IMPACT OF FOREIGN DIRECT INVESTMENT INFLOWS ON INCOME OUTFLOWS: A CASE STUDY OF PAKISTAN

IMPACT OF FOREIGN DIRECT INVESTMENT INFLOWS ON INCOME OUTFLOWS: A CASE STUDY OF PAKISTAN Author Names: Mahnaz Muhammad Ali Lecturer, Department of Economics Islamia University Bahawalpur (IUB), Pakistan Salma Shaheen Lecturer, Department of Economics University of Gujrat (UoG), Gujrat, Pakistan Corresponding Author: Mahnaz Muhammad Ali Lecturer, Department of Economics Islamia University Bahawalpur (IUB), Pakistan Abstract This study empirically investigated the effects of Foreign Direct Investment (FDI) inflows on Income Outflows (IO) for the Pakistan economy over the period 1993Q1-2011Q4. The Autoregressive Distributed Lag (ARDL) approach to cointegration was utilized to determine the long -term and short-term relationships between study variables. The study results indicated that FDI inflows are cointegrated with IO in case of Pakistan and thus exhibit a reliable long run relationship. Results also showed that there exists a short run positive and statistically significant relationship between FDI inflows and Income Outflows. The negative and significant coefficient of error correction term confirms long run relationship and convergence of actual series towards equilibrium. Based on empirical findings of the study, it is concluded that FDI inflows in Pakistan have negative implication for current account balance along with well cited positive effects on growth and employment. Policy makers need to attract FDI in Pakistan with special focus on sectoral composition to avoid or minimize its negative implications. Keywords: Foreign Direct Investment, Income Outflow, Cointegration, ARDL Model, 1. Introduction Foreign Direct Investment (FDI) is the category of international investment that reflects the objective of a resident entity in one economy obtaining a lasting interest in an enterprise resident in another economy (BPM 05, IMF). FDI is generally considered an important factor in achieving the country s socio-economic objectives including poverty reduction goals. In a capital-poor country 1 FDI considered as manna from heaven for recipient country (Mencinger, 2008). FDI can emerge as a significant vehicle to build up physical capital, create employment opportunities, develop productive capacity, enhance skills of local labor through transfer of technology and managerial know-how, and help to integrate the domestic economy with the global economy. This 1 like Pakistan COPY RIGHT 2013 Institute of Interdisciplinary Business Research 131

is the reason that over the last two decades, almost all developing Asian economies have progressively adopted more open policies toward FDI 1 and received considerable amount of FDI inflows in recent years 2. Like other developing countries Pakistan also received most of the FDI during last decade. However at the same time, it is also noticed that widening current account deficits is one of the less desirable macroeconomic effects of large capital inflows like FDI. So now-a-days, the economic impact of FDI on recipient country has been widely investigated. FDI is an important component of developing country s balance of payment (BoP) accounts and the increasing volume of FDI is directly related with the size of profit repatriation. Hussain (2007) showed that the initial impact of an inflow of FDI on BOP is positive but the medium term effect could become either positive or negative as the investors increase their imports of intermediate goods and services, and begin to repatriate profit. Similar to Hussain (2007), Fry (1993) found that FDI does not provide additional BOP financing for pre-existing current account deficit and concluded that FDI has negative impact on national savings and investment. Figure 1.1: Foreign Direct Investment Inflows and Income Outflows (Million US $) Empirical literature for developing countries suggests that FDI flows are significantly correlated with the current account deficit; Jansen (1995) has argued that the impact of FDI on the current account is complicated by the investment income payments that arise from FDI. The same kind of pattern can be seen between FDI inflows and income outflows in case of Pakistan (See figure 1.1). In case of Pakistan most of the available literature on FDI is related to the determinants of FDI and 1 Hossain (2007) 2 Global trend of FDI inflows shows that the share of developing and transitional economies in total FDI inflows reached record highs of 52% in 2010. (World Investment report 2011, UNCTAD) COPY RIGHT 2013 Institute of Interdisciplinary Business Research 132

