Impact of Performance Indicators and Labour Endowment on Traffic: Empirical Evidence from Indian Ports International Journal of Maritime Economics, The Holland By BUDDHAD EB GHOSH* & PRABI R D E** The nineties of the last century have witnessed the emergence of a new role the port infrastructure can play. Governments around the world have undertaken drastic measures for improving the operational efficiency of national ports through institutional reforms. Billions of private capital have been pledged for capacity expansion and service modernization in the world port system during 1991-1998. although India has recently opened this sector for private investments, the performance of the post system is not up to the mark. This paper attempts to find out the role played by port performance indicators (derived from principal component analysis) and labour endowment in determining port traffic ina comparative static framework. Even the use of OLS regression has come out with very satisfactory results. Both explanatory variables have been found to exert positive and significant impact upon port traffic. Finally, the use of a time dummy indicates some sort of stagnancy in Indian ports since liberalization in 1991. Keywords: Production function; principal component analysis; OLS; port performance indicators; port traffic. I. INTRODUCTION Since the deregulation of the US domestic transport networks, the transport industries across the world have witnessed major reorganization (Simon, 1996). 1 In the subsequent period, the port sector in many developing countries has been undergoing transformation from a subsistence infrastructure resource into a more capital intensive, commercially oriented facility. More than US$ 8 billion of private capital have been pledged for capacity expansion and service modernization in Far Eastern and South Eastern Asian ports (Peters, 1995). Ports in the United States have invested about US$ 8billion over the last 20 years; of this, $ 5 billion have been invested just between 1992-1996; and future investments in the subsequent period (1997-2001) are expected to total $ 7 billion in channel deepening and infrastructure projects (Tamaki, 1999). This process which was initiated in the nineties saw many of the ports across the world undertaking major policy changes in order to increase port productivity (Kuby and Reid, 1992). Interestingly, this structural transformation has coincided with significant increases in countries’ exports. In India, interest in the effect of public capital on port infrastructure has increased in recent years partly as a result of numerous reports regarding the fragile state of the nation’s public infrastructure. Estimates of investment needed to provide adequate port facility in 1 India range from US$ 4.5 billion to US$ 5.5 billion by 2005-2006 (Government of India, 1996). The importance of public capital for the growth of the port sector stems from the chain linkage between production, performance of individual ports and overseas transportation leading to exports. 2 Traditionally, by virtue legal stipulations, the government controls national ports in many developing countries. However, rising inefficiencies in ports have forced governments to deregulate the system (Estache and Carbajo, 1997; Bollard and Pickford, 1998). The effects of poor port performance on a country’s trade have become too obvious (Juhel, 1999). Inefficient ports, whether through lack of integrated transport network, outdated work practices or obsolete facilities, can stall a country’s growth even in a borderless world. Sadly enough, effects of infrastructural constrains on a countr’s foreign trade are really less discussed in mainstream economic circles. The unreliability and poor quality of infrastructural services often add to the costs of production of exportables and adversely affect a country’s international competitiveness (Marjit and Roychaudhury, 1997; Srinivasan, 2000). The present paper addresses these issues in the context of India. There are some studies, which have tried to understand the relation between a countr’s openness and port performance (Ghosh and De, 1999 for India), but few have touched upon the empirical causality between port performance and port throughput. Alternative view about the causal relationship between these variables of economic interest exist. Economic theory provides no firm basis to judge whether performance causes traffic or traffic causes performance. There may be a latent simultaneity between the two, given the overall economic perspective of a nation. The purpose of this paper is to find out relationship between port performance, labour endowment and port throughput. The organization of the paper is as follows. Section II deals with data and methodology. Individual performance of Indian ports, along with the construction of a port performance index (PPI) with the help of principal component analysis, is discussed in section III. Section IV tests the impact of port performance and labour endowment on port traffic. Policy implications are discussed in section V. finally, concluding remarks are briefed in section VI. II. DATA AND METHOLODGY The major exploit of this paper is port facility. In general, port facility can be taken as public infrastructure input from the supply side. Here, we have taken eight important variables for 12 major ports for four different time points over the period from 1985 to 1996, to judge relative port performance. These include: (i) ship turn-round time (TRT); (ii) pre-berthing waiting time (PBWT); (iii) percentage of idle time at berth to time at working berth (PITTWB); (iv) output per ship berth day (OSBD); (v) berth throughput rate (BTR); (vi) berth occupancy rate (BOR); (vii) operating surplus per ton of cargo handled (PTOS); and (iii) rate of return on turnover (RRT). The functional values of the port performance indicators appear to give a fairly representative and reasonable picture of port efficiency. Application of traditional production function methodology to spatial production units like airports and seaports, in order to find out the nature and strength of the explanatory variables, is not new, particularly when the sole independent variables used are labour and capital. But beyond the conventional wisdom in production economics, ports, unlike other manufacturing decision making units (DMU), represent a spatial production system which cannot be fully understood simply by the 2 quantity of labour and capital alone, even under equal demand conditions. One must probe deeper into the natural geo-navigational settings in which a port is situated. The adversity