The argument behind school choice is relatively simple. If parents can choose to send their children to any school, they will send them to the best schools. The best schools will expand and enrollment will drop at the poor schools until they shape-up or shut down.
The theory is persuasive in arguing that school choice would improve education, on average. The most obvious reason is that administrators would have greater incentive to attract good teachers and get rid of the bad ones. But the greatest gains may come from innovations that we can't even imagine at this point. Who knows what could be done if schools were free to experiment with new ideas?
At the same time, the concern about fairness is understandable. Perhaps all of the kids with bad parents would be left behind at the bad schools. Even if this were true, the theory suggests that bad schools would eventually shut down and choice would lead to a better education for these kids as well. Thus, even though the gains may be greatest for kids of affluent parents, most would likely gain. Still, if a policy would help nearly everyone, it should probably be pursued even if it doesn't help everyone equally. There are other means of promoting equality, including giving larger subsidies to poor families. I need to read more of the critics of this policy--I'm sure there are other issues.
Caroline Hoxby's has a nice primer on this topic. She explains why a lot of programs that look like they satisfy true school choice do not. Most importantly, schools must be able to expand and contract flexibly. For example, a state with magnet schools is not expected to reap the benefits of choice even though they don't limit enrollment based on residence. Why? They have no scope for growth. Instead, if there is excess demand, the spots are rationed and lots of kids may be left behind in bad schools. Hoxby also reviews the empirical evidence that points to the possible gains of school choice.
The theory is persuasive in arguing that school choice would improve education, on average. The most obvious reason is that administrators would have greater incentive to attract good teachers and get rid of the bad ones. But the greatest gains may come from innovations that we can't even imagine at this point. Who knows what could be done if schools were free to experiment with new ideas?
At the same time, the concern about fairness is understandable. Perhaps all of the kids with bad parents would be left behind at the bad schools. Even if this were true, the theory suggests that bad schools would eventually shut down and choice would lead to a better education for these kids as well. Thus, even though the gains may be greatest for kids of affluent parents, most would likely gain. Still, if a policy would help nearly everyone, it should probably be pursued even if it doesn't help everyone equally. There are other means of promoting equality, including giving larger subsidies to poor families. I need to read more of the critics of this policy--I'm sure there are other issues.
Caroline Hoxby's has a nice primer on this topic. She explains why a lot of programs that look like they satisfy true school choice do not. Most importantly, schools must be able to expand and contract flexibly. For example, a state with magnet schools is not expected to reap the benefits of choice even though they don't limit enrollment based on residence. Why? They have no scope for growth. Instead, if there is excess demand, the spots are rationed and lots of kids may be left behind in bad schools. Hoxby also reviews the empirical evidence that points to the possible gains of school choice.
13 comments:
In your third paragraph, you suggest affluent parents == good parents. I don't think you meant this, but some people (who might not support school choice) would make this equivalence. Somehow poor parents aren't good/smart enough to want their kids to have a good education.
If parents aren't good/smart enough to want the best for their kids, then there's bigger problems. Why stop at schools? For example, bad parents must be bad about feeding their kids, too. We should have the government mandate a certain diet for children who have bad parents. Its easy to see who is a bad parent... they're all the poor ones.
BTW, Fey Accompli often writes about school choice...
There are reasons that afluent parents might send their kids to better schools even if they're not "better" in some sense. Transportation costs, for example. They may also be less informed. Or they may be busier (since they're more likely to be single and/or have both parents working.) Of course, some of the less favorable explanations probably play some role as well.
And to continue to play devil's advocate, there are problems with your food analogy. The issue is with screwing with a kids future "too much." A person can likely recover pretty easily from having a poor diet as a kid. Can the recover from a poor education as easily?
Will check out the link.
I'm testament to the idea one can recover from a poor primary education... if, of course, you grant me that this thing I scrounge together and call a life is, in fact, a recovery of some sort.
But seriously, nutrition doesn't matter, but learning the State capitols does? Actually, don't answer that... just grant me that there exists a subjective test of what "matters" for outcomes, i.e. its not an objective fact which things "screw[] with a kids future" and which don't. If you give me that, then its not clear government should be in the business of making these choices for people.
It's hard to say if you "recovered" in the sense that we don't have a counterfactual. Maybe you could have been even more successful.
I agree that "what matters" is ridden with value judgments. But what matters to future earnings, for example, is not.
