ProfitOfEducation is off for a bit of a summer break!
- "What Freakonomics did in raising our collective economic literacy, this book does for the economics of schooling." -Kate Walsh, President, National Council on Teacher Quality
Suppose that the Vergara decision and its aftermath were to lead to the entire country replacing the worst two percent of teachers with average teachers? It’s pretty clear that students who benefited from the teacher upgrade would end up with higher wages when they enter the work force? How much is that worth?
Let’s begin with the idea that really bad teachers do incredible damage. (This is a good place to stop to remind everyone that really great teachers do incredible good.) Suppose that the effect of having really bad teachers for a kid’s K-12 experience is that he never earns a penny for his entire life. And suppose that if the same kid had average teachers throughout he would earn an average income. This would seem to over-estimate the teacher effect, but let’s go with it.
Median income in the United States is around $30,000 per year. There are around 160 million workers. So $30,000 times 160 million times 2 percent is just about 100 billion a year. That’s real money, but it’s not a giant slice out of our huge economy. (It’s a bit over half a percent of GDP.) And I suspect this estimate is an upper bound.
The Vergara decision, assuming it’s upheld, will make it far easier to dismiss teachers in California. In California, it’s very, very hard to dismiss a teacher for cause. Harder than in most of the rest of the country. While it seems logical that removing tenure rules will lead to moving out really bad teachers, it’s not so clear that’s what will happen.
California isn’t noticeable “behind” other states in moving teaches out. The fact is, California is already a pretty average state when it comes to dismissing teachers for cause. The average school district dismissed for cause 5.8/10ths of one percent of teachers in 2011-12. The number for California was 5.3/10ths of a percent. Here’s a little map I made.
California just doesn’t stand out. (The “state” that really stands out is D.C.) If California’s “tough” tenure laws didn’t result in low dismissal rates, why should removing it make much difference?
By the way, if you check the data on the fraction of teachers not renewed “for any reason”–on the theory that the map above misses teachers removed “for cause” but not reported as such to make everyone’s life easy–well, the California non-renewal rate is above the national average.
Making it possible to dismiss really bad teachers is important. Just changing the law may be more of a political statement than a cause of real change.
If the Vergara decision holds, schools will be able to turf out extremely ineffective teachers. How much difference this makes depends on the answer to a couple of hard questions.
One piece of evidence on these questions is that Washington, D.C. has had a system for several years that pushes out the very bottom of the teacher pool. D.C. has been turfing about 2 percent of teachers each year. So perhaps California will as well. So far as I know there hasn’t been a full, published evaluation of the D.C. experiment.
At the same time, the rationale behind the Vergara case wasn’t that getting rid of a few bad apples would improve education overall. The argument was that the worst teachers were disproportionately assigned to teacher poor and minority students. (Exactly those students most in need of good teachers.) Whether Vergara leads to an overall improvement, it does seem likely to help some kids on the bottom of the educational barrel.
Next time: Dollar estimates of cutting out the worst teachers.
The Vergara decision is a smack in the chops to how California does public education. Is education in California really all that bad?
Yes. It’s not fair to call California’s K-12 system a total failure, but California certainly way under performs. California is a rich state with below average academic achievement. Here’s a picture of math test scores for California and the nation as a whole. Children of more highly educated parents score better everywhere, of course. The take-away of the chart is that California trails in the country for every level of parental education.
If you’re a regular reader, this is where you should be expecting me to say that the problem is that California under pays its teachers. Well California does under pay teachers, but not nearly so badly as the rest of the nation. I’ve made a chart showing teacher salaries relative to average salaries for college-educated, full-time workers. California outpays the nation on this metric–by a lot.
I’m not a great fan of making public policy by court mandate, at least not in something as complicated as running the public schools. But the fact is education in California isn’t working, and the folks running it need to be shaken up.
The Vergara is the stunning education event of the year. Maybe of the decade. What is to be said about the economics of the decision?
Suppose that tenure really disappears. Whether you think tenure is a good idea or bad, a promise that you can’t be fired and that after a few years on the job you’ll never be laid off is valuable to a worker.
Providing foosball tables to high-tech employees makes for happy workers, which means Silicon Valley employers can pay employees a little less than if the workplace were a game-free zone. The wage saving may or may not be worth more than the price of a foosball table (plus the cost if workers would rather foos than work). But there will be some wage saving. Economists call this a “compensating wage differential.”
In the same vein, the job security that comes with the existing tenure/seniority system has economic value to teachers. If teaching jobs become no more secure than other jobs, schools will have to raise teacher pay to make up for the compensating differential implicit in the existing high-job-security world.
How much extra will we have to pay teachers? I think the answer is pretty clearly “who knows?” The one number I’ve found, estimated by John Aboud and Orley Ashenfelter, is 40 years old. For whatever it may be worth today, Aboud and Ashenfelter put the teacher compensating wage differential in the 1 to 5 percent range.
Because the outside-teaching labor market is so much more uncertain now-a-days, it would be reasonable to guess that the compensating wage differential is higher today. How high is anyone’s guess.
It will surprise you not at all to hear that students from higher income families get higher SAT scores. Here’s a picture of the 2013 facts.
