve Dış Ticaret Açıkları Gerçekten İkiz mi? Koentegrasyon ve Nedensellik Bulguları”, D.E.Ü. İİBF Dergisi, 1(18), 45–61. Sever, E. And M. Demir (2007) Türkiye’de Bütçe Açığı ile Cari Açık Arasındaki İlişkilerin VAR Analizi ie İncelenmesi, Eskişehir Osmangazi Üniversitesi İİBF Dergisi, 2(1), 47-63. Vamvoukas, G. (1999) The Twin Deficits Phenomenon: Evidence From Greece, Applied Economics, 31, 1093-1100. Yücel F. and Ata, A. Y (2003) Eş-Bütünleşme ve Nedensellik Testleri Altında İkiz Açıklar Hipotezi: Türkiye Uygulaması, Çukurova Üniversitesi Sosyal Bilimler Enstitüsü Dergisi, 12(12). Zengin, A. (2000) İkiz Açıklar Hipotezi (Türkiye Uygulaması), Gazi Üniversitesi Ekonomik Yaklaşım Dergisi, 11(39), 37–67. Journal of Economic and Social Studies Cox Regression Models with Time-Varying Covariates Applied to Survival Success of Young Firms 1(*) Aygül Anavatan, Akdeniz University, Faculty of Economics and Administrative Sciences, Department of Econometrics, 07058, Antalya, Turkey aygulanavatan@akdeniz.edu.tr, Murat Karaöz Akdeniz University, Faculty of Economics and Administrative Sciences, Department of Econometrics, 07058, Antalya, Turkey, mkaraoz@akdeniz.edu.tr AbstrAct The most widely used model in multivariate analysis of survival data is proportional hazards model proposed by Cox. While it is easy to get and interpret the results of the model, the basic assumption of proportional hazards model is that independent variables assumed to remain constant throughout the observation period. Model can give biased results in cases which this assumption is violated. One of the methods used modelling the hazard ratio in the cases that the proportional hazard assumption is not met is to add a time-dependent variable showing the interaction between the predictor variable and a parametric function of time. In this study, we investigate the factors that affect the survival time of the firms and the time dependence of these factors using Cox regression considering time-varying variables. The firm data comes from Business Development Centers (İŞGEM) which is a prominent business incubation center operating in Turkey. KEYWOrDs Survival Analysis, Cox Regression Model, Proportional Hazard Assumption, New Firms ArtIcLE HIstOrY Submitted:22 Jun 2012 Resubmitted:03 January 2013 Accepted:25 March 2013 Jel code: C41, C24, M13 1 This research paper has been an extension to the findings of the scientific research project “The Factors Affecting Survival and Growth Performance of Newly Established Enterprises in Business Incubators: A Survey on the KOSGEB Business Development Centers (İŞGEM)”, 109K139, which has been funded with grant from TÜBİTAK (The Scientific and Technological Research Council of Turkey). We also acknowledge the administrative support to the project from Turkish Small and Medium Entreprises Development Organisation (KOSGEB). (*) Volume 3 Number 2 Fall 2013 53 Stavros RODOKANAKIS / Vasileios A. VLACHOS Introduction Survival analysis deals with the probability of occurrence of a given event at a set of particular points in a time interval (Cox and Oakes, 1984; Sertkaya, Ata and Sözer, 2005). In the small business and entrepreneurship literature, survival analysis has been used to track the start-ups over the years. The typical survival anaylsis may include the reports of hazard rates, ratios and survival curves while relating a likely set of independent variables to a specific event. A survival curve of a cohort of newly established firms reports what percentage of the cohort continue to survive since its inception over time, indicating whether some of the firms are failed over the years (Karaöz and Albeni, 2011). In many survival studies, it has been examined whether some variables or risk factors are effective on survival or not. Cox proportional hazards (PH) model is the most preferred model in order to investigate the effect of variables on survival time. The key assumption of Cox model is that hazard rate related to different levels of the factors is constant throughout the follow-up period (Başar, 2006). Violation of the PH assumption requires additional measures for unbiased results of Cox survival regression. In this paper, Cox regression has been applied to investigate the survival of newly established firms under incubation. Violation of PH assumption has been tested and further Cox regressions are performed considering time-varying effects of independent variables to survival. Survival Analysis Survival analysis is a collection of statistical procedures for data analysis for which the outcome variable of interest is time until an event occurs (Harrell, 2001). This event may be failure, and for this reason, the analysis of such data is often referred to as survival analysis (Bellera et al., 2010). The main objectives of the survival analysis are i) to estimate and interpret survival characteristics: Kaplan-Meier plots, median estimation and confidence intervals (CI), ii) to compare survival in different groups: Log-rank test, iii) to assess the relationship of explanatory variables to survival time: Cox regression model (Yay, Çoker and Uysal, 2007). In a survival analysis, it is usually referred to the time variable as survival time, because it gives the time that an individual has “survived” over some followup period (Geiss et al., 2009). It is also typically referred to the event as a failure, because the 54 Journal of Economic and Social Studies Evaluating the employment probability: Men and women in comparative perspective in Attica and Central Macedonia event of interest usually is death, disease incidence, or some other negative individual experience (Kleinbaum and Klein, 2005). When survival time (�) is defined as a random variable with cumulative distribution function �(�) = ��(�� � ��) and probability density function �(�) = � �(�)⁄� (�), survival function �(�) is explained by Equation (2.1) (Yay, Çoker and Uysal, 2007); �(�) = �(� � �) = 1 � �(�) (2.1) Survival function �(�) gives the probability that the random variable � exceeds the specified time � (Kleinbaum and Klein, 2005). All survival functions have the characteristics that i) they are nonincreasing; that is, they head downward as � increases, ii) at time � = 0, �(�) = �(0) = 1; that is, at the start of the study, since no one has gotten the event yet, the probability of surviving past time 0 is one, iii) at time � = ∞, �(�) = �(∞) = 0; that is, theoretically, if the study period increased without limit, eventually nobody would survive, so the survival curve must eventually fall to zero (Kleinbaum and Klein, 2005). The hazard function �(�), with its complement of survival function �(�), is given by Equation (2.2), where �� denotes a small interval of time (Kleinbaum and Klein, 2005); �(�) = �im���� �(��������|���) (2.2) �� The hazard function �(�) gives the instantaneous potential per unit time for the event to occur, given that the individual has survived up to time � (Tabatabai et al., 2007). In contrast to the survival function, which focuses on not failing, the hazard function focuses on failing, that is, on the event occurring. Thus, in some sense, the hazard function can be considered as giving the opposite side of the information given by the survival function (Kleinbaum and Klein, 2005). The Cox Proportional Hazards Model The Cox PH model is usually written in terms of the hazard model formula� shown at Equation (2.3). This model gives an expression for the hazard at time t 2 Volume 3 Number 2 Fall 2013 55 Stavros RODOKANAKIS / Vasileios A. VLACHOS for an individual with a given specification of a set of explanatory variables denoted by �. That is, � represents a collection of predictor variables that is being modeled to predict an individual’s hazard (Kleinbaum and Klein, 2005); � �(�� �) = �� (�)� ∑��� �� �� (2.3) The Cox model formula says that the hazard at time � is the product of two quantities. The first of these, �� (�), is called the baseline hazard function. The second quantity is the exponential expression � to the linear sum of �� �� , where the sum is over the � explanatory � variables (Kleinbaum and Klein, 2005). In general, a hazard ratio (��) is defined as the hazard for one individual divided by the hazard for a different individual. The two individuals being compared can be distinguished by their values for the set of predictors, that is, the X’s. Hazard ratio is shown by the following formula, where �∗ denotes the set of predictors for one individual, and � denotes the set of predictors for the other individual (Kleinbaum and Klein, 2005); � * � � * � = h�t�X � = h0 (t) exp�∑ βi Xi � = exp�∑p β�i �X i* -Xi �� HR i=1 � � (t�X) � (t) h h0 exp�∑ βi X i � (2.4) �∗ = ���∗ � ��∗ � � � ��∗ ��and�� = ��� � �� � � � �� � denote the set of �’s for two individuals. Once the model is fitted and the values for �∗ and � are specified, the value of the exponential expression for the estimated hazard ratio is a constant, which does not depend on time. If we denote this constant by ��̂, then hazard ratio can be written as shown below (Kleinbaum and Klein, 2005); � �� = exp�∑��� ��� (��∗ � �� )� (2.5) If hazard ratio is greater than 1, the group which has the distinction of 1 category of the variable will higher significantly likely to be exposed to interest event by comparison 0 category of that variable. If the hazard ratio is equal to 1, chance of closing the two groups are equal; if it is between 0 and 1, the group receiving 0 category value has a lower closing probability by comparison 1 category. 3 56 Journal of Economic and Social Studies Evaluating the employment probability: Men and women in comparative perspective in Attica and Central Macedonia The basic assumptions of the Cox regression model can be explained as follows (Yay, Çoker and Uysal, 2007); i) the effects of independent variables on the hazard function are loglinear. ii) The relationship between loglineer function of independent variables and the hazard function is multiplicative. iii) In addition to these two assumption, observations should independent of each other and hazard ratio should remains unchanged with respect to time, ie., is constant. This assumption related to hazard ratio is known as proportional hazard assumption. A key reason for the popularity of the Cox model is that, even though the baseline hazard is not specified, reasonably good estimates of regression coefficients, hazard ratios of interest, and adjusted survival curves can be obtained for a wide variety of data situations. Another way of saying this is that the Cox PH model is a “robust” model, so that the results from using the Cox model will closely approximate the results for the correct parametric model (Kleinbaum and Klein, 2005). In addition to the general “robustness” of the Cox model, the specific form of the model is attractive for several reasons (Kleinbaum and Klein, 2005). First, the � exponential part � ∑��� �� �� of hazard model ensures that the fitted model will always give estimated hazards that are non-negative. Another tempting property of the Cox model is that, even though the baseline hazard part of the model is unspecified, it is still possible to estimate the �’s in the exponential part of the model. Lastly, it is preferred over the logistic model when survival time information is available and there is censoring. That is, the Cox model uses more information (the survival times) than the logistic model, which considers a (0�1) outcome and ignores survival times and censoring. Evaluating the Proportional Hazards Assumption For variables not satisfying the non-proportionality assumption, the power of the corresponding tests is reduced, that is, it is less likely to conclude for a significant effect when there is actually one. If the hazard ratio is increasing over time, the estimated coefficient assuming PH is overestimating at first and underestimating later on. For those variables of the model with a constant hazard ratio, the power of tests is also reduced as a consequence of an inferior fit of the model (Bellera et al., 2010). 4 Volume 3 Number 2 Fall 2013 57 Stavros RODOKANAKIS / Vasileios A. VLACHOS There are three general approaches to assess the PH assumption: 1) Graphical Approaches; Kaplan-Meier and log-log plots, observed versus expected plots, 2) Goodness of fit (GOF) test, 3) Statistical Methods; schoenfeld residuals, the logrank test and time-dependent covariates. Extension of the Cox Proportional Hazards Model An important feature of this formula, which concerns the PH assumption, is that the baseline hazard is a function of �, but does not involve the �’s. The �’s in the formula are called time-independent �’s (Kleinbaum and Klein, 2005). It is possible, nevertheless, to consider �’s which do involve �. Such �’s are called time-dependent variables. If time-dependent variables are considered, the Cox model form may still be used, but such a model no longer satisfies the PH assumption, and is called the extended Cox model (Kleinbaum and Klein, 2005). In the case of being time-dependent explanatory variables, Cox regression model expands to a model which contains time-independent variables and some functions of the time the product with these variables. Independent variables are, where �� � �� � � � ��� time-independent variables and �� (�)� �� (�)� � � ��� (�) time-dependent variables (Sertkaya, Ata and Sözer, 2005); �(�) = ��� � �� � � � ��� � �� (�)� �� (�)� � � ��� (�)� as shown. Accordingly, Cox regression model is, � and � which denote vector of coefficients of explanatory variables (Sertkaya, Ata and Sözer, 2005); � � � � ���� �(�)� = �� (�) exp �∑��� �� �� �(�)� �� �� + ∑��� (2.6) as written. Where �(�) is defined as a function of time. Selection of �(�) varies according to the state of the variables used and according to the information level of the researchers. This function usually is defined in the form of �, ���( �), ��(�) or step functions (Sertkaya, Ata and Sözer, 2005). The general hazard ratio formula for extended Cox model is shown below (Kleinbaum and Klein, 2005); 58 5 Journal of Economic and Social Studies Evaluating the employment probability: Men and women in comparative perspective in Attica and Central Macedonia � � (�) = ����� �� � ∗ (�)� � ����(�)� (2.7) �� � �� = exp �∑��� �� ���∗ (�) � �� (�)�� �� ���∗ � �� � + ∑��� An Application Into New Firm Survival Under Incubation Although the survival analysis extensively has been used in medical research on individuals, recently it becomes widely popular in business success and survival research. Thus, rather than on individuals, in this paper, we apply Cox regression to investigate the survival of newly established firms under incubation. There are studies applying survival violation of PH assumption has been tested and further Cox regressions are performed considering time-varying effects of independent variables to survival. Our 414 observations on firm characteristics acquired from 12 different incubators, İŞGEMs, located across Turkey, in Zonguldak, Tarsus, Ereğli, Eskişehir, Adana, Mersin, Van, Avanos, Samsun, Elazığ, Yozgat and Diyarbakır provinces. The data includes almost all firms that currently existing İŞGEMs or the firms that resided in the past yet left İŞGEMs by graduation or failure. The survey data consists of the total. A business incubator can be identified as an organization which mentors the development of newly founded firms by specialized services such as providing office space, specialized staff, machinery, equipment, facilities and business assistance (Aernoudt, 2004). Thus a business incubator is a framework organization which contains a collection of newly established firms. İŞGEMs are one of the significant business incubation concept operating in Turkey. Variables Used in the Analysis For our analysis, factors affecting the initial success of young enterprises can be summarized as i) Human capital characteristics of new enterprise's owner such as education level and sector experience, ii) Firm characteristics such as scale, age and human capital, iii) Industry characteristics such as market growth rate and entry barriers, iv) Incubation features, v) Other external factors such as macroeconomic fluctuations, regional factors and public policies (Hackett and Dilts, 2004; Aernoudt, 2004). All of the data and variables used in our analysis are taken from Karaöz and Albeni (2011) and descriptive statistics and definitions are presented at 6 Volume 3 Number 2 Fall 2013 59 Stavros RODOKANAKIS / Vasileios A. VLACHOS An Application Into New Firm Survival Under Incubation Although the survival analysis extensively has been used in medical research on individuals, recently it becomes widely popular in business success and survival research. Thus, rather than on individuals, in this paper, we apply Cox regression to investigate the survival of newly established firms under incubation. There are studies applying survival violation of PH assumption has been tested and further Cox regressions are performed considering time-varying effects of independent variables to survival. Our 414 observations on firm characteristics acquired from 12 different incubators, İŞGEMs, located across Turkey, in Zonguldak, Tarsus, Ereğli, Eskişehir, Adana, Mersin, Van, Avanos, Samsun, Elazığ, Yozgat and Diyarbakır provinces. The data includes almost all firms that currently existing İŞGEMs or the firms that resided in the past yet left İŞGEMs by graduation or failure. The survey data consists of the total. A business incubator can be identified as an organization which mentors the development of newly founded firms by specialized services such as providing office space, specialized staff, machinery, equipment, facilities and business assistance (Aernoudt, 2004). Thus a business incubator is a framework organization which contains a collection of newly established firms. İŞGEMs are one of the significant business incubation concept operating in Turkey. Variables Used in the Analysis For our analysis, factors affecting the initial success of young enterprises can be summarized as i) Human capital characteristics of new enterprise’s owner such as education level and sector experience, ii) Firm characteristics such as scale, age and human capital, iii) Industry characteristics such as market growth rate and entry barriers, iv) Incubation features, v) Other external factors such as macroeconomic fluctuations, regional factors and public policies (Hackett and Dilts, 2004; Aernoudt, 2004). All of the data and variables used in our analysis are taken from Karaöz and Albeni (2011) and descriptive statistics and definitions are presented at Table 3.1. The entrepreneur’s age, gender, education, professional career history and experience and family environment factors are the main factors in the literature in terms of the survival of firms (Karaöz and Albeni, 2011). 60 Journal of Economic and Social Studies Evaluating the employment probability: Men and women in comparative perspective in Attica and Central Macedonia Table 1. The variables used in analysis and descriptive statistics VARIABLE DEFINITION Number of Mean Minimum Maximum Observation EVENT OF INTEREST If the firm is closed (failed) during or after 414 the incubation 1, otherwise 0 exit - 0 1 41,52 2 158 414 - 0 1 414 - 0 1 367 3,64 3 4,25 414 - 0 1 414 5,83 0 40 414 - 0 1 414 1,24 1 4 DEPENDENT VARIABLE The elapsed time from the firm’s entry 404 into incubation until it’s closed (month) incubage INDEPENDENT VARIABLES If entrepreneur’s income only comes from the incubated firm 1, otherwise 0 If entrepreneur is female 1, male 0 (If gender there are both male and female partner 0) Entrepreneur’s age (If there is a partnerlnentage ship, it is taken as the oldest entrepreneur’s age-logarithmic scale) If entrepreneur is a college graduate 1, enteduuni otherwise 0 (if there is a partnership and one of the partners is college graduate 1) Entrepreneur’s prior experience before arentexp riving İŞGEM (year) If there is a role model for entrepreneurfamily ship in entrepreneur’s family or surrounding 1, otherwise 0 The number of partners within the estabpartner lished firm CHARACTERISTICS OF THE ENTREPRENEUR income export If the firm export 1, otherwise 0 414 - 0 1 lnempini initial firm size (logarithmic scale) 392 1,31 0 5,70 414 - 0 1 414 - 0 1 414 - 0 1 414 - 0 1 If the firm is a brand owner 1, otherwise 0 414 - 0 1 If firm’s founding capital is completely loan 1, otherwise 0 If entrepreneur is in cooperation with networkstakeholders within and outside the incuing bator 1, otherwise 0 If entrepreneur has made innovation 1, innova otherwise 0 If the firm has had an advertising 1, othadvert erwise 0 FEATURES OF THE FIRM onlyloan brand Volume 3 Number 2 Fall 2013 61 EXTERNAL FEA- INDUSTRIAL INCUBATION SERVICES TURES PROPERTIES AND PROPERTIES Stavros RODOKANAKIS / Vasileios A. VLACHOS comserv whenest incubsize sector compete prorank cycle If entrepreneur has used at least one of the common services offered by incuba- 414 tion 1, otherwise 0 If the incubated firm entered the incubation center within first 3 years (36 months) 414 of incubation center 1, otherwise 0 The number of incubation’s workshop If the firm is in the manufacturing industry 1, in the service sector 0 Intensity of competition in the sector (1-5 Likert scale) (%) Share of the GDP per capita of the province in the Country GDP where the incubation center is located If the firm has experienced an economic crisis 1, otherwise 0 - 0 1 - 0 1 414 43,14 14 84 411 - 0 1 410 - 1 5 414 1,51 0,59 2,07 414 - 0 1 Source: Karaöz M. and Albeni M.,” The Factors Affecting Survival and Growth Performance of Newly Established Enterprises in Business Incubators: A Survey on the KOSGEB Business Development Centers (İŞGEM)”, TÜBİTAK Project No: 109K139, Isparta, March 2011. (exit) variable is used as dependent variable. It takes the value of 1 if the firm is closed within the period in incubation or after the firm has graduated from incubation, the value of 0 in other cases. In addition to (exit), exit time (incubage) is the other main variable in our survival analysis. As seen at Table 3.1, for our dataset, the firms’ average life expectancy is 41.52 months. The maximum survival time observed as 158 months. Some of the firms failed either during or some time after leaving the incubator. Yet some of the firms still continue their activity either at incubator or outside the incubator. Survival curve of firms has been presented at Figure 3.1. According to the figure, surivors after 158 months diminish to about 20%. 62 Journal of Economic and Social Studies Evaluating the employment probability: Men and women in comparative perspective in Attica and Central Macedonia Figure 1. The survival curve of firms which is present or graduate from incubation (month) 0.00 0.25 0.50 0.75 1.00 Kaplan-Meier survival estimate 0 50 100 150 analysis time Results All Cox Regression results with and without considering time effects are presented in Table 3.2. (gender), (lnentage), (family), (export), (lnempini), (advert), (brand), (comserv), (sector), (compete) and (cycle) variables are insignificant in Model 1, which the time-dependent effects have not taken into account. According to Model 1 estimates, entrepreneur’s gender, age, whether s/he is affected family environment; initial firm size, whether the firm exports and does advertising, whether the firm is brand owner; whether the firm takes advantage of common services offered by incubators; the sector in which the firm, intensity of competition in the sector and whether the firm experienced any macroeconomic crisis are not significant on the firms’ survival times. Our tests indicate that further estimations are necessary using time-dependent variables. Thus we produce further new estimates and present most relevant two model results at Table 3.2. Model 2 includes the variables which in Model 1 and all of the interaction terms created by each of these variables multiplying , which is a function of time, in order to handle variable-time interaction. The Model 3 are obtained by removing the interaction terms of (lnempini), (innova), (enteduuni), (whenest), (export), (brand), (gender), (sector), (advert), (networking), (entexp), (income), (onlyloan), (partner), (family), (lnentage), (comserv), (compete) and (cycle) variables from the model. Model 3 is the best model that takes into account timedependent effects. The variables of (incubsize) and (prorank) are found to be the time-dependent variables. Volume 3 Number 2 Fall 2013 63 Stavros RODOKANAKIS / Vasileios A. VLACHOS Table 2. The estimates of the basic model and Cox model with time-dependent variables Variable Coefficient Variable income gender lnentage enteduuni entexp family partner export lnempini 1.18 -0.056 0.265 0.659 -0.084 -0.307 -1.71 0.827 0.214 0.010*** 0.892 0.732 0.044** 0.042** 0.402 0.015** 0.308 0.278 5.16 1.14 6.45 0.289 -0.307 -4.63 1.07 6.47 0.274 0.253 0.745 0.422 0.924 0.342 0.205 0.844 0.576 0.847 1.75 -0.093 0.721 0.762 -0.101 -0.249 -2.39 0.951 0.196 0.000*** 0.826 0.364 0.024** 0.013** 0.518 0.001*** 0.303 0.298 income gender lnentage enteduuni entexp family partner export lnempini -0.959 -0.352 -1.48 0.139 0.055 1.23 -1.08 -1.25 -0.023 0.427 0.699 0.474 0.861 0.505 0.212 0.457 0.678 0.954 - - - - - - - - - - - - - - - - - - Model 1 Model 2 Model 3 Model 2 (cont.) Model 3 (cont.) *, **, and *** indicate significance at the 1, 5, and 10% levels, respectively. Log-likelihood and prob values of Model 1, 2 and 3, respectively, are -190.632 [0.000***], -165.552 [0.000***] and -173.255 [0.000***]. 64 Journal of Economic and Social Studies Evaluating the employment probability: Men and women in comparative perspective in Attica and Central Macedonia onlyloan networking innova advert brand comserv whenest incubsize sector compete prorank cycle -1.03 -1.47 -1.67 0.636 0.865 0.264 -1.18 0.063* 0.004*** 0.006*** 0.17 0.275 0.592 0.013** 0.002*** 0.738 0.416 0.013** 0.21 -6.97 4.45 -2.24 -2.74 -4.22 -6.63 -0.985 -0.261 19.4 0.215 0.333 0.74 0.522 0.685 0.217 0.813 0.015** 0.291 0.23 0.003*** 0.085* -1.88 -1.54 -2.46 0.615 1.61 0.638 -2.25 -0.253 16.9 0.002*** 0.004*** 0.001*** 0.198 0.074* 0.234 -0.02 -0.156 -0.157 -4.39 2.92 -0.425 -0.341 0.000*** 0.009*** 0.362 0.099* -1.16 0.46 10.5 0.791 0.002*** 0.040** onlyloan networking innova advert brand comserv whenest incubsize sector compete prorank cycle 1.3 -1.7 -0.147 0.924 1.48 1.99 -0.415 0.059 0.378 0.177 0.937 0.418 0.582 0.161 0.718 0.030** 0.344 0.17 0.001*** 0.108 - - - - - - - 0.058 - - -5 - - - - - - - 0.017** - - 0.001*** - Volume 3 Number 2 Fall 2013 1.03 -0.903 -5.74 -2.5 - 65 Stavros RODOKANAKIS / Vasileios A. VLACHOS Also considering the Model 2 and 3, we obtain various results regarding the variables. The possibility of failure of the firms, whose owners only dependent on earnings coming from its new-born firm, is about 6 times higher than other firms. In this case it has been seen that the entrepreneurs having income from other sources are more likely to be successful in start-up business. It is interesting to see the result that the firms whose owners are university graduates have about two times higher risk of failure than other firms. Yet there is a plausible explanation. Most of the incubator residents are specialized in low-technology industries, which have higher likelihood of failure. University graduates, who later realized that the new business has not much prospect, close the firm immediately and return looking for a job related to his carreer. University graduates have higher chance of finding a better paying job than non-university graduates. By the same token, non-university graduates seem to strive more to keep the new business alive. An increase in the number of partners in the firm decreases the possibility of failure of firms to 20%. It is interesting to see that failure risk of firms, whose founding capital is formed entirely by loans, is only about %15 of the other firms, whose initial capital is partially or fully self-financed. If an entrepreneur is in collaboration with stakeholders within and outside the incubation, survival probability of the firm becomes approximately 5-times higher. Moreover, it has been seen from the estimates that innovation activity of new firms increases chance of survival approximately 12-times. Brand ownership also increases the chance of the firm’s survival. Establishing a firm within an incubation center that is within its first 3-years (36 months) increases survival probability. Finally, firms those experiences a macroeconomic crisis have nearly two times more likelihood of failure than others. Conclusions Cox proportional hazard model, besides others, rest on proportional hazards assumption that independent variables do not vary with time. When PH assumption is violated, Cox regression estimates become biased. Then, Cox survival estimates can be corrected by including the time-varying effects to the analysis. Identification and calculation of time-dependent effects give the opportunity to obtain some otherwise unseen valuable special time pattern information. In our analysis, initially, the Cox regression was performed by considering that all explanatory variables are constant over time. Then, extended Cox regression models 66 Journal of Economic and Social Studies Evaluating the employment probability: Men and women in comparative perspective in Attica and Central Macedonia were estimated by including the time-dependent explanatory variables in the model. Our extended model results have shown that it become useful to estimate the Cox Proportional Hazards regression by also including the time-varying explanatory variables to the analysis. Both the time-independent and time-dependent variables create significant effects on the probability of survival of the İŞGEM firms. Overall, our estimates suggest that entrepreneurial experience acquired before starting business at İŞGEM, higher number of partners in the firm, formation of the firm’s capital completely by loan, being in collaboration with stakeholders within and outside the incubator, innovative activities in the firm, starting the new business within first 36 months of an incubator (in a young incubator), higher number of office spaces, establishing the firm in an economically larger province, and the density of competition in the sector have positive impact on the probability of survival of the new-born firms within the incubator. Entrepreneurs whose only source of income comes from the young firm, who has college diploma, who has brand ownership at the firm, who experience a macroeconomic crisis are more likely to fail. References Aernoudt, R. (2004). Incubators: Tool for Entrepreneurship?. Small Business Economics, 23, 127-35. Başar, E. (2006, May). Orantılı Olmayan Hazard Üzerine Bir Çalışma. Paper presented at 5. İstatistik Günleri Sempozyumu, Antalya, pp.111-16. Bellera, C.A., MacGrogan, G., Debled, M., De Lara, C.T., Brouste, V., & Mathoulin-Pélissier, S. (2010). Variables with time-varying effects and the Cox model: Some statistical concepts illustrated with a prognostic factor study in breast cancer. BMC