or favourability of the inherent locational features of a port ultimately dictate the desired amount of capital expenditures which significantly affect the efficiency of operation at various layers of management, not necessarily all in the port complex. Hence, port performance efficiency is also expected to be highly contingent upon the indicators listed above. Thus, the mere amount of capital is not sufficient. What is important is how this capital is allocated and utilized in order to enhance port performance. Ideally, Y = f(L, K). but in the case of a spatial DMU like a port, the optimal combination of geo-navigational and strategic factors, along with current labour, are ultimately responsible for port performance rather than the mere amount of capital. In ny LDC like India, the problem lies in the fact that a very large proportion of public sector allocation is diluted through the political bureaucracy nexus (Bardhan and Mukherjee, 1998; Marjit, 1999), and the efficiency of utilization of public capital significantly varies across regions (Ghosh, Marjit and Neogi, 1998; ghosh and De, 1998). Thus, there is reasonable consensus among the economists that the mobility of goods, services and labour across the regions depends largely on the quality and quantity of various integrated facilities available, and not directly on the amount taxpayers pay. Naturally, therefore, use of the port performance indicators is likely to better reflect the input output relationship relative to capital in such a context. The point is not that capital is unimportant. Had there been complete information on capital accumulation and some quantifiable measure of public corruption at various layers of fund disbursement and execution across each region, it would have been justified to work with capital stock pertaining to the relevant infrastructures for measuring the performance of the ports. Thus, we propose to use Y = f (L, PPI), although for the purpose of verification we shall include capital as a third independent variable, thereby estimating the following production function: Y = f (L, PPI, K) The details of PPI and its construction are explained in section III. Labour endowment is represented by the total number of laborers employed in each port. Following Nijkamp (1984), one could use activity rate of labourers, i.e. share of each port’s labourers to total Indian port labourers, without any major difficulty. However, in our case, the specific form of the production function is as follows: Y i = f (L i α , PPI i β , K γ ) Taking log of equation (1) and adding a dummy, we get: Ln(Y i ) = a + α Ln(L i ) + Ln(PPI i ) + γ Ln(K i ) + φ D where Y = traffic of individual port; L = Labour; PPI = weighted port performance index; K = capital; I represents individual port (i=1,…,12); and D represent time dummy with D=1 for later years and D=0 for other. The time dummy will help us understand the role of natural rate of growth in terms of a shift effected over a time span. 3 From equation (2), the impact of PPI and labour on traffic can be estimated in this comparative static framework. In the case of a positive association between the independent and dependent variables, the values of α,β, and φ should be greater than zero. Here, α,β and represent the elasticity of traffic with respect to labour, port performance and capital respectively. If time as an independent variable positively influences port traffic, then φ must be greater than zero, too. Negative values of of α,β, and φimply opposite results. This function has been estimated for two different pairs of time points, 1985-1986 and 1996-1997 as well as 1991-1992 and 1996-1997, for all 12 major ports in a cross-section time-series pooled data set. 3 The choice of the initial year time upon a path of trade liberalization with special emphasis on the electronics sector. In 1991-1992, India undertook economic reform as explicit state policy which began with the announcement of the new Industrial Policy in July, 1991. a break at 1991-1992 helps evaluate the nature of the relationship among the variables in the post-liberalisation period.4 Due to the lack of adequate information on performance indicators, major ports have not been considered in this study. The main data have been collected from various issues of (i) Basic Ports Statistics of India, and (ii) Transport Statistics of India –all published by the Ministry of Surface Transport, Government of India. This data set has been supplemented by data from several issues of Major ports of India: A profile, published by the Indian Ports Association, New Delhi. III. MEASURES OF PORT PERFORMANCE We attempt to measure the performance of Indian major ports by developing a composite index, called port performance index (PPI) comprising indicators of operational performance (TRT, PBWT), asset performance (OSBD, BTR, BOR) and financial performance (PTOS, RRT). 5 We have taken eight individual variables, as described in section II, of the major ports, for three different time points, over the period from 1985 to 1996. The definitions of the performance indicators are given in Appendix II. The basic limitation of the conventional method of constructing a composite index a number of indicators is that, often, subjective and fixed weights have to be assigned to individual indicators, which actually vary time and space. To overcome this limitation, we employ the well-known multivariate technique of ‘factor analysis’ (or what is known as ‘principal component analysis’-PCA) from which the weights of the respective factors follow (Fruchter, 1967). In the PCA approach, the first principal component is a linear combination of the weighted variables and it explains the maximum of variance across space. Hence, here, the sole objective of the weighting mechanism is to explain the maximum variance for all the individual indicators across the ports at a point in time. the rationale for using principal component analysis is that it helps one reach an aggregate representation from various individual port performance indicators. Its broader objective is quite pari pasu with homogenizing the overall requirements for the individual indicators across the ports. Construction of PPI We have at our disposal values of eight port performance variables for three different years, 1985-1986, 1991-1992, and 1997-1998, across 12 major ports. The details of the factor loadings (weights) derived from the PCA of these performance variables are presented in 4 Table 1(a) and the estimated PPI in Table 1(b). PPI is a linear combination of the unit free values of the individual indicators such that: PPI ij = Σ W kj X