The effects of foreign direct investments and migration remittances on the Philippine economy: A macroeconomic approach

This paper presents a simple macroeconomic model of the Philippines using two types of estimation procedures, namely, the two stage least squares (2SLS) and the error correction model (ECM). Our macroeconomic model is composed of seven equations: national income identity, consumption function, investment function, remittances function, foreign direct investment (FDI) function, real exchange rate function and trade balance function. The 2SLS is used to analyze the relationship of these equations in a simultaneous equations system. The ECM, on the other hand, differentiates the short-run and long-run relationships of macroeconomic variables and estimates their speed of adjustment from the short-run to the long-run. Our study focuses on two major inflows-remittances and FDI since these two are the largest and the most stable form of inflows in the Philippines. Contrary to what is expected in the Philippine scenario, we found that remittances more significantly increase domestic investments rather than output in the long-run. Moreover, we found that remittances are largely determined by the degree of trade openness and the level of deployment in a country. FDI, on the other hand, does not seem to influence domestic investments significantly thus, the crowding-out or crowding-in effect suggested in the literature cannot be verified. Furthermore, we found that the government spending is the main determinant of FDI among the variables we used. Given that inflows are expected to lead to an appreciation, we confirmed that remittances indeed have a negative effect on the real exchange rate in the long-run. For FDI, however, we are not able to confirm this relationship. Finally, we found that the persistence of the effects of crises differs for most of the macroeconomic variables we used in our study hence, the dummy variables we included were not able to successfully capture the effects of the crisis. As such, we resorted to a graphic analysis of the trends observable in the variables