Using data from the 2002-03 survey, it is therefore possible to estimate the effect on
inequality under each of these four scenarios, by estimating the baseline income
distribution, and then simulating the effect of giving minimum wage workers a 10
percent pay rise, firing a random 10 percent of minimum wage workers, or both. I
then estimate the impact on the distribution of hourly wages, the distribution of
individuals’ weekly earnings, and the distribution of equivalized post-tax family
income.
In carrying out such an exercise, the question naturally arises as to how zero wages or
incomes should be treated, given that most inequality measures ignore zero values. If
under a particular simulation a worker loses his or her job, should her wage/earnings
be taken into account in calculating the inequality measure? I opt here for a halfway
solution, in which those who are simulated to have lost their jobs are coded to have
zero hourly wages (which are then ignored in calculating hourly wage inequality), but