I. Introduction:
Forty years ago, Titmuss argued against paying for blood donations by suggesting that monetary incentives are subject to backfiring. Monetary incentives operate on the assumption that people respond by modulating their activity to further their own best interest. That people respond to incentives is the cornerstone of standard economic theory. The basis of this theory shapes innumerable schemes that governments use to shape behavior—from tax schemes to fines for traffic violations. However, incentives do not always work as expected and anomalies in expected behavior, like those noted by Titmuss, occur. With the addition of incentives, the perception of the task is changed. These incentive anomalies do not happen at random, but instead follow consistent patterns and can be grouped as: 1) paying for inherently interesting tasks; 2) paying for prosocial behavior; 3) paying too much; and, 4) paying too little. For each anomaly, underlying mechanisms have been identified that explain the rationale behind this seemingly irrational behavior.
II. Patterns of Incentive Anomalies
Incentives can create doubt about the motivation for a prosocial behavior— is the individual completing the task because they are ‘good’ or because of the reward they will receive? When this happens, we see the prosocial behavior is crowded out by the monetary incentive. The United States and the United Kingdom operate their blood donation schemes differently. The United States system pays many of its donors while the United Kingdom offer no monetary reward to blood donors. Titmuss, in his now the classic example of monetary incentive anomalies, compared the two systems. The United Kingdom saw a steady uptick in blood donations per capita during the 20th century with no reported blood shortages. By contrast, the United States experienced blood shortages and saw a decline in blood per capita from the 1950s to 1960s (there was less available data to analyze from the United States). (Titmuss, 1970). More recently, Costa-Font et al noted that the response to blood donation was a result of prosocial behavior being crowded out by monetary incentives. (Joan Costa-Font, Jofre-Bonet, & Yen, 2013). An article in The Journal of Public Policy describes the hostile response to Titmuss paper. While some critics acknowledged the quality of blood in the United States was poorer than the United Kingdom, all disputed the idea that monetary incentives affected the altruistic behavior of blood donation (Mclean & Poulton, 1986). Today there are numerous examples from the literature that support Titmuss’s theory. Gneezy and Rushichini studied high school students who volunteered to collect donations and found that the students were less successful at soliciting donations when they allowed to keep a percentage of the donations for themselves. Interestingly, once the monetary incentive reached a certain threshold, effort increased (Gneezy & Rustichini, 2000b). The pro-social behavior had been completely crowded-out by the extrinsic incentive and the adolescents were now responding in a manner to maximize their own self-interest.
Another incentive anomaly seen is when monetary incentives are introduced for inherently interesting behavior While being rewarded for completing a task that an individual already enjoys may appear to be a win-win for an individual, it can instead lead to the individual questioning the motivation of the reward (Kamenica, 2012). Deci illustrated this in a series of experiments that measured the performance of individuals solving puzzles with or without a reward for completing the puzzle. After a reward was introduced (and subsequently removed) the motivation to complete the puzzle diminished. This suggests that the intrinsic motivation to complete a task decrease when extrinsic rewards are introduced (Deci, 1971).
Paying too much or too little as an incentive can lead to anomalies in expected behavior. However, the underlying mechanisms are different. The expectation of increasing monetary incentives is that it will lead to greater motivation and effort (Ariely, Gneezy, Loewenstein, & Mazar, 2009). This generally hold true. However, the increased stress associated with the large financial incentive can lead to poorer performance, a phenomenon known as choking may also be seen (Kamenica, 2012). The mechanism is discussed in more detail below. A commonly encountered example of this occurs while playing sports. A study of professional basketball players in Australia analyzed free-throw attempts. It found that the players’ performance was poorer during games than during practice (Dandy, Brewer, & Tottman, 2001). This is unsurprising to anyone who participates in sports. With the increased pressure and higher stakes comes a greater likelihood to become stressed and choke. When the offered reward (or penalty) is perceived as too little, we see a different response. When offering only a small incentive, individuals can perceive the task as not worth the effort or the reward as demeaning. A study of university students in the U.S. found that the likelihood of completing a survey diminished when offered a reward of fifty cents to complete the task. Students were, unsurprisingly, more likely to complete the task for a larger monetary reward, but also when no reward at all was offered (Heyman & Ariely, 2004). In a study of Israeli day-care centers, Gneezy and Rustichini found that after a fine was imposed for parents who picked up their children late, the number of late parents increased. In this case, the parents viewed paying the fine as absolving themselves of the social more of being on time (Gneezy & Rustichini, 2000a). The penalty for not picking up their children was not high enough to incentivize parents to always be on time and in fact had the opposite effect.
III. Mechanisms
In a 2012 article exploring incentive anomalies, Kamenica et al posits “…that [anomalous impacts of incentives], rather than being a disconnected string of idiosyncratic exceptions to the standard model, these findings constitute convergent evidence about a coherent set of principles that can help improve the design of incentive structures in a variety of settings.” The paper identified mechanisms to explain and help predict with incentive anomalies will happen including ‘contextual inference’ and choking (Kamenica, 2012). Contextual inference is based upon the assumption that people are imperfectly informed about their choices and use the environment to help guide their decisions. As illustrated in the previously discussed examples, individuals use incentives to gain context about the scenario. choking explains why excessive monetary reward can result in subpar performance. In these cases, the addition of the large incentive changes the perception of the individual regarding the importance and difficulty of the task. Of note, this definition of choking differs from psychological definition where choking occurs in “inherently stressful” situations. In contrast, economic choking occurs because the incentive has made the task stressful (Kamenica, 2012).
IV. Conclusion
Recognizing when incentives are likely to fail is crucial to structuring incentives. Only with this understanding of anomalies can previously implemented schemes be corrected (J. Costa-Font, 2011). The consequences of not understanding incentive anomalies extend beyond implementing ineffective interventions and programs. Titmuss argued that the United Kingdom’s system was not only more effective in collecting blood but was also collecting medically superior blood. He found that by paying for blood donations, the makeup of the donors changed. A controlled longitudinal study of patients in the cardiac unit of an American hospital found that the risk of developing Hepatitis B was greater than 50% amongst patients had received a blood transfusion from paid donor recipient. By contrast, no patients who received freely-donated blood became infected with Hepatitis B (Titmuss, 1970). To counteract these failures in a systematic way, Kamenica proposes different mechanisms by which incentives act in an unexpected fashion. With an understanding of these mechanisms, we see that seemingly irrational behavior is in fact rational and these incentive anomalies do not disprove standard economic theory, rather they support it.