Medical Research Methodology, 10:20. Cox, D.R., & Oakes, D. (1984). Anaylsis of Survival Data. London: Chapman and Hall. Demirgil, H. (2008). Firmaların Hayatta Kalma and Büyüme Performanslarını Belirleyen Faktörler: Göller Bölgesi Üzerine Bir Araştırma. Ph.D. Thesis. Department of Economics, Süleyman Demirel University, Isparta. Geiss, K., Meyer, M., Radespiel-Tröger, M., & Gefeller, O. (2009). SURVSOFT—Software for nonparametric survival analysis. Computer Methods and Programs in Biomedicine, Elsevier Ireland LTD., 96, 63–71. Hackett, M., & Dilts, D.M. (2004). A Systematic Review of Business Incubation Research. Journal of Technology Transfer, 29, 55-82. Volume 3 Number 2 Fall 2013 67 Karaöz, M., & Albeni, M., (2011), İş Kuluçkalarında Yeni Kurulan Girişimlerin Hayatta Kalma ve büyüme Performansını Etkileyen Faktörler: KOsGEb İş Geliştirme Merkezleri (İŞGEM) Üzerine bir Araştırma. The Scientific and Technological Research Council of Turkey (TÜBİTAK). (Issue Brief No. 109K139). Kleinbaum, D.G., & Klein, M. (2005). Survival Analysis: A Self-Learning Text (2nd Ed.). New York: Springer. Scheike, T. H. (2004). Time-Varying Effects in Survival Analysis. In: N. Balakrishnan & C.R. Rao. (Ed.), Advances in Survival Analysis, 61-85. Amsterdam: Elsevier North-Holland. Sertkaya, D., Ata, N., & sözer, M. T. (2005). Yaşam çözümlemesinde zamana bağlı açıklayıcı değişkenli Cox regresyon modeli. Ankara Üniversitesi Tıp Fakültesi Mecmuası, 58, 153-58. Tabatabai, M. A., Bursac, Z., Williams, D. K., & Singh, K. P. (2007). Hypertabastic survival model. Theoretical Biology and Medical Modelling, 4:40. Yay, M., Çoker, E., & Uysal, Ö. (2007). Yaşam Analizinde Cox Regresyon Modeli ve Artıkların İncelenmesi, Cerrahpaşa Tıp Dergisi, 38, 139-45. Journal of Economic and Social Studies Unit Root Properties of Energy Consumption and Production in Turkey Özgür Polat Department of Economics Dicle University, Diyarbakır, Turkey opolat@dicle.edu.tr Enes E. Uslu Department of Econometrics Ataturk University, Erzurum, Turkey ertad10@hotmail.com Hüseyin Kalyoncu Department of International Trade Meliksah University, Kayseri, Turkey hkalyoncu@meliksah.edu.tr AbstrAct This study analyzes unit root properties of total and sectorial energy production and consumption series of Turkey. This study is the first to analyze unit root properties of Turkish energy production and consumption in detail. The unit root analysis of energy production and consumption are tested by using unit root tests based on LM considering without structural break and with one and two structural breaks. According to unit root test without structural break, the unit root hypothesis is rejected only for consumption of natural gas. The unit root hypothesis is rejected for 15 out of the 33 series by the LS test with one structural break. When unit root test with two structural breaks are conducted, 25 out of the 33 series are found to be stationary around a deterministic trend. The production of hydraulic and the consumption of lignite, electricity, petroleum, coal and electricity, total energy and petroleum consumption in Transportation sector are found to be nonstationary, which indicates that the impacts of innovations on these variables will be permanent. The policy implication of the results suggests that the impacts of shocks on energy consumption and production will be temporary and not have a long memory for most of variables. KEYWOrDs Energy Consumption, Energy Production, Unit Root Analysis, Turkey ArtIcLE HIstOrY Submitted: 04 October 2012 Resubmitted: 24 December 2012 Accepted: 25 March 2013 JEL code: Q43, Q48 Volume 3 Number 2 Fall 2013 69 Özgür Polat / Enes E. Uslu / Hüseyin Kalyoncu Introduction The impact of unit root properties of energy variables for the formulation and consequences of economic policies are crucial in several aspects, especially on structural transitions from shocks in energy markets towards key macroeconomic variables (Mishra, Sharma, & Smyth, 2009; Narayan & Smyth, 2007). Impact of shocks on energy variable can be permanent or transitory according to its unit root properties. If the energy variable is stationary, impact of shocks will be transitory and long short term. On the other hand, if the energy variable is not stationary, the impacts of shocks will be permanent and have a long memory. Hendry and Juselius (2000) indicate that economic variables can inherit unit root properties from related economic variables and can in turn transmit this property to other related variables. They argued that relationship between economic variables can spread unit root properties throughout the economy. In this context, knowledge of unit root properties of an energy variable is of importance, since this property can be inherited by related macroeconomic variables. The impact of energy demand on economic activity can be serious. The literature has shown that energy price shocks, via their substantial impact on energy consumption, have significant impacts on output (Chang & Wong, 2003; Du, Yanan, & Wei, 2010; Hamilton, 1996, 2007; Huang, Hwang, & Peng, 2005; Jayaraman & Choong, 2009; Jiménez-Rodríguez, 2008; Lardic & Mignon, 2008; B. R. Lee, Lee, & Ratti, 2001; Lorde, Jackman, & Thomas, 2009; Zhang, 2011), inflation (Chang & Wong, 2003; Cologni & Manera, 2008; Cuñado & Pérez de Gracia, 2003; Ewing & Thompson, 2007), unemployment (Carruth, Hooker, & Oswald, 1998; Chang & Wong, 2003; Rafiq, Salim, & Bloch, 2009), employment (Papapetrou, 2001), stock market (Arouri, Lahiani, & Nguyen, 2011; Basher, Haug, & Sadorsky, 2012; Evangelia, 2001; Filis, Degiannakis, & Floros, 2011; Park & Ratti, 2008; Sadorsky, 1999), investment (Rafiq et al., 2009), the budget deficit (Rafiq et al., 2009), exchange rate (Ayadi, 2005; Basher et al., 2012; S. S. Chen & Chen, 2007; Narayan, Narayan, & Prasad, 2008; Özturk, Feridun, & Kalyoncu, 2008), interest rate (Lowinger, Wihlborg, & Willman, 1985; Park & Ratti, 2008), exports (Chiou Wei & Zhu, 2002; Faria, Mollick, Albuquerque, & León-Ledesma, 2009; Zhang, 2011), fluctuations in business cycle (Kim & Loungani, 1992) and money supply (Zhang, 2011). Besides shocks on energy demand, Hamilton (2007) showed that disruptions on energy supply can also have significant impact on economic activity by presenting a model based on Cobb-Douglas production function as below: 70 Journal of Economic and Social Studies Unit Root Properties of Energy Consumption and Production in Turkey Analysis for Selected OECD Countries Y = F ( L, K , E ) (1) where output (Y) is production, (L) is labour, (K) is capital and (E) is energy use of a firm. The profits (π) of a firm can be estimated as: (2) where P is the price of output per unit, W is the nominal wage paid for labour, Q is the nominal cost of energy used in the production process and r is the nominal rate of rented capital. The equilibrium energy price for rational firm will be at a level where marginal product of energy is equal to unit price of energy: FE ( L, K , E ) = Q / P (3) where FE(L,K,E) is the partial derivative of F(.) regarding E. The following equation will be obtained in case both sides of the equation (3) are multiplied by E and divided by Y: (4) Eq (4) indicates that the elasticity of output regarding change in energy consumption used in the production process can be derived from the cost of the energy used to produce the total output. Disruptions in energy production will affect energy prices and a change in energy prices used in production process will also have a significant impact on output of an economy as shown in Eq (4). Therefore, shocks on non-stationary energy production series will be permanent and affect economic activity perpetually , while shocks on stationary energy production series will be transitory and affect economic activity temporarily, via transmission mechanism (Narayan, Narayan, & Smyth, 2008). The unit root properties of energy variables are of importance for forecasting these variables. Accurate forecasts are crucial for energy planning and policy formulation. Future values of a stationary energy variable can be forecasted based on its past behavior (P. F. Chen & Lee, 2007), while past data about a nonstationary energy variable are useless in forecasting (Mishra et al., 2009). Stationarity of energy consumption can be due to a multitude of factors. Hsu, Lee, and Lee (2008) suggested that abundance of energy resources, less energy consumption, new environmental regulations and laws introduced by governments and middle income level may lead to stationarity of energy consumption. Volume 3 Number 2 Fall 2013 71 Özgür Polat / Enes E. Uslu / Hüseyin Kalyoncu The goal of this study is to analyze the unit root properties of energy consumption and production in Turkey by employing a Lagrange Multiplier based unit root test without structural break proposed by Schmidt and Phillips (1992) (SP) and a unit root test considering one structural break proposed by J. Lee and Strazicich (2004) (LS) and two structural breaks developed by J. Lee and Strazicich (2003) (LS). If the time series of the variable to be tested for the unit root properties has structural breaks, the unit root hypothesis cannot be rejected by conventional unit root tests (Perron, 1989). Monte Carlo simulations point that statistical performance of LS is better than other alternatives (Narayan, Narayan, & Popp, 2010). This study is the first to investigate unit root properties of Turkish energy production and consumption in detail. The next section briefly summarizes the literature on studies analyzing the unit root properties of energy consumption and production. Section 3 describes data used in the analysis. Section 4 summarizes the unit root tests used in this study. Section 5 presents results of the unit root test. Section 6 discusses main findings and implications of the results for policy formulation and implementation. Brief Overview of the Literature Although there have been numerous studies analyzing the unit root properties of energy consumption series, only a handful of studies have investigated energy production. Barros, Gil-Alana, and Payne (2011) examine the time series behavior of oil production for 13 OPEC member countries for the period of January 1973 and October 2008. They found that oil production series have mean reverting persistence with breaks identified in 10 out of the 13 countries. The results of the study indicate that the impact of shocks on oil production in these countries will be persistent in the long run for all countries. Narayan, Narayan, and Smyth (2008) analyze the unit root properties of crude oil production for 60 countries by conducting panel data unit root tests with and without structural breaks between 1971 and 2003. The results of tests without a structural break are inconclusive, while the results of test with a structural break are conclusive and indicate the stationary structure of production series of crude oil and natural gas liquids. Maslyuk and Smyth (2009) test for non-linarities and unit roots in crude oil production. They used monthly crude oil production for 17 OPEC and non-OPEC 72 Journal of Economic and Social Studies Unit Root Properties of Energy Consumption and Production in Turkey Analysis for Selected OECD Countries countries between January 1973 and December 2007. The results of their study show the presence of threshold effects on the crude oil production and unit root for 11 of the countries in both regimes and a partial unit root for the others. In contrast to the dearth of studies investigating unit root properties of energy production series, there are numerous studies on unit root properties of energy consumption. Narayan and Smyth (2007) analyze the stationarity properties of per capita energy consumption of 182 countries for the period of 1979 to 2000 by using annual data. The results of univariate unit root test indicate that the series of 56 countries are nonstationary at the 10% level or better. The panel data unit root test indicate that there is overwhelming evidence about stationary of energy consumption. P. F. Chen and Lee (2007) investigate energy consumption per capita series of 7 regional panel sets for the period of 1971 to 2002 by employing panel unit root testing procedure, and find stationary structure in all series. A substantial literature review about the unit root properties of energy consumption can be found in P. F. Chen and Lee (2007), Hsu et al. (2008) and Aslan and Kum (2011). Data The unit root properties of primary total production, total and sectorial consumption of various energy variables of Turkey covering different periods are explored in this study as shown in Table 1. Energy production data are obtained from TURKSTAT (Turkish Statistical Institution) and energy consumption data are obtained from Ministry of Energy and Natural Resources (MENR) of Turkey. The periods of analysis are determined by data availability. All data used in this study are transformed to natural logarithmic form prior to unit root tests. Descriptive statistics of the variables subject to analysis are presented in Table 1. Volume 3 Number 2 Fall 2013 73 Özgür Polat / Enes E. Uslu / Hüseyin Kalyoncu Table 1. Descriptive Statistics Series Code PCL PLT PPM PNG PHC PGL PWD PEY CCL CLT CPM CNG CHC CGL CWD CEY IND IND_PET IND_ECT IND_NGS TPT TPT_PET TPT_ECT RES RES_PET RES_ECT ACL ACL_PET ACL_ECT OSC OSC_PET OSC_ECT NEY Series Name Period Production Coal 1950 - 2008 Lignite 1950 - 2008 Petroleum 1950 - 2008 Natural Gas 1976 - 2008 Hydraulic 1950 - 2008 Geothermal 1963 - 2008 Wood 1950 - 2008 Electricity 1923 - 2009 Consumption Coal 1970 - 2009 Lignite 1970 - 2009 Petroleum 1970 - 2009 Natural Gas 1976 - 2009 Hydraulic 1970 - 2009 Geothermal 1970 - 2009 Wood 1970 - 2009 Electricity 1923 - 2009 Sectorial Consumption Industrial 1970 - 2009 Industrial (Petroleum) 1970 - 2009 Industrial (Electricity) 1970 - 2009 Industrial (Natural gas) 1976 - 2009 Transportation 1970 - 2009 Transportation (Petroleum) 1970 - 2009 Transportation (Electricity) 1970 - 2009 Residential 1970 - 2009 Residential (Petroleum) 1970 - 2009 Residential (Electricity) 1970 - 2009 Agricultural 1970 - 2009 Agricultural (Petroleum) 1970 - 2009 Agricultural (Electricity) 1970 - 2009 Other sectors 1970 - 2009 Other sectors (Petroleum) 1970 - 2009 Other sectors (Electricity) 1970 - 2009 Non-energy 1970 - 2009 Unit Mean Std. Dev. TEP TEP TEP TEP TEP TEP TEP 106 kWh 2097 5498 2325 293 1322 318 4405 34695 623 4569 1325 284 1329 309 764 53778 TEP TEP TEP TEP TEP TEP TEP 106 kWh 6411 8084 23066 9089 1971 354 4776 28122 3942 4061 7786 10980 1215 296 615 43089 TEP TEP TEP TEP TEP TEP TEP TEP TEP TEP TEP TEP TEP TEP TEP TEP TEP 15218 4699 2630 2508 8869 8637 36 16368 3346 2252 2083 1927 129 18450 4700 2381 1471 8461 1810 1773 2680 3826 3976 22 4926 2428 2079 1208 1058 149 6117 1670 2227 1152 Note: TEP indicates Ton Equivalent Petroleum Time series plot for the production of energy variables of Turkey are shown in Figure 1. Decrease in petroleum, coal and wood production series are remarkable in com- 74 Journal of Economic and Social Studies Unit Root Properties of Energy Consumption and Production in Turkey Analysis for Selected OECD Countries parison to other series in recent years. The decrease in wood production for energy usage indicates substitution for this resource with other energy resources such as natural gas. Trends for other series increase with some fluctuations over the periods analyzed and display steep increase thereafter. However, electricity and geothermal production series have no serious fluctuation indicating successful production policies on these energy variables and these production process variables’ structural strength towards disruptive shocks. Figure 1. Energy Production (Source: TURKSTAT) Natural Gas 0 0 200 Production - Electricity 50000 100000 150000 Production - Natural Gas 400 600 800 1000 200000 Electricity 1923 1933 1943 1953 1963 1973 Years 1983 1993 2003 1976 2013 1986 1996 2006 Years Hydraulic 0 0 1000 Production - Hydraulic 1000 2000 3000 Production - Petroleum 2000 3000 4000 4000 5000 Petroleum 1950 1960 1970 1980 Years 1990 2000 2010 1950 1960 1970 1990 2000 2010 Lignite 0 1000 1500 Production - Lignite 5000 10000 Production - Coal 2000 2500 3000 15000 Coal 1980 Years 1950 Volume 3 1960 1970 Number 2 1980 Years 1990 2000 Fall 2013 2010 1950 1960 1970 1980 Years 1990 2000 2010 75 Özgür Polat / Enes E. Uslu / Hüseyin Kalyoncu Geothermal 0 3000 200 3500 Production - Wood 4000 4500 Production - Geothermal 400 600 800 5000 5500 1000 Wood 1950 1960 1970 1980 Years 1990 2000 2010 1963 1973 1983 Years 1993 2003 Time series plot for consumption of energy variables of Turkey are shown in Figure 2. Only consumption of geothermal and wood series for energy usage significantly decrease among all energy variables. The decrease in wood consumption is consistent with its decrease in production, owing to alternative energy resources such as natural gas production. The increase in consumption of electricity, natural gas and petroleum are remarkable compared to other variables, and indicate the importance of these energy resources for economy in Turkey. Although the price of natural gas in Turkey is the highest in the world (Altunsoy, 2008), the remarkable increase in its consumption indicates it is still cheaper than other energy resources in Turkey. Figure 2. Energy Consumption (Source: MENR) Natural Gas 0 0 Consumption - Electricity 50000 100000 Consumption - Natural Gas 10000 20000 30000 150000 40000 Electricity 1923 76 1933 1943 1953 1963 1973 Years 1983 1993 2003 2013 1975 1985 1995 Years 2005 Journal of Economic and Social Studies Unit Root Properties of Energy Consumption and Production in Turkey Analysis for Selected OECD Countries Hydraulic 0 10000 Consumption - Hydraulic 1000 2000 3000 Consumption - Petroleum 15000 20000 25000 30000 4000 35000 Petroleum 1970 1980 1990 Years 2000 1970 2010 1980 2000 2010 2000 2010 2000 2010 Coal 0 0 Consumption - Coal 5000 10000 Consumption - Lignite 5000 10000 15000 15000 Lignite 1990 Years 1970 1980 1990 Years 2000 2010 1970 1980 Geothermal 1990 Years 0 3500 200 Consumption - Geothermal 400 600 800 Consumption - Wood 4000 4500 5000 1000 5500 Wood 1970 1980 1990 Years 2000 2010 1970 1980 1990 Years Time series plot for sectorial consumption of energy variables of Turkey are shown in Figure 3. The increase in energy consumption in industry indicates how the importance of industry has increased in the economy. At the end of 1990s, use of petroleum decrease significantly in industry. When compared to other energy resources, the significant increase in natural gas consumption in industry indicates a substitution between energy resources because of increasing oil prices and energy Volume 3 Number 2 Fall 2013 77 Özgür Polat / Enes E. Uslu / Hüseyin Kalyoncu policies promoting natural gas consumption. Total energy consumption in every sector increased with a positive trend indicating the rapid growth in the Turkish economy in the last decade. Structural breaks are clear in 1994, 1999, 2001 and 2008 when economic crises occurred. Figure 3. Sectorial Energy Consumption (Source: MENR) Industrial Petroleum 0 2000 Consumption - Industrial 10000 20000 Consumption - Industrial (Petroleum) 4000 6000 8000 30000 10000 Total Industrial 1970 1980 1990 Years 2000 2010 1970 2000 2010 Industrial Natural Gas 0 Consumption - Industrial (Natural gas) 2000 4000 6000 Consumption - Industrial (Electricity) 2000 4000 6000 0 1970 1980 1990 Years 2000 1975 2010 1985 1995 Years 2005 0 0 Consumption - Transportation 5000 10000 15000 Consumption - Transportation (Petroleum) 5000 10000 15000 20000 Transportation Petroleum 20000 Total Transportation 1970 78 1990 Years 8000 Industrial Electricity 1980 1980 1990 Years 2000 2010 1970 1980 1990 Years 2000 2010 Journal of Economic and Social Studies Unit Root Properties of Energy Consumption and Production in Turkey Analysis for Selected OECD Countries Total Residental 0 10000 Consumption - Residential 15000 20000 25000 Consumption - Transportation (Electricity) 20 40 60 80 30000 Transportation Electricity 1970 1980 1990 Years 2000 1970 2010 1990 Years 2000 2010 Residental Electricity 0 Consumption - Residential (Electricity) 2000 4000 6000 Consumption - Residential (Petroleum) 2000 4000 6000 8000 8000 10000 Residental Petroleum 1980 1970 1980 1990 Years 2000 1970 2010 1990 Years 2000 2010 Agriculture Petroleum 0 0 Consumption - Agricultural 1000 2000 3000 4000 Consumption - Agricultural (Petroleum) 1000 2000 3000 4000 5000 5000 Total Agriculture 1980 1970 1980 1990 Years 2000 2010 1970 1990 Years 2000 2010 2000 2010 Total Other Sectors 0 10000 Consumption - Other sectors 15000 20000 25000 30000 Consumption - Agricultural (Electricity) 100 200 300 400 500 35000 Agriculture Electricity 1980 1970 Volume 3 1980 Number 2 1990 Years 2000 Fall 2013 2010 1970 1980 1990 Years 79 Özgür Polat / Enes E. Uslu / Hüseyin Kalyoncu Other Sectors Petroleum 0 2000 Consumption - Other sectors (Electricity) 2000 4000 6000 8000 Consumption - Other sectors (Petroleum) 4000 6000 8000 10000 Other Sectors Electricity 1970 1980 1990 Years 2000 2010 1970 1980 1990 Years 2000 2010 Econometric Methodology The LS unit root test is based on Lagrangian Multiplier (LM) for trending data. J. Lee and Strazicich (2003, 2004) extended Schmidt and Phillips (1992) methodology by considering structural breaks. The form of the test allows endogenous determination of two structural breaks under both the null and alternative hypotheses for a change in both the level and trend. DYt =δ ′DZt + ϕ St −1 + ε t , (5) where Z t = [1, t , D1t , D2t , DT1t , DT2t ] is a vector of exogenous variables, δ = [ µ , γ , d1 , d 2 , d3 , d 4 ] is a parameter vector of Zt and the subsequent dummies, which allow two time changes in the level and trend, are as follows: D jt (6) 1 t ≥ TBj + 1 and DT jt  = 0 t < TBj + 1 t − TBj t ≥ TBj + 1 , j 1, 2. = t < TBj + 1  0 ψ x= Y1 − Z1δ and St =Yt −ψ x − Ztδ where δ are coefficients in the regression of DYt on DZ t . The null and alternative hypotheses are: H 0 : β − 1 = ϕ = 0 vs H1 : β − 1 = ϕ < 0 80 (7) Journal of Economic and Social Studies Unit Root Properties of Energy Consumption and Production in Turkey Analysis for Selected OECD Countries λ (= λ1 TB1 / T , λ= TB 2 / T ) ) LS To determine the location of the breaks (= 2 procedure utilizes a grid search as follows: LMτ = inf τλ ( λ ) (8) Break points are where the corresponding test statistic is minimal. Results The results of the unit root tests with one and two structural breaks and without structural break are presented in Table 2. Three distinct unit root tests are used in this study to distinguish the impacts of structural break(s) on the energy series. We considered breaks at level and trend of the series. The number of lags is determined according to the general to specific method up to specific number of maximum lag1 running by t-statistics significance at the 10% significance level. The unit root hypothesis is rejected only for consumption of natural gas by conventional unit root tests without structural break. The LS unit root test with one structural break rejected the unit root hypothesis for 15 out of the 33 series. When two structural breaks are taken into account, 25 out of the 33 series are found stationary. This series is stationary around deterministic trend with breaks. The production of hydraulic and the consumption of lignite, electricity, petroleum, and coal, total energy consumption in the transportation sector and consumption of petroleum in the transportation sector are found to be non-stationary. According to the results, structural breaks in energy variables of Turkey should be taken into consideration when the unit root properties are examined. If the time series of the variable with structural breaks are tested by conventional unit root tests, the unit root hypothesis may be not cannot rejected (Perron, 1989). Our results verify the theory that the number of rejection of unit root null hypothesis declines when the number of structural breaks is increased. 1 The number of maximum lag depends on number of observation of the series. Volume 3 Number 2 Fall 2013 81 Özgür Polat / Enes E. Uslu / Hüseyin Kalyoncu Table 2. Results of unit root tests Series PHC PGL PLT PWD PPM PCL PNG PEY CHC CGL CLT CWD CPM CCL CNG CEY IND IND_PET IND_ECT IND_NGS TPT TPT_PET TPT_ECT OSC OSC_PET OSC_ECT RES RES_PET RES_ECT ACL ACL_PET ACL_ECT NEY SP k 0 0 2 8 8 0 8 9 0 3 9 2 6 9 8 11 0 5 6 0 0 0 5 0 0 0 0 0 0 0 7 8 1 t statistics -1.52 -2.12 -1.45 -0.73 1.83 -1.99 -2.16 -0.45 0.14 -1.94 -1.93 -0.53 -0.93 -1.80 -3.13c -0.08 -2.22 -1.72 1.45 -1.53 -1.68 -2.56 -2.07 -1.40 -1.15 -0.98 -1.46 -1.07 -1.11 -1.64 -2.73 -1.94 -3.74a LS - one break k t statistics 6 -3.87 3 -3.58 2 -3.44 9 -4.20c 7 -3.37 9 -4.73b 8 -6.65a 9 -3.68 0 -2.33 2 -5.74a 9 -3.87 5 -4.26c 10 -4.30c 10 -2.94 5 -5.76a 11 -4.68b 5 -5.03b 9 -4.14 0 -3.98 4 -3.19 0 -4.11 3 -4.25c 9 -3.95 0 -2.53 1 -4.78b 9 -3.94 0 -2.42 9 -4.97b 9 -3.95 4 -3.61 7 -4.48c 8 -4.95b 2 -5.01b TB 1999 2006 1991 1990 1988 1994 1994 1987 1966 1989 1998 1989 1991 1989 1987 1981 2000 2003 1985 1993 1997 1997 2002 2001 2000 1985 2001 1993 1985 1995 1994 1992 2003 LS - two breaks k t statistics 6 -5.04 9 -8.71a 3 -6.35a 6 -5.79b 7 -5.40c 8 -6.46a 2 -16.39a 9 -4.51 7 -5.39b 2 -6.88a 9 -4.62 10 -6.18b 6 -4.65 5 -5.06 2 -9.17a 11 -4.87 5 -5.69b 9 -6.41a 5 -5.97b 6 -8.61a 8 -4.68 8 -5.15 6 -6.99a 6 -8.54a 2 -7.47a 9 -6.40a 6 -8.44a 9 -11.48a 9 -7.02a 7 -6.82a 7 -5.39c 9 -5.64b 3 -6.60a TB1 1983 1987 1987 1984 1988 1988 1988 1944 1968 1975 1979 1972 1965 1975 1988 1981 1991 1989 1984 1994 1987 1987 1986 1982 1996 1982 1982 1987 1982 1993 1993 1994 1997 TB2 1993 2006 2001 1996 1994 1994 2003 1973 1993 1987 1999 1990 2002 1993 1990 1989 2000 1994 2000 1999 1991 1991 2002 1999 2000 1995 1999 1993 1995 2006 1999 2006 2003 Notes: k indicates the number of lags. a, b and c denote significance at the 1% 5% and 10% level, respectively. TB denotes time breaks. 82 Journal of Economic and Social Studies Unit Root Properties of Energy Consumption and Production in Turkey Analysis for Selected OECD Countries Conclusion Specification of unit root properties of energy consumption and production is crucial for energy policy formulations and implementations. The impact of shocks on energy variables with a stationary process will be temporary and long short term, while impact of shocks on energy variables with a nonstationary process will be permanent and have a long memory. In this study, the unit root properties of total and sectorial energy production and consumption series of Turkey are investigated. This study is the first to investigate unit root properties of Turkish energy production and consumption in detail. The unit root structure for energy variables are tested by using the unit root tests based on LM without structural break and with one and two structural breaks. The results of unit root test without structural break show that the unit root hypothesis is rejected only for consumption of natural gas. In the case of one structural break, the unit root hypothesis is rejected for 15 out of the 33 series by LS test. When two structural breaks are taken into account, 25 out of the 33 series are found to be stationary around a deterministic trend with breaks. The production of hydraulic, the consumption of lignite, electricity, petroleum, coal, electricity, total energy consumption and petroleum consumption in the transportation sector are found to be non-stationary, which indicates that the impact of innovations on these variables will be permanent. The policy implication of these results suggests that the impacts of shocks on energy consumption and production will be temporary and not have a long memory for most of the variables. Therefore, the economic impact of energy stabilization and conservation policies will be temporary in Turkey. The results of this study, which found that most of the variables are stationary, are consistent the consensus about stationarity of energy variables found in numerous other studies (Narayan et al., 2010). In addition, the historical data on these stationary variables can be taken into account to forecast the future values of these variables. Volume 3 Number 2 Fall 2013 83 Özgür Polat / Enes E. Uslu / Hüseyin Kalyoncu References Altunsoy, İ. (2008, 1 May 2008). Turkey consumes most expensive gas in the world, TodaysZaman. Retrieved from http://www.todayszaman.com/newsDetail_getNewsById. action?load=detay&link=140592 Arouri, M. E. H., Lahiani, A., & Nguyen, D. K. (2011). Return and volatility transmission between world oil prices and stock markets of the GCC countries. Economic Modelling, 28(4), 1815-1825. doi: 10.1016/j.econmod.2011.03.012 Aslan, A., & Kum, H. (2011). The stationary of energy consumption for Turkish disaggregate data by employing linear and nonlinear unit root tests. Energy, 36(7), 4256-4258. doi: 10.1016/j.energy.2011.04.018 Ayadi, O. F. (2005). Oil price fluctuations and the Nigerian economy. OPEC Review, 29(3), 199-217. doi: 10.1111/j.0277-0180.2005.00151.x Barros, C. P., Gil-Alana, L. A., & Payne, J. E. (2011). An analysis of oil production by OPEC countries: Persistence, breaks, and outliers. Energy Policy, 39(1), 442-453. doi: 10.1016/j.enpol.2010.10.024 Basher, S. A., Haug, A. A., & Sadorsky, P. (2012). Oil prices, exchange rates and emerging stock markets. Energy Economics, 34(1), 227-240. doi: 10.1016/j.eneco.2011.10.005 Carruth, A. A., Hooker, M. A., & Oswald, A. J. (1998). Unemployment Equilibria and Input Prices: Theory and Evidence from the United States. Review of Economics and Statistics, 80(4), 621-628. doi: 10.1162/003465398557708 Chang, Y., & Wong, J. F. (2003). Oil price fluctuations and Singapore economy. Energy Policy, 31(11), 1151-1165. doi: 10.1016/s0301-4215(02)00212-4 Chen, P. F., & Lee, C. C. (2007). Is energy consumption per capita broken stationary? New evidence from regional-based panels. Energy Policy, 35(6), 3526-3540. doi: 10.1016/j.enpol.2006.12.027 Chen, S. S., & Chen, H. C. (2007). Oil prices and real exchange rates. Energy Economics, 29(3), 390404. doi: 10.1016/j.eneco.2006.08.003 Chiou Wei, S. Z., & Zhu, Z. (2002). Sources of export fluctuations: empirical evidence from Taiwan and South Korea, 1981–2000. Journal of Asian Economics, 13(1), 105-118. doi: 10.1016/s10490078(01)00114-2 Cologni, A., & Manera, M. (2008). Oil prices, inflation and interest rates in a structural cointegrated VAR model for the G-7 countries. Energy Economics, 30(3), 856-888. doi: 10.1016/j.eneco.2006.11.001 Cuñado, J., & Pérez de Gracia, F. (2003). Do oil price shocks matter? Evidence for some European countries. Energy Economics, 25(2), 137-154. doi: 10.1016/s0140-9883(02)00099-3 Du, L., Yanan, H., & Wei, C. (2010). The relationship between oil price shocks and China’s macroeconomy: An empirical analysis. Energy Policy, 38(8), 4142-4151. doi: 10.1016/j.enpol.2010.03.042 Evangelia, P. (2001). Oil price shocks, stock market, economic activity and employment in Greece. Energy Economics, 23(5), 511-532. doi: 10.1016/s0140-9883(01)00078-0 84 Journal of Economic and Social