After the JD Wave 0

Last month I wrote an article for the The Am Law Daily about the After the JD project, a longitudinal study that measures employment outcomes for people who passed the bar exam in 2000. I thought it might be interesting to offer, as an appendix, the Official Guide‘s employment outcomes for Y2K law grads (they’re in the ’03 edition). Obviously, this isn’t a perfect fit as some people who passed the bar in 2000 graduated earlier, but the overlap should be fairly significant.

Back in those days, though, the Official Guide wasn’t the treasure trove of knowledge that the ABA’s employment questionnaire reports are now, and it’s certainly not as detailed as the After the JD’s information is. However, for those interested in getting a sense of the legal market many of the After the JD cohort entered into by law school, look no further.

To conserve blog space, the tables will follow the jump.

Continue reading

Class of 2013 Employment Report

[UPDATE: Emory's data are now included in this page.]

Okay, this’ll be short as I have work to do.

The ABA has released its employment data for the class of 2013, which is defined as everyone who graduated between September 1, 2012, and August 1, 2013. Employment data are outcomes as of February 15, 2014, irrespective of interim job changes. Savvy readers will notice that the most mediocre law school ever, Generic University School of Law, made an appearance again. They’ll also note that Emory’s data have been removed from the ABA’s site, which I’m sure will be corrected soon. For those who are truly curious, the ABA has even broken down the employment data for each of Cooley’s branch campuses. There are a few other traps for the unwary, e.g. schools that haven’t received even provisional ABA accreditation, like Lincoln Memorial. Even Peking China made the list. None of these schools have any useful data.

Generally, graduates fared little better this year than last year. 55.3 57 percent were employed full-time/long-term in bar-passage required jobs. This is up 0.2 0.8 percentage points from February 2013.

On the other hand, the graduate unemployment rate grew while the non-response rate declined.

Tables below the jump to conserve blog space.

Continue reading

Site Update: Lawyer Overproduction Page

You can find it here or in the “original research (updated)” menu above. It’s long overdue as I’ve received requests for its sources.

I also delisted the “law schools and law students per capita” page. It hadn’t been updated in around three years, and the lawyer overproduction page pretty much supersedes it. It’s a little sad because that was the first research project I started on this blog back in the summer of ’10. Maybe I’ll come up with a reason to put it back into the mix, but not now.

Commerce Dept.: Legal Sector Hampered Recovery in 2012

Okay, it didn’t actually say that, but in its wholesale revision of its GDP-by-Industry data the Bureau of Economic Analysis (BEA) proclaimed this morning:

Overall, 20 of 22 industry groups contributed to the 2.8 percent increase in real GDP. … Professional, scientific, and technical services real value added—a measure of an industry’s contribution to GDP—increased 4.2 percent in 2012, continuing to reflect strong growth in computer systems design and related services.

Fifteen years ago, “computer systems design and related services” (computer services) was a junior partner to “legal services” in the “professional, scientific, and technical services” industry group. The legal sector’s share was a quarter of the total and computer services’ was about a tenth. Since the Great Recession, they’ve switched places: Computer services doubled its share; legal services lost a third.

Components of PSTS RVA (2009 $)

Computer services grew 12.6 percent in 2012, legal services -0.2 percent. Way to contribute to America’s recovery, lawyers!

However, in the price inflation race, the legal sector is definitely a winner. All services are 4.6 percent more expensive than in 2009, professional, scientific, and technical services grew 4.1 percent, computer services are 3 percent less expensive, and legal services are a whopping 12.7 percent more expensive than five years ago. I see this as more evidence that the most productive strata of legal profession are being wiped out because Americans are too poor to afford them at any price.

Regular readers might be curious why the above chart stops at 1997 and not 1977 as BEA data did in previous years. The reason is that the bureau excluded those years from its revision, and it used the term “superseded” to ensure that government data daredevils like me would think twice before using the older data. Nice try.

To see the impact the revision has had, here’s a rough comparison of demand for legal services (chain-type quantity indexes) between the last two releases and the new revision.

Legal Services Real Value Added Chain-Type Qty Index

The biggest difference is that the 2014 revision detects a spike in demand for legal services in 2008—probably bankruptcies and foreclosures—whereas in earlier GDP-by-Industry releases peak law was 2005. If you think 2008 is a fluke, a deviation from the general trend, then the “decline of law hypothesis” still holds: Aside from 2008, the year with the largest quantity output for legal services was 2004. (2007 came close though.)

