Economists and the ‘As If’ Argument

Many economists will admit that their models are not, and do not resemble, the real world. Nevertheless, when pushed on this obvious problem, they will assert that reality behaves as if their theories are true. I’m not sure where this puts their theories in terms of falsifiability, but there you have it. The problem I want to highlight here is that, in many ways, the conditions in which economic assumptions are fulfilled are not interesting at all and therefore unworthy of study.

To illustrate this, consider Milton Friedman’s famous exposition of the as if argument. He used the analogy of a snooker player who does not know the geometry of the shots they make, but behaves in close approximation to how they would if they did make the appropriate calculations. We could therefore model the snooker player’s game by using such equations, even though this wouldn’t strictly describe the mechanics of the game.

There is an obvious problem with Friedman’s snooker player analogy: the only reason a snooker game is interesting (in the loosest sense of the word, to be sure) is that players play imperfectly. Were snooker players to calculate everything perfectly, there would be no game; the person who went first would pot every ball and win. Hence, the imperfections are what makes the game interesting, and we must examine the actual processes the player uses to make decisions if we want a realistic model of their play. Something similar could be said for social sciences. The only time someone’s – or society’s – behaviour is really interesting is when it is degenerative,  self destructive, irrational. If everyone followed utility functions and maximised their happiness making perfectly fungible trade offs between options on which they had all available information, there would be no economic problem to speak of. The ‘deviations’ are in many ways what makes the study of economics worthwhile.

I am not the first person to recognise the flaw in Friedman’s snooker player analogy. Paul Krugman makes a similar argument in his book Peddling Prosperity. He argues that tiny deviations from rationality – say, a family not bothering to maximise their expenditure after a small tax cut because it’s not worth the time and effort – can lead to massive deviations from an economic theory. The aforementioned example completely invalidates Ricardian Equivalence. Similarly, within standard economic theory, downward wage stickiness opens up a role for monetary and fiscal policy where before there was none.

If such small ‘deviations’ from the ‘ideal’ create such significant effects, what is to be said of other, more significant ‘deviations’? Ones such as how the banking system works; how firms price; behavioural quirks; the fact that marginal products cannot be well-defined; the fact that capital can move across borders, etc etc. These completely undermine the theories upon which economists base their proclamations against the minimum wage, or for NGDP targeting, or for free trade. (Fun homework: match up the policy prescriptions mentioned with the relevant faulty assumptions).

I’ll grant that a lot of contemporary economics involves investigating areas where an assumption – rationality, perfect information, homogeneous agents – is violated. But usually this is only done one at a time, preserving the other assumptions. However, if almost every assumption is always violated, and if each violation has surprisingly large consequences, then practically any theory which retains any of the faulty assumptions will be wildly off track. Consequently, I would suggest that rather than modelling one ‘friction’ at a time, the ‘ideal’ should be dropped completely. Theories could be built from basic empirical observations instead of false assumptions.

I’m actually not entirely happy with this argument, because it implies that the economy would behave ‘well’ if everyone behaved according to economist’s ideals. All too often this can mean economists end up disparaging real people for not conforming to their theories, as Giles Saint-Paul did in his defence of economics post-crisis. The fact is that even if the world did behave according to the (impossible) neoclassical ‘ideal’, there would still be problems, such as business cycles, due to emergent properties of individually optimal behaviour. In any case, economists should be wary of the as if argument even without accepting my crazy heterodox position.

The fact is that reality doesn’t behave ‘as if’ it is economic theory. Reality behaves how reality behaves, and science is supposed to be geared toward modelling this as closely as possible. Insofar as we might rest on a counterfactual, it is only intended when we don’t know how the system actually works. Once we do know how the system works – and in economics, we do, as I outlined above – economists who resist altering their long-outdated heuristics risk avoiding important questions about the economy.


, , ,

  1. #1 by The Arthurian on April 3, 2013 - 10:01 pm


  2. #2 by Min on April 3, 2013 - 10:47 pm

    I think that “as if” arguments are fine, in their place. I like the example of a bird that dives for fish. Obviously, such birds would not long survive if they aimed at the apparent position of the fish, because of the refraction of light waves. In fact, they must be rather accurate in their aim. Therefore, if I wish to predict their diving behavior, I solve the physical problem using the refractive index. I may not know how the bird has learned to adjust its aim, but I know that I can determine its adjustment if I do the calculations. I do not believe that the bird goes through such calculations, but it acts AS IF it did. That is good enough for my predictions.

