Clearer Thinking on Collective Action

We all know of important issues that require collective action to mitigate or solve, with climate change as the archetypal example. At a national scale, individual states might all wish to set a standardized corporate tax system but feel compelled to undercut other states’ rates in order to attract business. Maybe a bunch of disgruntled workers all want to strike, but have no union to coordinate and know that unilateral action will be punished.

One thing you’ll hear in the public discourse is a vigorous disagreement about who is responsible for these types of problems, and who is responsible for solving them.

In particular, many on the left point out that emphasizing individual action and responsibility is not just an unjust scheme by the corporate world to evade blame for climate change, but is actually counterproductive. Think you need to stop keeping your home a comfy 77 degrees in the middle of January to reduce your personal “carbon footprint”? Think again - that’s just Exxon’s way of keeping under the radar.

What This Really Means

Let’s consider what this type of critique is implicitly arguing. Clearly, the world - and all of us - would be better if every individual really did take individual action to reduce their energy use. At the same time, it’s pretty unlikely that any amount of cajoling is going to get that to happen. So, we need some type of state initiative and/or coercion to get a comparable result.

One obvious and effective (and not so long ago, bipartisan) solution is a carbon tax. Let’s pretend for one moment that Congress was mildly functional and a certain amount of political pressure would yield a national carbon tax, which would in turn put the same measure on the legislative agendas of other nations.

It’s plausible that an emphasis on individual action would crowd out the type of political pressure necessary for such a coordinated solution. Of course, one wishes we could encourage both types of action, but in reality political capital and public attention are finite resources. The ideal ratio of personal action messaging to carbon tax messaging depends on the public’s relative responsiveness to these message types and the degree of political pressure necessary to pass the tax.

Leftist critiques of ‘personal climate responsibility,’ then, are implying that the expected benefit of calling for personal behavior change are smaller than the probability-weighted benefit of pushing for a political collective action solution. Well, there’s probably a bit of demand for righteous retribution as well, but that’s the technocratic Steelman.

Clashing Intuitions

I have some seriously clashing intuitions about whether this “personal action is futile” critique is legitimate. One one hand, even so-called “collective action problems” can be partially solved with individual action. It might be that Joe’s decisions to go vegan and stop flying results in an expected global temperature decrease of .00002 degrees, corresponding to an expected reduction in $4,000 indamages and 0.1 lives saved.

The facts that it would be better if everyone took such action, that Joe isn’t responsible for past societal failures, that corporations responsible for much of the issue aren’t involved, and that Joe only internalizes an infinitesimal fraction of these benefits don’t change that fact. From a utilitarian perspective, Joe should take these actions if the costs to him are lower than the social benefits produced.

On the other hand, there’s no way we can expect more than, say, 0.1% of the world population to take such measures. If this happens. maybe global temperatures will increase 3.12 instead of 3.14 degrees. Congratulations?

Perhaps the fact that individuals only internalize a tiny fraction of benefit of their action really does matter. The only way to get a reasonable degree of damage mitigation is to make it selfishly worthwhile for people to take socially-responsible actions, and this almost certainly comes in the form of government-led collective response a la carbon tax.

Confusion Resolved, Sorta

After thinking a bit about it, I think these two intuitions correspond to two different types of collective action problem, and two different questions to ask.


In the first type, the social benefits from action are continuous and roughly proportional to the amount of action (or even decreasing with the amount of action), even though individuals only capture a tiny bit of those benefits. In the second, the social returns to action are increasing, perhaps to the point that some threshold number or proportion of individuals must act before there is any social benefit at all.

Type 1: Decreasing-Constant Returns

Type 2: Increasing-binary returns

Obviously, real problems will be much more complex than these graphs suggest, with slopes changing on many different intervals.


Separately, it’s easy to knot together the question of which type of public messaging will best solve the problem and the question of whether you, as an individual, have a moral responsibility to unilaterally act.

It is totally coherent and possible that emphasizing collective action at the social level will in fact do more to solve the problem than urging individual action, but that you have a personal moral duty to act unilaterally.

Combining the types and questions, we get a nice little two-by-two with answers to each question-type combo. Importantly, these answers are relative to the other box in the same column, all else equal. The “true” answer depends on real-world factors like the effectiveness of different messaging campaigns and how important the problem is.

Upper-left corner: if social returns are decreasing (like in the first graph), it probably makes less sense to try to organize collective action because most of the gains can be won by convincing a few people to act on their own. These people are probably those who are least burdened by the action, and most convinced that the problem is important to solve, so getting them to act likely imposes a relatively small cost and is relatively easy.

Upper-right corner: If returns are decreasing, it’s pretty likely there is some low-hanging altruistic fruit. This is why many effective altruists emphasize how easy (i.e. cheap) it is to save a life, for example. It only costs a few thousand dollars to prevent someone from dying (probably from malaria), in expectation, but this is partially because a pretty tiny proportion of world resources are dedicated to the task. If everyone in the world making $40,000 annually was donating 10% of their income to charity, the easiest deaths to avert would already be averted and the cost of saving a life would increase.

Lower-left corner: just the opposite of the box above, increasing returns means that it’s likely impossible to win most of the gains without getting a lot of people to act, and that entails some sort of organization or coercion.

Lower-right corner: again the opposite of the box above, there is less likely to be ‘low hanging fruit’ because gravity works the other way around. As an extreme example, suppose society wins all the gains if 50% of people act, but nothing before then. Unless you happen to be the single person on that margin, your action will have no direct effect on society.

The Multipliers

As alluded to above, real-world facts about the issue in question matter. If there is a functional political system with engaged citizens willing to organize and vote, it’s more likely that promoting collective action via politics will be the way to go - even if the problem in discussion has decreasing returns. To generalize, the answer to the public-messaging question seems to depend on the relative effectiveness of each type

If the problem is really important, it’s more likely that you have a moral responsibility to act. Even if the fruit isn’t particularly close to the ground, it’s probably worth climbing the tree if you’re starving.

Separately (and obviously), more important issues deserve more public messaging altogether. Importance also affects messaging effectiveness. People are probably more willing to tolerate, say, governmental coercion for really important issues, and so these issues will be more responsive to coordination.

There’s More

So far, everything has been described in abstract utilitarian terms. Even though it’s beyond the scope of this post, more complex ethical issues involving the distribution of costs and benefits and those involving the political philosophy of coercion/coordination are plausibly important. Maybe those more responsible for climate change really should do more to mitigate it, independently of the other factors I’ve described. Perhaps coercion is never a legitimate solution to certain issues.


So what should we do about climate change? Well, I don’t really know and am frankly frightened and pessimistic about the whole thing, but here’s my best guess based on the whole schema:

I don’t even know where to begin to evaluate whether each additional unit of effort yields increasing or decreasing returns in terms of expected temperature change, but it seems pretty likely that each additional degree of change is even more catastrophic than the one before it. All in all, this makes the issue one of increasing returns to effort.

In terms of relative messaging effectiveness, I’m pretty confident that personal messaging has been both the dominant modality and has been virtually useless over the past few decades. Probably worth assuming that effort is better spent on politics or other forms of organization.

In terms of importance, it’s really important. For the case, I’ll refer you to The Uninhabitable Earth. Where does this leave us? If the two-by-two is any guide, we should prioritize collective action in public, but you should probably do what you can to reduce emissions.

It can feel intellectually dishonest, almost a lie by omission, to just keep secret that there’s something important everyone else should be doing. Just as it’s often wise to let someone be wrong on the internet, though, I think we ought to resist this urge and focus on what will work best, ideally while holding ourselves to the highest standard we can.