In all three releases the impact of the Great Recession is substantial. The revision posted this morning shows the legal sector contracting 12 percent since 2004, and there are no green shoots in sight.

Why the Stagnant Labor Productivity in the Legal Sector?

In my last Am Law Daily article, I discovered that contrary to all expectations, the quantity of private legal services provided per hour worked has stagnated for more than two decades. I think this is one of the most important insights I’ve come across, and I think the macroeconomic data on the legal sector is greatly undervalued in commentary on the future of law and law school. The lack of productivity increases certainly deserves a bit more discussion than I gave it in the article.

Legal Sector Labor Productivity

(Sources: BLS multifactor productivity tables, BEA, author’s calculations)

So, were lawyers (okay, legal sector workers, but I may use the term interchangeably) more efficient when they used paper reporters than LexisNexis?

Unlikely. I think there are other things going on here.

One, (private) legal services aren’t homogeneous outputs. An hour spent on a criminal defense matter isn’t the same as an hour spent on a landlord-tenant dispute or a corporate merger. It’s possible that many lawyers (and firms) have become more productive over the years, but the composition of the average hour of legal services Americans are buying has changed.

All this observation does, though, is raise the question of how the average hour has changed, and there isn’t much direct evidence on the subject. There are some indirect data. For example, American Bar Foundation (PDF) statistics show that since 1980, large firms have crowded-out smaller ones, though curiously solo practitioners are about as abundant as they ever were.

Percent Private Practice Lawyers by Firm Size

In 1980, only 7 percent of all private practice lawyers worked in a firm larger than 50 lawyers; in 2005, 20 percent did. But this doesn’t show much because in most circumstances we expect larger firms to be more efficient than smaller ones, so they’d cost less per client. On the other hand, Bill Henderson tells us that the employment composition of biglaw has changed since the 1980s as well, which I think can support the idea that the average hour of legal services has changed.

We also know that in general legal services have become more expensive over the years, even as the number of employed lawyers has grown. Legal services cost twice as much on average in 2011 as they did in 1985 in real terms.

Real Legal Sector Price Index

(Source: BEA, author’s calculations)

So how do you square the circle of more lawyers yet costlier legal services? In two ways: First, you can argue that demand for legal services is wealth and income elastic, i.e. the more money one has, the more they like spending it on lawyers and they buy different kinds of legal services too. And wouldn’t you know it, just this morning The National Law Journal reported that $1,000 per hour billable rates for “in-demand partners at the most prestigious firms” aren’t rare anymore. Second, you can argue that as Americans’ incomes and wealth decline, they’re unable to purchase the legal services that they used to, and the lawyers who formerly served them either move to a higher-priced market or go out of business.

The only alternative you’re left with is that there’s an attorney shortage, and all those “excess” law school graduates are really just lazy, greedy, entitled, and unwilling to make the tough sacrifices like abandoning their current lives and moving to rural America to serve the poor. However, this argument requires throwing out rational behavior assumptions and leads us to wonder why the supposedly efficient large firms won’t serve the poor if the greedy grads will not. The “market failure” line is ever wanting for an explanation.

If the composition of the average hour of legal services hasn’t changed, then the only way I can think of where we get the same (or less) output for the same effort despite technological advances and rising prices is by systemic fraud and cartelized behavior on a spectacular scale. Lawyers lie in lockstep about the cost of their services. Then they work fewer hours a week and take Fridays off. New lawyers attempting to enter the market, work more, and charge lower prices face threats and sabotage.

I’m sure plenty of people believe the above is true, but there’s scant evidence of it.


So we have stagnant legal sector labor productivity. I think it reflects wealth concentration rather than fraud.

Incidentally, clever observers might wonder how much it matters. Maybe increases in overall labor productivity have a long way to go to catch up to the legal sector’s, an observation that didn’t make the cut in my Am Law Daily piece. Indeed, most productivity gains over the last few decades have not gone to workers. Here’s a different measure, output per worker (instead of per hour, which is more precise).

