Mid-Sized Italian Manufacturing Firms: A Panel Data Analysis on Profitability

ASERS Journal of Advanced Studies in Finance J ournal of Advanced Studies in Finance Biannually Volume VII Issue 2(14) Winter 2016 ISSN 2068 – 8393 Journal DOI https://doi.org/10.14505/jasf Volume VII Issue2(14) Summer 2016 Journal of Advanced Studies in Finance is an advanced e-publisher struggling to bring further worldwide learning, knowledge and research. This transformative mission is realized through our commitment to innovation and enterprise, placing us at the cutting-edge of electronic delivery in a world that increasingly considers the dominance of digital content and networked access not only to books and journals but to a whole range of other pedagogic services. In both books and journals, ASERS Publishingis a hallmark of the finest scholarly publishing and cutting-edge research, maintained by our commitment to rigorous peer-review process. Using pioneer developing technologies, ASERS Publishing keeps pace with the rapid changes in the e-publishing market. ASERS Publishing is committed to providing customers with the information they want, when they want and how they want it. To serve this purpose, ASERS publishing offers digital Higher Education materials from its journals, courses and scientific books, in a proven way in order to engage the academic society from the entire world. 92 Winter 2016 Volume VII Studies in Finance Journal of Advanced Issue 2(14) Editor in Chief Laura GAVRILĂ (formerly) ŞTEFĂNESCU Spiru Haret University, Romania Co-Editor Rajmund MIRDALA Technical University of Kosice, Slovak Republic Contents: 1 Government Block holder Ownership, Sovereign Wealth Fund, and Firm Performance Kerry LIU …95 Editorial Advisory Board Mădălina Constantinescu Spiru Haret University, Romania 2 Wavelet Based Analysis of Major Real Estate Markets Adil YILMAZ, Gazanfer UNAL, Cengiz KARATAS Rosaria Rita Canale University of Naples Parthenope, Italy Francesco P. Esposito AlliedIrish Bank, Group Market Risk Management Lean Hooi Hooi Universiti Sains Malaysia, Malaysia Terence Hung United International College, Hong Kong Renata Karkowska Faculty of Management, University of Warsaw, Poland Kosta Josifidis University of Novi Sad, Serbia Ivan Kitov Russian Academy of Sciences, Russia 3 Establishment and Development of Tax system – Disadvantages and Advantages of Taxes: The Case of Kosovo Driton BALAJ, Teuta MULAKU-BALAJ Carlo MIGLIARDO, Daniele SCHILIRÒ Impact of Domestic Institutional Investors on 5 Indian Stock Market S. A. Atif SALAR Andreea Pascucci University of Bologna, Italy Daniel Stavarek Silesian University, Czech Republic Wing-Keung Wong Department of Economics, Institute for Computational Mathematics, Hong Kong Baptist University …117 Mid-Sized Italian Manufacturing Firms: A Panel 4 Data Analysis on Profitability Piotr Misztal Jan Kochanowski University in Kielce, Faculty of Management and Administration, Poland Laura Ungureanu Spiru Haret University, Romania …107 ASERS Publishing http://www.asers.eu/asers-publishing ISSN 2068-8393 Journal's Issue DOI: https://doi.org/10.14505/jasf.v7.2(14).0 93 …129 ...146 Journal of Advanced Studies in Finance DOI: https://doi.org/10.14505/jasf.v7.2(14).04 Mid-Sized Italian Manufacturing Firms: A Panel Data Analysis on Profitability Carlo MIGLIARDO Department of Economics. University of Messina, Italy cmigliardo@unime.it Daniele SCHILIRÒ Department of Economics, University of Messina21, Italy dschiliro@unime.it Suggested Citation: Migliardo, C., Schilirò, D. (2016). Mid-sized Italian manufacturing firms: A panel data analysis on profitability. Journal of Advanced Studies in Finance, (Volume VII, Winter), 2(14): 129 - 145. DOI:10.14505/jasf.v7.2(14).04. Article’s History: Received September, 2016; Revised November, 2016; Accepted December, 2016. 2016. ASERS Publishing. All rights reserved Abstract This paper aims to provide an empirical analysis concerning the different aspects of profitability of the Italian manufacturing firms of intermediate size, namely medium and medium-large size companies, for the period 2004-2010. It analyzes various aspects of firm profitability relating corporate structures, that is, capital structure, risk component, asset composition, and growth opportunities. The study investigates firm profitability by using econometric panel-data techniques, such as the system dynamic GMM estimator that assures the robustness of our empirical analysis. One of the main results of our investigation is that we find asignificant and negative impact ofcapital structure (i.e. leverage and tangibility) and risk component on firm profitability. Keywords: Mid-sized Italian firms; profitability; panel data; (GMM) generalised methods of moments. JEL Classification: C 23, G 32, L 25, L 60. 1. Introduction The paper aims to provide an empirical analysis of profitability of the Italian manufacturing firms of intermediate size for the period 2004-2010, and uses the System GMM estimator for the estimation of dynamic panel data models. The present study contributes to extend the empirical literature on profitability of mid-sized Italian manufacturing firms, byfocusing onthe firm-specific determinants on profitability,as well as the relationship between corporate structure and firm performance. Thus, this contribution is an empirical investigation that focuses on profitability of Italian manufacturing firms of intermediate size, which are essentially medium and medium-large size joint stock companies. They are characterized by family ownership, high degree of internationalization, continuous product innovation, specialization in high quality products, andstrong ties with the Italian industrial districts; also, these firms usually operate under conditions of monopolistic competition (Colli 2005, Coltorti 2006, Coltorti et al. 2013, Schilirò 2011 and 2012). In addition, this paper investigates the behavior of Italian manufacturing firms of intermediate size and their peculiar corporate structure, by focusing on risk component, asset composition, capital structure, growth opportunities, and the relationship of these factors with the firms’ performance. The empirical results of this study provide evidence that firms appear to operate in condition of monopolistic competition, since profits tend to persist over time, which may be a consequence of product differentiation and specialization in highquality products, as the empirical literature on the mid-sized Italian manufacturing firms seems to suggest (Colli 2005, Coltorti 2006 and 2014, Coltorti et al. 2013). Also profitability, for this category of firms, is inversely related to the leverage ratio, and the relationship between firm asset 21Piazza Pugliatti, 1 98122 Messina, Italy 129 Volume VII Issue2(14) Summer 2016 tangibility and profitability is negative. Finally, profitabilityis inversely related to the risk variable. The paper is structured as follows. Section 2 discusses the topic of profitability of firms in light of the theoretical literature; section 3 introduces the framework of our analysis by describing the mid-sized Italian manufacturing firms and their peculiar features; section 4 analyzes the data and the estimation method; section 5 shows the empirical model; section 6 discusses theempirical results; section 7 provides the conclusions and proposes the objectives for future research. 