I posted about this already before finding this article, which in fact is about one of the best I have seen that helps explain the technical side of things
In healthcare software there’s risk management too, so can you imagine the programming to be set up to create an entire practice of sick patients to produce more money? There is a certification process with healthcare software, and without standards and someone auditing the entire process it would be possible. The overall business intelligence would create the results on a dashboard based on the prerequisites programmed as well as reports.
Were there any standards or audit trails here? In software there’s always an area to change settings to customize a software application to where it works best for the user and those features are built in, but the over all calculation queries are run at the server level, beyond what the user can do from their PC computer, and the comments from the IT folks here explain a bit more of the process, with many throwing their arms up with what they were presented with to create the ultimate outcome for the brokers and staff to use.
Programmers by nature want everything to work properly and know what accuracy is, otherwise as coders, they are the ultimate providing source, so here there were IT departments being presented with code they did not write, but contained the algorithms written by those higher up to program into the calculation processes, some IT folks were put in a pretty touchy spot to say the least here, program or no job, and as some mention below some of it was pretty sloppy, which being a coder myself, when I get in a hurry, I can certainly turn out some pretty sloppy stuff too. I know exactly how they feel though as far as being put in situations where I could not create the desired result, right or wrong in nature some stuff just won’t fly. BD
Below are some of the comments made from the IT side of the coin, and what is presented to the programmers and folks running the networks and I am glad many commented and spoke up here.
1. “Been in rooms where as an IT guy, where we had some portfolio manager explain their process, so we could automate it. After he left the room, there was dead silence, we looked at each other in shock and bewilderment, and not a word was said. Needless to say, we are no longer working there. I agree to the willful simplicity of the models, many knew that the assumptions were crazy. There is a mathematical measure that can be used to determine the health of a mortgage borrower, it is called the FICO score, for many portfolios there was no way of incorporating that into the model. The FICO score measures the health of a borrower, and the goodness of the score it is updated frequently to account for the borrower’s financial health.
2. Like some of the other people commenting here, I was one of the IT people that tried to model the work the quants were doing. On more than one occasion we were given computer code that the quants had already written to model the risk. The quants were already using the code, we were just supposed to put it in a production environment. Most of the time the code didn’t really do what the quants thought it was doing. As smart as they were supposed to be, they wrote some of the worst computer code that I have ever seen. One one occasion, I pointed out an error to a quant that has a Ph.D in physics. It turned out that he was underestimating risk by over 10%.
Part of the problem is that these quants think that they are the smartest people in the world and don’t have any patience. That attitude would put people like us (IT people) under extraordinary pressure to try and do things that were impossible. We were never given the proper time to vet the work that they had done or to try to understand the nuances of what their models and trading strategies were.
3. The long and short of it is that complicated mathematics do not necessarily equate to accuracy. We will always need seasoned veterans to way in with a qualitative analysis, something that seems to have been ignored this time around.
4. It’s just GIGO all over again. Computers just make it worse. I know; I spent my career in I.T. No amount of fancy formulas substitutes for judgment and experience. It appears neither were applied here and we are all paying the result.
5. The fact that firms could sell these securities only fueled the most liberal of lending terms. Most of the commenters are dead on. As a former regulator I do wonder where the IT auditors were? Well maybe investment banking firms didn’t bother with them.
6. Maybe the SEC should set standards for these systems that assure that the software systems can warn the firms of extremely risky positions, instead of leaving it up to the firms’ management to set the standards.”
Before I started covering the Internet in 1997, I spent 13 years covering trading and finance. I covered my share of trading disasters from junk bonds, mortgage securities and the financial blank canvas known as derivatives. And I got to know bunch of quantitative analysts (”quants”): mathematicians, computer scientists and economists who were working on Wall Street to develop the art and science of risk management.
In other words, the computer is supposed to monitor the temperature of the party and drain the punch bowl as things get hot. And just as drunken revelers may want to put the thermostat in the freezer, Wall Street executives had lots of incentives to make sure their risk systems didn’t see much risk.
Lying to your risk-management computer is like lying to your doctor. You just aren’t going to get the help you really need.