Welcome to the Profit of Education website. Continuing the conversation begun in the book Profit of Education, we discuss the latest economic evidence on education reform.

Gender hostilities, disparities among economics professors keep women from ascending ranks

My most recent post on the BROWN CENTER CHALKBOARD at the Brookings Institution.

Women receive 57 percent of the bachelor’s degrees in the United States, but only 52 percent of the doctorates. In looking at recent datafrom the National Center for Education Statistics, we find: Women receive only 23 percent of Ph.D.s in computer science; in mathematics and statistics, the share is 28 percent; in engineering, 23 percent. STEM fields are often a path to high-paying careers, and the presence of women in the professoriate can serve as a role model for both undergraduates and graduate students.

One might hope that the gender disparities in STEM fields are legacies of a bygone era, and that there is little active opposition to women in the professoriate today. Sadly, harassment of and hostility toward female academics continues—at least in economics, the area we know best. All this is illustrated by a major contretemps in economics set off by, of all things, an undergraduate senior thesis from Berkeley.

The story begins with Alice Wu’s senior thesis, “Gender Stereotyping in Academia: Evidence from Economics Job Market Rumors Forum.” Economics Job Market Rumors (EJMR) is a web forum notionally focused on discussing the job market for economics Ph.D.s, but more often a site for gossip and discussion by self-described internet trolls. Wu web-scraped text from a million-plus posts and then used machine-learning techniques to look for correlations. She found two main results: First, nine of the top 10 words predictive of a post about a woman are explicitly sexual references (The Washington Post even bowdlerized some of these terms); second, posts about women contain 43 percent fewer academic or professional terms and 192 percent more terms related to personal information or physical attributes. Observation of the EJMR site shows that these terms tend to be used in the context of a hostile or sexual comment about a female economist.

Discussions of economists and economics as an academic profession aren’t usually the stuff of headlines. The explosion came when Justin Wolfers wrote about Wu’s paper in the Upshot column of The New York Times, followed by a piece by Elisabeth Winkler in The Washington Post’s Wonkblog. Wolfers was blunt, writing that Wu’s paper “quantifies a workplace culture that appears to amount to outright hostility toward women in parts of the economics profession.” Winkler wrote, “Women economists are hardly surprised. …. Wu has now managed to quantify that misogyny using men’s own words.”

It should be understood that EJMR is, using the words of one prominent male economist, “a cesspool.” To be clear, there’s nothing subtle about EJMR: Postings go so far as to suggest sexual assault of named female economists. Anonymity on the internet permits the kind of behavior that would get one arrested if done in person. Despite this, some economists dismiss the problem as so-called locker room talk. As another prominent male economist has written, “I personally find the forum refreshing. There’s still hope for mankind when many of the posts written by a bunch of over-educated young social scientists illustrate a throwing off of the shackles of political correctness.”

Compared to other academic fields, economists have little sympathy for political correctness. What is surprising is how many economists still don’t think that gender-based hostility has any effect on the underrepresentation of women in the profession. This might be less concerning if the professoriate in economics were making significant progress toward gender balance, but this is not the case. Even though women slightly outnumber men in getting bachelor’s, men overwhelmingly outnumber women as economics concentrators. The figure below shows that about a third of first-year graduate students and new Ph.D.s in economics are women. These numbers have been essentially unchanged for two decades. So the input end of the pipeline to the professoriate is narrow, and not approaching gender parity.

Unfortunately, the pipeline is also leaky, and female Ph.D.s do not become assistant professors nor advance to tenured positions at the same rate as men, per the Committee on the Status of Women in the Economics Profession’s 2016 Annual Report. Women full professors still comprise less than 15 percent of full professors at Ph.D.-granting institutions (the green line in the chart below), implying a very small cohort of senior women who can act as role models and mentors for female grads and undergrads. This fraction has risen very slowly, increasing by about five points in 20 years. If the female proportion of full professors were to continue at this growth rate, it will reach the graduate-school level (30 percent) in about 2080. Approaching gender parity would probably require tackling the pedagogic and information barriers that appear to dissuade women from majoring in economics at the undergraduate level.

Pipeline for departments with doctoral programs: Percent of doctoral students and faculty who are women

Source: CSWEP 2016 Annual Report | Note: (T) and (U) indicate tenured and untenured, respectively.

The recent explosion following a memo widely viewed as hostile toward women in tech written by a Google engineer suggests that problems are not at all limited to the discipline of economics. We would all like to think that having the STEM professoriate reflect something like the gender balance in society will happen naturally, or that it will happen if only K-12 would somehow make math equally attractive to girls and boys, or that at least it will happen when STEM majors learn to attract more women. Unfortunately, female academics still have to deal with open hostility from some of their peers, and with a see-no-evil attitude from many more.

Posted in Uncategorized | Leave a comment

Leave a Reply

Your email address will not be published. Required fields are marked *

Education programs and (un)selective colleges

My most recent post on the BROWN CENTER CHALKBOARD at the Brookings Institution.

The pipeline for training America’s teachers has been the subject of much discussion, with questions and disagreements about both quantity and quality. A largely overlooked aspect of the pipeline question asks which colleges train teachers. Today’s part of that question asks: Are education majors trained at relatively less selective colleges, or is the distribution of education majors pretty much the same as the distribution of other majors? The answer turns out to be the former. Graduates with education majors are disproportionately found at schools where students have lower SAT scores. The difference in SAT scores is large enough to be thought-provoking, at the least. (For a recent look at SAT scores for teachers see Goldhaber and Walch in EducationNext.)