Studies Unit Root Properties of Energy Consumption and Production in Turkey Analysis for Selected OECD Countries Ewing, B. T., & Thompson, M. A. (2007). Dynamic cyclical comovements of oil prices with industrial production, consumer prices, unemployment, and stock prices. Energy Policy, 35(11), 5535-5540. doi: 10.1016/j.enpol.2007.05.018 Faria, J. R., Mollick, A. V., Albuquerque, P. H., & León-Ledesma, M. A. (2009). The effect of oil price on China’s exports. China Economic Review, 20(4), 793-805. doi: 10.1016/j.chieco.2009.04.003 Filis, G., Degiannakis, S., & Floros, C. (2011). Dynamic correlation between stock market and oil prices: The case of oil-importing and oil-exporting countries. International Review of Financial Analysis, 20(3), 152-164. doi: 10.1016/j.irfa.2011.02.014 Hamilton, J. D. (1996). This is what happened to the oil price-macroeconomy relationship. Journal of Monetary Economics, 38(2), 215-220. doi: 10.1016/s0304-3932(96)01282-2 Hamilton, J. D. (2007). Oil and the Macroeconomy. In S. N. Durlauf & L. E. Blume (Eds.), New Palgrave Dictionary of Economics. Hendry, D. F., & Juselius, K. (2000). Explaining cointegration analysis, part1. Energy Journal, 21, 1–42. Hsu, Y. C., Lee, C. C., & Lee, C. C. (2008). Revisited: Are shocks to energy consumption permanent or temporary? New evidence from a panel SURADF approach. Energy Economics, 30(5), 23142330. doi: 10.1016/j.eneco.2007.09.007 Huang, B. N., Hwang, M. J., & Peng, H. P. (2005). The asymmetry of the impact of oil price shocks on economic activities: An application of the multivariate threshold model. Energy Economics, 27(3), 455-476. doi: 10.1016/j.eneco.2005.03.001 Jayaraman, T. K., & Choong, C. K. (2009). Growth and oil price: A study of causal relationships in small Pacific Island countries. Energy Policy, 37(6), 2182-2189. doi: 10.1016/j.enpol.2009.01.025 Jiménez-Rodríguez, R. (2008). The impact of oil price shocks: Evidence from the industries of six OECD countries. Energy Economics, 30(6), 3095-3108. doi: 10.1016/j.eneco.2008.06.002 Kim, I. M., & Loungani, P. (1992). The role of energy in real business cycle models. Journal of Monetary Economics, 29(2), 173-189. doi: 10.1016/0304-3932(92)90011-p Lardic, S., & Mignon, V. (2008). Oil prices and economic activity: An asymmetric cointegration approach. Energy Economics, 30(3), 847-855. doi: 10.1016/j.eneco.2006.10.010 Lee, B. R., Lee, K., & Ratti, R. A. (2001). Monetary policy, oil price shocks, and the Japanese economy. Japan and the World Economy, 13(3), 321-349. doi: 10.1016/s0922-1425(01)00065-2 Lee, J., & Strazicich, M. C. (2003). Minimum Lagrange Multiplier Unit Root Test with Two Structural Breaks. Review of Economics and Statistics, 85(4), 1082-1089. doi: 10.1162/003465303772815961 Lee, J., & Strazicich, M. C. (2004). Minimum LM Unit Root Test with One Structural Break. Department of Economics, Appalachian State University. Retrieved from http://econ.appstate.edu/ RePEc/pdf/wp0417.pdf Lorde, T., Jackman, M., & Thomas, C. (2009). The macroeconomic effects of oil price fluctuations on a small open oil-producing country: The case of Trinidad and Tobago. Energy Policy, 37(7), 27082716. doi: 10.1016/j.enpol.2009.03.004 Volume 3 Number 2 Fall 2013 85 Özgür Polat / Enes E. Uslu / Hüseyin Kalyoncu Lowinger, T. C., Wihlborg, C., & Willman, E. S. (1985). OPEC in world financial markets: Oil prices and interest rates. Journal of International Money and Finance, 4(2), 253-266. doi: 10.1016/02615606(85)90047-6 Maslyuk, S., & Smyth, R. (2009). Non-linear unit root properties of crude oil production. Energy Economics, 31(1), 109-118. doi: 10.1016/j.eneco.2008.09.004 Mishra, V., Sharma, S., & Smyth, R. (2009). Are fluctuations in energy consumption per capita transitory? Evidence from a panel of Pacific Island countries. Energy Policy, 37(6), 2318-2326. doi: 10.1016/j.enpol.2009.02.022 Narayan, P. K., Narayan, S., & Popp, S. (2010). Energy consumption at the state level: The unit root null hypothesis from Australia. Applied Energy, 87(6), 1953-1962. doi: 10.1016/j.apenergy.2009.10.022 Narayan, P. K., Narayan, S., & Prasad, A. (2008). Understanding the oil price-exchange rate nexus for the Fiji islands. Energy Economics, 30(5), 2686-2696. doi: 10.1016/j.eneco.2008.03.003 Narayan, P. K., Narayan, S., & Smyth, R. (2008). Are oil shocks permanent or temporary? Panel data evidence from crude oil and NGL production in 60 countries. Energy Economics, 30(3), 919-936. doi: 10.1016/j.eneco.2007.07.005 Narayan, P. K., & Smyth, R. (2007). Are shocks to energy consumption permanent or temporary? Evidence from 182 countries. Energy Policy, 35(1), 333-341. doi: 10.1016/j.enpol.2005.11.027 Özturk, İ., Feridun, M., & Kalyoncu, H. (2008). Do Oil Prices Effect the USD/YTL Exchange Rate: Evidence from Turkey. 18(115), 48-61. Papapetrou, E. (2001). Oil price shocks, stock market, economic activity and employment in Greece. Energy Economics, 23(5), 511-532. doi: 10.1016/s0140-9883(01)00078-0 Park, J., & Ratti, R. A. (2008). Oil price shocks and stock markets in the U.S. and 13 European countries. Energy Economics, 30(5), 2587-2608. doi: 10.1016/j.eneco.2008.04.003 Perron, P. (1989). The Great Crash, the Oil Shock and the unit root hypothesis. Econometrica, 57(6), 1361–1402. Rafiq, S., Salim, R., & Bloch, H. (2009). Impact of crude oil price volatility on economic activities: An empirical investigation in the Thai economy. Resources Policy, 34(3), 121-132. doi: 10.1016/j. resourpol.2008.09.001 Sadorsky, P. (1999). Oil price shocks and stock market activity. Energy Economics, 21(5), 449-469. doi: 10.1016/s0140-9883(99)00020-1 Schmidt, P., & Phillips, P. C. B. (1992). LM Tests for a Unit Root in the Presence of Deterministic Trends. Oxford Bulletin of Economics and Statistics, 54(3), 257-287. doi: 10.1111/j.14680084.1992.tb00002.x Zhang, Q. (2011). The Impact of International Oil Price Fluctuation on China’s Economy. Energy Procedia, 5(0), 1360-1364. doi: 10.1016/j.egypro.2011.03.235 86 Journal of Economic and Social Studies Journal of Economic and Social Studies Measuring and Reporting Cost of Quality in a Turkish Manufacturing Company: A Case Study in Electric Industry Hilmi KIRLIOĞLU, Sakarya University, Faculty of Business Administration, Department of Accounting and Finance, / 54187 Esentepe/Sakarya, Turkey. hilmik@sakarya.edu.tr Zülküf ÇEVİK, Sakarya University, Faculty of Business Administration, Department of Accounting and Finance, / 54187 Esentepe/Sakarya, Turkey. zcevik@sakarya.edu.tr AbstrAct Contemporarily, the competition in the markets has thoroughly heated up. Many companies try to decrease their costs in order to survive in this cruel market. In this respects, the quality costs gain importance in all over the world and in Turkey, too.In this study, the implementation of quality costs measuring and reporting system has been performed in a company. Accordingly, the data has been collected from a Turkish manufacturing company. The data gathered from this company’s accounting department has been used for studying on quality costs measuring and reporting system.Consequently, it is found out that the company cannot measure its quality costs adequately, for this reason quality reporting system in the company is not efficient. The company needs to give more significance to the quality costs measuring and reporting. JEL codes: L15, M41, M49, Volume 3 Number 2 Fall 2013 KEYWOrDs Total Quality Management, Quality Costs, Managerial Accounting. ArtIcLE HIstOrY Submitted:23 September 2012 Resubmitted:16 November 2012 Accepted:25 December 2012 87 Hilmi KIRLIOĞLU / Zülküf ÇEVİK Introduction In recent years, competitive environment of companies has been getting harder and harder. In order to have sustainable competitive advantage, companies should produce their products to entirely supply customers’ needs, wants and demands. Subsequently, companies need to have more quality products to remain competitive with other companies. To gain a competitive advantage over rival companies, a company should produce high quality products. While producing high quality product, the company should also take into account its quality costs. Shortly, companies need to produce high quality products in a low quality costs. As a result, quality and quality costs gain vital importance for a company to survive in a highly competitive market. The significance of this study is to comprehend the necessity of the quality system for a company which operates in the global and local markets. Another gist of the study is to provide recognition of quality costs system benefit to the profit and brand name. The quality costs system causes decreasing in the production cost and increasing in the brand name which will be perceived as producing qualified products. The aims of this study are to show the importance of the quality costs for a company which competes in a highly competitive market, and also demonstrate the necessity of quality costs system so as to have high qualified product with a low quality costs. As it is well known, the quality cost is not the responsibility of a department or an individual, on the contrary, every person in an organization should be responsible for quality. Highly qualified products can be reached by the collaboration of all departments in an organization. In this sense, the main aim of this study is to demonstrate the function of accounting department in quality costing activities. Those activities can be summarized as; the measurement of quality costs, the classification of this costs and the reporting techniques of the quality costs. In this respect, the purpose of the current study is to show the importance of quality costs’ reporting. The paper contributes to the literature by documenting the concepts of quality, quality costs, and the classification of quality costs and quality costs measurement. On the basis of literature review, a case study will be handled and lastly, the analysis and results will be given in last section. 88 Journal of Economic and Social Studies Measuring and Reporting Cost of Quality in a Turkish Manufacturing Company: A Case Study in Electric Industry Literature review To be able to analyze the measuring and reporting costs of quality, some concepts should be clear first. The three basic concepts of this paper will be introduced. These concepts are quality and quality costs, classification of quality costs and lastly, quality costs measurement. The concept of quality has been defined for many quality gurus. So, there are many definitions for quality. Quality is the features of products which meet customer needs and thereby provide customer satisfaction. Quality means freedom from deficiencies (Juran & Godfrey, 1998). According to D. C. Montgomery, Quality means fitness for use, and also he defined quality as inversely proportional to variability (Montgomery, 2005). In addition to those definitions, some of other quality gurus defined quality as; - Crosby (1979, p. 7)defines quality as “conformance to requirements” - Feigenbaum’s(1983, p. 7) definition of quality is “the total composite product and service characteristics of marketing, engineering, manufacture and maintenance through which the product and service in use will meet the expectations of the customer.” - As Ishikawa (1985, p. 45) suggests, quality means “quality of work, quality of service, quality of information, quality of process, quality of division, quality of people, including workers, engineers, managers and executives, quality of system, quality of company, quality of objectives, etc.” - Pirsig’s definition (1984,p. 206) of quality is that “Quality is a characteristic of thought and statement that is recognized by a nonthinking process. because definitions are a product of rigid, formal thinking, quality cannot be defined.” To sum up those definitions, quality is the whole good and service characteristic features of fulfillment power for stated and demanded needs. In other words, many quality gurus defined quality in terms of the degree of the product’s conformance to its requirements to maintain customer satisfaction and in terms of a product that contains no defects (Ömürgönülşen, 2009). Quality Cost is a cost for detection and anchoring of low quality about goods and services. Simply, costs of quality are the costs which occur because poor quality may or does exist (Hansen & Mowen, 2006). Quality costs are a measurement of the costs particularly related with the accomplishment or non-accomplishment of Volume 3 Number 2 Fall 2013 89 Hilmi KIRLIOĞLU / Zülküf ÇEVİK product or service quality. To make those explanations more specific, Jack Campanella(1999, p. 4) defined cost of quality as; “More specifically, quality costs are the sum of the cost incurred by (a) investing in the prevention of non-conformances to requirements, (b) appraising a product or service for conformance to requirements, and (c) failing to meet requirements.” In the definitions of Campanella, it is understood that the quality costs consist of three main parts; Prevention Costs, Appraisal Costs, Failure Costs. The required quality activities would incur costs and quality costs arecategorized into three main parts – Prevention, Appraisal and Failure Costs – Those can be also stated as PAF (Prevention-Appraisal-Failure) model (Jaju & Lakhe, 2009). Failure costs should be taken into consideration as two subtopics which are called internal and external failure costs. (Figure 1. Classification of Quality Costs(Rodchua, 2006 Cost of Quality Prevention costs AppraisalCosts Failure Costs Internal Failure External Failure In Figure 1, three main classifications of quality activities costs have been shown. Those costs do not occur at the same period of the production process. So, it should be also classified as time periods in which they occurred. Prevention Costs are the preliminary activities’ costs to reach quality goals for producing goods and services and avoid deviations of those goals (Kırlıoğlu, 1998). Prevention costs are occurred to prevent low quality in the goods or services being produced (Hansen & Mowen, 2006). Prevention costs are related with quality planning, designing, implementing and managing the quality system, auditing the system, supplier surveys, and process improvements (Rodchua, 2006). Appraisal Costs are activity costs of measuring the suitability of the product to customers’ needs. It is incurred to identify non-conformance to requirements (Oliver & Qu, 1999). Those costs are related with the supplier’s and customer’s assessment of 90 Journal of Economic and Social Studies Measuring and Reporting Cost of Quality in a Turkish Manufacturing Company: A Case Study in Electric Industry purchased materials, processes, intermediates, products and services to assure conformance with the specified requirements (Tsai, 1998). Internal Failure Costs are the costs of low quality product which are realized before sales of the product. In other words, these costs arise when the outcomes of production fail to meet stated quality specifications and are noticed before transfer of those low quality products to the customers (Vahevanidis et al., 2009). External Failure Costs are failure costs which come up after delivering the products to the customers(Kaner, 1996).Those costs take place for the reason that the products and services do not conform to specification or requirements and those products do not satisfy customer needs after being delivered to customers (Hansen & Mowen, 2006). It is also incurred by amending failures after transferring the finished goods and products to the customers (Low & Yeo, 1998). Additionally, Quality cost classification can be grouped in time periods. For example, prevention costs encompass the stage of both pre-production and during production and appraisal costs cover the three stages of production –preproduction, production and after production stage. Failure costs are divided into two subtopics which internal failure costs and external failure costs. Internal failure costs encompass the period of both production and after production stages. External failure costs just related with the stage of after sale (Barfield et al., 2002). Figure 2. Time-Phased Model for Quality Costs(Barfield et al., 2002) Before Production During Production Prevention Costs After Production After Sale Appraisal Costs Internal Failure Costs External Failure Costs Feedback Loop The Quality Costs Measuring helps to find out where unnecessary quality costs are occurred, thus management can take actions to eliminate that kind of costs and this Volume 3 Number 2 Fall 2013 91 Hilmi KIRLIOĞLU / Zülküf ÇEVİK elimination will reduce the occurrence of poor quality costs. In other saying, the quality costs measurement serves management to determine which area of operation requires preventive measures (Low & Yeo, 1998). To measure quality costs, one should collect related data from quality activities of a company. After the collection of data which are related with quality costs components, they should be analyzed before using in an action. This analysis consists of the relationship between a costs component and other costs components and searches the effect on total costs. Quality costs are analyzed in weekly, monthly, quarterly, yearly, etc. periods. Company structure should be taken into account in determining the period of analysis (Şimşek, 2001). In order to analyze quality costs, companies need to use some techniques. The analysis techniques for quality costs can be listed as; I. Pareto Analysis, II. Ratio Analysis, III. Correlation Analysis, IV. Trend Analysis, V. Regression Analysis. Pareto Analysis is one of the most used techniques in quality costs analysis. This technique was developed by Wilfredo Pareto who is a nineteenth century Italian social scientist and economist. He gave his surname to the technique. Pareto principle is universally known as the 80/20 rule. Pareto found out this principle by pinning down that 80 percent of Italy’s national income is shared by 20 percent of the Italy’s populations. With the help of Pareto diagrams, problems can be put in order of importance, problems of costs analysis can be easily performed and relative occurrence numbers could be searched simply (Sarıkaya, 2003). In other words, Pareto analysis can be utilized to recognize cost drivers which are accountable for the most of cost occurred by ranking the cost drivers in order of value (Tsai, 1998). The Technique of Ratio Analysis is aimed to identify the aspects of the quality costs’ performance to aid decision making.Ratio analysis consist of rationing quality costs to revenue, production costs, direct labor costs and rationing total quality costs within themselves (Özcan, 2012). Correlation analysis represents the direction and the power of the relationship between variables. In correlation analysis, the results do not give cause-effect relation- 92 Journal of Economic and Social Studies Measuring and Reporting Cost of Quality in a Turkish Manufacturing Company: A Case Study in Electric Industry ship, because there is no dependent and independent variable in this technique (Altunışık et al., 2005). Trend Analysisis a useful picture of how the quality improvement program has been doing since its inception. It provides management with information concerning the within-period progress measured relative to specific goals (Hansen & Mowen, 2006). Regression analysis examines the relationship between one dependent variable and one or more than one independent variables. In other words, this technique tries to explain the changes in dependent variable with the help of independent variables (Altunışık et al., 2005). Data and Methodology The data of this study is gathered from X Electric Inc. Company which was founded in 1990 in Adapazarı, Turkey. The company is a Low Voltage Circuit Breaker manufacturer company. The data has been collected from this company’s accounting department. The company’s accounting director gave the raw data of the company quality costs. We have analyzed these costs for reporting quality costs. According to the data, we drew table 1, 2, and 3. With the help of these tables, we made ratio and trend analysis of company’s quality costs. The data consists three years which are 2008, 2009, 2010. In the study, trend and ratio analysis have been performed for measuring and reporting the firm quality costs. Analysis of Quality Costs in the Firm The table below displays the company’s sales and production amount in Turkish Liras (here after TL). The sales and production amount have been given for three years. Additionally, the table contains of total quality costs in the firm for three years. Volume 3 Number 2 Fall 2013 93 Hilmi KIRLIOĞLU / Zülküf ÇEVİK Table 1. Some Ratios and Ratio Components in the Firm Years Data 2008 2009 2010 Total Sales (TL) 629.053.415 695.866.750 786.859.486 Total Production Costs (TL) 515.326.274 563.708.245 643.590.468 Total Quality Costs (TL) 10.028.516 11.712.822 12.642.655 1,59% 1,95% 1,68% 2,08% 1,61% 1.96% The Ratio of Quality Costs to Sales The Ratio of QC to Production Costs According to the firm information, the ratios of total quality costs to total sales have been calculated for given three years. And the ratios of total quality costs to total production costs have also been calculated. In the aspect of the information in the previous section, these calculations have been performed as follows. In 2008, the company’s total sales are 629.053.415 TL. In the same year, total quality costs are 10.028.516 TL. So the ratio of total quality costs to sales can be found out as follows; 10.028.516 629.053.415 = 1,59% It can be concluded that the amount of total quality costs is only 1.59% of the total sales in 2008. 10.028.516 515.326.274 = 1,95% The calculation above demonstrates that the ratio of total quality costs to total production costs is about 1.95%. For the years of 2009 and 2010, total quality costs to sales and total quality costs to total production costs have been calculated by the same way and written down in the figure above. This ratio is not too much for an early stage of quality costs analysis applier’s company. In other words, the firm analyses its quality costs not long ago, so the rates is in the acceptable limits. Besides this ratios can be reduced for more efficient quality costs system. When analyzing quality costs data for year 2010 as quality costs components, it will be useful for monitoring quality costs. Regarding this classification, quality costs component will be given as costs items. With the help of this costs items, the percentage amount of each costs item will be also calculated and given for the year. 94 Journal of Economic and Social Studies Measuring and Reporting Cost of Quality in a Turkish Manufacturing Company: A Case Study in Electric Industry Table 2. Total Quality Costs as Each Cost Items for the Components in the year of 2010 Components of Quality Costs Prevention Costs Quality Planning Quality Circle The Training of Quality Inspection and Tests Instructions Supplier Quality Planning Preventive Maintenances Other Prevention Costs Appraisal Costs Inspection and tests of purchased materials Costs (TL) 1.782.614,36 518.348,86 75.855,93 202.282,48 113.783,90 214.925,14 480.420,89 176.997,17 5.031.776,69 1.036.697,71 Ratio (%) 14,1 4,1 0,6 1,6 0,9 1,7 3,8 1,4 39,8 8.2 Control, maintenance and calibration of measurement instruments 101.141,24 0,8 Process inspection and tests Consumable materials for laboratory and tests Products inspection and tests Other appraisal costs Internal Failure Costs Salvage Reproduction and Repairs Re-inspection Corrective actions External Failure Costs Products Returns Transportation Damage Warranty Costs 1.150.481,61 581.562,13 1.984.896,84 176.997,17 4.450.214,56 2.225.107,28 1.656.187,81 480.420,89 88.498,59 1.378.049,40 998.769,75 50.570,62 328.709,03 9,1 4,6 15,7 1,4 35,2 17,6 13,1 3,8 0,7 10,9 7,9 0,4 2,6 Total Quality Costs 12.642.655 100,0 In the figure 2010, the non-conformance costs are under the half of the total quality costs. This demonstrates that the firm is going in the right way. The company gives more importance for conformance costs day by day, so the non-conformance costs decreases naturally. These changes will benefit the company in more ways than one. The figures below should be reported to the managers for monitoring quality costs activates by management. The importance of quality costs increases day by day. The next table shows the total quality costs as categorization groups. As it is mentioned before, the quality costs have two components which are conformance and non-conformance costs. And these components costs are given in the chart. Moreover, conformance costs are divided into two cost elements that are prevention costs and appraisal costs. The amounts of these costs are also given yearly in the table. The Volume 3 Number 2 Fall 2013 95 Hilmi KIRLIOĞLU / Zülküf ÇEVİK conformance costs are increasing for the given years. It rose up two times from the amount of 2008 to 2010 amount. It is good for a company to increase its prevention activities in order not to confront defects after selling the products out. Besides, the amount of prevention costs in conformance costs is too small. The firm should concentrate more on prevention costs. Table 3. The Amount of Quality Costs in Classification through the Years Quality Costs Conformance Costs Prevention Costs Appraisal Costs Non-Conformance Costs Internal Failure Costs External Failure Costs Total Quality Costs 2008 3.098.811,45 631.796,51 2.467.014,94 6.929.704,56 4.693.345,49 2.236.359,07 10.028.516 2009 5.177.067,33 1.147.856,56 4.029.210,77 6.535.754,67 4.767.118,55 1.768.636,12 11.712.822 2010 6.814.391,05 1.782.614,36 5.031.776,69 5.828.263,96 4.450.214,56 1.378.049,40 12.642.655 On the other hand, in the table above, the non-conformance costs have been shown in two parts which are internal and external failure costs. The company has endured too much internal failure costs. And, the company should increase its preventive activities and decrease the internal failure costs. When it comes to external failure costs, the firm is going in a right way, because the amounts of external failure costs are going down for each given year. With the help of Table 3, it can be seen that conformance costs – prevention and appraisal costs – are increasing for each year. Additionally, non-conformance costs – internal and external failure costs – are decreasing for each year. It also shows that the huge amounts of total quality costs are occurred after production stage. The internal failure costs are the biggest costs in the total quality costs for every year. This situation represents that the defects are realized after the stage of production. Figure 3. The Trends of Quality Costs’ Categorization in percentage 96 Journal of Economic and Social Studies Measuring and Reporting Cost of Quality in a Turkish Manufacturing Company: A Case Study in Electric Industry In general, the movements of quality costs components are in a right way, even though the non-conformance costs are more than conformance costs. In the chart, it can also be seen that the amount of prevention costs is under the 15% of the total quality costs which means the firm do not pay enough importance for the prevention activities. Although the trend of external failure costs is declining, the external failure costs have too much portion of total quality costs. Having too much external failure costs brings more costs than the firm can measure. Figure 4. The Trends of Quality Costs’ Categorization in TL In Chart 2, it is again shown the trend of quality costs components. Chart 1 shows the trends as percentage value; Chart 2 shows these trends as Turkish Liras amounts. The inferences of the Chart 2 are similar to Chart 1. The Application Results The quality costs activities in X Electric Inc. are concentrated in non-conformance activities. In other words, the firm is highly interested in internal and external costs. So, non-conformance costs are monthly reported to management. Besides, the firm does not give required importance for prevention and appraisal costs’ measurement. Therefore, conformance costs are just reported yearly period, even though the firm is giving more importance to conformance costs than before. On the other hand, while paying the non-conformance costs more importance than the conformance costs in the firm, the company endures more costs than it can measure. That is to say, the firm can bear the quality costs more than in numbers; there may be a non-visual negative effect on the firm. For instance, the firm may confront the loss of customers, bad brand recognition and poor employee motivation and so Volume 3 Number 2 Fall 2013 97 Hilmi KIRLIOĞLU / Zülküf ÇEVİK on. Furthermore, the efficient quality system causes to benefit the company in more ways than one. It decreases the non-conformance costs and increases profitability of the firm. The firm would have sustainable competitive advantage in the market. It is found out that the company cannot measure its quality costs adequately, for this reason quality reporting system in the company is not efficient. The company needs to give more importance to the quality costs measuring and reporting activities. Conclusion Previously, the company thought that quality control was just a waste of time. But this thought changed in course of time. With the help of effective quality control system, company can reduce the salvage, loss of labor hours and so on, which increase the productivity level. Yearly reporting of total quality costs is not efficient for making decisions on these costs. The measurement of prevention and appraisal costs is not made appropriately. The allocation key is generally labor costs which are not suitable for measurement of every costs item in the quality costs. The firm should establish an efficient quality costs system and determine this system specification for effective measurement and reporting of quality costs. By having reliable and sufficient data in quality costing, company can reduce its non-conformance costs. And it causes to reduce total quality costs and increase profitability of the firm. The firm should prepare instructions and procedures for making measurement more efficient especially in prevention and appraisal costs. Every person in the firm should be informed about these instructions and procedures. On the other hand, only the quality assurance department is responsible for quality costing in the firm. As it is mentioned in the previous parts, quality is not a person or a department job; it should be responsibility of every person and every department in the firm. The quality costs reports should be prepared and reported monthly. The accounting department should determine more suitable allocation keys for the measurement of quality costs and according to this measurement, the accountants in the firm should make the necessary journal entries for these costs. 