Reading Length: 7 Minutes
Keywords: Behavioral Economics, Psychology, Social Policy
The most obvious limitation to a correct anticipation of consequences of action is provided by the existing state of knowledge.
In a short essay dated from 1936 titled The Unanticipated Consequences of Purposive Social Action, the legendary social scientist Robert Merton considers the reasons why things often go awry when humans set out to accomplish things. Merton lists off ignorance, error, and self-interested delusion as the main culprits behind unanticipated consequences.
The attempt and failure to predict human behavior is the impetus of the social sciences writ large, and has been commented on by scholars since the genesis of the movement to study human behavior formally.
In some one of its numerous forms, the problem of unanticipated consequences of purposive action has been treated by virtually every substantial contributor to the long history of social thought.1
The problem looms particularly large in political science. In an influential book titled Seeing like a State: How Certain Schemes to Improve the Human Condition Have Failed, the political scientist James Scott chronicles how governmental plans to improve the human condition so often fail and even backfire.
If ignorance is one of the roots of this problem, is there any hope of clearing up this ignorance and improving our ability to anticipate the consequences of purposive action? Rather than grand theories of human behavior, a more fruitful approach would be to attack the problem piecemeal by isolating specific contexts in which human behavior is unexpected.
One example of the aforementioned approach is the research that the economists Uri Gneezy and Aldo Rustichini conducted in the early 2000s. Now a staple of the behavior economics literature, the pair of researchers discovered a surprising ‘unanticipated consequence’ of imposing fines on people.2
Late to Daycare
Traditional economic theory would predict that imposing a fine on some action disincentives that behavior. Imposing fines for speeding, for example, should reduce the amount of speeding that occurs. The authors summarize this ‘deterrence hypothesis’ as:
When negative consequences are imposed on a behavior, they will produce a reduction of that particular response. When those negative consequences are removed, the behavior that has been discontinued will typically tend to reappear.
The researches decided to put this theory to the test by working with a group of daycare centers. Prior to the study, parents who were late to pick up their child from daycare were warned not to be late but incurred no penalty. During the study, a fee of about $30 was imposed for picking up a child 10 minutes past the end of daycare.
The deterrence hypothesis would of course predict that imposing a fine should reduce the amount of parents that are late to pick up their children. Before the fine was announced, the average daycare had about 9 late pickups per week.
As shown above, imposing the fine actually increased the amount of latecomers relative to the control group with no fine. Rather than decreasing the behavior, the penalty did the opposite.
What’s more, the fine was removed in week 17, but the behavior stayed elevated. Contrary to the orthodoxy of the deterrence hypothesis, removing the fine did not lead to a resumption of previous behavior either.
What could explain this? Gneezy and Rustichini offer two theories.
The Incomplete Contracts Theory
One possible explanation refers back to the detail that before the fine was imposed, it was not clear what threat would be carried out if parents picked up their children late, nor was it clear how many times a parent had to show up late before that threat was carried out.
The ‘contract’ then was implicit and incomplete because of this ambiguity. There may be an unlikely but devastating penalty, such as possible expulsion from the daycare. this would explain why late pickups are relatively deflated before the imposition of the fine:
In order to avoid this unspecified and uncertain but possibly more serious consequence, parents abstain from “too many” delays.
Once the fine is imposed, it is now clear what the penalty will be. The potential threat of expulsion is eliminated:
The introduction of a fine makes the sure consequence of a delay a little worse, because parents now have to pay fore ach delay, but is also provides information.
The authors develop a brief game theoretical model mathematically to illustrate that the behavior would in fact be an expected equilibrium, but then turn towards a different theory altogether.
Clarifying Social Norms
Rather than worrying about whether the daycare would impose mild or severe consequences, it is possible that parents viewed a late pickup as a ‘gift’ to the parent because no fine was imposed.
Because the behavior of staying late is a gift, parents feel compelled not to take advantage of this gift by showing up late more frequently than they typically would.
Once the fine is imposed, however, the gift transforms into an economic exchange. No longer is the generosity of the daycare being extended, because they are now being compensated for the inconvenience. Since there is no longer a gift, parents will not feel guilty showing up late:
Parents feel justified in their behavior by a social norm that states, approximately,: “When help is offered for no compensation in a moment of need, accept it with restraint. When a service is offered for a price, by as much as you find convenient.”
In other words, “a fine is a price” that qualitatively transforms the moral valence of an action. A fine can remove the guilt from a behavior, licensing the further ‘purchase’ of even more of that behavior.
Finally, to explain the data showing a continued increase in late arrivals even after the fine was removed in week 17, they offer a third social norm: “once a commodity, always a commodity”.
The authors do not test these theories, but they do point to a parallel experiment they conducted with college students.
The Cost of Paying for Performance
An inverse theory of the deterrence hypothesis is that paying for a behavior will engender more of that behavior. If I want someone to do something, I should pay them, and if I want them to do it well, I should pay them even more.
The same authors Gneezy and Rustichini also put this orthodoxy to the test. In a set of simple laboratory tests, they vary how much they pay the subjects, who must answer some questions correctly.3
It would be normal to expect that subjects who are offered more pay are incentivized to do better. Surprisingly, however, the researchers find an exception:
The evidence we have presented seems to indicate that the effect of monetary incentives can be, for small amounts, detrimental to performance.
Although larger incentives can improve results, there is in fact a ‘discontinuity’ at the threshold from zero compensation to a small compensation.
As can be seen above, performance actually decreases when a small monetary incentive is offered, until recovering as the incentive gets bigger and bigger. This explains why the study is titled ‘Pay Enough or Don’t Pay at All’. If you only pay a little bit, you are actually making the situation worse off.
The relationship between incentive and performance can be said to be nonmonontonic, because it is neither everywhere increasing nor everywhere decreasing. The slope flips from negative to positive, so you can not make a simple prediction that “incentives increase performance”.4
Conclusion
The ‘incomplete contracts’ theory presented here have been controversial since it was introduced. Some researchers have investigated and corroborated the effect as recently as 2020, while others have not been able to successfully replicate the phenomenon, so for now the debate rages on.5
Gneezy and Rustichini provide a potential warning for policy makers in organizations and lawmakers in politics: efforts to incentivize or disincentive certain behaviors can backfire and lead to the exact opposite of the intended effect. Or as Merton would put it, the manifestation of The Unanticipated Consequences of Purposive Social Action will rear its head anew.
In a footnote, Merton lists “Machiavelli, Vico, Adam Smith (and some later classical economists), Marx, Engels, Wundt, Pareto, Max Weber, Graham Wallas, Cooley, Sorokin, Gini, Chapin, von Schelting” as some examples.
Gneezy, Rustichini, ‘A Fine is a Price’, 2000.
Gneezy, Rustichini, ‘Pay Enough or Don’t Pay at All’, 2000.
An important caveat of the study is that the decrease in performance occurred only when the monetary reward was contingent on performance. A flat fee for completion did not cause a decrease in performance.
For successful results, see: Kornhauser, Lu & Tontrup, ‘Testing a fine is a price in the lab’, 2020.
For unsuccessful results, see: Metcalf et al. ‘Is a fine still a price? Replication as robustness in empirical legal studies’, 2020. Note that this study is predicated on survey results rather than experimental evidence.