Go Hard or Go Home? Analyzing the Effects of Working from Home on Productivity and Promotion

The theoretical and empirical literature on the effect of working from home on workers' performance is ambiguous. Some studies show that working from home has a positive effect on workers' performance. According to these studies this positive effect originates from a quieter work environment and higher motivation. Other studies argue that working from home can have a negative effect on workers' performance and explain that shirking is the main driving force of this effect. By giving workers appropriate incentives, i.e. implementing performance pay, firms can eliminate the shirking problem. In such an environment working from home has a positive effect on workers' performance. Thus, the joint application of working from home and performance pay seems to be an attractive practice to improve workers' performance. However, the success of working from home does not only depend on the firm that offers it, but also on workers who decide to either work from home or stay in the office. Workers individual utility maximization drives this decision with all its consequences for the firm. Various factors influence the decision to work from home such as reduction of commuting costs, improvements in work-life balance and career prospects. While cost savings and work-life balance have been frequently associated as main drivers of the decision to work from home, results for career prospects are scarce. Some studies show that workers who work from home are less often promoted, as they have reduced social interaction. Analyzing workers strategic decision to self-select into working from home has important implications for firms' choice to offer working from home. If working from home increases workers' performance in the short-run but leads to suboptimal promotions, the firm might face the problem of inadequately qualified managers in the long-run. In this paper we analyze workers self-selection into working from home and its consequences for firms' promotion decisions. Our contribution is twofold. First, we develop a decision model that explains the sorting of workers into working locations conditional on ability and promotion probability. Second, we estimate the decision to self-select into working from home using data from a randomized field experiment. Workers are initially randomly assigned to two locations: working from home and working in the office. Thus, effects on productivity are solely driven by location. Afterward workers can decide whether they want to work from home or in their office. Therefore, sorting is affected by workers' learning on their own productivity and their promotion probability in these two locations. Results show that working from home improves workers' productivity over the entire productivity distribution. If workers can freely decide where they want to work, high performing workers sort into working from home, as this location maximizes their individual utility. For high performing workers the utility gain from an increased productivity at home is higher than the expected utility loss through the lower promotion probability at home. In contrast, low performing workers self-select into working in the office.