    If I am interested in how the bird learns to make adjustments, then I dispense with the as if argument. That’s a different matter.

    If we find, as you indicate with the Krugman example, that our as if argument does not enable us to make accurate predictions, then so much for the as if argument. There is no longer a warrant for it. Back to the drawing board.

    • #3 by Unlearningecon on April 4, 2013 - 10:45 am

      Am I incorrect in saying that those types of birds generally position themselves directly above their prey to avoid problems with refraction?

      I am hesitant to use the as if argument in economics for the same reasons you outline: prediction. Predictions in economics are rarely as precise as they need to be in order to justify an ‘as if’ argument, so we’re probably better off trying to find out what’s actually going on.

      • #4 by Min on April 4, 2013 - 7:34 pm

        IIUC, the birds typically dive at a small angle, for other reasons as well as refraction.

        As for predictions and projections in economics, my usual take is GIGO! GIGO! GIGO! 75 years into the future? Give me a break!

  3. #5 by W on April 3, 2013 - 11:17 pm

    You may state that “A bird acts AS IF it went through physical calculations”; in such a case, you would be acting in a way like that of the bird: it goes towards an aim, just as you go towards an aim.

    In the later case, yours, the aim seem to be the exchangeability of, say, facts (you may even ask…”but what are facts in the end”, or “but…are you talking of facts in the same sense that Hume did?”, and that would be great, indeed), and the objects (either material of psychological) intended to represent such facts (this is not very far away from magic, as far as i can see…).

    Then, from that moment onwards, your friends might start calling you “Birdie”.

    But finally the suspicion would remain about science to be rather ad-hoc. If so, that would be taken as a huge problem or not?

  4. #6 by avastmick on April 4, 2013 - 1:25 am

    I like this blog. It supports my thinking. I am reinforcing my beliefs about the practice of economics by reading it.

    While the above statements are true they do also mirror the problem with much of the thinking of modern economics. The poorest science is that that adheres to a certain “School of Thought”: it spends more energy on proving the applicability of the concepts of the school of thought generally than the application specifically. Hence the pejorative term religiosity; it is just ideology and dogma. It is not moving the world on

    “As if” cannot be a means to scientific understanding; and I assume that is the goal of economics, to be (or become) scientific. “As if” can only be a heuristic approach, a rule of thumb to encapsulate a complex system. Heuristics are only a slight step away from “convention”, then “tradition”, not far then from religion.

    Science has to be objective. “As if” is a subjective position. If a position is subjective (and heuristically-based) it cannot be falsifiable; it is built on faith.

  5. #7 by Protesilaos Stavrou on April 4, 2013 - 7:28 am

    Yes, you are right. One of the major “as if” hypotheses, especially in economic analysis, is the exclusion of political, cultural and historical factors and the consequent insulation of the “economic” sphere from anything else in the human world.

    For instance, we see direct comparisons between countries whose institutional morphologies, comprising both the social and the legal institutions, are profoundly different. The countries are fallaciously placed under the same denominator, usually by isolating the parameters that are being examined from the totality of factors that are contributing to them, permeating them, influencing them, or even engendering them.

    The underlying assumption in this case, is none other than the existence of that magnificent homo economicus, whose actions and methods are homogeneous throughout time and across places/cultures, allowing for the direct comparisons between different sets of such rational (and omniscient?) agents.

    In a nutshell, one of the primary “as if” assumptions is the one which blithely asserts that the *context* either does not exist or, if it does, it is irrelevant to the mechanics of the economy and the interpretation of the phenomena therein.

    • #8 by Unlearningecon on April 4, 2013 - 10:47 am

      This sounds a lot like my post institutions and economics. Have you read it?

      • #9 by Protesilaos Stavrou on April 4, 2013 - 11:28 am

        No I had not, as I must confess that I have subscribed to your blog only recently. Now that I have read it though, I find myself in agreement with your remarks.

        I would nonetheless wish to take the argument a step further, suggesting that this kind of approach is not limited to economics (certainly not Friedman), or indeed the social sciences, but rather traces its roots to philosophical doctrines.

