At the heart of any earnings-at-risk measurement is a base forecast or projection.  In addition to the base forecast we apply some sort of stress-test (usually a rate shock up and down).  I frequently come across bankers who endeavor to get the stress-test “right” or “wrong”.  They are convinced that they can design a “realistic” projection and test to measure their exposure.

As I’ve mentioned in more than one of my presentations and also here on my blog – like it or not this projection and stress-test often times comes down to something only slightly better than crystal ball gazing.  There’s a good article in last Friday's New York Times with an introduction that sums this up quite nicely.  The article itself is an opinion about increased regulatory oversight.  It is a warning that the country should not be overly optimistic about how successful improved oversight will be.  The initial comments about the reliability of forecasts are quite insightful:

"Trying to Tame the Unknowable" - NY Times, March 26, 2010

Perhaps the best place to start is to acknowledge what we cannot do. If recent events have taught economists and policy makers anything, it is the need for humility.

One thing we cannot do very well is forecast the economy. The recent crisis and recession caught most economists flat-footed. This is nothing new. We have never been good at foretelling the future, but when the news is favorable, others forgive our lack of prescience.

Some critics say the Federal Reserve should have foreseen the bursting of the housing bubble and its financial aftershocks. A few of them, having made the correct call themselves, are enjoying newfound celebrity.

Yet at any time, there are many forecasters with a large range of views. After the fact, a few will turn out to be right, and many wrong. Policy makers at the Fed don’t know in advance who will be the lucky few. Their best course is to rely on the consensus forecast and to be ready for the inevitable surprises.

from www.nytimes.com, March 26, 2010