Home > The Economics Profession > Econometrics – still lacking a valid ontological foundation

Econometrics – still lacking a valid ontological foundation

from Lars Syll

Important and far-reaching problems still beset regression analysis and econometrics – many of which basically are a result of an unsustainable ontological view.

Most econometricians have a nominalist-positivist view of science and models, according to which science can only deal with observable regularity patterns of a more or less lawlike kind. Only data matters and trying to (ontologically) go beyond observed data in search of underlying real factors and relations that generate the data is not admissable. All has to take place in the model of the econometric mind, since the real factors and relations according to the econometric (epistemologically based) methodology are beyond reach, since they, allegedly, are both unobservable and unmeasurable. This also means that instead of treating the model-based findings as interesting clues for digging deepeer into real structures and mechanisms, they are treated as the end points of the investigation.

As mathematical statistician David Freedman writes in Statistical Models and Causal Inference (2010):

In my view, regression models are not a particularly good way of doing empirical work in the social sciences today, because the technique depends on knowledge that we do not have. Investigators who use the technique are not paying adequate attention to the connection – if any – between the models and the phenomena they are studying. Their conclusions may be valid for the computer code they have created, but the claims are hard to transfer from that microcosm to the larger world …

freedman2Given the limits to present knowledge, I doubt that models can be rescued by technical fixes. Arguments about the theoretical merit of regression or the asymptotic behavior of specification tests for picking one version of a model over another seem like the arguments about how to build desalination plants with cold fusion and the energy source. The concept may be admirable, the technical details may be fascinating, but thirsty people should look elsewhere …

Causal inference from observational data presents may difficulties, especially when underlying mechanisms are poorly understood. There is a natural desire to substitute intellectual capital for labor, and an equally natural preference for system and rigor over methods that seem more haphazard. These are possible explanations for the current popularity of statistical models.

Indeed, far-reaching claims have been made for the superiority of a quantitative template that depends on modeling – by those who manage to ignore the far-reaching assumptions behind the models. However, the assumptions often turn out to be unsupported by the data. If so, the rigor of advanced quantitative methods is a matter of appearance rather than substance.

If econometrics is to progress it has to abandon its outdated nominalist-positivist view of science and the belief that science can only deal with observable regularity patterns of a more or less law-like kind. Scientific theories ought to do more than just describe event-regularities and patterns – they also have to analyze and describe the mechanisms, structures, and processes that give birth to these patterns and eventual regularities.

  1. August 1, 2013 at 11:30 pm

    Agree that economists “also have to analyze and describe the mechanisms, structures, and processes that give birth…”. May be that is they are only things econometricians can do successfully.

    The idea that it is possible (with current methods) to discover “objectively” simple empirical relationships about a complex economy has proven to be futile. It does not help when the economists do not understand their data, and less about the statistical packages they are using. Their bogus claims are accepted by even more clueless referees largely confirming their own particular biases.

    Econometrics should spend more/some effort testing theories rather than live in a parallel universe of objective empiricism, divorced from theory.

  2. August 2, 2013 at 8:05 am

    I fully agree with David Freedman in his saying, “In my view, regression models are not a particularly good way of doing empirical work in the social sciences today, because the technique depends on knowledge that we do not have. Investigators who use the technique are not paying adequate attention to the connection – if any – between the models and the phenomena they are studying.” They know nothing about what they are studying and they expect to find the philosopher stone from the data which they did not collect or understand

    Regression models are simply not powerful enough to discover anything you don’t already know. I use econometrics and regression models to verify (or falsify) what I already know. For example, from working in the industry, personally collecting data on costs of the pension fund industry and talking to management and data suppliers, I already know key factors which determine the costs of running pension funds.

    Regression models of measured costs against the identified determining factors of costs give R-squares of greater than 90 percent and p-values (probability of accidental result) of near zero (to more than 4 decimal places). The academic boneheads at an economic journal, never having seen statistical results of such high quality, simply waffled on about “not meeting sufficiently high academic standards”! They made me sick with their pretense.

    • Newtownian
      August 4, 2013 at 9:13 am

      Lyonwiss and Lars, I have to say thanks again for these comments and discussions on your struggles for economic enlightenment. Some observations and suggestions – which you may find of use, if not no harm done.

