This is the third of three posts on the current state of the legal job market. In the first post, I discussed the preliminary results of a study I did of employment of recent graduates from my law school and explained why nine-month job data does not tell the whole story on employment. In the second post, I explained why I am skeptical of the idea that the current state of the legal job market is the result of structural change. In this post, I discuss Bureau of Labor Statistics projections about the future of the legal job market.
The Bureau of Labor Statistics Projections.
Recent projections by the Bureau of Labor Statistics have received a lot of attention in discussions of the legal job market. For example, Eric Posner recently wrote on Slate that “the Bureau of Labor Statistics estimated that 218,800 new legal jobs would be created between 2010 and 2020.” There is a significant mismatch between this number and the projected number of law graduates over the same period, and it is understandable why many people have raised concerns about them when talking about the future of law schools.
I think we need to be careful, however, in reading too much into these projections. I have two specific concerns about the numbers. Before I get to them, I want to note that the BLS is a remarkably user-friendly government agency. Fearing a huge, impenetrable bureaucracy, I asked a research assistant to try to find me someone to talk with about the projections. She came back ten minutes later saying that I could just call and they would connect me to the right department. When I did call, a human being picked up the phone, and connected me to Michael Wolf, the Branch Chief of the National Employment Matrix, who was happy to talk to me about the statistics. It was incredibly nice to not have to find my way through an automated phone maze to find someone to talk to.
Mr. Wolf explained to me that the job statistics are based on two surveys. The first is the Occupational Employment Statistics survey. This survey goes to employers throughout the economy (including government employers). The responses are typically filled out by Human Resources people at each employer. The second is the Current Population Survey, which is designed to help catch people who are self-employed. For our purposes, the Current Population Survey should capture solo practitioners. The data from these surveys are combined to form the base year data on employment – that is, a picture of employment in the current year. Projections of future employment are then based on macroeconomic factors, tailored to each specific industry. Estimates of job openings factor in both expected new jobs and expected retirements.
The BLS job statistics therefore have two basic components – the base year data, which is based on surveys of current employment, and projections of future employment, which are based on applying certain projected macroeconomic assumptions onto the base year data. I think there are reasons to be cautious in interpreting both sets of data.
We need to be cautious in interpreting the base year data because the people included in the category “Lawyer” are people who are identified by the HR people filling out the survey as lawyers. I should emphasize that I am not criticizing the HR people who complete the survey – the survey is designed to be based on their responses, and the “Lawyer” category is not designed to be tied in any way to having a JD degree or admission to the bar. Some jobs typically held by lawyers have their own, separate category. Take a look at this table. Scroll down to category 23-1010, Lawyers and Judicial Law Clerks. Look over towards the right hand side, and you will see 218.8 – this is the 218,800 job openings projected for 2010-2020. This category, as its name implies, combines both Lawyers and Judicial Law Clerks, each of which has their own category immediately below. The number for category 23-1011, Lawyer, is right below, with 212,000 job openings projected for 2010-2020. Look right below that to 23-1020, Judges, Magistrates, and Other Judicial Workers. Many (though not all) of these people will have a J.D., but are not included in the Lawyer category. Nor are people with J.D.s who are legislators, other elected officials, law professors, community organizers, lobbyists, people in policy positions, people in business positions, people working in non-lawyer positions for NGOs, bloggers at legal tabloids, FBI agents, or any of the other positions that people with J.D.s might fill but not be described as a “Lawyer” by the person filling out the survey. I would imagine that most of the people who take “JD Preferred” or “Professional” jobs, as described in my first post, would not be categorized as lawyers in these surveys. As I mentioned in my first post, some of these other jobs are good, some are not. The crucial point here is the one I mentioned above – the “Lawyer” category is not synonymous with either having a J.D. or being a member of a bar.
