Economists And Their Assumptions

The standard (or neo-classical) view of economics makes a lot of assumptions. The main ones are that people have rational preferences, they are self-interested, they are utility maximisers and they have access to all relevant information (including information about the future). The economy is assumed to be in equilibrium, markets are efficient and perfect competition reigns (of course this is a simplification). These assumptions come in for a lot of criticism but they are defended as necessary simplifications. However, the assumptions economists make have a huge effect on the world of economics and therefore world economies.

One response to these assumptions that is popular among non-economists is to deride economists and call for their assumptions to be dumped. A great many people feel that viewing the economy as governed by self-interested utility maximisers as so unrealistic as to be useless. They say we should ditch it all and start afresh. Most of the key pillars of neo-classical economics have been debunked. There are numerous studies showing that people are not rational, preferences are not stable, people do not have perfect information, markets are not efficient and people do not have utility curves. Particularly since the Financial Crisis there has been anger at economists, many of whom are seen as out of touch with reality.

miracle_cartoon

Economists respond to this by either ignoring complaints or arguing that the assumptions must be simple to make their work easier. A common phrase is that “All models are wrong, but some are useful”. Economists claim that the economy is too complex to accurately model, so they must use a simplified model in order to provide insights. They claim they can still learn a lot about the economy from these simplified assumptions. One lecturer I had begun her class on utility functions by saying that none of this was true in the real world, but we had to learn it in order to construct models. Of course no one really believes in perfect competition, she said, but it is too complicated to create an accurate model and a simplified one gets us as near to the mark as possible.

I have two main objections to this. First of all, economists have a tendency to forget they are simplifying. It is common to hear economists criticise minimum wages or unions or taxes based your standard economic textbook explanation, seemingly forgetting about the unrealistic assumptions behind such theories. For example, take a look at this post. I choose Bryan Caplan as he is one of the more respected and well-known economists. Yet his post essentially argues that when history contradicts basic economics (the unrealistic assumptions no economist supposedly takes seriously) he believes that history must be wrong. He assumes that all workers had perfect information and were paid their marginal productivity of labour. When heterodox economists criticise these theories they are sometimes accused of attacking a straw man. (In case you think that was a once off, see here)

Secondly, it is somewhat absurd for a discipline to base itself on foundations generally agreed to be false. If we know people are not rational, should it not make more sense to build theories on how people really do behave? Is it not better to be roughly right than precisely wrong? Think about what unrealistic assumptions do to economics. It pushes away students disappointed that studying economics is of no help in explaining what is happening  in the world around them (even Masters in economics are based on unrealistic assumptions, it is only academic economists who come close to describing the real world).  It leads to self-selection so that those who don’t mind or even agree with unrealistic assumptions so that those who end up becoming academic economists see no reason to change things.

Finally it leads to stagnation. If we close off large parts of economics and refuse to listen to criticism of the core assumptions, then how is the field of economics supposed to grow? If new ideas or changes to neo-classical economics is dismissed or ignored how can economics advance? Is economics supposed to be forced into more and more obscure areas or more and more complicated (though no more accurate) models? Are neo-classical assumptions to be treated as non-falsifiable truths even when they’re wrong?

Other economists would say that perfect competition etc is only a starting pointing and that distortions and market failures get added on to models. However, your starting point has a strong impact on your end point, just as there is a large difference between assuming innocence until proven otherwise and assuming guilt. If markets are assumed efficient and consumers rational until proven otherwise this leads to an overly rosy view of markets and consumers, giving them powers they do not have. If you don’t believe me, imagine what it would be like if we assumed that the free market was dominated by monopolists. This would be easy to model and no less accurate than assuming perfect competition. But this view of the market would be far more negative with a much greater emphasis on government intervention.