2. Literature on profitability of firms The economic literature on profitability of firms is vast and involves many fields like strategic management, accounting and finance, and industrial economics. Since firm profitability is affected by numerous factors, for the purpose of our empirical analysis, we find convenient to look at different theoretical strands. Thus, to consider firm specific effects on profitability, we look at strategic management literature, and in particularto theresourcebased view. This theoretical strand claims that the bundle of resources, organizati onal structure, and management practices of the firmestablish a link between the internal characteristics of the company and its performance (Teece 1981, Barney 1991, Peteraf 1993, Levinthal 1995). Consequently, according to the resource view model, the heterogeneity in profitability across firms,is the result of the persistent differences in their specific characteristics (Rumelt 1991, Hawawini et al. 2003). Empirical literature shows that in this type of models aboveaverage profits are the result of tangible(financial and physical factors of production) and intangible (e.g. technology, reputation) resources thatreflect the distinctive capabilities of the firm, which are rare or costly to be copied or imitated. Empirical findings broadly confirm the dominance of firm-specific factors in determining the firm’s profitability with respect to industry-specific and country-specific aspects (Rumelt 1991, Mc Gahan, Porter 1997, Claver, Molina, Tarí 2002, Hawawini et al. 2003, Brito, Vasconcelos 2006, Kachlami and Yazdanfar 2016). Another strand of literature welook at is corporate finance. It is well known that corporate finance literature has traditionally focused on the study of financial decisions, mostly long-term decisions, there for corporate structure and dividends, fall among the topics it studies. We are particularly interested, in the empirical literature regarding the effect of capital structure on firm’s performance (Berger Bonaccorsi di Patti, 2006, Zeiturn and Tian 2007), but also the analysis of corporate sustainability performance (Artiach et al. 2010). The latter author support the view that leading firms have higher levels of growth and a higher return on equity than conventional firms have, but do not have lower advantage than other firms. Furthermore, we consider the literature on firm growth and the relationship between growth and profits (Goddard et al. 2005, Coad 2007). In particular, Goddard et al. (2005) have examined the financial drivers that could affect firm’s profitability in European manufacturing and services. These authors looked at the relationship between size and profitability, and adopted the return on asset (ROA) as a measure of profitability. They found thata firm’s profitability is negatively related with size and gearing ratio and positively related with market share and liquidity. Finally, we take into consideration the literature on the persistence of profit approach (Mc Gahan, Porter 1999, Goddard et al. 2005) which may reflect the influence of both industry-level and firm-level factors22. 3. Characteristics of mid-sized Italian manufacturing firms Mid-sized Italian manufacturing firmsare the object of our investigation on profitability. These firms, which are joint stock companies, have some special characteristics. They are generally distinguished for being familyowned, organized in groups, for having links with industrial districts, a tendency to innovate the product continuously, a strong international presence, and a commercial model specialized in niches of high quality products in order to defend against manufacturers in low-cost countries and large-size multinationals (Colli 2005, Coltorti 2006, Coltorti et al. 2013, Schilirò 2011, 2012). As Colliargued (2005) – and the literature on mid-sized Italian companies tendsto confirm 23 – the intermediate size of companiesis not to be intended as a size of transition between the small and large, is meant to persist over time and represents a well-structured economic corporate system. This type of companies, originated in the 1970s, derives from the gradual transformation of Italian industry following the globalization of market competition and the declining performances of the largest Italian industrial groups. 24 In addition, intermediate size firmshave been consolidating as a result of changes in Goddard, Tavakoli, Wilson (2009), analyzing the sources of variation in profitability and growth for manufacturing firms located in 11 European countries, have found that the firm-level effects are the most important class of effect in explaining the variation in performance. 23 Coltorti (2006, 2008, 2014), Coltorti et al. (2013), Marini (2008), Schilirò (2010, 2011, 2012). 24 For the transformation and disintegration of large industrial firms due to the globalization, see Feenstra (1998). 22 130 Journal of Advanced Studies in Finance productdemand due to radical changes in consumer behavior and consumption pattern. Most of these firms are involved in the production of “Made in Italy” goods and express Italy’s long-standing supremacy in food production, clothing, home furnishing, and light mechanic (Schilirò 2010, Coltorti 2014). Italian manufacturing firms of intermediate size are both medium-sized firms employing between 50 and 499 workers, and medium-large companies employing more than 499 workers25. The following features are most commonly found in their business model. These companies make a limited use of capital and manage to succeed in international markets by focusing on the quality of their workmanship, on the maximization of the value produced per employee (and this translates into innovations), on a corporate finance characterized by few debts, and on the ability to manage the corporate organization. At the same time, these firms focuson competitive advantages, with the intangibles (brand, communication, customer relationships, etc.) as the factors that have increasingly acquired importance26. Since they are able to create market niches, these firms usually operate under monopolistic competition. The literature seems to indicate that the Italian manufacturing firmsof intermediate size enjoy a good and time-persistent profitability compared to both largerand smaller companies (Colli 2005, Coltorti 2006, 2012, 2014, Schilirò, 2011). The aim of this paper is toinvestigate the profitability of the Italian manufacturing firms of intermediate size using panel data econometrics in order todetect the determinants of profitability, to evaluate their impact, to highlight the presence ofpersistent profits, and to investigate the firm specific factors. 4. Data and estimation method Our study investigates the profitability of Italian manufacturing firms of intermediate size (medium and medium-large firms) using panel data analysis. To identify this category of firms we followed the criteria previously adopted in Italy by Mediobanca. These criteria are based on two parameters: turnover and number of employees27. More specifically, the dataset considered takes into account only firms with a turnover ranging from €50 million to €3 billion and with a number of employees starting from 49. In our analysis, we refer to data published over the years in the Annual Survey of the Leading Italian Companies carried out by the Research Unit of Mediobanca28. We focus on the manufacturing firms excluding from the sample both public and service sectors companies. Thus, the sample includes 1066 Italian manufacturing firms of intermediate size for the period 20042010, all joint stock companies, although most of them are not publicly traded29. In addition, since the panel used in the analysis is unbalanced, the total annual observations considered are 426730. As confirmed by empirical literature, the choice of using panel data modelsis justified sincethis approach often tends tooutperform time series or cross-section analysis. In fact,longitudinal data increase the number of observations by pooling several times of data for each firm. This determines a more accurate inference of model parameters and a larger set of available estimators.At the same time it allows the opportunity of deepening the dynamics of firm behavior.31 Hereunder, Table1 shows the variables of the model that we adopt, their expected effect, and summary statistics. In our model, there are three alternatives to measure a firm’s profitability and efficiency: ROA, ROE, and PROFIT32. More specifically, ROA reflects average return on total gross assets and it is calculated based on earnings with respect to company assets consisting of both debt and equity. It represents the ability to generate turnover by exploiting the available resources. Therefore, ROA indicates the return offered to all the firm’s financial stakeholders. ROE is the return available to shareholders after considering tax and others claimants.Finally, PROFIT is calculated as the EBIT divided by total assets.The first tworatiosare the most commonly used indicators of profitability in empirical studies on firm profitability. However, ROE is a sensitive See Mediobanca-Unioncamere (2005-2011). Coltorti and Garofoli (2011); Schilirò (2012). 