Real Value Added Per Person Engaged in Production

(Source: BEA, author’s calculations)

By this measure, the legal sector has become 13.5 percent less productive between 2009 and 2011. This could mean that either all lawyers who didn’t lose their jobs in the Lesser Depression organized a work slowdown, or the Lesser Depression laid-off the most productive legal sector workers. In a variant of the latter, some practice areas might be more productive than others—especially those benefitting poorer clients and not the wealthy—and once poor people become destitute, an otherwise productive chunk of the legal sector goes out of business. Regardless, there may be significant deadweight workers in the legal sector.

What does this mean for the as-yet unmentioned prospective law student? As I wrote in the article, law will only consistently pay off for those who can serve the wealthy, at least in the short term. Also, as the largest purchaser of law graduates’ labor, the legal sector sets their wages. If legal sector productivity has stagnated while other sectors become more productive, there are probably better long-term opportunities than law.

A crucial independent variable here is how long the depression will last. Given the dismal December 2013 Establishment Survey figures (and the Obama administration’s shameful spinning of them), things are bad. Here’s the employment-population ratio for 16 to 54 year-olds and my projection of them based on the positive growth since September 2011.

Civilian Employment-Population Ratio (16-54)

(Source: BLS)

If the legal sector is sensitive to employment levels and wealth concentration, then it’s going to be a long while until law starts paying off consistently from the demand side. From the supply side, obviously, there are far fewer people going to law school, and I’m hopeful (yes!) that the median graduate in a few years will have less law school debt than in previous years. It’s also obvious that the trivial accounting identity holds true: The first person who doesn’t go to law school is the first not to be unemployed after graduating. The real questions are (a) whether the demand-side factors really are as bad as I think they are, and (b) whether some law schools’ reputations are so insubstantial that legal employers would rather hire an unemployed graduate from a better-regarded school.

New BLS Employment Projections: Only 196,500 New Lawyer Jobs by 2022

Despite the shutdown, the Bureau of Labor Statistics managed to update its employment projections for 2012 to 2022 a tad earlier than I expected. You can find them here.

One bit of good news for the legal profession is that between 2010 and 2012, the BLS estimated some growth in the number of lawyers employed in the U.S., 728,200 in 2010, 759,800 today. It’s about the same as in 2008 (759,200). However, back in 2002, the BLS projected 813,000 lawyer jobs for 2012, so once again, the projections were overoptimistic.

BLS OOH Lawyer Employment Projections

The bad news is that the projected number of employed lawyers in 2022 (834,700) is lower than in previous years, e.g. 857,700 in 2018. Moreover, the job growth rate is declining. In 2020, the BLS projected a total of 212,000 jobs due to growth and replacement. Between 2012 and 2022, the total is 196,500 lawyer positions. Dividing by ten, this means that 19,650 jobs are predicted to be created annually. Despite the law school applicant nosedive, the number of jobs law graduates and lawyers will be competing over appear to be diminishing, mainly due to fewer positions being replaced.

Here’s the master lawyer oversupply and law graduate overproduction chart. The 2000 edition of the Official Guide lists the number of new lawyers, and I supplemented that with National Conference of Bar Examiners data on lawyer licensing (which includes some people who were admitted by examination in more than one jurisdiction, so there’s some overcounting).

Lawyer and Law Graduate Projections (1983-2022)

I’d comment more, but…

I have to give a lecture on college education and student loan debt at the Henry George School of Social Science at 6:30, TONIGHT (i.e. Friday, December 20, 2013). DETAILS HERE.

Full-Time Private Law School Tuition Projections for 2017, 2022

It’s the season for the LSTB’s perennial full-time private law school tuition projections. This is the third year I’ve made such projections, and I always try to improve the methodology to produce accurate and precise forecasts. The biggest changes this year are a large increase in source data—last year I scoured the Web for older copies of the ABA-LSAC Official Guide to the ABA Law Schools (Official Guide)—and I’ve improved the projections methodology.

In previous years, I forecasted future costs via linear regression based on law schools’ previous prices. I chose this model because it offered the lowest average costs, but it proved woefully imprecise. This year, I’ve changed to using law schools’ average annual (numeric) growth rates because it is both more accurate and precise than the linear regression methodology.

You can read about my test of the methodologies here.

