Determinants of private investment in Indian states: an application of panel fixed effect model.
Article Type:
Investments (Research)
Economic growth (Management)
Mallick, Jagannath
Pub Date:
Name: Indian Journal of Economics and Business Publisher: Indian Journal of Economics and Business Audience: Academic Format: Magazine/Journal Subject: Business; Economics Copyright: COPYRIGHT 2011 Indian Journal of Economics and Business ISSN: 0972-5784
Date: March, 2011 Source Volume: 10 Source Issue: 1
Event Code: 310 Science & research; 200 Management dynamics Computer Subject: Company business management
Geographic Scope: India Geographic Name: India Geographic Code: 9INDI India

Accession Number:
Full Text:

The private investment driven economic growth is persistently diversified in nature among the Indian states. There is no detailed study which examines private investment at the state levels due to the paucity of data. The present study uses appropriate data and analyses the determinants by applying panel data model for 15 major states during the period from 1993-94 to 2004-05. The principal component method is used to construct the physical infrastructure and quality of governance indices in order to avoid the problem of multicolinearity in the multiple regression analysis. This study viewed that the private investment depends upon physical infrastructure, quality of governance, availability of finance, economic uncertainty and labour productivity along with the state specific factors. This paper suggests that, the above determinant factors should be given high importance while designing policy for the equitable allocation of private investment and hence to achieve the balanced regional growth across states in India.


Economic reform of Indian economy in 1991 included, among other measures, the liberalisation of state control over the private sector, such as, the abrogation of the Industrial Licensing Act with investors at liberty to choose their preferred investment destination amongst the various states in India, there arose an intense competition across states to obtain a share in the upcoming investment projects. This resulted in most state governments striving to better the investment climate prevailing in their respective states to make them conducive to inflows of domestic and foreign investment. Inflow of private investment in a state has the potential to generate state income and output along with providing employment opportunities to local residents. Inter-state private investment is the crucial determinant factor for inter-state economic growth in India [Alhuwalia, 2002; Baddeley et al., 2006; Rao et al., 1999; Krishna, 2004]. Therefore, states do compete with each other to attract private investment to achieve high economic growth. An additional benefit emanating from investment inflows is an upsurge in tax revenues. The direct tax returns from profitable investment projects (and indirect returns from central transfers through rising individual incomes within states) have the potential to improve the fiscal positions of individual states.

The role of states in determining the inter-state allocation of private investment can't be ignored. Though the federal framework of India causes all states to face certain common macroeconomic policies such as monetary policy and trade policy, states do retain extensive control over local administrative regulations, provisioning of infrastructure, state taxation, and provision of basic social services such as health and education. A question arises as to, how do states compete with each other in order to attract private investment in Indian economy during the economic reform periods. This is the un-explored area in the field of empirical research in India. The theoretical literatures argue that the factor prices determine the inter-regional allocation of private investment (Carlberg, 1981). In this context, the specific objectives of this paper are as follows: (1) to analyse the patterns of inter-state private investment and (2) to examine the determinants of inter-state private investment.

Fifteen major states are used for this study, viz., Andhra Pradesh (AP), Assam (ASM), Bihar (BH), Gujarat (GU), Haryana (HA), Karnataka (KA), Kerala (KE), Madhya Pradesh (MP), Maharashtra (MR), Orissa (OR), Punjab (PN), Rajasthan (RA), Tamil Nadu (TN), Uttar Pradesh (UP) and West Bengal (WB). Bihar, MP and UP are adjusted due to the partition of the states in 2000. These states account for a little over 90 per cent of the population, 82 per cent of states' incomes and 80 per cent of national private gross fixed capital formation (GFCF) in India from 1993-94 to 2004-05. Following the introduction, the measurement of state-level private investment in India is reviewed in section II. Inter-state variation in private investment is analysed in section III. Sections IV, V and VI describe the theoretical framework, methodology, data sources and results on determinants of allocation of private investment, respectively. The concluding remarks from the paper are presented in section VII.


The investment is the creation of capital or net addition of capital stock. Investment is usually measured by the Gross capital formation (GCF). It has two components i.e. GFCF and change in stocks or inventories (Central Statistical Organisation (CSO), 1989 and 2007). The concept and measurement of GFCF and inventories are detailed in CSO (1989 and 2007) and Mallick (2008). However, due co the fluctuating nature of inventories, the investment is measured by the GFCF only ((Mallick, 2008; Khan and Reinhart, 1990; Blejer and Khan 1984; Wai and Wong 1982).

CSO estimates capital formation (GCF and GFCF) at the national level by three approaches, i.e., production (or commodity flew), saving (or flow of funds approach), and expenditure approach. In all the three approaches, change in stocks is added to estimates of GFCF to arrive at GCF. However, the above national level approaches are not directly applicable at the state level due to non-availability of data. For instance, the Regional Accounts Committee (RAC) in 1972 recommended compilation of estimates of GFCF only at state level, rather than compilation of estimates of gross capital formation (GCF), as estimation of change in stocks is not conceptually viable or feasible at the state level because of the open boundaries of the states. The problem is mainly on account of non-availability of data on goods and services transacted across the state boundaries. Report by National Statistical Commission headed by Dr. C. Rangarajan in August 2001 and High Level Committee (Government of India (GOI), 2009) have recommended several measures for improvements of the state level official estimation of public and private GFCF in India.

The public sector GFCF is compiled by most of the states. In fact, the non-availability of state-wise details on capital expenditures made of the two sectors, namely, (i) private corporate sector and (ii) household sector, limits the compilation of estimates of private sector GFCF at the state level. Although data on GFCF in respect of private corporate sector at national level are available from the Studies of Company Finances conducted by the RBI, such details cannot be worked out as companies do not maintain location-wise capital expenditures in their accounts. In the same way, detailed data on GFCF for the unincorporated enterprises or household sector are not available, although data for benchmark years for households are available from the All India Debt and Investment Surveys (AIDIS). In addition, there are problems with respect to AIDIS results, as the investment data from this survey shows considerable underestimates.