        One major element of such a worldview is the essentially Aristotelean (and then Enlightenment) notion of the “hypokeimenon”, the indivisible subject, whose robust ontological reality persists in spite of whatever circumstantial conditions may have been bestowed on it.

        If we were to take this axiom at face value, we would, nolens volens, draw the inference that things can be conceptually removed from their milieu; we would, in other words, be led down the garden path of adopting a method of decontextualizing contextualization that would then permit us to place the “thing” in an imaginary framework where its “underlying” truths—those assumed by Aristotle et al—would be revealed to us.

        Of course, I’m not suggesting that your argument is anyhow incomplete. I am merely noting down this opinion of mine.

      • #10 by Unlearningecon on April 4, 2013 - 7:04 pm

        Yeah, fair enough. The one philosophy I am aware of that doesn’t fall into that trap is Marxism. Marx always thought his theories would only be true under capitalism, and once socialism or communism were achieved, things like class and the LTV would be meaningless.

      • #11 by noteconomist on April 4, 2013 - 12:38 pm


        Basically all academic fields have philosophical doctrines at their root.

        Hell, that’s why they call it a PhD.

      • #12 by Protesilaos Stavrou on April 5, 2013 - 7:26 am

        Yes, marxism, as diverse as it might be in its parts, does not fall into this particular slough of fallacies. It does however go to the other end of elevating factors such as class and history into determinist, objective constants. While this may run off to another tangent, a book worth reading, in case you haven’t done already, is The Imaginary Institution of Society by Cornelius Castoriades.

    • #13 by Luis Enrique on April 5, 2013 - 10:50 am

      ” in economic analysis, is the exclusion of political, cultural and historical factors”

      well you’ve pulled that claim out of your backside haven’t you. How many examples of economics incorporating political, cultural and historical would it take before you acknowledge that claim is false?

      • #14 by Luis Enrique on April 5, 2013 - 10:51 am

      • #15 by Unlearningecon on April 5, 2013 - 1:43 pm

        Maybe I should do a post on how economics and econometrics could be thought of as separate in many ways, and how the former could learn a thing or two from the latter.

  6. #16 by Boatwright on April 4, 2013 - 4:55 pm

    Economists with physics envy…………………

    I find the bird analogy interesting, but flawed. A diving bird or a ball player unconsciously calculating the necessary trajectory is in fact calculating in a remarkable (in the sense that the evolution of such skills is complex and interesting) way. However as a matter of mathematical physics the event is rather simple to model. We observe a limited number of variables — time, velocity, gravity, refraction, etc. — that the complex biological computer of the bird or human brain (with billions of nodes) has evolved to perceive and predict. Result: the bird swallows the fish, the ball meets the hand.

    To say this is equivalent to theories of economic optimization is simply wrong. In the case of the bird and the ball player we have a physical situation of sight, target and muscle. In the case of economic actors we have intractable complexity, with huge numbers of of variables including many such as emotion that are themselves hugely variable. Huge x huge = huge. Meaningful economic statements, therefore, will always be constrained to be probabilistic and heuristic.

    “As if” theorizing says nothing because it makes anything and everything “true”. The sooner economic theorists except the principle of falsifiability the better.

    • #17 by Unlearningecon on April 4, 2013 - 7:09 pm

      Yeah, I think this is true. A bird – or snooker player’s – mind is a lot closer to optimising the situation than a consumer facing 12320 products, or a firm looking to make a long term business is, given the complexity of the task and the inherent uncertainty.

    • #18 by Min on April 4, 2013 - 7:48 pm

      “To say this is equivalent to theories of economic optimization is simply wrong.”

      That was more or less my point. 🙂

  7. #19 by Luis Enrique on April 5, 2013 - 10:25 am

    The only time someone’s – or society’s – behaviour is really interesting is when it is degenerative, self destructive, irrational. If everyone followed utility functions and maximised their happiness making perfectly fungible trade offs between options on which they had all available information, there would be no economic problem to speak of.

    this does’t seem right to me at all. Firstly, if there are parts of the economy which function reasonably smoothly, or are well approximated or best understood that way, that’s still interesting and we still want to study and understand them. Secondly, we know that even rational actors with full information can produce all manner of economic problems – the whole panoply of market failures fits in that mould – and, further, the “as if” modelling approach of economics can readily incorporate information problems and departures from rationality. Many macro models use a learning algorithm or “adaptive expectations”, for example, in place or Rat Ex, which are still “as if” constructs. Krugman is no doubt correct that small deviations from rationality can have important consequences, but that might just mean the right approach is to model agents “as if” their behaviour involves small deviations from rationality. The interesting stuff is how these small deviations from rationality interact.

    you may be aware that snooker is also a canonical example when teaching chaos theory. The idea is that even very good physics models will involve some error, but in games like snooker small errors can send the game off down completely unpredictable directions – the idea being that a small millimetre error somewhere will mean one ball missing another ball that it would otherwise have collided with, sending the course of the game down a whole different path.

