The Impact of Financial Incentives on Individual Performance: An Experimental Approach

Diplomarbeit von Steffen Hetzel

Introduction:

Many jobs are today paid via performance-based payment or at least have a component including financial payments based on individual performance. Especially, as ones performance is comparably easy to investigate, financial incentives dependent on visible achievements seem useful and practicable and in consequence are multiply used in today’s society. This means, that numerous employees, especially at a particular level (Managers) or with a substantial part of identifiable success (Sales, Consulting) are rewarded for their effort based on observed measures of performance. The intention behind performance-based compensation is to stimulate individuals to increase their motivation and effort spent on tasks, and hence their output, or in other words, profitability for the company. By using this way of payment, special incentives are placed and a change in one’s behavior is intended to be reached for receiving the maximum possible outcome for the company.

On the other hand, the usage and effectiveness of financial incentives is not always seen as a positive instrument at all. The application of those tools and resulting behavior changes are seen as one of the major reasons for the financial crisis beginning in 2007. Wrong incentive setting for the employees and hereby misleading behavior of the staff yielded to specific behavior which lead to a fundamental crash of the world economy and threw the world into a global recession.

After those incidents, the mode of functioning of financial incentives and the most favorable usage of this motivation instruments came forcefully to the spotlight. Especially, comparably high ‘bonus payments’ for bankers were in consequence of billions paid by governments for bank rescue and avoiding a collapse of the global financial system publicly discussed.

The existing and fundamentally logical point of view on incentives is that incentive payments lead to increased motivation which resulted in higher performance of an individual. Furthermore, a positive correlation between incentives and performance is supposed to exist. This means, the higher incentives are, consequently the higher the motivation for conducting a task and followed by this the performance on a task can be expected to be.

The need of designing optimal contracts has always been a challenge, not only for companies. Economical research is delivering tools for understanding and minimizing problems arising through incomplete contracts between employer and employee. Previous to the design of optimal contracts, the efficiency of financial incentives needs to be investigated and proven.

Psychological research is laying the focus on analyzing motivation and performance of employees and how the connection between intrinsic and extrinsic motivation can be affected through the application of different rewards.

From an economic perspective, the Principal-Agent-Theory is dealing with the relationship between employee and employer and suggests solutions for minimizing the problem of asymmetric information and consequently providing optimal contracts for receiving maximum performance from the employee.

Theoretical approaches on applying small financial incentives in comparison to their absence have been analyzed and supplementary, as incentives increase, the effects on individual willingness for achieving additional earnings has been investigated. However, those models are hardly to be reviewed via empirical analysis as data are difficult to survey.

An experimental surrounding is predestined for the issue of analyzing and deepening findings on individual behavior in a tightly controlled environment and by this enriching economic research. So far, conducted experiments give useful information on the effectiveness of varying financial incentives but do not come up with an explicit explanation on the width mode of functioning.

The thesis on hand is enlarging the branch of research by suggesting the design of an experiment for further analysis of the way financial incentives work. The approach of experimental research is chosen in consequence of the advantages in using experiments for analyzing people’s behavior and in testing suggested theoretic models under supervised conditions.

A special perspective has been laid on lately conducted experimental studies to offer a current insight into the state of research. Additionally, the theoretical backgrounds on incentive effectiveness are introduced, which include psychological research on motivational development into economical optimization constraints.

The scientific relevance of this thesis is justified through the target of enhancing the branch of research of experimental economics by expending the perception on the effectiveness of financial incentives on individual performance. Consequently, it features assistance on promoting the understanding of individual behavior and principal-agent problems and enlarges the findings on analyzing the incentive-performance relation from an experimental perspective. By suggesting the design of an experiment to be conducted, the bases for subsequent analysis are being laid and successive research is relieved.

For the purpose of enlarging the research, this thesis is arranged in four parts:

First, an intuition on the theoretical research on the effectiveness of financial incentives is given. For this purpose, Chapter 2 provides an overview of the psychological literature, followed by an introduction of economic theory.

In Chapter 3, performed empirical studies by means of analyzing data sets and findings from realized experiments on financial incentives are described in more detail.

