Reducing Absenteeism in the Workplace

We collaborate with a retail chain. Management in one region considered to implement a monetary attendance incentive, following the example of a comparable incentive scheme already in use in a different region. We offered to evaluate this instrument and proposed a variation that is not only novel to the academic literature, but may also reduce possible adverse effects associated with monetary incentives. The intervention takes place in a region. With the exception of final year apprentices, all of the firm's apprentices in the respective region participate in the experiment. The firm distinguishes between two different store types. For administrative reasons expressed by the firm, stores of type II are not eligible for one of the treatment conditions (treatment Time).

2018-01-01

2018-12-31

Mean number of days absent due to sickness

The mean number of days absent due to sickness is standardized according to actual exposure.

Mean number of single day absences due to sickness
Mean length of absence spells

The mean number of single day absences is standardized according to actual exposure.

There are three different treatment conditions, henceforth referred to as Time, Money and the control group. All apprentices receive a personal letter sent to their home address from the firm’s HR office. For apprentices in the control group, essentially nothing changes during the treatment. They are only made aware of a project called “Attendance Bonus” and that this project will be relevant to them at some later stage. Apprentices in the remaining two treatment groups are notified that they receive a “bonus point” for each month without a single day of absence due to sickness. The total duration of the experiment is twelve months. After every three months in the treatment period, the incentivized apprentices receive feedback on their current point score. However, bonus points are only transformed into actual rewards after the end of the experiment. Importantly, only the total number of accumulated points is payoff-relevant. The reward function is identical for both the Money and the Time condition. However, the two treatments differ with respect to the nature of the employed reward domain.

Time
In the Time treatment, one reward unit corresponds to an extra holiday. A maximum number of four extra holidays can be achieved. The respective apprentices are granted their extra holidays two months after the end of the treatment period. From then on, the apprentice is free to choose when to make use of the additional holidays. Only store type I apprentices are eligible for the Time treatment for administrative reasons.

Money
In the Money treatment, one reward unit corresponds to a bonus payment of €60. That is, a maximum bonus payment of €240 can be achieved. This conversion rate is chosen such that the values of a single reward unit are approximately equivalent for both reward domains. The respective apprentices receive their bonus payments in addition to their regular salary two months after the end of the treatment period.

Control
Apprentices in the control group are not incentivized. Yet, they receive a notification prior to the beginning of the treatment period vaguely indicating that randomly chosen groups of apprentices will receive some sort of attendance contingent reward. Furthermore, they are notified that they will receive a comparable reward at a later point in time. In the context of the retail chain, informing the control group is inevitable, since apprentices in different treatment groups may yet attend the same vocational training school and thus exchange information. Therefore, informing those apprentices who are not incentivized that they are not systematically disadvantaged may prevent them from alternating their (attendance) behavior as an act of defiance.

Stratified Randomization

Store
Treatments were assigned seperately for type I and type II stores, albeit following the same procedure.

Yes

315 Stores:
151 Type I Stores
164 Type II Stores

542 Apprentices: 268 Type I Store Apprentices 274 Type II Store Apprentices

Time: 90
Money: 234
Control: 218

Monetary incentives are widely used to align employee actions with employer objectives. We conducted a field experiment in a retail chain to evaluate whether an attendance bonus could reduce employee absenteeism. Apprentices in 232 stores were randomly assigned to a control group or one of two treatment groups in which a monetary or time-off attendance bonus was introduced for one year. We find that neither variant of the attendance bonus led to a systematic reduction in absenteeism. On the contrary, the monetary attendance bonus increased absenteeism substantially by around 50% on average, which corresponds to more than five additional days absent per employee and year. This effect was driven by the most recently hired apprentices. Survey results reveal that the monetary attendance bonus shifted the perception of absenteeism as acceptable behavior. The backfiring effect persisted beyond the end of the experiment, indicating a lasting erosion of social norms.

Jakob Alfitian, Dirk Sliwka, Timo Vogelsang (2023) When Bonuses Backfire: Evidence from the Workplace. Management Science 0(0).

https://doi.org/10.1287/mnsc.2022.00484