The effects of foreign private investment on economic growth in Pakistan.

Shabbir, Tayyeb ; Mahmood, Azhar

1. INTRODUCTION

Studying the impact of foreign financial flows such as foreign private investment and aid on economic growth or domestic savings of a country has been a topic of considerable academic as well as practical interest. In fact, considering the expected decline during the 1990s in the supply of commercial flows particularly to the developing countries, the question of the impact of foreign capital on economic growth and other important macro variables such as saving and investment assumes a renewed significance.

The various forms of foreign capital inflows are being increasingly viewed as a potentially significant source of capital to supplement domestic investment. According to the latest World Development Report, in 1988, Direct Foreign Investment surpassed all other forms of lending as a source of foreign capital to developing countries (1) [World Bank (1991), p. 95].

In theoretical terms, there has been no consensus on the role of foreign capital in promoting economic growth in developing countries. [See Mosley (1980) for an historical review of the important literature on this issue.] In short, whereas early writers such as Chenery and Strout (1966) were able to show that foreign capital exerted a favourable effect on growth, their results may be sensitive to the assumption that foreign inflows fully augment resources available for capital formation and that the incremental capital-output ratio is independent of such inflows.

However, later work by the so-called "Displacement Theorists" questioned both these assumptions. [For instance, see Griffin (1970); Leff (1969) and Weisskopf (1972).] Foreign capital may in fact substitute for domestic saving on two accounts: (a) foreign capital inflows could lead to government becoming less enthusiastic about its revenue generation efforts, the nation could increase its consumption expenditure and/or liberalise imports and (b) to the extent that saving is determined by available investment opportunities, by crowding out domestic investment, foreign private investment could cause domestic savings to fall. Besides, foreign capital may lead to inefficiencies due to inappropriate technology and management styles. Thus, it is conceivable that foreign capital could adversely affect economic growth.

In terms of the existing empirical evidence, single equation estimates of the saving equation have found a negative relationship between foreign capital and domestic saving [Fry (1984); Griffin and Enos (1970) and Weisskopf (1972)].

Similarly, single equation estimates of foreign capital's effect on growth rate report favourable impact [Dowling and Hiemenz (1983); Stoneman (1975) and Papanek (1973)].

Obviously, the above results become suspect if one believes that saving and growth are embedded in a simultaneous system. There have been few studies that specify saving and growth in the context of a system of simultaneous equations [one exception is Gupta (1975)]. However, there are no such studies for Pakistan. Additionally, in terms of the literature on foreign private investment and/or other foreign capital inflows, very rarely has the analysis been conducted so as to concentrate on a single country; most studies tend to pool data across countries and time to lessen the degrees of freedom constraint. However, this procedure is valid only if certain restrictive assumptions on temporal and/or spatial constancy of model structure are made. It is perhaps most meaningful to bite the bullet and concentrate on time series analysis of each country individually.

Taking our own advice, in this study, we specify and estimate a two-equation simultaneous model of rate of real economic growth and saving ratio using annual time series data on Pakistan for 1959-60 to 1987-88. Here we plan to explore the answers to the following two propositions:

Proposition I = Net foreign private investment (FPI) promotes economic growth as measured by annual rate of growth of real GNP.

Proposition II = FPI displaces savings of a country as measured by its National saving to GNP ratio.

The rest of the paper is organised as follows:

Section 2 presents the specification for a simultaneous equation model of economic growth and saving ratio. Section 3 describes the data while the next section reports and discusses the empirical results. Finally, Section 5 contains some concluding remarks and notes some policy implications of our results.

2. MODEL SPECIFICATION

Consider the following simultaneous equation model of the rate of growth of real GNP (GR) and the ratio of National Saving to GNP i.e. so (Note: The time subscript is being omitted for convenience).

[MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII] ... (1)

[MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII] ... (2)

The above model specifies the relationships (2) amongst two endogenous variables, GR and S, and four exogenous variables: DISB, FPI, XM and RINT. (These variables have been defined in the next section). The a priori expected signs of the parameters are noted below each of the equations. In light of the controversy amongst economists, the signs of FPI and DISB coefficients cannot be determined a priori and thus are an empirical question. The XM variable in the GR equation is posited to have a positive effect since it is often maintained that growth in exports (or generally, 'openness of economy') can stimulate economic growth [for instance, see Balassa (1978) and Tyler (1981)]. Again, the RINT or the real rate of interest variable is the standard variable used in the specifications of the saving equation to represent the 'reward for saving'.