Let me explain how we went about measuring this. Every university reports to the federal government the number of degrees granted in different disciplines. (Some universities leave blanks in their reporting forms, but in general the coverage is quite good.) Universities also report the 25th percentile and 75th percentile of SAT scores of their student body. We averaged the two SAT numbers, by subject, and then added together the scores for math and reading. That gives us our measure for selectivity of each college. Note we’re not looking at SAT scores of education or non-education majors separately; we’re looking at how competitive the college is.

Next we take the average of each college’s combined SAT score, weighted according to the number of education majors it graduates. This gives us what you might think of as an “education selectivity” score, or the average SAT score of admits at colleges that generate the most education majors. The education selectivity score penciled out at 1060 out of 1600. We did the same for non-education majors, producing a “non-education selectivity” score, giving a result of 1116.Let me explain how we went about measuring this. Every university reports to the federal government the number of degrees granted in different disciplines. (Some universities leave blanks in their reporting forms, but in general the coverage is quite good.) Universities also report the 25th percentile and 75th percentile of SAT scores of their student body. We averaged the two SAT numbers, by subject, and then added together the scores for math and reading. That gives us our measure for selectivity of each college. Note we’re not looking at SAT scores of education or non-education majors separately; we’re looking at how competitive the college is.

Education majors are trained at less selective colleges. In other words, relatively few come from flagship universities while more come from regional colleges with lower bars for entry. But is the difference large enough to be meaningful? The education selectivity score corresponds to about the 60th percentile of college-bound seniors in reading and the 55th percentile in math. The numbers for non-education majors are the 70th and 64th percentiles respectively. The combined difference is nine or 10 percentiles. Thus the difference is not gigantic, but it’s certainly large enough to be meaningful.

While there is a noticeable difference, there is also a great deal of overlap between the colleges that train education majors and colleges that train other disciplines. That’s not very surprising. After all, even colleges with big teacher training programs train students in many other majors as well. The picture here, using bell curves as approximations, shows the selectivity distribution for education versus non-education. While the selectivity difference is noticeable, so is the overlap.

Selectivity distribution for education versus non-education

The data presented so far is for the most recent year for which data is available, 2015. Technically, the selectivity gap between education and non-education graduates appears to have risen over the last decade, but just by a couple of SAT points. For all practical purposes, as shown in the next figure, the selectivity difference is quite constant.

Difference in SAT score for education and non-education majors, by year

In the data, the “education” category presented here encompasses a number of different majors within the field. There is some difference in college selectivity across such majors, but not really all that much. The selectivity numbers for “general” education training and “specific subject areas” are about 25 SAT points higher than for “specific levels and methods” and “special education.” I will admit I was a little surprised to find the slightly low numbers for special education, given how much a teacher needs to know in this area. One should remember, though, that these are selectivity numbers for colleges—not for the individual students who select specific sub-majors.

Having documented that the college pipeline for training teachers runs through less selective colleges, how concerned should we be? We certainly don’t need all teachers to major in education at Harvard. (This happens to be a good thing, since Harvard doesn’t have an education major—although students can earn a Massachusetts middle/secondary school teaching certificate on top of their regular major.)

More-selective colleges do generally provide more academic opportunities than their less-selective brethren. They also often provide better networking opportunities. Both these elements contribute to developing future education leaders in education. (Do remember that alternative certification programs and master’s degrees are not included in any of these numbers.)

What’s more, the difference in selectivity is higher at the upper end of the selectivity distribution than it might appear based on the average numbers. The 90th percentile numbers for college selectivity are 1195 for education majors and 1380 for non-education disciplines. Those numbers correspond to SAT percentiles of 80th and 75th (education majors) versus 94th and 92nd (non-education majors), respectively. That’s starting to be a much larger difference.

So while the lower selectivity of colleges that train teachers is not so large as to be a cause of panic, it is large enough to think about whether this is the best way our system should run.

Isabel Steffens provided research assistance for this post.

Posted in Uncategorized | Leave a comment

Leave a Reply

Your email address will not be published. Required fields are marked *

Immigrant teachers play a critical role in American schools

My most recent post on the BROWN CENTER CHALKBOARD at the Brookings Institution.

America is engaged in an active discussion about reducing the flow of immigration. Perhaps surprisingly, immigration matters a lot for the supply of K-12 teachers. About 8 percent of American teachers were born abroad. If the supply of teachers were to be reduced by 8 percent, schools would be in deep trouble.

Some care is needed in this discussion. There are two different discussions ongoing with regard to immigration: undocumented people and people here legally. There are some teachers who are undocumented, but I suspect there are not many. In comparison, there are presumably many more teachers who immigrated to this country legally. There are also teachers here on temporary work visas, but again, I suspect not many. We are likely talking about a long-run issue, in which the supply of immigrant teachers and immigrants who might become teachers in the future teachers would be cut off, reducing the (already stretched) K-12 teacher pool.