98 Journal of Economic and Social Studies Measuring and Reporting Cost of Quality in a Turkish Manufacturing Company: A Case Study in Electric Industry By measuring and reporting quality costs, the managers can recognize that there is a huge amount of costs which they do not take into consideration while making managerial decisions. They can realize that the non-quality issues increase the evitable costs by too much. In the short run, investing in preventing activities can increase total quality costs in the firm, but in the long run, these investments will cause decreasing in failure costs. So, the firm will reduce its evitable costs in the long run. In the globalizing world, the firm should take the products quality into account. Also, it needs to be noted here that a company cannot survive in a highly competitive market with its low quality products. Last, but not the least, it must be kept in mind that the amounts of quality costs never excess the amount of poor quality costs. References Altunışık, R., Coşkun, R., Bayraktaroğlu, S., & Yıldırım, E. (2005). Sosyal Bilimlerde Araştırma Yöntemleri: SPSS Uygulamalı. Sakarya: Sakarya Kitapevi. Barfield, J., Raiborn, C., & Kinney, M. (2002). Cost Accounting: Traditions & Innovations. South-Western College Pub. Campanella, J. (1999). Principles of quality costs: principles, implementation and use. American Society for Quality press. Crosby, P. B. (1979). Quality is Free. New York: McGraw-Hill Book Co. Feigenbaum, A. V. (1983). Total Quality Control. New York: McGraw-Hill Book Co. Hansen, D. R., & Mowen, M. M. (2006). Cost Management: Accounting and Control. Thomson SouthWestern. Hoyer, R. W., & Hoyer, B. B. (2001, July). What Is Quality. Quality Progress, 53-62. Ishikawa, K. (1985). What is Total Quality Control? The Japanese Way. Englewood Cliffs, New Jersey: Prentice-Hall Inc. Jaju, S. B., & Lakhe, R. R. (2009). Quality costs in a manufacturing industry: a gateway for improvement. International Journal of Applied Engineering Research , 945-954. Juran, J. M., & Godfrey, A. B. (1998). Juran’s Quality Handbook. New York: McGraw-Hill Professional. Kaner, C. (1996). Quality cost analysis: Benefits and risks. Software QA, 23. Volume 3 Number 2 Fall 2013 99 Hilmi KIRLIOĞLU / Zülküf ÇEVİK Kırlıoğlu, H. (1998). Kalite Maliyetleri Muhasebesi. Değişim Yayınları. Low, S. P., & Yeo, H. K. (1998). A construction quality costs quantifying system for the building industry. International Journal of Quality & Reliability Management, s. 329 - 349. Montgomery, D. C. (2005). Introduction to Statistical Quality Control. New York: John Wiley & Sons. Oliver, J., & Qu, W. (1999). Cost of quality reporting: Some Australian evidence. International Journal of Applied Quality Management, 233-250. Ömürgönülşen, M. (2009, May). A research on the measurement of quality costs in the Turkish food manufacturing industry. Total Quality Management & Business Excellence, 547-562. Özcan, S. (2012). An Empirical Research about Quality Cost Analysis and Its Impact on Managerial Decisions in Automotive Supplier Industry. International Conference on Management, Applied and Social Sciences (ICMASS’2012) March 24-25,(363-367. Dubai. Pirsig, R. M. (1974). Zen and the Art of Motorcycle Maintenance. New York: William Morrow & Co. Rodchua, S. (2006, Oct-Dec). Factors, Measures, and Problems of Quality Costs Program Implementationin the Manufacturing Environment. Journal of Industrial Technology, 510-515. Sarıkaya, N. (2003) Toplam Kalite Yönetimi. 1. Baskı. Sakarya: Sakarya Kitabevi. Şimşek, M. (2001). Toplam Kalite Yönetimi. İstanbul: Alfa Yayınları. Tsai, W. -H. (1998). Quality Cost Measurement Under Activity-Based Costing. International Journal of Quality and Reliability Management, 719-752. Vahevanidis, N. M., Petropouolos, G., Avakumovic, J., & Mourlas, A. (2009). Cost Of Quality Models And Their Implementation In Manufacturing Firms. International Journal for Quality Research, 27-36. 100 Journal of Economic and Social Studies Journal of Economic and Social Studies Social Anxiety and Usage of Online Technological Communication Tools among Adolescents Bilal Sisman Economics and Administrative Science Faculty Afyon Kocatepe University, Turkey bilalsis@hotmail.com Sinan Yoruk Education Faculty, Department of Education Science Afyon Kocatepe University, Turkey syoruk@aku.edu.tr Ali Eleren Economics and Administrative Science Faculty Afyon Kocatepe University, Turkey aeleren@hotmail.com AbstrAct With the growing popularity of Internet communication among adolescents, the Internet, social media, instant messaging and cell phones have become important social tools in their life. This study examines teens’ use of social interactive technologies and the role that social anxiety plays on how adolescents communicate with others (technology or face-to-face). A questionnaire was designed and distributed to selected sample in the cities of Afyonkarahisar, Manisa and Uşak in order to analyze the relationship between adolescents’ social anxiety and their preference of communication tool. The data were gathered from 544 respondents among High School adolescents (ranged from 15-18; freshman, sophomore, junior and senior). Findings show that adolescents rarely use messenger sites and mail addresses. They generally send instant messages with their cell phones. They spend 1-2 hours for listening music and averagely 30 minutes for Facebook in a day. More than half of teens have hi-tech cell phones that enable to call, send message and access to Internet. The findings of the present study also reveal that females use text messaging more than males. However, males spend much more time than females to play games. In addition, females feel themselves more uncomfortable than males for face-to-face talking with others. And, on the contrary to males, females also prefer to some extent, to communicate with other on internet instead of face-to-face talking. Similarly, females prefer more than males to make new on internet. KEYWOrDs Social Anxiety, Communication Tools, Technology, Adolescents ArtIcLE HIstOrY Submitted: 22. Jun 2012 Resubmitted: 25 September 2012 Resubmitted: 9 October 2012 Accepted: 22 November 2012 JEL codes: C12, I12 Volume 3 Number 2 Fall 2013 101 Bilal Sisman / Sinan Yoruk / Ali Eleren Introduction The use of socially interactive technologies, such as social media, text messaging or instant messaging rises among young people (Pierce, 2009). Among adolescents, the Internet has become indispensable for instrumental purposes such as school work information gathering as well as for communication purposes. The communication applications of Internet, such as e-mail, instant messaging (IM), blogs and chat rooms have entrenched themselves in the lives of adolescents and Internet has become an important social context in the lives of adolescents today. In fact, a national survey of adolescents (10-17 years of age) revealed that in the year before they were surveyed 25% of Internet users had formed casual online friendships (Wolak et al., 2002; Subrahmanyam and Lin, 2007). Instant messaging is also an increasingly popular form of communication. A study by the Pew Internet and American Life Project found that 53 million Americans use IM, and 36% of these users reported using IM on a daily basis (Shiu and Lenhart, 2004). IM is a text based form of communication in which two or more people exchange text messages in real time using the Internet (Lenhart, 2005). Some researches suggest that socially interactive technologies allow users to avoid or replace face-to-face communication (Nie and Erbring, 2000). If the person is shy (socially anxious) and feels uncomfortable with face-to-face interactions, these technologies may serve as a useful tool for avoiding such unpleasant situations and therefore may replace face-to-face communication (Pierce, 2009). US, Canadian and Dutch studies have shown that the vast majority of adolescents spend several hours daily online. Further surveys show that adolescents consider the Internet a highly important medium in their everyday social life and use it to form and maintain social relationships (Selfhout et al., 2009). Motivation Today, social anxiety, social phobia and shyness are features that seen every individuals of all ages. Both adolescents and adults who have social anxiety against life and their environment scope, has led to more research on them. This study is related on teens 102 Journal of Economic and Social Studies Social Anxiety and Usage of Online Technological Communication Tools among Adolescents that high school level about 15-18. The beginning of puberty and the end of puberty is the reason why we selected this age range. Because, puberty is a vital phase for growing of young people and determine their personalities. This study is not been applied only Afyonkarahisar teens but also Uşak and Kütahya provinces are included. The purpose of this study is to research adolescents’ use of social media, computers, text messaging, cell phones and instant messaging related to social anxiety effective’s technological tools versus face-to-face communication. Because, especially 15-20 years-old-teens have been noticed to use on-line communication means having faster and interactive features rather than having face-to-face communication while talking to their friends or their parents. Consequently, in this study, the reason for this choice has a connection with whether social anxiety or gender will be researched. Literature Overview Internet Use among Teens Use of Internet continues to increase worldwide. Colley and Maltby (2008) indicated that, 57% of households now have access, in comparison to 46% four years ago in the UK (National Statistics, 2006). The Digital Future Project in the US has found that 78.6% of Americans went online in 2005, with an accompanying increase in the amount of time spent per week on the Internet (Center for the Digital Future, 2005). Many of the studies in this area have been conducted with university students, in part because of their widespread use of the Internet as a social tool; indeed such use among students is increasing. In addition, a number of studies have found a high prevalence of social anxiety among university students who may then turn to the Internet as a way of regulating, challenging or escaping their social fears (Shepherd and Edelmann, 2005). Many university students commenting that they are shy in social situations, but are more open, easy going on the net, in chat rooms. They are shy about approaching people at a party or in large group, for example, but not about talking with people in chat rooms. Volume 3 Number 2 Fall 2013 103 Bilal Sisman / Sinan Yoruk / Ali Eleren Cell Phone Use among Adolescents Auter (2007) researched that some of them related to adolescents and young adults use the cell phone differently than their parents and other older users. They prefer to consider it their primary phone –traditional landline phone – for its. Some studies have found little or no difference in use based on gender (e.g., DeBaillon, 2003; DeBaillon and Rockwell, 2005). Other studies have found, however that boys tend to be more intrigued with technical aspects of the devices, while girls tend to prefer the interpersonal connectivity – and spend more time using their phones for voice calls (Henderson et al., 2002; Rakow and Navarro, 1993; Skog, 2002). In fact, while women in one study felt the phones resulted in more freedom, male teens found the additional connectivity a constraining inconvenience. A recent Pew Internet research study found that approximately 33% of teens have a cell phone. Of the 33% of teens who reported owning a cell phone, approximately 64% said they had sent text messages. In addition, the Pew study found that of the young cell phone users, teenage girls tended to use text messaging much more than their male counterparts. Older teens (17 year olds) also reported sending more text messages on average than younger teens. Although cell phone use and text messaging have risen among U.S. teens, their popularity remains considerably less than European teens (Lenhart et al., 2005; Pierce, 2009). Seventy-five percent of teenagers now own cell phones, and 25% use them for social media, 54% use them for texting, and 24% use them for instant messaging. Thus, a large part of this generation’s social and emotional development is occurring while on the Internet and on cell phones (Hinduja and Patchin, 2007). Instant Messaging The two previously mentioned theoretical approaches might adhere to different types of Internet activities that may have differential effects on well-being. One type of activities surfing, which can be described as visiting web sites on the Internet for non-communication purposes. In contrast to surfing, Instant Messaging (IMing) consists of sending messages directly to others one has invited to the online conversation. Because of its dyadic, real-time, and private format, IM-ing has 104 Journal of Economic and Social Studies Social Anxiety and Usage of Online Technological Communication Tools among Adolescents been as an excellent ‘training ground’ for adolescents in terms of their social skills (Morgan and Cotten, 2003; Valkenburg and Peter, 2007). A survey study showed that among freshmen college students whereas increased time spent IM-ing was uniquely associated with less reported depression, increased time spent surfing was uniquely associated with more reported depression (Morgan and Cotten, 2003). Social Anxiety and Gender Differences Social anxiety has variously been referred to as both shyness and social phobia. Social phobia, (also known as social anxiety disorder) is at the more extreme end of the continuum of social anxiety and “includes a collation of symptoms which unlike shyness can vary in severity throughout one’s life” these symptoms are only elicited in certain social situations (Leary, 1983). In contrast shyness while at the less severe en of the social anxiety continuum is more likely to be a life long characteristic of an individual’s temperament which is experienced in every social occurrences (Lecrubier et al., 2000; Shepherd and Edelmann, 2005). Individuals are often motivated by a need to feel a sense of belonging; however, those with social anxiety may find it difficult to fulfill this social need because of the Internet. For socially anxious individuals, the Internet and socially interactive technologies (text messaging) can have both positive and negative results. Kraut et al (1998) found that online interaction greatly reduced face-to-face social interaction. Females usually have more negative attitudes toward computers and greater computer anxiety than males. Researches on computer self-efficacy in general also revealed that males on average have better computer self-efficacy than females (Hackett, 1985). Similarly, one might expect adolescent girls to be more worried than boys about peers’ negative evaluations of them. Indeed, surveys have found that adolescent girls are more concerned than boys about others’ judgments of their appearance and behavior. Moreover, rates of internalizing problems are higher among girls than boys and adult women are about twice as likely to be socially phobia as men (Schneier et al., 1992). Thus, adolescent girls may be more vulnerable than boys to feelings of social anxiety, and this may have implications for their social functioning (La Greca and Lopez, 1998). Jackson et al. (2001) predicted that women would use e-mail more and men use the Web for information more, based on the greater interpersonal orientation of women Volume 3 Number 2 Fall 2013 105 Bilal Sisman / Sinan Yoruk / Ali Eleren and greater task orientation of men. This prediction was supported in a large sample of Anglo-American undergraduates, even after computer self-efficacy, loneliness and depression were controlled for. Wasserman and Richmond-Abbott (2005) found that women use e-mail slightly but not significantly more than men but that men use chat rooms more. Shepherd and Edelmann (2005) study social anxiety and Internet use to explore in relation to regulation of social fears. It is hypothesized that high social anxiety will be associated with low ego strength as well as greater scores on the Internet. Methods Purpose It is very important to analyze of instant messaging, text messaging and other socially interactive technologies among teens due to significant popularity. Since social anxiety can affect one’s type of social interaction, it is prominent to examine if social anxiety is influencing how adolescents communicate with others and which technology. The purpose of this study is to research adolescents’ use of online social sites, cell phones, test messaging and instant messaging and if social anxiety influences technology versus face-to-face communication. So, the following hypotheses are proposed: • • • • • • 106 H1: Females use socially interactive technology (text messages, cell phones, e-mail, and instant messages) more than males. H2a: There is a relationship between not feeling comfortable talking with others in person and feeling more comfortable talking with others online. H2b: There is a relationship between not feeling comfortable talking with others in person and feeling more comfortable messaging with others. H2c: There is a relationship between the amount of their families’ monthly income and social anxiety. H3: Males are more social anxiety with face-to-face interactions than females. H4: Males are feeling more comfortable talking with others through a social interaction technology than females. Journal of Economic and Social Studies Social Anxiety and Usage of Online Technological Communication Tools among Adolescents Participants and Data The data was gathered from 544 students among High Schools in Afyonkarahisar, Kütahya and Uşak cities. The age of the participants ranged from 15 to 18. 32.4% were freshman, 24.1% were sophomore, 28.1% juniors and 15.4% were seniors. The sample consisted of 227 (41.7%) males and 317 (58.3%) females. We first were contacted the principals and asked permission in order to survey some classes on school. After that, teacher was chosen to gain a sample of students from each school year. After obtaining permission from teachers, each student was warned about survey able to tell it their parents. The survey was applied in the students’classrooms and teachers were present during all testing. All students received the same instructions and their identity would remain secret. Each participant completed a report questionnaire. Measures The first two questions of the questionnaire consisted of primarily demographic questions such as age, sex, year in school, and general information. The next series ten questions included items of the students’ comfort level in interacting with others face-to-face, online talking, cell phone or text messaging. Questions were prepared by using both nominal (yes or no) and 5 point likert-type scales. All hypotheses were developed by us to measure the relationship between social anxiety and usage of technological tolls among teens. The reliability of tests confirms our measurement (Cronbach α=72.4). This test can be acceptable according to rule of thumb for describing internal consistency 0.7 ≤ α ≤ 0.8 Results 3.7% stated the amount of children who separate from their families is not adequate for measuring the relationship between social anxiety and family situation. First of all, participants were asked about type of technological tools they had use. Volume 3 Number 2 Fall 2013 107 Bilal Sisman / Sinan Yoruk / Ali Eleren • 88.9% reported having personnel computer. • 79.2% reported having Internet on their PC. • 92.6% stated that having messenger address. • 73% reported having e-mail accounts. • 96.7% reported having cell phones. • 72.2% reported having text messaging capabilities on their cell phones. • 84.3% stated having social media sites and most of them are using Facebook (49.2%). Most of the other has more than one membership in social sites (twitter, MySpace, LinkedIn etc). The following next questions asked how much time adolescents use each of their time on their technological tools on average day. Table 1 shows the results. Table 1. Technological tools use 9 h + (%) 7-8 h (%) 5-6 h (%) 3-4 h (%) 1-2 h (%) Msn Instant message Cell phone Social sites Music Game Mail Internet 0.4 3.7 0.6 0.7 3.9 0.7 0.9 1.1 0.4 2.2 0.2 1.7 2,4 0.9 0.2 0.4 0 4.6 0.4 1.5 5.0 1.3 0.2 1.1 0.6 10.8 2.2 8.0 10.3 4,3 1.1 6.7 4.3 17.5 11.9 29.0 39.0 17.7 3.9 26.8 30 min (%) 13.6 23.6 56.9 33.3 31.8 24.0 12.8 40.1 None (%) 80.7 37.5 27.8 25.8 7.6 51.0 80.9 23.8 As shown in Table 1, adolescents rarely use messenger sites and mail address. They generally send instant message with their cell phones. They spend 1-2 hours for listening music and averagely 30 minutes for Facebook in a day. More than half of teens have cell phones that enable to call, message and access to Internet. All following hypothesis were examined to 95% level of confidence. Regression analysis was made to see the relationship between feeling anxiety and family income with technological tools usage among adolescents; T-test, Chi-square test and one way ANOVA test was performed in the study to see gender differences with technology use, face-to-face interaction, talking online, making friends online and playing games. 108 Journal of Economic and Social Studies Social Anxiety and Usage of Online Technological Communication Tools among Adolescents Table 2. Regression estimates and correlation Variables Dependent Gender (technology use) Gender (face-to-face interaction) Gender (talking online) Gender (making friends online) Game Feeling anxiety İncome (monthly) Males Females Males Females Males Females Full sample (n:544) Standart Significant Coefficient deviation (p) 1.52 2.31 0.019 1.08 1.90 1.86 0.99 0.003 2.12 0.92 3.48 1.05 0.009 3.71 0.96 Males 3.62 1.19 Females Males Females Instant message Social sites Msn Cell phone Instant message Social sites Msn Cell phone 4.14 0.30 1.06 0.069 0.139 -0.005 0.308 173,737 -270,341 -8,104 -38,099 1.04 0.75 1.66 0.054 0.052 0.025 0.045 119,286 115,816 55,636 99,021 Independent Correlation coefficient (r) 0.000 0.000 0.201 0.008 0.851 0.000 0.146 0.020 0.884 0.701 0.449 0.125 After T-test was made for hypothesis 1, table 2 shown that females (M: 1.08, SD: 1.9) reported that they are using text messaging more than males and significant relationship between text messaging and interactive technology (M: 1.52, SD: 2.31, p: 0.019). In addition, males (M: 0, 3, SD: 0.75) spend much more time than females to play games and there is a gender differences between playing games on computer and social technology using (M: 1.06, SD: 1.66, p: 0.000). There are not gender differences in use of instant messaging, e-mail, social network sites and chat rooms. When someone have to talk face-to-face, feeling anxiety was positively correlated with feeling comfortable with others through cell phones and social media sites (r:0.449, p: 0.000). There isn’t a relationship between social anxiety of face-toface interaction and feeling comfortable with others via messaging (p: 0.204). The regression analysis showed, there isn’t a relationship between the amount of families’ monthly income with social interaction technologies (p: 0.230, r: 0.125). Volume 3 Number 2 Fall 2013 109 Bilal Sisman / Sinan Yoruk / Ali Eleren According to the results of Chi-square test, significance was not found for hypothesis 3. However, a female feels more uncomfortable than males for face-toface talking than males (Levene’s test p: 0.017). There is just significant differences feeling comfortable by talking with social media sites with gender (p: 0.003), that is; famales (M: 2.12, SD: 0.92) are more uncomfortable than do males (M: 1.86, SD: 0.99, p: 0.000). There is significant difference for hypothesis 4, after One-way Anova test was made, teens prefer talk with someone on computer instead of talking and make new friends with someone on computer. Females (M: 3.71, SD: 0.96) prefer talking with some on computer instead of talking face-to-face than males (M: 3.48, SD: 1.05, p: 0.009). Similarly, females (M: 4.14, SD: 1.04) prefer making new friends with someone on computer more than males (M: 3.62, SD: 1.19, p: 0.000). Conclusion In Turkey, cell phone, text messaging, instant message and social sites membership are recognized as the most important medium of communications among adolescents. Internet access via PCs and mobile phones plays very significant role Turkey adolescents. Technology tools effect teens’ social emotions against their families and friends. This study demonstrated that teens are using socially interactive technology to communicate with others and it serves social anxiety and decreasing face-to face communication. The findings of the current study found that females are using text messaging more than males. However, males are playing games for a long time than females. In addition, females are feeling more uncomfortable talking with others face-to-face than males. And, females also prefer talking with some on computer instead of talking face-to-face than males. Similarly, females prefer making new friends with someone on computer more than males. Beyond the results of this study, we examined something very important situation related to adolescents’ today currents. While the age of participants were increasing, the time spend in social media is decreasing. The reason for such conclusion revealed that, senior students have to work more for university exams during education session. So, they can’t spend much more time on social media sites. 110 Journal of Economic and Social Studies Social Anxiety and Usage of Online Technological Communication Tools among Adolescents Many studies in this area is conducted for university students. The use of Internet as a social tool is more common among at this age adolescents. But this study is related to high school level for students and we examined the relationship between social anxiety and the use of Internet, cell phone as a technological tool. Further studies may apply for university students. Finally, adolescents who may be shy, have social phobia, and feel anxiety about talking with someone, various and different technologies provide reliable opportunity for them to contact with them. It is possible to reproduce the contributions of technology for our world. Although this technology facilities the lives of people, it sometimes breaks the peace. Technology has changed the way we live today and also changing our communication skills with high costs. Only time will tell what is going on in the future. Parents and teachers have great responsibilities for this subject. References Auter P. J., (2007). Portable Social Groups: Willingness To Communicate, Interpersonal Communication Gratifications, and Cell Phone Use Among Young Adults, Inetrnational J. Mobile Communications, 5(2) 139-156. Colley, A., and Maltby, J., (2008). Impact Of The Internet On Our Lives: Male and Female Personal Perspectives, Computers In Human Behavior, 24, 2005-2013. Hinduja S, Patchin J. (2007). Offline Consequences Of Online Victimization: School Violence And Delinquency. J Sch Violence. 6(3): 89 –112. La Greca, A. M. and Lopez N., (1998). Social Anxiety Among Adolescents: Linkages With Peer Relations and Friendships, Journal of Abnormal Child Psychology, 26(2), 83-94. Lenhart, A. (2005). Protecting Teens Online. Pew Internet and American Life Project. URL: http:// www.pewInternet.org/. Morgan, C., and Cotten, S. R. (2003). The Relationship Between Internet Activities and Depres-sive Symptoms in a Sample of College Freshmen. CyberPsychology and Behavior, 6(2), 133–142. O’Keeffe, G. S., and Pearson, C., (2011). The Impact of Social Media on Children, Adolescents and Families, Journal of American Academy Of Pediatric, 800-805. Pierce, T., (2009). Social Anxiety and Technology: Face-To-Face Communication Versus Technological Communication Among Teens, Computers İn Human Behavior, 25, 1367-1372. Selfhout, M. H. W., Branje, S. J. T., Delsing, M., Bogt, T. F. M., and Meeus W. H. J., (2009). Different Types Of Internet Use, Depression, and Social Anxiety: The Role Of Perceived Friendship Quality. Journal of Adolescence, 32 819-833. Volume 3 Number 2 Fall 2013 111 Bilal Sisman / Sinan Yoruk / Ali Eleren Shepherd, R. M., and Edelmann, R. J., (2005). Reasons For İnternet Use and Social Anxiety, Personality and Individual Differences, 39, 949-958. Subrahmanyam S. and Lin, G., (2007). Adolescents On The Net: Internet Use and Well-Being, Adolescence, 42(168), 659-677. Valkenburg P. M., and Peter J., (2007). Online Communication and Adolescent Well-Being: Testing the Stimulation Versus the Displacement Hypothesis, Journal of Computer-Mediated Communication, 12(4), 1169-1182. Appendix 1 The purpose of this questionnare is to research 15-20 year old adolescents’ use of online social sites, cell phones, test messaging and instant messaging and if social anxiety influences technology versus face-to-face communication. Assoc. Prof. Ali Eleren Asst. Prof. sinan Yörük reserch Assist. bilal Şişman 1. What is your sex? 2. What is your year in school? (select only one) ------(9) ------Male ------Female ------(10) ------(11) -----(12) 3. What is your families’ total income? ------------TL 4. Does your mother and father live together? ---Yes ---No 5. Do you have a daily computer? ---Yes ---No 6. Do you have a Internet access on your computer? ---Yes ---No 7. Is your Internet connection wireless? ---Yes ---No 8. Do you have a instant messaging address? ---Yes ---No 9. Do you have e-mail address use actively? ---Yes ---No 10. Do you have a cell phone? ---Yes ---No 11. Do you have access to the Internet on your cell phone? ---Yes ---No ---Yes ---No 12. Do you have a social networking account? (Twitter, Facebook, MySpace etc) 13. If yes on 12, which of the following social networking do you have? ------Facebook ------Twitter ------Linkedin ------Youtube ------Classmates ------Other (please specify)--------------------14. On an average day, how much time do you spend with each of the following? 