Pervasive market failures

Most people reading this blog already get it that neoclassical economics is deeply flawed. But I’m still amazed at its persistence in the classroom and in the blogosphere. My blood boils when I see the standard neoclassical workhorse brought out of the stable when really it ought to be put to pasture.

In a nutshell, neoclassical modeling ignores market failures, seeing them as the exception rather than the rule. This leads to all manner of ivory tower neoclassical nonsense (elegant, but deeply flawed algebraic models) that unfortunately gets taken seriously in the real world. So it is common to hear economists going on about removing “barriers” to efficient market functioning in order to improve efficiency, without acknowledging that those “barriers” might be there for a reason – to correct for market failures or to achieve some other policy goal.

The teaching of microeconomics is still based on a stylized market (“widgets”), rooted in fundamental assumptions about how that market works. These lead to an “answer”, and therefore any interventions by government (like taxation, regulation, minimum wages, unions, etc) necessarily move the outcome away from that ideal. But if underlying assumptions are questionable in a real-world setting – that is, almost all of the time – neoclassical models are a little more than fantasy world that barely resembles the capitalist economy we live in. Let’s take a quick look:

Perfect competition – In neoclassical models there is no market power among producers, just an abundance of quaint “firms” all in competition with each other, and there are no barriers to entering the market. But in the real world, many markets are dominated by a small handful of companies, due to substantial barriers to entry in the form of upfront investment costs, trade secrets and intellectual property rights. In the neoclassical world, there are large numbers of buyers and sellers, and thus no need to be concerned about negotiating power. In the real world, one side often dominates a transaction, whether a company hiring labour (unless unionized), and between companies and their customers.

No transaction costs – In the neoclassical world, exchange is costless. There are no search costs or time costs associated with making a purchase, largely because perfect information is assumed. If there are no asymmetries of information between buyer and seller, then everyone knows everything and can calculate all of the probabilities associated with decision-making and come up with a utility-maximizing choice. Stiglitz and Akerlof won the “Nobel” a number of years ago for pointing out that the logic of market efficiency falls apart when information is asymmetric; more recently, Kahneman and Smith won for demonstrating that people do not behave rationally and consistently, as economic models predict.

No externalities – The real world is heavily characterized by external costs and benefits to the original market transaction. Negative externalities include air and water pollution, toxic chemicals, and greenhouse gases, which means almost every economic activity receives an implicit subsidy that grows the more they pollute and get away with it. In the real world, this justifies government intervention to regulate or price polluting activities. But there are also positive externalities – there is a social benefit to education and public health above and beyond the private benefit, which justifies government expenditures. Ditto for basic research. Even in housing, my landscaping efforts serve to improve property values for my neighbours.

No economies of scale – Large upfront investment costs of going into business mean that as production increases, the average cost of production falls. This is true in almost every industry, and those upfront costs can include branding and advertising (also not considered, as it is assumed that consumer preferences are simply given). Neoclassical theory assumes rising marginal cost (of producing one more unit of output), and this leads to the famous upwards-sloping supply curve. But in the real world marginal costs generally do not rise as production increases, and in the case of anything digital, the marginal cost is effectively zero.

Fixed technology – While economists speak of “firms”, the processes by which a firm produces for the marketplace are treated as a “black box”. There are no advances in technology, no competition among companies to innovate using technology in their processes or their products.

Most markets do not adhere to these neoclassical specs in one or more dimensions. To be fair, microeconomics does take each up in turn, it is only as one imperfection amid an otherwise well-functioning market. Lipsey and Lancaster, in their 1957 paper, The General Theory of the Second Best, derived that if one of those imperfections existed, then the whole solution falls apart. That is, interventions by government can improve the situation since the “first best” market solution does not exist.

So why not try to model economies as they really are, rather than act the proverbial drunk searching for his keys under the streetlight when they were dropped elsewhere? Keynes was the greatest economist of the 20th century not just because he made deficit spending acceptable, but because aspired to understand and model capitalist economies as they really are. His General Theory set out a framework within which classical economics would be a “special case”, the characteristics of which “happen not to be those of the economic society in which we actually live, with the result that its teaching is misleading and dangerous if we attempt to apply it to the facts of experience.” (ch. 1)

In modern context, the tendency to assume that markets work (and deviations are only a rarity) is why most economists had nothing to say about the housing bubble or the subsequent financial market havoc until it was well underway. And yet there remains a certain righteousness that economics has the answer, when many economists know little about specific markets in question, or the other issues that are of interest to policy makers. Economics programs would be a whole lot more useful if they took a look at how actual sectors of the economy work, what their market failures are, and then talk about policy solutions in a broader context that includes the range of objectives and their winners and losers.


  • That’s a great list of counterfactual assumptions. I think you missed one, though. The basic neoclassical model as I understand it does not include time. Everything happens in a single moment of transaction-clearing. As a second-order issue, having perfect information and lacking a future it also doesn’t include the idea of risk.

    Some of the failed assumptions are related. For instance, not only do real people not match the maximizing construct neoclassical economics describes, but presumably absent perfect information (and infinite calculating power) it’s actually theoretically impossible for them to do so.

    I also think that in a way your critique doesn’t go far enough. Surely a model that incorporated real life assumptions rather than the neoclassical counterfactual ones would not just show market failures, but would in general display a completely different picture of how economies operate. Totally new dynamics would emerge. For instance, I suspect that incorporating time, expensive and asymmetric information, and increasing returns to scale in a model would show a tendency (absent political action) to increasing inequality for both individuals and firms.

    Proofreading niggle: “more recently, Kahneman and Smith won more recently”

  • Marc,

    I think it would be great if the economists formed a political party and eventually a majority government. I am all for it to the point that I would organize for them. Think about it they form government and they start making textbook laws. I bet in five years they could destroy the economy in five years. It would be painful but it would finally be the end of talking head economists and perhaps the type of departments that produce them.

Leave a Reply

Your email address will not be published. Required fields are marked *