'As explained in the text, the Heckman regression model (Equations 20.18 and 20.19, page 632) implies that

ðYijζi > 0Þ ¼ α þ β1Xi1 þ β2Xi2 þ***þ βkXik þ βλλi þ ni where βλ [ σεrεδ, λi [ m &ci ð Þ, and ci ¼ γ0 þ γ1Zi1 þ γ2Zi2 þ***þ γpZip Show that the errors ni are heteroscedastic, with variance VðniÞ ¼ σ2

ε 1 & r2

εδλiðλi þ ciÞ " #

where σ2

ε is the error variance in the regression equation (Equation 20.18), and rεδ is the correlation between the errors of the regression and selection equations. (Hint: See Equations 20.17 on page 631 for the variance of an incidentally truncated variable in a bivariate-normal distribution.)