Economists studying remote or hybrid work typically think of the time spent working remotely as a “treatment” on employees and use data to study the outcome (or impact) of this on employees and employers, via either an “observational” approach or an “experimental” one.
In the two previous issues of Labor & Employment Literature Watch we examined observational studies, in which the data analyzed was collected ex post, i.e., employees or employers decided (jointly or individually) whether or not to engage in remote work. Economists use the observational data to study outcomes either by comparing the outcomes of remote workers to their own outcomes before treatment (i.e., before they were working remotely), or by comparing the outcomes of remote workers to the outcomes of similar, non-treated employees (i.e., to employees who chose not to work remotely).
The results of observational studies are prone to “selection bias” because employees who opt to work remotely may be systematically different from employees who choose not to work remotely. By contrast, the studies we discuss in this Labor & Employment Literature Watch looks at an experimental approach, in which economists randomly assign treatment (i.e., working remotely) to employees. The experimental approach eliminates the possibility of selection bias because employees working remotely have not self-selected into remote work. As a result, these studies provide strong causal evidence on policy outcomes.
This Labor & Employment Literature Watch provides the final in a three-part series reviewing recent economic research on remote and hybrid work. Each part concludes with tips for employers to consider when assessing policies for remote and hybrid work.
