Are Critics of Economics Inconsistent?
A little while ago on Twitter, Ben Harrell expressed exasperation at economics critics for a reason I’ve seen a few times: the critic first argues economists don’t include some important feature of the world — say, bounded rationality (which we were discussing) — but when faced with examples which purport to include this feature within the mainstream framework — say, behavioural economics — they switch to methodological critique of how this important feature is modelled, instead of saying it isn’t modelled at all. As Harrell put it:
the initial criticism is rarely about method, but about unrealistic features of the method. If I make the method more realistic, THEN it becomes a critique about method…If your critique is about realism, don't retreat into an orthogonal critique about methodology when someone accommodates your original critique.
On a rhetorical level Harrell has a point, also made by Mathijs Krul a long time ago: it’s usually a losing tactic to insist that economists are completely unaware of common critiques of the standard model and are swimming in irrelevant and outdated ideas. As Olivier Blanchard reminded us, a quick look at the list of pertinent topics economists research is very long these days (you’d still be hard-pressed to find anything seriously wrestling with environmental collapse, corporate power and the slide into fascism, all of which are actually within the purview of the discipline, but I digress).
On the other hand, I don’t think argument this is necessarily as inconsistent as Harrell thinks. Put simply, the critic may not have in mind the same thing the economist does. So while the economist and the critic use the same words, what they refer to may be different. Thus, when the critic says “you don’t model that the right way” what they really mean is “you aren’t modelling what I said; you’re modelling something else”.
Bounded rationality is a paradigmatic example. Early criticisms of rationality from people like Herbert Simon were broad ranging criticisms of the idea that people could even begin to behave in the way depicted by the standard utility-maximising model given that they had neither the information, the capacity to process that information, nor the ability to act upon it that would be required. Thus they would use rules of thumb, ‘satisfice’ instead of ‘optimise’, and and so forth.
Mainstream behavioural economics bills itself as a response to Simon-type critiques and often calls its models ‘bounded rationality’, but they retain the utility-maximising framework and simply tweak it to add some ‘bias’ such as limited attention. As Berg and Gigerezner have nicely pointed out, the end result is that behavioural utility functions are more complex to solve than standard utility functions, making it less believable that real world agents would behave in line with their predictions. Paradoxically, this results in behavioural economists adhering even more strongly to the as-if doctrine of ‘it doesn’t matter that models are not accurate descriptions of real life’ despite behavioural economics being an overt attempt to make the models more realistic. (I except Richard Thaler from this because he has always been more reluctant to put things into an optimising framework).
Or, in Simpsons form:
If we instead take the bounded rationality criticism to heart we might want to abandon the assumption of optimising altogether. John Maynard Keynes famously applied this kind of logic to financial markets: because they are completely due to radical uncertainty, people will tend to herd, extrapolate from future trends, or just outright guess. As I wrote some time ago, seriously accounting for peoples’ inability to be ‘rational’ in a complex world leads us to a completely different framework to the standard model; simply trying to incorporate it into the mainstream leads us to miss the point entirely. (Check out evolutionary theorist Jason Collins if you want a more detailed account of how the ‘bias’ approach leads us astray).
Let me reiterate this point in painful philosophy-paper-style detail to make it clear. When stripped of its confusion, the debate takes the following form:
Critic: you don’t include X in your models
Economist yes I do, here is a model which includes X’
Critic: that’s not what I meant, you’ve modelled X as X’
Economist: why are you switching from a substantive to a methodological critique?
‘Imperfect information’ in mainstream economics is another example: in a sense individuals have more information than under perfect information, since they know all the probabilities and possible states of the world instead of just one of the latter. And as I have previously argued, this point applies even to the name of macroeconomic models — which claim to be Dynamic, Stochastic, General and Equilibrium, but are not really the first 3 in any meaningful sense. Critics would like to see their issue at the core of theories instead of bolted on at the end.
Why Oh Why Am I Discussing Macroeconomics Again?
Let’s take one more detailed example, again from the lush wellspring of opinion that is Twitter. A while ago Prof. Steve Hanke tweeted “No wonder economists are lowly regarded. Before 2008 crisis, their models ignored money and banking.” Salim Furth replied “I was in grad school before 2008 and can confirm this is false.” Ryan Decker (‘Updated Priors’) added “here comes the goalpost moving. “But they didn’t include money the way I think they should have” etc.”
Putting aside the issues I have with this amnesiac account of economists’ view of finance pre-2008, Hanke’s tweet is rhetorically unwise as an absolute statement because it is belied by there being a single pre-2008 model which tries in any way, shape or form to include money and banking, something that is not hard to find. On the other hand, he is getting at a crucial point: at their base level macroeconomic models do not include a financial sector. Like utility-maximisation and bounded rationality, attempts to model money within this framework often will profoundly miss the point.
Before 2008 the dominant way of including the financial sector was the ‘financial accelerator’, a mechanism that served to amplify non-financial shocks (eg to productivity) through the debt of non-financial firms. So this was not suited to anticipating or understanding financial crises, in which problems come from the financial sector itself. While it is inaccurate to say economists ignored money and banking entirely, saying that they ignored the most crucial characteristics of money and banking is much more defensible. While I possess no window into Hanke’s mind, given his mention of 2008 it is reasonable to think that a model of ‘finance’ which had no space for ‘financial crises’ would not really ease his concerns.
Because someone will ask about post-2008 macroeconomics: until it includes banks extending credit as a fundamental part of what drives the so-called real economy, until crises are generated endogenously rather than imposed from without through imaginary productivity shocks, and until these things are not attributable to any ‘frictions’ but are just how the model works at its core, I will not consider it a meaningful representation of the real financial sector. Otherwise it’s like saying you have a model of antlers when the antlers in your model are spherical.
So, two lessons here. For critics, try to be more specific and avoid broad rhetorical statements like “economists ignore X”, where X is something broad like bounded rationality or finance. Instead try “economists have a very limited notion of X” or “economists ignore x”, where x is something more specific like non-optimising bounded rationality or endogenous generation of financial crises. For economists, don’t think critics are moving the goalposts when your model which purports to represent X actually misses what the critic was trying to say. Just because you use the same words as them, doesn’t mean you’re getting at the same thing.