kij where PPI ij = port performance index of the i th port in j th time; W kj = weight of K th indicator in j th time; and X kij = unit free value of the K th indicator for the i th port in j th time. Each individual indicator was made unit free dividing by the standard deviation across the ports. a few observations are worth noting here. first, no single variable has emerged as the most influential factor in all the years. However, by looking at the two consecutive good rankings of OSBD in 1985-1986 and 19911992, it becomes obvious that OSBD has played an influential role in determining the performance of individual ports in the first two years. Similarly, BOR has been found to be an important determinant of performance in 1991-1992 and 1996-1997. thus, tow of the asset performance indicators (OSBD and BOR) have emerged as influential factors in determining the PPI. Second, in the post-liberalisation period, India has witnessed a rise in her overseas trade volumes, particularly between 1993-1994 and 1996-1997. 6 The largest part of this overseas trade has passed through the ports, and this is directly reflected in the port capacity utilization rate. During the period from 1991-1992 to 1996-1997, Indian ports have been over-utilised at an average rate of 102% per annum. Port congestion is, thus, quite likely to prevail in the coming years, if this rate of progression continues. Broadly speaking, when there is high congestion at a port, performance is influenced by efficiency in general as well as by the speed of clearance of vessels from the berth. In the short run, when capacity is overutilised, coordination among various intra-port activities and speed of clearance of vassels are the only ways out for tacking congestion. As a matter of fact, this has been reflected in the year 1996-1997 when two of our operational performance indicators (PBWT and TRT), and one assets performance indicator (BOR) became the first three weightiest factors in determining performance of an individual pert. Table 1(a): Weights of Port performance indicators: PCA Variables Weights Rank Weights Rank 1985-86 1985-86 1991-92 1991-92 TRT -0.536 7 0.131 6 PBWT 0.414 6 0.457 4 ITTWB -0.639 8 -0.655 8 OSBD 0.892 1 0.808 2 BOR 0.561 5 0.698 3 BTR 0.769 4 0.891 1 PTOS 0.789 3 -0.561 7 RRT 0.881 2 0.116 5 1 Eigen Value 1.637 1.666 Exp. Variance 2 0.536 0.555 Weights 1996-97 0.778 0.903 -0.544 -0.183 0.831 0.122 -0.155 0.416 1.873 0.624 Rank 1996-97 3 1 8 7 2 5 6 4 Notes: 1.The factor loading of the eight port performance indicators for three different years is derived by the formula eigen vector = (factor loading)/√(eigen value). Eigen value is the first value of the ‘variance explained’ column in the unrotated factor loading (pattern). 2. Explained variance as % of total. 5 Third, unlike popular belief, financial performance indicators like PTOS and RRI have emerged as factors of low importance in determining PPI in the last two years. But these two financial performance variables played a key role in determining PPI in 1985-1986. Fourth, Table 1(b) presents the values of the performance index of 12 major ports over three different time points. The coefficient of variation (CV) of PPI has substantially declined in 1996-1997 (0.34) even after rising from 0.58 in 1985-1986 to 0.69 in 1991-1992. thus, there has been a tendency towards equalization of inter-port performance index after liberalization. In some sense, this reflects a rise in competitiveness among Indian ports in the post-reform period. Sr. No. Major Ports 1985-86 PPI 1985-86 Rank 1991-92 PPI 1991-92 Rank 1996-97 PPI 1996-97 Rank 1 Kandla 13.45 1 9.34 2 13.62 1 2 Mumbai 7.83 5 1.75 11 10.57 2 3 Jawarlal Nehru * * 3.09 9 7.96 6 4 Mormugao 11.32 2 9.37 1 5.45 10 5 New mangalore 3.69 8 3.35 8 5.04 11 6 Cochin 2.45 10 1.83 10 4.10 12 7 Tuticorin 6.11 7 3.67 7 7.75 7 8 Chennai 9.14 4 4.72 4 9.79 3 9 Vizag 7.74 6 6.41 3 8.11 4 10 Paradip 3.27 9 4.19 5 8.03 5 11 Calcutta 0.68 11 -0.48 12 5.97 9 12 Haldia 9.34 3 3.89 6 6.57 8 Mean 6.82 4.26 7.75 SD 3.97 2.93 2.65 CV 0.58 0.69 0.34 Note: * Not in operation Fifth, although the average performance of the west coast ports has been slightly better than that of the east cost ones, the first four performance ranks have remained unchanged since 1991. in order of ranking, these ports are Kandla, Mumbai, Chennai and Vizag. They also hold the first four positions in india’s total port traffic since 1970-1971. their current (19981999) shares are as follows: Kandla 16.16%, Vizag 14.18%, Chennai 14% and Mumbai 12.30%. table 1 (c) and Figure 1 present the picture more clearly. Therefore, it may not be out of merit to find some sort of a scale economy to exert its positive impact on the performance index. 6 Table 1 (c): Share of major ports in total Indian port traffic (%) Ports 1970-71 1980-81 1990-91 1998-99 Kandla 2.08 10.88 12.96 16.16 Mumbai 18.56 21.10 19.02 12.30 * * 1.34 4.66 14.22 17.10 9.81 7.17 * 1.19 5.28 5.65 6.21 6.50 4.79 5.04 Tuticorin * 3.18 3.34 4.04 Chennai 8.95 12.90 16.14 14.00 Vizag 11.27 12.57 12.78 14.18 Paradip 27.90 2.78 4.53 5.21 Calcutta 7.98 5.04 2.71 3.63 Haldia 2.80 6.76 7.31 7.96 Jawarlal Nehru Mormugao New Mangalore Cochin Note: * Not in operation Figure 1: Scatter of cross section data on port traffic against PPI (1985-1986 and 1996-1997) 7 Finally, ranks of individual ports in terms of individual performance indicators given in Table 1(d)- make it clear that except Chennai, Vizag and Haldia the other ports of east coast have been gradually retreating in terms of most of the indicators. Kandla port always hovered around the top position. Mumbai port performed well in the post liberalization period and reached the second position in 1996-1997. Port performance retrogression is very significant for Mormugao and Haldia. However, these two ports performed well in 1996-1997 compared to 1985-1986 in areas like PBWT, OSBD, BTR (Mormugao) and PBWT, BTR, PTOS (Haldia). Table 1 (d): Rank of ports in individual performance indicators Year = 1985-86 Ports Year = 1991-92 TRT PBWT ITTWB OSBD BOR BTR PTOS RRT TRT PBWT ITTWB OSBD BOR BTR PTOS RRT Kandla 6 2 11 2 1 2 2 1 2 1 12 4 1 2 11 3 Mumbai 2 1 3 8 3 9 1 5 3 5 2 9 8 11 4 7 Jawarlal Nehru * * * * * * * * 10 2 11 11 6 7 1 10 Mormugao 5 4 9 1 7 1 7 4 6 10 8 1 5 1 12 11 New Mangalore 4 7 1 10 6 8 8 8 9 3 1 7 10 8 7 2 Cochin 8 11 2 9 10 10 9 9 12 11 2 8 12 10 8 9 Tuticorin 10 10 3 7 9 6 6 3 11 7 7 10 4 9 10 6 Chennai 9 9 3 5 2 4 3 2 5 4 5 6 3 4 5 1 Vizag 7 6 6 4 8 5 5 6 8 9 9 2 2 6 9 8 Paradip 3 3 8 6 11 7 10 10 4 8 9 5 9 5 3 4 Calcutta 1 5 7 11 4 11 11 11 1 12 6 12 11 12 6 12 Haldia 11 7 10 3 4 3 4 7 5 5 2 3 6 2 2 4 Note: * Not in operation Year = 1996-97 Ports TRT PBWT ITTWB OSBD BOR BTR PTOS RRT Kandla 2 1 11 4 1 2 11 2 Mumbai 1 2 7 11 5 11 4 10 Jawarlal Nehru 9 5 10 9 7 7 2 4 Mormugao 5 12 8 1 9 1 12 12 New Mangalore 11 9 2 2 12 6 6 1 Cochin 12 10 3 6 11 10 7 11 Tuticorin 10 6 5 10 2 9 10 3 Chennai 3 3 6 8 4 8 8 5 Vizag 7 6 12 3 6 4 9 9 Paradip 8 6 9 7 3 5 5 8 Calcutta 4 4 4 5 10 12 1 6 Haldia 5 10 1 11 7 2 3 6 8 To conclude, according to the PPI in 1996-1997, the best three ports are Kandla, Mumbai and Chennai, and the worst three ports are Mormugao, New Mangalore and Cochin. To be more specific, values of these scores are much more important than mere ranking. High variance in both PPI and individual factors, among other things, represent the level of differential interport performance. Except for the first three ports, lower scores for the remaining ports represent their potential for further rise. To understand the nature of data, it may be useful to review the inter-port variations of each of the eight performance variables as they are given in the form of raw data over time. the values of mean, SD and CV of the raw indicators of port performance are given in Appendix III. Except OSBD, BTR, PTOS, and RRT, the coefficients of variation (CV) for the rest of the variables have been rising over the years. Among these four variables, only RRT has become more equitable across the ports over time. That is, the value of CV of this variable has fallen from 0.64 in 1985-1986 to 0.41 in 19911992 and 0.19 in 1996-1997. in contrast to this, PBWT displays the highest disparity-more that doubling from 0.33 in 1985-1986 to 0.76 in 1996-1997. This being a reasonable good indicator of port congestion suggests that there has been high variance in congestion across the ports. IV. LINKAGE BETWEEN TRAFFIC AND PERFORMANCE As deemed by the present study, we have tried to find out the nature of the relationship between port performance, labour employment and capital with port throughput over different time spans. The estimated results of the OLS estimation of equation (2) are presented in Table 3 along with the values of the coefficients, t-statistics, adjusted R 2 , Durbin-Watson statistic (serial correlation), standard error of estimates and F values for two combinations of data sets – (i) 1985-1986 and 1996-1997 (case-1), and (ii) 1991-1992 and 1996-1997 (cse-2)separately. While the former explains the causality between endogenous and exogenous variables over an interval of 11 years, the later case evaluates the immediate impact of trade liberalization undertaken in July 1991. in case 1, Calcutta port was omitted as an outlier and Jawarlal Nehru was excluded because the port was not in operation in 1985-1986. in case 2, both Calcutta and Mumbai ports came out as outliers on the basis of Cook’s distance statistic (Cook and Weisberg, 1982). Hence, values of the parameters have been estimated after omitting these two ports. Regression results are highly satisfactory. A look at the correlation matrices given in Table 2 makes certain features about the independent variable obvious. First, in no situation, L and PPI are significantly correlated. Second, in one situation (case 1, 1985-1986), K is significantly correlated with L. interestingly, the same correlation in 19911992 is also marginally significant. Finally, K does not cause any significantly high correlation with PPI in any year. This creates a ground for including K as an coefficient variable with PPI and L in both pairs of years. Therefore, no correlation coefficient is significant enough to warrant the presence of multicolinearity among the variables. Table 2: Correlation matrix (a) Case 1 (1985-1986 and 1996-1997) PPI (96-97) PPI (96-97) L (96-97) K (96-97) PPI (85-86) L (85-86) K (85-86) 1 0.311 p=.351 0.517 p=.104 0.538 p=.088 0.375 p=.256 0.703 p=.016 9 L (96-97) 1 K (96-97) 0.443 p=.173 0.113 p=.742 0.646 p=.032 0.273 p=.416 1 0.391 p=.234 0.164 p=.630 0.307 p=.359 1 -0.213 p=.530 0.065 p=.850 1 0.698 p=.017 PPI (85-86) L (85-86) K (85-86) 1 (b) Case 2 (1991-1992 and 1996-1997) PPI (96-97) L (96-97) K (96-97) PPI (96-97) L (96-97) K (96-97) PPI (91-92) L (91-92) K (91-92) 1 0.322 p=.308 0.528 p=.078 0.380 p=.223 0.276 p=.385 0.525 p=.079 1 0.452 p=.140 -0.373 p=.233 0.946 p=.000 0.413 p=.183 1 -0.157 p=.627 0.596 p=.041 0.993 p=.000 1 -0.388 p=.212 -0.125 p=.698 1 0.571 p=.052 PPI (85-86) L (85-86) K (85-86) 1 Notes: 1. L = Labour, K = Capital, PPI = Port Performance Index 2. N = 22 in case 1, and 24 in case 2 3. Correlation coefficients are significant at p,0.05 10 Table 3: Regression results (a) Case 1 (1985-1986 and 1996-1997) Dependent Independent Coefficient Variable Variables Traffic (Y) Constant 8.057 tvalues 1.476 Ln(PPI) 0.530 2.458 Ln(L) 0.308 2.463 Ln(K) 0.205 0.697 D 0.626 3.830 Constant 0.139 0.028 Ln(L) 0.241 1.722 Ln(K) 0.645 2.409 D 0.577 3.099 Constant 11.821 15.108 Ln(PPI) 0.621 3.692 Ln(L) 0.365 3.893 D 0.671 4.543 Constant 0.790 0.150 Ln(PPI) 0.414 1.715 Ln(K) 0.675 2.627 D 0.500 2.807 Traffic (Y) Traffic (Y) Traffic (Y) Table 3: Regression results (b) Case 1 (1991-1992 and 1996-1997) Dependent Independent Coefficient Variable Variables Traffic (Y) Constant 13.196 t-values Ln(PPI) 3.581 Traffic (Y) Traffic (Y) Traffic (Y) 0.692 3.398 Ln(L) 0.643 5.638 Ln(K) -0.154 -0.807 D 0.228 1.404 Constant 7.251 1.567 Ln(L) 0.695 4.657 Ln(K) 0.148 0.655 D 0.523 2.837 Constant 10.145 11.577 Ln(PPI) 0.623 3.634 Ln(L) 0.621 5.674 D 0.238 1.488 Constant 12.567 1.893 Adj. R Sqr. 0.760 SEE F DW SC N 0.331 16.036 2.240 -0.176 20 0.684 0.380 14.726 1.968 -0.026 20 0.768 0.326 21.924 2.327 -0.229 20 0.684 0.380 14.703 2.189 -0.133 20 Adj. R Sqr. 0.789 SEE F DW SC N 0.307 18.716 2.312 -0.258 20 0.632 0.404 11.893 2.503 -0.352 20 0.793 0.303 25.289 2.314 -0.249 20 0.382 0.524 4.911 2.480 -0.267 20 11 Ln(PPI) 0.830 2.532 Ln(K) 0.111 0.350 D 0.101 0.367 Notes: 1. Calcutta port was deleted as outlier following Cook’s Distance Statistics, and JNP was excluded as the port was not in operation in 1985-1986. 2. In case 2, Mumbai and Calcutta ports were detected as outliers following Cook’s Distance Statistic. 3. SEE, F, DW, SC and N stand for standard error of estimate, F value, Durbin Watson statistic, serial correlation, and number of observations. 4. D = Time Dummy (=1 for later year, and 0 for other). 