As always, I exclude the two private law schools in Puerto Rico and Brigham Young’s tuition for LDS students. Unlike last year, though, I will not try to make projections for public-in-name-only law schools because there aren’t enough data to make reasonably accurate predictions. [Mini-update: The projections are in current dollars.]

Like last year, the “relative variance” column on the far right of the projections table is the percent difference between the law schools’ projected costs for 2012 (based on their 1999-2011 prices) and what they actually charged in 2012. The point is to provide an indicator for distinguishing between outliers and reasonably accurate projections. In 2012, the average private law school raised its price by less than in previous years, suggesting that going forward, costs will plateau. Here’s a chart of the distribution of tuition increases over the years.

Dispersal of Full-Time Private Law School Tuition Price IncreasesDispersal of Full-Time Private Law School Tuition Price Increases (Zoomed In)

Analysis of the Official Guide data also shows that the percent of full-time students paying full freight has dropped considerably in the last decade:

Percent Full-Time Law Students Paying Full Tuition

It might be the case that the real tuition today are less for the median student than a few years ago, but it’s still important for people to know that law schools will continue to depend on students who pay full freight. Because it’s harder for them to choose to drop out, 2Ls and 3Ls are especially likely to be asked to shoulder higher costs as their scholarships are rescinded.

The Official Guide data can be found on the LSTB data page, which I’ve updated to include the 1999-2000/2003-2004 academic years and inflation-adjusted tuition instead of percent increases. Okay, here are the projections. Enjoy.

Continue reading

Another Note on Private Law School Tuition Projection Methodologies

It’s the time of year again in which I offer my projections of how much private law schools will charge full-time students five and ten years from now. This requires a note (or digression) on the methodology I use. Last year, I went to great lengths to justify my decision to use law schools’ tuition data in individual linear regressions for the projections. Two things are different this year:

(1)  Because I scoured the Internet for older copies of the Official Guide, I now have more law school tuition data. My data go back to 1999 instead of 2004. The data alter the projections.

(2)  I may’ve found a better projection methodology than individual linear regression.

The added data speak for themselves, but it occurred to me that another potentially effective way of projecting future law school tuition is by using the average annual growth rate.

This table ought to explain what I’m talking about:

1999 20,980 N/A
2000 22,055 1,075
2001 23,194 1,139
2002 24,413 1,219
2003 25,968 1,555
2004 27,617 1,649
2005 29,229 1,612
2006 30,915 1,686
2007 32,841 1,926
2008 34,710 1,869
2009 36,510 1,800
2010 38,097 1,587
2011 39,697 1,600
2012 41,132 1,435

The questions are (a) is the average annual growth rate more accurate and precise than linear regression, and (b) if so, is it better to project each law school’s tuition individually or apply the overall average ($1,550) to each law school to snuff out outliers (Faulkner, I’m looking at you)?

On average the linear regression and average annual growth rate methodologies produce the same results.

Private Law School Tuition Projection Methodologies Comparison (2012-2022)

So let’s take this to the individual law school level by asking, “Which methodology on average most accurately and precisely predicted law schools’ 2012 costs based on their 1999-2011 prices?”

No. Private Law Schools by Tuition Projection Percent Variance by Methodology (2012)

Here are the average percent variances between the projected tuition prices and actual results (closer to zero is better) and the average deviations of the percent variances as a share of the average percent variance (smaller is more compact, which is better).

  • Linear Regression: 0.2% (average variance); 1,072.5% (deviation share of average)
  • Average Annual Growth Rate: 0.4%; 382.2%
  • Average Annual Growth Rate (Individual): 0.5%; 308.7%

You can see why I’m dissatisfied with the linear regression methodology: great average, horrible distribution.

The two average annual growth rate methodologies raise a conundrum: Applying the average law school’s average annual growth rate to each school is more accurate (closer to zero) while calculating the average annual growth rate for each school individually is generally more precise (tighter deviation). Here’s what it would look like projected out to 2017.

No. Private Law Schools by Tuition Projection Percent Variance by Methodology (2017)

  • Linear Regression: 20.7%; 15.4%
  • Average Annual Growth Rate: 19.3%; 12.8%
  • Average Annual Growth Rate (Individual): 19.8%; 9.4%

I’m pleased enough with the individual average annual growth rate’s precision that I’m willing to adopt it for this year. The only other revision to this methodology I can imagine is weighting the average annual growth rate on recent tuition increases on the basis that future tuition increases are going to be more like recent ones than those of the mid-2000s. I’d like to wait and see how much (or little) tuition costs increased this year before running that experiment.