Further, the benchmark enterprise surveys conducted by the CSO/NSSO do not give reliable estimates of GFCF at state level, at present due to variety of reasons. For instance, the manner in which disposed assets are valued, conceptually these should be measured at their depreciated values, but presently they are valued on historical costs, leading many times to negative figures of net acquisition of fixed assets. Besides, data collected in the Asset Block of enterprise surveys is based on verbal responses, rather than from the books of accounts, leading to considerable under-reporting of GFCF. The data on capital expenditures from the enterprise surveys are not reliable as well. The focus of these surveys is on collection of data on incomes and expenditures, rather than on capital expenditures. However, Government of India (2009) noted that the latest NSSO survey (62nd Round) gives the requisite data on assets and liabilities and acquisition of fixed assets during the reference period. Though the sampling design used for the surveys are based on sound statistical procedures, but the problem is due to non-sampling errors.

National Statistical Commission (2001) reported that 9 states have estimated the size of total GFCF and 16 states have estimated only for their public sector. EPWRF (2003 and 2009) published availability of the estimates of GFCF at the state level in India. EPWRF (2003) quoted that there is limitation in the estimates of capital formation in DES estimates. First, there is no information on GFCF by private and public sectors for all states. Second, "the details of the actual methods adopted by the states are not known". However, it is generally believed that the states apply the same broad methodology as recommended by the sub-group on "State Gross Domestic Product and Expenditure Account". At the same time, there are some variations in methods adopted depending upon the economic structure at a particular point of time and data availability in the individual states. Most of the states estimate only public sector investment.

The status of un-availability of the estimates of private investment is highlighted in Lakhchaura (2004), EPWRF (2003 and 2009), Mallick (2008) and GOI (2009). Most recently, GOI (2009) reviews the status on the estimation of capital formation at the state level in India. Both public investment and private investment are estimated in AP, ASM, HA, MP, RA, TN and UP at the state level in India. BH, GU, KA, MR, OR and PN estimate the public investment only. GOI provides the estimates of GFCF at the state level by the types of institutions i.e., private, public and household for the year 2004-05 at current prices. Further, Lakhchaura (2004).

Lakhchaura (2004) estimates GFCF by types of institutions from 1993-94 to 1999-2000 at current prices. Rajeswari et. al .(2009) estimates inter-state private GFCF at current prices from 1999-2000 to 2005-06. Mallick (2008) estimates the GFCF in private sector and public sector for the major fifteen states from 1993-94 to 2004-05 at constant prices 1999-2000, by adjusting Bihar, MP and UP with their divided states since 2000-01. Further, Mallick (2010) focuses on the estimation of combined capital expenditure of states and centre at the state level, and re-estimates private investment and public investment, which are found to be more robust in nature. In addition he extended the series on private foreign investment till 2004-05. This study uses data on private GFCF from Mallick (2010) due to the longer length of availability of data on private investment than Lakhchaura (2004) and Rajeswari et. al., (2009). The data length in Lakhchaura (2004) and Rajeswari et. al., (2009) are too short for the empirical analysis. Further, they can not be combined each other as they are based on different methodologies, data sets and assumptions.


There is the high variation in the allocation of private investment across states in India. The variation in the allocation of private investment is measured by the coefficient of variation (CV) of per Capita private investment, which ranges from 46 per cent to 62 per cent during the period 1993-94 to 2004-05. The annual average of per capita private investment and its descending order rankings of states are presented in Table 1. The annual average of private GFCF in MR is Rs. 6210 crore and highest across the states in India, followed by GU, TN and others. ASM ranks lowest in terms of the annual average of private GFCF and has only 16 per cent of MR during this period. Further, TN is most consistent (lowest CV) in terms of inflow of per capita private investment during this period. In contrast, ASM, PN, OR and UP are inconsistent. This indicates that there is high variation of private GFCF across the states in India during the economic reform periods.

States are classified into three groups on the basis of economic growth i.e., high, medium and low growth states in order to explore its relation with patterns of allocation of private investment. Following, Adabar (2005), high growth states are GU, HA, MR and PN, medium growth states are AP, KA, KE, TN and WB and low growth states are ASM, Bihar, MP, OR, RA and UP. The per capita private GFCF for the three groups of states are presented in Table 2. It shows that, the per capita private GFCF is Rs. 2730 and Rs 1901 in the high growth and medium growth states in 1993. In contrast, it is much low (i.e., Rs 999) in low growth states in 1993. There is a positive trend of per capita private GFCF in all the three categories of states. The growth of per capita private GFCF between 1993-94 and 2004-05 is 2.89, 2.57 and 2.87 for the high growth, medium growth and low growth states, respectively.

The foreign investment is the part of total private investment (Mallick, 2010). The CV of foreign private investment during the period 1993-94 to 2004-05 across the major states are presented in Table 3. CV ranges from 86 per cent to 179 per cent from 1993-94 to 2004-05, which is very high. This indicates that there is the high variation of allocation of foreign private investment in the states. The patterns of foreign private investment is analysed by splitting the whole period from 1993-94 to 2004-05 into two sub-periods with equal time length i.e. 1993-94 to 1998-99 and 1999-00 to 2004-05 in Table 4. The results show that the CV of the foreign private investment for TN across the years 1993-94 to 2004-05 is lowest (i.e. 56 per cent) among the major states. It indicates that TN is most consistent in terms of foreign private investment during the period 1993-94 to 2004-05. In contrast, it is 346 per cent, 171 per cent, 151 per cent, 148 per cent and 129 per cent for ASM, PN, BH, OR and RA, respectively. It indicates that these states are highly sinconsistent in terms of foreign private investment dinting the period 1993-94 to 2004-05 among the major states of India.