112 Journal of Economic and Social Studies Social Anxiety and Usage of Online Technological Communication Tools among Adolescents 9+ 7-8 5-6 3-4 1-2 30 min- None hours/ hours/ hours/ hours/ hours/ utes/ daily daily daily daily daily daily Instant messaging Text messaging with cell phone Talking with cell phone Social networking Listening music Playing game E-mail Online (surfing on the Internet) 15. How comfortable are you talking with friends, family and teachers face-to-face? ---Very comfortable ---Comfortable ---Normal ---Uncomfortable ---Very uncomfortable 16. Do you find yourself getting anxious when you have to talk with someone face-to-face? ---Always ---Frequently ---Sometimes ---Rarely (very few) ---Never 17. How comfortable are you talking with others using text messaging? ---Very comfortable ---Comfortable ---Normal ---Uncomfortable ---Very uncomfortable 18. (If yes on 12) How comfortable are you talking with others on your social networking sites? ---Very comfortable ---Comfortable ---Normal ---Uncomfortable ---Very uncomfortable 19. How comfortable are you talking with others on your instant messaging? ---Very comfortable ---Comfortable ---Normal ---Uncomfortable ---Very uncomfortable 20. How comfortable are you talking with others on your cell phone (talking)? ---Very comfortable ---Comfortable ---Normal ---Uncomfortable ---Very uncomfortable 21. How often dou you text message someone instead of talking to them face-to-face? ---Always ---Frequently ---Sometimes ---Rarely (very few) ---Never 22. How often do you talk with someone on the computer instead of talking to them in person? ---Always ---Frequently ---Sometimes ---Rarely (very few) ---Never 23. How often dou you text message someone instead of talking with them on the cell phone? ---Always ---Frequently ---Sometimes ---Rarely (very few) ---Never 24. How often dou you make new friends with someone on the computer? ---Always ---Frequently ---Sometimes ---Rarely (very few) ---Never 25. When confronted with a difficult stuation with another person, which of the following would you typically use to talk with other person? ------Text messaging with cell phone ------Talking with cell phone ------E-mail ------Instant messaging ------Social networking sites ------Face-to-face communication Volume 3 Number 2 Fall 2013 113 Journal of Economic and Social Studies State as the Source of Wealth: In Ottoman Economic Thought: A different approach to reflections in the aftermath of the global crisis Birol Çetin Gaziosmanpasa University Faculty of Economics and Administrative Sciences, Department of Economics / 60110 Tokat TURKEY birol.cetin@gop.edu.tr AbstrAct This study aims to deal with the ways of creating wealth by economic activities, presenting experiences within Ottoman state tradition and a role of Ottoman state during this period. In this context, the economic power achieved by the state will be explained through examples of practices. The role of the state in the economy has been raised with the latest global crisis and despite the historical expriences, this role has been started to debate in the economics. In fact, the corrupted state concept should be re-evaluated and re-established. Otherwise, re-evaluation of fundamental issues such as market system or freedom of enterprise wouldn’t contribute much to the solution of the problem. KEYWOrDs Ottoman, Global Crisis, Market Sistem ArtIcLE HIstOrY Submitted:21 Jun 2012 Resubmitted:08January 2013 Accepted:25 March 2013 JEL codes: N00,N20,B10 Volume 3 Number 2 Fall 2013 115 Birol Çetin Introduction In the traditional Islamic thinking, economy was subsistence and economic activities were mainly based on the supplement of daily requirements. Although there was no prohibition of becoming prosperous or accumulating wealth, owners of the wealth were advised to spend it for charity. According to Holy Quran and the Sunna, words and acts of the prophet Mohammad, basic sources of Islamic knowledge, and principles of justice were important in economical affairs, and legal way of earning were accepted equal to worshipping. However, after the establishment of sultanate tradition, the state transformed a source of income mechanism into a distrubition of wealth mechanism. Ibn-i Haldun,a famous philosopher of Islamic world,addressed the both normative and positive sides of the issue and stated that the state got power in the area where it shouldn’t have had it. According to Ibn-i Haldun, there were three ways of subsistence and income: agriculture, industry (crafts) and trade. It seems that he considered agriculture and industry affirmative while trade was doubtful. He considered agriculture and crafts honorable, because they rely on one’s direct labor and effort. But they both did not make a person rich. In addition, Ibn-i Haldun mentioned public service as a means of getting rich, despite the fact that it was not a regular way of subsistence. He thought that people should get close to either the state or trade in order to be rich. Imaret, described as a state governing by Ibn-i Haldun Imaret, was the way of collecting taxes through confiscating someone’s wealth using political power and administrative office. Therefore, state work is not a natural way of make for a living. He described the civil cervants as feminine people who cannot do their own business. According to Ibn-i Haldun the ones who cannot come close to the state in order to protect his capital and to be rich, in his own words “the ones who do not indulge and lick someones’ boots” cannot have wealth because wealth is produced by public office. The ones who seize the power and who come close to the rulers became rich. Therefore, this study aims to describe the ways of getting rich appropriate to Ibn-i Haldun and reveal the formation of them in Ottoman Empire. Trade, public services and companies were predominantly analized in this study. 116 Journal of Economic and Social Studies State as the Source of Wealth: In Ottoman Economic Thought: A different approach to reflections in the aftermath of the global crisis Role of the State in the Economy It is not possible to state that trade was excluded in the powerful central administration of Ottoman Empire. In terms of the state intervention to the markets, the central administration was parallel with the system of administration apart from some applications.The purpose of the state was to suppress all kind of power and influence of local authority over the economic life and to obtain a great influence over all kind of goods as well as people by means of eliminating the intermediate agents between the subjects and the Sultan (Güçer, 1987). Although, the main purpose seems beneficial for the citizens as it is frequently appears on the documents such as “public good”, “development and prosperity of the country” or “increasing the welfare of the citizens” (Tabakoğlu, 1994), it was actually decreasing the state’s reserves as it was frequently appears on documents such as “working to increase the state’s wealth without taking inadequate taxes”. Some interventions aimed to take taxes while others aimed to buy goods and services below the market price by the state. Dominant state Policy in Markets: Although,the prices were formed by a system determined by the state,the markets were not competitive in mant big towns,notably İstanbul had had a dynamic economic life.It can be claim that the narh system,the state intervention on the market prices,had negatively affected this dynamism. Aims of The Narh system: In addition to economic aims,there were political,moral and administrative aims at the narh system.A brief study of these aims will show the functions of the narh system.Although to separte these aims is diffucult,significant ones can be determined based on the classification of examples.These can be listed as follows: Administrative and Politicial Aims: Sultans and “sadrazam”s sometimes had attempted to lower the narh or to keep it low in order to reinforce their positions as well as to get the support of the general public. These attempts were embodied in expressions found in “hatt-ı humayun” (decree of sultans) issued by Selim III to “sadrazam” Musa Paşa (“make sure that the price checks are strict and satistaction of the citizens is ensured. Also affairs of citizens should be facilitated through your day-andnight efforts. Blackmarketing should be prevented in order to supply the needs of people Volume 3 Number 2 Fall 2013 117 Birol Çetin in Istanbul. The behaviours against rules should be punished (Pakalın, 2004). According to Tevkii Abdurrahman Paşa, a regulator, the narh should not be haphazardly dertermined. According to Şehit Ali Paşa, the form of “putting into effect the prices by the state should not be controlled by the judges and municipal police, also it should be checked on a daily basis. Overcharging should be eliminated. You should be aware of the fact that the narh is the most crucial work of the state.” Therefore,the narh system was not only for regulating the price stability but also played a significant role in maintaining the order of the community. In his book entitled “Fusul-i Hall’ü Akd and Usul-i Harc-i Nakd”, Gelibolulu Mustafa Ali points the following justifications as the purpose of the book: “... with increasing population, narh system could not be effectively practised. Thus, order of the society deteriorated, judiciary corrupted, and the system got into the hands of the dishonest. The sultan cannot be thought of accepting this( Yücel, 2003). ‘’ In this concept, deterioration of the narh, judiciary and social order are accepted as the same. Since the price instabilities which had the same meaning with the weakness of the state authority were taken as an indicator of administers’ and public’s weakness. This issue was taken seriously. It was also taken serious because speculations were forbidden religiously for the public. The narh system was not only to protect the consumers but also was necessary for the protection of produces and traders.Thus, the aim was to prevent non-standard and fraudulent good’s trade. There were various restrictions for traders who acted against the narh. For example, in a decree issued to Judge of Istanbul and municipality police in 1676, the Sultan himself dealt with the packaging and selling procedures for pears brought to marketplace: “Order to the Judge of Istanbul and Municipality Police Chiefs: As “Pazarbaşı” Veli and his deputies complained, since pears come into the marketplace, they are put into small boxes and sold as boxes, the poor cannot buy these boxes. The sales should be like before, i.e. Pears should be brought to the marketplace in big boxes and sold by kilograms based on consumer’s need.”(Çağatay,1981).The administrative interest abaout conforming to the narh system also involved quality and standard of the goods. For example, when it was noticed that the quality of fabrics produced in Bursa deteriorated, an inspector was sent to Bursa in 1477. The inspector found out that a cheaper production method was adopted due to increased raw material prices. This reason was approved and new product was allowed (İnalcık, 1993). 118 Journal of Economic and Social Studies State as the Source of Wealth: In Ottoman Economic Thought: A different approach to reflections in the aftermath of the global crisis Financial reasons: Fixed prices due to the narh system provided major advantage for the state, because state was the major buyer. The state could buy for lower prices. For example, grain prices formed as free market and the state had had to buy on that price. Free market prices were usually monitored on heritage notebooks and the prices on “tereke” notebooks were always higher than official ones. The reason was the higher “resm-I kısmet” (a kind of heritage tax) taken by judges because of higher “tereke” price. Although, price of wheat per kg was 11 “guruş” and the price of barley per “kile” (a weight in unit) was 5 “guruş” in 1799 at Tokat, according to official purchases by the state, price of wheat flour was 4.0 “guruş” and barley was 2.25 “guruş” (Öztürk,1987). However, it would not be realistic to say that official prices were always kept low and the narh prices were continuously suppressed by the state. Considering the fact that vast amount of agricultural products were sold in the market, it would be true to assume that suppressed prices were only common in extraordinary situations. Direct Interventions To Trade: Other than mentioned above,there were another intervention methods in trade. These were “yedd-i vahid” (monopoly) practices, regulations about market places and trade regions, gedik method (a kind of quota imposed on the craftsman about the number and location of businesses) and various confiscation procedures applied to traders. The aim here was to prevent speculations, blackmarketing, deceptive sale practices, and to establish the order in the country. Food supply in big cities was an issue to which much significance was attached and attention was paid. Ottoman administrators were deeply taking into account that a well running economic life would make other state affairs smooth. Otherwise,there would be many social and administrative problems. “Gedik” and “yedd-i vahid” practices aimed to facilitate the work of the tradesman and to keep things well organized establishment content. Although, confiscation procedures mainly involved state afficials, it applied traders who were thought to have become unjustly rich. Even in financially difficult times, state did not covet traders’ wealth. There was a significant explanation about this subject in “hatt-ı humayun”s issued by Selim III.During 1787-1792 AustrianRussian Wars,the state was in financial difficulty and Sultan asked for financial solutions. This was explained by Selim III as follows: “Kaymakam Paşa” what I say to you is: “There can be no answer if the state has no money, then how can we live such a pretentious lives and what did we consume? We showed our enemies that we are weak. And looking at your situation,I am surprised. Just tell me how it could be possible. You have been in public service for so long, you are the one who knows where the extra five akçe(a currency unit) is, and you are asking me. You said that this is away from here, and Volume 3 Number 2 Fall 2013 119 Birol Çetin didn’t I take it? (Karal, 1999) Although, upon this decree, Sadrazam wanted to confiscate heritage and merchandise of some traders, the Sultan strongly prohibited and saying: “Seizing the belongings of orphans is something contrary to my opinion. I can not allow to touch them.Whatever accumulated by someone should be left to his or her heirs after the death, whether it is accumulated through trade, farming or crafts, should be left to his heirs, after his death. Don’t ever touch them! Only the wealth of public servants belong to state treasury, neither to me or to his heirs. All of them should be spent for public. Our ascendants did so. If the accountant is afraid of God, he shouldn’t touch the heritage of craftsmen, otherwise he will be severely punished.” One of the clearest examples of direct interventions to market was obligatory migrations practiced upon traders in order to develop trade in some areas. However, these exchanges were provided some tax exemptions. After the conquest of Kefe, 300 traders, and their families, were deported to Istanbul. It is known that livestock traders and butchers worked under coercion in order to keep meat supply running in large towns, and free trade conditions were not applied in this area of work (Pamuk, 1994). Corporate Movements and Status of the State Being dominant in trade through the narh system, the state closely followed developments in the west regarding market economy and corporate movements, also tried to be involved in these activities.As a matter of fact,the first corporation “Şirket-I Hayriye” was established by the state. In this subtitle, the state’s effort to manage the process of capitalist development without losing power will discuss. Partnerships in Ottoman Trade Life: Although, tradesmen and craftsmen were confined with the strict laws which entrepreneurs had sufferd a lot,it seems that Ottoman State closely watched the corporate movements in the west.There is no doubt that partnership culture and practices always survived as an indicator of a deep rooted tradition. However, there were major differences between the corporate tradition of the west and Ottoman applications. There are three types of contract based on partnership in Islamic law consisting of capital, labour and credibility.These were in the form of “Şirket-i Inan” (a kind of 120 Journal of Economic and Social Studies State as the Source of Wealth: In Ottoman Economic Thought: A different approach to reflections in the aftermath of the global crisis labor-capital partnership) and “Mufavada Şirketi”, a kind of cooperation based on the equal rights of family members. Records about these partnerships were taken from “Mecelle” was common in “Şeriyye” Records. “Şirket-i Inan” practice was common among vegetable oil and honey markets, especially in far away trades. They had started from the supply of the product and ended when the product reached to Istanbul where it was shared. In some practices, it was seen that people became the joint-owner of ships through giving loans to ship owners as similar to the western practice (Farooqhi, 1993). Apart from this small scale partnerships which did not require much capital. Especially after beginning of the “Iltizam” practice, privatization of tax revenues, partnerships involved the merging of large companies had started to be common (Gedikli,1999). However, this kind of partnerships which require great amount of capital accumulation was formulated based on some privileges provided under certain circumstances and uncertain conditions (Cezar,1986). Private entrepreneurship and corporate culture had not developed adequately due to the fact that the capital accumulation was in the hands of high officials and profitable business was possible only between privilegers. As a result of “iltizam” practice, owner of “dirlik” (money obtained from the state land used for soldier training (similar with latifundia in Rome) were disappeared. Life-long “mültezim” concept appeared and provinces were left to “mültezim”s. Therefore, non-muslim traders gained power by means of the governor’s policy. Towards the end of the 18th century, some of this “sarraf ” s had fortunes more than a million pounds. These minority members were in close contact with foreign traders and they played significant role about providing loans to the state (Tezel, 1986). They acted like the treasury of the state during the hard times for financing military expences and like financial crisis. For example, in 1878 460,000 pounds were taken from banker Zarifi for military expences. Another document shows that “miri” farms in Yanya were given away for paying a previous loan of 25,000 Turkish liras. 1 Embassies and consulates of foreign countries started to market their own goods in modernization and especially in the later half of the 18th century. In this period, partnerships with minorities had become more and more common.These minority traders were also benefiting from various privileges given to foreign traders through merging. There were many minority members who changed their nationality. After securing minority rights through “Küçük Kaynarca” Treaty, this practice had become even more common and disguised traders became registered ones.Thus,the state could not stop this trend and gave the rights to its own minorities which were 1 - B.O.A. Yıldız Tasnifi, Sedaret Resmi Maruzat Evrakı no:1-40 Volume 3 Number 2 Fall 2013 121 Birol Çetin previously given to foreign traders and “European Mechanthood” was established. In order to keep balance,these privileges were given to Muslim traders with some extras,thus “Hayriye Merchanthood” was established. Unlike the former traders, later ones were also given the privilege of trade to Persia and India (Tabakoğlu, 1994). Corporate Movements in Ottoman Empire: There were some opinions in “Tanzimat” period that the economic problems country experienced could be overcome through corporate development. Sultan Abdülmecid stated that he struggled a lot to take loans but he failed and and revenue-increasing measures, such as forming companies as in other countries, were necessary to overcome the financial problems (Koraltürk, 1999). Ministry of Trade was established in 1840 and “Kanunname-i Ticaret” act of 1850 decelerated some developments about corporate development.”Kanunname-I Ticaret” was a translation of French law titled “Code de Commercé” of 1807 with some minor modifications.This act was current until 1926 and provisions about companies were mentioned in its 11th,13th,14th,19th articles (Şekerci,1981). According to an expression in 13th article, constitution which was formed in 1876 “Ottoman subject has rights to establish all kind of companies for trade,crafts and agriculture provided that they obey the rules and order”. Cevdet and Fuat Paşa established the first Ottoman share company, “Şirket-i Hayriye”,in 1851. At first, operating ferries were between Bosphourus and Karaköy. “Şirket-i Hayriye” increased its activities year by year, and became a modern business after obtaining the privilege of Bosphourus transportation. As a consequence of these developments, many maritime businesses were formed using state protection and practices after 1910 (Eldem,1994). Board of Industrial Development operated between 1860 and 1873 focues on corporate development and worked on restoration of craftsmen unions.The purpose was to collect the craftsmen operating as scatterd under share company system with a common capital investment. However, when the board started to dominate every decision from appointment of company executives to profit margins and sale prices, desired benefits were not achieved. An example of these efforts was a share company run by leather craftsmen of Istanbul with a capital of 10.000 golden coins. Corporate development gained speed in the second half of the 19th century. Following Şirket-i Hayriye, Bank-ı Osmani and Şirket-i Osmaniyye that run Aydın-İzmir railway were established. A majority of companies formed in this period were privileged foreign companies and they conducted banking, insurance, railways, mining, utiliy gas aand tramway operations.These were foreign companies.Until 1908, the 122 Journal of Economic and Social Studies State as the Source of Wealth: In Ottoman Economic Thought: A different approach to reflections in the aftermath of the global crisis number of share holding formed and involved in economic activities in Ottoman land was 86. However, a total of 236 companies were established in 1908-1918 period only (Toprak, 1997). In the process of corporate development and strengthening the capital structures Galata Bankers who had been operating since the time of Sultan Mehmed made significant contributions. During the reign of Abdülhamid,not only public debt securities but also shares of companies which had been operating in transportation business such as railways, tramway , tunnel, and in mining, water, natural gas, electricity and manufacturing areas were being traded in Galata Stock Maket. Galata Stock Market started to buy and sell state bonds and company shares.Sultan Abdülhamid,himself,was involved in these affairs by means of his private banker Zarifi and it took public attention to this area (Kazgan,1995). By the act of “Development of Industry” which had been put into effect after “Meşrutiyet”, corporate development was encouraged and domestic companies were started to be established. Most of these companies were included in the list of Shares and Bonds Stock Market. Companies quoted in the stock market were in different ındustries such as such as Thread and Linen Goods Share Company (with a capital of 10.000 Lira), Ottoman Anonymous Ittihat Company (with a capital of 100.000 Lira), Ottoman Anonymous Rubber Company (with a capital of 20.000 Lira), Ottoman Marble Company (with a capital of 25.000 Lira), Turkish Naphta Industry (with a capital of 250.000 Lira) and “Dersaadet Cheesemaker Teavün” Company Company (with a capital of 25.000 Lira)( Fertekligil, 2000). Figures about factories and manufacturing plants in a statistical record dated 1897 gives information about trade life. Under the title of “Memalik-i Mahrusa-ı Şahane’de mevcut fabrika ve değirmen vesairenin nev’i” (number and kinds of factories and mills in Istanbul) 23.837 business places were mentioned (Güran, 1997). Corporate concept in Ottoman Empire appeared as privileges of share corporations. In the study of Mehmet Cavid Bey entitled “İlm-i İktisat” after explaining various corporate types, the following information regarding share corporations were given (Mehmet Cavid Bey,2001). “A majority of companies are share companies. They performed profitable activities and much of work especially in trade and industry is done by these companies.As other European countries establishment of these companies are subjected to Sultan’s approval.Besides the company has to follow some procedures in order to get privilege”.Therefore,the key condition for the company was to obtain a privilege. While the corporates in the west formed around the capital accumulation.They were quite different in Ottoman Empire in terms of their logic since they did not Volume 3 Number 2 Fall 2013 123 Birol Çetin fit to partnership cooperation.Ottoman corporate development process completely relied upon privileges, and getting rich through trade which was possible only via the permission of the state. Public Service and Its Monetary Advantages Strong bureaucratic structure in Ottoman Empire was a result of an effort which was to keep permanently the government in the hand of centralized administration.Therefore, the public service officers were pleased financially to create a strong bureaucracy.The richest top ranking officers known as “devletlü” were remarkably in large numbers.The methods to obtain a fortune for high ranking officers will be clarify in this chapter. Monetary Awards: The issues such as great military achivements, excellent pursuance of assigned duties, and neutralization of oppositions taken place against the Sultan were the prominent awarding examples within the bureaucratic structure of Ottoman Empire. Since entering the public service had been a matter of good chance,young people were accepted to the palace after passing all tests.They had been hardly training. Entering to the palace was not easy and selected ones had to have certain qualifications (Ricaut,1686). According to Gökyay ‘’When Architect Mehmet Ağa, who built the Blue Mosque, entered the office in the Palace, he saw the attention paid towards the musicians and wanted to be a part of this profession. He convinced his master to give him lectures on music. However, after a dream he gave up this profesion and decided that it would conflict with his belief.After getting know the various artisans in the palace, he was interested in the art of “sedefkar”s, inlay workers using mother-of-pearl, and decided to perform this art.” Although the training given in the palace and criteria the young people were subjected to decide the occupations to be assigned were out of the present study’s scope.It is important to note that the trainees could select the occupation themselves. It seems that this practice had aimed high productivity from the very first day in public service (Gökyay,1988). In fact, there was no need for additional awards since Ottoman administration system. The upper office was quite awarding financially and the state officers had done their best to get upper offices. In order to understand the income of the high officers the following incidence reported by Mustafa Nuri Paşa is interesting “During the end of 124 Journal of Economic and Social Studies State as the Source of Wealth: In Ottoman Economic Thought: A different approach to reflections in the aftermath of the global crisis the reign of Sultan Mahmud II, Paris ambassador Ahmet Fethşi Paşa was appointed to “Sancak” of “Aydın” and then the commander of division and former chief commander Namık Paşa was appointed as a “müsellim” with 25,000 “guruş” monthly salary. After a while,Ahmet Fethi Paşa came to Istanbul and then went to Aydın. He offered Namık Paşa 25,000 “guruş” in addition to income of all “mukataa”s and under-door revenues and other income for his sustenance. Since Namık Pasha did not accept the offer, he appointed Ahmet Efendi, one of his relatives as his müsellim. After earning two million guruş of income that year, Sultan Mahmud II notified the Minister for Mukataas Nazif Pasha and allowed Ahmed Fethi Pasha to keep only the half of the income and returned the other half to treasury. The aim of this example is to show that governors and mutasarrıfs of sancaks with good incomes had very high incomes and they hadn’t needed extra money from taxes imposed on the people.Here, salary differences between administrator and deputy classes were significant. According to accountant records of Istanbul Palace in the period of 1555-1556,daily income of “Dar’ül Saadet Ağası” was 45 “guruş”,while 24 “ağa”s which had worked under his supervision was only 5 guruş (Barkan, 1979). In statuebooks, chapters had not only the job description but also had the salary aspects of the officers. For example, “and beğlerbeğilik is the path of four people: province chief accountants, approval officer on behalf of the Sultan, judges whose incomes were five hundred “akçe” and province “bey”s whose incomes reached four hundred thousand akçe (Akgündüz, 1990)”. Therefore, significance of officers were defined based on the income income obtained. Besides,statue changes were also remarkable.High officals called “devlet erkanı” such as Prime Minister, “Şeyh’ülislam”, Navy Commander, Ministers, Minister “Kethuda”s,accountants and lawyers had residences and as many as 200 service personnel such as kethuda, mühürdar (sealer), cook, stableman, boatman, residence keepers and tailors. These service personnel attended to wars and conducted their service during the war. Although this practice was criticized heavily, it was not changed (Aksan, 1997). Wealth of high officers were staggeringly high. When a messenger from India said to Halil Hamid Pasha that they could give five thousand sacks of golden coins to Ottoman Empire in return for helping them, the Pasha said “I and Kaptan Paşa present here alone could give that much”. The amount mentioned was about two hundred fifty thousand golden coins. It is seen that regular payments were done for the central officers until 18th century. However, along with the deteriorating financial structure, a “grant based service fee” system was adopted in place of regular salary payment. In this system, the grant received by bureucrats increased as he worked harder and did his job properly (Findley, 1996). Volume 3 Number 2 Fall 2013 125 Birol Çetin Distinction should be made between two aspects of material award system: first, appointing one or more level above offices in return for the accomplishments and the second giving big bonuses. The former was common and effective in use. This practice continued after Tanzimat, although regulations regarding public officials were subjected to certain rules Ahmet Muhtar Pasha was promoted from colonel to Marshall after his extraordinary success in Yemen only in 9 months. He talks about this: “When the boat which was carring the document showing my promotion and accompanying officers came to Kunfuda and Şakik Edip Efendi and aide of the sultan set out to come near me. However,I had already a herald from Kunfuda,so I was higly affected and cried. I was closed up in my tent and did not accept anybody for an hour. The reason not having enough time to think about deserving this duty or not and having enough qualification for this duty. I was scared of being humiliated later. For I was a colonel nine months ago. My brigade and major general period are only nine months. Giving me these three big ranks meant that they expect a lot from me.” Another example of this was promotion