Michael Simkovic (Seton Hall) & Frank McIntyre (Rutgers Business School), The Economic Value of a Law Degree, 43 J. Legal Stud. 249 (2014):
We investigate the economic value of a law degree and find that for most law school graduates, the present value of a law degree typically exceeds its cost by hundreds of thousands of dollars. The median and 25th-percentile earnings premiums justify enrollment. We track lifetime earnings of a large sample of law degree holders. Previous studies focused on starting salaries, generic professional degree holders, or the subset of law degree holders who practice law. We incorporate unemployment and disability risk and measure earnings premiums separately for men and for women. After controlling for observable ability sorting, we find that a law degree is associated with median increases of 73 percent in earnings and 60 percent in hourly wages. The mean annual earnings premium is approximately $57,200 in 2013 dollars. Values in recent years are within historical norms. The mean pretax lifetime value of a law degree is approximately $1 million.
From Michael Simkovic:
Here's a description of some of the changes to the final version compared to the last working paper:
We updated the SIPP data through 2013 (we have partial data for 2013, but seems to be about as good as full year).
The other changes are all pretty minor. The one to note is probably the instrumental variable literature review under (2).
The other changes we made, besides updating to 2013, are:
(1) Imputation and data quality improvements
We got rid of horizontal imputation, where the census fills in missing months of earnings data based on earnings of other people with similar demographics (race, sex, full or part-time work status) and instead we impute vertically, based on the months that come before and the months that come after, for that individual. This makes more sense than the census method, because census is imputing data based on other people with different levels of education.
We increased the quality of the data we used by eliminating "imputed" law degree holders (i.e., people who didn't necessarily say they were a law degree holder but the census filled it in) and requiring law degree holders to also report having a professional degree.
These changes increase the annual earnings premium a little bit. It's about $57,000 per year on average, now, as opposed to around $53,000 before. There were also increases in the annual percentage earnings premium and percentage wage premiums, and in the earnings from the quantile regression estimates (25P, 50P, 75P). See also number 3.
We added lots and lots of robustness checks, including checking for the spread of the earnings premium widening over time, which we did not find evidence of.
We also included additional information about instrumental variable (IV) studies (high tech methods of teasing out causation from observational data), the results of which suggest that OLS (the lower-tech technique we use) generally produces good estimates of the causal effect of education on earnings. Most studies actually produce higher estimates using IV than using OLS, so if we'd used a higher tech method we might have gotten a higher earnings premium. Twin studies generally suggest a slightly lower estimate than OLS. So based on what others have found in the literature, IV would probably be higher; twins would probably be a little lower; overall we've probably got a pretty good estimate.
(3) Quantile regression age groups
We replaced our 5-year age group dummy variables with a quadratic term for age in the quantile regression (this is a standard technique in labor econ). There were too many dummy variables in the quanitle regressions before. The quadratic works better and still reflects the non-linear nature of the age earnings relationship (i.e., initially earnings increase with age; then they decrease with age).
(4) Gender Break down
We include more info broken down by gender. For the quantile regressions and lifetime value of earnings, the men have a bigger spread than the women. So the 25th P for women is higher than the 25th P for men, and the 75 P for men is higher than the 75 P for women. On average, the present value of the earnings premium is about the same for men and women, because the law degree increases work hours for women more than for men.
(5) Sensitivity Analysis
We noted that with a 9 percent nominal discount rate, a 25th percentile man might not get a good private return on a law degree (at least not without IBR with debt forgiveness). We don't think this discount rate makes much sense--we use 6 percent nominal as our base case--but that's the way the numbers work out.
We see a decline in the point estimate of the earnings premium from 2008 to 2013. In 2013, the point estimate was at about the average level from 1996 to 2013. In 2008, it was way above average, so it looks like mean reversion.
Most of the rest is just changes in tone and cutting things to save space, plus updating to the latest available data, usually 2013, sometimes 2012.
Stephen F. Diamond (Santa Clara), Leading Study of JD’s Million Dollar Value Published in Journal of Legal Studies:
As the economic recovery from the collapse of the 2008-10 period continues its momentum there is some evidence that applications to top tier JD programs remain strong with applicants with very LSATs now having increased. Nonetheless, second and third tier schools remain challenged to survive the prolonged economic cycle. This study, however, is likely to reinforce the argument that the JD and law schools remain a viable and important economic institution.