Use the data in 401KSUBS.RAW for this question, restricting the sample to fsize = 1.
(i) To the model estimated in Table 8.1, add the interaction term, e401k ( inc. Estimate the equation by OLS and obtain the usual and robust standard errors. What do you conclude about the statistical significance of the interaction term?
(ii) Now estimate the more general model by WLS using the same weights, 1/inei, as in Table 8.1. Compute the usual and robust standard error for the WLS estimator. Is the interaction term statistically significant using the robust standard error?
(iii) Discuss the WLS coefficient on e401k in the more general model. Is it of much interest by itself? Explain.
(iv) Reestimate the model by WLS but use the interaction term e401k ( (inc - 30); the average income in the sample is about 29.44. Now interpret the coefficient on e401k.