Inaccurate forecasting

Experimental visualization of narrower problems
Other Names:
Unrealistic forecasting models
Irrelevant predictions
Surprise-free forecasts

Forecasting and planning for the future in quantitative terms is the activity of almost everyone, as it is a virtual necessity in private life as well as in the management of organizations and activities of much larger dimensions. However, there are two very great differences. The first is of scale. Error in personal forecasting may occasionally be disastrous, but error in large-scale planning for the future is almost certainly disastrous. Military defence, jobs requirements, and energy and food production planning, are some of the areas where error has serious ramifications. The second difference between personal and large scale planning is of kind and method. Simply stated, the enormity of the variables and growth which is proportional to the growth of the size of the plan, both as it extends in time and in elements contained, requires increasing levels of methodological sophistication. One popular technique is to develop alternate scenarios, or at least those that can be typified as optimistic and pessimistic so that a mid-range of expectations can be established. These scenarios can be identified in part by statistical evaluations of probability. The lack of application of the statistical laws of probability to the possibilities of discontinuities in trends, to interferences in processes, and to the possible aggregate error in the total plan's content of calculations and assumptions leads to exaggeratedly optimistic or pessimistic forecasting.


In a 1993 analysis by UNCTAD of the recent record of forecasting, it was reported that most forecasts, including those of OECD, the IMF and UNCTAD itself, have been quite inaccurate. There was persistent overprediction of growth in 1992 and 1993 and successive downward revisions had to be made since the outset of the current recession. The analysis suggested that it could be argued that these forecasting failures contributed to governments being insensitive to the risks of a major recession as the credit bubble of the 1980s burst.

Model builders failed to understand the changes in relationships between wealth and spending, as well as in the way monetary policy affects economic activity, a result of deregulation and integration of financial markets in the 1980s. They also misread the implications of the resulting debt accumulation. When interest rates started to rise sharply in 1989, pressure on firms and households to cut back their debt commitments became intense. This, together with the resulting fall in asset prices, led to a process of debt deflation which both deepened and prolonged the recession. However there was initially very little quantitative evidence of the magnitude of these effects. Forecasters had to make an unusually important act of judgement, by modifying the model to take account of the new behaviour. Such a judgement requires a high degree of experienced intuition which was clearly inadequate to the challenge.

Broader Problems:
Inconclusiveness of science
Related Problems:
Problem Type:
F: Fuzzy exceptional problems
Date of last update
04.10.2020 – 22:48 CEST