PRI's The World: Technology Podcast 211

September 22, 2008

This week's tech podcast (WTP 211), we assure you, makes absolutely no attempt to play up the fact that today is International Talk Like a Pirate Day. No, instead, we begin the podcast with a sea shanty, a regular wet-blanket of a dirge, of a different sort. We mourn the loss of our retirement accounts, and Pink Floyd's Richard Wright, in equal measure. Just listen in...it all fits together somehow.

The big news is, of course, the meltdown currently plaguing the US financial markets. The curse of being a journalist is that you get curious, maybe too curious, about some things. You see, I'd heard tell that there exists some different types of risk management software that these financial giants use to, well, help them NOT get into these kinds of dire straits. I couldn't help myself. I wanted to find out if it was the machines, or the people feeding the machines. So, I called up Richard Lindsey, co-author of a book called How I Became a Quant. I now see why quantitative analysis is both an art and a science.

And let's say that you did go and look at your retirement account this week. That might make you angry, or depressed; or, you may be gleefully hugging all those gold doubloons (oops, pirate reference) you've got stashed under your pillow. But if anger or depression are your fate, you can log onto this website. My Mobile Guru will send messages of comfort and encouragement directly to your cell phone. All you have to do is send a text message...oh, and be in the United Kingdom. Sorry.

There's a sampler at left, and an audio sampler on the podcast...which is well worth the listen. Six bucks may sound expensive for cell phone therapy, but hey, think about how expensive a shrink can be. Personally, I can't believe the Brits came up with this one first. Deepak Chopra, where are you when we need you...now...on our cell phones? The BBC has a discussion between the website creator, and a licensed psychotherapist.

We then pack our bags for India, where we hear about another website looking to help out. Shaadi is the Hindi word for marriage, and the website looks to bring busy Indians who don't have time for the usual arrangements together online.

And we end with news near and dear to my heart. The Hitchhiker's Guide to the Galaxy is one of my all time favorites, and I'll admit to complete and utter devastation when Douglas Adams passed away in a very untimely and selfish manner in 2001. H2G2 is great in just about any form, be it radio, TV, film, or the "trilogy" of five books. Two previous podcasts should convince you of my near fanboy-dom (WTP 18 and 19).

At the time of his death, Adams was writing, possibly, a sixth book to the Hitchhiker's trilogy. But now it's been announced that a sixth book has been officially commissioned. It will be written by Eoin Colfer, the Irish author of the popular Artemis Fowl series of books. The BBC interviews Colfer, who enters into the H2G2 world with equal parts excitement and intimidation.

Someone get the man a towel, and remind him:

(Screen grabs courtesy of mymobileguru.co.uk, and the home of The Guide)

Comments

[Kristen]

Gosh, I listened to this podcast on my shuffle while running and it got me so fired up I think I must have run a lot faster. I'm not sure if Richard Lindsey has some vested interest in misleading people or what, but I could not believe what I was hearing a supposed "quant" say. I'm no quant, I have a Masters in math and traded options for a few years, but it doesn't take a rocket scientist to understand how risk management was either purposefully manipulated or, more likely, was ignored. In trading, the risk management software helps the trader know where his or her risk in given certain circumstances. For example, on our sheets (yes, I'm dating myself, paper sheets) we would have on the bottom the amount we would make or lose given a ten percent, twenty percent or forty percent move in our portfolio of stocks. If those numbers got too big, the risk manager would give you a call and you would either have to justify the risk or hedge some it to reduce those numbers. Now, of course, some of those inputs into the algorithms were not rock solid, like historical volatility if a stock were new, but overall, unless you purposely lied, the algorithms would tell you the risk of certain things happening within a certain time frame. The liquidity point is interesting and certainly at a point of singularity like the current two weeks where there is almost no liquidity it cannot be measured, but normally that is a risk factor that is taken into account by risk managers and traders. If we had a stock that only traded a thousand shares a day, we knew we had to keep our markets wide and pricey because if you bought calls and had to sell stock to hedge you could send the stock price down by dollars because no one was trading it. So, I can only assume that in mortgage trading the illiquidity is measured and factored in to any risk assessment. And I am sure the science behind mortgage risk management is well-established since banks have been doing it for hundreds of years. So all it would take to measure the risk of one of these mortgages would be this. First, figure the probability of a family who makes X being able to pay Y mortgage at Z interest rate, which they have for their 5 year arm. I guarantee the banks have that figured even down to the zip code. Then, you say, ok the family has a 5 year arm, given the historical interest rates, what is the most likely interest rate for years 6-10 and what is the expected monthly price of the mortgage. Given this price what is the probability that the family who makes X (plus whatever inflation) will default. Again, banks have these kinds of things figured out quite well. They could even figure out the historical house prices in the zip code and what the likelihood is of the value of the house diminishing. Then all the quant has to do is amalgamate all that information and figure out the probability of defaults over time. All these numbers were available. The math is not hard. So it wouldn’t have been difficult for the banks to do risk-assessment on the packages of mortgages they were buying because presumably for each mortgage these numbers existed. Then to end with the doctor metaphor, that put me over the top! One would hope your doctor would have the risk assessment information to at least tell you that if you are this age and this percentage overweight and smoke and drink then you have a 60% or whatever chance of developing heart disease. I hope doctors at least tell you that. No, he or she can’t say you will get cancer on the third Thursday of the fifth month of 2010, but we are not asking risk assessment to do that. We are asking them to say given a move of different magnitudes what is the probability of loss and what is the probability of moves, based on historical trends, within certain time periods. So nothing was very hard to figure out. Either the banks purposefully ignored the risk assessments, which is likely because risk assessors don’t bring in the big bucks so banks tend to let traders, who do, do what they want or for some reason risk assessment didn’t do their jobs. Everyone knows that packaging all of the same types of mortgages doesn’t mediate the risk. Just like buying ten of same types of stock does not give you a balanced portfolio. It is simply silly to say that good basic risk management couldn’t have prevented this debacle.

[Kristen]

Now where the models get complicated is how much better they do all this than what I described. My caveman model above is done on five year intervals, but their models, using complicated algorithms can figure it out moving day by day and can include even more moving inputs to give much more accurate numbers, but even a back-of-the-envelope type of model would have measured an immense amount of risk exposure.