We've written before about why schools continue the practice of favoring legacy admissions - accepting the sons and daughters of wealthy alumni. Now there is some empirical evidence of the economics that drive this practice. Slate Magazine recently ran an article about the puzzle of charitable giving in economics - if markets are driven by individuals rationally pursuing their own best interest, where does charity come from? A new study by Jonathan Meer of Stanford and Harvey S. Rosen of Princeton shows that when it comes to donations to one's alma mater, charity isn't altruism. Alumni with kids are 13 percent more likely to donate, and they are more and more likely to donate as their kid reaches age 14. At that point there's a big split - for those parents who's kids go on to apply to the school, donations continue to increase. The parents whose kids do not apply to the alma mater drop off giving. It seems pretty clear that many parents give to their schools because they think it will influence their kids' chances of getting in. Colleges an universities benefit from this, but the study did not examine whether or not the donations worked. This whole process strikes most people as unfair, mostly because the focus on GPAs, SAT scores, and admissions essays makes it look like it's supposed to be a meritocracy. Americans love democracy (where everyone gets an equal say and an equal chance) and stories about unlikely success stories and self-made men. Allowing external factors to secretly skew admissions is so unpopular that affirmative action has been continuously attacked. Legacy admissions are affirmative action for rich people. So my advice to schools is to either do away with the practice (not very likely), or make it public. Why not set aside a certain number of admissions, and just let parents bid on them in an auction? The regular admissions will be more of a meritocracy, and auctions are pure capitalism, something Americans love. Heck, put the admissions up on eBay, that way you don't have to build your own infrastructure.