    A Friedmanite model of snooker could very easily incorporate errors. Standard econ models frequently do so. Game theorists look at “trembling hand” equilibria. However, whilst such a Friedmanite model of snooker, as if the players are probability-of-winning-maximising expert geometry-physics calculating machines that make some errors, might give very good short-run predictions of which shots skilled players will select and what those shots will look like, the long-run outrun of the game is much harder to predict.

    Now a lot of people are very attracted to thinking of the economy in those terms, and I agree in some sense that’s deep unpredictability is probably true. But it’s not obvious that’s the right way to think of all economic problems. For example, if you want to know what’s going to happen if the Tories cut government spending, rather that witter on about choas, it’s probably better to use an “as if” model in which the economy attains a new short-run equilibrium with higher unemployment.

    • #20 by Luis Enrique on April 5, 2013 - 10:26 am

      oh balls, html fail again. please do fix if so inclined.

    • #21 by Unlearningecon on April 5, 2013 - 1:40 pm

      Not really much to disagree with here. One thing I’ve cottoned onto is that I don’t think economists tend to define the appropriate ‘domains’ for their theories often enough: e.g. if x, y and z are true then this model will apply. They do sometimes, but overall it seems something like demand supply, comparative advantage and even DSGE models are not presented with a clear domain in mind. I know DSGE models generally explore a particular issue, and you might say this is a domain, but these ‘issues’ rarely occur in isolation the real world, and are often mixed up with other ‘issues’ which are explored in different models, which may be contradictory. How do we know which DSGE model is appropriate if they take interwoven ‘issues’ one or two at a time?

      You are correct that a more Friedmanite model may make decent short run predictions in the snooker example. However, an appropriate chaos model would make short run predictions that were at least as good, as well as having the added benefit of telling us about long run behaviour, sensitive though that may be to initial conditions. So as I imply, many of the economic models to which I object aren’t completely useless; they have useful truths that can nonetheless be capture din more comprehensive models that would supersede them.

  8. #22 by Boatwright on April 5, 2013 - 11:46 am

    All roads lead to epistemology.

    There is a deep pattern in social “science” to model theory after established physical science. It is no mystery that in the post Newtonian world J.S. Mill, Marx and others derived a simple, deterministic theoretical framework.

    We must indict mainstream economic theory for still being stuck in a three-hundred year old view of reality. Economic theory soldiers on like a twenty-first century Lord Rutherford as if relativity, quantum mechanics, chaos theory, ecological biology, etc. never happened.

    We see a visceral aversion to heuristic and probabilistic methodologies. There seems to be fundamental confusion about the heuristic method. Heuristic research is not “as if”, but “what if?”.
    The principal is to start a line of inquiry of a complex situation with an approximation – a best guess if you will – and see where it leads, testing results against real-world data, and adjusting the model accordingly. Heuristic research models are open ended, with no limit to the evolving complexity of a model other than how it measures against the real data.

    A good example of the heuristic approach is the rapid progress over the last few decades in meteorology and weather forecasting, where multiple competing forecasting models have been continually refined, with more and more real data points folded into the models, and increasing accuracy over longer and longer futures.

    This far different from arbitrary, fixed, and timeless assumptions as a starting point.

    • #23 by Luis Enrique on April 5, 2013 - 11:58 am

      “We see a visceral aversion to heuristic and probabilistic methodologies”

      I see a visceral aversion to acquiring some empirical knowledge about economics before sounding off on the subject.

      For just one example of economists using approaches you claim they do not, head to google scholar and look for “Bayesian model averaging economics”

      • #24 by Boatwright on April 5, 2013 - 12:25 pm

        Point taken. I should have qualified my criticism to neo-classical macro-economic theory. In fact some of the best research in contemporary economics is in detailed statistical modeling of micro-economic events, using the Bayesian methods you mention.