What are the facts? The figure below gives the fraction of primary and secondary school teachers born outside the United States using data taken from IPUMS USA, composed of various years of the U.S. Census and the American Community Survey by the University of Minnesota. In 1950, shortly after the end of World War II, about 4 percent of teachers were foreign born. The fraction dropped to 2 percent by 1960. Since then, the foreign-born fraction has increased steadily, having just topped 8 percent in the most recent data. The fraction of foreign-born teachers is lower than the fraction of immigrants in the general population. However, the teacher fraction is only a little lower than the fraction of legal immigrants in the general population. In this sense, the number of foreign-born teachers is in rough proportion to the foreign-born population.

Immigrant teachers are not spread evenly across the United States. If we didn’t have teachers from abroad, Alabama and Vermont would be not much affected. On the other hand, California and Florida would have teacher supply disasters if there were no foreign-born teachers. The next picture shows the distribution of foreign-born teachers across the U.S.

A U.S. map indicating each state's percentage of immigrant teachers. (Source: Dick Startz/IPUMS)
A U.S. map indicating each state’s percentage of immigrant teachers. (Source: Dick Startz/IPUMS USA, University of Minnesota)

To a considerable extent, there are high numbers of foreign-born teachers in states that have high numbers of foreign-born residents. (Interestingly, Alaska has an unusually high number of foreign-born teachers relative to its overall immigrant population.)

If you were to guess where foreign-born teachers migrated from, you would probably guess right. The largest single “donor country” is Mexico, followed by China. The next figure shows the major homelands of teacher immigrants to the U.S.

It isn’t surprising that many teachers are from China, since it has the world’s largest population. It might be surprising that so many teachers are from the lightly populated West Indies, although of course many of the islands have strong historical ties to the U.S.

The quick summary is that teachers who have migrated to the U.S. are a vital source of teachers. That’s especially true in the parts of the country that have a large immigrant population. None of the proposed changes to immigration policy is likely to cause any short-run crises because few of the changes would affect legal immigrants who are already teachers. (The future for undocumented teachers is much less clear.) But if we were to severely restrict immigration as some have proposed, we might be facing a critical teacher shortage a few years down the road.

Posted in Uncategorized | 3 Comments

3 Responses to Immigrant teachers play a critical role in American schools

Leave a Reply

Your email address will not be published. Required fields are marked *

Sealing the border could block one of America’s crucial exports: Education

My most recent post on the BROWN CENTER CHALKBOARD at the Brookings Institution.

The last election made clear that Americans want to improve our balance of trade in order to protect jobs. Something not often realized is that education—particularly higher education—is a major American export. If new border controls prevent the entry of foreign students, or simply makes them feel unwelcome so they go elsewhere, American jobs and American students pay the price.

If you are not an economist, the idea that America’s education of foreign students is an export may seem strange. What makes a service like education an export is that foreigners pay U.S. institutions, so money flows into the U.S. from abroad. Education of a French student in the United States is an export in exactly the same way that the visit of a Chinese tourist to Disneyland is an export, or the provision of stock brokering services on the New York Stock Exchange to a German financial company is an export. When we provide a service that leads to foreigners sending money into the U.S., that’s an export with exactly the same economic effects as when we sell soybeans or coal abroad.

The value of education exports is not a small number. According to the government’s Bureau of Economic Analysis, education accounts for 5 percent of the entire national export sector. (Most of the education spending is presumably in higher ed.) I will explain in a moment why this is an underestimate. First, here’s a picture of the official value of exports and imports in education.

Exports and imports of education

Source: BEA.gov; GDP deflator/Dick Startz

Exports in education topped $35 billion in 2015. Not only is the blue (export) line much higher than the green (import) line, the gap in our favor is growing rapidly. By the way, this is one of the few sectors where the trade gap is in America’s favor.

As I wrote above, there is good reason to believe that the official numbers understate the importance of education exports. The numbers reported measure direct spending on education—tuition and the like. They do not include spending on room and board, used cars, pizza shops, or anything else college students spend their money on. Research by John Douglass, Richard Edelstein and Cecile Hoareau from 2011, “US Higher Education as an Export,” reports that the estimated economic impact of higher education, taking out financial aid and adding in cost-of-living expenses, raises the direct economic impact by 43 percent. (The extra 43 percent measures direct spending only, excluding indirect effects of job creation and the like.) Unless that ratio of extract spending has changed, the total value of education exports today tops $50 billion.

There is a good argument to be made that higher education is one of our most successful export sectors. We’re good at higher education, and the rest of the world flocks to our shores and pays for the chance to study in America. Just under 5 percent of college students in the United States are visiting from abroad. As argued above, this is critical for exports. It’s economically critical for something else: Foreigners subsidize American students. Foreign students are generally eligible for very little foreign aid. They pay full freight. Especially in public universities, such as the one at which I teach, those dollars are critical in replacing the money that used to be allocated by state legislatures.

Blockading our borders risks driving away students and their dollars in two ways. The first is straightforward: Effectively excluding students from Muslim countries would mean we lose a lot of customers. The United States had 60,000 Saudi Arabian students in 2014-15. While well below the number from China and Hong Kong, Saudi Arabia was the fourth-largest source of foreign students, just edged out for third by South Korea. Other predominantly Muslim countries also send us a considerable number of students. Taken together, Indonesia, Turkey, Iran, Kuwait, and Pakistan send roughly another 45,000 students.