3. DATA DESCRIPTION

In this paper we have used time series data for the period 1959-60 to 1987-88. The names of relevant variables and their definitions are given below where the respective source of data is being noted in the parentheses.

Variable's Name Definition (Source)

GR Annual growth rate of Gross National Product (GNP) at constant
 market prices of 1959-60. (Source: Pakistan Economic Survey
 1989-90).
S National Savings as a ratio of GNP. (Source: Pakistan Economic
 Survey 1988-89, PIDE's Statistical Paper Series No.4, and
 Annual Report, State Bank of Pakistan 1987-88).
FPI Net Foreign Private Investment into Pakistan as a ratio of GNP
 where the former is the sum of cash brought in, capital
 equipment brought in and reinvested earnings. (Source: Foreign
 liabilities and Assets and Foreign Investment in Pakistan, (3)
 State Bank of Pakistan, Statistic Department, Various issues).
DISB Total disbursements including Grants during a year as a ratio
 of GNP. (Source: Pakistan Economic Survey, 1989-90).
XM Exports of goods and services as a ratio of GNP. (Source:
 Pakistan Economic Survey, 1989-90).
RINT Real rate of interest, computed by subtracting growth rate
 of consumer price index from the nominal interest rate on
 twelve month time deposits. (Source: Pakistan Economic Surve,
 1988-89, the PIDE Econometric Model of Pakistan's Economy
 (1959-60 to 1978-79), Annual Report, State Bank of Pakistan,
 1987-88.)
GPC GNP per capita. (Source: Pakistan Economic Survey, 1989-90).
REM Workers' Remittances as a ratio of GNP. (Source: Pakistan
 Economic Survey, 1989-90).

4. EMPIRICAL RESULTS/DISCUSSION

The simultaneous equation model given by Equations (1) and (2) is exactly identified (4) and is estimated by 2SLS. These estimates are presented in Table 1 whose columns 2 and 4 contain estimates that correspond exactly to Equations (1) and (2) which are for GR and S respectively. The remaining two columns i.e. colunm 1 for the GR panel and the column 3 for the S panel pertain to a specification which is only a slight variation on the model in the text since here we exclude the DISB or the disbursement variable. Incidentally, before we start discussing the results one may be reminded that except for RINT, all the variables in the model are ratios of GNP.

The results for the GR equation (columns 1 and 2) show that the coefficient estimate of net foreign private investment (FPI), the variable of central interest in this study, is positive and statistically significant at 90 percent level for one-tailed t test for 25 degrees of freedom (column 1) and almost significant at the same level in column 2. Further, the results show that the remaining three variables i.e. disbursements of grants and external loans (DISB), saving (S) and exports (XM) too have a positive impact on GR. However, with the probable exception of the coefficient of S, the significance levels of the coefficient estimates are not very high. In our view, this issue of the lack of significance is related to the problem of insufficiently long time series whenever we restrict ourselves to a single developing country most of them having come into being only after the Second World War. Lack of sufficiently long time series is also partly responsible for the relatively low [R.sup.2] for the GR equation. However, this is a common problem in studies of this kind. For instance, Mosley (1980) laments that 'depending on the sample of LDCs which is chosen and whether or not the independent variables are lagged, between 4 and 25 percent of growth in LDCs is explained by domestic savings and capital flows from abroad. (5) [In this regard also see Rana and Dowling (1988) and a recent study by Otani and Villanveva (1990).]

In terms of the saving equation, the evidence favours the 'Displacement Hypothesis'--that net inflow of foreign private investment and even financial flows by way of disbursements of grants and loans leads to a lessening of the country's saving ratio. (6) Besides our main variables of interest, FPI and DISB, the other variables which have been entered primarily as controls are RINT and GR. While RINT, as a measure of the real interest rate, is positive and significant at the 95 percent level, the coefficient estimate for GR is positive but not significant. (7)

While these results are consistent with those obtained by Fry (1984); Weisskopf (1973) and Griffin and Enos (1970) for saving and Dowling and Hiemenz (1983), Stoneman (1975) and Papanek (1973) for growth, unlike these single equation results which are susceptible to simultaneity bias, ours are perhaps more robust since they have been obtained using a simultaneous equation model.

