Market Failure 4: Collective Action Problems (Tragedies and Catch 22s)

“From now on I’m thinking only of me.”
Major Danby replied indulgently with a superior smile: “But, Yossarian, suppose everyone felt that way.”
“Then,” said Yossarian, “I’d certainly be a damned fool to feel any other way, wouldn’t I?”

Joseph Heller, Catch 22

In Part One, I extolled the virtues of the market as a coordination tool. It is very hard to get a large number of people to work together on the same problem. Any organization with a large number of people will inevitably have a significant fraction of its staff whose sole function is to make sure everyone else is doing what they should be. The market avoids this need by bundling a signal with an incentive – the price both tells you what you should do and gives you an incentive to do it. This is the essence of Adam Smith’s famous / infamous metaphor of the Invisible Hand.

So far, I have discussed what happens when the price leaves out a crucial piece of information, or what happens when the market cannot adjust to a given signal. But there can be problems on the incentive side too. Sometimes the invisible hand points people in the wrong direction as far as collective wellbeing is concerned. These situations are called collective action problems.

Often the best way to talk about collective action problems is to use game theory, which models the decisions of individual agents to figure out how a group of rational, self-interested agents would act. Without question the most famous game in game theory is the prisoner’s dilemma, but while it is a collective action problem, for our purposes I’m going to use a different game: the public good game.

The textbook version has 4 players, each of whom makes a simultaneous decision to commit up to 5 units of resources to a common pool. The total contributions to the pool are doubled and then split evenly between the players regardless of how they contributed. The payoff for each player is the amount they keep for themselves, plus half of the total contributions to the total pool. The total payoff for all players (which is the measure of overall social welfare) is the total amount each player kept for themselves plus double what is contributed into the total pool.

What this means is that the players are collectively better off if all 4 players put all their resources in, but each player does best by keeping their resources to themselves and benefiting from the contributions of the other players. The problem is that if everyone thinks that way, then no one contributes. So while the socially optimum behavior is for each player to contribute 5 units and receive a payoff of 10 (total social payoff of 40), the equilibrium for the game is for each player to contribute nothing and receive a payoff of 5 (total social payoff of 20). In other words, in this game following incentives cut everyone’s wellbeing in half.

While the public good game is not particularly realistic, introducing real world considerations doesn’t make the problem go away. While people who play this game for real don’t hoard all their resources and do contribute something to the common pool, they don’t contribute as much as is optimal. Adding more people to the game makes it worse – as the number of players goes up, the average contribution goes down. This is especially unfortunate since collective action problems in the real world can involve thousands or millions of people, not just four.

The key characteristic that makes this game go wrong is that there is a disconnect between the allocation of benefits and costs. In normal market transactions the benefits follow the costs. If you buy an apple, you pay the cost of the apple and receive the benefits of the apple. If you don’t buy it, you get no benefit and pay no cost. But if a large number of strangers were to all club in together and buy apples, then each buyer may decide that it won’t matter if they chip in a little less. After all, there’ll still be plenty of apples to go around, right? And if you think everyone thinks that way then of course you’re not going to contribute much.

This is not merely a blackboard curiosity – there are goods whose benefits are diffuse even when their costs are not. We call these goods public goods because the public nature of their benefits makes it very difficult for the market to create them in the socially optimal quantity. The key characteristics of a public good are that it is non-rival (adding extra consumers doesn’t use up the good) and non-excludable (you can’t stop people using the good). A good example is street lighting. One more person on the street doesn’t increase the cost of providing the lighting, and you can’t stop people from benefiting from the lighting even if they didn’t help pay for it. Perhaps on a small street a sense of community feeling (and informal social pressure) would be enough to get the residents to pay for street lighting, but the more street users there are, and the less cohesive a community they are, the less money each resident will be willing to pay. On a larger scale, goods like a justice system, military defense, and some public health interventions would qualify as public goods.

This may be a good point for a quick digression on terminology. The definition above is the policy economics definition of public good, but this is not necessarily how the word is used by the general public. In my experience, non-economists use the term public good in up to three ways:

1. The same way economists use it.
2. To suggest that society is better off if more of a good is consumed or produced, without the good being non-rival or non-excludable. Economists would call this a positive externality, which we covered in part 2.
3. To suggest that people themselves would be better off if they consumed more of a good. This could be an alleged case of irrational behaviour (which we’ll cover in Part 6) or a belief that there are higher-order preferences that should be prioritised over market decisions. The normal economics term for this is a merit good, and it may not reflect a market failure at all.

While ultimately you use can use whatever labelling system you want, I prefer the way economists label these phenomena because they are actually different things. When someone calls something a public good it can be hard to work out exactly why they think government intervention is called for and since the proper intervention is different in each case, it’s very hard to engage with proposals backed by vaguely-worded policy arguments. Worse, sometimes people don’t realise that there are potentially three different phenomena at play here, which leads to the policy arguments themselves being fundamentally confused.

For public goods the prescription is public funding (though not necessarily public production). By collecting revenue through involuntary taxation and purchasing the public good directly, a government can defeat the coordination problem and deliver a socially optimal amount of public goods. Note however the word “can” in that last sentence. There is no guarantee a government will automatically reach a socially-optimal amount of the good as it doesn’t have access to the feedback mechanism of market pricing. Governments should carefully consider how much of a public good they should provide and try to account for the preferences of voters, as well as their own ability to deliver the service. Also, governments should think about what aspects of a particular good are actually public goods. A vaccination campaign and emergency medicine are both healthcare, but only the first has a public good quality to it. Also, public goods whose benefits are limited to small, cohesive communities may not require any government intervention.

The other major type of coordination problem is the public good game in reverse – where benefits are private, but costs are shared across the community. Consider a lake that has fish living in it. If too many fish are caught, the fish population will dwindle, harming future yields for everyone who fixes the lake. However, each fish caught is only of direct benefit to the one who catches it. This gives fishers an incentive to overfish. If one entity owned the lake, then they would have an incentive to control fishing so as to maximize the value of their asset, but without some entity acting as a gatekeeper, the fishery is at risk of being depleted.

This problem is most commonly referred to as the tragedy of the commons, after William Forster Lloyd’s paper that first identified it. But this is not actually a very good description. Aside from being a trifle melodramatic, common ownership is not actually the problem here. As political scientist and economics Nobellist Elinor Ostrom has shown, there are any number of ways common ownership can manage a depleteable resource effectively through traditions and social norms that limit individual extraction and prevent entrants who do not abide by those norms. For another thing, corporations are a kind of common ownership, and they can also be used to prevent overuse. The real problem is when anyone can take whatever they want from the resource, and for this reason the common academic term for the problem these days is the open-access problem.

As the problem is open access, the solution is to restrict access. This can be achieved through many ways – licensing restrictions are a common type of solution, but a lot of these are not particularly smart. Equipment or season-length restrictions can turn into an arms-race between regulators and the regulated with farcical results. By contrast, a tradeable quota system (such as New Zealand’s Individual Transferable Quota system) turns the right to use the asset into a valuable property like any other. This ensures that extraction will tend to be done by those who can extract the resource most efficiently, as they will be able to pay the highest prices for quotas.

Another solution would be collective management. For a localized resource like an aquifer, set up a body to own and control the resource. The body can be controlled by the people with a claim to the resource. By imposing rules that are binding on all users, the coordination problem can be overcome.

Our next part will be on the limits of peoples’ knowledge, and the implications for market decision-making. BYO lemons.