The second way that building barriers hurts us is by making students from all countries feel unwelcome. Ignoring the issue of building goodwill toward the United States, the fact is that universities around the world would love to out-market us. Universities in the United Kingdom, Canada, and Australia teach in English and already have a higher ratio of foreign to domestic students than does the United States. (New Zealand and Ireland also have higher ratios than we do, but as exporting countries are too small to be major competitors.) In addition, many universities in Europe have switched part or all of their instruction to English to be more customer-friendly to foreign students.

Right now, the United States is the leading supplier of slots for international students, with about a fourth of the world market. You can see in the following chart that the U.K. and Australia combined trail us only by a little bit. This is a market we lead, but it is also a market we could lose.

Major competitors for international college students
Source: OECD data; author’s calculations/Dick Startz

Education as an economic export is likely not the first concern about how we operate at the border. Still, blocking American exports and moving the associated jobs overseas doesn’t seem in step with our national goals.

Posted in Uncategorized | Leave a comment

Leave a Reply

Your email address will not be published. Required fields are marked *

Teacher prep ratings: 2016 edition

My most recent post on the BROWN CENTER CHALKBOARD at the Brookings Institution.

Today, the National Council on Teacher Quality (NCTQ) released its latest rating of elementary school teacher preparation programs. NCTQ’s review of 875 traditional undergraduate programs will make some folks at some schools happy (those who got high ratings) and some unhappy (programs with low grades). Many schools don’t like NCTQ’s review; some schools disagree with NCTQ’s process and some may disagree with the criteria used. And then there are those schools who just don’t want to be held to any external standard at all. However, there are some findings in the NCTQ report that inform us about the overall state of teacher education in our nation. Regardless of whether you agree with the rating of a particular school, these overall findings are important. Maybe they’re even shocking. (Full disclosure: my daughter worked for NCTQ one summer as an intern.)

Everyone understands that a key element in learning to be a teacher is actually getting in the classroom and teaching. Book learning does matter (more on that below), but every teacher I know talks about the importance of learning-by-doing. A critical question, then, is whether that learning begins on the first day of the first job or whether the teacher gets a head start through a quality student teaching experience.

Shockingly, most teacher ed programs don’t do a very good job of arranging high-quality student teaching experiences. The criteria that NCTQ uses to rate schools in this area are very hard to disagree with. NCTQ emphasizes two criteria: the extent to which student teaching mentors observe their mentees and provide written feedback, and the extent of control that the teacher training program has on picking good mentors. Despite the fact that NCTQ sets a pretty low bar for getting a good grade on arrangements for student teaching, very few schools come anywhere close to what we would hope for. To quote NCTQ:

Just 3 percent earn an A (and 2 percent a B) by making an effort to match student teachers with strong cooperating teachers and requiring program supervisors to provide student teachers with at least four observations incorporating documented feedback (five is the minimum shown by research to be effective).

Does that sound bad? NCTQ also finds:

Only about seven percent of programs collect any meaningful information on each cooperating teachers’ skills, and only about one percent screen cooperating teachers for both their mentorship and effectiveness as a teacher.

These findings make me wonder just how many student teaching experiences consist of little more than throwing student teachers in front of a class with little or no useful guidance. While I’m sure most student teachers do receive informal feedback, I share NCTQ’s position that more formal feedback is also important.

Let’s turn to the book learning question. Here, NCTQ has findings that are quite clear…but that are also controversial. Education schools are notorious for their low academic standards. Of course some education schools do have perfectly high academic standards, but even these schools often find their reputations tarred by unfair association with weaker peers. NCTQ rates schools based on the selection of undergraduates into teacher prep programs, using a criterion that entrants should have measures of academic ability that come from the top half of college applicants. This does not mean every entering student should be from the top half, but at least on average entering students should be from the top half.

Selectivity criteria are debatable for two reasons. Some people don’t think that pure academic ability is all that important for elementary school teachers. Others are concerned with the effect that selectivity standards have on diversity. I have little sympathy for the former argument. Surely academic ability is not the only ability that an elementary teacher should have. A high Emotional-Quotient is also necessary. Just as surely, though, academic ability matters.

The arguments about diversity are a little different. One argument is that standard measures of academic ability, especially standardized tests, are biased against members of some minority groups. To the extent that the tests measure ability inaccurately, the concern is legitimate. A different argument is that schools need a diverse set of teachers, and in admitting prospective teachers or hiring teacher applicants, preparation programs and schools should prioritize workforce diversity, even if it comes at the expense of compromising performance or admissions standards for some minorities who are in short supply.

Whichever side of the debate you favor, NCTQ’s findings are very, very interesting. A bit over half of the 875 programs studied do pretty well on NCTQ’s selectivity standard. While only 6 percent of teacher training programs get an A or B rating based on admitting students with high GPAs or test scores, another 50 percent of programs get a high grade because they draw students who are above average based on their own institution’s admission standards—even though their own institution’s admissions are not high by national standards. This is good, and NCTQ points out that progress has been made over the last couple of years. However, to the extent that teacher prep programs are concentrated in low selectivity colleges, the fact that the ed schools are selective within their institution’s not-too-selective pool isn’t all that reassuring. In other words, those ed schools are doing a good job on the aspects of selectivity that they can control, but we might still worry about whether nationally too few teachers come from more selective schools.

Here’s the very good news: Many programs are both selective and diverse! 113 of the 875 programs earned an A+ for this accomplishment. NCTQ writes:

Programs can earn this A+ by being more diverse than the program’s institution as a whole or more diverse than their state’s teacher workforce. These 113 programs prove that teacher prep programs can be both selective and diverse.