5. CONCLUDING REMARKS

The main conclusions of this study and their relevant policy implications are being noted below:

1. It is found that both net foreign private investment (FPI) and disbursements of grants and external loans (DISB) has a positive impact on the rate of growth of real GNP (GR) in Pakistan. (However, the coefficient estimate of the DISB variable is not significant.)

2. In the saving equation, the coefficient estimates for both FPI and DISB are negative lending support to the "Displacement Hypothesis" i.e. foreign financial inflows may discourage domestic public and/or private saving behaviour and resource mobilisation efforts.

3. By specifying a simultaneous model of GR and S, we are able to control for a possible simultaneity bias which many of the earlier single equation studies of growth or saving were susceptical to.

What are the policy implications of the above results? First, on the grounds of empirical evidence, foreign private investment should be encouraged since it enhances country's rate of economic growth. However, these foreign financial inflows tend to discourage the saving ratio at home. Since we find that saving is positively affected by marginal changes in real interest rate, an optimal policy may be to encourage foreign investment for its favourable impact on growth and at the same time rationalise the domestic credit structure and interest rate policy to mitigate FPI's negative impact on saving and resource mobilisation effort.

Comments on "The Effects of Foreign Private Investment on Economic Growth in Pakistan"

I. FRAMEWORK OF ANALYSIS

The analytical framework adopted in the paper concentrates on testing, through regression, two hypotheses about foreign private investment (FPI): first, that it promotes economic growth; and second, that it displaces savings. This was justified by the authors by referring to the findings of some of the earlier studies which failed to reach any consensus on the role of foreign capital.

In my opinion, the very basis on which the hypotheses were formed is incorrect. Firstly, almost all the studies quoted in the paper are those dealing with foreign capital inflows in the form of loans or disbursements, rather than FPI. Secondly, the FPI, bulk of which is usually in the form of direct foreign investment, is distinctly different from loans in terms of its nature and impact. The loans foster economic dependence through debt servicing liability, and have import raising impact in general. The FPI, on the contrary, involves dividend and has an export generating and technology transfer impact. Even the attitude of the developing countries regarding the two is different: dependency in the former case, and sharing of power in the latter.

It is strange that the authors treat FPI like foreign loans and form hypotheses about FPI by generalising the findings in respect of foreign loans. In fact, if one surveys the literature on FPI, one hardly finds any evidence of disagreement on its role in economic growth. As an example, three such references are quoted below:

(i) Pearson Report (1969). (1)

"In our judgement available facts do suggest that direct foreign investment has added substantially to the real national income of the developing countries" [p.104].

(ii) The classic study on foreign private investment by Reuber (1973), (2) based on data for 109 countries, concludes; "The results rather clearly indicate a positive association between the stock of private direct investment per capita and GNP per capita" [p. 136].

(iii) More recently, Cynthia Day Wallace (1988) (3) concludes: "... the evidence does appear to indicate a positive ratio between direct (foreign) investment density and the economic development of the host country".

In the light of such evidence, undertaking a study on FPI in terms of simply testing the hypotheses through regression, as proposed by the authors, neither has any utility nor contributes anything to the literature on the topic. The appropriate framework for undertaking such a study is the one based on micro analysis on the impact of FPI on production, employment, exports and balance of payments, tax revenues, and transfer of technology (see, for example, Reuber's (1973) framework of analysis on the subject). (4) These are the issues which a developing host country, like, Pakistan, would be interested to have some knowledge about.

II. SURVEY OF RELEVANT LITERATURE

It is strange that none of the studies reviewed in the paper specifically pertains to FPI. Almost all of them deal with foreign capital inflows in general. There is much to be quoted on the subject from the experience of the newly industrialised countries (NICs) in South East Asia and Latin America who benefitted a great deal from direct foreign investment. There is even no mention of the emerging NICs in the Far East, namely, Malaysia, Indonesia, Thailand, and the Philippines, who indicate a rich profile in terms of FPI. It may be noted that in 1989, the FPI, compared to Pakistan's level of $177 million, was 4 times larger for Indonesia, 5 times for Philippines, 7 times for Thailand, and 10 times for Malaysia.

