An AP Stats Lecture

hilarious, i have a stats final on friday. powers, box plots, correlation, regression, null hypotheses, t statistics, z statistics, chi-squared tests.....ahhhhhhhhhh!!!!

Given the magnitude of the current wave of defaults, and the multiplier effect that it produces, there may well be a good number of people who would have cleared whatever threshold you set at the time their loan was made but would not clear the same bar now. If these people default, then worrying about whether we made Type I errors and Type II errors is a little less important than a static analysis would suggest. The problem is that consumers were betting on a traditional business cycle which would have produced inflation to stabilize some of the impact of the increased financing on aggregate demand. They were receiving mixed pricing signals. Many well-payed economists thought that the world had changed. Or, to put it another way, they believed that stabilizing housing prices were indicative of a new plateau, not an approaching cliff. I disagreed, but that's just me, and who am I?

We have to consider the environment in order to interpret people's responses to it. The fact that two sets of people were interpreting the same environment in two different, yet symbiotically destructive ways should make us rethink pricing signals more than Type I versus Type II errors. If the lender doesn't have the appropriate information to set the bar at an appropriate level to minimize defaults, the consumer certainly doesn't have enough information to decide whether or not to try to jump over it. How and why would they be receiving pricing information that would lead both parties to believe this was a good idea?

@El - You say that people who passed the test would have defaulted anyway. I say that loaning money to test failers catalyzed the problems that now result in problems with test passers' loans. Those loans helped inflate the bubble. No bubble; no problem.

Larry,

Understood.

What if we had some mild inflation two years ago to devalue the existing ARMs and tame the asset bubble? Bubble goes bye-bye and lending slows as rates rise. I'm not saying we should have wished for inflation to solve this problem. I'm simply trying to say that there are a number of possible scenarios that could have occurred, and very many people chose the wrong one.

In retrospect, you are correct. There was a bubble and then a crash. Marginal loans contributed to it. My assessment was correct during the run-up too; that this was all getting out of hand and was a pretty ominous circumstance that could lead to very negative consequences. I don't think we disagree about what occurred, or even why. I assume you aren't stuck with a financial mess, and neither am I. Good for us.

What I'm saying though is that understanding people were responding to signals helps us to think about what signals we are sending now and would like to send in the future. It also helps us understand what expectations they have and gives us a better sense of our audience. This should all help us to craft better policy. I think we could have done much better given what we knew at the time. Many of us chose not to. I'm just doing a little postmortem examination.

The reason I thought this was all a losing proposition is that I think we did run up against a resource constraint: time. With two-income households, ballooning consumer debt, and stagnant real wages, we had already nearly maxed-out our capacity to work and maintain any sort of a healthy lifestyle. Borrowing against future wages that were not going to materialize was not wise. Doing so for consumption instead of investment was doubly foolish on the part of many. I watched them do it, in person no less. That's what troubles me. I think they'd do it again.