Lies that economics is built on

18 Oct, 2014 at 10:38 | Posted in Statistics & Econometrics | 2 Comments

Peter Dorman is one of those rare economists that it is always a pleasure to read. Here his critical eye is focused on economists’ infatuation with homogeneity and averages:

You may feel a gnawing discomfort with the way economists use statistical techniques. Ostensibly they focus on the difference between people, countries or whatever the units of observation happen to be, but they nevertheless seem to treat the population of cases as interchangeable—as homogenous on some fundamental level. As if people were replicants.

You are right, and this brief talk is about why and how you’re right, and what this implies for the questions people bring to statistical analysis and the methods they use.

Our point of departure will be a simple multiple regression model of the form

y = β0 + β1 x1 + β2 x2 + …. + ε

where y is an outcome variable, x1 is an explanatory variable of interest, the other x’s are control variables, the β’s are coefficients on these variables (or a constant term, in the case of β0), and ε is a vector of residuals. We could apply the same analysis to more complex functional forms, and we would see the same things, so let’s stay simple.

notes7-2What question does this model answer? It tells us the average effect that variations in x1 have on the outcome y, controlling for the effects of other explanatory variables. Repeat: it’s the average effect of x1 on y.

This model is applied to a sample of observations. What is assumed to be the same for these observations? (1) The outcome variable y is meaningful for all of them. (2) The list of potential explanatory factors, the x’s, is the same for all. (3) The effects these factors have on the outcome, the β’s, are the same for all. (4) The proper functional form that best explains the outcome is the same for all. In these four respects all units of observation are regarded as essentially the same.

Now what is permitted to differ across these observations? Simply the values of the x’s and therefore the values of y and ε. That’s it.

Thus measures of the difference between individual people or other objects of study are purchased at the cost of immense assumptions of sameness. It is these assumptions that both reflect and justify the search for average effects …

In the end, statistical analysis is about imposing a common structure on observations in order to understand differentiation. Any structure requires assuming some kinds of sameness, but some approaches make much more sweeping assumptions than others. An unfortunate symbiosis has arisen in economics between statistical methods that excessively rule out diversity and statistical questions that center on average (non-diverse) effects. This is damaging in many contexts, including hypothesis testing, program evaluation, forecasting—you name it …

The first step toward recovery is admitting you have a problem. Every statistical analyst should come clean about what assumptions of homogeneity are being made, in light of their plausibility and the opportunities that exist for relaxing them.

Firmly stuck in an empiricist tradition, econometrics is only concerned with the measurable aspects of reality. But there is always the possibility that there are other variables – of vital importance and although perhaps unobservable and non-additive, not necessarily epistemologically inaccessible – that were not considered for the model.

Real world social systems are not governed by stable causal mechanisms or capacities. If economic regularities obtain they — as a rule — do it only because we engineered them for that purpose. Outside man-made “nomological machines” they are rare, or even non-existant. Unfortunately that also makes them rather useless.

Remember that a model is not the truth. It is a lie to help you get your point across. And in the case of modeling economic risk, your model is a lie about others, who are probably lying themselves. And what’s worse than a simple lie? A complicated lie.

Sam L. Savage The Flaw of Averages

2 Comments

  1. Balderdash or discourse?
    Comment on ‘Lies that economics is built on’
    .

    Scientific communication is guided by the code true/false, moral communication by good/bad-evil, and social communication by like/dislike. Economists have traditionally some problems to keep these spheres apart. And this is the main reason why economics is still at the stage of a proto-science.

    The title ‘Lies that economics is built on’ signals that the author commits a category mistake. The body of the text deals with the application of a statistical tool. Now, it is well known that tools may be inapplicable. This is often the case when they are transferred from one domain to another, say, from physics to economics. No doubt, economists, more often than not, apply statistical tools incorrectly. This, though, is a question of professional competence but not a moral problem. Who hammers his thumb is a botcher not a villain.

    Moreover, what economics is actually built upon are behavioral axioms (Debreu, 1959; Arrow und Hahn, 1991; McKenzie, 2008). These foundational propositions may be false but they cannot be characterized as lies.

    Lies belong to social and political sphere. In scientific matters we take it for granted that the dialog partner thinks and acts within the framework of true/false, well knowing that we cannot be perfectly sure.

    The argument that the other economist tells a lie is entirely misplaced in a discourse because it diverts the attention away from the question at hand. Researcher and detective are different occupations.

    “Scientific economics inquires only into the How and Why, not into the Good or Bad, of what is. From the scientific point of view preoccupation with Good and Bad is worse than useless since it not only fails to illumine anything but keeps the lightbeam of inquiry from being turned in directions where answers to significant questions can be found.” (Murad, 1953, p. 2)

    Or, as Schumpeter put it:

    “Remember: occasionally, it may be an interesting question to ask why a man says what he says; but whatever the answer, it does not tell us anything about whether what he says is true or false.” (1994, p. 11)

    Standard economics is built on false premises. That is a perfectly acceptable statement — both in form and content. Let politicians excel in vituperation, science is content with refutation.
    .
    Egmont Kakarot-Handtke

    .
    References
    Arrow, K. J., and Hahn, F. H. (1991). General Competitive Analysis. Amsterdam,
    New York, NY, etc.: North-Holland.

    Debreu, G. (1959). Theory of Value. An Axiomatic Analysis of Economic Equilibrium. New Haven, London: Yale University Press.

    McKenzie, L. W. (2008). General Equilibrium. In S. N. Durlauf, and L. E. Blume
    (Eds.), The New Palgrave Dictionary of Economics Online, pages 1–18. Palgrave
    Macmillan, 2nd edition. URL http://www.dictionaryofeconomics.com/article?id=
    pde2008_G000023.

    Murad, A. (1953). Questions for Profit Theory. American Journal of Economics
    and Sociology, 13(1): 1–14. URL http://www.jstor.org/stable/3484955.

    Schumpeter, J. A. (1994). History of Economic Analysis. New York, NY: Oxford
    University Press.
    .
    For the correct formal foundations of theoretical economics see the web page
    http://www.axec.org/#!axioms/cy2g

  2. Where is Sam L. Savage at home?
    Addendum to the comment on ‘Lies that economics is built on’
    .

    “Remember that a model is not the truth. It is a lie to help you get your point across. And in the case of modeling economic risk, your model is a lie about others, who are probably lying themselves.” (Sam L. Savage, source-link see above)

    A model is a mental construct that touches reality at crucial points. It has correctly been pointed out that a model is like a map and that, trivially, the map is not the landscape.

    While there is consensus that the map is different from the landscape, the question is to what degree.

    “Can any model be true? I do not think so. Any model, whether in physics or in the social sciences, must be be an over-simplification. … I think we have to admit that most successful scientific theories are lucky over-simplifications.” (Popper, 1994, pp. 172-173)

    Physicists can be proud of many lucky over-simplifications that come reasonably close to the ideal of truth.

    In economics things are different. Here we have first to discriminate between political economics and theoretical economics. In political economics it may happen quite often that a model “is a lie to help you get your point across.” Theoretical economics, on the other hand, complies with scientific standards.

    To recall, it were the ancient Greeks who first introduced the distinction between doxa and episteme, opinion and knowledge. And then they drew the line of demarcation between non-science and science. Any doubts where they would have situated political economics? And where they would have situated Sam L. Savage’s rhetorical shell game?
    .
    Egmont Kakarot-Handtke
    .
    References
    Popper, K. R. (1994). The Myth of the Framework. In Defence of Science and
    Rationality. London, New York, NY: Routledge.


Sorry, the comment form is closed at this time.

Blog at WordPress.com.
Entries and comments feeds.