Class size and student performance: major pitfalls in measurement

The literature on the impact of class size on student performance is largely inconclusive. Yet, there is still a public perception that size matters. Is the econometrician wrong? Miguel Urquiola and Eric Verhoogen argue that there is a subtle issue with class size that can seriously affect estimates.

Many studies exploit the following discontinuity. Jurisdictions mandate that class sizes cannot be larger than some number. For schools, this means when the number of students overtake a multiple of this number a new class needs to be opened. The consequence is a sudden decrease in the average number of students in classes, because the number of students is close to continuous, the number of classes is, however, a low integer. So far so good.

But the problem is with the endogenous reaction to this discontinuity. Schools are very reluctant to open a new class just because of one student, and will find ways around this, for example by increasing tuition. Also, parents select schools according to the number of students in classes. The big question is whether this matters in any significant way. Urquiola and Verhoogen build a model of schools maximizing profits and parents choosing schools and test it with data from Chile. There, private, subsidized schools are for profit and thus fit nicely with the model and indeed the above mentioned effects are important.

The schooling environment in Chile is particular, because a large portion of schools are for-profit. This may not be true elsewhere, but the fact is that schools are still reacting to incentives, and parents, too. So while the data and motives may not be as easy to model as in Chile, this needs to be taken into account for any estimate of the impact of class size.

Donations

As you may notice on the left bar, I have set up a donation button. Why? With the time I spend on this blog, not only writing but also cleaning spam comments, I figure that sooner or later the "lost" time will amount to a publication and the associated merit pay. So, I want to compensate this opportunity cost by allowing user to donate through PayPal. Not that I expect to strike it rich this way. I get personal satisfaction from the blog, so I do not expect complete compensation.

I have had Google ads as well for quite a while now, and they bring only peanuts. I have also had some requests for ads, which I have declined so far because they did not seem appropriate. But that may change.

Yes, tax capital income

A long standing result of taxation theory is that the optimal tax rate for capital income is zero. The only exceptions to this are in situations where private decisions lead to capital accumulation that is higher than socially desirable, for example when households self-insure too much in the face of borrowing constraints, or when accumulated capital is treated as a proxy for age and it is optimal to tax older people more.

Juan Carlos Conesa, Sagiri Kitao and Dirk Krueger present another good reason: borrowing constraints and life-cycle cum no age discrimination. Why would this make the lead article in the American Economic Review (after the presidential address)? Maybe because they come up with a high optimal tax rate of 36%, which comes close to reality in many jurisdictions. While I am sounding sarcastic here, this is a good paper with very interesting contributions. While this result is robust to any reasonable change on the calibration, it hinges one some critical modeling features. For example, endogenous labor supply is required, and so is progressivity of labor income tax and, or course, the life cycle. I cannot think of another paper that would have these (reasonable) features and would thus have had any chance of obtaining a significantly positive capital income tax rate.

Risk sharing among states: Canada does it well

While individual and firms can, at least to some extend, insure against various eventualities hitting them, governments do this remarkably little. Of course, they have some policies available, like fiscal policy, monetary policy and in particular borrowing against future tax revenue, but they do not exploit risk sharing. The only exception are the facilities provided by the International Monetary Fund. But within countries, it appears some good mechanisms are in place.

Faruk Balli, Syed Basher and Rosmy Louis look at Canada at the various ways the provinces have to share risk. Interestingly, only 19% of the risk is left after: capital markets (40%), federal government transfers (25%), and credit markets (16%). The large share of capital markets is attributable to the fact that the relevant banks are present nationwide. Federal transfers include direct transfers to provinces through perequation as well as federal social programs. Finally, the direct access of provinces to credit markets has become less important for smoothing, likely because of more chronic provincial deficits (instead of cyclical ones) and the more positive correlation of them across provinces.