I should emphasize that there is nothing whatsoever wrong with the BLS doing the survey this way. We just need to be careful in how we interpret their data. The one thing I would criticize the BLS for is this statement: “Employment of lawyers is expected to grow by 10 percent from 2010 to 2020, about as fast as the average for all occupations. Competition for jobs should continue to be strong because more students are graduating from law school each year than there are jobs available.” According to Mr. Wolf, this statement is based on a comparison between the projected number of Lawyer jobs and the number of law school graduates taken from Department of Education statistics. The overall conclusion may or may not be correct, but there is a mismatch between these two sets of data. Not all people who graduate from law school go on to get a job that would fit in the Lawyer category. (There was some discussion in the comment thread to my second post about whether the dramatic decine in law school enrollment has brought law school enrollment into line with what we might expect in the future job market.).
We need to be cautious in interpreting the predictions of job growth because they are just that – predictions. If you have any experience working with long-term projections, you will know one thing with certainty – they are pretty much always going to be wrong in one direction or the other. The assumed growth rate has a tremendous impact on the numbers, and actual reality is almost certain to differ from the assumed growth rate. To illustrate how projections can differ, take a look back at the 2002 survey data. In that year, the base year survey had 695,000 lawyer jobs. Based on an assumed 17% growth rate, the report projected 813,000 lawyer jobs in 2012. (Look at page 85 – under category 23-1011, you will see 813 under the 2012 column.). The 2010 base year data had 728,200 lawyer jobs, with 801,800 projected for 2020. (Look at row 23-1011 – the Judicial Clerk stand-alone category is new to 2010, but the 23-1011 Lawyer category is the same in both the 2002 and 2010 surveys.) In other words, in 2002 the BLS projections had more lawyer jobs for 2012 (813,000) than there in fact were in 2010 (728,200) and are projected for 2020 (801,800). The difference is largely in the assumed growth rate. The significant point here is that projections will typically differ from reality, often in major ways. Maybe lawyer jobs in 2020 will be significantly higher than the projected 801,800. Maybe they will be significantly lower. We won’t know for sure until we get there. The projections are useful in some ways, but they do not tell us with certainty about what the future will hold.
I have made three major points in this series of posts. First, I have argued that based on data on what graduates of the Widener-Harrisburg classes of 2010 and 2011 are doing now, many more recent law schools graduates are getting legal jobs than the nine-month job data would suggest. Second, I have explained why I am skeptical that the current anemic state of the job market is the result of structural, as opposed to economic, factors. Third, I have explained why we should be cautious in our interpretation of Bureau of Labor Statistics data on the legal job market.
I will close with two other points. First, I think that an important concrete step that we could take to improve graduate employment rates is to move the timing of the bar exam from the summer after graduation to the summer after the second year of law school. As I explained in my first post in this series, I think that bar timing is one part of the story of why nine-month data does not provide a full picture of graduate employment. Moving up the bar exam would help even in a more robust economy where more students are getting jobs within nine months of graduation, because graduates who landed jobs could start working sooner.
Second, nothing in this post suggests that we should not be concerned about student debt. At a few points in this series, I mentioned as an aside that I think it is wrong to focus on first year salaries when talking about law school affordability. I do think that – many entry level legal jobs (clerkships, ADA positions) have relatively low salaries but typically provide graduates with experience that can lead to higher paying jobs later. Entry level small firm jobs also often don’t pay a high salary to start, but the salary goes up over time. I also think that statements about graduates’ ability to service their debt that do not take cost of living into account paint with too broad a brush. A given salary goes a lot farther in, say, Harrisburg PA, than it does in New York City. This said, we could always use better and more thorough salary data, especially data that captures salary changes as a lawyer progresses through her career. I hope to include some data on salary in future iterations of my alumni study. Further, legal academics should be concerned about cost and student debt issues. Constant tuition increases of above the rate of inflation are inherently unsustainable. Finally, cost issues have come up at various points in the comments to this series. I plan to address cost issues in a future post.
[Update - Comments policy. I welcome pushback on any point that I made in any of the three posts in this series. I have a think skin, and I welcome hard questions. That said, I will delete any comment that doesn't actually engage in the subject of the post. If you think I am wrong on, say, the BLS data, explain why. Find another place on the internet to post things like "All law professors are lying scumbags."]