of Topal Osman Nuri Paşa from colonel to marshall only in three years. In the practice, awards called the state awards and promotions were given collectively as well as privately. Collective awards called tip from the Sultan were given during ceremonies for the Sultan’s claim of throne and during another celebrations after big military triumphs.Examples of individual awards were decorations of historians appointed by the Sultan to “Divan-ı Humayun” and some other offices. Along with the formal start of the state historian statue in 1700, famous historian Naima was rewarded 500 “guruş-u attiye” and one “guruş” payment rise to his daily wage at customs (İlgürel, 1991). Sometimes there were payments for the officials relieved of the duty in order to prevent them to get offend and to make them feel good about the trust of his service. After Halil Paşa removed from the duty,Sultan Mehmed II (the Conqueror) gave him a village in Filibe Sancak as his own property in order to dispel his doubts and to decorate him (Çarşılı, 1988). The available documents show that sometimes there were extravagancies in rewarding. For example, after the Crimmean war, the British and French saw that saw that Ottoman gun powder was better than the ones produced in their countries. They offered big rewards to chief of gur powder manufacturing officer Ohannes Dadyan and wanted to bring him to their countries. He turned down their offers and said that “I was born in Ottoman Empire and raised here. I dedicate my life to this country”. After heard about it Sultan Abdülmecid donated him a land in Yeşilköy as far as his eyes could see (Pamukciyan, 1958). This land was in the area known Florya 126 Journal of Economic and Social Studies State as the Source of Wealth: In Ottoman Economic Thought: A different approach to reflections in the aftermath of the global crisis today and consisted of 210 pieces of plots, which is almost all of today’s Florya. Dadyan’s inheritors had disagreements over the land in 1950’s and their cases went on for years in courts (Dadyan, 1954). Warrenties and Eliminating the Concerns for the Future: Incentive and awarding practices applied during the public service continues after termination of the service. Thus, officials conducted their lives without concerns. Besides a pension, the officials could retain all the wealth they accumulated during their service in their retirement. Although in Ottoman Empire only the Sultan could own property.”Paşa”s and “Vezir”s also could have properties such as large farms, palaces, residencies, whose incomes completely belong to them.In order to understand the practices on this issue, it is necessary to study Has (big Timars such as latifundia in Rome) concept. Although, “has” term is used for Sultan’s property in Great Seljuks,it has a meaning of private property land and “dirlik” which had returns at least 100,000 “akçe” in a year was belonged to servants of Sultan and palace officials such as “Bostaniyan-ı hassa”, “Doğanciyan-ı hassa”, “Mehteran-ı hassa”, as well as dynasty members, ministers, beylerbeyis, sancak beyis and accountants. For example, there was a rule regarding the sell of grain from “has”es and farms dated H.945 in Bursa (Gökbilgin,1988). “Has”es appropriated to “Vezir”s and high officals were abolished year by year and added to “Mir-i Mukataa”.After “Tanzimat” the principle of making all expenses through the state treasure was put in practice and this system was cancelled totally after starting regular salary payments. Divan-I Malikane” (life-long mültezim), a significant type of private property, was also an example of this privileged land property practice. “Malikane” owners had right to sell,donate,establish foundations and leave them to their children when they die (Barkan,1939).Giving to own a property right ,which exclusively belong to Sultan,to the state officals means that officers could demand nothing from the state since they had some privileges of sultan. As a result of having all financial needs that they could dream of the state officials could only focus on their duties.In addition belongings could have been passed to their heirs, so they did not have any concern for the future. Against the danger of confiscation or contraversions among family members, some officials disregarded inheritance rules of Islamic law and established Foundations special their children, so taking guarantee even the future of their children.In practice, in some cases children of some officials were paid salary.There was no doubt that confiscation was the main concern of state officials who had considerable amount of wealth accumulation. Volume 3 Number 2 Fall 2013 127 Birol Çetin “Müsadere” (confiscation) system took its legitimacy since “devşirme”s were in the same statue of slavery, the state had right to take their share of heritage according to Islamic law.Besides, hey had not pay any tax.they had to help the state in hard times just as the Sultans had paid from their own wealth.If all the wealth were confiscated, the state had to pay for the livelihood of statesman’s household. There were not any age limitations for statesman. Service was unlimited except for dismissal, dementia and impairment. There was an amount given in cash or dismissed officers called “arpalık”.This could be an annual income of a given land or a given sum appropriated from the treasury (Tabakoğlu, 1994). Conclusion Ottoman Empire represents a strong central state tradition and this subject is the source of its economic power.These facts literally show that the state is the source of wealth.The fact that the state was the sole owner of the land restricted the ways of getting rich by trading and working in public service.Strict state control and getting low profit margins made it difficult to be rich through the trade. However, “sahaf ”s operating in financial markets and especially later period’s non-muslims dealing with trade were exceptions. On the other hand civil service semms to be the only way of getting rich in Ottoman Empire.Confiscation establishment was one which could eliminate all concerns.The state was not complaining about high officials who were getting rich because it was thought that this wealth belonged to the state.The main reason for a conviction was the fact that statesmans could not leave this wealth to their inheritors. According to Ottoman political system took its power from economic strength.The state had to be as strong as possible within national borders.The state administration under control of financial system debilitated the state and thus hurt the citizen whose protector is the state. This became apparent with the financial depression happened in the 17th century. Present day situation is no different. In the recent recession, states must reconsider their relationships with financial system and make some reorganizations. It should not be forgotten that 90% of the public revenues come from taxes. Although cooperation with financial quarters can be conceived to the benfit of the states at first, it 128 Journal of Economic and Social Studies State as the Source of Wealth: In Ottoman Economic Thought: A different approach to reflections in the aftermath of the global crisis will result in drying of tax sources as well as the losses in production strength in long term. The most pressing question to be faced and to be answered is whether it is the state’s or the bank’s future that is important. References Akgündüz, A. (1990). “Osmanlı Kanunnameleri ve Hukuki Tahlilleri”, Fatih Devri Kanunnameleri Kanunname-i Al-i Osman Bab-ı evvel madde 14, Fey Vakfı Yay. İstanbul Aksan, V. (1997). “Savaşta ve Barışta Bir Osmanlı Devlet Adamı Ahmet Resmi Efendi 1700-1783”, Tarih vakfı Yurt Yayınları, İstanbul. Barkan, Ö. L. (1939). “Türk-İslam Toprak Hukuku Tatbikatının Osmanlı İmparatorluğunda Aldığı Şekiller Malikane Divani Sistemi”, Türk Hukuk ve İktisat Tarihi Mecmuası Barkan, Ö. L. (1979). “İstanbul Saraylarına Ait Muhasebe Defterleri”, Türk Tarih Belgeleri Dergisi, T.T.K. Basımevi, c.9, sayı 13, Ankara. Bey M. C. (2001). “İktisat İlmi”, Çev.Sema Alpun Çakmak, Liberte Yayınları, Ankara. B.O.A. Yıldız Tasnifi, Sedaret Resmi Maruzat Evrakı no:1-1 B.O.A. Yıldız Tasnifi, Sedaret Resmi Maruzat Evrakı no:1-40 Cezar, Y. (1986). “Osmanlı Maliyesinde Bunalım ve Değişim Dönemi, 18. yüzyıldan Tanzimata Mali Tarih”, Alan Yayıncılık, İstanbul. Çağatay, N. (1981), “Bir Türk Kurumu Olan Ahilik”, Selçuk Üniversitesi Basımevi, Konya. Çarşılı İsmail Hakkı Uzun, “Çandarlı Vezir Ailesi”, T.T.K. yayınları, Ankara,1988. Çetin. B, (2001). “Osmanlı İmparatorluğu’nda Barut Sanayi 1700-1900”, Kültür Bakanlığı Yay. Ankara. Eldem, V. (1994). “Osmanlı İmparatorluğu’nun İktisadi Şartları Hakkında Bir Tetkik”, T.T.K. Yayınları, Ankara, 1994. Farooqhi, S. (1993) “Osmanlı’da Kentler ve Kentliler”, Tarih Vakfı Yurt Yayınları, İstanbul. Fertekligil, A. (2000). “Türkiye’de Borsanın Tarihçesi”, İ.M.K.B. Yayınları, İstanbul. Findley, C. V. (1996). “Kalemiyeden Mülkiyeye Osmanlı Memurlarının Toplumsal Tarihi”, Tarih Vakfı Yurt Yayınları, İstanbul. Gazi Ahmet Muhtar Paşa, “Sergüzeşt-i Hayatım’ın Cild-i Evveli”, Tarih Vakfı Yurt Yayınları, İstanbul, 1996. Gedikli, F. (1999). “Osmanlı Şirketleri”, Osmanlı, C.3, Yeni Türkiye Yayınları, Ankara. Gökbilgin, T. (1988). “Kanuni Sultan Süleyman Devri Müesseseler ve Teşkilatına Işık Tutan Bursa Şer’iye Sicillerinden Örnekler”, T.T.K. yayınları, İsmail Hakkı Uzunçarşılı’ya Armağan Sayısı, Ankara. Volume 3 Number 2 Fall 2013 129 Birol Çetin Gökyay, O. Ş. (1988). “Risale-i Mimariyye-Mimar Mehmet Ağa Eserleri”, T.T.K. yayınları, İsmail Hakkı Uzunçarşılı’ya Armağan Sayısı, Ankara. Güçer, L. (1987). “16. ve 18. Asırlarda Osmanlı İmparatorluğu’nun Ticaret Politikası”, Türk İktisat Tarihi Yıllığı, İ.Ü. İktisat Fakültesi Türk İktisat ve İctimaiyat Tarihi Araştırmaları Merkezi , S.1, İstanbul. Güran, T. (1997). “Osmanlı devleti’nin İlk İstatistik Yıllığı 1897”, Tarihi İstatistikler Dizisi C.5, D.İ.E Yayınları Ankara. Süleyman, U. (2009). Ibn-i Haldun “Mukaddime” Cilt II. Dergah Yayınları İstanbul. Hülagü, M. M. (1994). “Topal Osman Nuri Paşa Hayatı ve Faaliyetleri 1840-1898”, Ankara Üniversitesi Osmanlı Tarihi Araştırma ve Uygulama Merkezi Dergisi OTAM. Ankara. İlgürel, M. (1991)., “Vakanüvislerin Taltiflerine Dair”, İ.Ü. Edebiyat Fakültesi Tarih Araştırma Merkezi Prof. Dr. Bekir Kütükoğlu’na Armağan Sayısı, İstanbul. İnalcık, H. (1993). “15. Asır Sanayi ve Ticaret Tarihine Dair Vesikalar, Osmanlı İmparatorluğu içinde”, Eren Yayıncılık, İstanbul. Karal, E. Z. (1999). “Selim III’ün Hatt-ı Hümayunları”, T.T.K. Basımevi, Ankara. Kazgan, H. (1995). “Tarih Boyunca İstanbul Borsası”, İ.M.K.B. yayınları, İstanbul. Koraltürk, M. (1999). “Osmanlı Devleti’nde Şirketleşme, İlk Anonim Şirket ve Borsanın Kuruluşu”, Osmanlı, C.3, Yeni Türkiye Yayınları, Ankara. Mustafa, N. P. (1987). “Netayic ül-Vukuat Kurumları ve Örgütleriyle Osmanlı Tarihi”, (sadeleştiren Neşet Çağatay), Türk Tarih Kurumu Basımevi, Ankara, 1987. Öztürk, M. (1987). “Tokat’ta Fiyatlar 1772-1823, Türk Kültüründe ve Tarihinde Tokat Sempozyumu 1986”,Tokat Valiliği Şeyhülislam İbn Kemal Araştırma Merkezi, Ankara . Pakalın. M. Z. (2004). “Osmanlı Tarih Deyimleri ve Terimleri Sözlüğü”, MEB yayınları, İstanbul. Pamukciyan, K. (1958) “İstanbul Ansiklopedisi”, Tan Matbaası, İstanbul. Pamuk, Ş. (1994). “Osmanlı Ekonomisinde Bağımlılık ve Büyüme (1820-1913)”,Tarih Vakfı Yurt Yayınları.İstanbul. Ricaut, (1686). “Türklerin Siyasi Düsturları”, Tercüman 1001 temel eser serisi, (This book was written by Ricaut, Secretary of English Embassy, and its originally published in Amsterdam in 1686). Şekerci, O. (1981). “İslam Şirketler Hukuku Emek Sermaye Şirketi”, Marifet Yayınları, İstanbul. Tabakoğlu, A. (1994), “Türk İktisat Tarihi”, Dergah Yay. İstanbul. Tezel, Y. S. (1986). “Cumhuriyet Döneminin İktisadi Tarihi”, Yurt Yayınları, Ankara. Toprak, Z. (1997). “İktisat Tarihi, Osmanlı Devleti 1600-1908( Türkiye Tarihi C.3)”, Cem Yayınevi, İstanbul. Yücel A. S. www.cumhuriyet.edu.tr/akademik/fak_ilahiyat/makaleler/fusul.htm, (23.4.2003). 130 Journal of Economic and Social Studies Journal of Economic and Social Studies Impact of Military Expenditure and Economic Growth on External Debt: New Evidence from a Panel of SAARC Countries Khalid Zaman, Department of Management Sciences, COMSATS Institute of Information Technology / Abbottabad, Pakistan iqtidar@ciit.net.pk Iqtidar Ali Shah Department of Management Sciences, COMSATS Institute of Information Technology / Abbottabad, Pakistan Muhammad Mushtaq Khan Department of Management Sciences, COMSATS Institute of Information Technology / Abbottabad, Pakistan Mehboob Ahmad Department of Management Sciences, FUIEMS, Foundation University, New Lalazar, Rawalpindi Cantt, Pakistan. AbstrAct This paper examines the impact of military expenditure and economic growth on external debt for a panel of five selected SAARC countries including Bangladesh, India, Nepal, Pakistan and Srilanka, over the period of 1988-2008. Using Pedroni’s (2004) test for panel cointegration, it was found that there is a long-run relationship between external debt, economic growth and military expenditure. The study finds that external debt is elastic with respect to military expenditure in the long run and inelastic in the short run. In the long run, 1% increase in military expenditure increase external debt between 1.18 % and 1.24%, while 1% increases in economic growth reduce external debt between 0.64% and 0.79%, by employed DOLS and FMOLS estimator respectively. In the short run, 1% increase in military expenditure increases external debt by 0.15%, while 1% increase in economic growth reduces external debt by 0.47 %. KEYWOrDs External Debt, Economic Growth, Military Expenditure, Panel Cointegration, SAARC Countries. ArtIcLE HIstOrY Submitted: 2 April 2012 Resubmitted: 28. February2013 Accepted: 25. March 2013 JEL codes: O1, O4, O5 and H5 Volume 3 Number 2 Fall 2013 131 Bilal Sisman / Sinan Yoruk / Ali Eleren Introduction and Literature Review The relationship between military expenditure and economic growth has been examined extensively in the literature. However, the effect of military expenditure on external debt has received less attention. In countries with large military expenditure, the role of military spending in contributing to external debt is important because of the potential adverse economic effects of external debt as excessive foreign debt accumulation can cause deterioration in the terms of trade, an overvaluation of the domestic currency and slower economic growth. Benoit (1973, 1978) in his pioneered study examined the relationship between military expenditure and economic growth in 44 less developed countries and found that there is a positive correlation between military expenditure and economic growth. Dakurah et al. (2000) studied 62 LDCs and found 13 countries showing unidirectional causality from military expenditure to growth; 10 countries from economic growth to military expenditure; 7 countries suggest bidirectional causality and the rest 18 countries displaying no meaningful relationship. Yildirim, Sezgin & Ocal (2005) examined the effect of military expenditure on economic growth for 12 Middle Eastern countries and Turkey using cross-sectional and dynamic panel data estimation techniques from 1989 to 1999 and found that military expenditure enhances economic growth in the Middle Eastern countries and Turkey as a whole. So for as the South Asian Regional Cooperation Council (SAARC) countries are concerned, a study was carried out by Hassan et al. (2003) to show the relationship between military expenditure and economic growth. They examined the impact of the military expenditure on economic growth and FDI covering five out of seven SAARC nations using panel data over the 1980-1999 periods. Interestingly the result suggests positive relationship between military expenditure and economic growth, and thus supporting the view that military expenditure can bring positive impact on growth. Other studies which have also found a positive relationship between military expenditure and economic growth include Mueller and Atesoglu (1993); MacNair et al. (1995), Chlestos and Kollias (1995), Sezgin (1999b, 2000) and Yildirim and Sezgin (2002). Equally military spending may have a negative effect on economic growth through reducing the availability of public funds for spending in the supposedly more productive civilian sector and creating inflationary pressures. Deger (1986) found negative relationship between military expenditure and growth in the less developed 132 Journal of Economic and Social Studies Social Anxiety and Usage of Online Technological Communication Tools among Adolescents countries citing that defense expenditure takes resources away from productive investments and fails to mobilize and create additional savings. Other empirical studies that found significant adverse effect of defense spending on the economy include studies by Deger and Smith (1983), Deger and Sen (1983) and Faini et al. (1984), Antonakis (1997), Heo (1998), Linden (1992), Dunne and Mohammed (1995), Sezgin (1999a) and Dunne, Nikolaidou & Smith (2002). Aizenman and Glick (2006) studied the long-run impact of military expenditure on growth and suggested that military expenditure induced by external threats should increase growth, while military expenditure induced by rent seeking and corruption should reduce growth. Abu- Bader et al. (2003) found that military expenditure had a negative effect on economic growth in Egypt, Israel and Syria over the period 1972 to 2001 within a Granger causality framework. DeRouen (2000) reaches the same findings in a single country study of Israel. Smyth and Narayan (2009) have examined the relationship between external debt and military expenditure nexus in the six Middle Eastern countries and found that external debt is elastic with respect to military expenditure in the long-run while inelastic in the short-run. In this paper an analysis has been carried out to find a panel cointegration between external debt and military expenditure along with economic growth in SAARC countries, using secondary data from 1988 to 2008. This paper does not include all dimensions and factors of the external debt and military expenditure problem from an econometric perspective, the small panel (T=19, N=5) is only sufficient to accommodate two explanatory variables without a substantial loss in power. The objectives of this paper are: 1. To empirically investigate the relationship between external debt, economic growth and military expenditure using a panel unit root and panel cointegration framework in selected SAARC countries. 2. To empirically investigate, whether there is a long-run or short-term relationship between the external debt, economic growth, and military expenditure. This paper is organized as: after introduction and literature review above, a brief overview of external debt, economic growth and military expenditure of the selected SAARC countries is given followed by data source and methodological framework. Next results and discussion has been carried out and conclusion of the study is given at the end. Volume 3 Number 2 Fall 2013 133 Bilal Sisman / Sinan Yoruk / Ali Eleren Overview of External Debt, Economic Growth, and Military Expenditure in the Selected SAARC Countries External Debt: Debt service liability as percentage of export of goods and services has decreased considerably in all the Member States of SAARC countries. In Bangladesh and India debt service liability has been reduced from 25.8 percent and 31.9 percent in 1990 to 3.7 percent and 7.7 percent in 2006 respectively. Pakistan and Sri Lanka have witnessed decline from 22.9 percent and 13.8 percent in 1990 to 8.6 percent and 12.7 percent in 2006 respectively. Figure 1 below shows the trend. Figure 1. External Debt in SAARC Region (1988-2008) EDBAN EDIND 320 150000 280 140000 130000 240 120000 200 110000 160 100000 120 90000 80 80000 40 70000 0 60000 88 90 92 94 96 98 00 02 04 06 08 88 90 92 94 EDNEP 96 98 00 02 04 06 08 02 04 06 08 EDPAK 4000 40000 3500 36000 32000 3000 28000 2500 24000 2000 20000 1500 16000 1000 12000 88 90 92 94 96 98 00 02 04 06 08 02 04 06 08 88 90 92 94 96 98 00 EDSRI 16000 14000 12000 10000 8000 6000 4000 2000 88 90 92 94 96 98 00 Sources: Asian Development Bank, 2008; World Banks, 2008. 134 Journal of Economic and Social Studies Social Anxiety and Usage of Online Technological Communication Tools among Adolescents Economic Growth: SAARC member states have maintained GDP growth rate in 2006 at 8.9%. GDP growth in South Asia is significantly higher compared with other developing regions. However, trickle down effects of growth would take time to effect the population of the region. Real GDP growth rate has increased in almost all the countries. Country-wise analysis shows that Bangladesh’s real GDP growth at 6.2 percent in 1990 increased to 6.6 percent in 2006. Bhutan during the period 1990-2006 witnessed sharp increase from 5.6 percent to 7.8 percent. India maintained its growth momentum from 5.6 percent in 1990 to 9.6 percent in 2006. Nepal economy has witnessed low and high GDP growth from 2.3 percent in 2005 to 2.8 percent in 2006. Pakistan was maintaining its growth but has witnessed a low growth rate of 5.8 percent in 2006. Sri Lankan economy has showed an increase 7.7 percent in 2006 (see, SHRDC, 2008). Figure 2 below shows the trend. Figure 2. Economic Growth of SAARC Countries (1988-2008) GDPBAN GDPIND 24000 1000000 22000 900000 800000 20000 700000 18000 600000 16000 500000 14000 400000 12000 300000 10000 200000 88 90 92 94 96 98 00 02 04 06 08 88 90 92 94 GDPNEP 96 98 00 02 04 06 08 02 04 06 08 GDPPAK 11000 200000 10000 160000 9000 8000 120000 7000 6000 80000 5000 4000 40000 3000 0 2000 88 90 92 94 96 98 00 02 04 06 08 02 04 06 08 88 90 92 94 96 98 00 GDPSRI 40000 35000 30000 25000 20000 15000 10000 5000 88 90 92 94 96 98 00 Source: Human Development Report 2007-08; UNDP, 2008. Volume 3 Number 2 Fall 2013 135 Bilal Sisman / Sinan Yoruk / Ali Eleren Military Expenditure: An increasing trend has been noticed in military expenditures over the time period. Pakistan and India are in the competitive zone, therefore, both have increased their military expenditure. In terms of military expenditure as percentage of GDP, Sri Lanka spent 4.1%, Pakistan 3.5%, India and Nepal 2.5%, Bangladesh 1.5%. Figure 3 shows individual country assessment of military expenditures over a time period. Figure 3. Military Expenditure in SAARC Region (1988-2008) MEBAN MEIND 70000 26000 24000 60000 22000 20000 50000 18000 40000 16000 14000 30000 12000 20000 10000 88 90 92 94 96 98 00 02 04 06 08 88 90 92 94 MENEP 96 98 00 02 04 06 08 02 04 06 08 MEPAK 160 4800 140 4400 120 4000 100 80 3600 60 3200 40 20 2800 88 90 92 94 96 98 00 02 04 06 08 02 04 06 08 88 90 92 94 96 98 00 MESRI 1000 900 800 700 600 500 400 300 200 100 88 90 92 94 96 98 00 Source: UNDP, 2008 and Human Development Report 2007-08. 136 Journal of Economic and Social Studies Social Anxiety and Usage of Online Technological Communication Tools among Adolescents Data Source and Methodological Framework The data set for five SAARC countries is collected from International Financial Statistics (IFS, 2008), World Bank (2008), SHRDC report, (SHRDC, 2008); Stockholm International Peace Research Institute (SIPRI, 2008) and Economic Survey of Pakistan (2008-09). The dependent and independent variables used in this study are listed in Table 1. External Debt is used as a dependent variable for the study. Independent variables are Economic Growth (GDP) and Military Expenditures (ME). Table 1. Variables used for the External debt-Military expenditure Model Variables Symbol Dependent Variable: External Debt Independent Variable: Economic Growth Military Expenditure Expected Sign ED GDP ME Negative Positive Panel Econometric Model: There is lack of panel cointegration to explain the relationship between external debt and military expenditure in the SAARC context. This paper uses panel cointegration analysis to test this relationship in Bangladesh, India, Nepal, Pakistan and Sri Lanka during 1988-2008. The model used to test the relationship between external debt and military expenditure is as follows: ln(ED) = f ln(GDP, ME) The general representation of the equation mentioned above is as follows: Log (Yt ) = C + β 1t log( X 1t ) + β 2t log( X 2t ) + ε t (1) Where: Yt C βt Xt T i εt β1 β2 Volume 3 = = = = = = = = = dependent variable; intercept; slope of the independent variables; independent variables (GDP and ME) 1, 2…21 periods; 1, 2…5 countries; error term; coefficient of economic growth; coefficient of military expenditure; Number 2 Fall 2013 137 Bilal Sisman / Sinan Yoruk / Ali Eleren In the above model, the sign of β1 is expected to be negative as it is argued that SAARC countries might have a capacity to repay external debt. Similarly, β 2 is hypothesized to be positive as it is argued that large military expenditure can result in large external debt. This paper uses a panel cointegration method to examine the long-run relationship between external debt and military expenditure in the selected SAARC countries. Thus, three different panel unit roots tests [(i.e. Levin-Lin- Chu (LLC) test, ImPesaran-Shin (IPS) test and Maddala-Wu (MW) test)] have been used in this study. Panel Unit root tests: Panel unit root tests could be considered as an extension of the univariate unit root test. The LLC test is based on the pooled panel data as follows (Levin & Lin, 1992); Dy it = ρy it −1 + α 0 + σ t + σ i + θ t + ε it (2) Where ρ , α 0 , σ are coefficients, α i is individual specific effect, θ t is time specific effect. According to Levin & Lin (1992), the LLC test could be conducted by the following steps. In step1, subtract the cross-section average from data; N y = 1 / N ∑ y it (3) i =1 In step 2, an ADF test is applied to each individual series and normalizes the disturbance. The ADF model could be expressed as; Pi Dy it = ρ i y it −1 + ∑ δ ij Dy i ,t − j + α i + ε it j =1 (4) Maddala and Wu (1999) argued that this is equivalent to perform two auxiliary regressions of Dy it and y i ,t −1 on the remaining variable n equation (3). Let the residuals from these two regression be eˆi ,t and Vˆi ,t −1 respectively. The, regress eˆi ,t on Vˆi ,t −1 . eˆi ,t = ρ iVˆi ,t −1 + ε it 138 (5) Journal of Economic and Social Studies Social Anxiety and Usage of Online Technological Communication Tools among Adolescents Levin & Lin (1992) suggest the following normalization to control the Heteroskedasticity in error. σˆ 2 ei = 1 (eˆi ,t − ρˆ i − Vˆi ,t −1 ) 2 ∑ T − Pi − 1 t = p + 2 eˆi ,t e~ = σˆ ei v~i ,t −1 ~ vi ,t −1 = σˆ ei In the next step, the LLC test statistic could be obtained from the following regression; ~ ei ,t = ρv~i ,t −1 + ε~i ,t ~ = 0 is given by The t-statistic for testing o tδ = δˆ STD (δˆ ) Where N δˆ = T ∑ ∑ vˆ i =1 t = 2 + p eˆ t −1 it N ∑ ∑ vˆ i =1 t = 2 + p 2 it −1 Next, the paper also employs the IPS test which is based on the mean value of individual ADF statistics or t-bar (Im, Pesaran and Shin, 2003). The IPS test provides separate estimation for each i section, allowing different specifications of the parametric values, the residual variance and the lag lengths. Their model is given by: n DYi ,t = α i + ρ i Yi ,t −1 + ∑ φ k DYi ,t − k + δ i t + u it (6) Volume 3 Number 2 k =1 Fall 2013 139 Bilal Sisman / Sinan Yoruk / Ali Eleren The null hypothesis and the alternative hypothesis are formulated as: H 0 : ρi = 0 H A : ρi < 0 for at least one i Thus, the null hypothesis of this test is that all series are non-stationary process under the alternative that fraction of the series in the panel are assumed to be stationary. IPS also suggested a group mean Lagrange multiplier test for testing panel unit roots. Maddala & Wu (1999) attempted to improve to the same degree the drawbacks of all previous tests by proposing a model that could also be estimates with unbalanced panels. Basically, Maddala and Wu are in line with the assumptions that a heterogeneous alternative is preferable, but they disagree with the use of the average ADF statistics by arguing that it is not the most effective way of evaluating stationary. Panel cointegration tests: Finally, this paper employs Pedroni’s (1999, 2004) panel-co integration method in order to examine the long-run relationship between external debt and military expenditure. If the independent and dependent variables are co-integrated or have a long-run relationship, the residual eit will be integrated of order zero, denoted I(0). Pedroni used two types of panel cointegration tests. The first is the “panel statistic” that is equivalent to a unit root statistic against the homogenous alternative; the second is the “group mean statistic” that is analogous to the panel unit root test against the heterogeneous alternative. Pedroni (2004) argued that the “panel statistic” can be constructed by taking the ratio of the sum of the numerators and the sum of the denominators of the analogous conventional time series statistics. The “group mean statistic” can be constructed by first computing the ratio corresponding to the conventional time series statistics, and then computing the standardized sum of the entire ratio over the N dimension of the panel. This paper uses two panel co-integration tests as suggested by Pedroni (1999, 2004), namely the “panel ADF statistic” and “group mean ADF statistic”. The two versions of the ADF statistics could be defined as: Panel (7) Group Mean (8) 140 Journal of Economic and Social Studies Social Anxiety and Usage of Online Technological Communication Tools among Adolescents Where eˆi ,t represents the residuals from the ADF estimation, ~ s NT is the contemporaneous panel variance estimator, and ŝ i is the standard contemporaneous variance of the residuals from the ADF regression. The asymptotic distribution of panel and group mean statistics can be expressed in: K N ,T − µ N ⇒ N (0,1) v Where K N ,T is the appropriately standardized form for each of statistics, µ m ADF regression is the mean term and v is the variance adjustment term. Pedroni provides Monte Carlo estimates of µ and v (Pedroni, 1999). These statistics are based on the estimated residuals from the following regression: Where ξ it = η i ξ i ( t −1) + µ it are the estimated residuals from the panel regression. The null hypothesis tested is whether η i unity is. The finite sample distribution for the test statistics have been tabulated in Pedroni (2004) using Monte Carlo simulations, if the test statistic exceeds the critical values in Pedroni (2004), the null hypothesis of no cointegration is rejected, implying the variables are cointegrated. Panel Long-run relationship: If long-run relationship among the variables were found then the long-run and short-run coefficients of economic growth and military expenditure on external debt will be estimated. To estimate the long-run effect of economic growth and military expenditure, the panel FMOLS, proposed by Pedroni (2000) and DOLS developed by Stock and Watson (1993) have been used. Results and Discussion To test whether each of ED, GDP and ME contain a panel unit root, the panel unit root tests proposed by Levin, Lin and Chu Test (2002), Im, Pesaran & Shin (2003) and Maddala & Wu (1999) have been applied. The results are reported in Table 2 where they are divided into three panels. Panel A consists of results from the Levin, Lin and Chu (2002); panel B consists of the results from the Im, Pesaran & Shin (2003) test and panel C consists of results from the Maddala and Wu (1999) test. For each of these tests, *, ** and *** indicates the statistical significance at 1 Volume 3 Number 2 Fall 2013 141 Bilal Sisman / Sinan Yoruk / Ali Eleren percent; 5 percent and 10 percent respectively. The results from all three tests, with or without linear trends; suggest that ED, GDP and ME contain a panel unit root as mentioned in Table 2. Table 2. Panel Unit Root Test Levels First Differences Individual Effects Individual Effects Individual Effects and Linear Trends Panel A: Levin, Lin, Chu Test (2002) Variables (in logs) ln(ED) 2.807 2.118 -1.400*** ln(GDP) 2.448 1.725 -1.621** ln(ME) -0.432 0.022 -4.294* Panel B: Im, Pesaran, Shin Test (2003) Variables (in logs) ln(ED) 3.664 2.712 -1.815** ln(GDP) -0.846 0.356 -3.586* ln(ME) -0.071 0.352 -3.256* Panel C: Maddala and Wu (1999) Variables (in logs) ln(ED) 0.015 1.985 20.712** ln(GDP) 12.285 8.253 38.361* ln(ME) 10.234 6.213 29.665* Individual Effects and Linear Trends -1.580*** -1.812** -3.827* -2.495* -3.201* -2.665* 21.321** 32.170* 27.424* * indicates significance at the 0.01 level. ** Indicates significance at the 0.05 level. To examine whether there is a long run relationship between the three variables for the panel of five selected SAARC countries, Pedroni’s (2004) panel Phillips-Perron (1988) type rho-statistic and group Phillips-Perron (1988) type rho-statistic have been employed. The panel rho-statistic and group rho-statistic are 2.2 and 2.7, respectively and the associated one-sided p-value is less than 0.01. Thus, both test statistics suggest that there is panel cointegration between ED, GDP and ME at the 1% level of significance. After finding that a long-run relationship exists between ED, GDP and ME, the long-run effect of GDP and ME on ED have been estimated using the panel FMOLS estimator suggested by Pedroni (2000) and panel DOLS estimator proposed by Kao & Chiang (2000). The results are reported in Table 3. 