The provision of public services—governance, health care, education—depends critically on the effort of the workers tasked with providing them. Effort can be increased in two ways: through incentives that directly reward it and through the selection of workers who are predisposed towards exerting it—that is, who have high intrinsic motivation. These two mechanisms are linked because incentives may affect not only effort on the job, but also who self-selects into a job. The effect of incentives on self-selection into public service jobs remains largely unexamined.
In 2010, the Government of Zambia launched a national effort to create a new civil service position: the Community Health Assistant (CHA). The government aims to train 5,000 new CHAs by 2017—a massive investment in a country with only 6,000 nurses. CHAs undergo a year of formal training, and then return to their rural home communities to work. The majority of their work consists of household visits, but they also spend one day a week in the community health post and organize community health-education meetings. They are the first line of healthcare for Zambians living in the most remote regions of the country.
Researchers tested how private incentives, in the form of opportunities for career advancement, affect the skills and motivation of applicants for a new health worker position and how, in turn, this self-selection affects job performance.In the first year of the program, the Ministry of Health aimed to recruit 330 CHAs, two from each of 165 rural health posts in 48 districts of Zambia. In a randomly selected half of the districts, recruitment posters emphasized the “social” benefits of becoming a CHA, such as serving and being a leader in one’s community. In the other half, recruitment materials emphasized the “career” benefits of becoming a CHA, such as opportunities for promotion and further professional development.
Other than the differing emphasis on social or career benefits, the recruitment materials and selection processes were identical. They specified that applicants needed to be Zambian nationals, ages 18-45, with a high school diploma and passing grades for at least two subjects on their secondary school graduation exams. The government received over 2,400 applications and interviewed 1,585 eligible candidates. Three hundred and fourteen attended a year-long training, and 307 graduated and began work as CHAs. Once deployed, actual benefits were identical between the two treatment groups—that is, all CHAs had the same job description (and thus the same social benefits) and the same private incentives, including the possibility of promotion after two years of service. As a result, any difference in performance was due to the selection effect of the incentives.
Career incentives attracted CHAs that were more qualified and had the same level of pro-social preferences as CHAs recruited by making social incentives salient. These CHAs consequently performed better on the job. By highlighting opportunities for career advancement, governments may be able to recruit better skilled and better motivated applicants to work in public service.
Pro-social and career preferences: Applicants recruited under both approaches displayed similar levels of pro-social preferences: about 84 percent of both groups perceived their self-interest as overlapping with their community’s interests, and a little over half planned to stay in the same community for the next 5-10 years. However, applicants recruited by making career incentives salientwere more ambitious regarding career advancement: a larger portion aspired to hold a higher-ranking government position in the next 5-10 years.
Ability: Applicants recruited by making career incentives salient were 6 percentage points more likely to qualify for university admission (compared with 71 percent of applicants recruited by making social incentives salient). They had higher secondary school graduation exam scores overall and in natural sciences.
Performance:During the first 18 months of work, CHAs recruited by making career incentives salient conducted 94 more household visits than those recruited in the social incentives group, who conducted an average of 319 visits. They did not achieve these gains by targeting easy-to-reach households or by spending less time on each visit. CHAs in the career benefits group also hosted more than twice as many community meetings (an average of 38 meetings, compared with 17 meetings in the community benefits group). The effect on performance was driven by high-performing CHAs in the career incentives group—workers who would not have been recruited if they had not seen materials emphasizing opportunities for career advancement.
Retention: There was no evidence that career incentives improved performance at the expense of retention. During the first six months after the CHAs’ one-year commitment period, the number of dropouts was similar across both groups. None of the CHAs who left did so for a higher-ranking position within the Ministry of Health because the Ministry requires CHAs to work for two years before applying for promotion.