        My criticism is pointed at the sort of non-sense we see put forth by popular political economists about the money supply, public debt, employment, etc.. using simple linear math such as IS-LM.

      • #25 by Luis Enrique on April 5, 2013 - 12:41 pm


        ah, I see what you mean. Well I’d say there’s still a place for very simple models. And even if you construct something simple with a couple of lines, you could be quite sophisticated about how you estimate their position and slope.

        neoclassical macro does this stuff too though – again head to google scholar and look for “macroeconomics dsge bayesian model averaging”

  9. #26 by Apu Garnesh on April 10, 2013 - 11:27 pm

    Very interesting point you make here! However, I must disagree when you say that

    “The only time someone’s – or society’s – behaviour is really interesting is when it is degenerative, self destructive, irrational.”

    As counterexamples, the Prisoners’ Dilemma, the Tragedy of the Commons, Adam Smith’s Invisible Hand, the theory of comparative advantage, etc. All these are “really” interesting too but do not rely on “degenerative, self destructive, irrational” behaviour.

    • #27 by Unlearningecon on April 11, 2013 - 9:13 pm

      Yeah, perhaps that was overstated. ‘Often’ behaviour is interesting when this happens. The idea I had in mind was utility functions, which aren’t particularly useful.

  10. #28 by Joe Seydl (@JoeSeydl) on April 11, 2013 - 8:47 pm

    “All too often this can mean economists end up disparaging real people for not conforming to their theories”

    This is dead on. Performativity in economics is a huge deal. Economists like to think of themselves as objective observers, but they are actually prescribers, influencing how people make decisions by pushing this very narrow utility-maximization model. The ones that do recognize this influence are probably OK with the fact that RCT doesn’t predict very well — at least it’s changing people’s behavior, which is more important, in their view.

  11. #29 by Blue Aurora on April 18, 2013 - 2:02 pm

    Let’s try this again…

    Speaking of Professor Paul R. Krugman and the “as-if” argument…

    Unlearningecon, have you seen my question to Daniel P. Kuehn on his thoughts on Daniel Ellsberg’s contributions to economics? I asked him that question back in 17 November, 2011. If not, then please see the following link, and the comments that ensued.

    Nearly a year later, Kuehn did a post on his thoughts on the Ellsberg Paradox, after being prompted by a section of a YouTube video where Professor Paul R. Krugman and Professor Joseph E. Stiglitz were asked for their views on the role of Subjective Expected Utility decision theory in academic economics. Please see the following post and the comments that followed.

    Kuehn then made another post in response to a term (ambiguity aversion) that I and another commentator (Ivansml) made. Please read the following post and its comments section.

    Kuehn last shared his thoughts on the Ellsberg Paradox briefly in the following post, before going onto another subject.

    I don’t mean to be rude or snarky to Daniel P. Kuehn, but did you find that my experience with Daniel Kuehn’s thoughts has a parallel with the contents of this post, Unlearningecon? (And just for the record, I consider myself neither “orthodox” nor “heterodox”.)

    • #30 by Unlearningecon on April 19, 2013 - 2:32 pm

      Typing out maths really bothers me!

      But yes, it seems Kuehn was really stretching out the body of SEU there. In fact I’m still confused about precisely what he was saying after multiple readings. He seems to concede that the paradox is a paradox but it doesn’t really matter.

      The question for me would be this: is there a theory which captures both situations. Economists don’t seem to be bothered about this; only about whether their theory might have some application, somewhere.

      • #31 by Blue Aurora on April 21, 2013 - 4:29 pm

        Actually, regarding the development of a decision theory that is more general than SEU and explains deviations like the Ellsberg Paradox…

        There are some versions. Look up “Choquet integration” and anything by Itzhak Gilboa and David Schmeidler. There’s also research done by notable figures like Peter Wakker and Mark Machina. Of course, besides people with background in mathematics or economics, philosophers have also contributed to decision theory (like Isaac Levi), and so have people educated in psychology (the famed Daniel Kahneman and his late partner, Amos Tversky, are notable examples).

  1. Economists and the ‘As If’ Argument | Pressed Words from Terry Hancock
  2. Thursday Morning Links |
  3. A not-so-dismal “science” | This is Ashok.