Based on the findings of those recently conducted experiments, a guideline for the body of an experiment for further research is introduced in Chapter 4. Therefore, results of the presented researches are compiled and the possible course of a graph expressing the relationship between financial incentives and individual performance is displayed. Questions that arose through conducted studies are pointed out and hypotheses for further analysis are introduced. Adapted from the hypotheses, the design of an experiment for enlarging the research is displayed. In connection with this, the requirements for the execution of the experiment are introduced and discussed. Suggestions for analyzing collected data within the experiment and reporting raised findings are listed. In the end, possible occurring challenges throughout the procedure of the experiment are mentioned and potential options for tackling them are suggested.

In Chapter 5, conclusions on the introduced analysis are drawn and expected research findings as well as further research areas are pointed out.

Table of Contents:

	Abstract	v
	List of Tables	ix
	List of Figures	x
1.	Introduction	1
1.1	Scope of Research	2
1.2	Outline of the Thesis	3
2.	Theoretical Background	5
2.1	Psychological Background	6
2.2	Economic Theory	11
3.	Research Findings	19
3.1	Empirical Findings	19
3.2	Experimental Findings	24
3.3	Critical Acclaim and further Questions	30
4.	Proposals for the Structure of an Experiment	34
4.1	Objective	34
4.2	Design	37
4.3	Organization	46
4.4	Subsequent to the Realization	48
4.5	Potentially occurring Challenges	51
5.	Conclusions	53
5.1	Summary of the Results	53
5.2	Further Research Suggestions	55
	Appendix A	57
	Appendix B	60
	References	66

Text sample:

Chapter 3, Research Findings:

The subsequent chapter is providing an overview on research findings. For this, the first part contains empirical findings concerning the impact of financial incentives on individual performance and in which way theoretical research results have been proofed by data analysis. The second part is presenting an insight into conducted experimental research and illustrating relevant experimental settings in more details to develop, based on so far conducted experiments, the setting for an advanced experimental design in Chapter 4.

3.1, Empirical Findings:

Prendergast stated as a result of his literature overview on the effects of financial incentives that agents respond to incentives and that a strong effect is achieved by pay-for-performance on output, admittedly in settings where measures of overall performance were available. He points out that those effects are not necessarily beneficial.

As there are publicly available data on the management level, it is not surprising that much of the work on incentives has been conducted on this level. Prendergast highlights that many of the constraints on the literature have been imposed by data limitations, as there are simply no easily accessible databases with personnel data. This illustrates the major problem of empirical research on this topic, as informative data sets on incentive payments are hardly available and surveyed.

The results of three meta-analyses concerning the findings of experiments conducted by economists on the impact of financial incentives on individual performance are introduced in the first part of this subchapter. This offers a survey of the research and prepares the ground for deeper specifications of the conducted experiments. Subsequently, studies analyzing the change of incentive schemes within companies and consequential results are presented.

Meta-Analysis on the Effects of Financial Incentives:

By introducing conducted meta-analyses and literature reviews an idea of the variability of research on this topic should be achieved, providing reliable and solid data for understanding. Research on incentive effects is diverse as they can influence quality and quantity of work, can be shaped in different amounts of the financial incentives or have a motivating effect. Those differing targets of research cause evaluating meta-analysis to be difficult and to be handled with care.

In their meta-analysis Jenkins et al. evaluated 39 experiments with 47 studies conducted between 1969 and 1996 to analyze the relation of financial incentives and individual performance. For this, laboratory experiments, field experiments and experimental simulations have been considered.

They figured out a 0.34 correlation between financial incentives and performance quantity , whereas they did not detect a significant correlation between financial incentives and performance quality. Furthermore, the authors investigated an influence by the setting of the experiment but no impact as a result of the performed tasks.

Camerer and Hogarth analyzed 74 experiments with varying financial incentives. Their focus relied on zero, low, or high performance-based financial incentives. As a result, they declared that the hypothesis of financial incentives not to make any difference in performance is false. Furthermore, they investigated that higher incentives often improve the individual performance. In the case of judgment tests , financial incentives did improve performance of the participants in the experiments. Whereas in other tests the intrinsic motivation was high enough or additional effort did not matter as the task was too hard, thereby monetary incentives were irrelevant. The authors furthermore investigated several tasks where financial incentives even hurt and worsen performance . The authors suggest conducting further research especially within firms as the natural setting and the companies’ incentive schemes may yield to different results than laboratory research does.