So, a suggestion for ed school deans: If you’re worried about how to become more selective without losing diversity, get some advice from one of the 113 peer institutions who’ve demonstrated that you can pull this off.

Posted in Uncategorized | Leave a comment

Leave a Reply

Your email address will not be published. Required fields are marked *

An Education Thanksgiving

My most recent post on the BROWN CENTER CHALKBOARD at the Brookings Institution.

This time of year we ought to reflect on the things that are so right in American education. News stories—and, yes, blog posts—tend to focus on the negative: what’s gone wrong and what needs to be fixed. While working for the better is the American way, it’s also important to recognize that which is going well. It is the month for Thanksgiving.

So today’s post offers solely data on good news: not perfect news, not the-job-is-all-done news—but good news.


Over the last fifteen years, high school graduation rates have risen. The vast majority of adults have a high school diploma. White, Black, and Hispanic graduation rates have all moved up (and notably, the increases have been larger among minorities, which has slightly narrowed gaps).

High school completion rates, age 25 and over


Source: Digest of Education Statistics 2015 Table 104.10


Over the last fifteen years, a noticeably increased portion of the adult population has earned a college degree. This is true among adults of all races, though it’s not clear gaps have made any progress during this time.

Bachelor’s degree completion rates, age 25 and over


Source: Digest of Education Statistics 2015 Table 104.10


There’s a tendency to focus on degree completion at the high school and college level. Community colleges and associate degrees matter too, and they are also up. Here are the numbers by sex. If you’re wondering, the numbers have also risen for all racial and ethnic groups.

Associate degree completion rates, ages 25-29


Source: Digest of Education Statistics 2015 Table 104.65


Violence against students is way, way down from the 1990s and seems to be continuing to slowly decline. Keeping our kids safe at school is surely something to be thankful for.

Rate of violent incidents in school per 1,000 students age 12-18

capture4Source: Digest of Education Statistics 2015 Table 228.20

School is also a less scary place than it once was—reductions of more than 50 percent for both males and females.

Percentage of students age 12-18 who reported being afraid of attack at school


Source: Digest of Education Statistics 2015 Table 230.70


Alcohol use among high school students is declining. I think this is a good thing, although I do wonder what the French would think of the notion that most high school students had not had a glass of wine in the last month.

Percentage of students in grades 9-12 who report using alcohol at least 1 day during the previous 30

capture6Source: Digest of Education Statistics 2015 Table 232.10


Perhaps the most telling good news comes from the upward trend in the EdNext/PEPG poll in which the public is asked to grade schools. For their own local schools, presumably the ones the public knows best, a majority now assign a grade of A or B. We can’t pin down what exactly is making the public more satisfied, but things appear to be moving in the right direction.

Fraction of public giving schools grades of A or B: Local schools and nation as a whole

capture7Source: EducationNext polls

In this month for Thanksgiving we do have much to be thankful for in our kids’ schools. But you will see that I have left out something, something I don’t have a picture for, something which is the most important thing to be thankful for.

Our schools are filled with a great many very good teachers and very many great teachers (and principals and staff and so many parent volunteers too). Thanks gals and guys.

Posted in Uncategorized | Leave a comment

Leave a Reply

Your email address will not be published. Required fields are marked *

An underfunded disaster awaiting taxpayers? Pension puzzles part II

My most recent post on the BROWN CENTER CHALKBOARD at the Brookings Institution.

In the first part of “Pension Puzzles” I wrote “Everyone knows that taxpayers are getting stuck with a huge bill for teacher pensions and that the money paying for those pensions is money not going into direct education support. The thing everyone knows doesn’t come close to describing how bad the situation really is.”

You probably have a sense that teacher pensions have gotten very expensive. That is true, but things are worse than you may think for two reasons: (1) Much of the large amount that we pay into pension systems is currently allocated to make up for failures to adequately fund pension systems in the past—such spending doesn’t support current education at all; and, (2) The accounting rules for public pension systems are bent, so the real deficit is far larger than it appears to be.

Are we really spending that much more than we used to on teacher pensions? Robert Costrell, of the University of Arkansas, has put together the numbers that show just how much spending has risen. Costrell’s first graph shows employer contributions for teachers’ retirement benefits (in red) and for private sector managers and professionals (in blue) from 2004 through the fall of 2015.


The first thing that strikes you may be that the red teacher line is substantially higher than the blue private-sector line. Remember though that the height of the lines are determined by employer contributions divided by salary. The denominator (the salary) is generally higher in the private sector (as I’ve written about before), which makes the blue line lower than it would be if we were looking at employer contributions for equal salary levels. So the relative heights of the lines don’t necessarily mean that pension payments are higher (in actual dollar value) for teachers. However, Costrell’s work does make two things clear:

  1. The mix of pension spending to current salary is much higher for teachers than it is for the private sector. This suggests, but for reasons discussed below doesn’t prove, that the mix between current and deferred compensation for teachers is wrong.
  2. Spending on pensions versus current compensation has skyrocketed for teachers while remaining more or less unchanged in the private sector.

Does this pension spending amount to a big deal in the overall scheme of education budgets?


Costrell has calculated that school contributions to retirement plans (for all employees, not just teachers) has doubled in inflation-adjusted terms over the last decade. Here’s his picture:

gs_20161012_teachers_benefitsCostrell estimates that schools now devote almost 9 percent of all current spending to pension contributions.

