Nava Ashraf, Oriana Bandiera, Virginia Minni, Luigi Zingales
9 min
This paper evaluates a worker-centered intervention called Discover Your Purpose (DYP), which helps employees reflect on their life purpose and align it with their professional activities. Moving away from traditional monetary incentives, the firm offered readings, writing exercises, and workshops to help staff understand what makes their effort personally meaningful. The study uses a randomized controlled trial (RCT) across 2,976 white-collar employees in 14 countries, tracking outcomes over two years.
The intervention significantly increases the firm's overall performance. This occurs through two distinct channels: first, it encourages low performers to leave the firm as they realize their current job does not align with their purpose (selection); second, it improves the motivation and effort of the employees who remain (treatment). Unlike traditional performance pay, which often benefits only top performers and increases inequality, DYP boosts performance from the bottom up. Treated employees also show higher job satisfaction and report better alignment with both their colleagues and the firm.
This research challenges the standard principal-agent model, which typically assumes that worker effort is inherently costly and must be extracted through financial rewards. The authors show that when employees are empowered to define their own meaning, they can find greater utility in their work, which in turn benefits the firm. Crucially, the authors demonstrate that standard observational data often hide these benefits because they are confounded by self-selection: top performers are more likely to participate in such programs voluntarily, masking the fact that the intervention is actually most beneficial for the lower tail of the performance distribution.
Sam: And someone could conclude that neither their duties nor possible changes fit them. That makes leaving part of the proposed mechanism, rather than necessarily a failure.
Alex: The program explicitly asks whether employees can live their purpose in their current job. It doesn’t prescribe particular tasks. In the authors’ model, the ability to recognize meaning belongs to the worker, so it also changes how they assess alternatives outside the firm.
Sam: Let’s separate that model from what they established. How did the experiment make participation informative rather than just identifying people who already liked reflective workshops?
Alex: They randomized invitation emails within each of fourteen countries. Employees in the comparison group would be invited after the study ended. Participation stayed voluntary, and most invitees accepted. The researchers followed personnel records, performance assessments, pay, and surveys over two years.
Sam: If attendance remained voluntary, comparing attendees with non-attendees would still mix the workshop with personal differences. Which comparison carries the causal claim?
Alex: The cleanest comparison is between employees randomly invited and those not invited. That estimates the effect of offering the program. The authors also use the invitation to estimate participation effects for employees whose attendance depended on receiving it. Those estimates aren’t effects for every possible worker.
Sam: What happened to departures first? That’s the outcome an employer might initially dislike.
Alex: The annual exit rate rose from about thirteen percent in the comparison group to sixteen percent among those invited. The gap emerged soon after invitations and remained stable. That pattern suggests the program changed some employees’ decisions, rather than merely bringing forward departures that would soon happen anyway.
Sam: Then the performance result needs unpacking. If weaker employees leave, the remaining workforce looks better even without anyone changing how they work.
Alex: The authors explicitly separate those channels. They use each employee’s pre-experiment performance to estimate how departures changed workforce composition. Then they compare that with the total performance change. By their decomposition, roughly half the reduction in below-standard performance came from departures, and half from improvement.
Sam: How large was that reduction relative to the problem the firm started with?
Alex: About ten percent of comparison employees initially performed below the expected standard. Receiving an invitation reduced that share by roughly a quarter. So the gain was concentrated at the bottom, not in turning already strong employees into even stronger performers.
Sam: What counts as performance here? I’d trust an independent sales measure differently from a manager’s impression.
Alex: The main measure is an annual manager assessment informed by colleagues’ feedback. The firm standardizes it across functions and offices, and uses it to determine bonuses. The authors also show that these scores correlate positively with sales performance where sales data are available. But the main result remains based on assessments, not a universal direct-output measure.
Sam: Did employees share in the gains, or did higher scores mainly benefit the employer?
Alex: More employees received performance bonuses, and bonus amounts increased. Some of that reflected low performers leaving; some reflected better performance among those staying. Fixed pay did not change. Survey respondents also reported higher meaning, job satisfaction, and happiness, although fewer than half answered the endline survey.
Sam: That makes the administrative outcomes more convincing than a claim that the workshop definitively reduced the psychological cost of effort.
Alex: That’s how I’d read it. The reduced cost of effort is the model’s interpretation, supported by the outcome pattern and surveys, not directly measured. The surveys also show narrower gender gaps in stated job priorities. Administrative records suggest invited men were more likely to take parental leave.
Sam: The firm still loses working time and pays to replace people. Does the economic case survive those costs?
Alex: Under the authors’ assumptions, lasting performance gains make the difference. Their estimated internal rate of return—the discount rate at which benefits equal costs—is seventy-two percent if gains last through the follow-up period. But that’s a calculation, not an observed financial return. It translates performance scores into productivity and includes replacement costs and compensation changes.
Sam: And this company already ran the program internally. A newly purchased consultant workshop isn’t necessarily the same proposition.
Alex: The authors’ consultant-cost scenario gives much lower returns. More broadly, this was an established program in one multinational, among entry-level white-collar workers. The authors argue that trust matters: employees must believe personal disclosures won’t be misused, or improved alignment exploited to cut their pay.
Sam: There’s also an adoption puzzle. If the program helps, why wouldn’t the firm’s earlier experience have made the benefit obvious?
Alex: Earlier voluntary participants tended to be stronger performers and less likely to leave. Those differences obscured the causal pattern: departures among weaker performers and improvements among those who stayed. A simple attendee comparison could therefore miss what made the program useful.
Sam: Who should read the full document, and where should they start?
Alex: Researchers studying labor, management, or workplace well-being should start with Worker Performance, especially the separation of departures from improvement. Then read Understanding Adoption for the experimental-versus-observational comparison. Anyone using the financial estimates should inspect the cost-benefit assumptions. For everyone else: better workplace performance can mean helping people improve—or helping them recognize they should leave.
Sam: And those are different outcomes worth keeping separate.