Wednesday, October 29, 2008

On "Inscrimination"

Via Ilkka, a post at L'Hote defending affirmative action:

[I]t's the same thing that Ward Connerly and other affirmative action foes want. A truly open and free and fair society, with neither preference or discrimination for those outside the norm. A high school, a culture, where difference really doesn't matter.

But life, I'm afraid, is not that simple. These old ways of discrimination have a way of outliving anyone who would openly endorse them. And the stark question remains: what if denying this preference really does ensure that these people will have no opportunity to excel at all? Affirmative action foes want to keep the conversation centered purely on principle. But what about the consequences? What if ending widespread affirmative action means that black college attendance rates become extremely low-- as evidence from the California public education system seems to suggest they might? Can't even the most hardened anti-AA warrior concede that there are practical public disadvantages to having college attendance-- and effectively, participation in the American middle class -- be drawn along racial lines? I know that many would say quite openly that they don't care if any black people at all go to college, as long as the selection is based on equitable and fair criteria. For myself, I think the existence of a permanent black underclass has been a major detriment to a just and secure American society, and I can't imagine a way in which ending affirmative action-- without some concurrent effort to ameliorate black poverty and joblessness-- could have a positive effect on that reality.

Freddie skips over a lot of detail here:

How much, and at what costs? Contra Freddie, a color-blind society would not exclude blacks from the middle class. It might focus them into fields in which they enjoy competitive advantages (marketing and sales) over those in which they do not (genome sequencing), but which still allow them to make middle-class money. It might also place them in schools (Cal State) where they stand a good chance of actually graduating with a degree in a substantive field, as opposed to schools (U.C.) where the odds are against completing any but the most worthless programs (i.e. "Black Studies").

But let's stipulate that, given the I.Q. distributions, and assuming that U.C. imposes an appropriate I.Q. cut-off (115, let's say), that the black share of seats in the U.C. system falls to 3.6%.* Even if we further stipulate that a U.C. seat automatically assures higher earnings than a Cal State seat, this scenario isn't exactly the exclusion of blacks from the middle class.

But what share is Freddie proposing? It's population quota of 13%? That is a non-trivial cost, not only in opportunity denied to whites (and Asians), but also in reduced performance, productivity, and competitiveness as the same affirmative action at U.C. percolates through the entire professional world.

Freddie may honestly think we can bear this cost; he should be given the chance to prove it. But having admitted the principle of racial quotas, does he honestly think we will limit them to blacks? Indeed, Hispanics are already standing in line to receive their "share" of the affirmative action pie, and if the census projections are to be believed, the beneficiaries of this pie will exceed 50% of the population in about 30 years. Let Freddie prove that cost to be non-trivial.

What is the distribution of costs and benefits? Affirmative action didn't keep Chelsea Clinton out of Columbia. On the contrary, affirmative action comes at the expense of the most vulnerable whites members of the non-preferred races, those whose backgrounds do not include the money or connections that buy entrance to the Ivy League. Their talent and hard work are all they have. In contrast, affirmative action accrues to the benefit of the most well-placed blacks: Berkeley's quota would have done well at Riverside; Riverside's quota would have made it at Long Beach, etc. Contra Freddie, affirmative action does nothing for the black underclass.

Are the benefits real? Let's remove the stipulation of my first point. In the real world, the Berkeley admissions quota does NOT go on to graduate. We know from controlled studies that there is an optimum pace at which a student assimilates new material; exceed that pace, and the student doesn't learn more, he learns less. If the body of knowledge is cumulative, this puts quotas at a crippling disadvantage. And let's not pretend that "it is better to have tried and failed"; again, in the real world, the college drop-out is not only out the wages and work-experience that he missed while in college, but he's now encumbered with non-dischargeable education debt.

Whither the Law? Freddie is counting on some omnibenevolent entity that will approve of discrimination he likes, while squelching discrimination he doesn't like. But that is not what the Civil Rights Act summons forth. What it actually requires is NO RACIAL DISCRIMINATION in education and hiring! Never mind that this was not and is not the goal of the majority of Civil Rights activists; it was the goal that America actually bought. That Freddie wants to change that goal is certainly his right, but he must change the law.

Repeal the Civil Rights Act, Freddie! Allow firms and colleges to hire and admit by their own lights, whatever they are, and remove this intrusive federal oversight into private and state-level concerns. And sell it to the public.

Alternatively, let the law reflect your true intentions. Specify that it is okay to discriminate against whites (and Asians) in favor of blacks (and Hispanics) at Berkeley, but NOT okay to do the reverse at BJU.** Sell that to the public.

But your omnibenevolent entity doing that from which legislation shrinks is tyrannical, and in any case, the Supreme Court has signalled an end to its patience with willful violations of the law as written. So pick a course. But your time is running out.