discuss its positive impact on economic growth and very few studies are found which investigated negative impact of FDI on current account balance through its implication on income account and trade account. So the present research would investigate the impact of FDI on income outflows in case of Pakistan. The remainder of the paper is organized as follows: Section 2 describes the literature review. Section 3 discusses data and methodology. Section 4 presents estimation results. Concluding remarks are discussed in the final section. 2. Review of Literature Considerable amount of available literature on FDI helps us to scrutinize the different aspects of FDI as observed throughout the world. Most studies on FDI discussed its positive impact on growth and other macroeconomic variables and have been concerned with the question: how to attract FDI? [Baharom, et al. 2008; Borensztein, et al. 1995; Atique, 2004; Basu, et al. 2003; Nurudeen, et al. 2010; and Ericsson, et al. 2001; Liu, et al. 2002,]. However, only a few studies were conducted on the possible negative effects of FDI [Seabra, et al. 2005; Fry, et al. 1995; Kumar, 2007; Lehman, 2002; Woodward, 2003; Siddiqui and Ahmad, 2007; Ashfaq and Kim, 1999]. A paper by Seabra and Flach (2005) investigated the nature of the causal relationship between FDI and profit remittance in Brazil using the Granger causality test procedure developed by Toda and Yamamoto (1995). The findings of the paper indicated that FDI causes profit remittance and emphasize significant adverse long run effects of FDI attraction policies for the Brazilian economy. Whether capital flows, particularly FDI inflows, Granger cause current account deficits or vice versa this hypothesis is investigated by Fry et al. (1995). An error-correction model (ECM) with one lag and Granger causality test is used to check the direction of causality. Bidirectional Granger causality is found in some countries and no causality in other countries. It is found that when FDI is independent of other balance-of-payments flows it increases capital formation in the host country by considerably more than it does when FDI is Granger-caused by some other balance-of-payments flow. Although FDl could still crowd out domestically financed investment even if it were independent, the study found that the majority of independent FDI increases capital formation; for each $100 increase in FDI; capital formation is increased by $64. On the other hand, endogenous or interdependent FDI raises capital formation by only $13 for every $100 FDI inflow. Lehman (2002) found that structural change in external accounts of a country takes place due to FDI inflows. Author studied that trade openness and host country risks are found to increase affiliate profitability of FDI and concluded that earning repatriations are not determined through constant dividend payout ratio. By studying the experience of Brazil and Argentina for the sample period of 1996-2000 when three quarters of current account deficit of these countries was caused by FDI inflows, Lehman claimed that FDI is causing incomes and profits repatriations and thus increasing the current account deficit rather than balancing it. A similar view is discussed by COPY RIGHT 2013 Institute of Interdisciplinary Business Research 133

Kumar (2007) that FDI inflows can be risky for developing countries economies. As the FDI being foreign capital may lead to capital flight in times of extreme financial crisis. FDI may be accompanied by distress sale of domestic assets and prove harmful for the economy. The profits earned through the investment are repatriated to the countries of origin of that foreign investment. Woodward (2003) claims that FDI flows may contribute substantially to current account deficits and thus to dependence on foreign capital and vulnerability to crises. By taking the example of six economies he has shown that FDI has contributed a major portion in the current account deficits of these countries. By making FDI analogous to loan Woodward has explained that subsequent repatriation of the capital from the recipient country is same as repayment of loan. In order to examine the impact of FDI on Pakistan s imports, Ashfaq and Kim (1999) tested an import demand function. Results suggested that the inflow of FDI increases imports with a lag of one year. The coefficient is statistically significant with a positive sign and suggested that a 10% increase in the inflow of FDI increases imports by 1.8 percent. Income elasticity of import demand is less than unity (0.8) indicating that a 10% increase in real GDP increases imports by 8 percent. Hafeez et al. (2010) analyzed the impact of foreign inflows on equilibrium real exchange rate of Pakistan for the period 1993 to 2009. Monthly data is used for analysis for the sample period. Cointegration estimation technique is used for analysis. The study concluded that massive foreign direct investment inflows and workers remittances have significantly appreciated equilibrium real exchange rate of Pakistan. Results also indicated that FDI inflows in Pakistan appreciated the equilibrium real exchange rate which reflects the existence of Dutch Disease 1. The appreciations of equilibrium exchange rate affect the export competitiveness of Pakistan in world market and ultimately worsen current account balance. 3.Data and Methodology Data The series used in this study are Foreign Direct Investment (FDI) inflows and Income Outflows (IO) of Pakistan measured in US $. The sample used in this empirical investigation covers quarterly data for the period from 1993Q1 to 2011Q4 with total of 76 observations. All the relevant data is obtained from electronic and published data of International Financial Statistics (IFS) and State Bank of Pakistan. The logarithmic form of the variables has been used for the analysis. Model The main purpose of this study is to analyze the impact of FDI inflows on income outflows of Pakistan. To test the hypothesis empirically the model used can be specified as follows: Where LIO: logarithm of income outflow LFDI: logarithm of FDI LIO = α +β LFDI + µ 1 See international economics by Salvatore 9 th edition COPY RIGHT 2013 Institute of Interdisciplinary Business Research 134