How to insure against natural disasters in the absence of insurance

Survivalists experience a revival these days as all sorts of doomsday scenarios circulate. But even if we are to experience and economic meltdown, it does not compare to natural disasters in developing economies. How do people cope when there is no financial or insurance market to help you out when you are in dire straits. Looking at Bangladesh, Ethiopia and Malawi, Futoshi Yamauchi, Yisehac Yohannes and Agnes Quisumbing show that individuals with better nutrition and human capital fare better. While self-insurance in financial assets is not possible, one can invest in biological assets. Good nutrition allows you to make it physically through bad times, and education allows you to transfer wealth through time, wealth you cannot lose to a disaster or to robbers, and it is very protable should you have to move. Note that the latter in not limited to oneself, but extends to one's children, as they can help out as well.

The Bible and the price volatility

There has been a lot of talk about irrational exuberance in the housing market, of irrational fears in the current recession, and animal spirits in general. Given this, it seems natural to study whether religiosity would have had any impact on these developments. Concentration on evangelical protestants, Christopher Crowe finds that they are much more level-headed than the general population: wherever they population share is higher, house price volatility is lower.

The reasoning here is that their behavior is countercyclical. Evangelical protestants believe that we live in the end times, so any bad news is good news to them, and vice-versa. But the argument needs to be more subtle than that. Imagine that the current bad economic news is some sign of the end of the world. Then asset price should be going further down, as the horizon on which future expected returns are cumulated is shortened. This would lead to more volatility. Crowe argues instead that evangelical protestants are taught to live normally through the end times, and that they are more "joyful" when bad news, such as 9/11, occur. And this would make them spend more, including on housing.

I find this argument rather hard to swallow. But clearly, the empirics seem to be consistent with that. Also, interesting to see that the IMF is interested in this type of topic.

How to best auction natural resources

It is now well known that many natural resources suffer from the tragedy of the commons: because they are not owned, they are overused and depleted. The solution is for the state to sell rights to them. Those rights were typically set administratively and often with political considerations, thinks of the ultra-low grazing rights on federal lands. But governments can obtain more for the natural resources, not only because it increases revenue, but also because it encourages an efficient use of scarce resources. After all, the market price is still a wonderful signal. Thus auctions come into play. But auctions can go horribly wrong as well, as documented by Paul Klemperer. Every auction is different, and details are important.

Peter Cramton offers a handy guide for the design of efficient natural resource auctions. It focuses on oil and mineral rights in developing economies, but there are lessons to be learned for other auctions as well. There is an incredible array of auction designs, which helps accommodate a multitude of market situations and government goals. So, the next time you have an oil field to sell, you know where to look!

Sharing NIMBY

The NIMBY (Not in my backyard) problem is well known: a globally social good that involves a local private bad is impossible to provide if locals can have veto power. One would think that the Coase Theorem would apply: there is a transfer scheme that compensates the hurt community, paid for by those who benefit. The problem, however, is that the proper compensation is difficult to determine due to the public good and the incentive to lie for all parties. Think about the localization of a new power plant, where locals claim they will all die within a year and the other dispute any significant private benefits.

Jérémy Laurent-Lucchetti and Justin Leroux claim to have solved the problem. The mechanism they propose is the following. Every community proposes a compensation, but before a location is selected. Thus, it is not clear from the onset whether a community will be hosting the public good or not. And in the event it is, it wants to make sure the compensation is good. While this induces truth-telling, it does not necessarily balance the budget: there may not be sufficient compensation for the host to give in. However, the public can be provided if one imposes some information structure in the game. The authors suggest that it works by assuming that every community knows something about the preferences of the other communities. That is perfectly reasonable as long as the number of communities is limited.

Specifically, it should be know which community has the lowest hosting cost, and in equilibrium this is also where the public good is located. The game proceeds in two rounds. First every one announces what the lowest hosting cost overall is. Then, all but the one indicating the lowest cost (the "optimist") reveal their own cost. The one with the lowest cost in that second group is selected, as long as its accepts the compensation determined by the optimist. If not, the optimist is selected. This is a nice example of a game where everyone is kept in check, akin to a divide-and-conquer strategy.

The evil empire strikes again (II)

Following up on yesterday's post about Elsevier's strange practices, it appears that a firestorm is erupting, and it seems to uncover other not so nice things about the Evil Empire:

• While I reported about one fake journal, it appears there could be at much as fourteen of them, according to Open Reading Frame. Elsevier so far conceded six, but without naming the corporate sponsors.
  
