Since the 1920s, members of the Vienna Circle have developed a philosophical program that can be best described as “logical positivism”. The primary goal of the program was to separate truly science from pseudo-science by demarcating scientific knowledge. Logical positivists therefore only accepted two forms of propositions as scientific knowledge, namely, analytical and synthetic a prior, any proposition that is not true by definition is supposed to pass the test of empirical research. In addition, empiricism has two key aspects from their point of view. First, all evidence relates to synthetic propositions are based on sense perception, second, predicates must be empirically verifiable. However, the problem of induction brought difficulties to the program. According to Hume, “observations only allowed the acceptance of singular statements about the properties of particular things, times and places”. The quote indicates the possibility that an unobserved thing is not consistent with a past generalization always exists, thus there is no certainty on the process of induction. As a result, none of universals or laws has the opportunity to be inferred from sense data then science would only contain a provision of descriptions without laws.
Econometrics was created in the 1930s with the aim of developing empirical methods, and new challenges arose almost the same time. While laws are significantly important for either explaining theories or predicting events in the logical positivism program, the inability to do experiments in the case of economics makes laws have to be found outside the setting of the laboratory. It brings lots of problems with regard to controlling the environment, moreover, the causal relationships between variables have to be studied independently.
“The probability approach in Econometrics” which is written by Haavelmo (1994) discussed how to tackle the above problems. Since it is impossible for phenomena of real economic be separated from “other influences” artificially, the problem was explained in the light of “judging the degree of persistence over time of relations between economic variables”. When we are not able to conduct a control experiment and the behavior of the variable can be only observed passively, the distinction between potential and factual influences becomes the foundation to assess “the degree of persistence of a relationship over time”. In a passive observation setting, only factual influence can be measured as factors with potential influence cannot be observed. Haavelmo stated that factors that contain relatively important factual influences might be limited to a small degree by nature, so the actual observed value of y would be explained by only a certain amount of factors with factual influences.
Over decades there are many debates in economics and no agreement has achieved for now. For example, to solve the problem of incompleteness, Cowles’s Commission proposed to develop models that as complete as possible a description of reality, but Friedman was an opponent and disagreed with any realistic models. A positive economics science should one day get rid of subjectivism and be able to construct a systematic model for an economy as a whole.