But the crucial point is that unrealistic foundations don’t bring you closer to reality. Not only is a large core of economics (especially microeconomics) unhelpful in describing the real world, but economists are largely looking in the wrong direction. Economists could study how consumers really make decisions in all their irrational glory. It would be imprecise, inexact and not very mathematical, but it would be real. Or they could keep drawing indifference curves and constructing mathematically impressive but functionally useless models. They could look at how economies really function, with their inefficiencies, market failures or they can create perfect economies in the clouds. They can either look in the weeds where the key is or under the streetlamp where the light is better.

I once went to a talk by the lecturer I mentioned above, who admitted that the utility curves she had to teach were not real. She was presenting some game theory research of hers where she examined interactions between inventors and firms in the world of innovation. It was an interesting paper, but there was one glaring hole in it. None of it was real. She didn’t interview any people, study any business or examine any data. All she did was make a lot of assumptions, create an algorithm and get a result (which in this case was that private sector funding of research was more efficient than public sector). I use her as an example not because she was a neo-classical, but because she wasn’t. She voted for a left wing European party and was critical of many economic assumptions. Yet she too created a model with little connection to the real world.

There are many serious problems affecting economies over the world. They have been badly damaged by the Financial Crisis, the policies of austerity and most have still not recovered. If we want to effectively combat these recessions, we need economists who are willing to examine the world as it really is, not as they wish it was. Simplified models with only one consumer in the economy or world where people have perfect information, leads economics in the wrong direction. It brings us away from the real world of bubbles, busts, panics and euphoria’s and instead economists solve problems that do not exist while the real problem of millions of unemployed is ignored. If you take a perfect economy as your starting point, then the idea of systematic failure will never cross your mind. The problem is these crises occur whether we are ready for them or not. It is time economists climbed down from their towers and got to work on the real problems of society.

Advertisements

14 thoughts on “Economists And Their Assumptions”

  1. Agreed. In particular, it’s fascinating how little economists consider the “firm” as an economic and legal construct. I know some do, but it’s still assumed away. Yet when lobbying firms can literally write legislation, and create markets and monopolies, and do, then it’s pretty inexcusable to ignore it.

  2. I remember Caplain’s post on how women really were much better off in the 19th century than now because the total lack of legal rights didn’t really mean anything but whoa, economic freedom!

  3. I can only hope that in about fifty years we’ll look back on all the meaningless mathematical models of the neoclassicals in much the same way we look back on medieval theologians arguing about how many angels could dance on the head of a pin.

    Both involved scholars becoming so suffocated by their own rhetoric that they ended up constructing an alternate universe for themselves abstracted from anything resembling the operations of the real world.

  4. “All models are wrong” is a truism that contains no informational value whatsoever and is just used to deflect substantive criticisms. Case in point: Krugman versus Lars Syll on IS/LM. Syll makes a number of points about internal coherence, methodology and empirical relevance; Krugman replies with:

    “Well, of course it doesn’t. It’s a model – a simplification of reality designed to provide useful insight into particular questions. And since 2008 it has done that job, yes, brilliantly.”

    http://krugman.blogs.nytimes.com/2014/03/26/dare-to-be-silly/

    ..then goes on to a bait-and-switch about hyperinflationists, as if Syll or other post-Keynesians ever predicted that.

    Having said this, Robert, I do feel like you focus a lot on the core theories of perfect competition in your critiques. A standard undergraduate curriculum contains plenty of discussion of information asymmetry, moral hazard, monopoly etc. Now, these are all flawed in their own way (eg the assumption that probability distributions and outcomes can be known by agents), but if you only focus on perfect competition you risk the accusation that your critique of economics is limited or narrow. Just my 2c.

    1. Hey UE, you’re right that I do focus a lot on core teaching of economics, especially in the early posts, but I’m slowly moving outwards. I could feel the focus was a bit narrow when I was writing this so I tried to broaden it in the later paragraphs. Thanks for your input, its always good to hear from you and you should drop a comment more often.