5. Each port’s capital was deflated by gross domestic capital formation deflator (base: 19801981=100), and accumulated over time up to the concerned time point. For each pair of years, we have tested four combinations of independent variables. Some findings are as follows. First, it is observed from Table 3 that in both cases best results, relatively speaking, are obtained from the combination of PPI and L and time dummy as can be judged by the standard error of estimate. Interestingly, the inclusion of K as a third variable not only marginally reduces the value of F-statistic. And while doing so it fails to come out with any significance as can be judged from lower t-values. Second, the traditional combination of L and K makes the regression results substantially poorer. Third, as analysed by us earlier, the best result is obtained only with the combination of PPI and L with the presence of time dummy in both cases. Given the limited number of cross section observations, the high values of adjusted R 2 (0.768 and 0.793) along with the required values of DW and F-statistic confirm the fact that only two factors, performance index and labour employment, explain a very high proportion of traffic across Indian ports. Port traffic is highly contingent upon port performance, the coefficient (or elasticity) of which is not only very high (0.621 in case 1 and 0.634 in case 2). The elasticities of traffic with respect to PPI for both cases indicate that every 1% increase of PPI causes 0.62% higher traffic. Despite the marginal difference between the coefficients of PPI in the two cases, there is still scope for further development of performance levels of ports in the coming years. We have also found positive and significant association of labour employment over time with traffic. The coefficient of labour in case 1 (0.365) is lower than that of case 2 (0.621). while in case 1, for 1% increase in labour employment, traffic increases by 0.37%, for the same increase of labour input in case 2, traffic is higher by 0.62%. it is also found that the elasticity of output with respect to PPI has been higher than that of L in both cases. This has become more transparent in case 2 where labour endowment appears to be significant due to the fact that, in the post-liberalisation period, total labour force of major ports has been declining fast. This may also be substantiated with the help of table 4. it is seen that annual growth rate of traffic of almost all major ports has been much higher than that of labour. In fact, except Mumbai, New Mangalore and Haldia, all the major ports have registered negative growth of labour force. The above relationship also indicates that a few ports need to consolidate their labour pool. Inclusion of the time dummy has helped in understanding the production system in port between each pair of years. As it can be seen from Table 3, the value of the coefficient of the dummy vcariable is high, positive and significant between 1985-1986 and 1996-1997. 12 but, as expected, it was insignificant between the later pair of years. This suggests that no significant breakthrough has occurred in india’s port performance behaviour ever since the reform in 1991. alternatively, the significant positive value of the time dummy for the earlier pair of years implies that natural rate of progression has occurred over a time lag of 11 years. Table 4: Average annual growth rates of traffic and labourers Major ports 1985-86 to 1996-97 Traffic Labour (%) (%) Kandla 9.50 -0.79 Mumbai 1991-92 to 1996-97 Traffic Labour (%) (%) 12.12 -3.49 3.52 0.52 5.69 3.26 * * 37.85 5.93 Mormugao 0.67 -0.63 2.93 0.03 New Mangalore 21.58 7.16 10.11 14.21 Cochin 11.12 -1.72 11.39 -2.39 Tuticorin 10.69 -1.26 11.28 -2.88 Chennai 6.86 -1.20 5.43 -1.84 Vizag 10.62 -3.80 12.06 -8.60 Paradip 22.52 -2.09 11.73 -2.42 Calcutta 4.06 -4.49 8.94 3.18 Haldia Note: *Not in operation 10.41 2.38 8.89 3.18 Jawarlal Nehru It may be concluded that to attain higher traffic, ports obviously should give highest priority to their performance by improving operational performance factors like PBWT, TRT, and asset performance indicators like BOR. These three most important determinants of port performance indicators like BOR. These most important determinants of port performance as appeared in the post-liberalization period from principal component analysis are related to the degree of port congestion. Hence, for attracting higher traffic, ports have to monitor these crucial factors very closely so that higher efficiency indues higher traffic. On the other hand, under the present trend of traffic, there is sufficient hint in favour of such traffic. On the other hand, under the present trend of traffic, there is sufficient hint in favour of such policy measures by which to create further port capacity. Thus, the use of a quasi-production function demonstrates that port performance index and labour endowment alone explain a very large proportion of traffic in Indian ports. V. POLICY IMPLICATIONS The findings of this study, among other things, have strong implications for future policy changes pertaining to Indian ports, and as a matter of fact, to the ports of developing nations in this era of globalization. It is observed that Indian ports need to emphasise more on ‘decongrestion’ for making exports competitive. The lagging ports may not be able to improve their performance instantaneously because the resilient external factors, which have very strong influence on performance, are beyond the control of the port authorities. Generally 13 speaking, a nationally favourable environment would be more congenial for improving port performance lavels before undertaking any drastic policy measures. Sometimes, lack of coordination between railways, roadways and coastal shipping authorities put the port in a very dismal position, which in subsequent periods negatively impacts upon the growth of export by raising transaction costs. It has been observed that there has been a shift in long distance coastal cargo transportation from ports to railways and/or roads, due to the poor performance and differential freight systems across ports. the average ship turnaround time in 1996-1997 was 7.8 days. It was even higher, 8.1 days, in 1990-1991 (Government of India, 1998). Compared to Singapore’s port where the turnaround time is counted in hours, Indian ports are just globally uncompetetitive in terms of any crude measures.according to Peters (1997): ‘…… the reason for this sorry state of affairs is that our ports while other types of traffic specific categories of cargo which have configurations were not adjusted to the new categories of cargo. In almost all ports productivity levels are extremely low by international standards. Due to this, the added costs to our exporters and importers are substantial’. There is no denying the fact that the question of efficiency is inextricably linked to the appropriateness of the chosen technology, and more