27 The conventional criteria limiting the sample have been set forth by the Research Unit of Mediobanca that provided the dataset. 28 Mediobanca, Le principali società italiane, (2005 - 2011). These surveys document financial statements of the leading individual Italian companies. 29 The period under consideration allows to cover different phases of the business cycle; a first phase (2004-2007) and a second phase (2008-2010). 30 The panel is unbalanced since it contains firms entering or leaving the market during the sample period (e.g. due to default, mergers). Unbalanced panels are very common in studies of a specific country’s firm profitability (Baltagi 2008). 31 Cameron and Trivedi (2011) show that panel data analysis achieve consistent estimations by controlling for unobserved individual heterogeneity and the associated biases. 25 26 32 Both ROA and ROE reveal how well a company uses its financing and assets to create income. 131 Volume VII Issue2(14) Summer 2016 indicator ofdebt expansion and buy-backof shares: these two actions willaffect the measure of profitability.In order to examine the robustness of our empirical findings, we prefer toadopt three alternative ratios as financial performance variables. From Table 1, we can derive few interesting clues. Firstly, the average return to equity for the sample is about 10.2%, while the average return to asset as whole is 6.16%. Consequently, ROA is smaller than ROE, which means that financing costs less than the profit it makes. In other words, the firm is making sufficient profit on borrowings to cover the cost of the interest on those funds. Mid-sized Italian manufacturing firms often overcome thisimportant financialstress test. Furthermore, for these financial ratios, there are not universal value benchmarks, but they should be assessed concerning the sector and time period taken into account. Lastly, PROFIT − the variable that represents the net profits on total assets− appears the least volatile profitability indicator, but is also the one with the lowest average yearly return (2.27%). Table 1 - Definitions, notations, and the expected effect of the explanatory variables of model on firm profitability, 2004-2010 Variable Definition Endogenous Variables ROE Return on equity Profit Net profits divided by total assets ROA Return on assets (T.A.) Explanatory Variables - Dimensional Dummy sized1 Dummy equal to 1 if T. A. is less than 100 million Dummy equal to 1 if T. A. is greater than 100 million and less sized2 than 500 mill. Dummy equal to 1 if T. A. is greater than 500 million and less sized3 than 1 bill. Dummy equal to 1 if T. A. is greater than 1 billion and less sized4 than 5 bill. Tang Fixed assets to total assets Leverage Financial Debt to total assets Risk Standard deviation over time of the firm’s return on equity Growth Rate of Growth of Sales opportunities No. obs 4267 No. of Firms 132 Expected effect mean s. d. 0.1025 0.0227 0.0616 1.1149 0.0727 0.0804 Benchmark 0.4473 0.4973 Positive 0.4001 0.4900 Positive 0.0941 0.2920 Positive 0.0585 0.2348 Undefined Negative Positive 0.2230 0.2660 0.2133 0.1538 0.1935 0.7371 Positive 0.0520 0.3316 1066 Journal of Advanced Studies in Finance Table 2 - Pearson’s correlation matrix of the variables, during 2004-2010 roe Prof. Roe 1.000 0.035** Prof. ROA sized1 sized2 sized3 1.000 ROA 0.030** 0.784*** 1.000 sized1 0.016 -0.049*** -0.026* 1.000 sized2 -0.012 0.029* 0.013 -0.746*** 1.000 sized3 -0.004 0.026* 0.023 -0.291*** -0.257*** 1.000 sized4 -0.003 0.013 0.002 -0.219*** -0.193*** -0.075*** 1.000 tang -0.027* -0.182*** -0.192*** -0.133*** 0.087*** 0.040*** 0.055*** 1.000 Leverage risk growth N° workers -0.032** -0.142*** 0.040** -0.007 -0.397*** -0.093*** 0.098*** -0.013 -0.347*** -0.080*** 0.118*** 0.004 -0.025 0.010 -0.039** -0.361*** 0.011 0.000 0.033* -0.105*** 0.006 -0.004 0.003 0.364*** 0.027* -0.019 0.011 0.547*** 0.214*** -0.080*** -0.053*** 0.089*** Note: t statistics in parentheses Note: Statistically significant at the *10%, **5% and 1% level. 133 sized4 tang Leverage Risk growth nworkers 1.000 0.042** -0.034 -0.058*** 1.000 -0.018 0.000 1.000 -0.005 1.000 Volume VII Issue2(14) Summer 2016 Table 3 - OLS, fixed effect and Random effect for the profitability ROE ROE Growth Risk Tang sized2 sized3 sized4 Leverage Constant OLS 0.0948 (0.06) -0.3388* (0.19) -0.4541*** (0.10) 0.0189 (0.05) 0.0390 (0.05) 0.0035 (0.06) -0.6728*** (0.14) 0.4387*** (0.07) Random-effect 0.0659* (0.04) -0.2523 (0.24) -0.4682*** (0.13) 0.0582 (0.09) 0.0691 (0.08) 0.0461 (0.07) -0.8725*** (0.17) 0.4742*** (0.08) Fixed-effect 0.0452* (0.03) 0.0 (0.00) -0.5816*** (0.20) 0.1205 (0.15) 0.2287 (0.15) 0.2781* (0.15) -0.9183*** (0.18) 0.3882*** (0.10) GLS 0.0904*** (0.01) -0.2234*** (0.02) -0.3618*** (0.01) 0.0357*** (0.00) 0.0243*** (0.01) 0.0048 (0.00) -0.5355*** (0.01) 0.3549*** (0.01) 3092 839 0.072 3092 839 3092 839 0.007 3092 839 L.roe L.growth No. obs. No. of firms R2 Wald-test AR(1) AR(2) Sargan-test No of instruments Note: Statistically significant at the *10%, **5% and ***1% level. Robust standard errors in parentheses System GMM 0.4162*** (0.08) -1.3617** (0.61) -0.2826 (0.46) -0.0706 (0.13) -0.0837 (0.10) 0.1875 (0.15) -0.9633*** (0.40) 0.633763*** (0.21) -0.5546*** (0.22) 0.6016*** (0.12) 2217 664 2(9)=61.15 z=-1.1723 p-value=0.24 z= -1.13 p-value=0.26 2(59)=59.38 p-value=0.46 69 Table 4 - OLS, fixed effect and Random effect for the profitability variable PROFIT Prof. Growth Risk Tang sized2 sized3 sized4 Leverage Cons OLS 0.0161* (0.01) -0.0104*** (0.00) -0.0671*** (0.02) 0.0097*** (0.00) 0.0149** (0.01) 0.0142* (0.01) -0.1370*** (0.01) 0.0694*** (0.00) Random-effect 0.0153** (0.01) -0.0106*** (0.00) -0.0831*** (0.02) 0.0130*** (0.00) 0.0238*** (0.01) 0.0237*** (0.01) -0.1609*** (0.02) 0.0756*** (0.01) Fixed-effect 0.0125** (0.01) 0.0 (0.0) -0.1668*** (0.03) 0.0224*** (0.01) 0.0531*** (0.01) 0.0631*** (0.01) -0.1908*** (0.03) 0.0924*** (0.01) L. Prof L2. Prof 134 GLS 0.017*** (0.00) -0.0075*** (0.00) -0.6291*** (0.00) 0.0072*** (0.00) 0.0098*** (0.00) 0.0119*** (0.00) -0.1254*** (0.00) 0.0663*** (0.00) System GMM 0.0563*** (0.01) -0.0349* (0.02) -0.0901*** (0.03) -0.0067* (0.01) 0.0333** (0.02) 0.028 (0.02) -0.0985** (0.04) 0.8682 (0.01) 0.3465*** (0.12) 0.0268 (0.11) Journal of Advanced Studies in Finance Prof. No. obs. No. of firms R2 Wald-test OLS 3092 839 0.197 Random-effect 3092 839 Fixed-effect 3092 839 0.105 GLS 3092 839 System GMM 2217 664 2(13)=164.97 z=-4.03 p-value=0.00 z=-0.69 p-value=0.49 2(22)=15.86 p-value=0.82 AR(1) Ar(2) Sargan-test Number of instruments Note: Statistically significant at the *10%, **5% and ***1% level. Robust standard errors in parentheses 36 Table 5 - OLS, fixed effect and Random effect for the profitability ROA ROA Growth Risk Tang sized2 sized3 sized4 Leverage Cons OLS 0.0230 (0.01) -0.0108 (0.01) -0.0838*** (0.02) 0.0079* (0.00) 0.0130* (0.01) 0.0119 (0.01) -0.1312*** (0.01) 0.1112*** (0.01) Random-effect 0.0211** (0.01) -0.0082 (0.01) -0.0981*** (0.02) 0.0089* (0.00) 0.0201*** (0.01) 0.0168** (0.01) -0.1371*** (0.01) 0.1137*** (0.01) Fixed-effect 0.0187* (0.01) 0.0 (0.0) -0.1468*** (0.03) 0.0144** (0.01) 0.0413*** (0.01) 0.0440*** (0.01) -0.1553*** (0.02) 0.1225*** (0.01) GLS 0.0247*** (0.00) -0.0089*** (0.00) -0.0803*** (0.00) 0.0081*** (0.00) 0.0131*** (0.00) 0.0116*** (0.00) -0.1214*** (0.00) 0.1066*** (0.00) 3092 839 3092 839 3092 839 3092 839 L.ROA L.ROA 2 No. obs. No. of firms Wald-test AR(1) AR(2) Sargan-test No of instruments Note: Statistically significant at the *10%, **5% and ***1% level. Robust standard errors in parentheses System GMM 0.0927*** (0.02) -0.0489 (0.04) -0.0931** (0.04) -0.0055 (0.01) 0.0269** (0.01) 0.0143 (0.02) -0.0777* (0.04) 0.0939*** (0.03) 0.4609*** (0.15) -0.1312 (0.13) 2217 664 2(12)=189.91 z=-5.54 p-value=0.00 