  5. perfect comp is a teaching tool that you would never find in a journal. Which model says agents have access to information about the future? I think rational expectations says that agents don’t make systematic errors that are correctable. You won’t turn up 5 minutes late for the bus every day once you know what time it’s at.
    I don’t think equilibrium is something alot of economists cling to either. Your characterisation of economists as people who cling to false beliefs in the hope of learning one minor insight into the economy is wrong too. A lot of the assumptions you point out are pedagogical tools used to get students thinking about concepts.
    Criticising economists for not talking to business people is weird. Will information gathered like that be as useful as statistical inference gleaned from actual behaviour? It’s possible that these people could lie and mislead the researcher. Just some thoughts.

    1. “perfect comp is a teaching tool that you would never find in a journal”

      Well, you see perfect competition crop up sometimes in papers if the writer is exploring something else. But in any case, does the fact that PC is a “teaching tool” insulate it from criticism? What if it is so far away from reality that it does not even serve this purpose? Is there any piece of evidence that could make you believe this?

      “You won’t turn up 5 minutes late for the bus every day once you know what time it’s at.”

      Well, some people do! But more importantly, knowing exactly how the entire economy works is not equivalent to knowing when a bus will arrive (which is a clear, single piece of information, transparently stated at the bus stop with some random error). I mean, even economists don’t yet know how the economy works, so why would agents? The whole thing is an exercise in circularity: assume the economy works this way, then say agents have to conform to that model because it’s how the economy works.

      “Will information gathered like that be as useful as statistical inference gleaned from actual behaviour?”

      Well, yes, because you have to make assumptions about probability distributions and data that may not be true to do this. There is also the ever-present issue of causality, which surveys do more to uncover by discovering people’s actions and motives instead of inferring them.

      “It’s possible that these people could lie and mislead the researcher. Just some thoughts.”

      It must be pretty depressing being in this kind of mindset, where everybody else is assumed to be a self-serving liar until proven innocent.
      It *could* be the case that survey respondents mislead; but it could also be the case that they simply tell the truth. I mean, why would you lie in a survey?

      In any case, the other social sciences (sociology, anthropology) know how to design surveys to avoid these kinds of pitfalls and many others. Maybe economics could learn something from them.

      1. Could you cite a paper where the person uses PC please. How close it is to reality doesn’t matter. More realistic models of production are framed using the same ideas- MC, production functions etc.

        If you provide evidence I’ll consider it. Are you taking the piss with that comment? Have I intruded on some sort of cult?

        Some people might turn up late all the time but are they a significant enough portion of the whole to worry about? hopefully evolution will weed them out. Consistently making an error is pretty damning evidence that someone is stupid.

        I think you make a fair criticism about imbuing people with knowledge about the economy. It’s orthogonal to my point that no models (that I know of) have people in them that know the future. Is that what people call switch and bait? I’m not having it.

        I think it’s really cool that you read somewhere that correlation is not causation and consider econometrics completely flawed as a paradigm for understanding behaviour. It has its problems, no doubt, but is a more useful road to go down than assuming things about survey respondents.

        It’s cute that in a column about assumptions you make a pretty huge (HUGE) assumption that survey respondents don’t lie. It’s based on your intuition? your own behaviour? They could lie or they could tell the truth, they have two options. Misleading the surveyor could be intentional or a genuine error but misleading nonetheless. This is a big problem that probably doesn’t have a satisfactory solution.

        I misunderstood the point made by the op. He was criticising a theoretical model simply because it was a theoretical model. The above article is a pastiche of bland unspecific criticisms about assumptions. If you can give me concrete examples of a situation where an assumption is used that completely mis-charactersises a behaviour of someone/something I would be happy to read it.

        It’s pretty funny that the critics of something that is incredibly specific ( and benefits a lot from this) can be so vague.

        I think the tone of your reply (unlearning econ) was unhelpful. You used a semi colon, you’re obviously a well educated man, can we keep it civil?

        1. “Could you cite a paper where the person uses PC please.”

          “How close it is to reality doesn’t matter.”

          What? So you can just use any model as a starting point, no matter how absurd?

          “More realistic models of production are framed using the same ideas- MC, production functions etc.”

          These ideas are themselves flawed. For MC, see below on pricing theory. For production functions, see the Cambridge Capital Controversies. If that’s too dense (and I’m with you if you think so) then some of the more basic issues with it are here:

          http://www.arnoldkling.com/blog/whats-wrong-with-the-neoclassical-production-function/

          “Consistently making an error is pretty damning evidence that someone is stupid.”

          An error relative to what? Again. you’re assuming there’s some sort of ‘optimal’ solution towards which people can converge, but often reality is just too complex for anything like that. I mean, do you optimise much in your life? I don’t even know what that would mean in most situations!

          There’s also the point that while individuals can be (loosely speaking) rational, their actions might aggregate into an outcome that is irrational, such as investors being forced to bid on a bubble even though they know it’s a bubble, simply because there are short term gains to be made from riding the wave.

          “I think it’s really cool that you read somewhere that correlation is not causation and consider econometrics completely flawed as a paradigm for understanding behaviour”

          This is pretty off the mark – I’ve taken a number of econometrics classes, and it is actually my favourite area of economics, in that I think it is the most useful. However, it is only one tool and statistical evidence is not the only evidence that should be considered sound.

          “It’s cute that in a column about assumptions you make a pretty huge (HUGE) assumption that survey respondents don’t lie. It’s based on your intuition? your own behaviour?”

          No, I simply questioned the assumption they were *necessarily* lying. I don’t really understand why there’s an incentive to lie in a survey – I mean, if you’re unhappy with the survey, why bother answering at all? One reason to tell the truth *is* self-serving: to get your opinions and voice heard. Would *you* lie in a survey?

          “This is a big problem that probably doesn’t have a satisfactory solution.”

          I agree that survey design is a problem for some of the reasons you state: confusion, honesty etc. Yet the correct answer is not nihilism, and it’s silly to assume people who conduct surveys have not thought of these issues. You can correct for them in numerous ways: paraphrasing a question and asking it repeatedly to check for consistency; cross-tabulating results against different types of evidence; defining terms and making sure questions are not ambiguous or too long.

          It’s also worth noting that statistical information is often imperfect, too, and is actually sometimes based on survey results! GDP is calculated using a lot of survey evidence, and for example the famous Card-Krueger minimum wage paper used survey techniques. It’s also worth noting CK later found their survey findings were consistent with other types of data, which suggests those who answered their survey were not lying at least.

          “It’s orthogonal to my point that no models (that I know of) have people in them that know the future. Is that what people call switch and bait? I’m not having it.”

          Well, standard overlapping generations models posit that agents have perfect foresight so that they can plan their consumption intertemporally. However, you are right that the ratex used in other types of macro models isn’t the same as predicting the future. Whether it is more or less ridiculous is up for debate, though…

          “If you can give me concrete examples of a situation where an assumption is used that completely mis-charactersises a behaviour of someone/something I would be happy to read it.”

          I’ll give you some blog posts to save reading through 20 page papers or entire books.

          1. Utility maximisation. People do not behave in the way this theory predicts:

          http://curiousleftist.wordpress.com/2013/08/30/debunking-economics-part-1-5-the-myth-of-the-rational-consumer/

          2. Profit maximisation. Firms often don’t maximise profits and they generally do not even use the concepts of marginal cost:

          http://socialdemocracy21stcentury.blogspot.co.uk/2013/12/post-keynesian-price-theory-101.html

          The top few links deal with the competing theories best. An alternative is return-seeking firms:

          http://ckmurray.blogspot.co.uk/2014/01/time-for-new-theory-of-firm.html

          3. The Euler Equation. Consumers do not behave in the way it predicts:

          http://noahpinionblog.blogspot.co.uk/2014/01/the-equation-at-core-of-modern-macro.html

          Etc.

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 )

Twitter picture

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

Facebook photo

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

Google+ photo

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

Connecting to %s