1. My way of thinking about the price/rent ratio for housing (see this post from 2003) caused me to think in terms of values that might be reasonable, not historical norms. When the price/rent ratio went above historical norms, I did not consider this in and of itself as an alarming sign. If I had, I would have started worrying much sooner and much more about high house prices.
Until very late in the game, I thought that the biggest threat to the housing market was an increase in the real interest rate, as opposed to a purely internal bubble/collapse.
2. I was sure that the stress tests used by Freddie and Fannie and the capital regulations at bank were sufficient to keep those institutions from taking on excessive credit risk. I thought that the sub-prime crisis would only cause newer, peripheral institutions to go bankrupt.
Incidentally, the Report released yesterday by the regulator overseeing Freddie Mac and Fannie Mae was very disappointing to me in that it says nothing about stress tests. What I would like to know is this: as of December, 2007, how much capital should Freddie and Fannie have been holding in order to conform to the regulatory stress test requirements? Did they have enough? Too little? More than enough? Then, subsequently, how bad was the housing price outcome relative to the stress tests? And what do the stress tests say now?
I strongly suspect that the actual house price outcome was not dramatically worse than what is used in the stress testing methodology. Instead, I suspect that Freddie and Fannie were way under-capitalized all along based on the stress tests, and that the regulator was not aggressive in dealing with the problem. My guess is that the regulator is not terribly eager to bring this to our attention.
3. I thought that the economy had become less susceptible to cyclical downturns. In April of 2003, I wrote The Elastic Economy, which argued that "the private sector has become more chaotic but more robust." Read the whole thing. I also believed that inventory corrections would be less of a problem, for two reasons. First, the share of GDP represented by automobiles and other durable goods is much lower. Second, computer systems have made inventory management less mistake-prone.
4. I'll talk about derivatives, not because I was wrong but because Brad and Tyler bring them up. From the late 1990's until 2008, I was not interested in financial markets, so derivatives were off my radar screen. Back in 1986, the Fed published a book-length staff study on "Financial Futures and Options in the Economy," and I wrote a chapter called "Futures Markets and Transaction Costs." My view was that derivatives markets on organized exchanges serve to reduce transaction costs. Then and now, my thinking tended to downplay the role of derivatives as hedging and risk-management tools. So I would probably not have drunk the Kool-Aid that said that these were making financial markets more able to handle risk.
Back in the 1980's, I got the sense that a lot of financial executives who played around with derivatives had insufficient understanding of option-pricing models. (You should have seen some of the S&Ls that bought "CMO residuals" for yield back in the 1980's. I had not yet coined the term "suits vs. geeks divide," but the phenomenon certainly existed.) So the fact that some companies blew up because of derivatives is not a shock to me. And I don't think that regulated exchanges are the answer. Google for "Arnold Kling credit default swaps exchange" to see why.
I first heard about credit derivatives when my former Freddie Mac colleague Frank Vetrano mentioned them during a break at one of our fantasy baseball auctions one year at Dave Andrukonis' house. This was some time around 1999 or 2000, after I had left Freddie, and I did not think that credit derivatives made much sense or would amount to anything. Judging by subsequent market volume, I was clearly wrong in thinking they would not amount to anything. As to whether they make sense, I think one could say that is still an open issue.
5. I would say that I have become less of a Keynesian since the crisis took place. Before the crisis, I would have stuck up for Keynesian macro, with the proviso in (3) that I would have thought that Keynesian demand policies would be needed less going forward. I was always a skeptic on monetary policy, and I continue to have a lot of skepticism. But I developed the whole Recalculation Story and related ideas as kind of a delayed, semi-subliminal response to a Tyler Cowen blog post as well as thinking about various empirical phenomena, such as the JOLTS data. However, whether I was most wrong in believing in Keynesian economics before or whether I am most wrong now in believing the Recalculation Story is certainly an open issue.