You’re part of a group—maybe family, maybe fellow investors or professionals, maybe a team of managers or employees—working together to complete a project that none of you could complete on your own. It could be a small project, like beautifying the land enclosed by a cul-de-sac shared by five homes or providing dishes for a family gathering, or a larger one, like researching and writing a report about a new product line or community center. The dynamics at work are even present on far larger scales, like determining contributions to the U.N. or N.A.T.O., or what effort individual countries contribute to fighting climate change.

Most of you are doing your best to push the project along, but a few seem to keep probing to figure out how little they can get away with doing and to what extent others will pick up the slack for them. If replacing them with others who’ll pull their weight isn’t feasible, should you keep contributing your own effort or funds, pull back to try to teach the laggards a lesson, or give up and stop throwing good money (effort) after bad?

Dilemmas of Cooperation

Both economists and psychologists have conducted thousands of studies of dilemmas of cooperation that include people in this sort of position. One of the most pervasive themes of those studies is that many people intensely dislike having others free ride on their efforts, and that cooperation often founders on the nearly ubiquitous presence, along with these detesters of free riding, of individuals who try to get away with doing less.

Studies by behavioral experimental economists establish that most individuals are inclined to contribute more the more that they know others are contributing, which seems like good news compared to the assumption that everyone is determined to contribute nothing at all. People tend, that is, to be “conditional cooperators.” However, the large majority of conditional cooperators are also “selfishly biased” in the sense that their ideal is to increase their effort roughly but not quite fully in proportion to that of others—in other words, to try to more or less keep up, yet do “just a little less” when possible. If everyone’s goal is to match others’ contributions roughly 80 or 90 cents to the dollar, the long-run result can still be a race—or at least a slow jog—to the bottom, one in which everyone’s effort slowly approaches zero as each tries to do about as much as, but just a little less than, the others.

However, it’s actually far from clear that everyone is equally inclined to free ride on the efforts of others. It’s also not clear that all have as strong a tendency to withdraw effort on seeing that one or more others are contributing less. Some may value the goal more than others, take pride in doing more, or treat the presence of a few slackers as a fact they’re willing to live with provided that the group contains enough cooperators to get the job done.

In a laboratory game, for instance, five people may each be given $10 and asked to divide it between an amount kept private and an amount pooled with the others’ money, tripled, and divided five ways. If all pool their full $10, the combined $50 becomes $150, and divides into $30 for each—a 200 percent profit. If only four put in their $10 and the fifth puts in nothing, the pool has $40, which becomes $120 and each of the five gets $24, which is still 140 percent profit for those who contributed. Would the four contribute $10 each if they knew that it would give each of them $24 instead of the original $10, and that there would be a hold-out who gets the same $24 share and also keeps their own $10, thereby earning $34? Perhaps no. But refusing to contribute, in view of such knowledge, and earning $10 instead of $24 due to unwillingness to be taken advantage of, would be a case of “the best being the enemy of the good.”

Recent Experiment

In recent research, my collaborator Arhan Ertan of Bogazici University (Istanbul) and I invented a variant on the conventional public goods experiment with which to study the conflicts between the desire for joint profit and the resistance to being taken advantage of. Five anonymous players are randomly grouped together, and each gets 10 currency units (CU) in each of 10 periods of play. If any combination of the five puts a total of 110 CU into a joint account, all five receive 10 CU during each remaining period. Earnings are thus highest if each puts their 10 CU into the project in the first two periods and puts in 2 of 10 CU in period three, since they then earn 0 in periods 1 and 2, 18 (their kept 8 plus 10) in period 3, and 20 in each of periods 4 to 10, for a total of 158 CU rather than the 100 CU earned if each simply keeps all CU available to them.

But the players are unable to discuss their situation and unable to form a binding agreement. Predicting what people will do in this situation is nearly impossible, since so much depends on what each believes about one another’s beliefs and motives. One of the most important things one needs, to form a good prediction, is a sense of whether, in the face of evidence that not everyone is doing their “fair share,” any cooperating group members will pull back or quit entirely, versus forging ahead so long as there are enough other cooperators that they can make themselves better off by completing the project despite knowing that some others are “making out like bandits” by free riding on their efforts. With no clear prediction of what would happen, we set out to see what a few hundred study participants would do in fact.

While there is great variation among the outcomes and while our analysis is still ongoing, it’s striking that in not one of the more than 70 groups of five participants studied did most members give up on putting the necessary funds into their group project, although every single group had at least one “slacker,” which caused the benefits of cooperation to be unequally shared. Reaching the target followed different courses in different groups, with 30 percent of them reaching it in the minimum possible time, hence coming close to maximum earnings. The records of the five players’ individual decisions on how much to put into the project, period by period, read like dozens of sagas in which there are always at least one or two acting as if daring the others not to keep going, and often a majority displaying hesitation at one or another point in the process.

The key point, however, is that almost all players ended up improving their earnings—scarcely any earned less than the 100 CU available without project completion, and on average they raised their earnings to more than 142 CU, which is 90 percent of the maximum possible. In a sense, it was only possible because enough members of each group turned out to be willing to live with being taken advantage of a little by whichever group members chose more selfish decisions. Cooperation in spite of some holdouts can occur, and can benefit the vast majority, so refusing to go on because one or two don’t carry their weight was not the main order of the day among these study participants. There may be a lesson from this for both family dynamics and global cooperation.