Experimental Design Details
We examine the consequences of reducing reporting paperwork, that is, the forms and protocols the top management of a firm requires employees and team managers to fill in as part of the reporting process. The effects of reporting on performance are theoretically ambiguous: while reporting helps deal with agency issues, it costs valuable time and is not conducive to building trust between the firm and its employees. We therefore approach our research question empirically, using an RCT as a means of creating exogenous variation in exposure to reporting requirements.
Setting. Our study will take place in a family-owned bakery chain in Germany that operates 148 bakeries employing 1,300 people. In a way similar to many other firms in the food sector, the firm has a hierarchical structure: 148 shop managers report to 16 regional managers, who report to three sales directors, who report to the executive board. There are many reporting duties and paperwork imposed on bakery employees in addition to their core activities (sales). Evidence from a pre-treatment employee survey hints at dissatisfaction with excessive reporting paperwork, and suggests a link between it and turnover intentions. The survey results prompted the firm’s executives to contemplate possibilities of reducing the burden of reporting for the employees.
Treatment. We will remove two elements of reporting paperwork, described below, in 73 randomly selected bakeries, the remaining seventy-five operating as usual.
Our proposed treatment is informed by a thorough study of the reporting processes in our study firm. In Summer 2020, together with the firm’s management and the head of the workers’ council, we surveyed dozens of pieces of reporting paperwork in the bakeries. Subsequently, we visited twentey randomly selected bakeries and asked their managers and employees to evaluate each piece in terms of its informational and monitoring benefits to the firm and time costs to the workers. The costs relative to the benefits were particularly high for two pieces:
i. The form in which employees document whether they smiled when interacting with customers, put products in the required order on the counter, and followed marketing campaigns (to be filled several times per day).
ii. The daily protocol in which the store manager and employees document their daily activities.
According to the firm’s top management, the above two pieces of reporting paperwork were initially implemented to monitor essential processes. Our survey evidence, however, suggests that they are excessively costly and signal firm’s distrust to the workers.
We performed stratified randomization by shop location and size. The control and treatment shops are balanced across a number of dimensions including performance outcomes we expect to be affected by our treatment, as outlined below.
Expected outcomes. We will monitor a number of behavioural and performance outcomes that could be affected by the reduction of paperwork. First, we expect changes in shop managers’ time use, from reporting to sales activities, and possibly changes in leadership styles and in the interaction between store managers and employees. Next, time freed up from reporting and spent on sales activities could improve sales, especially in the hours when reporting paperwork is usually done (we measure this information in pre-RCT surveys). On the other hand, less compliance with technological and service procedures might reduce sales, either directly or through lower customer satisfaction.
Less work and stress from less reporting could reduce employee turnover and absence, as well as improve employees’ attitudes towards the firm, especially their perceptions of being trusted by the firm. We will also examine the effect on physical appearance and customer service standards as measured in mystery shopping scores.
Heterogeneous treatment effects. We will measure how treatment effects on the above outcomes vary along several dimensions, as we now outline. For initial intuition on how the effect may vary, as well as to control for possible partiality towards specific shops, we conducted interviews with regional managers where we asked them to predict in which shops our treatment will and will not work, and to explain why they thought so. The results of the interviews suggest that regional managers expect treatment to work positively in small teams, where there is less need for formal coordination and reporting, and in longer-established teams since they have better knowledge of the work routines. Accordingly, we will measure treatment effect heterogeneity with respect to shop team size and tenure, as well as regional managers’ expectations we learned from the interviews.
Additionally, we will explore variations in the treatment effects with survey measures of employee reciprocity, as more reciprocal individuals may react to the treatment more strongly, and the identity of the sales director, since not all sales directors were equally optimistic regarding our proposed treatment.