The problem with back in the dayComments
I keep getting emails from people regarding the much-discussed Simkovic/McIntyre paper, so I thought it would be worth reiterating a few basic points:
(1) S&M’s data indicate that, over the past few decades, the median law graduate has made, post-tax, about $420,000 more over the course of a career than an arguably similar college graduate who does not go on to acquire a graduate or professional degree. (Their figures are discounted to present value, but do not include the cost of acquiring a law degree).
(2) Their calculations are based on what labor economists term a synthetic estimate of work-life earnings. Here’s a simplified version of what that means. Suppose that your survey includes 25-year-old lawyers who make an average of $60,000 per year, 35-year-old lawyers who make an average of $100,000, 50-year-old lawyers who make an average of $150,000 per year, and 65-year-old lawyers who make an average of $100,000. In attempting to extrapolate the return that people going to law school today are going to get from their degrees, a synthetic estimate of work-life earnings assumes that the average career earnings of the people in the survey will be replicated (in real dollars) by the people going to law school now, over the course of their future careers.
(3) There is one very obvious problem with this methodology, which is that it will prove to be accurate only to the extent that the field whose future earning potential is being extrapolated is not undergoing any sort of fundamental change. For example, a synthetic estimate of future work-life earnings for new scriveners in 1870 or budding journalists in 1980 would have ended up generating an extremely inaccurate extrapolation of the value of a future career in those fields for people entering them in those years.
(4) S&M address this problem by assuming it away. They argue – very implausibly in my view – that there is no more reason to believe that the value of a law degree is decreasing or will decrease than there is to believe that it will increase (“One can just as easily list reasons why the value of a law degree will increase.”). The evidence they offer for this is that the earnings premium they observe in their most recent data, which run through 2011, is consistent with the average earnings premium between 1996 and 2011, i.e., the years covered by their study. But this is a weak argument, for several reasons.
First, they have no data at all on earnings for people who have graduated in the last five years, and very little data on anyone who has graduated in the last decade. Beyond this, career earnings for anyone who graduated in the last ten years must be almost completely extrapolated, which is to say that the only way to estimate these earnings is to assume that somebody who graduated from law school in in 2003 or 2013 is going to earn the same premium over the course of his or her career as someone who graduated in 1973. But this, of course, essentially begs the very question at issue – which is the extent to which a law degree is a good investment for someone contemplating getting a law degree now, not in 1973.
Second, S&M claim their data demonstrate that the earnings premium for law degree holders is “cyclical,” rather than manifesting any signs of structural change. What data they do have really show is that the earnings premium has been basically flat between 1996 and 2011, with year to year variations that, given their abruptness, appear to be noise rather than signal (For example, per their data the law degree earnings premium was at its lowest point in 1999 and 2002, but at one of its highest points in 2001).
Third, there are specific reasons to doubt that the overall earnings premium enjoyed over the past 15 years by all people with law degrees (many of these degrees were earned 30 or 40 years ago) is a good guide to the earnings premium, if any, which will be enjoyed by recent graduates and future matriculants.
For one thing, as the authors concede, the number of law graduates produced each year by law schools remains far higher than the number of jobs for lawyers being generated by the American economy. (The BLS estimates that there will be an average of 21,800 new jobs for lawyers over the course of the decade. ABA-accredited law schools are currently graduating more than 46,000 people per year, although after this year that number will decline somewhat). In other words, the oversupply of lawyers gets markedly worse with every passing year. It seems implausible that an increasing level of over-saturation in a labor market will not produce a decline in the relative earnings in that labor market, all other things being equal.
For another, there is no question that the earnings of new law graduates have in fact declined sharply. How sharply? Let’s compare the earnings of new law graduates in 2002 – note that 2002 marks the single worst year for the earnings premium earned by all law degree holders in the authors’ study – with the most recent class for which we have data, 2012 (All figures below are in 2012 dollars):
Median salary for 2002 graduates (53% of all graduates had a reported salary): $73,383
Median salary for 2012 graduates (44.7% of all graduates had a reported salary: $61,250
In other words, reported entry-level salaries for law graduates have declined by 20%, while the real decline is probably larger, given the significantly lower percentage of 2012 graduates who had a reported salary. (I estimate the real median salary for 2012 graduates was in the neighborhood of $45,000).
Furthermore, in the course of this decade, the cost of law school attendance has continued to increase dramatically, with private law school tuition rising by 42% in real terms, and resident tuition at public law schools nearly doubling, again in inflation-adjusted dollars.
(1) Over the time period covered by the authors’ study, their data indicate that earnings for all law degree holders have remained flat, while the cost of law school has risen sharply. This means that, per their analysis, the value of a law degree has declined significantly in absolute terms during the time period covered by their study, even if one assumes that recent graduates will not suffer any decline in their career earnings relative to older graduates (This latter assumption, again, is simply a methodological axiom the authors adopt, rather than a conclusion supported by any data).
(2) Recent law graduates have in fact suffered a sharp decline in their earnings at the beginning of their careers, relative to the initial career earnings of older law graduates.
(3) Law schools continue to graduate two JD holders for every available legal job.
So law students are paying vastly higher prices to get much-lower paying jobs, assuming they get jobs at all (19% of the class of 2012 did not have reported employment of any kind nine months after graduation, and only slightly more than half the class has a legal job, loosely defined).
In response to these grim facts, Simkovic and McIntyre enhance their data with the methodological magic of synthetic work-life earnings, thereby suggesting strongly that, back in the day, law school was quite a bargain.
But the problem with back in the day is that it was back in the day.
. . . Further thought: S&M can concede their study shows the absolute value of a law degree has been declining over the entire time period they study, yet still argue that the relative value of the degree (relative to stopping school after undergrad) remains steady, because the value of an undergrad degree continues to deteriorate at a similar pace. There are two problems with this response: (1) It highlights that, when calculating opportunity cost, the paper doesn’t take into account postgraduate educational opportunities other than law school; and (2) If the cost of a degree continues to rise and the return from it remains steady, at some point the value of that degree versus no degree at all becomes negative. In other words, if the NPV of the average college degree becomes negative relative to not going to college (we may be approaching that point with many college degrees), pointing out that a law degree is more valuable than something that has negative value becomes an increasingly weak argument. Extrapolation is always dangerous, but if one doesn’t have to extrapolate the current trends in cost/benefit ratios very far before both higher education in general and law school in particular become indefensible investments.