often than not the fault may lie in the institutional preparedness for scientific management which is necessary for the smooth functioning of the new technology. As for the manufacturing sector so also for the port sector, economists in the developed countries have long been persuaded about the role of R & D in fostering appropriates technological manoeuvre which in turn helps achieve faster allround productivity growth in an economy (Giriliches, 1980; Nadiri, 1980; Ishikawa, 1981; Ghosh and Neogi, 1993). But the governments as well as the economists of the LDCs bothered little about the role technology can play in intergrating manufacturing, awareness for trade, transaction and trade itself. India is no exception. Port sector in India is not only one of the most ignored sectors but also technological obsolescence epitomizes the official attention received by Indian ports since the First Five Year Plan in 1951 to the beginning of the Ninth Plan in 1997. it is in very recent years that failure in the export fromt has been sporadically linked to the non-price factors in general and inefficiency of ports in particular. Effective utilization of existing facilities in ports not only helps tackle congestion more efficiently but it has also a tremendous spill over effect among manufacturers enabling them to turn to overseas markets. Despite tremendous data limitation relating to technological level of Indian ports, we have been able to present in Table 5 some information about the equipment and its utilization rates across the major ports. the argument is not just that the utilization rates of all three types of cranes for handling general cargo are unequivocally suboptimal. It rather implicitly suggests the level of obsolescence of equipment in use in Indian ports. even the Ninth Five Year Plan (1997-2002) acknowledges this: ‘…….equipment utilization has been low in most of the major ports. low productivity is mainly due to the operational constraints such as equipment breakdown, time spent on service and power failures, etc. Over-aging of installed equipment is another area of concern. Out of the total fleet strength, 88% of wharf cranes, 66% of mobile cranes, and 31% of fork lift trucks have crossed their economic life’. The net result of all this is that some ports handle much less cargo than what they did 50 years ago. Interestingly, the same conclusion was reached by Haralalmbides and Behrens (2000). According to them, Indian ports are currently characterized by the existence of obsolete and poorly maintained equipment, hierarchical and 14 bureaucratic management structures, excessive labour, and in general an institutional framework that is considerably in variance with Government’s overall economic objectives. Table 5: Average monthly utilization of general cargo handling equipment Major ports 1 Wharf Mobile Cranes 2 Cranes 19851991199619851991199686 (%) 92 (%) 97 (%) 86 (%) 92 (%) 97 (%) Kandla 29.17 30.14 41.11 5.28 2.92 0.28 Mumbai 18.19 17.50 28.47 24.44 17.22 Mormugao 15.69 24.17 8.89 4.03 New Mangalore 3.33 4.31 0.97 Cochin 9.44 6.94 Tuticorin 7.92 Chennai Vizag Paradip 3 Calcutta 4 Notes: 1. 2. 3. 4. 198586 (%) 24.31 Fork Lifts 199192 (%) 27.22 199697 (%) 27.22 19.03 35.97 24.72 20.42 9.58 4.86 10.69 4.58 28.06 4.03 2.78 0.97 6.39 3.19 4.44 11.39 21.94 15.97 8.75 16.94 18.61 19.31 22.08 17.78 3.47 14.58 3.33 5.00 10.97 10.69 57.36 38.47 24.03 47.08 35.28 38.19 46.67 37.50 30.69 17.22 25.97 31.11 23.61 11.81 19.72 21.39 18.61 20.42 13.89 38.33 20.14 13.89 17.64 8.06 4.86 19.03 9.72 29.72 22.36 11.53 34.44 30.42 17.08 36.25 35.97 21.94 Jawarlal Nehru and Haldia ports were omitted due to back of required data. Counts only electric wharf cranes of 3-10 tonnes. There were no 3-10 tonnes wharf cranes working and all were above 10 tonnes cranes. Only 2-10 tonnes wharf cranes were considered. Relising the urgent need for inter-port and intra-port competition to attain higher performance level, the Government has recently invited the private sector to finance new port facilities (Government of India, 1999). But success to date has been very limited. Continued deliberation of the same approach by the Government can be expected to lead to more disappointments. Many nations have provided lots of examples for alternative ways of port development through ‘win-win’ strategies. Time is ripe for India to adopt a commercially viable approach so that investments flow according to the market requirements. VI. CONCLUDING REMARKS This paper uses some individual port performance indicators, which are largely internal to each of the ports, for explaining port throughput. These factors alone explain a substantial proportion of traffic. Hence, even though we did not consider the external factors, it may be concluded that performance does lead to increased traffic. Efficient management can lead to significant increases in port throughput, which may then increase the demand for port services. If a port performs better by improving its operational and asset performance initially, then it is going to get higher traffic. However, we have not considered here the spatial concentration of manufacturing activities in and around the ports, differential geographical attributes, logistics networks, and differential nation-wide infrastructure stocks associated with each region which may have very substantial bearings on port traffic. The findings of this paper may be further substantiated incorporating these exogenous factors. 15 This will help us evaluate the performance of the port sector in a general framework under a liberal economic regime. END NOTES --------------1 However, empirical studies of the transport industry have provided little evidence of the presence of scale economics. Specialized research works in this field have been made by keeler (1974, 1983) on railways; Caves, Christensen and Tretheway (1984), Bauer (1990), Khumbhakar (1992) on airlines; McMullen (1987), Winston, Corsi, Grimm and Evans (1990) on trucking industries; and Coto-Millan et. Al. (2000), Notteboom et. Al. (2000) on ports. 2 Public capital in port sector (such as radio wireless station, pilot vessels, mooring boats, tugs, cranes and trailers of various types, warehousing and storage facilities, water supply and energy facilities, fire fighting utilities, conservancy, etc.) could be viewed as inputs in the service process which contribute independently to the cargo handling. Under the present institutional set up, the role of private capital in Indian ports is very limited. Yet, it is generally taken for granted that decades of ‘sheltered market’ phenomenon have not created among Indian industrialists any ‘outward looking’ tendency. Moreover, high ‘gestation lag’ in the port sector is believed to be the main cause for low level of private capital in this sector even under the selective liberalization initiated by the Government in the post 1991 period. 