A Biased Evaluation of Employees’ Performance Can Be Useful for Employers

© iStock

In assessing an employee’s performance, employers often listen to his immediate supervisor or colleagues, and these opinions can be highly subjective. Sergey Stepanov, an economist from HSE University, has shown that biased evaluations can actually benefit employers. An article substantiating this finding was published in the Journal of Economic Behavior and Organization.

The model described in the article ‘Biased Performance Evaluation in A Model of Career Concerns: Incentives versus Ex-Post Optimality’ was developed within the ‘career concerns’ framework pioneered by Bengt Holmström. His paper represents the relationship between an employee (often called an agent by economists) and an employer, or principal (broadly speaking, this can be the market as a whole). This modelling considers three components of performance: talent, effort and random factors. An agent’s incentive to exert effort arises from the fact that better performance results in a higher evaluation of the agent’s talent by the market, which, in turn, can help to increase his future wage.

In the canonical model, an employer (or the market) observes the results of an agent’s work. Sergey Stepanov, Assistant Professor of HSE University’s Faculty of Economic Sciences, modified the model by adding an intermediate party – an evaluator. If the principal is busy or has many employees, it would be difficult for her to monitor each agent individually, and thus she will often rely on the evaluation of an agent by his supervisor or peers. For a variety of reasons, their assessments are likely to be biased, either in favour of the agent or against. With this in mind, the question the researcher sought to answer in this study was: ‘what should the best direction and degree of the bias be?’

Sergey Stepanov
Assistant Professor of HSE University’s Faculty of Economic Sciences

In classic career concerns models, the principal observes the performance of an agent directly. However, we know that this is often not the case, and principals receive such information through ‘evaluators’. However, the interests of these people may not coincide with those of the principal. And I thought: maybe it’s actually a good thing that they don’t? Objective evaluation is, of course, optimal from the point of view of making correct decisions about an agent (e.g., to promote him or not), but such an evaluation may create sub-optimal incentives to exert effort.

Agents who are very talented a priori will lose motivation if they are evaluated fairly, because they know they will most likely clear the performance bar even with a low effort. Similarly, agents who are initially believed to be below average will lose motivation because they are unlikely to succeed even with a high effort. Hence, an ideal evaluator should be stricter on employees who seem to be capable and talented, but more lenient towards those who are less capable. In addition, the greater the degree of career concerns of an agent, the less objective the optimal evaluator should be, while the performance of those whose abilities are initially very uncertain, for example, without a prior track record, should be judged most objectively.

Thus, the ‘unfair’ opinion of an evaluator may prove to be more useful in motivating an employee than an objective assessment.

The model may be useful, for example, for organizing internships. This proves that stronger interns with good CVs should indeed be given more demanding supervisors, whereas those for applicants with very brief CVs (which tell very little about their experience of skills) should be more balanced in their assessments.

The results of this research will be useful in evaluating the performance of government officials working on public projects or senior corporate managers, as well as in making internal promotion decisions.