Firms in developing countries find it hard to recruit adequately skilled workers, and to motivate and retain them (World Bank, 2013; IGC, 2014). Yet, recent evidence suggests that many firms do not use standard financial rewards to attract high quality workers and increase their performance (Bloom et al., 2010; Fafchamps and Woodruff, 2014). Are financial rewards too costly to be profitable? Or are managers setting financial incentives inefficiently?
Inefficient incentives at the firm level can contribute to the low productivity challenge facing policy makers in several developing countries around the world. For example, when workers are not properly incentivised to apply for the positions where the returns to their skills are highest, labour can be misallocated and productivity can fall.
We have designed a field experiment in Addis Ababa, Ethiopia to address these questions. The experiment has two connected components. In a first “selection experiment”, we will measure the productivity of the workers that apply for a real position under different wage offers. In a subsequent “hiring experiment”, we will ask real firm managers to make a simulated, incentivised decision. The manager has to choose what wage to “offer” for the position we have advertised. If she offers a high wage, she will be assigned to a worker from the pool of applicants in the "selection experiment" who responded to the high wage offer. If she offers a low wage, she will be assigned to a worker who responded to the low wage offer. Managers will earn a real payoff that increases with the productivity of the candidate they have been assigned to, and decreases with the wage. To maximise this payoff, managers need to draw upon their knowledge of the labour market. In particular, they need to consider whether the increase in the productivity of the candidates attracted at higher wages is greater than the pay rise.
We will investigate whether managers choose the effective level of financial incentives in the hiring experiment and, if they fail to do so, whether this is due to inaccurate expectations about the productivity of the workers selected at higher wages. Building on a panel dataset of firms that we have created in the last two years, we will be able to relate our experimental measure of managerial efficiency to the level of competition in the sector where the firm operates.
Furthermore, the “selection experiment” will allow us to quantify the extent of the skill shortage that firms face. While existing evidence on this issue is mostly based on qualitative reports, we will estimate the monetary cost that firms must pay in order to attract better candidates. This is of great interest in a fast-growing country like Ethiopia, where human capital constraints may stifle the performance of expanding sectors. Finally, we plan to use the results of the “selection experiment” to structurally estimate the relative magnitude of the application costs faced by different groups of jobseekers (for example, men and women, high and low socioeconomic background, etc…) and the distortions that these costs may create to the process of matching between workers and firms.