Consider the following regression (robust standard errors in parentheses) log(profit) = 0.1309 +0.5344 log(invest) + (0.0854) (0.1962) 0.2117 log (FDI) + (0.0956) 0.2301 tech. (0.0526) In the above, profit represents net profit (measured in million USD); invest and FDI are domestic and foreign direct invest, respectively (both measured in million USD); tech is a binary variable, = 1 if the company is considered high-tech. (a) Consider high-tech companies. Based on the regression estimates, what is the change in net profit for a 1% increase in domestic investment? How about a 1% increase in foreign direct investment? (b) Does your answer for part (a) change if we instead consider firms that are not high-tech? Another researcher decides to run a different regression (robust standard errors in parentheses) log (profit) = = 0.1022 + 0.6201 log(invest) + 0.1985 log (FDI) (0.0979) (0.2538) (0.1022) +0.1821 tech+ 0.1053 log(FDI) x tech. (0.0736) (0.0899) Based on the new regression estimates, (c) Consider the following claim: The elasticity of foreign direct investment is the same for high-tech and non high-tech firms. Formally state the claim as a hypothesis testing problem. Conduct a statistical test.
