<<<	articles
The naked economic forecaster Samuel Brittan: Financial Times 03/01/03 New year is the time for crystal-gazing about the future and for spoilsports to have fun in exposing the abject failure of past prophecies. I should like to do something more modest, which is to examine, in all its nakedness, the basis of the short- to medium-term forecasts of the whole economy that - whether we like it or not - lie at the heart of economic policymaking. Such forecasts are neither mumbo-jumbo nor science but have a simple core that is disguised by all the mathematics. There are three essential elements. The first is an estimate of the trend rate of sustainable output growth. The second is an estimate of the "output gap" indicating how far from trend the economy is. Third, there is the estimate of how fast we are moving towards or away from the trend path. There are, correspondingly, three ways in which forecasts can go wrong - quite apart from the effects of international catastrophes such as oil price explosions. They can overestimate or underestimate the extent and speed to which the economy is moving towards or away from trend. They can make a false estimate of the trend rate itself. Or they can make a mistake about the sustainable level of economic activity, looked at in terms of physical capacity or the unemployment rate and related labour market indicators. Most vulgar attention is devoted to the immediate movement away from or towards trend. This can be affected not only by underlying forces but also by all kinds of temporary forces, such as changes in indirect tax or in the prices of imported commodities. The Bank of England rightly believes that it would be both difficult and harmful to try to offset such temporary shocks and that if the economy can be steered towards trend growth, these short-term factors will come out in the wash. The further ahead you look, the more the fundamentals dominate the picture. That is why it uses a two-year horizon. Short-term mistakes in predicting the immediate prospect for demand are in my view the least important source of error. If the central bank gets this wrong, it can remedy matters next time round by easing or tightening policy. It probably would not greatly matter if it mainly looked at what is happening now, which is no easy matter to discover. The second, and more serious, source of error is to underestimate - or, more usually, overestimate - the trend rate of growth. The latter is determined by the growth of productivity plus the change in the working population. In the UK, productivity growth has been astonishingly stable for many decades despite all government efforts to improve it. The big source of error has been in estimating changes in the active labour force, whether caused by immigration or changes in participation rates. Another source has been wishful thinking by governments that believe that they have fundamentally improved productivity. Owing to the wonders of compound interest arithmetic, modest errors of a fraction of a per cent per annum in estimating the trend growth rate can add billions to government deficits, even on a cyclically adjusted basis. The third, most important and least discussed source of error is in estimating the sustainable level of economic activity. Without such a view it is impossible to estimate the output gap. Overoptimism here can lead to accelerating inflation, and excess pessimism to errors of a deflationary kind, although because of labour market rigidities there is not the neat symmetry that some economic theorists like to posit. Lord Burns, who was a distinguished academic forecaster before he became a chief economic adviser to the government, has remarked that the forecasts have not become any better over the 30 or more years he has been watching them, despite improved techniques and much larger personnel. The world is an uncertain place and no amount of effort or cleverness on the part of the forecasters will prevent significant mistakes. Business people who criticise inaccurate forecasts are oblivious of the fact that they are paid - pretty generously - to deal with uncertainty; and while good policy can reduce unnecessary uncertainty there remains an irreducible amount - which, incidentally, is also found in hard sciences such as physics. Many official economists would be delighted if they could dispense with conventional forecasting. But unfortunately attempts to use short-cut methods such as money supply targets have come to grief. Monetarism was never intended for short-term steering. The basic contention was merely that sticking to a modest and stable growth of the money supply would achieve low average rates of inflation over a period and help to reduce booms and slumps in real output. Unfortunately there has never been enough confidence in the behaviour of any particular measure of the money supply for it to be used in this long-term way. My own view is that the core model of trend and output gaps is too useful an aid to thinking to be dispensed with altogether. But the detailed forecasting process needs to be relegated to a minor role. The chief enemy of reform, however, is the insistence of so many people on a crystal ball for viewing the future. They might as well use tea-leaves or consult a pet astrologer.
<<<	articles
Site designed and managed by Andrew Heavens - aheavens@ftnetwork.com