In May of 1968, the U.S. Navy lost a submarine as it was going in for repairs to Newport News, VA. While many commanders suggested the best way to find the missing sub was by using instinct, Navel Officer John Craven came up with a different solution. He gathered estimates from a diverse group of experts and asked them to answer 3 questions: 1) what was the cause of the malfunction, 2) what was the sub's speed at the time of trouble, and, 3) what was the sub's angle of decent. Craven then took all the estimates and averaged them to project where the sub could be located. Using Craven's data, the Navy found the ship 220 yards from where he had predicted it would be.
Interestingly, no single estimate given to Craven was accurate, but the collective judgment was. This case study was in the back of our minds when we heard of a mid-sized bank that does quarterly polling on 25 of its line personnel (regardless of title). The bank uses the data to form its interest rate consensus and offers small prizes for those that end up closest to reality. In addition to creating an inexpensive rate forecast, the process serves to get the whole bank talking about interest rates and gives each employee a sense of ownership in the process. This bank claims great accuracy. We will quickly add that we have tracked the top 40 market economists for a long time and have not been impressed by any single forecaster's accuracy. In fact, a substantial number don't even get the direction right.
Because of our research, we adopted a "group" methodology as well. We average the top 40 economists and combine it with forward market rates. We publish this data monthly in our ALCO economic update. In using a methodology like this, it is important not to throw out the outlier predictions, as these extreme forecasts are often what give the average its accuracy. We highlight these forecasting methods as a better alternative to using the forecast of a single source (such as a local University or broker-dealer).
If you are interested in seeing both the forward rates and the compilation of economist predictions of rates to aid in your planning, simply let us know and we will sign you up for our complementary Monthly ALCO package. This package is concise and timely and is delivered via e-mail on approximately the 3rd day after the end of the month. In addition to economist surveys from around the nation, the package also has narrative on recent activity in securities, loans, CDs and capital.
Of interest to many banks, the package also contains average loan pricing by maturity and sector. This gives banks a data point of where certain par priced loans were originated during the prior month. Similar to the Craven analysis, data is gathered from loan pricing from around the country, bringing you the collective pricing insight of a large group of community banks.
While we are sure you get many sources of information each month to share internally, when it comes to forecasting, sometimes having a collective insight produces a superior result.