Tuesday, April 7, 2009

A Bridge to Where?

Canadian economic forecasters have recently been drastically lowering their expectations for Q1 GDP from simply a very bad quarter, say around -4% to -5%, to expecting one of the worst quarters in Canadian history. So what gives? Is incoming data really that bad? How can one reconcile current data with quarterly forecasts?

Well, it turns out that there is a fair amount of literature on updating quarterly forecasts with monthly information. One technique that caught my eye (because of its simplicity) is the construction of a so-called “bridge” equation. A bridge equation is basically a simple regression of monthly economic variables, aggregated to quarterly values and then used to forecast quarterly GDP. An example of such an equation can be found in this Bank of Canada working paper.

The model in the BoC paper includes the consumer confidence index, hours worked, retail sales, housing starts, 3-quarter lagged Canadian GDP and US industrial production as explanatory variables. Based on the most recent data for those variables, the bridge model predicts Q1 real GDP growth of -8.3%. This compares with a quarterly model prediction of -4.4%.

So, using the bridging equation we end up at the same bad place arrived at by Bay Street and others– a really really really bad Q1.

No comments: