Friday, March 16, 2018

Unintended consequences

Unintended consequences of well-intentioned policies, unexpected behavioral changes in response to ignored incentives, unusual supply (or demand) responses to demand (or supply) interventions, and clever new pathways for changes to happen are the sorts of mechanisms that make economics fun, and I hope useful to cause-and-effect understanding of human affairs.

A case in point is an Atlantic article from 2012 that a friend pointed me to last week, by Richard Sander and Stuart Taylor Jr.
... UCLA, an elite school that used large racial preferences until the Proposition 209 ban [on overt racial preferences] took effect in 1998... Many predicted that over time blacks and Hispanics would virtually disappear from the UCLA campus.
And there was indeed a post-209 drop in minority enrollment as preferences were phased out. Although it was smaller and more short-lived than anticipated, it was still quite substantial: a 50 percent drop in black freshman enrollment and a 25 percent drop for Hispanics...
...The total number of black and Hispanic students receiving bachelor's degrees were the same for the five classes after Prop 209 as for the five classes before.
How was this possible? 
Indeed, I too would have guessed, if I didn't think hard about it, that eliminating racial preferences would have to have reduced the number of minorities who graduated, and that the affirmative action argument would have gone on to other pros and cons. But that's wrong.
First, the ban on preferences produced better-matched students at UCLA, students who were more likely to graduate. The black four-year graduation rate at UCLA doubled from the early 1990s to the years after Prop 209.
Yes. Half the admits but double the graduation rate leaves constant the number of graduates.
Second, strong black and Hispanic students accepted UCLA offers of admission at much higher rates after the preferences ban went into effect; their choices seem to suggest that they were eager to attend a school where the stigma of a preference could not be attached to them. This mitigated the drop in enrollment.
Third, many minority students who would have been admitted to UCLA with weak qualifications before Prop 209 were admitted to less elite schools instead; those who proved their academic mettle were able to transfer up to UCLA and graduate there.
Thus, Prop 209 changed the minority experience at UCLA from one of frequent failure to much more consistent success. The school granted as many bachelor degrees to minority students as it did before Prop 209 while admitting many fewer and thus dramatically reducing failure and drop-out rates. 
To be absolutely clear, this post is about pathways. I do not wish to wade into a perilous pro or anti affirmative action debate, a basically radioactive topic for white male economists. (Though I am pleased to report a quick Google search that suggests both Sanders and Taylor still employed, something that might not happen if their book were published today.)

And a proponent of affirmative action could nonetheless make many arguments consistent with this work.  Perhaps dropping out of UCLA is good for people. Perhaps more minorities on campus is useful for white students' social perceptions, even if it harms its intended beneficiaries. Perhaps things were going on at other universities that drove minority upperclasspeople UCLA's way. UCLA is part of the California state system, which encourages transfers at year two, which is not the case everywhere. I also don't know how the numbers are holding up post 2012. Finally, racial preferences seem to have advantaged whites by keeping asians out, which is an interesting scandal by the silence surrounding it.

Today's post is not about this larger argument.

I'm willing to bet Brad DeLong still blogs I'm racist for even mentioning the topic, but that will be an interesting test of today's political climate.


  1. Was there a counteracting growth in supply? Another argument says affirmative action was a separating wedge, selecting the better. When released from the wedge, the other colleges responded with more, better supply.

  2. care to look at the percents of people being granted A's.....grade inflation is stunning.

  3. >>Perhaps more minorities on campus is useful for white students' social perceptions, even if it harms its intended beneficiaries.

    Nitpick: I don't think it's possible for minorities to be harmed by AA if they have rational expectations. Since 75% of the value of college is signaling and not human capital accumulation, if I choose to go to a college I might not be qualified to attend (in the sense that there's a high chance I won't graduate), then I'm still better off on average than going to a lower ranked college where I will graduate for sure. Ex-poste I can be worse off having failed UCLA, but ex-ante I am better off on average going to UCLA vs community college.

    1. I'm not sure your argument fits with signaling. A UCLA dropout is better off than someone with a sheepskin from a lesser university? I'm skeptical.

  4. I agree with the thrust of this post and I especially agree with the sentiment that people should be able to express opinions or insights without fear of retribution.

    There are many unintended consequences of policy, including in the areas of property zoning, or foreign expeditions to lands such as Vietnam, Iraq or Afghanistan.