3 In India, there are 12 major ports in total; namely Kandla, Mumbai, Jawarlal Nehru, Mormugao, Cochin, New Mangalore, Chennai, Tuticorin, Paradip, Vizag, Calcutta, and Haldia. One may consult Appendix I for classification of Indian ports and their institutional set-up. 4 India had adopted central planning ever since the First Five Year Plan (1951-1956). 1991 (July) marks the beginning of the era of market economy away from centralized planning. The new economic policy became effective from this year. The salient features of this policy, to start with were: (i) Permitting direct foreign capital in industries, trading companies and banking up to 51% of share capital (in some cases, 100%); (ii) Automatic clearance for capital goods import; (iii) Automatic approval of foreign technology agreements in high priority areas; (iv) settingup of Foreign Investment Promotion Board for negotiating with MNCs and granting single point clearance; (v) Permission of private sector banking; (vi) Other measures including the abolition of industrial licensing except for limited areas, abolition of MRTP and FERA, closing down of chronically sick public sector units etc. sadly enough, no major fundamental policy change pertaining to port sector could be visualized in 1991. 5 For a critical and comprehensive review of the methods for measuring performance of a port system, one may consult Chung (1993); Frankel (1987); Fourgeard (1999); Plumlee (1982); UNCTAD (1983, 1987). 6 The liberlisation process during 1991-1992 to 1995-1996 has enhanced the importance of international trade in the so far closed economy of India. To be precise, the share of India’s trade in GDP has increased to more than 24% in 1995-1996. During the 90s, higher growth has been recorded in India’s imports leading to decline of the trade deficit from US$ 5-6 billion per annum in the 80s to around two billion in 1994-1995. The revival of exports began in 1992-1993 followed by three years of strong ranging between 18.4% and 20.7% in US$ terms. A large number of measures for the control of imports have been dismantled. Also a number of quantitative restrictions, which were imposed in the earlier protectionist regime, have been done away with. APPENDIX I ----------------- India is endowed with an extensive coastline of about 6000 kms along nine coastal states. These states are Gujrat, Maharastra, Karnataka, Goa, Kerala (west coast) and Tamil Nadu, Andhra Pradesh, Orissa, West Bengal (east coast). These nine states have in total 12 major and 179 minor ports. among these 12 major ports, six are located in the west coast (Kandla, Mumbai, Jawarlal Nehru, Mormugao, Cochin, New Mangalore) and six in the east coast (Chennai, Tuticorin, Paradip, Vizag, Calcutta, Haldia). Four of the major ports viz. Calcutta, Mumbai, Chennai, and Mormugao are more than 100 years old. Cochin and Vizag ports have recently celebrated their golden jubilee. The ports of Kandla, Tuticorin, New Mangalore and 16 Paradip came into existence after independence. Jawarlal Nehru port become operational only after 1989. ennore (in Tamil Nadu) is the first corporate port of India, having come into existence in 1999. Out of 179 minor ports including 13 non-working ports, 120 ports belong to west coast comprising 67% of total Indian minor ports; 24 ports belong to east coast; the rest 35 belong to Union Teritories. Due to the lack of overseas cargo, some maritime states like Andhra Pradesh and Kerala closed down a few minor ports which are called now non-working ports. maharastra has the highest number of ports-two major and 53 minor. Next to it is Gujrat, where one major and 40 minor ports are situated. West Bengal is the only maritime states along with corresponding major ports. Ports in India are classified as ‘Major Ports’ and ‘Other (minor) Ports’. Major Ports come under the jurisdiction of the Central Government and, by virtue of entry 27 in the List I of Schedule VII of the Constitution of India, form part of the Central subject. The Major port Trusts Act, 1963, and the Indian Ports Act, 1908, basically govern them. Ports other than Major Ports are included among the subject of the List III (concurrent list entry 31), and hence are controlled and operated by state government subject to certain Central Legislation being operative on them as well. The primary responsibility for the development and management of minor and intermediate ports rests with the state government within the purview of Indian Ports Act, 1908. The Indian Ports Act of 1908 describes the regulatory powers of the Port Authority whereas the Major Port Trusts Act of 1963 enables the port to conduct its regulatory as well as commercial functions. The Indian Ports Act applies automatically to all ports and parts of navigable rivers and channels leading to the ports, irrespective of category, i.e. major or minor. The Major Port Trusts Act is restricted to the port proper of the major ports. ‘Major Ports’ may be any ports which the Central Government may, by notification in the Central Gazetter, declare under the Indian Ports Act, 1908, or may under any law for the time being in force have declared to be a major port (ports of Calcutta, Mumbai and Chennai were declared as ‘MajorPort’, pursuant statutory enactment dated 16.12.1920, and were brought under the direct control of the Central Government by virtue of the Seventh Schedule of the Government of India Act, 1935). ‘Other Port’ means any port, which the state government may, by notification in State Gazette, declare as a minor port. APPENDIX I ----------------- 1. Ship turn-Round Time is the duration of the vessel’s stay in port and is calculated from the time of arrival to the time of departure. 2. Pre-Berthing Waiting Time is the time a ship has to wait before getting entry into a berth. 3. Percentage of Idle Time at Berth to Time at Working Berth is the ratio of total idle time to total working time while a ship is in the port. 4. Output per Ship Berth Day means total tonnage handled, or distributed over the total number of ship berth days. 5. Berth Throughput means total cargo handled by a berth in a port. 6. Berth Occupancy Rate is the time that a berth is occupaied by ships. 