Readers might recall several months back when I crafted a modified Lorenz curve of full-time law school applications against the U.S. News rankings. It occurred to me that I could expand my use of this rankings-adjusted Lorenz curve to law school employment outcomes for the class of 2014. Now, I’ve finally made the time to do it.
As a refresher, a Lorenz curve plots the distribution of two variables cumulatively, most commonly cumulative income against cumulative population, starting with the individuals with the lowest incomes. Such an analysis would be able to allow you to claim, for example, that the bottom 50 percent of households earn only 10 percent of the income. Because the dependent variable is rarely distributed equally, the Lorenz Curve isn’t a triangle, and the more concave the curve, the more unequal the distribution.
Furthermore, if you subtract the area under the Lorenz curve from the total area of the triangle representing perfect equality, and then divide that by the same triangle’s area, you get the Gini coefficient. ((Triangle – Area) / Triangle)
Normally, a Lorenz curve of each category of law school outcomes should be sorted with the smallest contributors first, but I want a commonly understood, “neutral” independent variable that allows comparisons among employment outcomes. As a result, the Lorenz curves below squiggle and deviate quite a bit. They also come out convex for some outcomes.
Here are the four major employment statuses that people care about, less school-funded jobs (to show the “real” demand for graduates’ labor) and excluding the three Puerto Rico law schools.
This doesn’t look too bad. Unemployment is a tad convex, and full-time, long-term bar-passage-required jobs are fairly evenly distributed. JD-advantage jobs look very evenly distributed.
Digging deeper into employment types, we find that jobs aren’t doled out evenly at all. Private practice jobs in particular are given out directly according to ranking.
The same goes for public interest jobs and federal clerkships.
I’ve also calculated the Gini coefficients for these outcomes irrespective of the U.S. News rankings. My idea was to see how some of these outcomes compare to income inequality in various countries, since that’s something people might be familiar with.
| EMPLOYMENT STATUS | GINI COEFFICIENT | PROPORTION OF TOTAL GRADS |
|---|---|---|
| Employed – Bar Passage Required FTLT | 0.31 | 25,344 / 43,195 |
| Employed – JD Advantage FTLT | 0.36 | 4,774 / 43,195 |
| Employed – Profession Position FTLT | 0.53 | 1,371 / 43,195 |
| Employed – Non Profession Position FTLT | 0.72 | 200 / 43,195 |
| Employed – Undeterminable | 0.93 | 21 / 43,195 |
| Employed – Pursuing Graduate Degree | 0.44 | 693 / 43,195 |
| Unemployed – Start Date Deferred | 0.64 | 313 / 43,195 |
| Unemployed – Not Seeking | 0.54 | 553 / 43,195 |
| Unemployed – Seeking | 0.43 | 4,103 / 43,195 |
| Employment Status Unknown | 0.67 | 841 / 43,195 |
| EMPLOYMENT TYPE | GINI COEFFICIENT | PROPORTION OF TOTAL GRADS |
| Solo-FTLT | 0.53 | 803 / 43,195 |
| 2-10-FTLT | 0.33 | 6,695 / 43,195 |
| 11-25-FTLT | 0.38 | 1,796 / 43,195 |
| 26-50-FTLT | 0.42 | 990 / 43,195 |
| 51-100-FTLT | 0.46 | 771 / 43,195 |
| 101-250-FTLT | 0.51 | 1,057 / 43,195 |
| 251-500-FTLT | 0.66 | 1,068 / 43,195 |
| 501-FTLT | 0.79 | 3,934 / 43,195 |
| Unknown-FTLT | 0.83 | 186 / 43,195 |
| Business/Industry-FTLT | 0.36 | 5,274 / 43,195 |
| Government-FTLT | 0.32 | 4,569 / 43,195 |
| Public Interest-FTLT | 0.52 | 1,744 / 43,195 |
| Federal Clerkship-FTLT | 0.68 | 1,247 / 43,195 |
| State Local Clerkship-FTLT | 0.57 | 1,982 / 43,195 |
| Other-FTLT | 0.96 | 26 / 43,195 |
| Academia-FTLT | 0.50 | 451 / 43,195 |
| Unknown Employer Type-FTLT | 0.88 | 51 / 43,195 |
At 0.31 full-time, long-term bar-passage-required jobs are pretty nicely distributed, more so than JD-advantaged jobs (0.36). This means that the rankings make JD-advantage jobs appear much more equally doled out than they really are, which is probably not a good thing for that employment status.
Likewise the convex curve and 0.43 Gini coefficient for unemployed-seeking grads means they’re very much shifted to low-ranked schools. Generally, convexity corresponds to undesirable employment outcomes.
Taking the cake are the high-Gini-coefficient, concave curves. Half of all public interest jobs went to schools ranked in the top 60; half of all federal clerkships went to grads from the top 20; and half of all 501+-lawyer firm jobs went to grads in the top 14. In fact, at 0.79, the Gini coefficient for 501+-lawyer firm jobs are distributed worse than income Namibia, which in 2010 had a Gini coefficient of 0.597.
If you’re wondering why I used this year’s rankings as the independent variable, even though they obviously could not have had any effect on the previous year’s graduates’ employment outcomes, it’s because I wanted to illustrate just how rigid employers’ regard for law school prestige is. It’s so persistent that you can see the relationship when temporally there shouldn’t be one.
To put it differently: In 2014, the distribution of quality law jobs was as bad as or worse than income in kleptocratic, HIV-riddled states. You might think that’s an unfair comparison, but it certainly resonates for me.
The Last Gen X American 
5 comments