Further, the results in Table 4 show that the annual average of volume of foreign private investment in 1993-94 to 1998-99 was Rs 1121.7 crore in Maharashtra. In contrast, there was no foreign private investment in ASM. It indicates the degree of scattereness of the foreign private investment across the states of India. Maharashtra ranks first in terms of the annual average of foreign private investment in 1993-94 to 1998-99 followed by KA, TN, GU, AP and others. The annual average of foreign private investment was Rs 4355.8 crore in MR in 1999-00 to 2004-05. MR has maintained to be first rank in terms of foreign private investment in the period 1999-00 to 2004-05 as well. Again, the annual average of foreign private investment is Rs 0.2 crore only in this period. MR ranks first followed by GU, KA, TN, AP and others in terms of foreign private investment in this period. The results show that there is change in the pattern of foreign private investment across the states. For instance, there is no change in rankings of the states of MR, AP, WB, UP, Bihar and ASM from the period 1993-1998 to 1999-2004. There is improvement in the rankings for Gujarat, HA, KE and PN from the period 1993-98 to 1998-2004. However, the ranking decreased for KA, TN and OR. Because of the high growth of the foreign private investment in GU in the period 1993-98, its rank increased from fourth to second, KA and TN came down from second to third, and from third to fourth from the period 1993-98 to 1999-2004. However, the rank correlation coefficient of the annual average foreign private investment between the period 1993-98 and 1999-2004 is 0.76. It indicates that the states which were in the higher ranks during the period 1993-98 remained in the higher ranks during the period 1998-2004.

Pal and Ghosh (2007) noted the implication of the concentration of FDI in a few states in India due to either a big domestic market or cheap and skilled labour. The concentration of FDI in a few states did not help to reduce inequality during the reform period. On the other hand, in a bid to attract FDI, many states have completely overlooked the rural sector and concentrated their development expenditures in the urban areas. This has resulted in increased rural-urban inequality.

In short, there is high variation and inequality in the allocation of both domestic and foreign investment at the state level in India. The inter-state allocation of private investment is associated with the state's GSDP (see, Table 1), the measure of the market size. The large market size involves high demand, which requires high private investment to fill the gap between demand-supply in the economy.

Further, due to economic growth facilitating nature of private investment, the high growth states try to maintain the existing trend of private investment in order to sustain their growth of economy. Nevertheless, the medium and low growth states also try to attract more private investment to achieve higher economic growth. Hence, all the states do compete with each other to attract private investment into respective territories. Particularly, the measures of economic reform such as (a) dismantling of industrial licensing except 14 industries, which are regulated under the ground of environmental hazard, security and social well-being, (b) abolition of Monopolies and Restrictive Trade Practices(MRTP) act, which is introduced for the private sector to seek permission from the central government to extend their units and for establishment of new industries (c) liberalisation of trade by abolishing licensing of imports and reducing tariff, and (d) privatisation and liberalisation of foreign investment regime. As a result, the competition to attract private investment is intensified amongst the states. This also compels the state governments to improve the investment climate conducive to the domestic and foreign enterprises (Dollar, Iarossi and Mengistae, 2002; Ferro, Rosenblatt and Stern, 2004). States and centre provide various types of incentives to attract private investment in India through Special Economic Zone (SEZ), Export Oriented Units (EOUs), Software Technology Park (STP), Industrial Growth Centres, Electronic Hardware Technology Parks (EHTP) and Free Trade & Warehousing Zones (FTWZs) etc. The incentives are in the areas of labour laws, environmental protection, taxation and administrative approvals and available in incentive packages as well as in special deals.

Promotional policies and incentives vary across sectors, industries and regions over the time. Due to the un-availability of data, we used incentive indices in the year 1992-93 to the foreign private investors (Venkatesan and Varma, 1998), represented in Table 5. It shows that, the incentive index for TN, WB, Maharashtra, Gujarat and Karnataka was higher than that of other states in India. The foreign private investment is the part of total private investment. The state-wise FDI approval for the year 1992-93 which is sourced from SIA (1999), compared with incentive indices using rank correlation coefficient. The correlation coefficient is found to be 0.39. It indicates that the foreign private investment is positively associated with incentive indices. The incentive index may not vary so much in the consecutive years. Hence, the state-wise incentive indices are compared with the inter-state private GFCF in 1993-94. The rank correlation coefficient is 0.66. Further, the rank correlation coefficients of the incentive index for the year 1992-93 with the annual average of private GFCF and annual average of foreign private GFCF for the years 1993-94 to 2004-05 are found to be 0.67 and 0.4, respectively. It suggests that the promotional policies and incentives do affect the inter-state allocation of private investment.


Following Carlberg (1981), inflow of capital into a region depends on the factor price of capital directly. The diminishing marginal product of capital equates the factor price in the long run. According to him, inter-regional growth is characterised by the free trade, labour and capital movements. Products are shipped to the region paying the best price. Capital and labour allocated through market mechanism is efficient as well. That means the allocation of capital and labour are determined by the rate of return and wage rate, respectively. The model is based on the following assumptions (time subscript "t" is dropped to simplify the notation.

(1) All regions producing a homogenous output ([Y.sub.i]) using two inputs (capital and labour)

(2) A distinct neo-classical technology in each region and spatial diffusion may spread technological knowledge.

(3) Full-employment of resources under perfect competition in product and factor markets.

Now, the regional production function, [Y.sub.i] = [F.sub.i]([K.sub.i], [L.sub.i]) (1)

Where, [Y.sub.i] = Aggregate output for the "i" region.

[K.sub.i] = is stock of capital "i" region.

[L.sub.i] = is labour inputs in "i" region.

[F.sub.i] = is neoclassical technology

Market allocation: Let the nation's endowment with capital K and labour L are given. For the simplification, consider there are two regions in an economy.

Then K = [2.summation over (i=1)] [K.sub.i] and L = [2.summation over (i=1)] [L.sub.i]

The profit function under perfect competition and full-employment of factor of production for the region "i" is as follow.

