Search This Blog

Tuesday, January 13, 2009

What does the global financial crisis and global warming have in common?

The largest banks in the world did not intend to commit suicide.

The Swiss banks in particular prided themselves on risk control. So why did the largest Swiss bank (and the world’s largest investment manager) manage to lose over $50 billion dollars in mortgage backed securities?

To understand what happened it’s useful to know why they almost killed themselves without knowing it.

Most mortgages are sold to institutions like Fannie and Freddie after they are issued. They are bundled together into a bond which is sold to people who want a steady income. Mortgage backed bonds were subject to some fancy financial engineering to create different levels of risk. Bonds were split into pieces (called “tranches”). Each piece had a different level of risk associated with it.

Let’s look at a typical mortgage backed bond. It’s essentially a bunch of mortgages. We know that mortgages are backed by real estate and the ability of the mortgage holder to pay principal and interest. We have a historical record that shows us what happens to mortgages when interest rates go up (they are paid off over a longer time period) and when rates go down (they get paid off faster because people refinance).

We also have good statistics to show how many mortgages go into default.

So based on these data, we can slice and dice these mortgage backed bonds to either increase income or reduce risk.
Let’s say we slice them into four “tranches.”
Tranche A has the highest interest rate but if a mortgage goes into default, this tranche absorbs all the loss.
Trance B is next with the next highest interest rate and it absorbs all the loss from defaults after A is done.
Tranche C is next in line to absorb losses.
Tranche D is last and safest.
Based on all the data that we have, and with trances A, B & C acting as a safety buffer, the bonds in tranche D were considered AAA because the computer models showed that they were virtually bulletproof. No chance of defaults hitting tranche D. These are the ones the Swiss, with their desire for safety and prudence, owned.

Computer had to be used of course because there were literally millions of individual mortgages and no one could be expected to examine every one. The banks that created them, the investment bankers who sold them and the hedge funds and financial institutions that bought them had faith that the models were right.

Until they were not.

Reality set in because the models were not built to handle the fact that thousands of minimum wage dishwashers earning $25,000 per year would be able to get $600,000 mortgages with no money down in the expectation of flipping the house and walking away with a large profit.

There are many reasons for the housing bubble and its collapse; the fundamental ones had to do with government policy. But computers played a large part. They persuaded people to abandon reason and substitute risk modeling generated by computers whose job it was to define risk.

The Swiss bankers believed that they were keeping an eagle eye on the risk they were taking with their money, but in reality they relied in computer models that failed, and failed spectacularly, bringing the world to the edge of financial collapse.

Computers can do that. They are wonderful machines which can take reams of data and give you answers with incredible precision. But precision is not the same as “right,’ as the bankers worldwide found out last year. They trusted their computers to tell them how close they were to the edge of financial ruin and the computers misled them, and they fell into the abyss.


In a very real way, you can blame the financial crisis on computer models.


Which brings us to man-made global warming theory which is nothing but carefully mined historical data fed into computer models. The only significant difference is that the computer models that the bankers used were good at predicting reality, and were amazingly accurate for decades, until reality and the assumptions built into the models drifted apart.

As far as I can tell, the global warming computer models have not had any predictive value since they were developed. And this winter, the model and reality have collided. The aspect of climate modeling that allows its acolytes to maintain their faith is that the models always predict the distant future. I wonder if we will have to experience a global climate crash along the lines of a financial crash before reality sets in and the models are exposed as a fraud and a delusion.

1 comment:

Anonymous said...

Great comparison. The masters of the universe, whether in business, politics or science, are rarely so.