142 Journal of Economic and Social Studies Social Anxiety and Usage of Online Technological Communication Tools among Adolescents Table 3. Panel Long-run Elasticity Panel Methods ln(ME) 1.182 (8.337)* 1.243 (3.210)* DOLS FMOLS ln(GDP) -0.638 (-11.783)* -0.796 (-3.974)* Note: Figures in parenthesis are t-statistics. * denote statistical significance at the 1 % level. For the DOLS estimator, 1% increase in economic growth decreases external debt by 0.638 %, while a 1% increase in military expenditure increases external debt by 1.18%. Both results are statistically significant at the 1% level. On the other hand, for the FMOLS estimator the coefficient on GDP is 0.796, suggesting that a 1% increase in growth (GDP) decreases external debt by 0.80%. The coefficient of military expenditure (ME) is 1.243, which implies that a 1% increase in military expenditure increases external debt by almost 1.24%. The results for the short-run impact of economic growth and military expenditure on external debt for the panel of five selected SAARC countries are reported in Table 4. Table 4. Panel Short-run Elasticities Variables Constant Coefficient 9.452 t-statistics 19.528* D ln(ME) 0.149 6.102* D ln(GDP) -0.471 -3.761* ECTt −1 -0.092 -2.183** Goodness of fit: R2 = 0.84; R 2 = 0.81 Note: *, ** and *** denotes statistical significance at 1, 5 and 10 % level. Table 4 indicates that economic growth has a negative impact on external debt while military expenditure has a statistically significant positive impact on external debt in the short-run. The coefficient of the military expenditure is 0.149, suggesting that a 1% increase in military expenditure increases external debt by 0.15% respectively. On the other hand, GDP decreases external debt by almost 0.47%. The one period lagged error correction term, which measures the speed of adjustment to equilibri- Volume 3 Number 2 Fall 2013 143 Bilal Sisman / Sinan Yoruk / Ali Eleren um following a shock to the system, has a negative sign and is statistically significant at the 5% level. Its sign and significance level suggests that external debt is able to revert to its equilibrium following a shock to growth and military expenditure. But, the magnitude of the coefficient, because it is very small, suggests that the speed of adjustment to equilibrium is very slow. Overall military expenditure has a positive and significant impact on SAARC external debt in the short and long-run. The relationship is elastic in the long-run, but inelastic in the short-run. Conclusion In this paper a short term and long term impact of military expenditure and economic growth the external debt for five selected SAARC countries; namely, Bangladesh, India, Nepal, Pakistan and Sri Lanka has been examined using data for the period 1988-2008 by applying panel unit root and panel cointegration framework. It was found that the external debt, economic growth and military expenditure were cointegrated for the panel of five SAARC countries. In the long-run, both estimators (DOLS and FMOLS) suggest that economics growth has a statistically significant negative effect on external debt, while military expenditure has a statistically significant positive effect on external debt. Using DOLS estimator, it was found that 1% increase in economic growth decreases external debt by 0.638 %, while a 1% increase in military expenditure increases external debt by 1.18%. Both results are statistically significant at the 1% level. On the other hand, using FMOLS estimator, it was found that 1% increase in growth (GDP) decreases external debt by 0.796%. While 1% increase in military expenditure increases external debt by almost 1.24%. In the short-run it was found that economic growth and military expenditures have a statistically significant negative and positive effect on external debt. In short run it was found that 1% increase in military expenditure increases external debt by 0.15% while 1% increases growth (GDP) decreases external debt by almost 0.47%. One important limitation on our finding is that, from an econometric perspective, the small panel (T=21, N=5) is only sufficient to accommodate two explanatory variables without a substantial loss in power. Future studies for the South Asia as well as other regions in the world could include more potential determinants of external debt within a panel cointegration framework subject to an increase in data availability. 144 Journal of Economic and Social Studies Social Anxiety and Usage of Online Technological Communication Tools among Adolescents References Asian Development Bank (2008). Asian Development Outlook, Asian Development Bank. Available Online at: http://www.adb.org/Documents/Books/ADO/2008/ Update/default.asp accessed on May 2, 2010 Abu-Badar, S. & Aamer S. Abu-Qarn, (2003). Government expenditures, military spending and economic growth: causality evidence from Egypt, Israel, and Syria, Journal of Policy Modeling, 25(6-7): 567-583. Aizenman, J. and Glick, R. (2006). Military expenditure, threats and growth, Journal of International Trade and Economic Development, 15(2), 129-155. Antonakis, N. (1997). Military expenditure and economic growth in Greece, 1960-1990, Journal of Peace Research, 34(1): 89-100. Benoit, E. (1973.). Defense and Economic Growth in Developing Countries. Lexington: Lexington Books. Benoit, E. (1978). Growth and defense in developing countries, Economic Development and Cultural Change, 26(2): 271-280. Chletsos, M. & Kollias, C. (1995). The demand for Turkish military expenditure 1960-1992’, Cyprus Journal of Economics, 8(1): 64-74. Dakurah, A.H., Davies, S.P. and Sampath, R.K. (2000). Defense spending and economic growth in developing countries: a causality analysis, Journal of Policy Modeling, 23, 651-658. Deger, S. (1986). Economic development and defense expenditure, Economic Development and Culture Change, 35,179-196. Deger, S. and Smith, R. (1983). Military expenditure and growth in LDCs, Journal of Conflict Resolution, 27(2),335-353. Deger, S. and Sen, S. (1983). Military expenditure, spin-off and economic development, Journal of Development Economics, 13, 67-83. DeRouen, K. (2000). The Guns-Growth relationship in Israel, Journal of Peace Research, 37(1): 69-83. Dunne, J. Paul & Mohammed, Nadir A. L. (1995). Military spending in Sub-Saharan Africa: some evidence for 1967-1985, Journal of Peace Research, 32(3): 331-343. Dunne, J. Paul; Nikolaidou, E & Smith, R.P. (2002). Military spending, investment and economic growth in small industrializing economies, The South African Journal of Economics, 70(5): 1-27. Economic Survey of Pakistan (2008-09). Finance division, economic advisor’s wing, Government of Pakistan, Islamabad. Faini, R., Annez, P. and Taylor, L. (1984). Defense spending, economic structure and growth: evidence among countries and over time, Economic Development and Culture Change, 32, 487-498. Hassan, M.K., Waheeduzzaman, M. and Rahman. A. (2003). Defense expenditure and economic growth in the SAARC countries, The Journal of Political, Social and Economic Studies, 28(3), 275293. Heo, Uk, (1998). Modeling the defence-growth relationship around the globe, Journal of Conflict Resolution, 42(5): 637-657. IFS (2008). International Financial Statistics 2008, International Monetary Fund, Washington, available at: http://www.imfstatistics.org/imf. Accessed on April 30, 2010 Im, K. S., M. H. Pesaran and Y. Shin (2003). Testing for unit roots in heterogeneous panels. Journal of Econometrics 115, 53-74. Kao, Chihwa and Min-Hsien Chiang (2000). On the estimation and inference of a cointegrated regression in panel data, Advances in Econometrics, 15: 179-222. Volume 3 Number 2 Fall 2013 145 Bilal Sisman / Sinan Yoruk / Ali Eleren Levin, A. & Lin, C. F. (1992). Unit root tests in panel data: asymptotic and finite-sample properties, mimeo, University of California, San Diego. Levin, Andrew; Chien-Fu Lin and Chia-Shang James Chu, (2002). Unit root tests in panel data: asymptotic and finite sample properties, Journal of Econometrics, 108(1): 1-24. Linden, M. (1992). Military expenditure, government size and economic growth in the Middle East in the period 1973-1985, Journal of Peace Research, 29(3): 265-270. MacNair, E.S.; James C. Murdoch; Chung-Ron Pi & Todd Sandler, (1995). Growth and defense: pooled estimates for the NATO Alliance, 1951-1988, Southern Economic Journal, 61(3): 846-860. Maddala G, Wu, S. (1999). A comparative study of unit root tests with panel data and a new simple test. Oxford Bulletin of Economics and Statistics 61: 631–652. Mueller, Michael J. & Atesoglu, H.S. (1993). Defense spending, technological change and economic growth in the United States, Defence Economics, 4: 259-269. Pedroni, P. (2004). Panel cointegration: asymptotic and finite sample properties of pooled time series tests with an application to the PPP hypothesis. Econometric Theory 20: 597–625. Pedroni, P. (1999). Critical values for cointegration tests in heterogeneous panels with multiple regressors, Oxford Bulletin of Economics and Statistics, Special issue, November, 61, pp. 653-70. Pedroni, P. (2000). Fully modified OLS for heterogeneous cointegrated Panels, Advances in Econometrics, 15: 93-130. Phillips P. and Perron, P. (1988); Testing for a unit root in time series regression, Biometrica, Vol. 75, No. 3, pp.335-346. Sezgin, S. (1999a). A causal analysis of Turkish defense growth relationships: 1924-1996, Ankara University Journal of Political Sciences, 55(2): 113-123. Sezgin, S. (1999b). Defence expenditure and economic growth in Turkey and Greece: a disaggregated analysis, Paper Presented at a The Arms Trade, Security, and Conflict conference at Middlesex University Business School, London, 11-12 June. Sezgin, S. (2000). An empirical analysis of Turkey’s defence-growth relationships with a multi-equation model (1956-1994), Defence and Peace Economics, 11(1): 63-80. SHRDC (2008). SAARC Human Resource Development Centre, 2008, Database Profile on Macroeconomic and HRD Indicators in the SAARC Region (1990-2006). Online available at: www. shrdc-isb.org.pk accessed on May 2, 2010. SIPRI (2008). Stockholm International Peace Research Institute, Oxford, Oxford University Press. Smyth, R and Narayan, K, P. (2009). A panel data analysis of the military expenditure-external debt nexus: evidence from six Middle Eastern countries. Journal of Peace Research, Vol. 46, No. 2, 235-250.. Stock, J. and Watson, M.W. (1993). A simple estimator of cointegrating vectors in higher Order integrated systems, Econometrica , 61(4) :783-820. UNDP (2008); Human Development Report 2007-2008, Washington D.C Human Development Report (2007-2008). United Nation Development Program. Available online at: http://hdr.undp.org/en/reports/global/hdr2007-2008/ accessed on May 5, 2010. World Bank (2008). World Development Indicators 2008. Washington D. C. Yildirim, J. and Sezgin, S. (2002). Defense, education and health expenditures in Turkey, 1924-1996, Journal of Peace Research, 39(5): 569-580. Yildirim, J., Sezgin, S. & Ocal, N. (2005). Military expenditure and economic growth in Middle Eastern countries: a dynamic panel data analysis, Defence and Peace Economics, 16(4): 283-295. 146 Journal of Economic and Social Studies Journal of Economic and Social Studies Government Expenditure on Nomadic Education in Nigeria: Implications for Achieving the Millennium Development Goals AKIGHIR David Terfa Department of Economics, Benue State University, Makurdi, Nigeria. akighirdavidterfa@ yahoo.com OKPE. I Department of Economics, Benue State University, Makurdi, Nigeria. AbstrAct The paper examines government expenditure on nomadic education in Nigeria and the implications for achieving the MDGs. Secondary data were used and the data were analyzed with the aid of descriptive statistics. The study revealed that government expenditure on nomadic education in Nigeria over time has been on the increase which has necessitated the increase in the number of nomadic schools and teachers in the country. The study further found out that there is a wide gap between male and female enrolments in nomadic schools in Nigeria; factors such as early marriages and teenage pregnancies, cultural and religious biases as well as economic issues were believed to be responsible for the gap. Also, it was discovered that the total increase in nomads’ enrolments in nomadic schools in the country is not proportionate with the increase in government expenditure on nomadic education. The study attributed this low school attendance by the nomads to the problems of under-funding, dearth of teachers, constant migration of nomads, the involvement of the children of nomads in the productive system, corruption, among others. The study concluded that the present form of implementation of the nomadic education would make it difficult for it to be a panacea for achieving the MDGs in the country. Recommendations were made on how to improve on the nomadic education system in the country. KEYWOrDs Education, Government Expenditure, Millennium Development Goals, Nomads, Nigeria. ArtIcLE HIstOrY Submitted: 20 October 2011 Resubmitted: 10 February 2012 Resubmitted: 12 March 2012 Resubmitted: 24 April 2012 Accepted: 21 May 2012 JEL codes: H5, 015 Volume 3 Number 2 Fall 2013 147 AKIGHIR David Terfa / OKPE. I Introduction Education is the spring board for social and economic change. It plays a major role in the socio-economic development of a nation. Education occupies an important place in most plans for economic and social development. It is important in the human development as a supplier of the trained man power as well as a requisite for the accomplishment of other development goals (Adebiye,2004). These roles played by the educational sector stimulate economic growth and development of a country. This explains why countries of the world expend so much on this vital sector in order to enhance the level of literacy of their citizenry. Inequality of access to education and educational marginalization have deleterious effects on the national development of a country. In Nigeria, however, available records have shown that expenditure on education is below the internationally acceptable standard. According to the UNDP Human Development Report (2008), Nigeria spends almost an insignificant proportion of its financial resources on education, the expenditure on education in Nigeria as a proportion of GDP averaged 5.84 percent, which falls below the UNESCO’s benchmark of 26 percent of the budgets of developing countries. This accounts for the sluggish educational growth rate of 0.59 in the country. The poor funding of education in Nigeria has over time deprived a lot of Nigerians access to education. According to Nafisatu and Abdu (2010), out of the estimated population of 9.4 million nomads in Nigeria,3.3 million are children of school age, but the participation of the nomads in the existing formal and non-formal education programs is abysmally low, with a literacy rate ranging between 0.2% and 2.9%. The Nigerian nomadic pastoralists are made up of the Fulani (5.3m), Shuwa (1.01m), Koyam (32,000), Badai (20,000), Dark Buzzu (15,000) and the Buduma (10,000). The Fulani are found in 31 out of the 36 states of Nigeria, while others reside mainly on the Borno plains and shores of Lake Chad. The migrant fishing groups account for about 2.8 million, comprising numerous tribes. They are found in the Atlantic coastline, the riverside areas and river basins of the country. These groups of people amongst others do not have access to functional education in the country over time. In the quest to remove the chronic illiteracy among this mobile population of Nigeria, the federal government of Nigeria introduced Nomadic Education Program (NEP) in 1986. NEP was designed to provide the nomads with the relevant and fundamental basic education that would improve their survival skills. This was expected to provide them with the knowledge and the skills that would enable them raise their productivity and income; as well as empower them to participate effectively in the socio-economic and political affairs of the country. In a bid to achieving these goals, 148 Journal of Economic and Social Studies Government Expenditure on Nomadic Education in Nigeria: Implications for Achieving the Millennium Development Goals the National Commission for Nomadic Education (NCNE) was established in 1989 with the mandate to: a) formulate policies and guidelines on all matters relating to nomadic education in Nigeria; b) provide funds for research and personnel development for the improvement of nomadic education; and develop programs on nomadic education and provide equipment, instructional materials, construction of classrooms and other facilities for nomadic education (Nafisatu and Bashir,2010). Over the years, the government has been spending money on the nomadic education program so as to provide an unfettered access to quality basic education for the nomads. The aim is to equip them with the skills and competencies that will enhance their wellbeing and participation in the nation-building process. The Nigerian government considers nomadic education as a veritable measure for the development of the universal basic education with a view to achieving the Education for All (EFA) goals and the Millennium Development Goals ( National Commission for Nomadic Education,2002 ). An assessment of the results of the program against its objectives thus far is imperative. Government expenditure on education in Nigeria has been widely studied. However, attention has not been paid specifically on assessing the impact of government expenditure on nomadic education as a measure aimed at achieving the MDGs of universal basic education in Nigeria. At best, available literature presents partial analyses of this issue. Thus, to be area specific, this paper seeks to evaluate the impact of government expenditure on nomadic education in Nigeria with a view to ascertaining whether or not, it will be the key for achieving the universal basic education of MDGs in Nigeria. Following the introduction, the paper is structured as follows: Section two deals with the review of government expenditure on education in Nigeria and the outline of the MDGs. Section three considers the evolution and strategies of nomadic in Nigeria; while section four presents the problems of nomadic education in Nigeria. Section five is the method of the study; section six presents and analyses the data. Section seven uncovers the findings of the study; and section eight contains the recommendations and conclusion of the paper. Government Expenditure on Education in Nigeria Government funding of education in Nigeria comes from different sources. The major one for all levels of government is the public revenue from taxation and oil (Saavedrea, 2003). Education funds are reported to be distributed among the primary, secondary and tertiary education levels in the proportion of 30%, 30% and 40%, respectively ( Balami ,2003). According to Hinclifte (2003) (as cited in Adewale, Volume 3 Number 2 Fall 2013 149 AKIGHIR David Terfa / OKPE. I Ajaji and Enikanoselu, 2005), Government expenditure on education includes direct government expenditure ( for teachers’ salaries and instructional materials) as well as indirect expenditure in the form of subsidies to households such as tax reductions, scholarships, loans and grants. It also includes payment from Education Tax Fund (ETF), mainly for capital expenditure. The main sources of funds that the Nigerian government has are federal taxes and duties on petroleum, profits, imports and exports, which form the revenue of the Federation Account, and the centrally collected Value Added Tax (VAT) introduced in 1996. The federal government allocations to the educational sector from 1995 to 2011 is presented in Table 1. Table1. Federal Government Allocation to Education between1995-2011 Years 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2008 2009 2010 2011 Capital (N) 3,017,900,000 3,215,800,000 3,807,900,000 9,739,600,000 8,291,800,000 35,000,000,000 35,183,789,000 22,100,000,000 15,723,260,401 21,550,000,000 50,540,287,898 40,005,096,425 53,667,933,553 304,670,538,799 Recurrent (N) 9,798,600,000 12,135,900,000 13,033,200,000 13,828,300,000 19,421,700,000 29,514,932,711 37,676,055,443 59,994,441,815 63,228,742,652 72,217,886,839 145,219,839,130 196,218,973,905 192,594,871,801 518,251,289,348 Total ( N) 12,816,400,000 15,351,700,000 16,841,200,000 23,666,100,000 27,713,500,000 64,514,932,711 72,950,836,443 82,094,441,815 78,952,003,053 93,767,886,839 195,760,127,029 236,224,070,330 246,262,805,354 356,495,828,145 Source: Budget Office of the Federation, Federal Ministry of Finance, 2011. The table above shows that government expenditure on education consists of recurrent and capital expenditure. In nominal terms, it can be seen from the table that the budgetary allocations to the education sector are on the increase but the growth rate is not impressive. For instance, in 2008 the allocation to the capital expenditure on education was 6.4% of the total budget and it was 3.9% in 2009, and 3.9% in 2010 representing 0.0% increase between 2009 and 2010; while the educational allocation to the recurrent expenditure was 15.1% in 2008,15.0% in 2009, and 14.4% of the total budget in 2010. This represents a decline of -0.6% in the allocations between 2009 and 2010. These allocations have grossly failed to meet the UNESCO’s 150 Journal of Economic and Social Studies Government Expenditure on Nomadic Education in Nigeria: Implications for Achieving the Millennium Development Goals conventional benchmark of 26% for the budgets of developing countries. Given the importance of this sector to human and economic development, it would be important to push the education allocations up to at least half of the international benchmark requirement so as to attain the universal basic education as described in the goal two of the Millennium Development Goals (MDGs). Outline of the Millennium Development Goals(MDGs) At the Millennium Summit held in September 2000, in New York, United States of America, members of the United Nations (UN) made the following declaration: We will spare no effort to free our fellow men, women and children from the abject and dehumanizing conditions of extreme poverty, to which more than a billion of them are currently subjected to”. This led to the acceptance and the formulation of the Millennium Development Goals which are expected to be fully achievable in the year 2015. These goals are: Goal 1: Eradication of extreme poverty and hunger by half in 2015; Goal 2: Achievement of the universal primary education by 2015; Goal 3: Promotion of gender equality and empowerment of women by 2015; Goal 4: Reduction of child mortality rate especially the under 5 by two-third in 2015, Goal 5: Improvement of the maternal health; Goal 6: Combat HIV/AIDS, malaria and other diseases; Goal 7: Ensure environmental sustainability; and Goal 8: Development of a global partnership for development (National Planning Commission,2004). The quest to achieve the universal basic education as described by goal two of the MDGs, the Nigerian government gave a rekindling interest to the Nomadic Education Program as one of the measures of achieving the MDGs in the country by 2015. Evolution and Strategies of Nomadic Education in Nigeria. The Nomadic Education program (NEP) started officially in November 1986, after The Yola National Workshop on Nomadic Education. The workshop resolved that : “… The nomads needed a fair deal through the provision of education and other social amenities to reciprocate their contribution to the nation-building ….” (Ismail, 2000). Consequently, the Federal Government promulgated the Decree No. 41 of Volume 3 Number 2 Fall 2013 151 AKIGHIR David Terfa / OKPE. I December 1989 (now CAP 243 LFN), which established the National commission for Nomadic Education (NCNE) with the responsibility to implement the National Nomadic Education Program. The NCNE was mandated to formulate policy and issue guidelines in all matters relating to nomadic education in Nigeria. These mandates include: providing funds for research and personnel development of nomadic education in Nigeria; the development of programs on nomadic education; the provision of equipment and other instructional materials, construction of classrooms and other facilities relating to nomadic education. Secondly, to establish, manage and maintain primary schools in the settlements and grazing reserves carved out for nomadic people. Thirdly, to determine standard of skills to be attained in the nomadic schools. Also, to arrange for effective monitoring and evaluation of activities of agencies concerned with nomadic education. It was mandated to liaise and cooperate with other relevant ministries and agencies. Furthermore, NCNE was to receive block grants and funds from the Federal Government or any agency authorized and allocate same to the nomadic schools based on any format approved by the Federal Executive Council; act as agency for channeling all external aid to the nomadic schools in Nigeria; ensure effective inspection of nomadic education activities in Nigeria through the sections in the Federal and State Ministries of Education performing duties relating to nomadic education. Finally, it was mandated to collate, analyze and publish information relating to nomadic education in Nigeria; and undertake any other action desirable for the promotion of nomadic education in Nigeria (National Commission for Nomadic Education, 1989). The commission has four departments, namely, Program Development and Extension; Monitoring, Evaluation and Statistics; Administration and Supplies and Finance and Accounts. It has six Zonal offices located in Bauchi for the North-East, Kano for the North-West, Minna for the North-Central, Ibadan for South-West, Enugu of the South-East and Benin for the South-South. Furthermore, it has four university-based nomadic education centers located in the University of Port Harcourt for migrant fishermen education, University of Maiduguri for teacher-training, University of Sokoto for curriculum development, and University of Jos for research and evaluation to cater for pastoral nomads (National Commission Nomadic Education, 2000). The Organogram of the Commission is shown in Appendix I. Nomadic Education Program Strategies In pursuance of its functions, the National Commission for Nomadic Education has between 1990 and 2006 evolved four distinct programs for the effective implemen- 152 Journal of Economic and Social Studies Government Expenditure on Nomadic Education in Nigeria: Implications for Achieving the Millennium Development Goals tation of the nomadic education program. The programs briefly highlighted are: Provision of primary education – this program is implemented in collaboration with States and Local governments, as well as local communities, Non-Governmental Organizations (NGOs) and Collaborative Body Organizations (CBOs). So far, all the 36 States and Abuja are participating in the program. However, their level of commitments to the program varies. Provision of academic support services through the University based centers. The centre at Jos is responsible for research and evaluation, the University of Maiduguri for teachers training and outreach programs, Usmanu Dan Fodiyo University for the development of curricula and textual materials and the University of Port Harcourt for research, curricula development and teachers training for the education of migrant fishing communities. The Commission has been working closely with the Centers for the development of curricula materials and pupils’ texts, conduct of research projects and the organization of teachers training workshops; Provision of infrastructural facilities – faced with the problem of inadequate infrastructure that has necessitated teaching and learning under trees, the Commission has adopted the following strategies for addressing the peculiar needs of all the groups: provision of permanent and semi-permanent structures, provision of mobile collapsible classroom structures, provision of boat schools and dug-out canoes. Provision of extension services: Actual intervention by the Commission in the provision of educational extension services to the nomads only began in 1996/97. The major driving force behind the NCNE’s intervention in this aspect of education was the realization that, the adoption of an integrated approach to education provision engender nomads participation in support for the program. However, the mandate of the Commission specifically restricted its operation to the provision of primary education to the children of the nomads. The dilemma necessitated the convening