[Q.sub.i] = [Y.sub.i] - [K.sub.i] [r-L.sub.i] w (2)

The first order condition for profit maximization:

[partial derivative][Y.sub.i] / [partial derivative][K.sub.i] = r and [partial derivative][Y.sub.i] / [partial derivative][L.sub.i] = w (3)

Where, [partial derivative][Y.sub.i] / [partial derivative][K.sub.i]: the marginal product of capital in the region 'i' and :the marginal product of labour in region 'i'.

The second order condition is satisfied since neoclassical technology exhibits diminishing return to substitution. The necessary condition implies that factor prices correspond to marginal products.

Optimal allocation: National output in a two regions economy is

Y = [[summation].sup.2.sub.i=1] [Y.sub.i] ([K.sub.i], [L.sub.i]) (4)

Let. [mu] and [lambda], two lagranzian multiplier such that 0 [less than or equal to] [mu] [less than or equal to] 1 and 0 [less than or equal to] [lambda] < 1. The objective function is to maximize,


From the first order conditions, we get

[partial derivative][F.sub.1] / [partial derivative][K.sub.1] = [partial derivative][F.sub.2] / [partial derivative][K.sub.2] = [lambda] = r (5)

The equation (5) indicates that the market determined rate of return on capital across the regions maximizes the national optimal output (income) in the long run. In other words, the market allocation of capital across the regions maximises the national total income. Hence, the allocation is efficient. Further, the equation (5) indicates that in the long run allocation of capital equates regions' marginal product with the factor price or higher the factor prices in region, higher will be the inflow of capital.

[K.sub.i,t] = f([]) (6)

The above equation can be expanded to include several other factors, following Kiabel (2003), Oshikoya (1994), Greene and Villanueva (1991), Aggrawal, (2005) and Rao et al. (1999) [see, Appendix Table 1]. Wilson (1999) assumes that in an economy using two factors i.e. labour and capita for production, while labour is immobile and capital flows from one region to another, the local governments compete to attract capital into their territories through tax competition and provision of public goods. The provision of public goods to increase the productivity and hence profit of the capitalist is financed through imposition of tax. The provision of public good is a positive factor, while imposition of tax is a negative factor for the flow of capital stock. Overall, the flow of capital stock depends on after tax rate of return on capital. The net rate of return will be equalised due to the diminishing marginal product of capital across the regions in the long run. The firms invest to the point where marginal product of capital equates with after tax rate of return. In turn, the gross rate of return on investment depends on the taxation and provision of public goods.

The imposition of taxes affects adversely and tax cut or concessions promote private investment. Kiabel (2003) argues that the using of taxation (i.e. tax incentives and tax cut) as an instrument is not an efficient strategy to attract private investors. The tax cut and tax incentives may send a wrong signal to the private investors and makes suspicious of the real intentions of the host region. The development of infrastructure may be a better instrument to encourage private investors by enhancing the productivity and hence rate of return on capital.

The corruption and stability of the government may explain the flow of private investment (Kiabel, 2003). The corruption and stability of government as the components of governance are the crucial factor in determining the inflow of private investment. For instance, an investor intending to go into mining, he/she has to obtain a mining license and to take the electricity connections and so on. Because of corruption, the cost to start the business increases. The investors want to operate in a peaceful environment, where they are free to make huge amount of investment targeting to operate for a long period of time. Kiabel (2003) argued that the private investors may not prefer to operate in the countries such as Angola, Somalia, the Democratic Republic of Cango, Sierra-Leone, Algeria, Nigeria, Yogoslovia, Sri-Lanka, which are potentially unstable as a result of civil conflicts and religious and ethnic clashes etc. Further, Kiabel (2003) viewed that the financial stability, the indicator of debt burden also in one component of governance. The increase in public debt indicates the increase in tax rate in future, and reduces the rate of return on private investment. Hence, the public debt is expected to influence negatively to the private investors.

Economic uncertainty may explain the inflow of private investment (Oshikoya, 1994) and (Greene and Villanueva, 1991). Greene and Villanueva (1991) argue that inflation increases the riskiness of long-term investment projects, reduce the average maturity of commercial lending and distort the information content of relative prices. The high fluctuation of prices affects the economy adversely. Whereas, a moderately rising price level serves as a stimulant for economic activities and by keeping the level of economic activity high, it creates suitable condition for the private investors. Though rising prices increase profits and induce the entrepreneur to invest more, but a sudden rise in prices is un-desirable for the investment. A sudden increase in prices leads to decline in demand for output, and hence investors incur loss. So a rapid change in price level creates uncertainty and discourages investment.

Market size is expected to have a positive influence on the allocation of private investment. The larger the market size of the host economy the greater is the possibility of reaping the advantages of the scale economies and reduces cost of production and hence increases the rate of return (Aggrawal, 2005). Further, labour cost is expected to have adverse impact on the flow of private investment. The higher is the labour cost, the higher will be the cost of production. Hence, other things remaining same, it reduces the rate of return on investment. The cost of labour is a factor of cost competitiveness. In turn, the higher is the labour productivity, the lower is the cost of labour and hence, maximises the rate of return on investment. Finally, the availability of finance affects the allocation of private investment positively by increasing the rate of return on investment by reducing the rate of interest. The availability of finance is the main source of private investment.


The panel data model is used for our analysis to control for individual heterogeneity of the states, more degree of freedom and more efficiency (Baltagi, 2001). In the case of panel data model, the error term [] is a composite residual consisting of time invariant state specific components [[mu].sub.i], and captures various characteristics of the state, which are not observable but have a significant impact on inter-regional allocation of private investment, and a disturbance term [[epsilon]], which is assumed to satisfy the Classical Linear Regression (CLRM) model assumptions. The estimable panel data model is specified as:

[] = [alpha] + [beta] [] + [[mu].sub.i] + [[epsilon]] (7)

i = 1 ............... 15 and

t = 1993-94, 1994-95, .............................., 2004-05.

[] is per capita private investment of states, [] is the vector of explanatory variables. [alpha] and [beta] are the parameters of the model.