z=-0.10 p-value=0.91 2(69)=28.14 p-value=0.17 36 5. The empirical model This section describes the empirical model for the estimation of the firm-level profitability. More specifically, in order to assess the many aspects of firm profitability, we conduct an empirical analysis by using an econometric model that relates the firm’s performance to a set of explanatory variables, which includes firm size, risk, and capital structure. More specifically, inspired partly by previous empirical literature (e.g., Hawawini et al. 2003, Goddard, Tavakoli and Wilson 2005, Berger and Bonaccorsi di Patti 2006, Coad 2007, Zeiturn and Tian 2007, Artiach et al. 2010), we specify a panel equation aiming at capturing the potentially relevant factors in determining firm profitability. The specification of the static model is the following: 135 Volume VII Issue2(14) Summer 2016 PRi ,t  0  1GRi ,t   2 Riski  3Tangi ,t  V j Sized j  4 Leveragei ,t  ui ,t , ( 4 j 2 The dependent variable PR stands for profitability of the firm i. As it has been pointed out in the previous section, we use three alternative proxies for this variable of interest. The three alternatives are ROA, ROE, and PROFIT, all continuous variables. The determinants of firm profitability, instead, are represented by GR, RISK, TANG, SIZE, and LEVERAGE. GR is the rate of growth of sales and measures the growth opportunities of the firm. Several empirical studies (e.g. Cowling 2004, Coad 2007) have shown a positive and significant relationship between firm growth and profitability. However, this result is not conclusive. In fact, Davidsson et al. (2009), who follow the resource view reasoning, cast doubt on that positive effect and tend to confirm that profitable low growth firms are more likely to reach the desirable state of high growth andhigh profitability as well as to have a decreased risk of ending up performing poorly on performance dimensions. Hence, we cannot establish ana priori effect for the variable GR. RISK is a control variable that measures the standard deviation of ROE over the six-year period for each firm. Since when we adopt the ROE ratio as dependent variable the risk component is a function of the return on equity, RISK is assumed to be a predetermined variable in the System GMM set-up. We introduce this RISK variable in accordance to corporate finance literature. Although riskier firms (i.e. firms with higher performance volatility) are expected to generate greater expected return, several empirical studies (Berger and Bonaccorsi di Patti 2006, Zeitun and Tian 2007, Lee and Li 2012) have found a negative effect of risk on profitability. These authors justified their finding as the result of higher operating risk that implies a higher probability of financial distress along with higher bankrupt costs, and thereby lowers firm’s performance. TANG captures the composition of the asset structure and is calculated as the ratio of physical capital divided by the total assets. The relationship of this explicative variable with firm performance may be positive, but the effect could revert to negative if the fixed asset is relatively high. This latter finding may be ascribed to the decreasing marginal efficiency of the capital. Prior studies (Zeiturn and Tian 2007, Artiach et al. 2010) have found an overall negative correlation between the tangibility asset and firm profitability. Since firms often invest part of their profits to boost fixed asset, one should also bear in mind a possible reverse causation from higher profitability to more tangibility. For this reason, we model firm tangibility as an endogenous variable in the System GMM set-up. SIZED represents a set of dummy variables included in the model to control for differences associated with firm’s size. As in Berger and Bonaccorsidi Patti (2006), we include size class dummy variables; in our case, the dummies are ranging from below €100 millions in gross total assets (Sized 1) to over €5 billion in gross total assets (Sized4). The control group is the category of the smallestfirms (Sized 1, under €100 millions in gross total assets), while the other estimated dummies (Sized 2, Sized 3, Sized 4) have to be compared to the benchmark category (Sized 1). Firm’s size influences the profitability since it can be a proxy of the firm’s efficiency and therefore the omitted (benchmark) variable may represent the most efficient firm. Even if some previous studies (e.g. Lee and Li, 2012) have found a non-monotonic effect of the firm’s dimension on profit rate or, alternatively, a negative correlation, other studies (Hall and Weiss 1967, Gleason and Mathur 2000, Claver et al. 2002, Tian and Zeiturn 2007, Artiach et al. 2010) in contrast, showed a positive and significant coefficient on firms’ profits33. Hence, the expected effect of the relationship between firms’ size and profits is uncertain. LEVERAGE is a determinant that captures the influence of corporate capital structureon the firm’s performance and is measured as the ratio of financial debt to the gross total assets. Generally, corporate governance models predict that leverage influences agency costs and, consequently, affects positively firm’s profitability (Harris and Raviv 1991, Myers 2001). However, a relatively high leverage indicates an anomalous firm’s structure and that the financial expenses became too high. Moreover, a further expansion of financial debt may produce significant agency costs of external debt that determine risk shifting, or a reduced effort to control risk that may, in turn, result in higher 33 Our dataset does not contain the very small firms, in fact the sample starts with firms more than €50 million in total assets and /or more than 49 workers. It also does not contain the large companies. 136 Journal of Advanced Studies in Finance expected costs of financial distress, default, or liquidation34. These agency costs translate intohigher interest payments for firms to reward debt holders for their expected losses. Furthermore, if we relax the Modigliani-Miller (1958) capital structure irrelevance principle, this willimplythe presence of an external risk premium, i.e. leverage would bemore expensive thanequity. Finally, several empirical analyses (e.g. Tian and Zeitun 2007, Jang and Park 2011) have found negative and significant effects of leverage on corporate performance. Thus, we expect both lever age and corporate return to be negatively related. As for tangibility (TANG), also LEVERAGE is treated as potential endogenous variable in the System GMM estimator, as lower profits induce higher financial debt and vice versa. The model also contains the constant term β0 and the disturbance component ui,t, the latter term consists of two components, the unobserved firm-specific effect vi, and the idiosyncratic errorεi,t. According to the persistence of profits literature (McGahan and Porter 1999, Goddard et al. 2005, Mcmillan and Wohar 2011), firm’s profits show a tendency to persist over time, because of markets imperfections, asymmetric information, and market power. Consequently, we also adopt a dynamic equation that includes the dependent variable with two lags35 among the explicative variables. Thus, the dynamic specification of our model is: PRi ,t  0  0 PRi ,t 1  1PRi ,t 2  1GRi,t  2 Riski  3Tangi,t  V j Sized j 4 Leveragei,t  ui,t , (2) 4 j 2 where: coefficient sλi represent the speed of adjustment to steady state equilibrium. In other words, if λi are close to 1, profits are highly persistent, denoting evidence in favor of scarcely competitive goods markets. Conversely, values of λi close to 0 implyahighly competitive environment. Finally, both Eq (1) and (2) do not include temporal dummies to control for time effects. The inclusion is not plausible because, across all the specifications, their coefficients are jointly statistically insignificant accordingly to the F-Test. 5.1 Univariate analysis of the model In this section,we examine the Pearson correlation coefficients and carry out a univariate analysis of the model adopted. Table 2 - Pearson’s correlation matrix of the variables, during 2004-2010 roe Prof. ROA sized1 