Poor track record for economic predictions

Economics: According to figures released this week, the Exchequer recorded a budget surplus of €33 million in 2004. This compares with the €2.8 billion deficit projected 12 months earlier. The big difference between forecast and outturn is a clear reflection of the turnaround in economic conditions in the intervening period.

There is lots of other evidence to show that the economy ended 2004 in much better shape than at the end of 2003.

For example, employment in late 2004 looks to have been growing at about twice the rate of a year earlier.

A similar picture emerges from aggregate measures of economic activity like GDP and GNP, which grew at annual rates of 5.8 per cent and 4.2 per cent respectively in the third quarter of last year.

These figures compare with equivalent rates of 2.8 per cent and 2.3 per cent in the second half of 2003.

On the other hand, it is evident that the economy is not firing on all cylinders. There is at least one notable area of persistent weakness.

The latest available data show that the industrial sector remains in decline. In the three months to October, the volume of industrial output was 4 per cent below its year-earlier level, with output volumes in the so-called "high-tech" sector down almost 6 per cent.

Not surprisingly, industrial employment continues to contract: in Q3, it was 3 per cent lower than a year before and 13 per cent below its peak of early 2001.

Merchandise exports remain very sluggish (in contrast to exports of services, which are rising strongly) and appear to have been growing at a slower pace in late 2004 than a year earlier.

Whatever about the travails of industry, the "consensus" view amongst analysts is that the impressive pace of overall economic activity that built up as 2004 developed will be broadly maintained in 2005 and 2006.

The December survey of Irish economic forecasters polled by Reuters shows that, on average, they expect GDP to grow by 5.2 per cent this year and by 5.6 per cent next year. They envisage that this will be accompanied by a further decline in unemployment and the continuation of moderate rates of inflation.

The average forecast for the annual rate of CPI inflation is 2.5 per cent in 2005 and 2.7 per cent next year.

It is worth recording a couple of observations about the forecasts included in the Reuters poll.

First, there is a remarkable sameness about them. For example, the most bullish forecast for 2005 envisages GDP growth of 6 per cent; the most bearish is a forecast of 4.5 per cent. (The range of GNP forecasts is 4 per cent to 5.6 per cent.)

When set against the economy's actual performance over the past five years when GDP growth ranged between a low of 3.7 per cent in 2003 and a high of 9.9 per cent in 2000 (the equivalent range for GNP growth is 1.5 per cent in 2002 to 10.1 per cent in 2000), this represents a pretty tight set of expectations.

Second, the forecasts are clustered around estimates of the economy's potential growth rate.

In other words, what the forecasters - pretty well without exception - are saying is that the economy will be growing on trend for the next couple of years.

This is a happy picture.

Should we be reassured? What has been the record of short-term economic forecasters over recent years?

To answer this question, I have looked at the range of independent forecasts published by the Department of Finance in the Budget booklet for each of the years 2001 through 2004.

This is a partial exercise because it includes only the forecasts of State or international organisations (the Department of Finance itself, the ESRI, the Central Bank, the OECD and the European Commission).

I can't say for sure whether including the forecasts of private sector firms would materially change the outcome, but I suspect not.

Anyway, the most important finding from this exercise is the following. For each of the variables covered (GDP growth, GNP growth, employment growth and CPI inflation) and for each of the years surveyed except one (2003 in the case of GDP, GNP and CPI; 2001 in the case of employment), the actual outturn fell outside the range of forecasts - in many instances, well outside that range.

Take 2004, for example. At the end of 2003, the forecasts of GDP growth for this year fell within the range 3.2 per cent to 3.7 per cent. The outcome, which of course we cannot yet be sure about, looks like being close to 5.5 per cent.

For GNP, the likely outcome of almost 5 per cent growth compares with a forecast range of 2.2 per cent to 3.1 per cent. For CPI inflation, the range of forecasts was 2.6 per cent to 3 per cent; the outcome 2.2 per cent.

While for employment growth, the forecast range was 0.7 per cent to 1.3 per cent compared with a probable outcome of about 2.5 per cent.

The lesson is that short-term forecasters often get it wrong.

Indeed, the evidence from recent years is that they get it wrong more often than they get it right, and they all get it wrong together.

The reasons are that accurate short-term forecasting is extremely difficult and that short-term forecasters, because of their aversion to being wrong and isolated, tend to hunt in packs. (I know more about this than most because I did short-term forecasting for the best part of 20 years.)

As far as 2005 is concerned, therefore, the chances are that the forecasts reviewed above will be wrong again and that the economy's performance will surprise once more.

A key question is whether that performance will be surprisingly good (as in 2004) or surprisingly poor (as in 2001). In other words, if GDP does not grow at a rate within the range 4.5 per cent to 6 per cent over the coming year, is it more likely to grow by something like 2 per cent to 3 per cent, or by something like 7 per cent to 8 per cent?

Or, put another way again, are the risks to the economy more heavily stacked on the downside than the upside?

Regrettably, I'm not paid enough to pose this question and answer it in the same article, so I'll come back to it at a future date. In the meantime, you can put it to your favourite short-term forecaster.

Jim O'Leary is currently lecturing in economics at NUI-Maynooth. He can be contacted at jim.oleary@may.ie