Traditional forecasting uses to be done by the "what if" method. One makes some assumptions and then calculates what will happen if the assumptions hold true.But what if the assumptions should be changed? Then there will be no forecast. So many scientists are sure that economic forecasting is principally impossible.

Those who believe are inspired by brilliant examples of forecasting in astronomy, physics, chemistry and other natural sciences. Economics so far does not fall into the category of natural sciences and this is not quite natural because economy systems are also liable to the laws of Nature.

In order to understand why forecasting in economy is viewed upon so pessimistically one has to ask why in other systems it proves so effective? And the answer is ready at hand. In natural sciences forecast is the logical consequence of the fact that basic properties underlying the behaviour of a system stay invariant in time. In economy the behaviour is determined by the resulting force of all human wills and forecast would be possible if the manifold of all wills has invariant properties. We list these properties as they are understood in cybernetics.

  1. The resulting force of all human wills pushes an economic system towards the state of minimum uncertainty.
  2. Minimizing the uncertainty of a system is possible if sufficient information is available.
  3. Minimizing the uncertainty of a system is feasible if the system is capable of receiving the necessary information.
  4. If minimizing the uncertainty of a system is impossible minimizing the uncertainty of a subsystem may be possible.
  5. Lack of information produces uncertainty.
Thus the invariant property of an economic system which consists in minimizing the uncertainty of the system can be used for economic forecasting. The corresponding procedure is described in Non-monetary analysis of economic way of Russia zipped file (WINWORD 90 Kb) and shortly outlined below.

Two levels of system definition are chosen: macro-level and meta-level. Macro-model of order 20 containing an input-output matrix, output vector, capacity vector, end product vector, a vector of displacements expressing the resulting force of social wills and an investment matrix describing content of materials in capacities. Macro-model is used for quantitative analysis in terms of highly generalized products. Meta-model classifies roles of people into generalized categories and is used together with the macro-model to evaluate the distribution of energy over forms of economic activities. The entropy of the distribution is the measure of uncertainty that is to be minimized through a proper choice of the vector of displacements.

Calculations can not be made automatically because numerous fuzzy parameters have to be input through expert judging. What is described in the linkable Russian texts is the ideology and methods of using a dialogue system exemplified by Russian history.

Go back