      I’m currently working in an analogous environmental biological/ecological context on a predominantly linear cause and effect system and seeing exactly the same issues you are commenting on albeit without perhaps the academic acrimony. Myself and colleagues has framed this is in terms of emerging ‘3rd generation’ ‘mechanistic’ models (a great handle for which I take no credit) with a view to the models eventually superseding more developed ‘2nd generation’ empirical models. The latters’ limitations seem very comparable to current economics models.

      Observations arising include:
      – Current 2nd gen empirical economics models may still be useful in distilling relationships which 3rd generation models need to predict approximately in the first instance.
      – Though our long term aim is utilitarian models which describe environmental/human impacts, the 3rd gen models ideally will be based on fundamental ecological understanding.
      – The mechanisms seem to include ones from sub-disciplines that most biologists don’t learn normally, e.g. diffusion biophysics based on the Stokes-Einstein equation. There is no malice in this not having been looked at early but rather modern biology is dominated by molecular biology leaving much less time for people to learn or go into depth or review basic processes.
      – Ironically there is evidence of useful research being done by pure physicists but they don’t seem to have a biological appreciation whereas I cant make head nor tail of their equations.

      Some recommendations for economics arising from what seems to me a science analogy:
      * Much better genuine cross disciplinary exchange and cooperation is needed. An obvious area would be more equal discussions between the physics quants apocryphally populating finance institutions and economists – exploit their imagination rather than treat them as black box calculators. Past interactions with Von Neumann and John Nash seem to provide a model but these shouldn’t be glorious exceptions but routine (RER seems open to this but I am still seeing mainly here traditionally trained economists in the comments).
      * Economists finding a way to ‘free their minds’ (at least to some degree) of older emergent concepts. (I cant help but feel money, value etc. are like earth air fire and water – useful but distracting from the real elements whatever these may be).
      * Recognize this will likely take a very long time and be a thankless task at first probably without financial rewards – to be undertaken for the love of it or by the independently wealthy – philosophically minded bankers – (?Jeremy Grantham wanna bes) rather than the academy.
      * Mainstream neoclassical economics may provide useful instruments (spreadsheets like the retorts of the old alchemists) but like the Vatican you will never deflect them – so don’t bother or expend your energy on them. They have too much vested interest and too much system resilience to challenge.
      * You need to return to understanding systems – possibly exploring candidate analogies from the hard science. Ecosystems are not the same as human economics but there are probably useful analogies there. In many ways this is what economics tried with equilibrium theory 150 years ago but it failed to move on further.
      * As you well know, deconstruct the history of neoclassical economics and get this story more out into the popular sphere to demystify economics for the masses and illustrate why its a house of cards. Steve Keen is one example – but there is a need for a greater communicator.

      • August 4, 2013 at 10:55 am

        Sorry, since you mentioned me. I must say I don’t know what you are waffling about, apart from dropping a few big names. What is your point in one or two sentences? No cliches, please.

  3. Newtownian
    August 6, 2013 at 1:47 am

    Dear Lyonwiss.

    I’ll try to make it clearer.

    I work in biology/ecology/environmental management. For some years I’ve been involved in modelling – quantitative risk assessment. The approach is powerful (e.g. quantitative estimates of risk and decision bases for remedial action). But underneath the models have been largely based on sophisticated regression analyses which unintentionally obscure our ignorance of the underlying more fundamental biophysical processes.

    This appears to be exactly the trap Lars describes current economics modelling as being in. Why my gratitude? I found your comments fascinating because they showed you were confronting similar challenges in modelling that I might learn from.

    Recently I’ve been looking at new models based not on regression of cause v. effect, but the underlying physics (models). As hoped the approach appears extremely promising and exciting and offers a way forward. And this lesson appears to be an old one in science e.g. Dyson F. A meeting with Enrico Fermi. Nature 2004; 427: 297-297.