There are three types of panel model. These are (a) pooled regression model (PRM), (b) fixed effects model (FEM), and random effects model (REM). Diagnostic test such as Breush and Pagan Lagrange Multiplier (LM) Test and the Hausman (H) Specification Test are used to choose between panel data models. LM test is used to test the null hypothesis of non-random individual effect. A high value of LM favours fixed effect model \ random effect model over pooled regression model. Hausman specification test is used to test null hypothesis of zero correlation between state specific effects and the explanatory variables. The significance of langrange multiplier (LM) test statistics indicates that the model estimated by using REM or FEM give better estimates than PRM. Further, the statistical significance of Hausman (h-test) specification test suggests preferring estimation by using FEM to REM. The standard statistical frameworks for estimation of these models are well known (Greene, 2006; Baltagi, 2001).


The determinants of allocation of private investment are estimated for major 15 states from the period 1993-94 to 2004-05 by using basic equation (7). The dependent variable is the estimates of private investment at constant prices (1999-00). Labour cost can be measured in two ways. First, labour cost per output (value added) and second, wage rate i.e., the labour compensation per unit of labour. The labour cost may be less due to the higher productivity of labour. Hence this study uses the average productivity of labour to analyse the determinants if private investment across the states. The labour productivity in the registered manufacturing sector of states is used as proxy to measure the labour productivity across states, due to the paucity of data.

States share in all states' public debt, the number of crime rates, the number of riots, and the number of man days lost due to strike are the indicators to capture the impact of stability and corruption on inter state allocation of private investment. Due to the unavailability of suitable measure, crime rate can be used to capture political and bureaucratic corruptions (Baddeley et al., 2006). The crime rate indicates the law and order situation in a state, which is measured by the number of crimes to 10,000 population. Further, the number of riots indicates the stability of state. The instability due to civil conflict, religious and ethnic clashes negatively influence the inflows of private investment. In addition, the man days lost is used to assess the industrial relation in the states. Aggarwal (2005) argues that the sound industrial relation stabilizes the economy. The high industrial relations in terms of low number of grievances, low strike activity, and so on promoted the productivity and hence investment.

The state governance variables are the financial status of states (public debt), crime rates, Riots and man days lost. The principal component analysis is used to construct one variable by using four indicators. Generally, the inflation is measured by the rate of change in wholesale price index (WPI). We use growth of GSDP deflator as proxy for the inflation rate, because the WPI is not constructed at the state level.

The measurement of infrastructure by public and private is a difficult task because of unavailability of data. Hence, the state wise total infrastructure is used in our study. The infrastructure has several components, such as; the Electricity, Road ways, Rail ways, Air ways, Telephone facilities and Banks. These are used to construct physical infrastructure index by applying the method of principal component analysis: The state wise electricity facility is measured by the per capita power consumption [Rao,2004: Kurian, 2000; Rolee, 1991; Shah and Patel, 2006]. The road facility can be measured by road density per 1000 square kilometres [Rao, 2004; Rolee, 1991; Shah and Patel, 2006]. Similarly, the status of rail facility is measured by the route in length in kilo meters per 1000 square kilo meters [Rao, 2004; Rolee, 1991; Shah and Patel, 2006]. Banking infrastructure is measured as the number of bank offices per lakh population (Rolee, 1991). The telephone facility is measured the telecom lines (direct exchange lines) per 1000 persons in (Kurian, 2000). Air ways facility is measured by the number of passenger handled per 1000 population.

The credit extended by the scheduled commercial banks across states is used as a proxy for availability of finance. The current values are deflated by the GSDP price. The states share of scheduled commercial banks credit extended in all states credit is incorporated into the model.

The measurement of data sources and variable need in the estimation are presented in Appendix Table 2. The descriptive statistics and simple correlation coefficients of the variables included in the analysis are presented in Appendix Tables 3 and 4. The correlation coefficients among physical infrastructure, per capita income and availability of finance are found to be very high. This suggests that the economic development of a state, measured by per capita income is highly associated with the development of physical infrastructure and the availability of finance. Hence, it may create muiticollinearity problem in the estimations, which makes one or more coefficients statistically insignificant (Gujurati, 1995). Further, the correlation coefficient between availability of finance and labour productivity is found to be 0.52. The inclusion of both the variables into one regression equation may create the multicollinearity problem in the estimations.

In the light of the existing theoretical and empirical literature, different combinations of the variables are estimated. Table 6 presents the estimated results for 5 models. Model 1 includes all the 6 independent variables. Model 2 drops the per capita income as an independent variable. Model 3 drops the physical infrastructure and per capita income as the independent variables. Model 4 drops the availability of finance and infrastructure as the independent variables. Model 5 includes the physical infrastructure, inflation, state governance and labour productivity.

The value of LM and h-statistics are found to be significant in all models. This indicates that the state specific factors are correlated with the explanatory variables. Hence, Regression Model with single constant term is inappropriate on empirical ground. We considered the Fixed Effect Models as the better choice for the interpretation. Table 6 contains the results of fixed effect model for estimation of inter-state allocation of private investment in India for the period 1993-94 to 2004-05. The F statistics is significant in all models, which indicate the overall significance of the regression models.

The results of Model 1 show that regression coefficients of infrastructure, governance, market size and availability of finance are statistically significant. But the coefficient of physical infrastructure is negative. It could be due to high correlation of market size with the physical infrastructure and availability of finance to the private investors. Hence, the physical infrastructure is dropped as the independent variable and estimated in Model 2. The result shows that the governance, market size and the availability of finance are statistically significant in explaining the inter-state allocation of private investment in India. The market size is found to be statistically significant in the model.

The results of Model 3 show that the state governance, availability of finance and inflation are statistically significant in explaining the inter-state variation in private investment of major states in India during the period 1993-94 to 2004-05. The statistical insignificance of labour productivity is due to the high correlation between the availability of finance and the labour productivity variables. This variable is dropped as the independent variable in Model 4.