*To the uninitiated, here is how I arrived at 3.6%: The mean black IQ is 85. the mean non-black IQ is 100 (it's probably lower given the increasing Hispanic share of the population, but let's assume worst case). The standard deviation of IQ for all races is 15 and the distributions are normal. Find a table of Z-scores on the internet, and discover that the percentage of blacks with an IQ over 115 is about 4% (with rounding), while the share of non-blacks with over-115 IQs is 16%.

Now we assume that the black share of the population is 13% (with rounding); ergo, the non-black share is 87%. Let's define our probabilities thus:

P(B) = Probability of being black = .13.

P(N) = Probability of being non-black = .87.

P(S|B) = Probability of being smart (IQ > 115) given being black = .04.

P(S|N) = Probability of being smart given being non-black = .16.

We can now apply the law of total probability and Bayes' Theorem to calculate the non-quota black fraction of U.C. admitees:

P(S) = Probability of being smart = P(S|B)P(B) + P(S|N)P(N) = (.13)(.04) + (.87)(.16) = .14.

P(B|S) = Probability of being black given being smart = P(S|B)P(B)/P(S) = (.13)(.04)/.14 = .036 = 3.6%.

**Yes, I know: Bob Jones University has abolished its racially discriminatory regulations. But you get my point.

UPDATE: In the comments, Bobvis links to a study showing that, indeed, the primary victims of affirmative action are Asians. I have corrected my above implication to the contrary.


bobvis said...

affirmative action comes at the expense of the most vulnerable whites

Do you have any evidence of this, by the way? This is a question, not an accusation, by the way. :)

bobvis said...

Also, one other question. Why use IQ scores when these aren't even available to universities? Wouldn't it make more sense to use something like SAT scores?

Φ said...

Bobvis: Good questions. The honest answer to the first is that I have no numerical evidence. But then, admission to the Ivy League was never about just grades and test scores.

Consider, for instance, the preference for the progeny of alumni, particularly alumni that have made large contributions to the university. I believe it's called "legacy admissions."

Consider also admissions criteria that give points for unpaid internships and expensive "broadening" experiences over, say, stocking shelves at WalMart because you need the money.

Universities may, and probably do, have self-interested reasons for these racially neutral policies. And I am not in favor of "class-based" affirmative action. But the point is that their net effect is to benefit the rich and connected over the poor regardless of race, and when it's time to use race to "plus-up" the jackets of the affirmative action recipients, it comes at the expense of the whites who do NOT have these advantages.

The honest answer to the second is that I made up the 115 cutoff because I was too lazy to look up U.C. SAT data. But I'll see what I can find.

bobvis said...

You don't have to look up the SAT data on my behalf. I was just wondering if there was a reason in particular.

Here was a post that I wrote that cites a paper that finds that Asians are actually the ones who give up most of the spots that are reassigned due to affirmative action.

bobvis said...

By the way, if you're not able to access the AA paper in the post I linked to and are interested, I can send you a copy.

(Hmm, the word verification thingy here is "coward". Is blogger trying to tell me something? It is very rare for it to actually spell out a real word.)

Freddie said...

Once again, the question remains, completely independent of questions of the efficacy or fairness of affirmative action: are you prepared to deal with the consequences of an African American college-going population that is close to zero? And are you willing to publicly say, "I'm totally cool with an effectively segregated system of higher education"? Do you think that public awareness of near-zero college attendance rates would go unnoticed or unchallenged by black Americans? What are the public policy consequences of having an African American population that is unable to partake in the most effective method of improving social class, getting a college education? Those are my central questions, and despite your bluster, you don't appear to have answered them.

Φ said...

Bobvis: good point about how the Asian share of U.C. seats was the most diminished by affirmative action. Ironic, in its way, because "Asians" are a pretty diverse bunch, and the "smart Asian" stereotype applies primarily to N.E. Asians (Japanese, Chinese, Korean), whereas S.E. Asian intelligence comes in at, or slightly below, the U.S. mean.

Freddie: to answer your question specifically, I am prepared to allow black representation on campus to settle at its natural level, which I take to be a lot more than 0%, and a lot less than 13%.

bobvis said...

In the case of Asians, I have to express my doubts in your use of the resident IQ figures of the native populations. There are at least two good reasons for this.

1) Oddly enough, the US government does not allow immigration of a representative set of Japanese, Chinese, Indians, or any other group in the planet.

2) It ignores cultural differences. To claim that all the differences between black and Asian college qualifications is IQ seems naive. Growing up Asian (any type of Asian) in the United States means that your entire childhood is dedicated to an obscene degree to getting into college. Many of their parents are advanced US degree holders. This makes a difference.

Φ said...

Bobvis: Thanks for the clarificaton. I did not intend to attribute their outsized representation in higher education entirely to their native intelligence. You are correct this is amplified by the selection effect and the cultural pressures.