A token incentive to encourage a behavioral change.
While financial incentives, as an intervention, are typically associated with neo-classical economic theory, there are cases in which financial incentives can be “behavioral.” When the incentive is small compared with the behavior change it causes—and when it is not enough money to relieve an actual financial constraint or market failure—we classify it as behaviorally motivated. Imagine, for example, a friend who drives to a store all the way across town just to redeem a one dollar coupon on a grocery item. In this case, the experience of receiving an incentive may be more important and relevant than the financial value of the reward itself. This is sometimes the case with conditional cash transfers, but there is no hard line for determining when a CCT is “behavioral,” and when it is not.
Simply the prospect of earning a reward can improve health-related behaviors. In contrast to the case described above of very small, but certain rewards, rewards with very small expected values due to low probabilities may change behavior if the maximum reward is high enough.
- For example, in the US, patients prescribed an anti-stroke medication were offered a lottery ticket as a reward for taking their pills (Volpp et al., 2008). The lottery offered a very low-probability opportunity to win a small relatively sum of money—the largest prize was $100, and each individual had just a 0.1% chance of winning. However, it succeeded in virtually eliminating non-adherence to drug treatment.
- A similarly small incentive in India—a bag of lentils, equivalent to a half-day’s wages for an agricultural laborer—almost doubled the fraction of women bringing their children to a vaccine camp for immunizations (Banerjee et al., 2013).
Incentives can also overcome behavioral biases (including status quo bias and present bias) in the context of RH providers. Potential interventions include:
- ‘Pay-for-performance’ contracting mechanisms, which provide modest institutional and provider incentives to improve the quality of care (Basinga et al., 2011; Lagarde, Haines & Palmer, 2007; Eijkenaar et al., 2013; Miller & Babiarz, 2013); and
- Vouchers for services distributed to targeted populations, especially women (Bellows, Bellows & Warren, 2011; Warren et al., 2011). The vouchers’ financial value may be modest, but they can incentivize providers to offer the vouchered services, streamline access to services, and increase women’s confidence in their right to services (Meyer et al., 2011).
The above incentives were not necessarily small enough to be classified as behavioral, but they do offer evidence that incentives can change behavior. Further research would be required to disentangle the behavioral effect of receiving the award from the financial benefit.
One caveat in the use of financial incentives is that dependent on context and size, they have the potential to crowd-out a health provider’s intrinsic motivation with extrinsic motivations, like money (Serra, Serneels & Barr, 2011), although they can also complement providers’ intrinsic motivation and act as a signal regarding expected social norms (Bowles & Polania-Reyes, 2012). Experiments have demonstrated that gifts-in-kind can increase productivity significantly more than monetary gifts; this result is driven by positive reciprocity (Kube, Marechal & Puppe, 2008), or a social preference like reciprocal fairness. This might explain why a randomized experiment in Zambia, aimed at increasing community-based provision of services, found non-cash incentives for health extension workers to be more effective than cash (although a combination of the two approaches was found to be even more effective; see Ashraf et al., 2013).
Providers can also be positively motivated by social norms, as health worker motivation is at least in part driven by how much social recognition they receive from their communities (Franco et al., 2004). An evaluation of BRAC in Bangladesh found that social prestige or recognition is a key determinant in retaining community health workers (Alam & Oliveras, 2014).