I recently heard an informal argument that went something like this:

  1. Through individual self-interest and freedom of production and consumption, the best interest of society, as a whole, are fulfilled. [Investopedia: What Is the Invisible Hand in Economics?]
  2. But there exist games in which the Nash equilibrium is inefficient in Pareto sense. E.g. the prisoner's dilemma shows that when each agent pursues their own self interest, the equilibrium outcome is worse for each agent than another outcome.
  3. Therefore, 1. must be incorrect, even though free-market proponents often refer to 1 as some kind of proof of optimality of the free market.

I had never thought about 1. and 2. together before, and connecting them on the spot was a bit challenging. My initial reaction was: I have to look up the assumptions behind each one, but even without that, mere existence of games with "bad" equilibria does not imply the "game" considered in 1. must have a bad equilibrium. But to be fair, given the complexity of the game in 1., the probability that a free market would necessarily lead to "the best" possible equilibrium seems vanishingly small. But my gut feeling is that a free-market solution may often be reasonably close to an equilibrium such that better solutions would be hard to come by.

What would be a good response to the 3-point argument above? (Nitpicking is unlikely to persuade the counterparty.)

Sorry if this is too loose/vague, but it came up in an actual discussion with a non-economist and got me thinking.

  • 1
    $\begingroup$ "What would be a good response to the 3-point argument" It is not really a "three point argument", point 1 is merely a claim without supporting reasoning given; also it is so undefined and overly general that is likely to be false! Point 2 provides what most would consider to be a good counterexample to the claim, thus depending on the exact definitions point 3 is correct. $\endgroup$
    – Giskard
    Commented Jul 14, 2023 at 13:55
  • $\begingroup$ If someone were to ask me about this argument, I would simply direct them to the classic market failure cases presented in most intermediate micro textbooks: market power, information asymmetry, public goods. These are very clear real world problem cases; in most models of them self-interest does not lead to Pareto-optimality. $\endgroup$
    – Giskard
    Commented Jul 14, 2023 at 13:56
  • $\begingroup$ @Giskard, the reasoning for 1. is (or at least should be) in Adam Smith's Wealth of Nations. 2. says counterexamples may exist in principle, but a prisoner 's dilemma is not necessarily characteristic to the markets in general (or at least this needs some elaboration). Classic market failures are good examples, though. $\endgroup$ Commented Jul 14, 2023 at 14:47

1 Answer 1


I think the response you suggest would be reasonable, but I think, the premises on which it is based lack some context and also its important to note that this cannot be resolved just by referring to theory (which is what you attempt to do).

The 1 is simply not correct argument, but it is also a bit of a straw-man of the actual arguments thinkers use. For example, Smith never claimed that 'invisible hand' always leads to the best interest of society. He claimed that generally this holds (see Smith Wealth of Nations Book IV Ch2). So 1 is simply a mistaken/simplified/straw-man version of an original claim. Investopedia is generally known to often contain mistakes so in their case its most likely mistake rather than intentional straw-man.

Next I believe the following will help to contextualize and hopefully answer your question:

  1. You can mathematically prove that if the markets are perfect (no informational asymmetry, competitive, secure property rights, no externalities or access to cheap enforceable contracts that allow for Coaseian bargaining etc.) the markets will lead to optimal allocation of resources and they will maximize 'vanilla' Marshallian social welfare. However, it is important to note that it does not necessary maximizes social welfare under ideologies that give different weight to different classes of people e.g. under Rawlsian social welfare function that only cares about welfare of the poorest and disregards welfare of any other member of society this does not necessarily holds.

    I am not going to provide the proof here, you can look up the first welfare theorem to see the proof. You can find it in any graduate textbook of public economics. You can also re-contextualize such proof in terms of Nash equilibrium, so for example, in the prisoner dilemma you get bad equilibria because people can't make binding contracts which would change payoffs.

  2. If markets are not perfect, there are informational asymmetries, non-competitive market structures such as monopolies, not secure property rights etc. which are situations that often arise, there are typically some non-market or government interventions that can lead to more optimal outcome that 'corrects' these market imperfections.

