Michelle Belot (Edinburgh)

Title: The Spillover effects of Monitoring: A Field Experiment

Abstract: We provide field experimental evidence of the effects of monitoring in a context where productivity is multi-dimensional and only one dimension is monitored and incentivized. We hire students to do a job for us. The job consists of identifying euro coins. We study the direct effects of monitoring and penalizing mistakes on work quality and evaluate spillovers on unmonitored dimensions of productivity (punctuality and theft). We find that monitoring improves work quality only if incentives are harsh, but substantially reduces punctuality irrespectively of the associated incentives. Monitoring does not affect theft, with ten percent of participants stealing overall. Our findings are supportive of a reciprocity mechanism, whereby workers retaliate for being distrusted.

Lea Cassar (U Zurich)

Title: Optimal Contracting with Endogenous Project Mission

Abstract: I present a model in which a principal selects one among many agents to develop a project and influences the agent's ex post level of effort not by outcome-contingent rewards, but by the choice of the project's mission. The closer the project's mission to the agent's preferred mission, the higher the agent's intrinsic benefit from exerting effort. The principal and the agents disagree on what the project's mission should be and the agents vary in how much they care about the project's mission, i.e. they have heterogeneous unobservable intrinsic motivation levels. I derive the optimal mechanism (allocation rule, project's mission, payment) to select and motivate the agent. I also consider situations where the project's mission must be chosen prior to the allocation of the project and where the agents face budget constraints. Several applications are discussed.

Erika Deserranno (LSE)

Title: Financial incentives as signals: experimental evidence from the recruitment of health workers

Abstract: Recent empirical research has documented the key role of financial incentives in determining who applies for a job. Theory suggests two channels through which financial incentives affect selection: a standard participation constraint channel when information is complete (the job attributes are know to potential candidates), and an additional signal channel when information is incomplete. I design an experiment that studies the effect of financial incentives in changing agent's perception of a task at the recruitment stage, and how this subsequently affects the size and composition of the application pool for a job. I do this by experimentally varying expected earnings during the recruitment of agents for a newly created health worker position in Uganda. I find that more lucrative positions are perceived as entailing lower positive externality for the community. While higher financial incentives help the recruiter attract more applicants and fill vacancies, this comes at the cost of displacing pro-social preferences in the pool of applicants and in the pool of appointed candidates. Over the course of two years, workers who exhibit pro-social preferences are found to be less likely to drop out and perform better.

Robert Dur (Erasmus)

Title: The Effect of Student Feedback to Teachers: Evidence from a Field Experiment

Abstract: We conducted a field experiment on the effects of student feedback to teachers at a large Dutch school for intermediate vocational education. Students evaluated all teachers, but only a randomly selected group of teachers received the feedback. Additionally, we asked all teachers before as well as after the experiment to assess their own performance on the same items. We find a precisely estimated zero average treatment effect of receiving feedback on feedback scores a year later. Only those teachers who highly overestimate their scores before the experiment improve significantly in response to receiving feedback. Lastly, we find that provision of feedback reduces the gap between teachers' self-assesment and students' assessment, but only to a limited extent.

Rema Hanna (Harvard)

Title: Dishonesty and Selection into Public Service

Abstract: Students who cheat on a simple laboratory task are more likely to prefer public sector jobs in India. We show that cheating on this task predicts corrupt behavior by bureaucrats, implying that it is a meaningful predictor of future corruption. Students who demonstrate pro-social preferences are less likely to prefer government jobs, while outcomes on an explicit game and attitudinal measures to measure corruption do not systematically predict job preferences. A screening process that chooses high ability applicants would not alter the average propensity for corruption. Our findings imply that differential selection into government may contribute, in part, to corruption.

Andreas Leibbrandt (U Monash)

Title: A Natural Field Experiment on Cheating and Sabotage in the Workplace

Abstract: We present a natural field experiment where we investigate the extent of cheating and sabotage in an actual job setting using different incentive contracts. We study day laborer who have been recruited to collect donations for a charity drive. During their shift they can cheat on their own performance, sabotage the performance of co-workers, and sabotage the employer.  Our findings show that relative-performance pay induces cheating to a larger extent than sabotaging co-workers. In addition, we observe that sabotaging the employer is rare and that cheating and sabotaging co-workers can be reduced if relative-performance pay is paired with a larger wage.