• The particular Elsevier branch that is mainly incriminated, Excerpta Medica, seems to have been very reputable until fairly recently. The sudden selling out on the goodwill seems consistent with other actions of Elsevier. Journal editors reported to me that when they were negotiating terms with Elsevier, they had the impression the Evil Empire was set to extract as much rent as possible over the next few years, and maintaining quality was not first priority. Hence increased pages and issues in various journals. And selling out to the highest bidder.
  
• Elsevier is not supervising what its editorial boards are doing and imposing standards. Two other examples: Secret blogging seminar discusses a mathematics journal where the chief-editor published dozens of his own articles, none of scientific merit. Elsevier was OK with it, as the journal had the highest impact factor in Mathematics, thanks to self-citations (some much for Thomson's shenanigans as well). The other is the rift between the Society of Economic Dynamics and Control and Elsevier over the board of its own journal, the Journal of Economic Dynamics and Control. This eventually lead to the creation of the Review of Economic Dynamics at Academic Press (which was then promptly acquired by Elsevier). The European Economic Association also left Elsevier in disgust over issues with the European Economic Review and created the Journal of the European Economic Association. The Journal of Economic Theory is similarly being deserted in favor of the open access Theoretical Economics. This is happening in other fields, too, in particular biology.
  
• Elsevier lead the creation of PRISM, a coalition battling the emergence of open access journals. Its main credo is "Government mandates that ignore the need for sufficient and sustainable financial support for peer-reviewed journals -- whether the source of support is from users, authors, or sponsors -- risk undermining the very fabric of the system of independent, formal peer-reviewed publication, a system that is of crucial importance for scholarly communication and the preservation of scientific knowledge." Well, Elsevier just showed this is not possible. Open access is the way to go because it is independent from commercial funding.
  
• In 2004, the CEO of Elsevier testified to a British commission making research available for free would hurt researchers and that payments from firms like Merck (explicitly cited!) are necessary for preserving the integrity of the system. Sure.
  

And what does Elsevier have to say? "It is an isolated case, it was long ago, the people responsible have long left, we are very ethical". We believe you, sure.

The evil empire strikes again

A critical aspect of the research publication process is the integrity of everyone in the business. While one may sometimes have doubts whether particular editors are biased one way or the other, or that journals have a history of favoring certain people (see: JPE and especially QJE), publishing houses should not have any reasons to meddle in this process except to redress editors who venture out of bounds. But could a publishing house, a major one at that, actually hamper the integrity of research and its dissemination?

Yes, Elsevier managed to do just that, according to this article in The Scientist. Let me summarize for those who do not want to go through the free registration. Merck commissioned Elsevier to publish a "fake" journal, the Australasian Journal of Bone and Joint Medicine, stuffed with articles giving positive reviews of Merck drugs, in particular the troubled Fosamax and Vioxx. Now, nothing prevents a drug company to print brochures advertising its products, but Elsevier disguised this like a real journal, with editorial board, subscriptions and real articles (they were a selection of articles published elsewhere, plus fake review articles), all this paid by Merck and without any disclaimer.

This is a particularly nasty practice from a publishing house that has already drawn the ire from librarians and academicians across the board for its journal pricing practices and its predatory acquisitions in the academic publishing market. Closer to home, Elsevier is also trying to hamper the success of RePEc by preventing it to use its bibliographies for citation purposes, as I mentioned before. While the latter practice is not as vile as faking journals and charging researchers exorbitant prices to access their own research, it shows that this company cannot be trusted to be doing what is good for research. It also shows that the profit motive in academic publishing can lead to some pretty nasty results. Non-profit societies and open access outlets need to step up and take over, and we authors need to dump the commercial outlets.

Followup here.

Mother's day special: money creation

This is a TV ad for a German bank. Somewhat racy, depending on your local culture.

The value of virginity

On the marriage markets, virgins have been valuable throughout human history. One can conjecture that this is due to the sexual exclusivity that the husband enjoys. Why? Some claim that rich people care, and thus daughters of rich people will try to remain virgin in order to be eligible for marriage within their class. Poor people cannot afford virgins, and thus do not care, and sexual promiscuity before marriage is common.

All these arguments sound crude, but they reflect the fact that virginity is indeed a trait more common, historically, in the upper class. One consequence is that virginity should be more valued in societies that are more stratified. This is what Fabio Mariani studies using a model with a marriage market where poor girls can move up through love and virginity. This model is capable of explaining the recent decrease of the value of virginity as a consequence of the stronger social stratification (which makes matching across classes more difficult), the increase of female labor market participation (which gives new opportunities for women to strike it rich) and the reduced inequality. Interesting ideas, and I would love to see all this put to data more systematically than through anecdotal evidence.

Who benefits from agricultural subsidies?

Plenty of governments are dishing out large subsidies to their farmers. But are they really benefiting from them? If they rent the land, classic theory would indicate the land owner who be able to extract the whole subsidy from the renting farmer, simply because of the inelastic supply of land, assuming perfect competition for land.

Barrett Kirwan answers this question using US data and exploiting changes in farm subsidies. Kirwan finds that tenants actually manage to keep 75% of the subsidy. Why? Because the rental market is not perfect competition after all, something that is confirmed by the fact that tenants manage to extract more where there is less competition.

While I find it hard to justify agricultural subsidies, they are targeted towards those who farm the land, not those who own it. And it appears that this is working.

The proper way to create education vouchers

Education vouchers are supposed to create competition among schools to improve the level of education within a school district. In particular, it is supposed to help students get out of particularly bad schools and into better ones. The reality is, however, quite different. The best schools get even better because they can afford to become more choosy, and the differences across schools become even larger. So how could this be fixed?

Dennis Epple and Richard Romano suggest that school voucher should not be just blank checks. You need to be subtle. If you want to achieve high and equal quality education, the amount of the voucher needs to decrease with student ability, and the school need to accept them as full tuition. This requires large vouchers, and thus high taxes to finance them. Epple and Romano show that a less expensive system is possible, all you need is attach various constraints to the use and amount of the voucher. And this still works if students or schools can choose to opt out.

The key to all this is to prevent schools from making too much profit from vouchers. Essentially, vouchers increase the paying capacity of schooling demand, and schools exploit this. To counteract this, they need to be constrained, either by disallowing them to accept payments in addition to the voucher (or they would just charge the usual tuition plus voucher and laugh all the way to the bank), or allow side payments with more strings attached. The former seems much easier to implement and monitor, though. Also critical is that voucher amounts should not depend on the income or wealth of parents. Then, one can prevent richer schools from getting even richer with more rich kids.

Mom should stay at home

New working mothers always face the question of when to return to work. While their concern is the immediate well-being of the newborn, what about its long term prospects? It is now well establish that adult outcomes (education, wages) are largely set in the first years of life, especially in pre-school years. The empirical literature on the subject is largely inconclusive, but suffers of several issues: 1) mothers who work and use child care may be different from the others; 2) the child's cognitive abilities may influence the mother's choices.

Raquel Bernal corrects for these issues by avoiding the ominous reduced-form regression. She estimates a full-blown dynamic employment and child care choice model using data from the National Longitudinal Survey of Youth, the same dataset others used without clear results. Bernal obtains clear and significant results, though, showing that using child care during one of the first five years reduces the test score for cognitive ability at the start of schooling by 1.8%. While this does not seem much, this amounts to one eighth of the standard deviation of this score. For high ability kids, the impact is even stronger.

The nice thing with a structural model is that one can perform meaningful policy experiment. For example, introducing a 35% child care subsidy encourages the use of child care, but also reduces test scores by about 1% (it ranges for 0.23% to 1.87% depending on the test). Not particularly encouraging. A maternal leave policy is detrimental as well: as the mother can then rejoin the workforce under the same conditions as when she left it, it increases the opportunity cost of staying at home and she rejoins the labor market even earlier. Cognitive skills of the child are reduced by 0.1% to 1%. However, giving a baby bonus (a quarterly $250 lump sum) increases significantly the number of stay-at-home moms and test scores.

So much for all these policies encouraging women to work. They have perverse effect to slow the development of the youngest. One needs thus to complement these policies with incentives to stay at home during the pre-school years.