µ: error term α: constant term Estimation Methodology Most of economic time series are non-stationary at levels. One way of getting the interesting information about the stationarity of time series is to plot the original series and making correlogram at both level and first difference. The second, most rigorous way is to use the Augmented Dickey Fuller (ADF) test which is the wider version of the standard Dickey Fuller (DF) test and the Phillips-Perron (PP) test is employed to verify the presence of unit root in the series. At the second stage ARDL approach to cointegration is used for empirical investigation. Recent empirical studies have indicated that the ARDL approach is more appropriate to other conventional co-integration approaches such as Engle and Granger (1987), and Gregory and Hansen (1996). The main advantage of ARDL modeling lies in its flexibility that it can be applied when the variables are of different order of integration (Pesaran and Pesaran 1997). It is more robust and better performer technique for small sample sizes as compared to other co-integration techniques. It is also argued that using the ARDL approach avoids problems resulting from nonstationary time series data (Laurenceson and Chai 2003). Another advantage of this approach is that the model takes sufficient numbers of lags to capture the data generating process in a generalto-specific modelling framework (Laurenceson and Chai 2003). Moreover, a dynamic error correction model (ECM) can be derived from ARDL through a simple linear transformation (Banerjee et al. 1993). The ARDL method estimates number of regressions in order to obtain the optimal lags for each variable, where p is the maximum number of lags to be used and k is the number of variables in the equation. The optimal model can be selected using the model selection criteria like Schwartz-Bayesian Criteria (SBC) and Akaike Information Criteria (AIC) The statistic described in this approach is familiar Wald or F-statistic in a generalized Dickey- Fuller type regression, which is used to test the significance of lagged levels of the variables under consideration in a conditional unrestricted equilibrium error correction model (ECM) (Pesaran, et al., 2001). In Pesaran et al (2001), the co-integration approach, also known as the bounds testing method, is used to test the existence of a co-integrated relationship among variables. The procedure involves investigating the existence of a long-run relationship in the form of an unrestricted error correction model for each variable is as follows: 1.1 Null hypothesis would be The decision related to cointegration between the variables exists or not is made by considering on the following criterion: If Computed F-Statistics Upper Critical Bound; Rejection of null hypothesis of no cointegration and if COPY RIGHT 2013 Institute of Interdisciplinary Business Research 135

Computed F-Statistics Lower Critical Bound Than it indicates the failure to reject the null hypothesis of no cointegration If the value of computed F-Statistic lay between lower and upper critical bound than results would be inconclusive. The second step involves the long run and then short run relationship between the study variables. The equation for long run relationship would be as follows. After the existence of long run relationship lagged error correction term (ECT) would be computed to test the short run behavior of variables. Equation for VECM model would be: It is documented that if the value of coefficient of lagged ECT is between 0 and -1, then adjustment to the dependent variable in current period is the ratio of error in the previous period. In such situation, ECT causes the dependent variable to converge to long span of time stable equilibrium due to variations in the independent variables. 4.Results and Discussion Unit Root Analysis Although there is no need to check the stationarity of the series for the ARDL Bound test, but this test is still conducted to check that none of the series is integrated of order 2 or higher, because inclusion of any variable, with I (2) complicates the F-statistics. In doing so, two unit root tests are used i.e. Augmented Dicky-Fuller (ADF) and Phillip-Perron (PP). Table 1.1: Unit Root Tests 1.2 1.3 Level (Intercept) First Difference (Intercept) Variable LIO -1.405007 0.5754 LFDI -1.558390 0.4987 ADF PP ADF PP Test Statistics (Prob) -3.027997 0.0368-2.177982 0.2159-22.38057 (0.0001)** -15.86599 (0.0001)** Test Statistics (Prob) Note: The asterisks ** showed that the coefficient is significantly different from zero at 1% level of significance. It is revealed from the results of unit root tests that both the study variables are stationary at I (1) at 1% level of significance. Lag length Criteria For co-integration appropriate lag length is required to be determined. So the next step of estimation process is to determine the optimal lag length. In order to find out the optimal lag order of the model we carried out a lag order selection criteria procedure. Among various criteria the Akaike Information Criterion (AIC) is selected for choosing the optimal lag order although other criteria also suggested the same lag length. The maximum lag length is found 2 which is depicted by * in the model (LIO LFDI). -22.19152 (0.0001)** -15.99585 (0.0001)** COPY RIGHT 2013 Institute of Interdisciplinary Business Research 136