Because of the change, here are revisions to past projections according to the older data and newer methodology. Note, the revisions do not include more recent years’ data to keep them consistent with the older projections.

Continue reading

Good News: Students Borrowing Less From Education Department

The bad news is that I just updated the LSTB’s student debt data page, but revising it again is my problem, not yours.

For those in the know, the Department of Education (ED) tracks the amount of debt the government lends out each quarter (and each academic year), going back to the late 1990s. Here’s total borrowing by loan program, which includes direct loans and guaranteed loans back when they were around.

Amount of Federal Loans Disbursed

Don’t let the 2012 data throw you. Because Congress stripped subsidized Stafford loans from graduate and professional students, the 2012 bars’ meanings completely changed from previous years. Now all subsidized Stafford loan borrowers are undergraduates only, and ED kindly separated graduate unsubsidized borrowers from undergraduates. Nevertheless, the total amount of Stafford borrowing is dropping. In the 2011-12 academic year it was $85 billion; in 2012-13 it fell to $78 billion.

The declining loan volumes imply that the Office of Management and Budget’s long-term direct loan projections, which are discussed in the aforementioned student debt page) are probably high.

Projected Direct Loan Balance (OMB)

Less money lent out means fewer dollars likely to be lost to the program, so I consider this good news. However, the amount of money lent in 2012-13 is still sky high compared to the middle of the decade, and we don’t know who’s not borrowing, why they’re not borrowing, or whether their parents are just taking out dubious 410(k) loans instead.

Parent PLUS loan borrowing is down as well (-170,000 recipients), but that’s probably due to ED tightening the eligibility requirements on those loans. Grad PLUS loans are down slightly too, with about 18,000 fewer recipients. At most 2,000 of these can be attributed to law school graduates who were not replaced in the 2012-2013 academic year.

Here’s a chart of the number of recipients by loan type:

No. Federal Loan Recipients Per Year

(Note: the data point for unsubsidized undergraduate Stafford borrowers overlaps with the point for all PLUS borrowers in 2012, 6.9 million (left) and 697,000 (right), respectively.)

…And here’s the amount disbursed per recipient:

Amount Disbursed Per Recipient

Splitting graduate unsubsidized Stafford borrowers from the undergrads reveals just how much more graduate and professional students borrow. If grads and professionals go add Grad PLUS loans to their unsubsidized Staffords, they’re taking on more than $37,000 in debt in one academic year. Since there were 335,000 Grad PLUS borrowers last year, we can expect that in the near future, the highest student debt brackets (e.g. >$100,000) that we’re told aren’t really a problem will increase more quickly than the lower brackets. This much is not good news.

I’d give an update on the freestanding private law schools, but for some reason Western State didn’t appear in the data and I’m waiting for an explanation from ED.

Census Bureau to America’s Youth: ‘Don’t Grow Up!’

In the famous original Star Trek episode “Miri,” the Enterprise happens upon a world that is exactly like Earth in the 1960s (boy that gimmick got old), except the only people alive are a handful of children. The grown-ups—”grumps”—all died due to a deadly disease that was supposed to make everyone immortal. “Miri”‘s children, however, were hundreds of years old because they only aged one month per year.

Wackiness ensued.

Worse than life imitating art, children in the United States of America could only dream of living on the pseudo-Earth in “Miri.” This week, the Census Bureau updated its “Income, Poverty and Health Insurance Coverage” page for 2012, and most of the focus has been on the cratered real median family incomes. Indeed, last year the median family income was less than what it was in 1989. Heckuva job.

But I’m not here to dumpster-dive into Census data to comment on families’ incomes; rather, the update’s relevance to me is the 2012 person income tables, which contains the ever-so-useful PINC-03: “Educational Attainment—People 25 Years and Older, by Total Money Earnings in 2012, Work Experience in 2012, Age, Race, Hispanic Origin, and Sex.” PINC-03 goes back in some form or another to 1994, and it gives us personal earnings data on Americans in the 25 to 34 age bracket by educational attainment, courtesy of the Household Survey. It is awesome.