7. Operating Surplus per ton of Cargo Handled is derived from total operating surplus divided by total tonnage of cargo handled by the port. 17 8. Rate of Return on Turnover is derived from operating surplus devided by operating income of a port. APPENDIX III ------------------Port Performance Indicators: Mean SD 1 and CV 2 Performance Variables 1985-86 1991-92 1996-97 Mean 10.27 6.37 6.76 SD 2.75 1.82 2.24 CV 0.27 0.29 0.33 Mean 2.92 1.46 2.38 SD 0.97 0.74 1.80 CV 0.33 0.51 0.76 Mean 35.73 34.17 31.63 SD 11.74 11.66 12.56 CV 0.33 0.34 0.40 Mean 3073.45 4458.42 5031.67 SD 1773.89 2506.94 2071.99 CV 0.58 0.56 0.41 Mean 70.40 66.87 73.01 SD 8.34 12.73 15.85 CV 0.12 0.19 0.22 Mean 99.63 116.33 143.10 SD 86.90 74.89 85.67 CV 0.87 0.64 0.60 Mean 1.21 2.34 5.71 SD 0.97 1.27 4.12 TRT (in days) PBWT (in days) PITTWB (in %) OSBD (in tones) BOR (in %) BTR (in %) PTOS (in Indian rupees) 18 CV 0.81 0.54 0.72 Mean 27.07 27.47 40.02 SD 17.33 11.33 7.55 CV 0.64 0.41 0.19 RRT (in%) Notes: 1. Standard Deviation. 2. Coefficient of Variation REFERENCES ------------------BARDHAN, p. AND Mookherjee, D. 1998: Expenditure decentralization and the delivery of public services in developing countries. Working paper No. C98-104. Berkeley: Centre for International and Development Economics Research, University of California. Bauer, P.W. 1990: Decomposing TFP growth in the presence of cost inefficiency, non-constant returns to scale, and technological progress. Journal of Productivity Analysis 1, 287-299. Bollard, A. and Pickford, M. 1998: Deregulation and competition policy in the transport sector in New Zealand. Journal of Transport Economics and Policy 32, 267-276. Caves, D.W., Christensen, L.R. and Tretheway, M.W. 1984: Economics of density versus economics of scale: why trunk and local service airline costs differ. Rand Journal of Economics 15, 471-489. Chung, K.C. 1993: Port performance indicators. Working paper No. PS-6. Wshington, D.C.: Transportation, Water and Urban Development Department, the World Bank. Cook, R.D. and Weisberg, S. 1982: Residuals and influence in regression. London: Chapman and Hall. Coto-Millan, P., Banos-Pino, J. and Rodriguez-Alvarez, A. 2000: Economic efficiency in Spanish ports: some empirical evidence. Maritime Policy & Management 27, 169-174. Estache, A. and Carbajo, J. 1997: Competing private ports: lessons from Argentina. In Klaus Tilmes (ed.), The Private sector in infrastructure: strategy, regulation, and risk. Washington D.C.: The World Bank, 129-132. Fourgeaud, Patrick 1999: Measuring port performance. Washington, D.C.: Transportation, Water and Urban Development Department, the World Bank.. Frankel, E.G. 1987: Port planning and development. New York: John Wiley & Sons, 141-216. Fruchter, B. 1967: Introduction to factor analysis. New Delhi: Affiliated East West Press. Ghosh, B. and De, P. 1998: Role of infrastructure in regional development: a study of India over the plan period. Economic and Political Weekly 33, 3039-3048. Ghosh, B. and De, P. 1999: Indian ports and globalization: need for new policy initiatives. Mimeo, ISI, Calcutta. Forthcoming in Economic and Political Weekly. Ghosh, B., Marjit, S. and Neogi, C. 1998: Economic growth and regional divergence in India: 1960-1995. Economic and Political Weekly 33, 1312-1321. Ghosh, B. and Neogi, C. 1993: Productivity slowdown. American Economic Review 70. Government of India. 1996: The India infrastructure report: Policy imperatives for growth and welfare. New Delhi: National Council of Applied Economic Research. Government of India. (1998) Economic survey 1997-98. New Delhi: Ministry of Finance. Government of India. 1999: Plan document: Ninth five year plan (1997-2002). New Delhi: Planning Commission, vol. II, 765-865. Haralambides, H.E. and Behrens, R. 2000: Port restructuring in a global economy: an Indian prespective. International Journal of Transport Economics 27, pp. 19-39. Ishikawa, S. 1981: Essays on technology, employment and industrialization in economic development. Tokyo: Kinokuniya Co. Juhel, Marc H. 1999: Global challenges for ports and terminal in the new era. Ports and Harbors March, 17-27. Keeler, T.E. 1974: Railroad costs, returns to scale and excess capacity. Review of Economics and Statistics 56, 201-208. Keeler, T.E. 1983: Railroads, freight and policy. Washington D. C.: The Brookings Institute. Kuby, M. and Reid, N. 1992: Technological change and the concentration of the U.S. general cargo port system: 197088. Economic Geography 68, 272-289. Kumbhakar, S. 1992: Allocative distortions, technological progress, and input demand in US airlines: 1970-1984. international Economic Review 33, 723-737. Marjit, S. 1999: Decentralised financing, governance and public-private cooperation. Economic and Political Weekly 34, 3315-3320. Marjit, S. and Roychaudhury, A. 1997: India’s export: an analytical study. New Delhi: Oxford University Press. 19 McMullen, B.S. 1987: The impact of regulatory reform on US motor carrier costs. Journal of Transport Economics and Policy 21, 307-319. Nadiri, M.I. 1980 Sectoral productivity slowdown. American Economic review 70. Nijkamp, Peter 1984: A multidimensional analysis of regional infrastructure and economic development. In Andersson, A.E., Israd, W. and Puu, T. (eds.), Regional and industrial development theories, models and empirical evidence. Amsterdam: North-Holland, 267-293. Notteboom, T., Coeck, C. and Broek van den J. 2000: Measuring and explaining the relative efficiency of container terminals by means of Bayesian stochastic frontier models. International Journal of Maritime Economics 2. Peters, Hans. J. 1995: Private sector involvement in east and south-east Asian ports: an overview of contractual arrangements. Working paper No. PS-10. Washington D.C.: Transportation, Water and Urban Development Department, the World Bank. Peters, Hans. J. 1997: Reforming India’s port system. A position paper presented at the Japan Chamber of Commerce and the World Bank workshop on the Power and transportation Sector: India, Tokyo. Plumlee, C.H. 1982: Port performance index. Washington, D.C.: Transportation, Water and Urban Development Department, the World Bank Simon, David. 1997: Transport and development in the third world. London: Routledge. Srinivasan, T.N. 2000: Eight lectures on India’s economic reforms. New Delhi: Oxford University Press. Tamaki, Desmond C. 1999: Consolidation in the maritime industry and its effects on North American trade: a U.S. east coast case study. Ports and Harbors September, 12-14. UNCTAD. 1983: Mannual on a uniform system of port statistics and performance indicators. Geneva. UNCTAD. 1987: Measuring and evaluating port performance and productivity. Geneva. Winston, C., Corsi, T.M., Grimm, C. M. and Evans, C. 1990: The economic effects freight deregulation. Washington D.C.: The Brookings Institute. 20