The Model 4 includes the market size, quality of state governance, economic uncertainty of states and the labour productivity as the independent variables. Except inflation, all other variables are found to be statistically significant. Further, the market size is replaced by the physical infrastructure along with quality of governance, inflation and labour productivity in Model 5. The results show that the all explanatory variables are statistically significant. Hence, the physical infrastructure, quality of governance, labor productivity and the inflation explain the inter-state allocation of private investment in India during the period 1993-94 to 2004-05.

In short, the market size or the per capita income, the physical infrastructure, the labour productivity, the availability of finance and economic uncertainty explain the inter-state variation in the allocation of private investment of the major states in India for the period 1993-94 to 2004-05. The physical infrastructure and labour productivity influence the private investor positively by increasing the profit and hence the rate of return on their investment by reducing the cost. The per capita income, the measurement of development affects the private investment positively. Bhattacharya et al. (2004) argued that the income, infrastructure and poor governance are the crucial factor in the inter-state variation in private investment and hence inter-state growth disparity. Further, the statistically significance of the state specific constant, shows that the state specific factors have important role in determination of inter-state allocation of private investment. Through this coefficient, we are capturing the state specific factors such as state wise availability of natural resources and other variables.


There is the high variation in the allocation of total private investment (and private foreign investment) during 1993-94 to 2004-05 across 15 major states in India. The high and medium growth states, such as, MR, GU, KA and TN attracted higher volume of private investment during this period. The investment incentives are positively associated with the allocation of private investment. The quality of governance, the physical infrastructure, the per capita income, the economic uncertainty and labour productivity are statistically significant to explain the interstate allocation of private investment. The un-availability of data restricted to use investment incentive as an explicit variable in our fmodeling. Nevertheless, the statistical significance of state specific coefficient justifies the importance of incentives in the inter-state allocation of private investment in India. Hence, in order to maintain balanced regional development through the allocation of private investment, the role and importance of the above factors should be looked by policy makers. The findings of this study corroborates with Bhattacharya et al. (2004) and Ahluwalia (2000). Ahluwalia (2000) argued that the decontrol of investment licensing is the cause of inequality in the allocation of private investment in the post reform periods. Hence, private investment is potentially highly mobile and flows to states which have a skilled labour force with a good 'work culture', good infrastructure especially power, transport and communications, and good governance generally" (Ahluwalia, 2000).



Adabar, K (2005), "The Regional Dimension of Economic Growth in Indian Federalism", Doctoral thesis submitted to University of Mysore, Karnataka.

Aggarwal, A. (2005), "The Influence of Labour Markets on FDI: Some Empirical Explorations in Export Oriented and Domestic Market Seeking FDI across Indian States", Department of Business Economics, University of Delhi, South Campus, available at online / india&globalization_A_Agarwal.pdf.

Ahluwalia, M. S. (2002), "Economic Performance of States in Post Reforms Period", Economic and Political Weekly, May 6, 1637-1648.

Annual Survey of Industries (Various years), Volume-I, Ministry of Statistics and Programme Implementation, Central Statistical Organisation, Kolkata.

Baddeley, M., K. McNay and R. Cassen, (2006), "Divergence in India: Income differential at the state level, 1970-97", Journal of Development Studies, Vol. 42, No. 6, 1000-1022.

Bhattacharya, B. B. and S. Sakhivel (2004), "Regional Growth and Disparity in India: Comparision of Pre-and Post-Reform Decades", Economic and Political Weekly, Vol. 39, No. 10, 1071-1077.

Baltagi, B. H. (2001), Econometric Analysis of Panel Data, West Sussex PO191UD, John Wiley & Sons, Ltd., England.

Carlberg, M. (1981), "A Neoclassical model of interregional economic growth" Regional Science and Urban Economics, Vol. 11, No. 2, 191-203.

CMIE. (Various years). Infrastructure, Centre for Monitoring Indian Economy Pvt Ltd, Mumbai.

--(Various years) Money and Banking, Centre for Monitoring Indian Economy Pvt Ltd, Mumbai.

CSO (1989, 2007), "National Accounts Statistics: Sources and Methods", Ministry of Statistics and Programme Implementation, Government of India.

Dollar, D., G. Iarossi and T. Mengistae (2002), "Investment Climate and Economic Performance: Some Firm Level Evidence from India", Working Paper No. 143, 'Center for Research on Economic Development and Policy Reform, Stanford University.

EPWRF[Economic and Political Weekly Research Foundation] (2003), State Domestic Products: 1960-61 to 2000-01, Mumbai, India.

Erumban, A. A. (2009), "Productivity and Unit Labour Cost in Indian Manufacturing: A comparative Perspective", Economic & Political Weekly, Vol. 44, No. 15, 39-48.

Ferro, M, D. Rosenblatt and N. Stern (2004), "Policies for Pro-Poor Growth in India, in India's Emerging Economy. ed. Kaushik Basu, page. 153-183. New Delhi: Oxford University Press.

Government of India (2009), "Report of the High Level Committee on Estimation of Saving and Investment, pp. 223-232. New Delhi.

Greene, W. H. (2006), Econometric Analysis, Dorling Kindersley, New Delhi.

Greene, J. and D. Villanueva (1991), "Private Investment in Developing Countries: An Empirical Analysis", IMF Staff Papers, Vol. 38, No. 1, 33-58.

Gujarati, D. N. (1995), Basic Econometrics, (3rd edition), McGRAW HILL, International editions, Singapore.

Kiabel, B. D., (2003), "Corporate Tax Incentives in Nigeria, American Economic Review", Vol. 5, No. 1, 60-70.

Krishna. K. L. (2004), "Patterns and Determinants of Economic Growth in Indian States", Working Paper. 144, ICRIER, New Delhi.

Kurian, N. J. (2000), Widening Regional Disparities in India; Some Indicators, Economic and Political Weekly, 538-50.

Lakhchaura, Prakash (2004), Capital Formation at State Level. Indian, Ministry of Statistics and Programme Implementation mimeo.

Mallick, Jagannath (2010), Private Investment and Economic Growth in India: An Empirical Analysis, Doctoral Thesis submitted to University of Mysore, Karnataka.

