3 ms·

After the Crash: How Software Models Doomed the Markets

6 points by prakash 18y ago

mixmax 18y ago
A friend of mine does financial software modelling that is used in a number of major banks. According to him these models have predicted a major crash for the last couple of years. They weren't able to see when the crash would occur, but they could see that certain relationships such as household income/mortgage payments were diverging further and further from a historical mean.
- pasbesoin 18y ago
  Just observing the property value increases in my area made me nervous. I could observe no inherent basis for their magnitude. Housing was experiencing a surge not reflected in the broader economy; it was feeding upon itself. The penultimate "house of cards". There was no sustaining model for the demand. The population wasn't increasing that rapidly, and people's productivity (or earnings) wasn't rising in a fashion to enable them to afford to occupy an ever increasing quantity of living space.
  - allenbrunson 18y ago
    nitpick: the word "penultimate" means "next to last."
    - pasbesoin 18y ago
      Thanks. Guess I'm getting a bit sloppy in my phrasing. Or maybe it was my subconscious? If so, I'm not sure I want to know / learn what is the ultimate house of cards. The current one is already proving challenging enough for me. ;-)
  - gamble 18y ago
    Interest rates were coming down, particularly since 2001. Given that people tend to size their house purchases to match the monthly payment they can afford, you would expect home prices to rise as interest rates fall. Whether that's sustainable in the long term is another question.
lutorm 18y ago
I bet Nicholas Nassim Taleb is getting a lot of calls these days... Seems his criticisms of economic quantitative modeling have been vindicated more than he probably wished to see.
run4yourlives 18y ago
Seriously? Any chimp with half a brain - including me - saw a crash coming. I didn't need software models, and I may have been unable to pinpoint it, but all you had to do was use a little common sense. When you're pretty confident that you're making market - more or less - and yet you can't possibly understand how your peers afford $500K houses and 2 $40K cars, the writing is on the wall.
jerf 18y ago
The more I learn about this crisis, the more I am convinced it all came down to one thing, and one thing alone: leverage. Now, the situation is obviously complicated and can't be pinned down to one factor like that, so let me be clear: I'm not saying that one fact explains every last detail about the crisis. What I am saying is that the grotesque overleverage we saw was sufficient to cause a crisis. Exactly how it manifested depended on computer models and dumb loans and all the other little details. But the situation where banks and large financial entities were leveraging 30x meant that a crisis of this type was inevitable; the only question was the details. When you set up a scenario such that any 3% drop in asset value has driven you bankrupt and 4% means you're in horrific debt, you have already lost. How exactly this loss manifests, and how many smart people there are with computers trying to put off the day of reckoning there are hardly matters; epic failure (and I mean that in the original sense, not the meme sense) is inevitable. Did the geeks tell the banks to overleverage? Quite possibly. If so, then yell at them too. This article doesn't support that claim, though. (In fact it's very vague in general.)
sh1mmer 18y ago
There was an article on HN a month or so I can't find a link to which discussed how a number of major banks didn't properly update their models to knowingly take greater risks. It explored how a number of banks were using outdated models to allow them to bundle risks to create short-term profits at increased risk without upsetting the feds. So, sure, the models were wrong but was the reason the inability of the modelers?

6 points by prakash 18y ago

A friend of mine does financial software modelling that is used in a number of major banks. According to him these models have predicted a major crash for the last couple of years. They weren't able to see when the crash would occur, but they could see that certain relationships such as household income/mortgage payments were diverging further and further from a historical mean.

pasbesoin 18y ago
Just observing the property value increases in my area made me nervous. I could observe no inherent basis for their magnitude. Housing was experiencing a surge not reflected in the broader economy; it was feeding upon itself. The penultimate "house of cards". There was no sustaining model for the demand. The population wasn't increasing that rapidly, and people's productivity (or earnings) wasn't rising in a fashion to enable them to afford to occupy an ever increasing quantity of living space.
- allenbrunson 18y ago
  nitpick: the word "penultimate" means "next to last."
  - pasbesoin 18y ago
    Thanks. Guess I'm getting a bit sloppy in my phrasing. Or maybe it was my subconscious? If so, I'm not sure I want to know / learn what is the ultimate house of cards. The current one is already proving challenging enough for me. ;-)
- gamble 18y ago
  Interest rates were coming down, particularly since 2001. Given that people tend to size their house purchases to match the monthly payment they can afford, you would expect home prices to rise as interest rates fall. Whether that's sustainable in the long term is another question.

lutorm 18y ago

I bet Nicholas Nassim Taleb is getting a lot of calls these days... Seems his criticisms of economic quantitative modeling have been vindicated more than he probably wished to see.

run4yourlives 18y ago

Seriously? Any chimp with half a brain - including me - saw a crash coming. I didn't need software models, and I may have been unable to pinpoint it, but all you had to do was use a little common sense. When you're pretty confident that you're making market - more or less - and yet you can't possibly understand how your peers afford $500K houses and 2 $40K cars, the writing is on the wall.

jerf 18y ago

The more I learn about this crisis, the more I am convinced it all came down to one thing, and one thing alone: leverage. Now, the situation is obviously complicated and can't be pinned down to one factor like that, so let me be clear: I'm not saying that one fact explains every last detail about the crisis. What I am saying is that the grotesque overleverage we saw was sufficient to cause a crisis. Exactly how it manifested depended on computer models and dumb loans and all the other little details. But the situation where banks and large financial entities were leveraging 30x meant that a crisis of this type was inevitable; the only question was the details. When you set up a scenario such that any 3% drop in asset value has driven you bankrupt and 4% means you're in horrific debt, you have already lost. How exactly this loss manifests, and how many smart people there are with computers trying to put off the day of reckoning there are hardly matters; epic failure (and I mean that in the original sense, not the meme sense) is inevitable. Did the geeks tell the banks to overleverage? Quite possibly. If so, then yell at them too. This article doesn't support that claim, though. (In fact it's very vague in general.)

sh1mmer 18y ago

There was an article on HN a month or so I can't find a link to which discussed how a number of major banks didn't properly update their models to knowingly take greater risks. It explored how a number of banks were using outdated models to allow them to bundle risks to create short-term profits at increased risk without upsetting the feds. So, sure, the models were wrong but was the reason the inability of the modelers?