Zdeněk Tůma, Governor, CNB
Forecasting Dinner 2006
Prague, 15th February 2006
Ladies and gentlemen, dear colleagues,
Last year, when I also had the pleasure of opening the Forecasting Dinner, I tried to discuss several points connected with forecast uncertainty and the role market forecasts can play as an information source in the decision-making of the central bank. Tonight, I would like to talk about the inevitable uncertainty that is associated with any attempt at forecasting, and about the fact that this feature of any forecast quite often seems to be overlooked.
The point I have in mind is that once you present your forecast, many people seem to take the numbers literally and tend to forget about their basic property, which is that they are based on the information available today. By definition, any forecast implicitly assumes that that there will be no surprises and that no unexpected shocks will hit the economy in the future. This follows the basic attribute that any good forecast should encompass, i.e. to be the most probable future scenario. However, we all know that we live in a world where surprises really happen and shocks really matter for economic development. Naturally, they force the economy off the forecasted path. It follows that all forecasters produce their forecasts based on the assumption that no unexpected shocks will occur, while at the same time being almost certain that something is going to surprise them. All in all, I must say I consider forecasting to be a rather difficult job.
In order to overcome the discrepancy between the assumption of no surprises in the future and the belief that something is certainly going to happen, forecasters have introduced the term 'forecast uncertainty". It is fair to say that forecast uncertainty does not cover real shocks, as these are something nobody can predict. It seems to me that uncertainty is more about your own opinion that there is a non-zero probability that a particular variable will move in different direction, but you are still not able to declare this to be the most probable scenario and thus your forecast. The definition of forecast uncertainty is thus more about the perceived risks than about the true shocks. The perceived risks are, however, far from being easy to share with forecast users, i.e. those who use the forecast for their decision-making. Indeed, my personal experience is that whenever, after a monetary policy board meeting, there is a governor's briefing where the CNB forecast is presented, it is relatively easy to present to the audience of journalists the forecasted numbers for inflation and GDP growth, but sharing the Board's uncertainty about these numbers is quite a difficult task. In fact, the difficulty in clearly conveying the risks is exactly why it may happen that the forecast users interpret the numbers as something that is going to materialize, completely forgetting that any forecast is conditional upon our present knowledge of the current state of the economy and a few basic rules that drive the economy forward.
At the same time, however, I consider personal perceptions and evaluations of the forecast's risks, i.e. the forecast uncertainty, as the main force driving the decisions of the forecast's users. Indeed, in good times, the perceived risks immediately start to build up, as we are all suspicious that this state of affairs cannot last for long. So, we try to guess which factor will be the one to move the economy out of good times and to decide in line with our evaluation of the risks. Anyone who would have been able to anticipate (at least partially) the Mexican or Asian crises would have made the right decisions with no doubts. But the rest of us could only wait and see which of the risks we were thinking about would actually materialize.
Having all this in mind, I see the evaluation of risks as something that is really critical for decision-making no matter whether you are a central banker deciding about monetary policy or a private investor deciding about your investment. We should remember this whenever we look at any forecast and always ask ourselves for a risk analysis and risk evaluation.
The topics of both today's main speeches lead me to a practical example. There is no doubt that the uncertainty and risks surrounding both the global labour market and the expected development of EMU are crucial to our decisions. For central bankers, for instance, the price decline that follows the relocation of industrial production to countries with low labour costs is a completely new phenomenon. If the risks of this development are not evaluated properly, monetary policy may lose credibility by systematically undershooting the inflation target. Similarly, for you, it may pose a risk of biased expectations about the average level of interest rates in the economy.
Ladies and Gentlemen, let me wish all of us the best possible risk analyses, so that we are only rarely caught completely off guard.