CAPITAL FLOWS AND ECONOMIC GROWTH IN NIGERIA: AN ECONOMETRIC APPROACH

On the front burner of developing countries’ economic policy is the pursuit of economic growth and development. This however has been hindered by inadequate resources needed to drive the process of growth and development. One of the key components fronting the movement in support of economic globalization and integration is capital flows considering its complementarity effect in bridging the gap between domestic savings and investment. This study therefore examines the impact of capital flows on economic growth in Nigeria using data covering the period 1981 to 2016 and sourced from the Central Bank of Nigeria. The method of error correction model framework and autoregressive distributed lag was adopted in estimating our specified model. Findings from our estimated model reveal that capital flows significantly affect economic growth in Nigeria. The study thus recommends that, sound, robust and vigorous economic policies be formulated with the sole purpose of attracting and drawing capital flows into the country that helps to bridge the needed capital for economic growth and development in Nigeria.


Introduction
The desire for capital by developing countries as complement to domestic savings for growth and development has existed for many decades. This is spurred by the gap between savings and investments required to sustain economic growth and evidenced by the attention given to the drive for foreign capital as an important source of augmenting the saving-investment gap in most resource deficient economies especially in developing countries (Adeola, 2017; Orji, Uche and Ilori, 2014; Nwosa and Amassoma, 2014). This has led to the arguments that external financing is critical if not inevitable for the sustained growth of countries like Nigeria (Adegboye, Ogbebor and Egharvba, 2014). As put succinctly by Essien and Onwioduokit (1999), the need for foreign capital arises when the desired investments exceed actual savings and investments with long gestation periods that generate monetary returns and growing government expenditure that are not tax financed. The long run development of an emerging economy like Nigeria would require persistent and massive investment expenditures that can match the dire need for capital (Adegboye et al, 2014). Due to the dynamism of the core economic underpinnings, several means are continuously being explored to attain these goals. One of the measures is to open the trade among nations and accelerate the smooth flow of foreign direct investment (Okafor, Ogochukwu and Chijundu, 2016). A priori, the current reforms engineered in most developing countries' economies such as Nigeria, spanning through the financial sector, public sector and the private sector supports the fact that foreign capital flow is a factor for economic growth and development (Ekwe and Inyiama, 2014).
However, controversy regarding the costs and benefits of capital flows through the concept of globalization has taken center stage in policy and academic circles. While concerns over the benefits of capital mobility once voiced by John Maynard Keynes during the design of the Bretton Woods System were nearly forgotten in the 1970s and 1980s, the crises of the last decade have revived the debate over the merits of international financial integration (Alfaro, Kalemli-Ozcan and Volosovych, 2007). The most powerful argument in favor of international capital mobility, voiced by, among others, Fischer, Obstfeld, Rogoff and Summers (2000), is that it facilitates an efficient global allocation of savings by channeling financial resources to their most productive uses, thereby increasing economic growth and welfare around the world. Nonetheless, massive capital inflows create a strong challenge for economic managers in the conduct of macroeconomic policies. This is because attempts at curbing exchange rate appreciation through monetary policy tightening, may even result in additional inflow of foreign capital into the domestic economy (given that higher interest differentials are signals for higher returns) and thereby putting further pressure on the exchange rate. Even if international capital flows do not trigger excess volatility in domestic financial markets, it is still true that large capital inflows can spark off inflation in the presence of a fixed exchange-rate regime (Kaminsky, 2005). Similarly, transitory capital inflows may distort relative prices, with the domestic economy losing competitiveness as a result of the appreciation of the real exchange rate (Kaminsky, 2005;Caruana, 2011 portfolio investment has a positive relationship with rate of growth of real nonoil gross domestic products. Modou and Liu (2017) in their study of 13 West African countries using data from 1985 to 2015 found that both trade and FDI significantly contribute to economic growth and the existence of unidirectional and bidirectional causality between FDI and trade on economic growth respectively. Saibu (2014) investigate the effect of capital inflow on economic growth using principal components analysis and autoregressive distributed lag (ARDL) bound testing methodology found that capital inflow when interacted with trade openness has significant impact on growth hence providing empirical support for the modernization hypothesis that capital flows and trade openness are growth stimulant. Ogbebor and Egharvba (2014) using vector error correction mechanism on Nigeria's quarterly data between 1981 and 2012 revealed that capital flows, with a focus on foreign direct investment, portfolio investment and external debt, have significant effects on economic growth within the period covered by the study. Baghebo