sized2 sized3 sized4 tang Levera ge Risk growt h Roe 1.000 Prof. 0.035** 1.000 ROA 0.030** 0.784*** 1.000 sized1 0.016 -0.049*** -0.026* 1.000 sized2 -0.012 0.029* 0.013 -0.746*** 1.000 sized3 -0.004 0.026* 0.023 -0.291*** -0.257*** 1.000 sized4 -0.003 0.013 0.002 -0.219*** -0.193*** -0.075*** 1.000 Tang -0.027* -0.182*** -0.192*** -0.133*** 0.087*** 0.040*** 0.055*** 1.000 Leverage Risk growth N° workers -0.032** -0.142*** 0.040** -0.397*** -0.093*** 0.098*** -0.347*** -0.080*** 0.118*** -0.025 0.010 -0.039** 0.011 0.000 0.033* 0.006 -0.004 0.003 0.027* -0.019 0.011 0.214* -0.080*** -0.053*** 1.000 0.042** -0.034 1.000 -0.018 1.000 0.004 -0.361*** -0.105*** 0.364*** 0.547*** 0.089*** -0.058*** 0.000 -0.005 -0.007 -0.013 nwork ers 1.000 The agency costs of debt are usually explained in terms of asset substitution or of risk-shifting issue. The latent conflict among debt claimants and equity is such that shareholders expropriate wealth from bondholders by investing in new projects that are riskier than those currently held in the company’s portfolio. In this case, shareholders acquire most of the gains (i.e., when high-risk projects payoff), while bondholders bear most of the cost (Fama and Miller 1972, Jensen and Meckling 1976). 35 In the specific case of ROE equation, the dependent variable PR is included as explanatory variable only with one lag; while the GR variable (i.e. rate of growth of sales) is included with two lags. 34 137 Volume VII Issue2(14) Summer 2016 Note: t statistics in parentheses Note: Statistically significant at the *10%, **5% and ***1% level. Table 2 reports the cross-correlation coefficient among the variables. In this regard, make several observation scan. First, firm’s leverage appears to be the most important factor associated with thefirm’s lowperformance, in so supporting the hypothesis that capital structure (agency costs) and financial debts influence negatively the profitability, as widely shown in the empirical literature. Second, corporate performance is also negatively and significantly related to tangibility, so it is consistent with another assumption regarding the decreasing marginal efficiency ofphysical capital. Third, a weaker support was found aboutthe effect of firm’s dimension on profitability; in fact, the correlation between size and performance is nearly always statistically insignificant and,in addition, the coefficients (Sized 2, Sized 3, Sized 4) are close to zero. This result is also confirmed when the number of employees has been adopted as alternative firm’s dimension proxy. However, as it has been argued in section 5, the dimensional heterogeneity of the sample is not so relevant here. Fourth, preliminary results of the other control variables, namely risk and rate of growth of sales, show opposite effects. The first variable (risk) displays a negative and significant correlation with profitability, while the second (the growth opportunity variable), positively impacts on corporate returns, as expected. From the point of view of robustness, even ifsome of the control variables are mutually correlated, showing few evidences of collinearity,the magnitude of the coefficients is small.36 5.2 Econometric specifications In our empirical analysis, we address the following issues regardingthe identification of the model. First, weestimatethe static specification (namely, equation (1) section 5) byusing traditional econometric methods such as Ordinary Least Square, Random or Fixed effect model, and the Generalized Least Square. Therefore, at this stage, by applying the Chow test, we examine the presence of unobserved heterogeneity that makes pooled regression results heavy biased. In our case, the Chow test always rejects the null hypothesis. Second, we try to assess, through the Hausman test, if the individual effects are fixed or random.The results of the Hausman test suggest adopting the fixed effects for the ROA and PROFIT, while for the ROE the proper specification is the Random effects specification. However, within our study,the models described abovemight produce biasedand inconsistent results, particularly for the dynamic specification in equation (2) section 5, because of the potential issue of endogeneity, i.e. because the disturbance term of the specification is correlated with the explanatory variables, and consequently, will produce biased coefficients and standard errors. To overcome this issue, Arellano and Bond (1991), Arellano and Bover (1995) and Blundell and Bond(1998) developed the Generalized Method of Moments (GMM hence forth) for panel data analysis, this estimator deals with the above mentioned biasedness and inconsistency of the standard model applied to the static specification. In particular, Blundell and Bond(1998) deal with the issues caused by endogeneity by recurring to lagged and differenced values of the explanatory variables as internal instruments37. Moreover, system GMM estimation in dynamic panel models is robustto control for reverse causality, simultaneity bias, and possible omitted variables; while it controls the individual and time specific effects. Hence, we address these econometric issues by using a two-step system GMM technique (Arellano and Bover 1995, Blundell and Bond 1998) estimating a level-equation as well as a difference equation38. Furthermore, the use of a dynamic equation − such as (2) in section 5 − is justified because we expect that firm’sprofits denote a tendency to persist over time. In order to evaluate the validity of our System GMM estimations, we ran two common tests that confirm the null hypotheses. The first test is the Arellano–Bond, which confirms the absence of second-order autocorrelation in the transformed idiosyncratic errors. The second one is the Sargan test, which strongly confirms the soundness of the imposed over-identifying moment conditions and consequently the validity of the instruments used. In addition, our econometric model takes into account Roodman’s advice (2009) concerning the excessive As Gujarati and Porter (2009) suggest, multicollinearity becomes a serious issue only when the correlation among the control variables exceeds the threshold of 0.8. Moreover, the Variance Inflation Factors (VIF) test showed no evidence for multicollinearity among the model variables (mean VIF values ranged between 1.07 and 1.66). 37 In the system GMM estimator, the endogenous explanatory variables are instrumented with their lags so that the instruments are uncorrelated to the disturbance. 38 We are aware that the system GMM estimator has some critical aspects, as Roodman (2009) has shown regarding the instrument proliferation. For this reason we used also alternative methods such as OLS, GLS, Random and Fixed effects estimators that confirm the robustness of the results. 36 138 Journal of Advanced Studies in Finance “proliferation” in the number of instruments that may cause over-fitting of the endogenous variables and could bias the specification tests of instruments’ joint validity.Therefore, we have evaluated the robustness of our GMM results by forcefully cutting the numbers of instruments byreducing lag length39. Moreover, all GMM regressions adopt the Windmeijer (2005) correction procedure for the estimation of standard errors40. Finally, to achieve the models dynamic completeness required for the System GMM estimator, we include two lagged dependent variables for PROFIT and ROA, whereas for ROE the lagged variable is one. We also checked the robustness of the estimates using a balanced sub-sample of our panel dataset and found similar results with the full sample. Therefore, we have undertaken a number of additional sensitivity analyses, in order to explore the robustness of the results across different model specifications. In sum, after controlling for the potential endogeneity problem, with the other specifications, our main findings of the GMM remain robust and consistent. 