On the misuse of probability theory in economics

Absztrakt

The aim of this paper is to give an outline of the sources, arguments and consequences of probabilistic approach to all types of economic data and phenomena. Probability theory is an efficient and useful tool of inductive research in those areas where the conditions of its applications prevail but it leads to illusory results where these conditions are invalid. The first part of the paper deals with the general objective conditions of the applicability of probability theory.
Probability theory can be applied to analyse disorganized, non-learning systems with very large numbers of particles if there are no aims and goals of elements and the behaviour of the elements is stochastic and unchanging in time or the change in time is negligible or stochastic, that is, the behaviour of elements of the system can be described by objective probabilistic terms. The second application of probability theory is the sampling theory. There is an important difference between the two applications. In the first case the examined process is itself random. In the second case randomness and independence is not a necessary characteristic of the population from which the sample is derived. Randomness and independence can be introduced by the sampling procedure, therefore in these cases statistics and probability theory can be used for descriptive historical
reasons and not as a tool of inductive theoretical research.
The second part of the paper deals with the history of that convention in economics (and first of all in econometrics), which treats population data as a result of a stochastic process or as one actualisation of a repeatable random
sample. The very justification of econometrics from Haavelmo’s influential paper is the dichotomy between deterministic and stochastic phenomena.
However, this dichotomy is invalid in social sciences because there is a third type of phenomenon, the uncertain phenomena, which is typical in economics. The uncertainty stems from the inherent characteristics of the research subject and it is not a deficiency that could be overcome by the development of scientific methods. In this case objective numerical probability of the events cannot be counted or observed. The last part of the paper gives theoretical and practical examples of the negative consequences of the misuse of probability theory in economics.