To get things rolling, here’s a chart showing the dispersion of incomes by education (with medians drawn in) in 2012 and a blow-up version to better show the LSTB’s favorite category, professional degrees:

Dispersal of Earnings by Education (25 -34) (Thousands, 2012) Dispersal of Earnings by Education (25 -34) [Blow-Up]

The PINC-03 tables measure incomes in $2,500 brackets, stopping at $99,999. After that they lump all incomes above $99,999 that into one category. Here’s a comparison:

Percent 25-34 Earning 100,000 by Ed (2012)

So what does the 2012 update add to the time series data beginning in 1994? Why, mass unemployment and declining incomes for young Americans!

Median Earnings by Education (25 - 34) (2012 $)

One of the neat tricks the government uses to buoy median incomes is to exclude people who have $0 earnings from the population (the denominator), figuring that we can’t tell whether people choose to leave the workforce or are involuntarily unemployed (because, you know, unemployment is voluntary). Excluding zero-earners makes less and less sense as we go up the education scale. Like, come on Census, you really expect us to believe that 12 percent of the people with professional degrees took on all that education debt and sacrificed prime working years just to earn nothing last year?

Percent 25-34 With Zero Earnings

As you can see in these last two charts, the professional degree-holder data are much noisier than the other categories, the reason being that they only comprise 1 percent of the 25 to 34 bracket. Small sample sizes bedevil us. The remaining categories I’ve depicted are all greater than 10 percent, so they’re likely more precise.

Now you might be asking, “Hey LSTB, how many college graduates are earning less than the median high school graduate?” and you would be clever for doing so, for the answer indicates that for many young Americans college does not provide a “premium,” unless you think being less unemployed than a high school graduate is an accomplishment. Because the PINC-03 tables provide the number of people by $2,500 earnings bracket, the next chart shows the floors and ceilings for how many people are above or below the median.

Percent Range College Grads Earning Less Than HS (25 - 34)


Now, another brief digression on professional degree-holders in the 25 to 34 bracket. Probably a substantial majority of them are law school graduates. I suspect that because according to the Digest of Education Statistics law grads dominate the 10-year degrees-conferred rate for professional school graduates.

10-Year Rate of Professional Degrees Conferred

(Source: Table 322, author’s calculations)

One thing to note is that while the 10-year degrees rate for the last couple of years has been nearly 900,000, the number of 25- to 34-year-olds in the PINC-03 tables with such degrees is usually half that. Either the Household Survey is consistently imprecise, wantonly excluding many professionals, or non-law school professionals don’t obtain their degrees until much later than 25. If the latter is true, then the “professional degrees” category in the 25 to 34 bracket is even more law grad-heavy than the rest of the population, justifying my hunch that “professional school grads” in these charts means law degree-holders and a handful of doctors.

So, here’s the percentage of professional school grads who earn less than bachelor’s degree-holders and high school graduates, aka The End of All Things.

Percent Profesisonal School Grads With Below-Median Earnings (25-34)

Three in twenty professional school graduates who have earnings at all earn less than the median high school graduate of the same age bracket, and one in eight of the total earns nothing. That’s a quarter of the total making less than the median high school graduate who has earnings. I’m not the first person to point this out, but no one goes to a professional school program to drop out of the workforce or be severely underemployed. If they do, then we’re way into credential inflation territory. And no, these people are not all in low-paying non-profit gigs, and if you say they’re going back to school to get better degrees, you get a grade-inflation immune “F”.

One more pair of charts, the percent of 25- to 34-year-olds with earnings greater than $100,000, adjusted for inflation.

Percent BA's With 100000 Earnings (25-34, 2012) Percent Pros With 100000 Earnings (25-34, 2012)

NALP, eat your heart out.

The collapsed earnings for young people depicted in these charts are among the most discouraging things I’ve ever seen since I started this blog. It’s as if the Grumps’ ultimatum to young Americans is: Either acquiesce to your stolen futures or don’t grow up at all. “Planet ‘Miri’ or Bust!”

Gene Roddenberry would roll in his grave if his ashes hadn’t been dumped into space.


Get every new post delivered to your Inbox.

Join 120 other followers