--(2009), "Trends and Patterns of Private Investment in India", Working paper, No. 226, ISEC, Bangalore, India.

--(2008), "Private Investment in India: Data and Measurement Issues at the State Level", The Journal of Income and Wealth, Vol. 30, No. 2, 44-63.

National Account Statistics (various issues), Central Statistical Organization, Government of India.

Oshikoya, T. W. (1994), "Macroeconomic Determinants of Domestic Private Investment in Africa:

An Empirical Analysis", Economic Development and Cultural Change, Vol. 42, No.3, 573-596.

Pal, P. and J. Ghosh (2007), "Inequality in India: A Survey of Recent Trends" Working Paper Number 45, Department Social and Economic Affairs, United Nations, New York.

Rao, A. S. (2004), "Inter State Disparities in the Development of Tertiary Sector in India", Indian Journal of Regional Science, Vol. 36, No. 2, 93-105.

Rao, M. G., R. T. Shand and K. P. Kalirajan (1999), "Convergence of Incomes Across Indian States: A Divergent View", Economic and Political Weekly, Vol. 34, No. 13, 769-778.

Rolee, K. (1991), "Regional Development and Planning in Madhya Pradesh", Indian Journal of Regional Science, Vol. 23, No. 1, 96-105.

Shah, F. and N. K. Patel (2006) "Economic Infrastructure and Socio-Economic Development in India", Indian Journal of Regional Science, Vol. 38, No. 10, 27-35.

SIA (1999), Newsletter, Secretariat for Industrial Assistance, Department of Industrial Policy & Promotion, Ministry of Commerce & Industry, Government of India, New Delhi, India.

Venkatesan, V. and S. Varma (1998), Policy Competition among the States in India for Attracting Direct Investment, NCAER, New Delhi.

Wilson, J. D. (1999), "Theories of Tax Competition", National Tax Journal, Vol. 52, 269-304.

JAGANNATH MALLICK, Indian Council for Research on International Economic Relations (ICRIER), Lodhi Road, New Delhi, India.

([dagger]) This paper forms part of the author's Doctoral Thesis. The author is grateful to his Doctoral thesis supervisor Prof. M.R Narayana for his valuable comments. The author is also benefited from the discussion with Mr. Ramesh Kolli. However, usual disclaimer applies. The author can be contacted at or
Appendix Table 1
Determinants of Private Investment

Sl. No.   Studies             Determinants

1         Carlberg (1981)     The factor price i.e. the rate of
                              return on capital determines the
                              inter-regional allocation of capital.

2         Kiabel (2003)       The competition through the provision
                              of infrastructure is more effective
                              than using tax competition to attract
                              private investors. Corruption,
                              political and financial stability of
                              the governments are crucial in the
                              inflow of private investment.

3         Oshikoya (1994)     Economic uncertainty adversely
                              influences inflow of private

4         Greene and          Economic uncertainty adversely
          Villanueva (1991)   influences inflow of private

5         Aggrawal (2005)     Market size expected to have positive
                              impact on the inflow of private
                              investment. Further labour cost
                              negatively influences negatively the
                              inflow of private investment.

6         Rao et al. (1999)   Infrastructure is crucial in allocation
                              of private investment across states.

Appendix Table 2
Variable Description And Data Sources

Variables               Measurement             Data sources

private investment      This is at constant     Mallick (2010)
                        prices 1999-00.

Labour productivity     The ratio of gross      ASI
                        value added to total
                        labour engaged in the
                        industries. Consumer
                        price for industrial
                        worker is used to
                        convert at current
                        prices to constant
                        prices (1993-94).

Market size             It is measured by the   CSO
                        state-wise share in
                        all states GSDP. This
Inflation               is expressed in terms   CSO
                        of percentage. Growth
                        of GSDP deflator

State Governance        This variable is        The public debt and
                        constructed by          crime rate are taken
                        including public        from RBI and
                        debt, the number of,
                        crime to 10,000         respectively. The
                        population, riots per   riots and man days
                        lakh population and     lost are taken from
                        the number of man       Statistical Abstract
                        days lost per 1000      of India (SAI) of
                        population. The         CSO.
                        principal component
                        analysis method is
                        used to construct
                        this variable.

Physical                This is constructed     The number of bank
Infrastructure          by including number     offices and power
                        of bank offices per     consumption are taken
                        lakh population, per    from CMIE report on
                        capita power            "Money & Banking" and
                        consumption, the rail   "Energy",
                        ways density per 1000   respectively. The
                        square kilo meters,     others are taken from
                        road ways density per   report on
                        square kilo meters,     "Infrastructure"
                        the number of           (CMIE).
                        passenger handled per
                        1000 population and
                        telephone lines per
                        1000 population. The
                        principal component
                        analysis method is
                        used to construct
                        this variable.

Availability of         Outstanding Credit of   Money & Banking, CMIE
Finance                 Scheduled Commercial

Other fiscal            Two alternative         Gross fiscal deficit
variables               variables i.e. share    and sales tax are
                        of gross fiscal         taken from report on
                        deficit in GSDP and     "State finances: A
                        sales tax rate are      budget of study", RBI
                        tried to capture the
                        impact of fiscal
                        policy on the inter-
                        state allocation of
                        private investment.