    From my professional reading of (environmental) economics the problem I’m seeing is that unlike the hard sciences which have diverse fundamental frameworks which emerging disciplines like risk assessment can build on – economics at its heart is still thrashing about with the equivalents of ‘earth, air, fire and water’ – e.g. private property, debt, money, value etc. – which are useful concepts in society day to day – but philosophically are very unsatisfying and suggest a failure of imagination on the part of modern economists.

    The challenge now seems how to expand economists’ minds. I mentioned Von Neumann not for pretence but rather because he seems to exemplify an expansive polymath, of whom we have few these days.

    The discussions in RER suggest the process may be happening. The thrust of my suggestions was simply that economists might benefit from becoming a bit more multidisciplinary in that polymath tradition and learn a bit from the history of science.

    • August 6, 2013 at 5:16 am

      Your comments agree with those of Lars, David Freedman and mine. My summary of what has been going on in econometrics for decades is as follows. Econometricians have been analyzing “objective” data in the hope of discovering for economics, the equivalent of Kepler’s laws of planetary motions for astronomy. (Note that Kepler’s empirical finding came after centuries of data collection by many people.)

      But nothing like a “law” has been discovered and in my view, never will, for several fundamental reasons. Economics is fundamentally not universal. There are never enough data even for a given period of time or geography which are sufficiently stationary to discover anything statistically significant. Moreover, the data are often compromised by government manipulation, due to a lack of integrity standards imposed by the profession.

      Regression models and statistical methods are too weak to detect clear signals relative to the noise. It would be more realistic for econometrics to set its sights much lower: to verify or falsify some causal economic processes, for which there are some theoretical hypothesis, formed from real-life experience. Laws of economics are beyond reach, at this stage, I believe.

  4. August 6, 2013 at 12:09 pm

    After all the bagging of Reinhart and Rogoff (R & R) from the paper of Herndon et al. and all the
    negative comments from readers on this blog, as well as R & R’s admission of deficiency of their own work, one would have thought R & R’s critical 90 per cent public debt to GDP would have been consigned to the economic dustbin and oblivion. But no, here is a proof that the rubbish continues to circulate in the economic pot (like all other rubbish never discarded):


    Sylvester Eijffinger is a professor of financial economics at Tilburg University in the Netherlands,
    who could not have read R & R critically (no need for Harvard wisdom). He accepts R & R as reason to justify financial repression, for governments to work down the real value of their debts, hence negative real interest rates are expected to rob savers for another decade at least.

    This is more proof that applied econometrics and “empirical evidence” in economics are merely noise.

  5. BFWR
    August 6, 2013 at 3:26 pm

    The two problems with economics are 1) it is riven with unexamined orthodoxy and 2) it fails to understand that money is basically accounting, and that cost accounting is the most basic data from which one can derive economic relationships and patterns. What can be more basic than individuals exchanging their incomes for products and services, i.e. liquidating prices? If there is a fundamental scarcity of individual incomes in ratio to prices in the act of production itself (which is what cost accounting reveals) that would mean that the productive process (economic system itself) is radically and fundamentally price inflationary. If the system itself is fundamentally price inflationary it requires that there be an equating of total individual incomes with total prices in a given period of time….that does not first re-initiate the price inflationary nature of the system. The best and only effective way to do this is to distribute (not re-distribute) a supplemental income to individuals first…before it is injected into the economy as for instance in the form of loans to businesses.

    This would have numerous advantages and positive effects:

    1) It would actually resolve the most basic economic problem of the system.
    2) It (along with a mathematically derived formula for a general discount on prices) would make for a more robust economic security for the individual.
    3) It would break up the monopoly on credit creation that the Financial system currently has, as well as the very restrictive paradigms of loan ONLY and loans ONLY for production.

    Orthodox objections to this resolution to the productive system’s most basic problem as well as a lot of other abstract theoretical falderal about problems, most of which would disappear and/or dissipate with the resolution of this most basic problem, would best be replaced by asking oneself an equally basic question:

    Are systems made for Man, or Man for systems?

    First fix the systems actual problems. Then take the attitude that one ought to adapt the system to the physical and psychological needs of individual, not the individual to a broken/problematic system.

  1. No trackbacks yet.

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Google+ photo

You are commenting using your Google+ account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )


Connecting to %s

This site uses Akismet to reduce spam. Learn how your comment data is processed.