Theoretical Framework
The theoretical framework for the for the study is the endogenous-growth AK model developed by Frankel (1962) and modified by Rebelo (1991) and Pagano (1993) in explaining the connection between finance and growth. The AK model is an extension of the endogenous model of economic growth that investigate the connection between capital flows and economic growth. The AK model exhibit endogenous growth assuming that capital is the only input into production and that there is constant return to scale. With these assumptions the production function can be given as: Where Y is total output in period t, K is stock of capital in period t and α is the total factor productivity growth or marginal productivity of capital.
Constant returns to scale ensure that output grows at the same rate as capital stock. However, to ensure sustained growth, we match increases in capital with equal growths in other inputs. As such, human capital with technical progress embodied in the quality of labor introduced into equation 3.1. Thus, equation 3.1 becomes; Where H is the level of human capital. When human capital is incorporated into the model the role for policy is extended. Endogenous growth can arise when capital and labor are augmented by additional inputs in the production function. However, if we assume human capital to be constant, that is equal to 1, equation 3.2 reduces to: Now, if we assume there are no growth in population, and only one good is produced that is consumed or invested. If this only one good is invested and capital stock is depreciating in every period at a rate of  , gross investment becomes:  Where CF is capital flows, S as domestic savings and I as in. Given equation 3.6, the steady-state growth rate is given as: Where there is international capital flows it is expected that * y g be greater than y g , * s be greater than s , and as a corollary * t I is greater than t I . Equation 3.7 describe how capital flows affect long run growth of an economy. Decoding this theory into empirical specification by integrating capital flows and other macroeconomic variables, equation 3.7 can be specified thus: Where y g is the growth rate, Z is omnibus variable capturing other macroeconomic variables affecting growth, CF is capital flows (like foreign direct investment, foreign capital flows, etc) and OPN is the variable capturing the degree of trade openness.

Model Specification
Following our theoretical framework, we specify our model thus: However, with the estimation of short and long run equilibrium impact of capital flows on economic growth, the static model is re-specified to capture the error correction mechanism that adjust the economy back to equilibrium in the event of shock. Thus, re-specifying equation 3.8 gives: Where; rgdp is the real gross domestic product used to proxy output growth.
fdi is foreign direct investment fpi is foreign portfolio investment exd is external or foreign debt topn is trade openness exr is exchange rate β's are the parameters  is error correction parameter  is the random error term t is time Considering the theoretical framework enunciated above, we expect all the parameters to be estimated to be positively related to economic growth. Essentially, we expect that:

Normality
The Jarque Bera statistic tests for normality of the distribution around the mean. It is a goodness of fit test that ascertain the skewness and kurtosis of the data. Skewness measures the extent to which a distribution is not symmetric about its mean value and kurtosis measures how it is fat the tails of the distribution. If the probability value of the Jarque Bera value is greater than 0.05, there is the presence of normality and if less than 0.05 it is otherwise.

Source: Researcher's Computation (2018) using E-Views 9.0
From the table above, the normality test on each of the variables shows that real gross domestic product, trade openness and exchange rate are normally distributed with the probability value associated with their respective Jarque Bera statistics greater than 5% (percent). Meanwhile, the trio of foreign direct investment, foreign portfolio investment and external debt have no normal distribution.

Correlation Analysis
Correlation measures the extent or size of linear association or relationship between variables. Values of the correlation coefficient are between plus one (+1) and minus one (-1). A correlation coefficient of +1 indicates the existence of a perfect positive association between the two variables, while a correlation coefficient of -1 indicates perfect negative association. A correlation coefficient of zero, on the other hand, indicates the absence of relationship (association) between two variables (Brooks, 2008). The table below shows the correlation matrix among dependent and independent variables.  Http://www.granthaalayah.com ©International Journal of Research -GRANTHAALAYAH [190] 0.0000*** 0.0000*** 0.0006*** 0.0814 -----EXR 0.914485 0.801079 0.415981 0.618152 0.874535 1.000000 0.0000*** 0.0000*** 0.0116*** 0.0001*** 0.0000*** -----**Significant at 5% level of significance ***Significant at 1% level of significance Source: Researcher's Computation (2018) using E-Views 9.0 The correlation results as reported in Table 4.2 above shows that foreign direct investment, foreign portfolio investment, external debt, trade openness and exchange rate have positive correlation with real gross domestic product. This implies that as these variables move in an upward direction, real gross domestic product move in the same direction. Foreign direct investment as shown in the above table 4.3 has positive correlation with foreign portfolio investment, external debt, trade openness and exchange rate. Similarly, foreign portfolio investment has positive correlation with external debt, trade openness and exchange rate. Also, external debt, as shown in the table, has positive correlation with trade openness and exchange rate. Lastly, trade openness is positively associated with exchange rate.

Autoregressive Distributed Lag Model
Autoregressive distributed lag model (ARDL) provides for the estimation of both short run and long run effect of the explanatory variables on the explained variables. Reported are the short run and long run impact of the control variables on the dependent variables.

Source: Researcher's Computation (2018) using Stata 13
The reported bound test examines the existence or otherwise of cointegration among the variables. It ascertains the existence of long run equilibrium in the estimated Autoregressive Distributed Lag model (ARDL). From the table above, the result of the bound test shows that the reported fstatistics and t-statistics are higher than the upper bound critical value of 3.79 and -4.19 respectively implying that the underlying null hypothesis of no cointegration for the model is rejected at 5% (percent) conventional level of significance and confirming the existence of long run equilibrium among the variables. Dependent variable is real gross domestic product. ***, ** and * indicate the coefficients are significant are 1%, 5% and 10% respectively.