III. EMPIRICAL USEFULNESS

Even if the proposed hypotheses had to be tested through regression at all, it has been done in a purely mechanical way, without looking at the meaningfulness of the results. All the discussion about single vs. simultaneous equation model, structural vs. reduced-form coefficients, and two-way vs. one-way simultaneity proved to be interesting only from an econometric point of view, as the results yielded by the model were technically too weak to be of any use. All the coefficients of the growth equation in Table 2 with which the authors compare the reduced-form coefficients, are statistically insignificant. One also wonders how the authors could justify their findings of FPI displacing domestic savings, in the light of the empirical experience of the East Asian Countries which have combined a rapid inflow of FPI with an increase in savings rates.

In fact, the results could have been improved if the misspecification of certain variables in the model had been avoided. Some of such misspecifications are listed below:

--GNP which, in 'growth rate' form, is the dependent variable in the growth equation, also appears in the 'level' form on the right-hand side as a denominator of all explanatory variables.

--The dependent 'growth' variable has been defined in terms of percent change in GNP at market prices. It is clear that net factor income from abroad and indirect tax (net) which from a part of GNP, may cause this change to deviate from the true growth rate defined in terms of production of goods and services in the economy, which is captured by the GDP at factor cost. For example, for 1982-83, the growth of GNP at market prices was 9.5 percent while that of GDP at factor cost was 6.8 percent.

--Exports, except through its effect on savings, is more an outcome rather than a cause of growth. Since savings are already explicitly included in the growth equation, there is hardly any justification for including exports. If the objective is to capture the openness of the economy, the imports in the first instance, or imports plus exports as a second alternative, is a more appropriate variable.

Shaukat Ali Niazi

Planning Commission, Islamabad.

(1) Laster B. Pearson et al. Patterns in Development: Report of the Commission on International Development. 1969.

(2) Grant L. Reuber Private Foreign Investment in Development 1973.

(3) Cynthia Day Wallace, "Foreign Direct Investment. A New Climate for Negotiations with the Third World", in Yochelson (ed.) Keeping Pace: U.S. Policies and Global Economic Change, 1988.

(4) Grant L. Reuber (1973), op. cit

Authors' Note: We would like to acknowledge but not implicate Dr Nadeem A. Burney and Dr Ashfaque H. Khan for useful discussions with regard to this paper. The present paper is an abridged version of the paper presented at the conference which is available with the authors.

REFERENCES

Balassa, B. (1978) Export and Economic Growth: Further Evidence. Journal of Development Economics 5:181-89

Chenery, H. B., and A. M. Strout (1966) Foreign Assistance and Economic Development. American Economic Review 56:4 Part I, 679-733

Dowling J. M., and U. Hiemenz (1983) Aid, Savings and Growth in the Asian Region. The Developing Economics 21:1 3-13.

Feder, G. (1983) On Exports and Economic Growth, Journal of Development Economics 12:1-2 59-73.

Fry, Maxwell (1984) Econometric Analyses of National Savings Rates in Domestic Resource Mobilization Through Financial Development. Vol. II. Manila: Asian Development Bank.

Griffin, K. B. (1970) Foreign Capital, Domestic Savings and Development. Bulletin of the Oxford University, Institute of Economics and Statistics, No. 32

Griffin, K. B., and J. L. Enos (1970) Foreign Assistance: Objectives and Consequences. Economic Development and Cultural Change 18:3.

Gupta, K. L. (1975) Foreign Capital Inflows, Dependency Burden, and Savings Rates in Developing Countries: A Simultaneous Equation Model. Kyklos 28: Fasc. 2.

Intriligator, M. D. (1978) Econometric Models, Techniques and Applications. Amsterdam: North Holland Publishing Company.

Left, N. H. (1969) Dependency Rates and Savings Rates. American Economic Review 59:5.

Mosley, P. (1980) Aid, Savings and Growth Revisited. Oxford Bulletin of Economics and Statistics 42:2 79-96.