Unless you’ve recently been in the college hunt, you may not have a good idea of how many points should be thought of as making a “big difference” on the SAT. So I’ve add to the graph very approximate average SAT score of admitted students at Cal State Long Beach, a very good university in California’s second academic tier, and UCLA–one of the world’s better public universities. (Please don’t try to read the graph backwards. This doesn’t give you average family income at these schools.)
The SAT scores at Cal State Long Beach correspond to typical scores from a kid from a fairly well-off, middle class family. Maybe an upper middle-class family. The scores at UCLA are off the chart (literally off the chart).
The lesson lies at the left end of the chart. Typical scores from the lower end of the income distribution aren’t going to get a student into any competitive four-year university.
We all know that more education leads to better life as an adult: higher earnings, lower unemployment, etc. What isn’t so obvious is how we know this to be true. How do we know that kids who were headed toward good adult outcomes anyway don’t also choose to get more schooling. Maybe the relation between more school and better outcomes is a correlation without causation?
One part of the answer as to how we know comes from looking at changes in compulsory attendance laws. The idea is to look what happens when states change rules to force everyone in the state to stay in school until a certain age. While many students are not affected (because they would have stayed in school anyway), we know the effect on those who are forced to stay in school longer is causal. In the first half of the 20th century there were many such changes. Statistical evidence based on these changes has shown that more schooling really leads to much better adult outcomes.
A new analysis by Melvin Stephens Jr. and Dou-Yan Yang, “Compulsory Education and the Benefits of Schooling,” calls into question much of the existing work. Basically, they point out that other things were changing at the same time as compulsory education laws, and that these other changes happened differentially in different parts of the country. For example, at one point 40 percent of Southern children had hookworm. The eradication of hookworm led to higher adult wages. This mattered in the South much more than elsewhere. What Stephens and Yang show is that once you allow for the possibility that outcomes were on different trends in different regions, much of the conventional evidence based on changes in compulsory schooling disappears.
For example, using conventional techniques the authors estimate the effect of another year of schooling one wages to be 10.5 percent–which is a large number. The range of uncertainty (95 percent confidence interval) associated with this estimate is 8.3 to 12.3 percent. So we would seem to have pretty clear evidence that the education has a large effect.
The authors then repeat the analysis allowing for different underlying trends in different parts of the country. The estimate of the effect of another year of schooling drops to -0.3 percent, with a range from -5.8 percent to 1.6 percent. That’s a tight enough estimate to suggest increases in compulsory schooling haven’t done much good.
Now there’s lots of other evidence that education pays off. But this most recent work surely shakes what had been thought to be one of the stoutest scientific pillars.
School Accountability, Postsecondary Attainment and Earnings offers some decidedly mixed evidence on the long-run effects of school accountability programs. David Deming, Sarah Cohodes, Jennifer Jennings, and Christopher Jencks looked at Texas’ school accountability program in the 1990s. They then tracked students in schools under accountability pressure to watch outcomes in terms of college attendance and earnings as young adults. The research findings have something to make everyone happy–or unhappy.
If you look at high schools that were at risk of being rated “Low-Performing” you see real improvements. Students scored significantly higher in the high-stakes 10th grade math tests. And these students were more likely to attend a four-year college and more likely to graduate from a four-year college. What’s more, at age 25 these students had higher earnings. Earnings were around 1 percent higher, which is a lot for this kind of intervention. Poor and minority students especially benefited.
That’s the good news for testing and accountability fans.
If you look at high schools that were already at a higher level, between “Acceptable” and “Recognized,” the results were just the opposite. Notably, these schools increased the number of students classified as eligible for special ed in order to get them out of the pool that counted toward the accountability goals. It appears that students in these schools actually ended up worse off because of the accountability standards. Students with low baseline scores performed worse in these schools. Students with high baseline scores saw no change at all. The effect of accountability in these better performing schools was to reduced high school graduation rates. There was no overall effect on wages.
If you look at students with in higher performing schools who had low baseline scores, the effect of accountability was to lower their earnings at age 25 by $700 a year.
None of this proves what might happen with a differently designed accountability program. What it does prove is that accountability programs have real results, some good and some bad. We need to be wary of those who think that accountability is a magic word that will resolve our educational woes.
Going off-topic (K-12) today. Otherwise my head may explode. I spend a fair amount of time trying to convince my college students to write clearly and simply. Sometimes this means helping them unlearn what they’ve been taught in other parts of the university. Apparently, I should be hoping my students learned to write elsewhere.
Here’s an excerpt from “Supporting Veterans in the Classroom: Veterans bring unique challenges that require a focused faculty response,” which appeared in the May-June 2014 issue of Academe, the journal of the American Association of University Professors.
Military communication guidelines differ greatly from academic writing. Communication training in the military, especially in the army, focuses on concise, clear writing. Army guidelines can seem more designed for the telegraph than an expository essay, with single words sometimes replacing sentences. …Not surprisingly, some veterans find academic writing daunting.
If you happen to know a veteran heading back to school who might find “academic writing daunting,” would you be kind enough to send them to the closest economics department. I promise we will not penalize them for “concise, clear writing.