of an Experts Meeting on viable strategies for implementing Nomadic Education in Nigeria in 1995. Having examined the condition of nomadic education in Nigeria, the meeting came out with recommendations to further strengthen, expand and sustain the program. Some of the recommendations were that, the provision of education should be for the children as well as the adults, and that tremendous efforts be made to positively alter the behaviour of nomads towards modern education. Current Program Implementation To support effective teaching and learning, the Commission also collaborates with Nomadic communities, CBOs, NGOs, at all levels, development partners and other Volume 3 Number 2 Fall 2013 153 AKIGHIR David Terfa / OKPE. I international support organizations. Over the years, the Commission has embarked on a number of activities and recorded modest achievements in the following areas, namely: Broadening access to basic education – providing access to basic education, the NEP has facilitated 2,354 schools in 36 States and FCT for pastoralist children (432,411), 451 for fisher folk children (88,288) in 9 states, 260 schools for migrant farmers in 8 states with 33,164 pupils; The Commission has facilitated the increase in the number of nomadic schools. The number of Nomadic schools increased from 2,094 in 2005 to 2,294 in 2006 and to 2,526 in 2007. There was progressive teachers recruitment and retention in nomadic schools, there was an increase in the number of nomadic schools teachers from 6,918 in 2005, to 7,989 in 2006 and 8,665 by 2007; there was an increase in Nomadic girl-child education initiative of the Commission which has increased female enrolment, progression, and transition in Nomadic schools. The female enrolment has increased from 153,489 in 2006 to 164,769 in 2007 representing 28% increase rate (Nafisatu and Abdu, 2010). Problems of Nomadic Education in Nigeria According to Nafisatu and Abdu (2010), the following are the problems militating against the smooth operation of the NEP in Nigeria: constant migration of the nomads, the involvement of children in the productive systems, unsuitability of the formal school curriculum, physical isolation and restriction of the nomads from social interaction with the larger society, unfavorable land tenure system, underfunding of nomadic education and late release of approved funds, unwillingness of State and Local Governments to make budgetary allocations for Nomadic Education Program, indiscriminate transfers of the teachers by LGEAs from the Nomadic primary schools to conventional primary schools without replacements, the dearth of teachers in terms of quantity and quality, relatively low level of enrolments in Nomadic schools, general lack of supervision and monitoring of nomadic schools by the local and state governments, relative exclusion of Nomadic schools from UBE and other intervention funds accruing to the States, Non-provision of funds for the Commission’s extension service programs, constant clashes and conflicts between farmers and herders and amongst fisher folks over fishing rights resulting in displacements. According to Ismail (2000), the under-funding of nomadic education is partly blamed on inaccurate demographic data. Lack of reliable statistics on the nomads leads to planning based on guessing; there was much confusion as to the actual number of the nomadic schools, types of school facilities and the 154 Journal of Economic and Social Studies Government Expenditure on Nomadic Education in Nigeria: Implications for Achieving the Millennium Development Goals number of teachers in various locations. Lack of authentic data in these areas has made planning for nomadic education very difficult. Schools are stationed inappropriately; few in densely populated areas, and many in sparsely populated areas. Malinga (2009), observed that, the major hindrances to school attendance are the daily grazing movements and the lack of labor substitutes. Unlike farmers who use child labor marginally, the Fulani rely heavily and continuously on the children for labor. A Fulani man will not send his child to school even if an adult is available to attend to the animals because the child needs to learn the herding skills. The reliance on juveniles for shepherding task, explains the poor participation of the pastoralists in formal education. From the foregoing, it is apparent that Nafisatu and Abdu, Ismail, and Malinga have identified various problems of nomadic education in Nigeria. However, in this study, the problem of underfunding of nomadic education, dearth of teachers in terms of quantity and quality, corruption, constant migration of the nomads, and active involvement of the school-going age children of the nomads in the productive system were ranked as the most important problems. Thus, these problems were critically engaged in the section of data analysis. Method of the Study The study used mainly secondary data that were obtained from the National Commission on Nomadic Education (NCNE) and the Federal Ministry of Finance as well as journals. The data collected were on the number of pupils enrolment, number of nomadic primary schools, number of teachers in nomadic schools and government expenditure on nomadic education in Nigeria from 1990 to 2008 as well as the corruption perceptions indices in Nigeria. The data were analyzed using tables, percentages, trend graphs, bar charts and rates. Data presentation and Analysis In order to evaluate the impact of government expenditure on nomadic education in Nigeria, data were collected on the number of schools, teachers and the pupils’ enrolments in nomadic schools from 1990 to 2008 as presented in table 2. Volume 3 Number 2 Fall 2013 155 AKIGHIR David Terfa / OKPE. I Table 2. Distribution of the of nomadic schools, Teachers and Pupils enrolment from 1990 to 2008 Years 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 No of schools 329 473 626 656 754 860 940 1,103 1,022 1,068 1,494 1,574 1,680 1,820 1,981 2,034 2,354 2,354 2,526 No of teachers 878 1,489 2,491 2,365 2,822 2,788 2,915 3,265 3,265 3,365 4,748 4,907 5,290 6,306 6,861 6,918 7,989 7,989 8,665 Pupils enrolment Male 13,763 25,942 33,463 38,335 42,738 56,759 63,638 71,695 69,578 75,601 112,958 118,905 134,930 175,962 211,931 222,061 224,304 224,304 235,064 Female 5,068 10,559 16,689 15,253 19,094 35,751 40,938 47,081 47,366 46,934 80,291 84,939 92,014 127,556 151,622 153,489 164,769 164,769 197,347 Total 18,831 65,019 50,152 53,588 61,832 92,510 104,576 118,776 116,944 122,535 193,243 203,844 226,944 303,518 363,553 375,550 389,073 389,073 432,411 Source: NCNE, 2008 The table reveals that the number of nomadic schools in the country has increased from 329 in 1990 to 2,526 in 2008 representing a cumulative increase of 87% in the number of nomadic schools. It further shows that, the number of teachers employed to handle teaching and learning in the nomadic schools has increased from 878 in 1990 to 8,665 in 2008 representing 90% increment in the number of teachers employed during the period. The table also depicts that the number of pupils’ enrolment in nomadic schools rose from 18,831 in 1990 to 432,411 in 2008, representing a cumulative increase of 96% in the total pupils’ enrolment over the period. To show clearly these increasing trends, data from table 2 were used to construct trend graphs as shown in Figures 1 and 2. The trends show clearly that both the number of nomadic schools and teachers have increased over time in Nigeria. The main reason for the trend may be the government’s continued grant-in-aid to the nomadic education. 156 Journal of Economic and Social Studies Government Expenditure on Nomadic Education in Nigeria: Implications for Achieving the Millennium Development Goals Figure 1 Trends of the number of Nomadic schools and the number of Teachers employed over time The chart in Figure 2 further shows that the number of pupils’ enrolment by gender in nomadic education in Nigeria is on the increase. Source: Constructed from the data in table 2 Volume 3 Number 2 Fall 2013 157 AKIGHIR David Terfa / OKPE. I Figure 2. Pupils’ enrolments in Nomadic school by gender A close examination of the chart reveals that there is a wide gap between male and female enrolments over the years. In a bid to explain the gender gap in nomadic education in Nigeria, we considered the general causes of low female school enrollments in Nigeria. Some of the factors are early marriages and teenage pregnancies. These are common experiences in the country, especially in the northern part of the country where the nomads are dominant. In that part of the country, many school-age girls often drop out of school because of pregnancies to marry. Secondly, cultural and religious biases adversely affect girl-child education in Nigeria. Many Nigerian parents, especially in large families with limited resources, tend to enroll boys in school instead or before girls. Some parents also keep their daughters out of schools due to misinterpretation of the tenets of the Islamic religion. This practice is typical of illiterate Muslims of which the nomads are part of. They generally believe that their female children will face sexual harassment in schools. Nonetheless, poverty and economic issues are equally contributory factors to this gap. Given the high level of poverty in Nigeria, many parents, including the nomads, often send their daughters to sell wares in the market or on the street in order to generate additional incomes for the families. For the Fulani nomads, their school-age daughters are commonly involved in the hawking of extracted cow milk. These factors are responsible for the disproportionate male-female enrolments in Nigerian schools, especially at the primary school level. Thus, it may be said that nomadic education in the country also faces these general problems. This male-female gap in school enrolments has a very serious implication for attaining the two educational Millennium Development Goals of Universal Primary Education(UPE) and the elimination of gender disparities in the primary and secondary schools in 2015. This is so because, the EFA goals and MDGs in Nigeria aimed at raising the gender parity rate to 80% in primary 1-6 and 50% in JS1-3 by 2015, using nomadic education as a potent tool. Furthermore, the growth rate of government expenditure on nomadic education in Nigeria was compared with that of pupils’ enrolments in nomadic schools. The results are presented in table 3. 158 Journal of Economic and Social Studies Government Expenditure on Nomadic Education in Nigeria: Implications for Achieving the Millennium Development Goals Table 3. Government Expenditure on Nomadic Education and Pupils Enrollment in Nigeria % D in EXP % D in Enrollment YEAR EXP(N) 1990 4227139.00 1991 5284802.00 25.02 65019.00 245.28 1992 2958582.00 -44.02 50152.00 -22.87 1993 11225544.00 279.42 53588.00 6.85 1994 6930438.00 -38.26 61832.00 15.38 1995 3153896.00 -54.49 92510.00 49.62 1996 8929536.00 183.13 104576.00 13.04 1997 8876172.00 -0.60 118776.00 13.58 1998 6613698.00 -25.49 116944.00 -1.54 1999 15676272.00 137.03 122535.00 4.78 2000 17382572.00 10.88 193243.00 57.70 2001 81352364.00 368.01 203844.00 5.49 2002 87872301.13 8.01 226944.00 11.33 2003 50000000.00 -43.10 303518.00 33.74 2004 51163143.22 2.33 363553.00 19.78 2005 59890663.01 17.06 375550.00 3.30 2006 70162576.31 17.15 389073.00 3.60 2007 70373063.00 0.30 389073.00 0.00 2008 70584183.00 0.30 432411.00 11.14 ENROLLMENT 18831.00 Source :NCNE, Annual Report,2000. The table shows that government expenditure on nomadic education has been on the increase over the years, even though the increase has not been consistent. As it can be seen from the table, in nominal terms, the expenditure has increased over time from N4,227,139.00 in 1990 to N70,548,183.00 in 2008. However, in terms of the growth rate, the trend of the expenditure on nomadic education has not demonstrated any definite trend. For instance, in 1991, the expenditure increased by Volume 3 Number 2 Fall 2013 159 AKIGHIR David Terfa / OKPE. I 25.05% and declined by 44.02% in 1992, it then rose tremendously by 279.42% in 1993 and thereafter, decreased by 38.26% in 1994. The declining trend continued and again, appreciated by 183.13% in 1996 and peaked in 2001 by 368.01% . The Expenditure dropped by 43% in 2003 and afterwards, increased moderately. Pupils’ enrolments as depicted in the table increased continuously during the review period except in 1992 and 1998 when a decline was recorded. The enrolments increased from 18,831 pupils in 1990 to 43,244 pupils in 2008 representing a cumulative increase of 56.45% in the number of pupils enrolled. In order to clearly see whether the increases in the government expenditure are commensurate with the changes in the enrolments, percentage rates for the expenditure and enrolments over time as contained in the table 3 were used to construct a trend graph as shown in Figure 3. Source: Constructed from the data in Table 3. Figure 3. Trends in Percentage change in Expenditure on Nomadic Education and School Enrollment in Nomadic school in Nigeria A close examination of the trends reveals that the percentage increases in school enrolments by the nomads are not proportionate with the increases in government expenditure on nomadic education over time. Having discovered this, we tried to ascertain the possible causes for the disproportionate relationship between government expenditure on nomadic education and nomadic school enrolment in Nigeria. In doing this, we have engaged some variables critically to see whether they are 160 Journal of Economic and Social Studies Government Expenditure on Nomadic Education in Nigeria: Implications for Achieving the Millennium Development Goals responsible for this relationship . First, we considered the problem of underfunding of nomadic education in the country. We used the ratio of government allocations to nomadic education as the proportion of the total education allocations in the country over the years as a proxy for underfunding problem as shown in the table 5. Table 5. The ratio of Government Allocations to nomadic Education as a proportion of the total allocation to the Education sector in Nigeria. Year Government allocation to Education (N) Government allocation to Nomadic Education (N) 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 12,816,400,000 15,351,900,000 16,841,200,000 23,666,100,000 27,713,500,000 64,514,932,711 72,950,836,443 82,094,441,815 78,952,003,053 93,767,886,839 195,760,127,029 236,224,070,330 246,262,805,535 356,495,828,145 3,153,896.00 8,929,536.00 8,876,172.00 6,613,698.00 15,678,272.00 17,382,572.00 81,352,364.00 87,872,301.13 50,000,000.00 51,163,145.22 59,890,663.01 70,162,576.31 70,373,063.00 70,584,183.00 Ratio of Government allocation to Nomadic Education as a percentage of Total allocation to Education (%) 0.02 0.06 0.05 0.06 0.03 0.03 0.11 0.06 0.06 0.05 0.03 0.03 0.03 0.02 Source: Computed from Tables 1 and 4 The table shows that government expenditure on nomadic education as a percentage of the total allocations to the education sector averaged 0.05%. This proportion is indeed ,too meager to exert any meaningful impact on nomadic education as per increased enrolment rates. Secondly, we considered the teacher-student ratio as a proxy for the problem of inadequacy of teachers in nomadic schools over the years as shown in the table 6. Volume 3 Number 2 Fall 2013 161 AKIGHIR David Terfa / OKPE. I Table 6. Teacher-Student Ratio in nomadic Education Schools in Nigeria Year 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 Number of Teachers 875 1489 2491 2365 2822 2788 2915 3265 3265 3365 4748 4907 5290 6306 6861 6918 7989 7989 8665 Number of Pupils 18,831 65,019 50,152 53,588 61,832 92,510 104,576 118,776 116,944 122,535 193,243 203,844 226,944 303,518 363,553 375,550 389,073 389,073 432,411 Teacher-Student ratio 25 44 20 23 22 33 36 37 36 36 41 42 42 48 53 55 49 49 50 Source: Computed from Table 2 By taking the average of the teacher-student ratio over the years as contained in the table 6, it was discovered that the ratio is 1:40, implying that there are 40 students to a teacher. These number of students to a teacher is relatively high since such a teacher may find it difficult to discharge responsibilities effectively as required. The net effect may be abstentism by the students due to ineffective control over them which could degenerate into subsequent drop- out of many nomads of school-age from school. Furthermore, we examined the menace of corruption, where we used the Corruption Perceptions Indices with a view to demonstrating how corrupt practices affect public funds in Nigeria, of which nomadic education is no exception. Transparency International introduced the corruption perceptions Index in 1995 to measure the level of corruption in countries. The CPI scores relate to the perceptions of degree of corruption as seen by business people and country analysts. The scores range between 10( highly clean) and 0 (highly corrupt). The CPI scores for Nigeria from 2003 to 2010 are shown in the table 7. Table 7. The Corruption Perception Indices for Nigeria Year CPI 2003 1.6 2004 1.4 2005 1.6 2006 1.9 2007 2.2 2008 2.2 2009 2.7 2010 2.8 Source: Transparency International, 2011 162 Journal of Economic and Social Studies Government Expenditure on Nomadic Education in Nigeria: Implications for Achieving the Millennium Development Goals From the above table, Nigeria has an average of 2.0. This implies that Nigeria is hypercorrupt. The implication is that, even with the meager allocations to nomadic education, corrupt practices may have weakened the supposedly positive impact. Thirdly, we looked at the constant migration of the nomads as a factor that may have caused low school attendance of the nomads. The nomads are typically people that travel from one place to another due to the nature of their occupations. These constant migrations have deleterious impact on school attendance of their children. Once they relocate from a given place, they abandon the schools their children were attending. Thus, it normally takes the children some time to settle down and start school in the new settlement. Lastly, we took cognizance of the factor of active involvement of the children of the nomads in the productive system. The mentality of nomads is such that they believe in training their young ones who are of school-age in their trades. This practice has made them to place more preference for their productive system than the education of their children. Consequent upon these analyzed factors, we submit that low school attendance of the nomads may be as a result of the synergy of these factors. Generally, the growth rates in school enrolments of the nomads in Nigeria is not encouraging because, of the estimated 3.3 million nomads of school going age, the available statistics show that only 432,411 nomads were enrolled for nomadic education as at 2008. This implies that, about 2,867,589 nomads representing 86.9% of the school-age nomads were not going to school. The implication is that, with all the efforts made by the government over time to ensure equal access to education among the various groups in the country with a view to achieving the MDGs in the country by 2015; only 13.1% of the total nomads of school-age were going to nomadic school as at 2008. This school enrolment rate of the nomads is too low to make any meaningful contribution to the overall achievement of the universal basic education as described in the Millennium Development Goal two. In Nigeria generally, the Millennium Development Goals are influenced by some of these socio-economic and political factors: First, there is a disconnection between the tiers of government in the implementation of the MDGs. However, the constitutional responsibility for the implementation of almost all the goals rest with the States and Local governments in Nigeria’s Federal structure; but in spite of remarkable strides at Federal level, appreciation of the requirements for meeting these goals, as well as institutional capacity remain relatively low at these levels of government. Poor governance and integration of the MDGs into national development strategies have also been a challenge. This is aggravated by policy inconsistencies, for instance, Obansanjo regime introduced NEEDS I and II, Yar’adua’s administration instituted Seven-point Agenda and Vision 20:20:20 and President Goodluck has now introduced Transformation Agenda. Volume 3 Number 2 Fall 2013 163 AKIGHIR David Terfa / OKPE. I All these policies are at variance in principle but targeted at achieving the MDGs in the country. Other challenges include, lack of transparency and accountability in ministries, lack of the political will; a weak monitoring mechanism for the MDGs and low stakeholders involvement( private sector and civil society organizations ). Lastly is the unavailability of up to date data on most of the indicators. This is compounded by the limited funding available for data generation and management in the country. Findings of the Study Emergent from the above discussion, it was found out that the government over the years has demonstrated concern to ensure equality in the literacy level among the various groups in the country by initiating nomadic education in order to boost the literacy rate among the nomads who are educationally disadvantaged. This concern can be seen from the perspective of continued and sustained expenditure on nomadic education over the years. It was discovered that increases in government expenditure on nomadic education have precipitated increases in the number of nomadic schools in the country across the states vis-à-vis the number of teachers employed by the government to ensure smooth teaching and learning in nomadic schools in the country. The study revealed that enrolments by the nomads in the schools have also increased over time but not proportionate with the increases in the government expenditure. Factors such as underfunding, dearth of teachers, corruption, constant migration of the nomads and active involvement of the children in the productive system by the nomads may have been responsible for this gap. It was also found out that, there is wide gap between male and female enrollments in nomadic schools over the years. Factors such as early marriages and teenage pregnancies, cultural and religious biases as well as poverty and economic issues may have been responsible for gender gap in the enrolments in nomadic schools in the country. Conclusion and Recommendations Based on the above findings, it is concluded that the present implementation of the nomadic education in the country may make it difficult for it to be a panacea for achieving Millennium Development Goals(MDGs) in terms of universal basic education attainment by 2015. This is so because, the growth focus of the Millennium Development Goals( MDGs) is more concentrated at the importance of achieving clear and real progress in human capital development measured through educational 164 Journal of Economic and Social Studies Government Expenditure on Nomadic Education in Nigeria: Implications for Achieving the Millennium Development Goals foundation. Consequent upon the foregoing, the following recommendations are submitted: First, there should be continuous mobilization and sensitization of the nomads to send their children to these schools especially the female children so as to bridge the gap between the male- female enrolments in nomadic schools. Second, the states and local governments should be made to supplement the federal government funding of the nomadic education, as this will go along way boosting the achievements in nomadic education in the country. Third, nomadic educational development initiatives should be planned and aligned with other community improvement and development programs such as agricultural extension, rural development and social welfare services. This approach will attract the interest and involvement of more stakeholders as this will encourage the stakeholders to support the program. Fourthly, there should be selection of more individuals from nomadic communities for training as teachers, this is because they are more acquainted with the cultural values of the nomads. Thus, they will be better placed to inculcate the necessary knowledge in the targeted nomads. Fifthly, the government should provide support to animal health issues especially on major diseases and vaccines as well as provision of support in the area of water development for improvement of livestock production and reduction of incident of conflicts between the nomads and the hosting communities. This practice will make the nomads more stable in a place to receive nomadic education. Sixthly, government should seek more new partnerships and greater collaboration with development partners and other stakeholders in other to boost the financing of nomadic education for higher results. Seventhly, NCNE should adopt a flexible timetable that adjusts itself to seasons favorable to nomads, as this would enhance their enrollments. Finally, the government should actively consider the issue of language barrier to communication, and find contextually appropriate language policies. References Adebiyi, M.A. (2004). Public Expenditure and human capital in Nigeria: An Autoregressive Model. International Journal of Economics, 1(3), 86-77. Volume 3 Number 2 Fall 2013 165 Adewale, T.M, Ajayi, K.O & Enikanoselu,O.A. (2005). Trends in the Federal Government Financing Education in Nigeria. Journal of Education Management, 1(1), 35-41. Association for the development of Education in Africa (1999). Ensuring Access and Equity in Basic Education for Nomads in Nigeria. www.adeanet.org/adeaportal/adea/program/.../pstr99-nigeria. pdf on 28th August,2001. Balami, D. (2002). Finance of Education in Nigeria. Paper Presented at the Forum on Cost and Finance of Education in Nigeria at Women Development Centre, Abuja,12th-14th May, Abuja. Fashima, D. (2005). Reforms in Nigeria University system: What Direction? National Freedom, 1 (6), 30-36. Ismail ,I. (2000). Nomadic Education and Education for Nomadic Fulani. Retrieved from www. gamji.com/fulani7.htm. on 15th August, 2011. Federal Ministry of Finance (2011). Federal Government of Nigeria Annual Budgets. A publication of Ministry of Finance, Abuja, Nigeria. Hinchcliffe, K. (2003). Public Expenditure on Education in Nigeria: Issues, Estimates and some Implications, Abuja, World Bank Publication. Malinga, M. F. (2009). Flexible education Reaching nomadic children in Africa. Retrieved from www. cedol.org/cgi-bin/items.cgi on 15th August, 2011. Nafisatu, P. & Abdul, B. (2010). Enhancing Livestock Development and Means of Existence for the Vulnerable Populations through Education and Training. A case of NEP in Nigeria. A Paper presented at the CILSS, RPCA and OCDE forum on Livestock breeding and food Security in Accra, Ghana. December 14th – 16th, 2010. National Commission for Nomadic Education (1989). “ Federal Government intervention in the Education of Nomads in Nigeria, 1996 – 1999”. Memorandum submitted to the panel on streamlining/ Rationalization of poverty Alleviation institutions/Agencies by the National Commission for Nomadic. A publication of NCNE, Abuja, Nigeria. National Commission for Nomadic Education (1989). Decree No.14 of 1989, NCNE publication Abuja, Nigeria. National Commission for Nomadic Education (2000). Action plan 2000-2005,NCNE Publication, Abuja, Nigeria. National Commission for Nomadic Education (2002). Education Financial Record 1990-2002, NCNE publication, Abuja, Nigeria. National Commission for Nomadic Education ( 2002). Monitoring Report 2002, NCNE publication, Abuja, Nigeria. National Commission for Nomadic Education (2009). Education Financial Record 2003-2008, NCNE publication, Abuja, Nigeria. National Planning Commission (2004). National Economic Empowerment and Development Strategy (NEEDS) Policy Document, Published by NPC, Abuja, Nigeria. Saavedrea, M.A. (2003). Debt service- Education Expenditure Nexus: The Nigerian Experience. Selected Papers for the year 2002 Annual Conference, The Nigerian Economic Society(NES) PP,243-267. Tahir, G. (1996). Vision and Mission of Nomadic Education in Nigeria: paper delivered during the NCNE, National conference on Vision and Mission of Nigerian Education; at the NTI 22nd -25th July 1996 , Kaduna. Tahir, G., Nafisatu, D.M & Ahmed, M.M (2005). Improving the quality of Nomadic Education in Nigeria: Going Beyond Access and Equity. A study conducted by the Federal Ministry of Education of Nigeria, 2005 in collaboration with Association for the Development of Education in Africa (ADEA). Retrieved at www.adea net.org. on 15th August 2011. Tahir,G. (1998). “ Nomadic Education in Nigeria: Issues, problems and prospects”. Journal of Nomadic studies,(1)1, 10-26. Transparency International (2011). Corruption Perception Index for Nigeria. Retrieved from www. nigerianmuse.com/2011201145701zg/nigeria-watch. 28th January,2012. UNDP(2008). Human Development Report on Nigeria, New York Appendix I Figure1. An Organogram showing the organizational Chart of National Commission for Nomadic Education ORGANISATIONAL CHART OF NATIONAL COMMISSION FOR NOMADIC EDUCATION MINISTER OF EDUCATION BOARD EXECUTIVE SECRETARY Finance & Accounts Administration General Admin. Establishment Transport Monitoring Supply Budget & Funds North West Zone North Central Zone North East Zone Academic Planning Computer Unit Documentation Unit Evaluation Statistics South South Zone Legal Services Unit Information Unit Animal Husbandry & Extension North East Zone Special Services Section Mobilization & Public Enlightenment University Centers Graphics Unit Usmanu Danfodiyo University, Sokoto Source: NCNE, 2000 Field Operation Final Accts. South West Zone Physical Planning & Maintenance Programme Development & Extension Monitoring Evaluation & Statistics University Of Jos University Of Maiduguri University Of Port-Harcourt Audit Unit LIST OF REVIEWERS FOR THIS ISSUE AKCA Hasan Cankiri Karatekin University, TURKEY AKIN, Zafer TOBB University of Economics and Technology, TURKEY AKYUZ, Mursel Suleyman Demirel University, TURKEY ALVAN, Arzu Gediz University, TURKEY COSKUN, Ali Fatih University, TURKEY DEMIROZ Fatih Florida International University, USA DRAGIČEVIĆ Mirjana University of Zagreb ERDEM, Cumhur Gaziosmanpasa University, TURKEY ERGUN, Ugur International Burch University, BOSNIA AND HERZEGOVINA HANDZIC Meliha International Burch University, BOSNIA AND HERZEGOVINA JOUINI, Fathi University of Sousse, TUNISIA KANIBIR, Huseyin Bursa Orhangazi University, TURKEY KORKMAZ, Turhan Bulent Ecevit University, TURKEY KOSEKAHYAOGLU, Levent Suleyman Demirel University, TURKEY MOHSIN Amena Independent University Bangladesh OGUMS, Raymond Investment Operations, USA OZLEN Kursad International Burch University, BOSNIA AND HERZEGOVINA PASIC Mugdim University of Sarajevo, BOSNIA AND HERZEGOVINA TATIC, Kasim Sarajevo University, BOSNIA AND HERZEGOVINA TEBERLER, Metin, Istanbul Technical University, Turkey