Question

#1 Consider the following regression which relates salary and performance for Major League Soccer

players.

salary =B₁ + B₁games + B₂goals + Bassists + u

where salary is the salary paid to a player for the season, games denotes games played, goals is the

number of goals scored, and assists is the number of assists scored. The coefficients ₂ and 33 relate to

player performance; that is, better players will score more goals and obtain more assists.

a) What is the expected sign on B₂? Propose a one-sided hypothesis for this coefficient where the

alternative hypothesis is in the expected direction of B₂.

b) Propose the hypothesis (both the null and the alternative) that player performance does not matter for

the determination of salary.

c) Suppose you want to run an F-test for the hypothesis P₁ = 0, B₂ = 0. What is the regression that is run

to obtain the SSR of the restricted regression?

d) If the error term were heteroskedastic in the regression, would using the usual t-statistic be sufficient to

determine statistical significance of a regression parameter? Explain.

Question image 1