Experts do not get too MAD

Economic forecasters don't get it wrong, they get MAD - mean absolute deviation. This is the measure of the difference between what they say will happen and what actually does.

And on the evidence, the Economic and Social Research Institute (ESRI) did not get too MAD in the period 1994-2004.

The mean difference between its predictions of economic growth and the actual rate was 2.2 per cent across the 11 years, according to an assessment of its own forecasting published with yesterday's Quarterly Review.

For the most part, and particularly during the boom in the late 1990s, the ESRI admits that its predictions were too cautious.

For example, in 1999, it predicted growth in gross national product (GNP), which is the value of goods and services produced in the State, of 5.75 per cent. But the actual rate turned out to be 8.5 per cent.

The following year, the ESRI went for 5.9 per cent growth, but the real result was a still unbeaten record of 9.5 per cent.

But in the more volatile conditions that followed, the ESRI got it the other way around. For 2001, it expected 6.1 per cent, but we got 3.9. The following year it forecast 3.1 per cent growth, but it actually reached 2.7 per cent.

The old pattern has since reasserted itself, albeit with less deviation. Last year, the ESRI predicted 3.3 per cent growth, and we got 4 per cent.

As growth outstripped expectations, unemployment also fell faster than the ESRI predicted. But its forecasts were more accurate than for GNP growth. In 1999, it said the Republic would end up with 6.5 per cent unemployment, the rate was 5.6 per cent. In 2000, it predicted 4.7 per cent joblessness and the final figure was 4.3 per cent.

It was also a lot closer when it came to inflation, but also tended to underestimate the rate at which prices would rise in the Republic's economy. In 2000, it forecast 4.0 per cent inflation, but the actual rate was a record 5.6 per cent. In 2004, it predicted 1.8 per cent, but the outturn was 2.2 per cent.