It's a very sad thing that there are tons of disadvantaged kids out there who are constrained by the circumstances provided by their parents (more so than their privileged counterparts.) We should be concerned with trying to help these kids. Education seems like the most sensible way.
I agree that, in principle, everyone should be given a fair shake.
I don't think you can measure "given a fair shake" by outcomes, though. Besides Those Things That Shall Not Be Named and other inherent differences, there's differences in effort and, most importantly, differences in motivation.
Because you chose to enter a field that pays much less than your IQ could muster (e.g. investment bankers aren't smarter than you but they make a hell of a lot more than you ever will), are we to assume you weren't given a fair start, that your parents sucked?
I'm trying to think of better measures of a "fair shake".
Maybe social mobility would be a good measure. Mobility suggests people are given opportunities and they're outcomes aren't determined by initial conditions.
If you look at an inner city, how is the younger generation doing compared to the older generation (being careful to account for immigration)? If the difference is greater than the national average, let's say, we know, controlling for income, something good is happening there.
"I don't think you can measure 'given a fair shake' by outcomes, though."
Despite your valid concerns with naive comparisons, I disagree. You allude to the solution yourself. We can look at "shocks" to initial conditions and see how they affect the children throughout their lives to understand the causal impact of the initial condition on the outcome. For example, there's compelling evidence that reductions in parent's income *causes* reduced earnings for children. This is especially true for the poorest kids (maybe due to diminishing returns). See Page, Oreopolous, and Stevens:
http://www.chass.utoronto.ca/~oreo/research/intergenerational_displacement/details.htm
My first paper (joint with Page and Stevens) replicated this paper with data from the US to find similar results (although a smaller sample size led to more imprecise estimates.)
Ohhh, nice...
Of course, I'd be skeptical about the natural experiment, though.
In the Page/Stevens study, I don't think the shock to income due to firm closures is a clean identification (I think I'm using that term correctly).
People that are less motivated are going to select into poor companies that, in turn, more likely to shut down. They control for industry, but I think there's a correlation within industries between companies with low quality workers and companies that go out of business.
To be honest, without knowing the variation of income, 8% seems to be a surprising small effect. Also, do you know how many standard deviations 8% of income is?
Regarding the identification, you're right that it's not completely clean but it's pretty darn good. Not only are there controls for parent's industry and state, but they control for pre-displacement income as well. If you thought hard about matching controls for the displaced kids, matching the pre-displacement income would be at the top of your list of priorities.
Also, most of the skepticism over the magnitude of the estimates are that they're surprisingly large. First of all, consider the fact that the parents income is only reduced by 14%. Secondly, the impact on kids whose dad's were in the lowest income quartile prior to displacement had their adult incomes reduced by 19%.
Yeah, I'd have to argue that the bad employees were making more than their marginal product before the firm shut down... but, then again, maybe this was true. That a firm is paying its employees too much seems like a good reason for a firm to shut down, no?
Could you control by some other measure of quality (of the worker)? Like IQ or something?
Also, I guess I don't understand the results very much... but this is entirely because I spent about 4.5 seconds browsing the article. :-)
You wouldn't *have* to argue that. While I acknowledge that it's possible, I don't see why that would be the case.
Furthermore, they focus on fathers with at least 4 years of tenure with their firm before being displaced--so they were arguably fairly stable men who could hold down a job.
(And yes, IQ would be nice but isn't found in many data sets.)
I should let you have the last word, but I can't help myself...
I'm suggesting that there's a correlation between profitability and the quality of the employees at a firm. (This seems to be a no-brainer, but whatever.) Given such a correlation and the fact they control for pre-whatsit income, I'd have to argue these folks were making more than they were worth before the closure, i.e. pre-whatsit income isn't a good proxy for quality.
If this is true, there's more low quality men in your treatment than your control. Low quality men produce, probably, low quality kids via inherited qualities.
You could say I'm nit-picking and you can always critique an empirical study by listing the infinite things it didn't control for, but the inherited/"fair shake" question is central to this issue. From a policy point of view, we can't do much about inherited factors (and even if we could, we shouldn't... wouldn't want to encourage folks to have more dumb kids), but, otherwise, we might be able to something to even the playing field.
I think the 4-year thing makes my argument pretty weak...
... not only would I have to argue some folks were making more than their marginal product, I'd have to argue that that situation persisted for more than 4 years.
Prolly not.
Post a Comment