6. Estimation results In this section, we comment our main results showed in Tables 3, 4, and 5 and also discuss the robustness checks for the hypothesis tests. Table 3 - OLS, fixed effect and Random effect for the profitability ROE ROE Growth Risk Tang sized2 sized3 sized4 Leverage Constant OLS 0.0948 (0.06) -0.3388* (0.19) -0.4541*** (0.10) 0.0189 (0.05) 0.0390 (0.05) 0.0035 (0.06) -0.6728*** (0.14) 0.4387*** (0.07) Random-effect 0.0659* (0.04) -0.2523 (0.24) -0.4682*** (0.13) 0.0582 (0.09) 0.0691 (0.08) 0.0461 (0.07) -0.8725*** (0.17) 0.4742*** (0.08) Fixed-effect 0.0452* (0.03) 0.0 (0.00) -0.5816*** (0.20) 0.1205 (0.15) 0.2287 (0.15) 0.2781* (0.15) -0.9183*** (0.18) 0.3882*** (0.10) GLS 0.0904*** (0.01) -0.2234*** (0.02) -0.3618*** (0.01) 0.0357*** (0.00) 0.0243*** (0.01) 0.0048 (0.00) -0.5355*** (0.01) 0.3549*** (0.01) 3092 839 0.072 3092 839 3092 839 0.007 3092 839 L.roe L.growth No. obs. No. of firms R2 Wald-test AR(1) AR(2) System GMM 0.4162*** (0.08) -1.3617** (0.61) -0.2826 (0.46) -0.0706 (0.13) -0.0837 (0.10) 0.1875 (0.15) -0.9633*** (0.40) 0.633763*** (0.21) -0.5546*** (0.22) 0.6016*** (0.12) 2217 664 2(9)=61.15 z=-1.1723 p-value=0.24 z= -1.13 p-value=0.26 As recommended by Roodman (2009), the number of instruments used in a dynamic GMM estimator should relatively low and smaller than the number of the number of observations. In our analysis we use 36 instruments for both ROA and PROFIT, while 69 instruments have been used for the ROE. Therefore, in both cases the number of instruments is small and lesser than our 2217 observations. The “optimal” number of instruments has been achieved by using the restriction of one lag for levels and two for differenced equations. In addition, we have done alternative estimations by reducing further the number of instruments. Nonetheless, these further reductions worsen the diagnostic tests (specifically, they resulted in a lower Sargan p-value), indicating that our selected number of instruments should be fairly “optimal”. 40 Windmeijer (2005) proposed a correction method for the commonly downward biased estimated standard errors produced by the two-step GMM technique. In particular, he corrects the finite sample biases by the estimated asymptotic variance of the two-step GMM estimator that produces the corrected adjusted Wald Statistics. 39 139 Volume VII Issue2(14) Summer 2016 ROE OLS Random-effect Fixed-effect GLS Sargan-test No instruments Note: Statistically significant at the *10%, **5% and ***1% level. Robust standard errors in parentheses System GMM 2(59)=59.38 p-value=0.46 69 Table 4 - OLS, fixed effect and Random effect for the profitability variable PROFIT Prof. Growth Risk Tang sized2 sized3 sized4 Leverage Cons OLS 0.0161* (0.01) -0.0104*** (0.00) -0.0671*** (0.02) 0.0097*** (0.00) 0.0149** (0.01) 0.0142* (0.01) -0.1370*** (0.01) 0.0694*** (0.00) Random-effect 0.0153** (0.01) -0.0106*** (0.00) -0.0831*** (0.02) 0.0130*** (0.00) 0.0238*** (0.01) 0.0237*** (0.01) -0.1609*** (0.02) 0.0756*** (0.01) Fixed-effect 0.0125** (0.01) 0.0 (0.0) -0.1668*** (0.03) 0.0224*** (0.01) 0.0531*** (0.01) 0.0631*** (0.01) -0.1908*** (0.03) 0.0924*** (0.01) GLS 0.017*** (0.00) -0.0075*** (0.00) -0.6291*** (0.00) 0.0072*** (0.00) 0.0098*** (0.00) 0.0119*** (0.00) -0.1254*** (0.00) 0.0663*** (0.00) 3092 839 0.197 3092 839 3092 839 0.105 3092 839 L. Prof L2. Prof No. obs. No. of firms R2 Wald-test AR(1) Ar(2) Sargan-test No instruments Note: Statistically significant at the *10%, **5% and ***1% level. Robust standard errors in parentheses System GMM 0.0563*** (0.01) -0.0349* (0.02) -0.0901*** (0.03) -0.0067* (0.01) 0.0333** (0.02) 0.028 (0.02) -0.0985** (0.04) 0.8682 (0.01) 0.3465*** (0.12) 0.0268 (0.11) 2217 664 2(13)=164.97 z=-4.03 p-value=0.00 z=-0.69 p-value=0.49 2(22)=15.86 p-value=0.82 36 Table 5 - OLS, fixed effect and Random effect for the profitability ROA ROA growth Risk Tang sized2 sized3 sized4 Leverage OLS 0.0230 (0.01) -0.0108 (0.01) -0.0838*** (0.02) 0.0079* (0.00) 0.0130* (0.01) 0.0119 (0.01) -0.1312*** (0.01) Random-effect 0.0211** (0.01) -0.0082 (0.01) -0.0981*** (0.02) 0.0089* (0.00) 0.0201*** (0.01) 0.0168** (0.01) -0.1371*** (0.01) 140 Fixed-effect 0.0187* (0.01) 0.0 (0.0) -0.1468*** (0.03) 0.0144** (0.01) 0.0413*** (0.01) 0.0440*** (0.01) -0.1553*** (0.02) GLS 0.0247*** (0.00) -0.0089*** (0.00) -0.0803*** (0.00) 0.0081*** (0.00) 0.0131*** (0.00) 0.0116*** (0.00) -0.1214*** (0.00) System GMM 0.0927*** (0.02) -0.0489 (0.04) -0.0931** (0.04) -0.0055 (0.01) 0.0269** (0.01) 0.0143 (0.02) -0.0777* (0.04) Journal of Advanced Studies in Finance ROA Cons OLS 0.1112*** (0.01) Random-effect 0.1137*** (0.01) Fixed-effect 0.1225*** (0.01) GLS 0.1066*** (0.00) 3092 839 3092 839 3092 839 3092 839 L.ROA L.ROA 2 No. obs. No. of firms Wald-test AR(1) AR(2) Sargan-test No of instruments Note: Statistically significant at the *10%, **5% and ***1% level. Robust standard errors in parentheses System GMM 0.0939*** (0.03) 0.4609*** (0.15) -0.1312 (0.13) 2217 664 2(12)=189.91 z=-5.54 p-value=0.00 z=-0.10 p-value=0.91 2(69)=28.14 p-value=0.17 36 Our inference analysis has been conducted with several specifications. Overall, the results are all robust, but –as it has been argued in section 5.2 – system GMM estimation is the optimal method that produces the more efficient and consistent coefficients. In support of this hypothesis, the coefficients of one period-lagged ROE, PROFIT, and ROA are found to be statistically significant in all cases,while they are the highest in terms of magnitude among all the explanatory variables.This seems to imply that the Italian economy is far from a perfectly competitive market structure, and that mid-sized manufacturing firms tend to segment the market by creating market niches, hence they operate in monopolistic competition.41 The profitability of mid-sized Italian manufacturing firms tends to be highly persistent over time and, therefore, the lagged dependent variable should be included in the regression models. However, the application of OLS or Fixed Effect estimators of a dynamic specification would result in biased estimated coefficients, because of possible endogeneity of the regressors. Consequently, these estimators are likely to perform poorly. Thus, the results of the static models have been reported simply for the purpose of comparison, while the last column of each table reports the coefficients of reference. On the other hand, all the estimated models denote the presence of some robust regularity among the several specifications and alternative profitability ratios used. As expected, firm profitability is shown to be larger when sales growth rateis positive. This suggests that higher sales growth generates income that partially influences firm’s profits. This positive impact is consistent with earlier studies that use the same proxy variable (growth rate of sales) to measure growth opportunity ratio42, but also with other empirical studies that adopt different variable ssuch as, for instance, the rate of change in total assets 43 . Moreover, in the case of ROE, past growth is observed to have a greater positive impact on the subsequent profit rate than contemporaneous growth; thus, growth seems to generate dynamic increasing returns. This evidence is in line with previous studies (e.g. Coad 2007). Furthermore, the estimated coefficients to capture the risk contribute to firm’s performance are negative and almost always statistically significant, i.e. safer companies show higher profits than riskier firms. This is consistent with prior studies (Bonaccorsi and Berger 2006, Tian and Zeiturn 2007, Lee and Li 2012). Moreover, the negative correlation is more than proportional in the case of the return on equity ratio, as a 1% increase in profit volatility determines a reduction in profits of 1.36% for ROE. In sum, volatility of earnings reduces the value of firms. This finding can be ascribed to the higher risk of default that induces a greater probability of financial distress and larger bankruptcy costs, and, consequently, downgrades firm profits. The effect of tangibility on profitability at the sample mean is statistically significant (for ROA and PROFIT) and negative (in all other cases). Firms with relatively large levels of tangible assets are less profitable, and this Migliardo (2012) found evidence of high degree of market power in the Italian firms over the same sample period taken into consideration in the present study. 