Problem 2

In order to examine how effectively the government subsidies for employee training raised the amount of training provided, we have collected observations from 50 firms. The estimation results of the following regression model are shown below.

 where Ch Trainingi = change rate of annual training hours per employee in firm i; SUBSIDY i = a dummy variable which is set to 1 if firm i was granted a subsidy, 0 otherwise; SALESi log = annual (logarithm) sales; EMPLOYi log = annual average (logarithm) number of employees; and WAGEi log = annual average (logarithm) wage.

OLS estimation results:

(a) Write the null and alternative hypotheses which evaluate the effectiveness of the subsidy policy.

(b) Is it correct to use the OLS estimation results for testing the hypotheses? If not, what is the problem with the OLS estimation results?

(c) What is your conclusion about the effectiveness of the subsidy policy? In other words, test the hypotheses specified in (a) at the significance level of 0.01.