Philip Keefer (World Bank)

Title: Was Weber right?  The effects of pay for ability and pay for performance on intrinsic motivation, ability and effort in the public sector

Abstract: We examine the effects of pecuniary compensation on the ability and motivation of individuals who select into organizations that have non-pecuniary missions.  Specifically, we compare the effects of several common pay schemes on the productivity and mission motivation of individuals who select into a mission-oriented organization.  We compare flat pay systems, unrelated to ability or effort, to high-powered, pay for performance schemes and more traditional, “Weberian” schemes that calibrate pay to ability, independent of effort.  Flat pay systems attract significantly less pro-social (pro-mission) individuals into the mission organization.  However, pay for ability yields no loss of productivity nor pro-social motivation compared to pay for performance.

Joanne Reinhard (Behavioural Insights Team)

Title: Behavioural Economics and Employment - Evidence from 3 Field Experiments

Abstract: In this paper we present the findings of three field experiments that apply behavioural science interventions to the labour market. In the first of these, we conduct a pilot study in which benefit claimants are randomly assigned either to receive a package of interventions, including the use of commitment contracts and a systematic streamlining of their experience, as part of their benefit claim journey. Following this, we report the results of a much larger randomised controlled trial, implemented as a stepped wedge design, rolling out the same interventions across a much wider area. In both studies, we detect positive and significant impacts of the treatments, and this now forms a part of national UK policy. In our final experiment, we attempt to encourage participants to employ for additional jobs using messages from their job centre advisor. In this trial, we find a strong effect of messages designed to induce reciprocity, compared to a more standard text message.

Michael Sanders (Behavioural Insights Team, UK Cabinet Office)

Title: On the dynamics of "Behavioural" interventions: 5 experiments on habits, habituation, and social contagion

Abstract: Governments around the world are beginning to make use of the lessons of Behavioural Economics and Psychology in the process of policy design. There has been little study, however, on the dynamics, either temporally or socially, of these interventions when they not just used, but used repeatedly and in environments with complex social dynamics. What evidence there is, reviewed by Rogers & Frey (2014), suggests that interventions that are initially successful may become less so in time, either due to their novelty eroding, or due to their use being increasingly seen as manipulative. In this talk, we present the results of several experiments that begin to investigate these effects: Experiments 1 and 2 investigate the effect of a single successful intervention on future behaviours. In experiment 1 we report the results of the effect of randomly assigning some participants to be accepted to a large government run volunteering programme on both the decision to volunteer in future and behaviour in a dictator game. In experiment 2, we test the impact of a small, unconditional gift on charitable giving. A year later, we return and repeated the same experiment. In addition to replicating our primary findings, we are also able to establish the effect of the same treatment being repeated within individuals. Experiments 3, 4 and 5 investigate social contagion. Experiments 3 and 4 take place in the same environment as experiment 2, and participants are asked to make a donation to charity. Through secondary analysis of a successful intervention administered within clusters at the individual level, we are able to search for social contagion in charitable giving. Finding none, experiments 4 and 5 attempt explicitly to create social contagion. This is done through ³network nudges², in which influential people in a social network are intervened upon and the response of others in the network measured.

Rupert Sausgruber (WU Wien)

Title: Experimental Evidence on Motivational Sorting into Jobs

Abstract: We run an internet experiment (lab-style experiment with a large sample of people from all walks of life in Denmark) in which subjects choose between a job with steep incentives (high piece rate) or flat incentives (low piece rate). The job (i.e. the real effort task) is gender neutral, i.e. men and women perform equally well. We observe that high-performing women choose the flat scheme much more often than men and show that this is primarily driven by men being overconfident in their abilities and to a lesser extent to women being more risk averse. To investigate the external validity of these results we show that the experimental data predict labor market outcomes. In particular, we find that people choosing flat in the experiment are also more likely to choose jobs in the field that have lower income variance.

Patrick Warren (Clemson)

Title: Crime Lab Control and Criminal Justice Outcomes

Abstract: The relationship between forensic laboratories and the other institutions of law enforcement varies widely over space and time in the United States. Some jurisdictions have their own local lab within the police or sheriff's department, others depend on a statewide lab system either independent or under the state police, and others still contract with a private lab to process their forensic evidence. These different organizational forms may shift the incentives lab technicians and managers have to provide timely and accurate analysis and testimony. In this paper, I investigate the relationship between one particular institutional variant, local police/sheriff control of the crime lab, and one particular outcome in the criminal-justice chain: the conviction of innocent defendants due, in part, to faulty or misleading forensics. In a sample of the 200 largest counties in the U.S., counties with locally controlled labs have lower rates of exonerations in which faulty forensics were implicated in the original trial than similarly-situated counties without locally controlled labs with similar rates of non-forensic exonerations. This difference is robust to state fixed-effects, a variety of observable characteristics, and nearest-neighbor matching. The difference seems to be driven by control, per se, as matched counties containing state-controlled labs have significantly higher exoneration rates.