Tag Archives: Macroeconomics

Economic science is in dire straits

This article brings out quite well and rather unintentionally a lot of what’s wrong with Economic science in India (and frankly, the entire world) today. Particular points in the article are especially revealing. Take this segment, for instance, where the author gives you a mental image of the typical macroeconomist.

…. he’ll tell you that he actually studies the impact of soft coal prices in the Ruhr on the velocity of the money supply in West Germany in the 1970s and something else you will never learn because you suddenly hear your phone ring.

So what IS the macroeconomist doing? He is building a model. A model is a mathematical framework that gives us the values of certain output variables for different combinations of input variables. The model does this by assuming certain mathematical relationships between the input and output quantities.

Why is the macroeconomist trying to build a model? The primary purpose of modelling is to be able to predict outcomes of actions and choices. To do this, a model first tries to explain observed correlations between the same input and output variables in one or more historical circumstances. The more the number of such instances in which the model makes accurate predictions and the lower the error, the more robust the model is said to be.

Where would such models be used? One key use of such models is in policy formulation and evaluation. The term policy in particular refers to the particular approach that government would take to the broader economy or a specific segment thereof, which in turn translates to the manner in which government chooses to intervene in the economy or a particular segment of it. Thus, we see that models are primarily used in shaping and evaluating government interventions in the economy.

Another key use is to help businesses in making better economic forecasts and therefore in making better business decisions.

What is the problem with this approach?

The problem is very fundamental. Model building is based on assumptions that are absolutely unsuited to the study of Economics. Every model assumes certain constant quantitative causal relationships between the input and output variables. As a social science that may be defined as the study of exchange, sound economic theory cannot assume any such constant quantitative causal relationships between variables.

Unlike in the natural sciences where twice the volume of water flowing into a leak-free pipe will mean twice the water flowing out as well, in Economics, the subject of the study is man. The singular trait of man is his unpredictability. No two human beings may be assumed to behave the same way in the same circumstances. The same human being may not behave the same way in the same circumstances at different points in time.

What this translates to is this – In the natural sciences, if a stone thrown up today falls back to the ground, a stone thrown up in identical fashion tomorrow will fall back to the ground as well. In economics, however, such automatic repetitive and repeatable behaviour cannot be expected because we are talking of people, not stones.

What models end up doing, therefore, is that they make unreal assumptions about human behaviour in order to make their predictions fit observations at some point in time. The problem in doing so is that past robustness of a model in explaining is no indication of its correctness or its future accuracy. People change, and with that change economic outcomes. Models cannot predict the manner in which people can change.

Modelling in economics is little different from voodoo. Personally, I like to call it the virgins in the volcano method of economics. In simple terms,

The model says that there is a strong correlation between the number of virgins that fell down the volcano and the amount of rain that fell every year. So in order to improve rainfall this year, let’s throw a couple more virgins down that volcano right away.

What we learn is that the very attempt at creating economic models is fundamentally and deeply flawed. What do we say about economics education when modelling is the core of the development of economic science? May I say “Heaven help us”?