Here’s another way to put the numbers. This figure shows we now spend nearly $1,100 per student on retirement benefits. The average public school student teacher ratio is 16 to 1. So we are spending about $17,000 per year per teacher in pension contributions. Do you suppose most teachers might prefer a little less in the way of pension contributions and a little more in the way of salary? The mix of current versus deferred compensation for teachers seems weighted too heavily toward pension and not enough toward current salary.

But here’s the problem. In many states, much of the spending on teacher pensions isn’t actually going to pay for pensions for current teachers. Instead, the payments into pension funds are going to make up for the failure to adequately fund those pensions in past years. (The inadequate funding is a problem for state and local pensions generally; it’s not limited to teacher pensions.)

How did we get into this situation? When politicians sit down to negotiate teacher compensation they face a choice: Pay good salaries now and raise taxes to fund them, or pay modest salaries now with the promise of big pensions later and figure that someone else will be in office when the bill comes due. Well, you could probably guess what happened in most states, and now the bill is coming due. The National Council on Teacher Quality writes,

In 2014 teacher pension systems had a total of a half trillion dollars in unfunded liabilities—a debt load that climbed more than $100 billion in just the last two years. Across the states, an average of 70 cents of every dollar contributed to state teacher pension systems goes toward paying off the ever-increasing pension debt, not to future teacher benefits (p. iii).

While we are spending a huge amount to fund teacher pensions, most of that spending doesn’t go to attracting the best teachers. It’s paying off past debts. (For a more detailed discussion, see Chad Aldeman’s “The Pension Pac-Man.”)

If all this sounds bad, that just means you don’t fully appreciate the situation. It’s not “bad,” it’s really, really bad. It turns out that public pension systems play by a different set of accounting rules than those allowed in the private sector, and this allows the pension funds to greatly overstate their reserves. (Again this is all local and state pensions, not just teacher pensions.)

The issue at hand is the interest rate that the public pension systems are allowed to assume in calculating how much money they need to fund future pension liabilities. The fundamental rule of finance theory is that if you guarantee a future payment—as pension funds do—then you need to figure your investment returns based on investments guaranteed not to lose money. In other words, the interest rate you should use in calculations is the so-called “safe rate of return.” Safe assets don’t pay as much as risky assets. When we invest for our own retirement, most of us put some of our money in stocks. We realize that if the stock market takes a turn for the worse we’ll be hurt, but the odds are good that we’re better off with some risk in the investment. The issue is that when we invest for ourselves we’ve decided to risk a smaller retirement income. Pension funds have promised a specified payment level—therefore they should be forecasting earnings only at the safe rate of return.

The way the law is written, public pension funds are allowed to assume they will earn at the higher, risky rate. Nobel laureate William Sharpe (Bill invented much of modern finance theory) explains that public pension funds can be valued in two ways. The “actuarial approach” uses the (legally permissible) higher, risky interest rate. The “market value approach” uses the (correct) lower, safe rate of interest. For example, CalPERS (the giant California public pension system for most workers other than teachers) assumes a 7.5 percent return on assets (actuarial approach) when deciding how much money it needs to set aside but assumes a much lower 2.56 percent return (“market value approach”) when negotiating buyouts from the system.

All this makes an unbelievably large difference. Suppose you wanted to guarantee a $20,000 a year pension with 30 years of payments. If you believe that you are guaranteed to earn 7.5 percent a year then you need to set aside $236,208. But if you are only sure to earn 2.56 percent a year then the required set aside rises to $415,276. In this example, a pension fund using the actuarial method has set aside only a bit over half of what it really needs.

If that sounds really, really bad…you’re still missing something. Most teacher pension plans provide significant inflation protection. That means that the safe investment return has to include an offset for future inflation. At the moment, the return on a 30-year Treasury bond that includes an inflation offset is only 0.67 percent. Using that as a truly safe rate of return, the pension fund in our example should really set aside $541,908 (2.3 timesthe original set aside value).

In summary, teacher pensions aren’t great for most teachers, and the system is an underfunded disaster waiting to happen for taxpayers. For teachers, we should reform the system to something equitable for all teachers and fairly funded going forward. It’s harder to say what we should do about the gross underfunding of the already incurred pension debt. But as the aphorism attributed to Will Rogers says,

If you find yourself in a hole, stop digging.

Posted in Uncategorized | Leave a comment

Leave a Reply

Your email address will not be published. Required fields are marked *

Are teacher pensions really that high? Pension puzzles part I

My most recent post on the BROWN CENTER CHALKBOARD at the Brookings Institution.

Everyone knows that teachers get great pensions and that those great pensions make up for the less-than-stellar salaries teachers earn. Everyone knows that taxpayers are getting stuck with a huge bill for said pensions and that the money paying for those pensions is money not going into direct education support. Well, the first thing everyone knows isn’t true and the second thing everyone knows doesn’t come close to describing how bad the situation really is.

The first thing you should know is that really good data on pensions that’s comparable for teachers and non-teachers is tough to come by. The best source for teacher pension data is probably TeacherPensions.org, a  group that’s trying to straighten out the teacher pension mess. I’ll show you some of their data together with data from other sources for non-teachers.Teacher pensions are a mess. Understanding what’s going on with teacher pensions is messy as heck. However, some progress can be made if we divide the topic into two pieces: Are teacher pensions really high? (Not so much.) Are teacher pensions really expensive? (Not only expensive, they’re even more expensive than it appears on the surface.) I’ll tackle the first question today, then focus on the second in my next Chalkboard post.

