Embarrass Me Now, Please
By Bryan Caplan
I’m now writing the most number-crunching parts of The Case Against Education: one chapter on education’s selfish return, another on its social return. The work is grueling, and haunted by the fear that I’ve made an early mistake that invalidates all my efforts. Several true friends have spent hours in my office watching me walk through my work. But I’m still nervous. I don’t want to toil for years only to be severely embarrassed when an eagle eye like Justin Wolfers exposes a critical error.
So I’m turning to you, EconLog readers, for help. I’d rather release a flawed draft and endure mild embarrassment now than released a flawed book and endure deep embarrassment in 2017. So embarrass me now, please. Here‘s an Excel spreadsheet on which all my calculations depend.
What am I doing? Snapping together the following pieces:
1. 2011 Census numbers on average income for full-time, full-year workers, broken down by high school dropouts, high school only, B.A.s, and advanced degree holders. I assume that the average dropout finished 10th grade, then quit school.
2. 2000-2013 St. Louis Fed numbers on unemployment rates, broken down by high school dropouts, high school only, B.A.s, and M.As. I assume that the average dropout finished 10th grade, then quit school.
3. The punchline from my ability bias literature review, which concludes that 45% ability bias is reasonable.
4. The punchline from my sheepskin effect literature review, which concludes that, in percentage terms, the last year of high school has 3.4 times the effect of one regular high school year, the last year of college has 6.7 times the effect of one regular college year, and the last year of an advanced degree has 6.2 times the effect of a regular advanced degree year.
5. The CBO’s numbers for private sector benefits as a fraction of pre-tax income, by education class.
6. Four student archetypes:
a. The Excellent Student, who has the raw ability of the typical advanced degree holder.
b. The Good Student, who has the raw ability of the typical B.A. without an advanced degree.
c. The Fair Student, who has the raw ability of the typical high school grad who never went to college.
d. The Poor Student, who has the raw ability of the typical high school dropout.
By construction, a person whose educational attainment and raw ability “match” gets the observed income and unemployment rate for people with his educational attainment. If you’re a Poor Student who finished 10th grade, then stopped, you get the observed averages for drop-outs: $21,107 income and 9.8% unemployment rate. If you’re a Good Student who finished a B.A. but not an advanced degree, you get the observed averages for B.A.s: $59,415 income and 3.4% unemployment.
However, if your educational attainment and raw ability don’t match, I use (1) through (5) to compute your outcomes. While an Excellent Student who drops out of high school earns much less than the typical advanced degree holder, he earns much more than the typical high school dropout thanks to his high raw ability.
7. Last, the spreadsheet calculates alternate “naive” estimates that set ability bias to 0%.
I’m happy to clarify my work in the comments or via email. All feedback is welcome, but I deeply appreciate the revelation of demonstrable errors. If you find any, bless you for embarrassing me now when the cost is mild.
P.S. Using Excel does not count as “demonstrable error.” It’s my comparative advantage.