Table 1.2: VAR Lag Order Selection Criterion (LIO and LFDI) Lag FPE AIC SC HQ 0 0.039194 2.436529 2.503994 2.463107 1 0.009233 0.990675 1.193070 1.070409 2 0.004378* 0.243840* 0.581166* 0.376730* * indicates lag order selected by the criterion LR: sequential modified LR test statistic (each test at 5% level) FPE: Final prediction error AIC: Akaike information criterion SC: Schwarz information criterion HQ: Hannan-Quinn information criterion ARDL Bounds Testing Analysis This test is based on F-statistics for determination of long run relationship. In ARDL approach, if value of calculated F- statistics is more than upper bound then it confirms the presence of cointegration. If it is lower than lower critical bound than there is no co-integration. If calculated F- statistics is between lower and upper bound than no decision can be made. Table 1.3: ARDL Bounds Testing Analysis Model Estimated (LIO, LFDI) F-Statistics 8.753708 Selected Lag Length 02 (Criteria) (AIC) Pesaran et al (2001) Critical bound values Lower Bound Value 10% 4.19 5% 4.87 1% 6.34 In the present study critical bounds by Pesaran et al. (2001) are used to check the existence of conitegration between the study variables. The tabular values are given in Table-1.3 and the calculated F-Statistics for model one (LIO, LFDI) is 8.753708 which is greater than upper critical bound at 1 percent level of significance according to Pesaran et al. (2001). It shows the existence of cointegration or long run relationship between LIO and LFDI. So the null hypothesis of no cointegration is rejected. Seabra and Flach (2005) also reported the same results by rejecting the null hypothesis of no co-integration between the FDI inflows and profit repatriation for Brazilian economy by using the Johansen cointegration (maximal eigenvalue) technique. Long Run Analysis Since there is evidence of cointegration in the model under study so it is possible to estimate the long-run impact of FDI on IO. In order to check the long run coefficients of the FDI and IO the series were normalized by taking log. COPY RIGHT 2013 Institute of Interdisciplinary Business Research 137

Variable Table 1.4: Long Run Analysis Dependent Variable: LIO Coefficient T-statistic Prob Constant 1.433660** 3.631366 0.0005 ln IO t 1 0.087652 1.052165 0.2963 ln IOt 2 0.604817** 7.390555 0.0000 ln FDI 0.106929** 3.568010 0.0007 2 R 0.877635 2 Adj R 0.872391 F-statistic 167.3537 Prob. (F-statistic) 0.000000 Note: The asterisks **denotes the significant at 1% level The long-run coefficients derived from the ARDL model is reported in Table 1.4. The results reveal that FDI has positive impact on IO and it is significant at 1 percent significance level. The coefficient of FDI shows that one percent increase in FDI inflows will lead to 0.106929 percent increase in income outflows in the long run. These findings are consistent with Seabra and Flach (2005); they also indicated long-run adverse effects of foreign direct investment on profit remittance. The implication of these findings in case of Pakistan is that the deterioration in income account mainly stemmed from repatriation of profit & dividends. The sectors which witnessed increase in repatriation of profit and dividends included power, petroleum refining and telecommunication 1. The coefficient of lagged value of Income Outflows, indicating that previous level of profit repatriation also positively affecting the current value of IO. Short Run Analysis The ECM coefficient explained the adjustment speed from short run to long run span of time. Its coefficient should have negative sign and statistically significant (Bannarjeeet al. 1998). It provided the support to confirm earlier found co-integration between the variables. The existence of an error-correction term between the co-integrating variables implies that changes in dependent variable are a function of both the levels of disequilibrium in the co-integration relationship (ECM) and deviations in independent variable. This provides the evidence that if there is any type of deviation in the equilibrium from long run how much force is needed to bring it back towards equilibrium in long run (Masih and Mashi, 2002). Table 1.5 reported the short run coefficient estimated from the ECM of ARDL. The results describes that LFDI is positively related to LIO and statistically significant at 10 percent level of significance. In short run, 1% increases in FDI inflows will lead to 0.084426% increase in income outflows. The coefficient of ECM (t-1) reported is negative and the coefficient is between zero to one. The coefficient of error term is statistically significant at 1 percent level of significance. This implies that error correction process converges to equilibrium with the speed of 89.11% from current to next time period. 1 Annual reports of SBP COPY RIGHT 2013 Institute of Interdisciplinary Business Research 138

