Endogeneity bias — fiction in a fictitious world (wonkish)

19 Feb, 2020 at 18:12 | Posted in Statistics & Econometrics | 10 Comments

endoThe bivariate model base and its a priori closure destines ‘endogeneity bias’ to a fictitious existence. That existence, in turn, confines applied research in a fictitious world. The concept loses its grip in empirical studies whose findings rely heavily on forecasting accuracy, e.g. a wide range of macro-modelling research as mentioned before. It remains thriving in areas where empirical results are evaluated virtually solely by the statistical significances of estimates of one or two predestined structural parameters in models offering highly partial causal explanations of the data at hand. These models are usually presented to serve the practical purpose of policy evaluation. Since conclusive empirical evidence is hard to come by for policies implemented in uncontrolled environments, making a good story becomes the essential goal … Use of consistent estimators actually enhances the persuasive power of the story by helping maintain the unfalsifiable status of the models …

From a discipline perspective, although belief in endogeneity bias has worked in favour of research topics where empirical findings are relatively hard to falsify, knowledge gain from data there is often dismally low, especially in studies working with large data samples … Econometric practice that disregards data knowledge in model design and camouflages deficiencies in model design by estimators which effectively modify key causal variables in non-causal ways against what was originally intended in theory, can only be called ‘alchemy’, not ‘science’.

Duo Qin / Journal of Economic Methodology

A great article that really underscores that econometrics is basically a deductive method. Given the assumptions (such as manipulability, transitivity, Reichenbach probability principles, separability, additivity, linearity etc) it delivers deductive inferences. The problem, of course, is that we will never completely know when the assumptions are right. Real target systems are seldom epistemically isomorphic to axiomatic-deductive models/systems, and even if they were, we still have to argue for the external validity of the conclusions reached from within these epistemically convenient models/systems. Causal evidence generated by statistical/econometric procedures may be valid in ‘closed’ models, but what we usually are interested in, is causal evidence in the real target system we happen to live in.

Advocates of econometrics want to have deductively automated answers to fundamental causal questions. But to apply ‘thin’ methods we have to have ‘thick’ background knowledge of what’s going on in the real world, and not in idealized models. Conclusions can only be as certain as their premises – and that also applies to the quest for causality in econometrics.


  1. Is this the best way to persue “the quest for causality in econometrics”?

    • Let’s hope he didn’t hurt himself …

  2. “‘alchemy’, not ‘science’”
    Professor Perry Mehrling talks of the “alchemy of banking” which basically violates standard economic assumptions about money being a zero-sum game to expand the money supply. He diagrams the alchemy in balance sheets and shows how they can remain expanded indefinitely, thus confounding the mainstream economist who thinks money has to be destroyed on loan repayments.
    Banking is like alchemy to conventional economists. So they simply ignore banking as Mehrling describes it, preferring their own brand of academic alchemy to justify their revered models.

    • Cough … Credit is superseded by contracts, predates physical currency, quantity and quality of contracts is a separate conversation.

      Your insinuation is that contracts are like alchemy ….

      • See Mehrling’s recent paper: https://www.ineteconomics.org/uploads/papers/WP_113-Mehrling-Payment-vs-Funding.pdf
        He describes three scenarios for what happens to new purchasing power created by a mortgage loan origination. The last scenario involves the new purchasing power staying in circulation indefinitely.
        “somehow the spread of that new purchasing power through the economy has had the effect of increasing demand for money balances. The new money did not disappear, but the excess supply of money did, because the demand for money grew into it.”
        Earlier: “this simple operation increases my purchasing power without decreasing anyone else’s. It seems like alchemy, or anyway a violation of some deep conservation law.”
        The alchemy of banking violates Pareto-optimality as well as thermodynamics.

        • As I said contracts E.g. mortgage, underwriting determines quality – see Black.

          Never new Pareto or thermodynamics had anything to do with contracts, sounds like scientism, social psychology might be a better discipline to inform.

          Worst is the suggestion of modeling to such human actions, by such methods, thought you would understand the Lévy distribution observations pointed out by Mandelbrot.

        • Put it another way Robert … for some to focus myopically on money, yet in the same breath call for sanctity of contracts, that for those other than the astute, for anything, especially RE, post wages and productivity diverging, is an ideologues game.

          Its quite the tell that my last response was not responded too, regardless of your prostrations over a period of time, contra to the known.

          Amends Lars.

          • Contracts can be written so that new purchasing power emerges for both sides, without anyone losing purchasing power. The new purchasing power meets demand for money balances, not real goods. That is why inflation should be no constraint on public policy. But inflation might occur for irrational reasons, so we should implement explicit COLAs and sell inflation swaps and TIPS so everyone can easily adjust to nominal inflation in the unlikely event that it returns due to irrational psychology.
            Sorry if I didn’t see your response sooner. I was out sleeping in my car, dealing with a flat tire and dead battery …

            • Your special emotional plea to capital is noted – immutable tokens of labour that runs afoul of Lars post on the corn economy perspective.

              You have still not responded to the mathematical realities offered above which refute you previous argument.

              • I’m lost. I was triggered by the negative use of the word “alchemy”, contrasted with “science”, in the Duo Qin quotation above. I merely wanted to point out that bankers routinely use alchemy to create money that debits no one except possibly, eventually, the Fed, at no cost to taxpayers.
                Economists ignore this alchemy.
                We should use the alchemy to free ourselves from relying on private capital.
                Money is not simply “immutable tokens of labor”. Lars gets paid more than me but what is he actually producing? Attention? Economists are well-paid in tokens but don’t grow corn themselves. The number of tokens is far in excess than the labor required to produce actual real consumable goods. The tokens have a value separate from labor.

Sorry, the comment form is closed at this time.

Blog at WordPress.com.
Entries and comments feeds.