    This can again be typically mathematically proven, however, it typically relies on assumptions that policy will be implemented by benevolent and omniscient social planner, that always tries to maximize given social welfare function. In addition, it will be usually assumed that social planer has perfect control over given instrument(s) to implement policy. Since this is broad literature I will again not go into individual models but you can read Stiglitz Economics of Public Sector for undergrad treatment or Hindriks and Myles Intermediate Public Economics.

  3. The same way as informational asymmetries etc disturb efficient operation of market they also disturb efficient operation of government. For example, once you relax assumption of omniscient social planer, even if you keep the benevolent part, a lot of results that show improvements from government action break down.

    In fact this was kinda part of Smith's argument as he mentions in the same chapter (emphasis mine):

By pursuing his own interest, he frequently promotes that of the society more effectually than when he really intends to promote it. ... What is the species of domestic industry which his capital can employ, and of which the produce is likely to be of the greatest value, every individual, it is evident, can in his local situation judge much better than any statesman or lawgiver can do for him.

  1. Real life social planner, even in developed and advanced democracies, are not generally benevolent. This is not to say they are evil, or that there is no difference between dictatorships and democracies but even best social planers have face incentives (e.g. pursuing re-elections etc), that may lead to adoption of sub-optimal policies (see many examples in Muller Public Choice III).

  2. As a consequence of 1, 2, 3 and 4 there is an ambiguity about whether markets benefit society, perfect markets do but not all markets are perfect. On other hand, 2, 3 and 4 suggest that cure for market imperfections has potential to be worse than the disease.

    So only way to solve this conundrum is to refer to some empirical observations. Empirical studies generally show that countries with open trade, free labor and goods market, free enterprise, secure property rights, centralized government with enough power to enforce contracts tend to be more prosperous than countries that lack most of the above (see Acemoglu & Robinson Why Nations Fail). Of course, most of the institutions above are not just binary, you don't just have secure property rights or no property rights, so there are still open questions about the optimal degrees of these institutions, but that does not contradict the general tendency that when these institutions are generally present societies tend to prosper (at least materially).

Now if you combine 1, 2, 3, 4 and 5 you get essentially your response, as already mentioned above. To sum it up, although perfect markets do lead to efficient allocation of goods and maximizes sum of people's welfare, real life markets are typically not perfect. This opens possibility for 'the best interest of society' to be achieved via various interventions. However, due to government failures and incentives that policy makers face interventions may not improve outcomes.

Then it becomes empirical question whether markets, on average, tend to beneficial outcomes or not. Generally, speaking evidence shows that in many cases, but certainly not all, they do. This is basically what Smith observed, since his argument was based not just on empty theorizing but on qualitative empirical assessment of history and contemporary society.

  • 1
    $\begingroup$ Sounds good. I would add that governments also need to be able to calculate an optimal course of action, and that is technically infeasible even up to a rough approximation. Therefore, central planning could not work well even if the government were fully benevolent and able to execute its policies perfectly (none of which is the case). $\endgroup$ Commented Jul 14, 2023 at 14:59
  • $\begingroup$ @RichardHardy ...are profit maxizimers and utility maximizers able to calculate optimal courses? Seems like an impossible standard. $\endgroup$
    – Giskard
    Commented Jul 14, 2023 at 21:00
  • $\begingroup$ @Giskard, no, but their problems are much easier than that of the central planner, so they (unlike the central planner) are able to get "close enough". $\endgroup$ Commented Jul 15, 2023 at 5:38
  • $\begingroup$ @RichardHardy Sounds interesting! Can you please provide a scientific reference for your claim? $\endgroup$
    – Giskard
    Commented Jul 15, 2023 at 13:46
  • $\begingroup$ @Giskard, no, this just a common sense argument that seems quite intuitive to me, but I think it is too general to have scientific backing. It could also be wrong. $\endgroup$ Commented Jul 15, 2023 at 20:05

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.