A Disaggregated Econometric Analysis of Domestic Investment in Bangladesh: Issues and Challenges

Abstract

This paper tries to assess the degree of influences of the components of domestic investment on it and to examine their causal relationships associated with them at the disaggregated level in Bangladesh so that the policy makers could formulate appropriate policies regarding the issue. In doing so, the domestic investment function has been estimated .The pre-estimating techniques (the Chow test, the Coppock Instability Index, the Jarque-Bera test, the correlation matrix) confirm that there is no structural break point of the data series in 1990 and they are more instable during pre-liberalization (1990). All the variables are positively correlated and majorly they are normally distributed for parametric estimation. In econometric analysis, the results of unit root tests (the ADF, the D-F (GLS), the Phillips-Perron and the correlogram tests) show that the data of the variables of the domestic investment function have been found non-stationary at their levels as the null hypotheses are insignificant. But, they have all been found stationary after the first difference. That is, the variables have been integrated of order one I(1). The Johansen’s Maximum Likelihood (ML) cointegration results show that there are 2 (two) long run stable cointegrating relationships between the pair-wise variables of domestic investment function. The estimated coefficients indicate that the GDP growth rate, FDI, real export and domestic credit have the positive impact on the domestic investment in Bangladesh of which real export affects it significantly. On the other hand, financial intermediation and human capital have negative impact on domestic investment but they are insignificant. The Wald test also confirms that the coefficients are jointly insignificant but some of them may be significant individually for domestic investment in Bangladesh. The VECM results show that the long run causalities exist between GDP growth rate, financial intermediation, real exports, human capital and domestic credit to the domestic investment. The short run effects exist between domestic investment and financial intermediation. There is short run dynamics to the long run equilibrium among GDP growth rate, real export, human capital to domestic investment otherwise, a divergence relation exist. The VAR estimation results indicate that the long run positive elasticities exist between real exports, domestic credit availability to domestic investment while long run negative elasticities exist between financial intermediation and human capital to domestic investment. The short run elasticities exist between GDP growth rate, FDI, financial intermediations, real exports and human capital to domestic investment in Bangladesh. Results of Granger causality test show that there are bidirectional causalities between pair-wise real export and domestic credit to domestic investment as they cause each other to grow. Otherwise, unidirectional causality exists. The response of all variables is either positive or negative in the short run but in the long run they all are responded towards the domestic investment in Bangladesh. The variance decomposition shows that the volatility of domestic investment is mainly caused by its own variation, as it always accounts for major portion (above 50%) of the fluctuations. Finally, the robustness of the results has been justified with the popular postestimation model diagnostic tests. The analysis of this paper indicates that most of the factors of investment are unfavorable for the domestic investment in Bangladesh that should be addressed properly.