Widura, Robert (2009). Commission-Based Pay -- Three Essays on German Insurance Tied Agents. PhD thesis, Universität zu Köln.


Download (1MB)


My thesis consists of three main building blocks. The first of these is the relationship between incentives, allocation of effort, and task performance. Second, I will analyze how person variables, i.e., different forms of labor-market experience, influence the agent's performance. Finally, I discuss the effect of compensation on agent satisfaction, using a model of expectancies and reference-point dependent utility. The data is on German insurance tied agents and was collected in the first half of 2006. It is based on an online questionnaire among German insurance companies. A total of 9,386 insurance tied agents from 10 large German insurance companies participated in the survey. The first part of my thesis focuses on the question of how incentives can influence agents' time allocation and whether they can lead to a change in performance. Many recent studies have suggested that incentives matter, and that piece-rates are the best way to endorse the desired behavior. Empirical evidence that supports this hypothesis, however, is scarce. Alongside the problem of if and how incentives increase the amount of effort exerted by the agent, there is also the question of how the effort is allocated. A common hypothesis among practitioners is that there is always a trade-off between acquiring new customers and caring for existing ones. I use the concept of a multi-task model to describe this trade-off between selling and customer care. First, I develop a two-period multi-task model and derive hypotheses regarding the influence of effort allocation on performance and how commission rates can be used to steer this allocation. Second, I test the hypotheses using data on auto insurance sales of three German insurance companies. I find a significant positive effect of the time spent on customer care activities both on the portfolio size and new business production. The new business commission rate has a positive effect on selling and a negative effect on customer care time. The results for the recurring commission rate are mixed. In the second part of my thesis, I turn towards the influence of individual characteristics of each worker. The human capital theory is well-known and seeks to explain the relationship between experience and pay, stating that more experience makes the employee more valuable and should thus lead to higher wages. Previous research has cast some doubts on this simple relationship and suggested that the type of experience needs to be differentiated from the total sum of experience. I differentiate three categories of experience: industry-specific, job-specific and general labor-market experience. Based on the standard wage model, I test the influence of experience using the data on tied agents. I find both the expected strong effect of tenure on pay due to the special structure of agent compensation in this industry, and that general- and industry-specific labor market experience also have positive effects. The latter deviate is dependent on the agent's tenure: it is strong for agents with low tenure, but becomes rather weak or even negative for high-tenure agents. A detailed look into the effects on different compensation components reveals a high degree of diversity in the size and direction of the experience-related influence on wage. By applying an instrumental variable (IV) approach and introducing proxy variables for the job- and industry-match, the chapter also contributes to the discussion regarding the omitted-variable bias that can arise from using simple OLS instead of IV. Finally, the third part of my thesis discusses the motivational aspects of compensation. Recent publications found that the expected wage acts as a reference-point and that performance drops after that target is met. One of the most often analyzed phenomena in the context of reference points is "loss aversion": people usually feel that a loss reduces their utility far more than a comparable gain would increase it. Based on this line of thought, I present empirical evidence for the existence of a reference-point for satisfaction with compensation. Information on the satisfaction (both with pay, and overall) of the tied agents is used to test the hypotheses. An ordered probit regression and a subsequent analysis of the marginal effects is applied to investigate the influence of compensation on the satisfaction score. I find a relationship between the agents' deviation from peer-group compensation and their satisfaction with pay. Surprisingly, the analysis shows a positive relation with satisfaction for deviations above the peer-group median but no statistically significant relation for deviations below the median. These results run contrary to the common notion of loss aversion presented in other publications. In line with previous work, no relation is found between compensation and overall satisfaction.

Item Type: Thesis (PhD thesis)
Translated title:
Provisions-basierte Vergütung -- Drei Essays zu gebundenen Versicherungsagenten in DeutschlandGerman
Widura, Robertrobert.widura@gmail.comUNSPECIFIED
URN: urn:nbn:de:hbz:38-29330
Subjects: Management and auxiliary services
Uncontrolled Keywords:
Versicherungsagenten , Provisionssätze , Zeitverteilung , Zufriedenheit , ReferenzpunktGerman
Insuranceagents , provision-rates , effort-allocation , pay-satisfaction , reference-pointEnglish
Faculty: Faculty of Management, Economy and Social Sciences
Divisions: Faculty of Management, Economy and Social Sciences > Seminar of Personnel Economics and Human Resource Management
Language: English
Date: 2009
Date of oral exam: 16 November 2009
NameAcademic Title
Sliwka, DirkProf. Dr.
Refereed: Yes
URI: http://kups.ub.uni-koeln.de/id/eprint/2933


Downloads per month over past year


Actions (login required)

View Item View Item