gwern comments on A Parable On Obsolete Ideologies - Less Wrong
You are viewing a comment permalink. View the original post to see all comments and the full post content.
You are viewing a comment permalink. View the original post to see all comments and the full post content.
Comments (272)
And why can't being part of a minority group be part of a track record? In both cases, the party thinking of giving the loan has learned new and material information.
So? If they had offered a loan at 100% interest, people would be bitching anyway about discrimination. Maybe it is a calculated PR move that people will complain less about not getting a loan than about getting a loan at +2% interest; maybe it has to do with the fixed overhead of servicing loans; maybe they only had so much appetite for risk even if the interest rate were raised to make the loan +EV again; maybe this is an unexpected consequence of the millions of words of regulation governing banks.
I doubt that and I suspect you have no good reason to believe that.
Maybe for people with a really thin track record. Not for people who've run a profitable business for a decade, have a solid business case, etc.. Just what independent information does being black provide, here? (No need to respond to this point here - you can put it in response to TheOtherDave's parallel comment)
Now, I freely grant that if you only look at the bankers' side of things, regulation doesn't make sense. They should be perfectly informed and make the ideal decisions as to who will succeed or not! And if they can't make perfect use of that information, that's their problem!
But of course, it's not only their problem, and the purpose of the law is not solely to help the bankers get money.
A 'being black' penalty even of only 2% on loans would be colossal. Think how much debt a growing business has to incur, and compound that over the years. Apply it to choices of housing, of auto financing. Everyone in the community has less money, so they can't buy as much, so businesses have a harder time growing, so people earn less. It's self-perpetuating.
But it wasn't that mild. Instead, the loans were simply turned down. The near-total absence of capital was crippling.
Do you think this is a problem? What might you do about it?
In the scales of values, where does "equal opportunity" rank up against "allowing each individual to make the best choices available to them personally without regard to the impact that choice has on what choices other individuals get"?
(Responded.)
To repeat myself:
Moving on:
So, are you arguing that the 'being black' penalty is not accurate here? That the 2% penalty is compensating for risk that does not exist? If so then that's quite an opportunity: a free 2% price cut. Especially in the current low-interest rate environment, that's a lot of money being left on the table.
Or are you arguing that the penalty is accurate and anyone who tried to offer at non-penalty rates would lose their shirts as tons of loans defaulted?
I thought you were arguing against the latter, before... If you've changed your mind, it'd be good to say so clearly and explicitly before going on.
You persist in only looking at the effect on the banker. As I said, the regulation won't make much if any sense from that point of view alone. Your pointing this out once more doesn't really change anything.
That said, considering that the regulated banks had a low default rate until everything went screwy because of the (non-regulated) predatory lenders, I think I'll stick with the theory that there was a market inefficiency.
Now, it wasn't a simple barrier-free market inefficiency. If only one bank had moved, it could well have lost its shirt. Being the sole investor in a dead area is rough. It's a bit like a many-player game with 'staying out' having a positive payoff from being able to invest elsewhere, and 'going in' having a payoff that starts out negative with no other cooperators, increases with cooperators for a while, peaks above 'staying out', then declines down to the 'staying out' level once the market is fully competitive (then below if it becomes saturated). When no one is there, you'd be crazy to move first.
Changing the rules so the banks had to serve their localities guaranteed a minimum level of going in, which made it no longer a bad move to go in.
Just so I understand that last part... are you claiming:
1) ...there is no factor or set of factors X such that X screens off skin color for purposes of calculating loan EV,
2) ...there may be such an X, but we don't know what it is and/or have no way of measuring it,
3) ...we know what X is, but the admissible parts of the track record don't include it,
...or something else?
#1 seems really unlikely to me, so if you think it likely I'm interested in your reasons.
#2 seems plausible.
#3 also seems plausible, but if true, suggests that talking about the admissability of skin color is a herring of unspecified hue; as a bank, I should be trying to get X admitted instead.
I am claiming that you have jumped to the conclusion that claims 1-n must all be false and hence the cancellation is a Bad thing. For my purpose, any of 1-3 can be correct, and I don't need to argue for any specific one especially if you already find plausible any of them.
If you want to overrule the bank's reaction, you need a lot of local domain-specific knowledge, and that's something neither you nor I have.
Here's an example of the sort of obscure knowledge you might need: are you familiar with the term "minority fronts"? This is when a government or other entity attempts a form of affirmative action in which minority-owned businesses or organizations get financially favored for grants or bidding on contracts; yet of course there aren't very many competent minority-owned businesses (that's why they're doing it in the first place), so this immediately leads to fraud in which a person meeting the minority criteria is put in place as the straw owner of a company actually run by other people of other groups and then the business bids on business as a minority-owned business, frequently simply passing on the work and money to a normal business minus the straw owner's share. Naturally, this is largely meaningless for a white-owned business as whites are not considered a minority. Now, remembering that the 'owner' in such cases has stooped to conspiracy and felony fraud to steal possibly millions, do you think that their shell business is a better or worse credit risk than an equivalent white owner matched on other observables like income?
Personally, I'd say worse. This is just a cute example of why race can be a salient factor in loans and why a white owner might be different from a black owner; I bring it up to stand for the larger class of relevant details and possibilities, not being a banker or economist or risk specialist. So before you can say that the 'discrimination' is purely irrational and unjustifiable, you need to know about all these details. Do you? I don't think you do.
EDIT: Ran into an international example of the abuse of affirmative action in The Economist:
OK.
If that's true, I haven't noticed.
If you feel like summarizing the evidence for believing I've reached that conclusion, though, I'll certainly listen.
I certainly agree that overruling the bank's decision about loans without knowing about as much about loans as the bank does is going to have unexpected consequences, so unless the bank is currently pessimizing for my values I'm probably going to be less happy with the result than I am now.
And I certainly agree that I don't know about all the details that go into making decisions about loans.