Risk Modeling: Can we blame Analytics for the current status of economy?

Scenario analysis in business involves investigating multiple alternative events and their impact on the final outcome. For example if I am trying to predict the revenue of Motorola in US based on growth in demand for a certain high-end model, than I need to look at alternative growth figures to create several possible scenarios and predict revenues for each of those scenarios. Hence if the likelihood of revenue growth is 8% this quarter based on a 15% growth in the new model, than I must look at all other possible growth figures in addition to the 15% number.This will provide alternate figures of revenue growth. Such analysis is quite common in the industry and decision makers often look at these alternatives and plan accordingly. 

Banks and other financial institutions whose business are dependent on prices in the "real estate" market must have looked at multiple scenarios of price change. Now the question is how do we determine these alternative growth numbers? Historically the housing market has experienced steady growth (in prices) due to persistent government policy of encouraging higher house ownership tom-toming it as a sign of prosperity. If the growth number is always positive and varies for example between say 1% to 15% on average, than an analyst is going to use only these numbers in the stress test. Consequently he/she will most likely miss out the -25% drop in average price that we have experienced in the last year. 

Now any decision based on the lowest growth of 1% is unlikely to cover the risks associated with a 25% drop in price. Is this how our sophisticated analysis failed to prepare the banks adequately for the present slump in real estate price and account for the risks associated with it? I am neither an expert in this process nor aware of the exact data that are used to perform the risk analysis. However inability to visualize an event which has never occurred is really a difficult situation. It is a challenge for the analytics community to look beyond historical data in a significant way, and apply those scenarios into the decision making process. 

Comments are Welcome!

Print this Article