Source: Researcher's Computation (2018) using Stata 13
From the above result, in the long run, foreign direct investment, foreign portfolio investment and external debt have positive impact on real gross domestic product and significant at 5% (percent) and 1% (percent) level of significance. Essentially, and as shown in the estimated model, a naira increase in foreign direct investment will in the long run significantly raise real gross domestic product by ₦332.78. Similarly, a naira increase in foreign portfolio investment will yield ₦57.44 significant increase in real gross domestic product in the long run. Also, a naira increase in foreign debt incurred by the government will bring about ₦23.81 increase in real gross domestic product in the long run. However, the duo of trade openness and exchange rate have negative impact on economic growth in the long run with the impact significant at 5% (percent) conventional level of significance.
The R-square and the adjusted R-square given as 0.99 and 0.98 respectively indicate that variation in real gross domestic product is largely explained by the control variables. Specifically, the adjusted R-square shows that 98% (percent) movement in real gross domestic product is well explained by the variables in the long run. The reported f-statistic shows that the overall long run model is statistically significant at 1% (percent) conservative level of significance.
The estimated short run model results of capital flow and growth is reported above. In the short run, foreign direct investment has a negative impact on economic growth with the impact significant at 5% (percent) level of significance. Similarly, the impact of external debt on economic growth is negative albeit not significant at both 5% and 1% (percent) level of significance. On the other hand, foreign portfolio investment has positive impact on real gross domestic product albeit not significant at 5% (percent) level of significance. Meanwhile, the impact of trade openness and Http://www.granthaalayah.com ©International Journal of Research -GRANTHAALAYAH [192] exchange rate on economic growth is positive with both significant at 5% (percent) conventional level of significance.
The coefficient of the cointegration factor or error correction term is negative and significant at 1% (percent) level of significance. The error correction factor measures the speed of adjustment towards long run equilibrium. The desired effect of the error correction factor implies that the rate of variation by which growth adjust to shock per time is not significantly different from zero. The error correction term shows that speed of adjustment of economic growth to long run equilibrium in the event of shock in the equation is 2.17% (percent). This implies that 2.17% (percent) of the divergence between long run and short run real gross domestic product is remedied within a year.

Discussion of Findings
Capital flows as shown above exhibit positive and significant impact on economic growth particularly in the long run. Essentially, whereas foreign direct investment might pose a negative impact on growth in the short run, it goes to have a significant impact on growth in the long run.  Rodrik (1992) that state that trade openness could cause macroeconomic volatility enhancing inflation, deflating exchange rates and resulting in balance of payment crisis that consequently impact on growth. Similarly, Matteis (2004) state that trade openness place exogenous constraints to economic growth by contributing to impose their reliance on global demand and to amplify their exposure to the volatility of international markets. Levine and Renelt (1992) also established that trade openness negatively impacts local investments. External debt has negative impact on economic growth in the short run, and turns positive in the long run. The result is consistent with findings of Adegboye, Garba and Egharvba (2014). It however negates the findings of Quattara (2006).

Summary
The main thrust of this research work is to ascertain the impact of capital flows on economic growth in Nigeria. Essentially, the work attempt to critically identify the impact of capital flows on Nigerian economic growth using data from 1981-2016. The study adopted an econometric method of analysis and data were sourced from the Central Bank of Nigeria. Descriptive statistical analysis was carried out on the data to establish the characteristics and/or attributes of our datasets. Correlation analysis was further tested for to determine the degree of association among the variables. Stationarity test using Augmented Dickey Fuller test was applied to ascertain the order Http://www.granthaalayah.com ©International Journal of Research -GRANTHAALAYAH [193] of integration and to check for the possibility of each of variables having unit root. Johansen cointegration test was performed to verify the existence of long run equilibrium relationship among the variables. Given the existence of long run equilibrium as evidenced by Johansen cointegration test, error correction model estimation technique was conducted on the specified model. Consequent on the result of the error correction model, given the error correction coefficient not in consonance with econometric theory by assuming a positive but significant value, autoregressive distributed lag model bound test was conducted to further prove the existence of long run equilibrium relationship between capital flow and economic growth. The result of the bound test further shows the existence of long run equilibrium relationship and consequently autoregressive distributed lag model techniques was applied on the specified model and the short run effect, long run effect and the error correction factor or term that correct for disequilibrium in the model reported.
Empirical findings from this study shows that: 1) Foreign direct investment has negative impact on economic growth in the short run, but in the long run, positively and significantly impact economic growth implying that an increase of FDI flows will affect growth rate negatively in the short run, and in the long run, positively stimulate growth rate in Nigeria. 2) Foreign portfolio investment has weak positive effect on economic growth in the short run, but impose strong and significant impact on growth in the long run. This uncover the importance of foreign portfolio investment in spurring economic growth. 3) External or foreign debt has negative and significant impact on economic growth in the short run but as proceeds of foreign debt are put into infrastructural and other development activities, external debt stimulate growth as evidenced by the long run estimates which is positive and significant. 4) Trade openness impact growth positively in the short run, but in the long run the degree of openness impact economic growth negatively.
Similar to what was obtained for trade openness, exchange rate positively impacts economic growth in the short run but turns to impact growth negatively in the long run. The implication of this is that volatility of exchange rate may not be felt in the short run, but the impact can turn to hurt growth rate in the long run.