The major project in my economics forecasting project is to forecast what effect, if any, an increase in the federal minimum wage would have on the hiring chances of a young worker. I’m going to isolate my cohort to 16 to 19 year olds.
Venturing into a somewhat unkown field, I would like to use panel data for my analysis, and I’m thinking of doing a comparison of Virginia, which uses the federal minmum wage, and North Carolina, which has a minimum wage currently set at 6.15 an hour, which went into effect on January 1, 2007. I haven’t yet settled on the actual cross-section, but I was thinking of dividing into NC_urban and NC_rural and VA_urban and VA_rural. I would like to find some way to have a suburban factor, too, but I’m not sure how to do that yet.
In any event, since it is a forecast I need to decide A) What exactly it is I intend to forecast and B) How I intend to explain changes in whatever it is I do forecast.
The variable to be explained could either be the unemployment rate of individuals aged 16 to 19, or it could be the probability that a person aged 16 to 19, whose current wage is inbetween the current minimum wage and the new minimum wage, will be “disemployed” – the term the literature uses a great deal. The latter is a more direct approach but will probably require a subtler model. So be it. I’ll roll with for now.
Edit: I realized not long after finishing this post that I could not use probability of disemployment by minimum wage as a dependent variable. That would require surveying individual workers. So I think I’ll stick with the unemployment rate for 16 to 19 year olds as my dependent variable.
How to explain this? That comes down to supply and demand for labor and also to the economic short run fluctuations.
Since what I’m attempting to do is explain the variation in unemployment of 16 to 19 year olds as a result of changes in the minimum wage, it is necessary to control for all other factors that will affect the level of employment of someone within that cohort. So first there’s the question of labor supply:
The quantity of labor supplied = f(wage, cohort size, season [christmas? summer?], educational attainment, in school? [yes/no], family income, race, geographical region[urban, suburban, rural]).
The quantity of labor demanded = f(wage, marginal rate of technical substitution, # of firms, turnover rate)
Of course, if we’re talking about trying to measure the effect that changes in the minimum wage, then we’re talking about a still smaller group of people who are “at risk” – that is, whose current wage is between the current minimum and the future minimum. Jobs that would hire people for such low wages are typically fast food and retail in general, as well as a few services here and there. Then there are those youngsters who work on farms and perhaps do even more complicated work but, since they live in rural areas, have lower wages in general due to a lower cost of living.