If we want to ask whether teacher pensions are high or not, we need to begin by asking “Compared to what?” The labor market for teachers is a subset of the more general labor market for college-educated workers. So let me give you two background numbers computed from the Current Population Survey. Among college-educated workers, only 57 percent report that they have an employer-provided pension plan. Almost all public schools provide a pension plan, so in terms of availability teachers are better off than other workers. Except, as we’ll see in a minute, many teachers never become eligible for the offered plan, so the difference isn’t as great as it may seem.

How large are pensions for college-educated workers in general? The number I’ll use as an average benchmark, computed from the Current Population Survey, is $33,281 a year. But this is more of a factoid than a hard number. It’s the average pension for college-educated workers who report being retired and who report pension income—except that I exclude IRA’s, 401(k)’s and the like from reported pension income because the data doesn’t separate out whether these sources include employer contributions or are based in some part on the retiree’s own savings. (A retiree’s own savings aren’t part of the teacher pension numbers, so they should be left out.) In other words, in order to exclude private savings I had to also exclude in 401(k)’s I had to exclude employer contributions as well. So $33,281 is somewhat lower than the truth.


How does $33k compare to teacher pensions earned by recent retirees? Here’s a map, based on data from TeacherPensions.org, that shows the average pension for teachers who retired in the last decade.

Average teacher pensions

In 35 of the states, teacher pensions are lower than our national average of $33,281 for all college-educated workers. In the middle-ranked state, the pension is only $21,355. Even though the comparison numbers are quite rough, what we have suggests that teacher pensions are not out of line with pensions of similarly educated workers. Maybe they’re even a bit on the low side. (Remember though, that the comparison numbers for college-educated retirees may include pensions from multiple jobs. Although, the comparison numbers don’t include anything from employer provided 401(k)’s. Many teachers will also have pensions from non-teaching jobs because they don’t teach their entire career. Since these “other job” pensions aren’t included, the comparison isn’t perfectly “apples-to-apples.”)

On the other hand, you can see that the states marked in dark blue have pensions paying over $60,000 a year! Oh, you can’t see it…it’s too tiny; that’s because the only place with such high teacher pensions is Washington, D.C. The average pension paid in D.C. is fully a third higher than the pension in the second-ranked state (Connecticut). Except, while D.C. is high, it may not be quite so crazy high as it may sound.


Here’s the next puzzling piece about teacher pensions: In 15 states, teachers are not eligible for social security. One of those “states” is D.C. So part of the explanation for high pensions in D.C. and those other states is that the high pensions are making up for the absence of social security payments. Country-wide, about 40 percent of teachers are left out of the social security system.

Is being left out of social security a big deal? A person who has earned $50,000 a year for the last 20 years would expect, roughly, $25,000 a year in social security benefits. So the absence of social security in D.C., and other states, makes a huge difference in thinking about pensions. (But remember that neither school districts nor teachers in those states have to contribute their share of social security tax, which is around six percent of salary each.) A Brookings study by William Gale, Sarah Holmes, and David John explains the reasons why it would be better to bring all teachers into the social security system. But for thinking about current teacher pensions, that’s not where we are now.

If you think that not being covered by social security is weird, since almost everyone else in the country is covered, let me make it a little weirder. Suppose a teacher has worked part of her career in the covered sector, paying social security taxes and earning credit toward social security on retirement. Now that teacher takes a job in a district which doesn’t participate in social security. It turns out that the teacher loses part of the social security benefits she earned and that she and her employer paid for in the covered sector. A study by Alan Gustman, Thomas Steinmeier, and Nahid Tabatabai finds that teachers in this situation lose about 20 percent of the social security benefits they had earned.


What we have to this point is the idea that teacher pensions might be roughly comparable to other pensions, although there is clearly enormous variability. Except, about half of teachers don’t get teacher pensions at all. As an extreme example, “high-pension” D.C. estimates that four out of five beginning teachers won’t get a cent in pension pay.

Two issues affect whether a teacher gets a pension and whether that pension is worth much. Both are related to the fact that many teachers have relatively short careers in education. In many pension systems, you have to participate for a minimum number of years for the pension to “vest,” i.e. for you to get a right to your employers’ pension contributions. In general, federal law requires private employers to either vest fully after five years or to begin partial vesting earlier, in which case full vesting can stretch to seven years. However, public sector pension plans are allowed longer waits.About a quarter of such plans require 10 years or more for full vesting. So the first short career issue is that many teachers leave teaching before being vested in their pension.

There is enormous variation across states in how many teachers end up with a pension. With a warning that the map isn’t perfect because some states have changed plans, here’s a picture again based on TeacherPensions.org data.


There’s a lot of light blue on that map. That’s a lot of states where many teachers walk away with no benefits at all.