42 Cowling (2004) and Coad (2007) have shown a significant and positive correlation between sales growth and profits. 43 E.g. Zeitun and Tian (2007), Nunes et al. (2009). 41 141 Volume VII Issue2(14) Summer 2016 suggests that the Italian manufacturing firms of intermediate size tend to use their fixed assets inefficiently. More specifically, they invest excessively in physical capital over the total assets, and since the marginal efficiency of capital is decreasing, this worsens corporate performance. In general, the estimated equations show that the firm’s size does not affect the pattern of firm’s performance44; the latter being almost stable along all size classes. This result might be a scribed to the sample structure of our analysis and justifies why the relationship SIZE-PROFIT is weak. On average and as expected, the leverage variable is negative and significant for all the measures of profitability used. An expansion in the debt ratio of 1% determines a reduction of corporate returns from 0.08% for ROA up to 0.96% for ROE respectively. This indicates that capital structure is too unbalanced toward the financial debt. There are several possible theoretical explanations for our findings. In more detail, such inverse relationship between firm value (ROE) and leverage is justified in literature by the pecking order theory (Myers and Majluf 1984)45. According to it, the firms finance their investments at first with internal resources; i.e. profitable firms use their earnings primary; once the endogenous funds are used up, the companies turn to debt financing. Finally, they opt to capital share increase as a last source of funding. Several reasons explain this hierarchy of financing sources: first, the asymmetric information in the financial markets increases the cost of issuing equity46; second, old shareholders tend to limit the emission in order to retain control of the company; third, the internal financing strategy allows transaction cost saving. In summary, a profitable company uses less lever age and in so doing determines a higher firm’s value that will be positively correlated with corporate performance and negatively linked with the debt. Therefore, the results of this study are consistent with the pecking order theory. Moreover, the significant and negative correlation of the other proxy (ROA and PROFIT) dependent variables with leverage can be explained by agency conflicts causing overleveraged firms and adversely affecting their profitability negatively. In the context of our analysis, the negative relation cans beascribed to several idiosyncratic reasons. On the one hand, mid-sized Italian manufacturing firms seem to choose debt (bank debt) instead of equity, owing to either legal market restrictions (e.g. company profile) and/or credit conditions, which do not allow the recourse to financial markets. On the other hand, tax purposes can address this strategic choice, i.e. firms opt to debt rather than internal capital, because companies benefit from debt tax shields (Modigliani and Miller 1963). Finally, for mid-sized Italian manufacturing firms, the previous year’s corporate performance has significant positive impact on ROA and PROFIT, while, by contrast, a large ROE implies a decrease of profitability in the following year 47 . This remarkable result can be ascribed to a weak forward-looking strategy of profit management, as it is highlighted by a strategy of over paid dividends that jeopardizes the performance in the long run. Conclusions This study provides a further contribution to the extensive empirical literature on firm profitability. It focuses on mid-sized Italian manufacturing firms, namely medium and medium-large enterprises, which represent the most dynamic and profitable companies characterizing Italian economy. Firm-specific determinants on profitability and the relationship between corporate structure and firm performance are highlighted. Several concluding remarks can be drawn from our results. First, since firm returns denote the tendency to persist over time, showing a good resilience, we find support to the persistence of profit hypothesis48. This may prove that mid-sized Italian manufacturing firms operate in a context of monopolistic competition, hence they are able to create market niches by specializing in high quality products, differentiating the products, but also being able to establish a customized relationship with their clients. Second, as referred to both leverage and tangibility, mid-sized Italian manufacturing firms’ capital structure looks unbalanced. Our findings prove that financial debt Alternatively, we also used either the raw size variable referred to total assets in logarithm term, and the natural logarithm of the number of employees, but in both cases size proved not to be an important factor for firm profitability. 45 The pecking order theory of capital structure affirms that, all other things being equal, companies seeking to finance a new project or product have a hierarchy of preferred financing options progressing from the most to the least preferred. 46 The new equity issuing leads to a firm’s stock price decline, because investors perceive that managers consider the company overvalued. Thus, the investors are monetizing this overvaluation. As a consequence, the firm’s value decreases and whilst the cost of external financing increases. 47In general, we would expect a positive sign for the lagged dependent variable. Nonetheless the negative effects for the specific case of ROE ratio (Table 3) is due to the high dividend payments. 48 Although we know that even the profitability of mid-sized Italian firms has deteriorated in the period 2007-2010 due to the global crisis (Cerved, 2014). 44 142 Journal of Advanced Studies in Finance (i.e. leverage) is relatively high with negative impact on firm profits. Physical assets (tang)t end to be outsized, this affects negatively firm performance, because tangible capitalis subject to decreasing return to scale. Moreover, our empirical evidence suggests that present growth of sales (and in the case of return on equity also growth occurred in the past) as well as profit volatility (i.e. risk) are important and significant determinants of firm profitability. In sum, our empirical analysis supplies additional insights for managers as well as for planners of economic policy. The results provide several managerial implications for Italian manufacturing companies of intermediate size. Our evidence indicates that profitable manufacturing firms could achieve better returns, if they adopted a diversification in their financial funding strategy and/or if they modified their allocation of business assets, e.g. by intensifying their intangible assets. From the point of view of the economic policy planner, our results could suggest the adoption of fiscal incentives in order to induce adjustments in the capital structure of this type of firms and, consequently, to enhance their profitability. Potential extensions of the present study might be an investigation on medium size European manufacturing companies and/or an evaluation of our results within a macroeconomic framework such as DSGE models. References: [1] Arellano, M., Bond, S. (1991). Some tests specification for panel data: Monte Carlo evidence and an application to employment equations. Review of Economic Studies, 58: 277-297. doi: 10.2307/2297968 [2] Arellano, M., Bover, O. (1995). Another look at the instrumental variable estimation of error-components models. Journal of Econometrics, 68: 29-52. Available at: http://dx.doi.org/10.1016/0304-4076(94)01642-D [3] Artiach, T., Lee, D., Nelson, D., Walker, J. (2010).The determinants of corporate sustainability performance. Accounting and Finance, 50: 31-51. DOI: 10.1111/j.1467-629X.2009.00315.x [4] Baltagi, B. (2008). Econometric Analysis of Panel Data. 4th Edition. New York, John Wiley & Sons. [5] Barney, J. (1991). Firm resources and sustained competitive advantages. Journal of Management, 17(1): 99-120. doi: 10.1177/014920639101700108 [6] Berger, A.N., Bonaccorsi di Patti, E. (2006). Capital structure and firm performance: a new approach to testing agency theory and an application to the banking industry. Journal of Banking & Finance, 30: 10651102. [7] Blundell, R., Bond, S. (1998). Initial conditions and moment restrictions in dynamic panel data models. Journal of Econometrics, 87: 115-143. Available at: http://dx.doi.org/10.1016/S0304-4076(98)00009-8 [8] Bond, S. (2002). Dynamic panel data models: a guide to micro data methods and practice, Working Paper 09/02. Institute for Fiscal Studies, London. [9] Brito, L. and Vasconcelos, F. 2006. How Much Does Country Matter? In A. Cooper, S. Alvarez, A. Carrera, L. Mesquita, and R. Vassolo (eds.), Entrepreneurial Strategies. Oxford, UK: Blackwell, pp. 95-113. [10] Cameron, C., Trivedi, P.K. (2011). Microeconometric, Methods and Application. Cambridge, Cambridge University Press (Virtual Publishing). [11] Claver, E., Molina, J., Tarí, J. (2002). Firm and industry effects on firm profitability: a Spanish empirical analysis. European Management Journal, 20(3): 321-328. [12] Coad, A. (2007). Testing the principle of growth of the fitter: the relationship between profits and firms growth. Structural Change and Economic Dynamics, 18: 370-386. Available at: http://dx.doi.org/10.1016/j.strueco.2007.05.001 [13] Colli, A. (2005). Il quarto capitalismo. L’Industria, 2: 219-236. [14] Coltorti, F. (2006). Il capitalismo di mezzo negli anni della crescita zero. Economia Italiana, 3: 665-687. [15] Coltorti, F. (2012). Italian industry, decline or transformation?A framework. European Planning Studies, 141. doi: 10.1080/09654313.2012.722972 [16] Coltorti, F. (2014). Il Quarto capitalismo: vero motore dello sviluppo italiano, MBRES, Milano Mediobanca. https://www.mbres.it 143 Volume VII Issue2(14) Summer 2016 [17] Coltorti, F., Garofoli, G. (2011). Le medie imprese in Europa. Economia Italiana, 1:187-223. [18] Coltorti, F., Resciniti, R., Tunisini, A., Varaldo, R. (eds.) (2013). Mid-Sized Manufacturing Companies: The New Driver of Italian Competitiveness, Berlin, Springer Verlag. [19] Cowling, M. (2004).The growth-profit nexus. Small Business Economics, 22: 1-9. [20] Davidsson, P. Steffens, P., Fitzsimmons, J. (2009). Growing profitable or growing from profits: putting the horse in front of the cart? Journal of Business Venturing, 24: 388-406. [21] Deloof, M. (2003). Does Working Capital Management Affect Profitability of Belgian Firms. Journal of Business Finance and Accounting, 30: 573-588. [22] Fama, E., Miller, M. (1972).The Theory of finance.New York, Holt, Rinehart and Winston. [23] Feenstra, R. C. (1998). Integration of trade and disintegration of production in the global economy. Journal of Economic Perspectives, 12 (4): 31-50. [24] Gleason, K. C., Mathur, L.K. and I. Mathur (2000). The Interrelationship between Culture, Capital Structure and Performance: Evidence from European Retailers. Journal of Business Research, 50: 185-191. [25] Goddard, J., Tavakoli, M., Wilson, J. (2005). Determinants of profitability in European manufacturing and services: evidence from a dynamic panel model. Applied Financial Economics, 15(18): 1269-1282. [26] Goddard, J., Tavakoli, M., Wilson, J., (2009). Sources of variation in firm profitability and growth, Journal of Business Research, 62(4): 495-508. [27] Gujarati, D., Porter, D.C. (2009). Basic Econometrics. 5th Edition. New York, McGraw-Hill. [28] Hall, M., Weiss, L. (1967). Firm Size and Profitability. The Review of Economics and Statistics, 49(3): 319331. [29] Harris, M., Raviv, A. (1990). Capital structure and the informational role of debt. Journal of Finance, 45: 321349. [30] Hawawini, G., Subramanian, V., Verdin, P. (2003). Is performance driven by industry or firm-specific factors? A new look at the evidence. Strategic Management Journal, 24:1-16. [31] Jang, S. S., Park, K. (2011). Inter-relationship between firm growth and profitability. International Journal of Hospitality Management, 30(4):1027-1035. [32] Jensen, M., Meckling, W.H. (1976).Theory of the Firm: Managerial Behavior, Agency Costs and Ownership Structure. Journal of Financial Economics, 3(4): 305-360. [33] Lee, B. S., Li, M.-Y. L. (2012). Diversification and risk-adjusted performance: a quantile regression approach. Journal of Banking and Finance, 36: 2157-2173. [34] Levinthal, D. (1995). Strategic management and the exploration of diversity, in C.A. Montgomery (Ed.), Resource-Based and Evolutionary Theories of the Firm, Norwell (MA), Kluwer. [35] Kachlami, H. Yazdanfar, D. (2016). Determinants of SME growth: The influence of financing pattern. An empirical study based on Swedish data. Management Research Review, 39(9): 966 – 986. Available: http://dx.doi.org/10.1108/MRR-04-2015-0093 [36] Marini, D. (ed.) (2008). Fuori dalla Media: Percorsi di Sviluppo delle Imprese di Successo. Venezia, Marsilio. [37] McGahan, A. Porter, M. (1997). How much does industry matter really? Strategic Management Journal, 18: 15-30. [38] McGahan, A. Porter, M. (1999). The persistence of shocks to profitability, Review of Economics and Statistics, 81: 143-153. doi:10.1162/003465399767923890 [39] Migliardo, C. (2012). Heterogeneity in price setting behavior, spatial disparities and sectoral diversity: Evidence from a panel of Italian firms. Economic Modelling, 29 (4): 1106-1118. [40] Modigliani, F., Miller, M. (1958).The Cost of Capital, Corporate Finance and Theory of Investment. American Economic Review, 48 (3): 261-297. 144 Journal of Advanced Studies in Finance [41] Modigliani, F., Miller, M. (1963).Corporate income taxes and the cost of capital: a correction. American Economic Review, 53 (3): 433–443. [42] Myers, S. C. (2001). Capital structure. Journal of Economic Perspectives, 15(2): 81-102. [43] Myers, S. C., Majluf, N. S. (1984). Corporate financing and investment decisions when firms have information that investors do not have. Journal of Financial Economics,13 (2): 187–221. [44] Nucci, F., Pozzolo, A., Schivardi, F. (2005). Is firm’s productivity related to its financial structure? Evidence from microeconomic data. Working Paper, Research Department,Roma, Bancad’Italia. [45] Nunes, P. J. M., Serrasqueiro, Z.M., Sequiera, T. N. (2009). Profitability in Portuguese service industries: a panel data approach.The Service Industries Journal 29: 693:707. [46] Peteraf, M. (1993). The cornerstones of competitive advantage: a resource based view, Strategic Management Journal, 14: 179-191. [47] Roodman, D. (2009). A note on the theme of too many instruments. Oxford Bulletin of Economics and Statistics, 71(1): 135-158. [48] Rumelt, R. (1991). How much does industry matter? Strategic Management Journal, 12(3): 167-185. DOI: 10.1002/smj.4250120302 [49] Schilirò, D. (2010). Distretti e quarto capitalismo. Milano, Franco Angeli. [50] Schilirò, D. (2011). Innovation and performance of Italian multinational enterprises of the “fourth capitalism”. Journal of Advances Research in Management, Volume II, 2: 89-103. [51] Schilirò, D. (2012). Italian medium-sized enterprises and the fourth capitalism. Journal of Applied Economic Sciences, Volume VII, 4: 436-446. [52] Teece, D. (1981). Internal organization and economic performance: an empirical analysis of the profitability of principal firms. Journal of Industrial Economics, 30:173-199. [53] Windmeijer, F. (2005).A finite sample correction for the variance of linear efficient two-step GMM estimators. Journal of Econometrics, 126: 25-51. Available at: http://dx.doi.org/10.1016/j.jeconom.2004.02.005 [54] Zeitun, R., Tian, G.G. (2007). Capital structure and corporate performance: evidence from Jordan. Australasian Accounting Business and Financial Journal, 1(4): 40-61. *** Cerved, (2014). 2014 Cerved SMEs Report, Milano, Cerved, New Copy Service, October. *** Mediobanca, (2005-2011). Le Principiali Società Italiane (various years), Milano. 145 ASERS Journal of Advanced Studies in Finance Web: www.asers.eu URL: http://www.asers.eu/publishing E-mail: asers@asers.eu asers2010@yahoo.co.uk ISSN 2068 – 8393 Journal DOI: https://doi.org/10.14505/jasf Journal’s Issue DOI: https://doi.org/10.14505/jasf.v7.2(14).00