This article "The Next Epidemic: bubbles and the growth and decay of securities regulation" has a nice historical survey of bubbles and stock prices and fraud with an interesting simplistic model of when fraud pays. It could be expanded to predicting individual stock prices or markets I think with some benefit. The conclusions about regulation of course one finds completely out of blue.

Russ Sears comments: 

I must tie this to the one to the Chair's "10 things I learned about the market" post. I have found that higher math is NOT useless in the markets, in fact I have built my career on it time and time again. Studying them has been quite fruitful for the firms I have advised. However, it is not by actually applying them; rather it is understanding how regulators use them and thereby create blind spots for the bubbles to develop. Why people build faith in their sophistication and regulators love their absolute answers, but miss the glaring obvious that is outside the model/numbers, but still numerable.

There is a simple rule that the risks you are most blind to is the risks that naturally builds up and is totally over-allocated in your portfolio.Now that we have GPS watches that accurately measure a course, it is clear that my favorite courses were all about 1/4 mile short of my estimate and the courses I avoided were estimated as short than reality. Bubbles build because the general public and the regulators are blind to the risk in their models.

I have heard that the regulators have pinned the crisis on everybody being driven to invest in the same thing at the same time, hence they all crashed together. To solve this problem, they have come up with the wonderful (not) idea of measuring the correlation of a bank with the correlation to the others to determine the capital required of any one bank. This might actually work if we had hundreds of banks all about the same size and none of them big enough that they could dominate anyone market. So now instead of nearest neighbor/ copula coefficients misunderstanding the correlations rather than cause and effect over/under-allocations. We soon will have a system built on distancing from the nearest neighbor driving them to exotic allocations. So soon instead of having all banks investing in a deep market like the housing market until it is overwhelmed by the cash pouring in due to cheap capital requirements on AAA rated bonds. We will have each of the big banks overwhelming the smaller asset classes, due to it non-correlation to the markets.

While I have used this successfully, I purposely left out some of the details on how its executed, I will leave it to the Chair to determine if this is too valuable information to the general public/ website.





Speak your mind


Resources & Links