MANY of the assumptions underpinning the Budget we shall hear today will be based on computer forecasts of the economy for the next 12 months or so. Like weather forecasts made by a computer, such predictions are very often right, but sometimes wrong, and many of the potential sources of error in an economic forecast have a direct analogy in meteorology.
Weather forecasting by computer is based on a mathematical model of the atmosphere - a description of its behaviour expressed in mathematical equations. Observed values of pressure and temperature at thousands of places in the atmosphere are fed into the machine and processed by this model, which then calculates expected values of these variables for some future time.
Economic forecasting is similarly based on a computer model which describes, essentially, human behaviour. Economists have distilled their beliefs and theories about how people act and react into a large number of equations which encapsulate statistical relationships between a myriad of economic variables. For example, it might be assumed that a 1 per cent rise in the total amount of wages and, salaries paid in Ireland in a given year will lead to a 1 per cent rise in the amount that people spend.
Obviously, in the case of both economics and the weather, the forecast will be wrong if the basic model is inadequate - which may well sometimes happen. But both predictions must also be based on accurate and timely data. In economies this comprises information about, for example, how people have spent their money in the past, how much they have earned, and how much they have saved; in a weather context, observations are the crucial data, and meteorologists know only too well that if the original observations are too sparse or are inaccurate, the resulting forecast will be poor.
Errors in economic forecasts also arise from assumptions made about "external" parameters - assumptions, say, about future oil prices or interest rates that are fed directly into the computer. To find the analogy for these in meteorology, we have to go to climate modelling, where the external variables might correspond to future estimates of CO2 emissions, or other man-made additions to the atmosphere. And the final reason why economic forecasts may be incorrect is if the interpreter, perhaps because of a nasty experience in the past, simply does not believe what the computer tells him, and uses personal judgment to nudge the final figures up or down. This happens, too, in the world of meteorology - and just occasionally, of course, the hunch pays off, just as it sometimes does in economics.