  5. Another possibility is that minority LA students, aware of the heightened standards of UCLA, improved their high school performance on average. Some students likely deliberately try about as hard as they expect they need to to get into the level of school they're interested in. If you know you can get into UCLA with a 4.0, and you're fine with UCLA, why spend the extra time studying to get a 4.3? But if UCLA raises its expectation to 4.2, then you study more to meet the new GPA necessary to get in. It'd be interesting to see if the average black or hispanic high school GPAs increased after UCLA abandoned racial preference.

    By the way, is opposition to affirmative action really that bad for one's career in academia? It didn't seem like a career-killer a few years ago. Didn't think things were changing that fast.

    1. Today's mob mentality has become a real source of fear, whether the threat is real or not. Academics can get away with anything unless they just happen to have the spotlight pointed at them, in which case, they can be in hot water for absolutely nothing. Or so it seems to many.

  6. you need to isolate how the professors reacted to the new student body. Did they try to help boost the new minority students with higher grades or perhaps they had a hidden bias against the old group with lower grades. Or perhaps professsors are immune to such potential conflicts or bias. This might be hard to study without upsetting a lot of people

  7. John;
    John McWhorter offered this tidbit in an affirmative action debate ( 4:30): At UC San Diego before the ban on racial preferences exactly 1 of 3268 black freshman was making honors, after the ban 20% were making honors.
    The argument has unfolded thus. Affirmative action mismatched students. Students who had been struggling at UCLA were now free to flourish at UCSD. I think part of the explanation for UCLA would be that Berkeley is a notch higher I think, and was probably poaching students for their quotas. After the ban those students were free to flourish at UCLA.

    I credit Sowell's book and writings for most of what I think I know. His explanations have been pretty good on affirmative action.

  8. I would like to see other actions taken at that time. Presumably, it was known for some time when ban would come into force. So, government, local communities, charities and activist organizations had ample time to prepare different actions that would help non-whites finish university. It is possible that with new measures, if number of non-white students were to return to the old level, twice as much students would get their degree.
    If something like that happened, it would show the economic inefficiency of protectionism. Because non-whites were "protected" (they had guaranteed number of places), there was no critical mass of those concerned to find innovative ways to improve their ability to get the degrees. Same with protectionism in trade - if you can keep some form of protectionism and force industry to modernize, it may be worth it. However, when tariffs or another means of protectionism are enacted, business in question is more likely to pocket fat profit than to invest to reach the efficiency of international competition.

    Another possibility that would warrant serious analysis: sociological and psychological side. The article states that without preferences, larger percentage of students accepted the offer. It could well be that some students thought they are being selected only because of their race, and not because UCLA thought the are capable of getting a degree, so they opted for safer universities where they could be sure of getting a degree. Without preferences, the are assured that they are good enough. Because of rejections, UCLA did not get half of the non-white students that were good enough and half that were preferred, they got 1/4 or 1/3 that were good enough.
    Socially, at least in the beginning, non-whites should be more likely to socialize among themselves than among the general population (there should be some resentment on part of those who "earned" their right against those who were preferred, and it is likely that white and non-white students came from different cultural backgrounds and even social circles, so it takes time to develop relationships; also "we have to stick together" mentality). That means that those non-white students who had disruptive element (someone who is more interested in partying and being drunk than studying) among their circle would find it harder to leave him/her and find other students to socialize, and exposure to such disruptive element can drag the whole group down.


    So another teacher gets fired for a non-PC outburst.

  10. I first encountered the existence of "Affirmative Action", around 1982, not long after moving to USA. It sounded like a good idea. i.e. affirming that we act without racism. But I was puzzled that the implementation had nothing to do with acting without racism, which would be an affirmative purification of attitude. It had only to do with producing a numerical account that would satisfy some faceless bureaucracy.

  11. I wonder if we think the case of university students is somehow unique or if the same thing would happen if explicit or implicit affirmative action was ended in other areas?


Comments are welcome. Keep it short, polite, and on topic.

Thanks to a few abusers I am now moderating comments. I welcome thoughtful disagreement. I will block comments with insulting or abusive language. I'm also blocking totally inane comments. Try to make some sense. I am much more likely to allow critical comments if you have the honesty and courage to use your real name.