The following numbers are accurate, but they can’t be taken seriously because we have no idea how the U.S. Treasury selected the 166 people in their Hollywood sample. There are 44 females and 122 males in the sample.
The mean salary of all 166 people is $79,897.63. The mean female salary is $84,841.36 and the mean male salary is lower by $6,726.72. It’s $78,114.64. However, this difference is not statistically significant.
Next, we can control for three kinds of positions: executives (mainly including directors, producers, etc.), writers and the rest (actors and actresses). The executive position made $25,153.28 more on average; this is almost statistically significant. It also is what one might expect, given the importance of directors and producers to a motion picture. The writers made $17,532 less on average, not significant by the usual statistical tests. Writers may not be a dime a dozen, but they are easier to recruit, despite their critical role, than people who actually make the movie. The latter often contribute to the final screenplays.
With these two occupational variables present, it turns out that males now make $12,589 less than females. This is not significant either. This is because all the variables together are only explaining a minuscule 3 percent of the variation in salaries.
Clearly, it’s an uphill battle to understand salaries for something as heterogeneous as acting. We do not know the related box office ticket sales, past and anticipated. We do not know the profitability of the movies, which depends on how large the production and marketing costs are in order to complement the actors.
There are lots of prices of labor in labor markets. Jobs are not homogeneous. Conditions vary a great deal even within a given city, not to say across states. There is international labor competition too. The Hollywood data do not prove anything, but they introduce a cautionary note or two, that proving discrimination by looking at statistics of pay, even with controls, is not going to be easy. Who knows what variables are really important? Who knows what’s been omitted? Why should discrimination be accorded a special position, when there are so many other economic and job-related variables of importance: education, experience, productivity, skill, on-the-job training, responsibility, leadership quality, ambition, supply of substitutes, overtime, medical costs, etc.
7:32 pm on April 5, 2019