The second short career issue is that many teacher pensions are rigged up to give disproportionately high payments to very long service teachers at the expense of quite low payments to teachers with “short” careers. Sometimes “short” means a couple of decades. Chad Aldeman, Daniel Fuchs, and Leslie Kan have looked at how the value of Illinois’ current pension system varies depending on how long a teacher works. Chad sent me their data.

 captureThat’s right, a teacher who retires after 25 years of teaching loses money. How can that be? Teachers, like most of us, make contributions to their pension as does their employer. In Illinois, the system is set up so that the value of a pension for a teacher with 25 years on the job is less than the value of that teacher’s contributions plus accumulated interest.

So what’s the bottom line? Some teacher pensions are indeed very generous, but many teachers end up with only a small pension—or no pension at all. This is a screwy way to run a retirement system, and is almost certainly not an effective way to spend taxpayer money to attract great people into the profession.

And on that issue of what this means from the taxpayers’ point of view, stay tuned to my next Chalkboard post for the bad news.

Posted in Uncategorized | Leave a comment

Leave a Reply

Your email address will not be published. Required fields are marked *

Making better use of data on free school lunches

My most recent post on the BROWN CENTER CHALKBOARD at the Brookings Institution.

A student carries her lunch tray at Salusbury Primary School

Feeding hungry children is the primary purpose of America’s free (or reduced price) lunch program. Just over half of public school students are eligible for the program, ranging from nearly three-quarters of Mississippi students to just over one-quarter of students in New Hampshire. However, there is also a secondary use of the free lunch program: eligibility for free lunch is often the only way for states, school districts, and education researchers to identify low-income students.

Schools don’t generally have records of exact family income—it being mostly none of their business—so eligibility for free lunch status is usually the best proxy measurement anyone has to identify low-income students. But the real educational issue is less whether a family is having a bad financial year right at the moment and more what the family’s finances look like over the long run. So it makes sense to look at more than current-year free lunch eligibility. Katherine Michelmore and Susan Dynarski, in a paper titled “The Gap within the Gap,” demonstrate that using multiple year measures of free lunch status better identifies long-term, low-income students. Bottom line: students who are on free lunch status for several years come from lower income families than students who just happen to be eligible for free lunch this year. I’ll explain here the intuition of what’s going on (having benefitted from hearing Sue talk about the paper) in a rather wonky way. Below, I’ll explain why this should be a real action item for those who distribute money and resources among schools.

In the figure below I show (in blue) a pattern of fluctuating income that might be representative of a particular student’s family over the course of primary school. The shape is the same in the left and right panels. The red line shows the cutoff for free lunch. So in the years in which the blue line dips below the red, the student gets subsidized lunch. You will see that the student depicted on the left is eligible for free lunch only once. In contrast, the student on the right is eligible in most years. (In the real world, thecutoffs are 130 percent of the federal poverty line for free lunch and 185 percent of the poverty line to be eligible for a reduced price. The exact numbers depend on family size and are slightly different for Alaska and Hawaii, but for a mother with one child free lunch kicks in at a yearly income of about $21,000 and the reduced-price line is about $30,000.)

Fluctuating income under two scenarios

As a guess, the fact that the student at the right has more years eligible for free lunch was probably the second thing you saw in the figure. You likely noticed first that the wiggly blue line on the right is lower than the wiggly blue line on the left. And these two observations exactly explain the intuition of why students who are more frequently eligible for free lunch generally come from families with lower long-term income. If a high income family has a one-year dip in income, the odds are their income is still too high to qualify for free lunch. In contrast a family of more modest means is often near the eligibility border, so a dip down makes their child eligible. And a very low income family’s child is probably always eligible.

Michelmore and Dynarski look at a massive administrative data set from Michigan and show that the pattern described above is what we see in the real world. (Michelmore and Dynarski then go on to show that the number of years eligible for free lunch is a better predictor of educational outcomes than you can get using a single year.) In the next figure, I’ll show you what the relation between income and years eligible for free lunch looks like in nationally representative data taken from the Early Childhood Longitudinal Study.Fraction of students on free lunch vs income

Students and their families were surveyed five times from kindergarten through eighth grade, and one of the items surveyed was whether their children were eligible for free (or reduced-price) lunches. About half of students were never eligible. Other students were eligible between once and every year of the survey. The big income drop-off is, unsurprisingly, between the “never eligible” and everyone else (where income is measured as average income for a family over all surveyed years). But—just as in the first picture above—there is a large difference in average incomes between those eligible one year and those eligible every year. The gap is close to a factor of two. So if you want to identify students who come from very low-income backgrounds, look for those who are regularly on the free lunch program.

I promised an action-item suggestion above. There are, in fact, two.

Very frequently the only income data available to education researchers is whether a student is eligible for subsidized lunch. Researchers who are looking for a proxy for low socio-economic status should use all years, not just the current year. (Wonk side note: Matt Chingos points out that Congress has changed the technical requirements for establishing free lunch eligibility in ways that are good for kids but which may make FRPL eligibility much less useful for researchers in the future.)

The more important action item is for state and school district administrators who direct money in part toward low-income districts or low-income schools. You should strongly consider adjusting your formulas to put heavier weight on those districts and schools in which many students have a long history of being on free lunch. That’s a more effective way of getting money to the lowest of the low income.

Garrison Schlauch provided research assistance for this post.

Posted in Uncategorized | 1 Comment

One Response to Making better use of data on free school lunches

Leave a Reply

Your email address will not be published. Required fields are marked *

The Kathleen Dunn Show

Posted in Uncategorized | Leave a comment

Leave a Reply

Your email address will not be published. Required fields are marked *