Why you shouldn’t trust forecasts…
…but, you need them anyway. A simple example showing why financial market forecasts aren’t worth much except in terms of historical forensics.
- Recently TD Bank suggests that oil prices will be 30USD in the 2nd quarter of 2009.
- In their prior forecast they suggested that the price would be closer to 45USD.
- In a report on agriculture from November 2008 they suggest that oil will average 65-75USD during 2009.
- In a report from April 2008 they had an outlook for prices still in the 80-90USD range for Q1 2009.
What does this mean? That TD doesn’t seem to have much luck in forecasting the price of oil and, mostly, that forecasts are bullshit.
Does this mean that forecasts shouldn’t be done? No, just that betting the house on the surmisings of a bank economics group isn’t a good idea. Forecasts are decent at illustrating trends but are generally very poor at accuracy. They’re mostly good at determining, after the fact, where your analysis was wrong so you hopefully don’t make the same mistakes twice. Of course, given the turnover in a bank’s economics group, there’s likely to be a loss of experience over time making the same mistakes reasonably likely.
Of course someone, somewhere will get a forecast right and they’ll be the prediction guru for the next while.
Anything here rankle you? Feeling overly perturbed or elated? Leave a comment below. or subscribe to the Sauce Captain feed.
Well put.
-Jonathan