|
I'm starting to notice a trend whenever I engage with economics on these matters. In fact this pattern shows up every time, it's like they are trained to do this. It starts with an acknowledgement that the theory is stylized beyond any practical purpose, or that it is merely a pedagogical "tool". They then point out that any fault I see with economic theory and/or methodology must be my own failure to understand the material. Like, sure, go read Varian or Mas-Colell's tomes before you try to bring up anything wrong with elementary supply-demand analysis. It's worth noting that they never go out of their way to explain what the supposedly superior methods are. To me it's just moving the goalposts. In fact, I just stumbled on this paper, the conclusion reads as something straight out of Monty Python's Flying Circus: Complexity and Economics: computational constraints may not matter
By F. Echeneque, D. Golovin and A. Wierman Conclusion:
"We show that the notion that economic agents have limited computational resources adds no empirical content to the theory of utility maximization. A data set of observed consumption at different budgets is either in contradiction with the hypothesis of utility maximization, or it can be explained using a utility function that is easy to maximize. Our paper is not a critique of the literature on complexity and economics in general; rather, we take issue with the idea that worst-case hardness of a model implies that the model is flawed. We emphasize that the existing results on complexity are useful for understanding how economics can be applied in a normative and algorithmic way — for example, to engineer economic systems with desirable properties. We posit that while computer scientists tend to think ‘algorithmically’ about economic models, economists tend to think ‘empirically’ about the models. There is a need for considerations of computational complexity in both views. In particular, an algorithmic view of economic models assumes that the model is fixed and literally true, and then proceeds to ask about the computational demands placed on the agents by the model. That is, it assumes that the agent is simply an implementation of the model and asks whether the agent can efficiently compute its decisions. In contrast, an empirical view takes the model as a tool for thinking about reality. One does not presume agents literally follow the model, only that the model provides a way to explain the observed behavior. In this view, a model still loses credibility if the agents must solve computationally hard problems; however, worst-case complexity is no longer the relevant concept. Instead, the question is whether data from an observed phenomenon can always be explained by the theory with the additional constraint that agents are not required to solve computationally hard problems. This is the case with the theory of the consumer. On the other hand, we expect complexity to matter empirically for other economic models. When that is the case, one would want to characterize the added empirical consequences of assuming that economic agents do not solve hard problems." Source pdf: http://users.cms.caltech.edu/%7Eadamw/papers/letter_sigexc.p... TL:DR; Yes the critique is completely valid, except in highly restricted environments that never occur in the real world. But look, that's not what we are actually trying to do, oh and by the way, those critics just don't understand economics. In any other field you'd be laughed out of the room, economists parade this stuff with a straight face, and if you object "you just don't understand economics". > > The world is a complex adaptive system, how does micro or macro equilibrium fit that any better than homo economicus? If anything it sounds even worse to me... > Chemistry is a field of complex and dynamic systems, yet we are still able to talk about equilibria there. Equilibrium doesn't mean nothing is changing - and in fact, it doesn't mean that the equilibrium corresponds to any actual observable state of the world. Due to aforementioned dynamism, observation effects, and other complications, it may not be possible to observe the true equilibrium state at all, but it's still a useful construct in our understanding of the chemical world. I might have worded that awkwardly. Yes equilibrium is a thing, I don't see how modeling markets as tending towards equilibrium is helpful in any meaningful sense. In fact, I think it actively obscures the dynamic nature of real world economics. > > I mean, I have never seen an econ 101 lecture mention things like path dependence > I'm sorry to hear that you've never experienced a good introductory economics course. However, that experience (or lack thereof, as the case may be) doesn't serve as testimony against the field. No comment :) |
I could say the same, in that I notice a pattern too. It's like HN commenters are trained to engage in extremely facile criticisms of fields that they have little knowledge of, understanding of, or respect for, and then reject any substantive critiques to the contrary due to their lack of understanding.
> It starts with an acknowledgement that the theory is stylized beyond any practical purpose, or that it is merely a pedagogical "tool"
Nobody has said either of those things. They've said that it's a model. As Box famously said, "All models are wrong, but some models are useful". The key to using a model, though, is understanding its scope and limitations - where it does not apply, but also where it does, and in which ways.
> They then point out that any fault I see with economic theory and/or methodology must be my own failure to understand the material. Like, sure, go read Varian or Mas-Colell's tomes before you try to bring up anything wrong with elementary supply-demand analysis. It's worth noting that they never go out of their way to explain what the supposedly superior methods are. To me it's just moving the goalposts
Nobody's saying that you have to do doctoral research in order to discuss the limitations of elementary supply-demand analysis. It's just that the critiques you're bringing up (in this thread, but with this pattern more broadly) are somewhere between "not very interesting, because they have no relevance or are mitigated elsewhere", and "based on complete misunderstandings of the fundamental concepts at play".
Ultimately, it ends up something like the physicist from XKCD #793.