Treasury admires its image in the mirror

Treasury has reviewed its own forecasting record, presumably in an effort to improve performance.

It found that:

Budget forecasts of nominal GDP growth exhibit little evidence of bias over the past two decades, with the average Budget forecast error being insignificantly different from zero over this period. That said, an examination of the patterns in forecast errors reveals a more variable performance, with the forecast errors being correlated with the economic cycle (Figure 1). Hence, with the benefit of hindsight, Treasury has tended to underestimate growth during economic upswings and overestimate growth during economic downturns.

The finding that Budget forecasts of nominal GDP growth exhibit little evidence of bias is in contrast with those of a recent study by Jeffrey Frankel of official government real growth rate (and budget balance) forecasts between 1985 and 2009 in 33 countries (including Australia).

There is a pattern here. Treasury has much more consistently under-estimated growth during Liberal government periods and overestimated in Labor government periods. Treasury’s own explanation, that its methodologies do not work so well at cyclical turning points, seems fair enough.

The Review also finds that Treasury’s macroeconomic forecasts have been reasonably accurate. Over the past two decades, Budget forecasts of nominal economic growth have exhibited a mean absolute percentage error (MAPE) of 1.6 percentage points. Treasury’s macroeconomic forecasting performance is comparable with that of other domestic forecasters. In fact, the overriding impression of the forecast errors of Treasury, the Reserve Bank and Deloitte-Access Economics (Access) is the similarity in the error patterns across agencies (with errors for each agency exhibiting significant variation across time).

On the other hand, Treasury’s forecasts are comparable with, or better than, those of official agencies overseas, although some caution is required in making cross country comparisons over a period as short as ten years, and given that official agencies prepare forecasts at different times in the year:

I don’t know about you but the way I read this chart, Treasury looks more consistently wrong than everyone but New Zealand. Performing better during one outlier year does not offset a decade of under-performance.

Economic forecasting is a fools game but it looks like US, Canadian and especially UK fools are better at it most of the time.