Variable Table 1.5: Short Run Analysis Dependent Variable LIO Coefficient T-statistic Prob Constant 0.010789 0.505280 0.6150 d ln FDI 0.084426* 1.677935 0.0979 ln FDI t 1 d -0.065416-1.309701 0.1946 ECM t 1-0.891156*** -5.402709 0.0000 2 R 0.338460 2 Adj R 0.309697 F-statistic 11.76735 Prob. (F-statistic) 0.000003 Note: * showed that the coefficient is significantly different from zero at 0.10 probability level ** showed that the coefficient is significantly different from zero at 0.05 probability level *** showed that the coefficient is significantly different from zero at 0.01 probability level Policy Recommendations In case of Pakistan, FDI could not contribute towards income generating activities; rather raising conspicuous import based consumption and it is concentrated to a few non-tradable sectors including oil and gas exploration, power, communications and financial services. FDI inflows in non-tradable sectors have worsened the balance of payments (BOP) problems for Pakistan through its impact on repatriation of profit. This has built pressure on income account of current account balance of the recipient country. So the policy recommendation is that once the investor s confidence in domestic economy builds, FDI should be sector specific. It should come only in those sectors where local investors cannot invest preferably in tradable sectors. COPY RIGHT 2013 Institute of Interdisciplinary Business Research 139

REFERENCES Atique, Z., M. H. Ahmed, and Azhar, U. (2004) The Impact of FDI on Economic Growth under Foreign Trade Regimes: A Case Study of Pakistan. The Pakistan Development Review, 43 (4), Part II, 707-718. Balance of Payment Manual 05. A Document of IMF. Banerjee, A., J. J. Dolado, J. W. Galbralth, and D. F. Hendry (1993) Co-integration, Error Correction and the Econometric Analysis of Non-Stationary Data. New York: Oxford University Press. Basu, P., C. Chakraborty, and D. Reagle (2003) Liberalization, FDI, and Growth in Developing Countries: A Panel Cointegration Approach. Economic Inquiry, 41, 510-516. Baharom, A. H., M. S. Habibullah, and R. C. Royfaizal (2008) The Relationship between Trade Openness, Foreign Direct Investment and Growth: Case of Malaysia. Unpublished. Borensztein, E., J. De Gregoria and J. Lee (1995) How Does Foreign Investment affect Economic Growth? Journal of International Economics, 45(1): 115 35. Ericsson, J., and M. Irandoust (2001) On the Causality between Foreign Direct Investment and Output: A Comparative Study. The International Trade Journal, Volume 15 No. 1. Engle, R. F., and C. W. J. Granger (1987) Cointegration and Error-Correction: Representation, Estimation, and Testing. Econometrica, Vol. 55, 251-276. Fry, M. J. (1993) Foreign Direct Investment in a Macroeconomic Framework: Finance, Efficiency Incentives and Distortions. World Bank Policy Research Working Paper, No.1141. Fry, M. J. (1996) Foreign Direct Investment in a Macroeconomic Framework: Finance, Efficiency Incentives and Distortions. World Bank Policy Research Working Paper, Debt and International Finance, International Economics Department, World Bank 1-30. Fry, M. J., S. Claessens, P. Burridge, and M. C. Blanchet (1995) Foreign Direct Investment, Other Capital Flows and Current Account Deficits What Causes What? Policy Research Working Paper, 1527 The World Bank International Economics Department Debt and International Finance Division. Gregory, A. W. and B. E. Hansen (1996a) Residual-Based Tests for Cointegration in Models with Regime Shifts. Journal of Econometrics, Vol.70, 99-126. Granger, C. W. J. (1969) Investigating the Causal Relations by Econometric Models and Cross- Spectral Methods. Econometrica, Vol 37(3). Hussain, M. (2007) Impact of Foreign Direct Investment on Bangladesh s Balance of Payments: Some Policy Implications. Jansen, K. (1995) The Macroeconomic Effects of Direct Foreign Investment: The Case of Thailand. World Development 23, 193-210. Khan, A. H., and Y. H. Kim (1999) Foreign Direct Investment in Pakistan: Policy Issues and Operational Implications. Asian Development Bank. COPY RIGHT 2013 Institute of Interdisciplinary Business Research 140

Kumar, A. (2007) Does Foreign Direct Investment Help Emerging Economies. Economic Letter: Federal Reserve Bank of Dallas. Laurenceson, J., and J. C. H. Chai (2003) Financial Reform and Economic Development in China. Edward Elgar, Cheltenham. Lehman, A. (2002) Foreign Direct Investment in Emerging Markets: Income, Repatriations and Financial Vulnerabilities. IMF Working Paper. Masih, A., M. Mansur, and R. Masih (2002) Propagative Causal Price Transmission among International Stock Markets: Evidence from the Pre- and Post Globalization Period, Global Finance Journal, 13, 63-91. Mencinger, J. (2008) The Addiction with Foregin Direct Investmnt and Current Account Balance. International Center for Economic Research(ICER), Working Paper No.16, Nurudeen, A., and O. G. Wafure (2010) Determinants of Foreign Direct Investment in Nigeria: An Empirical Analysis. Global Journal of Human Social Science, Vol. 10 Issue 1 (Ver 1.0). Pesaran M. H., Y. Shinb and R. J. Smithc (2001) Bounds Testing Approaches to The Analysis of Level Relationships. Journal of Applied Econometrics, 16: 289 326. Phillips, P. C. B., and P. Perron (1988) Testing for Unit Roots in Time Series Regression. Biometrika, 75, 335-346 Rehman, H., A. A. Jaffri, and I. Ahmed (2010) Impact of Foreign Direct Investment (FDI) Inflows on Equilibrium Real Exchange Rate of Pakistan. A Research Journal of South Asian Studies. Vol. 25, No. 1, 125-141. Seabra, F., and L. Flach (2005) Foreign Direct Investment and Profit Outflows: A Causality Analysis for the Brazilian Economy. Economics Bulletin, 6(1), 1-15. Siddiqui, D. A., and M. H. Ahmad (2007) The Causal Relationship between Foreign Direct Investment and Current Account: An Empirical Investigation for Pakistan Economy. MPRA Paper No. 19743. State Bank of Pakistan. (2008) Annual Report of State Bbank of Pakistan, Karachi, Pakistan. State Bank of Pakistan. (2009) Annual Report of State Bank of Pakistan, Karachi, Pakistan. State Bank of Pakistan. (2010) Annual Report of State Bank of Pakistan, Karachi, Pakistan. State Bank of Pakistan. (2011) Annual Report of State Bank of Pakistan, Karachi, Pakistan. State Bank of Pakistan. (2012) First Quarterly Report of State Bank of Pakistan, Karachi, Pakistan. State Bank of Pakistan. (2012) Second Quarterly Report of State Bank of Pakistan, Karachi, Pakistan. Toda, H. Y., and T. Yamamoto (1995) Statistical Inference in Vector Autoregressions with Possibly Integrated Processes. Journal of econometrics, Vol 66. Woodward, D. (2003) Financial Effect of Foreign Direct Investment in the Context of a Possible WTO Agreement on Investment. Presentation to NGO workshop on WTO negotiation on investment and new issues. Geneva, 18-19. COPY RIGHT 2013 Institute of Interdisciplinary Business Research 141