I saw this quote in your article, ‘”If you are making a high return, I guarantee you there is a high risk there, even if you can’t see it,” said Mr. Lindsey, a former chief economist of the Securities and Exchange Commission.’ I agree with the implication that high return in the markets are derived from high risk but “even if you can’t see it” makes no sense. Just follow the trajectory of Wall Street bonuses over the last 6 years. The tripling of compensation to the “masters of the universe” bonus pool should have been a warning and it was published in your paper several times a year over through the bubble. When bankers look to distribute bonuses this year, after collapse of several institutions, loss of value in the survivors, and a bail-out by taxpayers, can someone please ask the question, where it the outsized risk or leverage that justifies this payday. These bonuses of the investment banking bubble were the canary in the coal mine. If the bankers insists its necessary to retain talent, have them explain the risk that’s implicit in the compensation.

Is this simple implicit risk in high compensation a potential warning sign, particularly in a time where the rest of the economy was going sideways?

“The models failure” should probably read “the hip models failure.” There is an old — over 50 years — and distinguished financial model: the Discounted Cash Flow (DCF) model. It is so old and elementary it is scarcely ever used by financial analysts, at least of those working in the “finance” sector that is the concern of the financial crisis. (It is still used within companies and for almost all project-financed projects.)

Had someone done a DCF modeling of the bundled-mortgages industry they would have predicted the crisis. Indeed several of the Federal Reserve Banks’ research economists wrote of the worsening quality of mortgage loans whose conclusion was increased default rates which in turn implied housing price decreases which given the leverage of the industry implied the crisis (at least once the default rate hit some minimum level).

Without looking at the actual cash flows (the incomes of the borrowers, the interest and principal payments of the mortgage borrowers, the same of the mortgage holders (the bundlers), and the non-mortgage cash flows of the mortgage holders, it is hard to estimate the size of the quasi-systemic risk (systemic to the mortgage backed asset markets but not necessarily to the larger financial market). Just looking at housing prices, the rise after WW II probably looked worse on the surface. For instance, if the US had a 25% increase in population (say a large relaxation of immigration) and the doubling of total factor productivity, my guess would be that the housing price increase would make the 2000-2006 increase look like a little blip. The growth of GDP and its implied increase in expected net worth would justify a substantial growth in the prices of a good with a large fixed component — the classic Ricardian dismal prediction of increased land rents.

A large part of financial model failure is simply the failure of statistical trends or tendencies to change — the parameters or equilibrium conditions change (see Clements and Hendry “Economic Forecasting in a Changing World” The Berkeley Electronic Press for an example regarding macroeconomic forecasting). Our proclivity for financial crises is in part that the models we rely upon the most tend, by the nature of their current popularity, to use the same statistical evidence: here the estimated distribution of future housing prices.

The subprime crisis was a price bubble with an increasing leverage with moderately increasingly risky fundamental cash flow (the home owners net available cash flow), hardly something we haven’t seen before.

That’s very remarkable (that the analysts assigned a very low probability to a 10-20 percent drop in real estate prices).

According to this website (go to time = 5:50 for discussion of the U.S. housing bubble), the housing bubble should have been very visible as early as 2000…and certainly by 2005.

http://www.chrismartenson.com/crash-course/chapter-15-bubbles

So any analyst familiar with the hundred-plus year history of housing prices should have thought that a 10-20 percent drop was virtually certain after ~2005 (and maybe as early as 2000).

My guess is there weren’t enough analysts who had knowledge of long-term historical U.S. housing prices.