Absence of conflict of interest.
- The study's objective was to examine the impact of teleworking on job turnover rates. The other chapters of the study are linked here:
The study applied an instrumental variable design to estimate the impact of telework on employee turnover rates, using passage of the 2010 Telework Enhancement Act as an instrument for access to telework for federal employees. The Telework Enhancement Act required federal executive agencies to provide a telework policy, but did not affect non-executive federal agencies. Using federal administrative data on teleworking and turnover rates, the author compared outcomes for federal employees who worked in agencies that were and were not required to adopt a telework policy under the Act.
The study suggested that access to teleworking decreased employee turnover.
This study receives a low evidence rating because the author reported agency-level results but did not provide information on changes in the composition of employees working in agencies that were and were not affected by the Act. This means we are not confident that the estimated effects on turnover are attributable only to teleworking; other factors may have contributed.
2010 Telework Enhancement Act
Features of the Intervention
The 2010 Telework Enhancement Act was implemented with the aim of expanding telework among federal employees. Under the Act, covered agencies were required to establish a telework policy, assign a telework managing officer, identify eligible employees, and inform those eligible of their telework option. The Act applied to executive federal agencies only (those headed by a Cabinet secretary or the Office of the President); non-executive federal agencies (e.g., the Federal Reserve or Securities and Exchange Commission) were not covered but could choose to offer teleworking to their employees.
Features of the Study
The study used an instrumental variable design to estimate the impact of teleworking on a federal agency's turnover rate. The instrumental variable was the passage of the 2010 Telework Enhancement Act, which required executive federal agencies to implement a telework policy if they did not already offer one. The Act did not apply to non-executive federal agencies, although a substantial number voluntarily offered a telework policy, and that number increased after the Act was passed. The treatment group included 411 federal executive agencies (those required to adopt a telework policy) and the control group included 221 federal non-executive agencies (those not required to adopt a telework policy).
The study relied on agency-level data from the Office of Personnel Management and individual-level data from Fedscope. The former provided annual counts of total employees, employees eligible for telework, and employees who telework. The latter provided information on employee separations, and demographics and length of service. On average, approximately 25 to 30 percent of federal employees in these agencies had more than a bachelor's degree and about 45 percent were female; employees in the treatment group were slightly younger and more educated. A statistical model was conducted to compare annual turnover rates for federal agencies that were required to adopt a telework policy under the Telework Enhancement Act to those that were not.
Comparing federal agencies that were and were not required to offer a teleworking policy under the Telework Enhancement Act, the study suggested that access to telework lowers turnover rates.
Considerations for Interpreting the Findings
The author compared turnover measured at the agency level, not at the individual level, but did not provide information necessary to assess changes in the composition of employees working for the agencies in the treatment group versus the control group. Differential changes in the demographic characteristics of the employee groups could influence the difference in turnover rates between the treatment and control groups, and employees might leave a control-group agency to work at a treatment-group agency in order to access telework. Without being able to confirm that differential migration in and out of the treatment and control groups was within allowed limits, we cannot be certain that the estimate reflects only the impact of access to telework.
Causal Evidence Rating
The quality of causal evidence presented in this report is low because the author did not confirm that the level of migration in and out of the population of employees in the treatment and control agencies was within allowed limits. This means we are not confident that the estimated effects are attributable to telework; other factors may have contributed.