Appendix Table 3

Descriptive Statistics

Variables                   Mean   deviation    Minimum      Maximum

Private GFCF            19557.49    16585.13     810.98    108801.17
Infrastructure           0.00004        0.62      -1.05         2.01
Governance               0.00002        1.00      -1.73         3.15
Market Size             18190.86     5741.13    9101.40     33327.59
Availability of          3038.69     2509.40     221.92     15561.24
  finance (AVF)
Labour productivity    270111.36   460017.67   10373.01   3960854.68
Inflation                   6.08        3.60      -6.13        16.35

Appendix Table 4
Simple Correlation Coefficients

Variables         Private
                     GFCF   Infrastructure   Governance   Income

Private GFCF         1.00
Infrastructure       0.24             1.00
Governance          -0.14             0.18         1.00
Market size          0.33             0.74        -0.13     1.00
AVF                  0.68             0.70        -0.08     0.76
LP                   0.51             0.19        -0.06     0.26
Inflation           -0.28            -0.10         0.06    -0.23

                    AVF      Lp   Inflation

Private GFCF
Market size
AVF                1.00
LP                 0.52    1.00
Inflation         -0.24   -0.17        1.00

Table 1
Inter-state Variation In Per Capita Private Investment

States     Annual average of   Rank   CV (%)   Correlation coefficient
          per capita private                           of private GFCF
                GFCF (in Rs)                          with GSDP during
                                                    1993-94 to 2004-05

AP                   2374.78     10    27.12                      0.94
ASM                  1000.11     15    69.65                      0.92
BH                   1158.04     13    25.68                       0.9
GU                   5617.57      2    35.55                      0.94
HA                   4485.56      4    27.72                         1
KA                   3637.09      5    33.82                      0.98
KE                   3337.13      7    30.91                      0.99
MP                    2562.4      9    22.35                      0.97
MR                   6210.19      1    28.64                      0.96
OR                    1107.9     14    58.62                       0.9
PN                   3508.42      6    46.06                      0.86
RA                   2278.81     11    34.63                      0.94
TN                   4765.96      3    19.24                      0.96
UP                   2128.27     12    43.96                      0.92
WB                   2866.29      8    29.34                      0.97

Source: Author's calculation

Table 2
Patterns of Private GFCF by Income Groups of States

Years    1993   1994   1995   1996   1997   1998   1999   2000   2001

HGS      2730   3102   4090   4315   4492   4493   5277   4603   5052
MGS      1901   2162   2639   2765   3080   3218   3868   3599   3781
LGS       999   1029   1387   1433   1548   1714   1867   2174   2280

Years    2002   2003   2004

HGS      5555   6107   7902
MGS      4143   4195   4893
LGS      2293   2739   2870

Note: HGS; High growth states, MGS: Medium growth states and LGS; Low
growth states

Source: Author's calculation.

Table 3
Variation of Foreign Private GFCF of States

Years    1993   1994   1995   1996   1997   1998   1999   2000   2001

CV        160    135    103     86     94    122    152    179    166

Years    2002   2003   2004

CV        133    130    133

Source: Authors calculation based on the data from Mallick (2010).

Table 4
Patterns of Foreign Private Investment of States

States   CV (in %)         Foreign private         Foreign private
                     investment in 1993-98     investment in 1999-
                            (in Rs. crore)     2004 (in Rs. crore)

AP               81              402.7 (5)              1162.2 (5)
ASM             346               0.0 (15)                0.2 (15)
BH              151              22.0 (14)               31.3 (14)
GU              118              490.5 (4)              2317.7 (2)
HA              108             112.3 (10)               324.3 (7)
KA               81              703.2 (2)              2136.2 (3)
KE              124              28.7 (13)              216.8 (10)
MP              118              399.7 (6)              159.3 (12)
MR               69             1121.7 (l)              4355.8 (l)
OR              148              363.2 (7)               49.7 (13)
PN              171              90.5 (12)               512.7 (6)
RA              129             109.0 (11)              161.8 (11)
TN               56              618.7 (3)              1427.5 (4)
UP               77              143.7 (9)               275.8 (9)
WB               65              303.8 (8)               297.5 (8)

Note: (1) The figures in the 3rd and 4th columns in the above table
are in annual average.

(2) The figures in parenthesis are the descending order rankings of

Source: Authors calculation based on the data from Mallick (2010).

Table 5
Incentive to the Private Investors in 1992-93

States   AP   BH   GU   HA   KA   KE   MP    MR   OR   PN   RA    TN

Index     53   67   93   53   93   80   40   100   80   13   93   129

States   UP    WB

Index     93   107

Source: Venkatesan and Varma (1998)

Table 6
Determinants of Private Investment at the State Level

Independent variables               Dependent Variable: Real
                                         private GFCF

                                Model 1       Model 2      Model 3

Governance                   -4697.68 *       -5290 *    -5299.8 *
                              (1476.47)      (1548.8)    (1578.55)

Infrastructure              -8371.85 **

Market size                       1.2 *        0.73 @
                                 (0.44)        (0.47)

Availability of finance          3.47 *       3.14 **       4.13 *
                                 (1.34)        (1.41)       (0.81)

Inflation                       -131.27       -154.29    -301.7 **
                               (136.05)       (138.5)     (137.07)

Labour Productivity               0.007         0.009        0.006
                                (0.019)        (0.21)       (0.18)

LM                              353.7 *      372.84 *     474.67 *

H                             11.65 ***       12.5 **        6.9 @

Adj [R.sup.2]                      0.88          0.87         0.87

F model test                    69.27 *       69.13 *      71.15 *

F test for [U.sub.1] = 0        30.75 *       31.87 *         38.46

Independent variables           Dependent Variable:
                                 Real private GFCF

                                Model 4      Model 5

Governance                      -5954 *   -7485.94 *
                                 (16.01    (2022.12)

Infrastructure                             9868.25 *

Market size                      2.05 *

Availability of finance

Inflation                        -69.82    -615.69 *
                                (145.6)     (152.88)

Labour Productivity            0.033 **       0.05 *
                                 (0.16)       (0.18)

LM                             405.91 *     380.45 *

H                               21.15 *      13.18 *

Adj [R.sup.2]                      0.86         0.81

F model test                     61.8 *      44.54 *

F test for [U.sub.1] = 0        49.43 *      34.73 *

Notes: (1) *, **, *** and @ indicate statistical significance at
1%, 5%, 10% and 12 %, respectively.

(2) The figures in parenthesis are standard errors of estimates.

Source: Estimated using equation (7).
Gale Copyright:
Copyright 2011 Gale, Cengage Learning. All rights reserved.