In the Review of Austrian Economics, Tyler Cowen of The Marginal Revolution, writes The Costs of Cooperation (PDF) with Daniel Sutter.#

These costs associated with cooperation appear in situations where a group has the power to produce a public good, and thus also a public "bad."#

We focus on the link between an increased ability to produce public goods and an increased ability of some individuals to produce public bads. The same cooperative techniques which allow individuals to produce public goods also allow some individuals to combine and pursue their self-interest at the expense of others. By treating the beneficial and adverse consequences of cooperation separately, the extant literature overestimates the benefits of cooperative efficacy; in some cases increased cooperative efficacy brings net costs. 2

Examples of the costs of public goods production abound. The same mechanisms which support public goods production also lead to ostracism of minorities, discrimination on the basis of race, ethnicity, and gender, restrictive social norms, and collusive industrial practices. We use the term public bads to refer to collective products or outcomes which involve costs in excess of their benefits; such outcomes are public goods for some small, cohesive group but not for the broader society.

In the discussion of government involvement with this process, the authors make the point that governments are not completely coercive. While not essential to the paper's thesis, it is an interesting statement:#

Economists typically associate governments with coercive solutions to public goods problems. But government activity itself requires the voluntary production of public goods. Agents in governments must overcome collective action problems to achieve their ends. Citizens and government employees must cooperate voluntarily with government pronouncements. Tax systems rely on perceptions of government legitimacy to limit cheating and induce payment. The bureaucrats who enforce government regulations must believe that the regulations are beneficial or at least reasonable. Only easily monitored contributions to public goods production can be supplied through government coercion. Conscription can fill the ranks of an army, but inducing the conscripts to fight hard requires voluntary compliance. A government which does not receive significant voluntary contributions will be extremely weak. 8

Another interesting point related to government is the need for tight cooperative efficacy in order to constrain the power of government in the way that most libertarians would desire:#

While the anti-communitarian slant of our argument may be obvious, the costs of cooperative efficacy also challenge non-communitarian views, such as libertarianism and classical liberalism. Both libertarians and classical liberals favor relatively small governments, limited by binding constraints. The necessary means of limiting government to such levels, however, may imply a world with excess cooperative efficacy.

Libertarian views frequently ascribe large scale government intervention to the "concentration of benefits, diffusion of costs" logic presented by Olson (1965); Milton Friedman (1980) is one eloquent proponent of this view. In other words, libertarians believe that voluntary institutions do not necessarily produce the public good of mobilizing public opinion against excess government intervention. At the same time libertarians wish that this public good could be produced with greater effectiveness. If we translate this desire into a concrete trade-off, the level of cooperative efficacy must be higher if government is to be constrained.

[...]

Note that a government with low cooperative efficacy can coerce only through easily monitored contributions to public goods, like taxes. Such a government can be relatively unobtrusive. Libertarians with a strong revealed preference for privacy may well prefer provision of public goods through a large (but ineffective due to low efficacy) government to private provision through tight-knit communities. Similarly, in some cases libertarians should prefer a large and inefficient government to a smaller, more efficient government, given that the social structures needed to sustain the latter may involve excess cooperative efficacy.

This conclusion nicely sums up the position of the paper: Cooperative efficacy is a double edged sword that maybe desirable for the same reason public goods are desirable. The solution to public bads lies not in destruction of efficacy, but in a more accurate filtration system to ensure only authentic public good pass the gauntlet.#

Claims of institutional failure are more complex than first meet the eye. Solving most problems requires greater cooperation, but more cooperation does not always bring large benefits or even necessarily net benefits. Cooperative efficacy is not accompanied by a selection mechanism that filters out the public goods from the public bads. The costs of cooperative efficacy thus imply greater pessimism about our ability to improve the world through political change or superior provision of public goods.

Our analysis also suggests we should shift our attention towards mechanisms for selecting the outputs to produce and away from the level of cooperative efficacy per se. With imperfect selection (i.e., a mixed mechanism), increased cooperative efficacy leads to greater production of public goods and public bads. An improvement in the selection mechanism, on the other hand, filters out public bads without any corresponding offsetting disadvantage.