“Economic Methodology: Understanding Economics as a science” – Marcel Boumans & John B. Davis

Economists can’t conduct controlled experiments like natural scientists because humans are messy and we have no fully controlled environments. As a result, economists often study things after they’ve happened. In the 1930s, the field of ‘econometrics’ popped up to develop empirical methods tailored to economists. As economist Jack Johnston put it, the purpose of econometrics is to put some empirical flesh and blood on theoretical structures. Thus, an econometrics model normally has an explicit functional form (in other words, it’s written as a mathematical equation), it uses relevant data, and bridges the gap between theory and data using statistical methods. That way, a theory can be backed up and tested empirically.

The implicit assumption here is that all relevant factors should be included in the model and data. This led to the Keynes–Tinbergen debate in the late 30s, where Keynes criticised Tinbergen. Both economists wanted to explain the business cycle, which involves periods of economic growth and contraction. Tinbergen used statistical and empirical methods, while Keynes used theoretical and policy-driven methods. Keynes argued that relying heavily on formal modelling was problematic. Firstly, if some significant factors cannot be measured, how will you account for them? Secondly, even if your current model is correct, will these relations hold in the future?

The debate between theory, causal factors, and prediction continued; models include assumptions about human behaviour, which can be unrealistic but make the mathematics tractable. Economist Milton Friedman argued in the 50s that a model’s success lies in its ability to predict the future, not in the realism of its assumptions. If a model assumes humans are completely rational, it doesn’t matter if this is unrealistic as long as it predicts human behaviour correctly. In the model, humans act ‘as-if’ they were fully rational. Friedman likens this to a pool player who may not know mathematics or engineering but plays intuitively, taking into account angles, velocity, and force. These ‘as-if’ explanations pop up a lot when we talk about experiments, as we’ll see later.