Otani, Ichiro, and D. Villanveva (1990) Long-Term Growth in Developing Countries and Its Determinants: An Empirical Analysis. World Development 18:6 769-783.

Papanek, G. F. (1973) Aid, Foreign Private Investment, Savings and Growth in Less Developed Countries. Journal of Political Economy 81:1 120-130.

Rana, P. B., and J. M. Dowling (1988) The Impact of Foreign Capital on Growth: Evidence from Asian Developing Countries. The Developing Economics 26:1 3-11.

Stoneman, C. (1975) Foreign Capital and Economic Growth. World Development 3:1 11-26.

Tyler, W. G. (1981) Growth and Export Expansion in Developing Countries: Some Empirical Evidence. Journal of Development Economics 9:1 121-30.

Voivedas, C. S. (1973) Exports, Foreign Capital Inflow and Economic Growth. Journal of International Economics 3:4 337-349.

Weisskopf, T. E. (1972) The Impact of Foreign Capital Inflow on Domestic Savings in Underdeveloped Countries. Journal of International Economics 2:1 25-38.

World Bank (1991) World Development Report 1991. New York: Oxford University Press.

(1) However, much of the DFI tends to concentrate only in a few relatively better off countries. In the 1980s, fifteen countries received 75 percent of the total DFI. Typically, very poor countries like Pakistan manage to garner only a small share.

(2) While the GR equation could also be motivated using a more formal two-sector model comprising export and non-export sectors, [as a basis for such a specification see Feder (1983)] the S equation is the fairly standard saving function.

(3) While this source gives FPI on a calendar year basis, we have used a two year moving average to represent the fiscal year counterpart.

(4) While in the case of exactly identified equation the estimates of the 2SLS and the Indirect Least Squares (ILS) are identical, the 2SLS method also yields information on asympotatic standard errors which is not available using ILS. [See Intriligator (1978), p. 385.]

(5) While a low [R.sup.2] may just be a symptom of the degrees of freedom constraint, it could also be a 'measure of our ignorance' about the determinants of rate of growth which is affected by a myriad of factors such as climate, domestic political conditions etc. which are typically not included in the specification.

(6) The argument is often presented in terms of the ratio of Domestic Saving (DS) to Domestic Gross Product (GDP). However, the variable used by us i.e. National Saving (NS)/Gross National Product (GNP) amounts to the same thing because the Net Foreign Income, the term which denotes difference between the DS and the NS, is present both in the numerator as well as the denominator and is thus 'flushed' out. In any event, even when considered without relation to any measure of national output such as GDP or GNP, the national and the domestic saving are very highly correlated.

(7) However, we suspect that if we had a relatively longer time series, the positive coefficient of GR in the S equation would have been significant. In any case, since there is a strong theoretical reason for expecting GR to have a significant impact on S, the empirical estimate is still an open question until the degrees of freedom issue is resolved more satisfactorily.

Tayyeb Shabbir and Azhar Mahmood are respectively, Senior Research Economist and Staff Economist at the Pakistan Institute of Development Economics, Islamabad.

Table 1
2SLS Estimates of Growth and Saving Equations
(1959-60 - 1987-88; N = 29)

 1 2 3 4
Dependent Variable GR GR S S

Constant 0.004 -0.01 -0.12 * -0.13 *
 (0.10) (-0.27) (2.12) (2.94)
FPI 8.80 *** 7.89 -11.50 * -9.57 *
 (1.48) (1.26) (-2.76) (-2.42)
GR 0.24 0.07
 (0.24) (0.09)
DISB 0.15 -0.08
 (0.54) (-0.52)
S 0.29 0.37
 (1.19) (1.24)
XM 0.14 0.18
 (0.58) (0.71)
RINT 0.003 * 0.003 *
 (3.01) (3.17)
[R.sup.2] 0.12 0.13 0.65 0.69
D.W. 2.03 2.01 1.52 1.41
df 25 24 25 24

* Significant at 95 percent level for two-tailed t test
for the respective degrees of freedom (df).

** Significant at 90 percent level for two-tailed t test
or at 95 percent level for one-tailed t test for the
respective degrees of freedom (df).

*** Significant at 90 percent level for one-tailed t test
for the respective degrees of freedom.