Condly et al. analyzed 45 studies (with 64 acceptable incentive/performance research comparisons) conducted between 1960 and 2000 in which designed field and laboratory research on the use of incentives to motivate performance were reported.

The overall effect of all incentive programs in all work settings and on all work tasks was an average 22 percent increase in work performance compared to people who performed similar work and did not receive incentives. The influences of nine different factors on individual performance have been analyzed and average gains on performance due to monetary incentives (27 percent) were doubled compared to those by non-money but tangible incentives (gifts or travel) (13 percent). One more indication of the power of monetary incentives was the fact that the longer incentives were in place, the greater the realized performance gain was. Short programs (less than a month) created gains averaging 20 percent performance increase, but more than six months programs realized an average gain for all programs of about 44 percent. The main result of this study is a strong support for the claims that incentives can significantly increase work performance when they are carefully implemented and performance is measured before and during incentive programs. Money resulted in higher performance gains than non-monetary, tangible incentives.

In summary, the results of presented meta-analyses provide no obvious picture which could be drawn. One of the reasons is the variety of experiments and their classification as well as the difficulty to compare research findings. It is hard to find a common ground as studies are built up in different ways, have different target groups or work with different incentive schemes. Several more extensive reasons ought to be considered. First, meta-analyses can only take into account studies which have been published. This might yield to a publication bias, as a multitude of studies are only published when significant results are on hand. The introduced analyses give only an overview on those experiments which have been selected and through this, a selection-bias may be the cause for the differing results. Furthermore, the quality of the comprised studies could only be superficially controlled and including bad studies results in a questionable methodology of the meta-analysis.

Analysis on the Introduction of Financial Incentives in Firms:

Between 1994 and 1995 Lazear investigated the performance of workers on a production line to survey the effects of monetary incentives on individual output . In the analyzed company the paying scheme for glass installers was changed from a time-rate wage (pay per hour) to a piece-rate wage (pay per unit) over the course of a year and a half. In total, the productivity of 3,000 different workers over a 19-month period was considered.

The outcome of the analysis originated an increase in individual productivity for workers who started in the time-rate scheme and switched to the piece-rate scheme of 22 percent. Furthermore, an increase of about seven percent in compensation of the workers had been detected. By comparing the numbers of productivity gains and compensation increase, overall profitability gain for the company has been observed . Through those results and improved payments and profitability, a win-win situation for the company and the workers has been realized. Whereas, an increase of profit in combination with the adjustment of the compensation scheme is due to the theory not a compulsorily effect.

Lamere et al. conducted a study at a wage disposal firm in Michigan for nearly 4 years, which changed the compensation scheme to a more performance-based model. For this, 22 drivers in the roll-off division were assigned into two groups. The two groups were composed for analyzing differing effects of the financial incentives. The members of group 1 switched to the new incentive scheme 14 weeks prior to those of group 2. After the introduction of the new incentive scheme, drivers received their basic salary and an additional incentive if their performance exceeded the baseline average.

As a result of this study the introductions of the incentive program lead to an improved performance at both studied groups. However, performance did not constantly increase as the value of incentives increased from the initial level. As a side effect of the introduction of the incentive program, the job satisfaction of the workers moved up and the spending on net labor costs for the company was reduced.

Shearer analyzed data from a field experiment to estimate the gain in productivity that is realized when workers are paid piece rates rather than fixed wages. The experiment was conducted within a tree-planting firm and provided daily observations on individual worker productivity under both compensation systems. Unrestricted statistical methods estimate the productivity gain to be 20 percent. Since planting conditions potentially affect incentives, structural econometric methods are used to generalize the experimental results to out-of-sample conditions. The structural results suggest that the average productivity gain, outside of the experimental conditions, would be a productivity gain of at least 21.7 percent.

Chang investigated the composite effect of extrinsic motivation on work effort. For this, from July through October 2000, data from 401 employees from 29 companies in Korea have been collected . The author analyzed the impact of the extrinsic motivation effect of money and job security.

Chang found a positive correlation between individual performance and monetary incentives, as well as between performance and job security. The most important result from his analysis is the positive combined effect of the two extrinsic motivation concepts. As workers receive higher extrinsic motivators, they are supplementary motivated. However this effect is reduced as the composite level becomes extremely high.

Supplementary empirical research has been conducted leading to the result of improved performance of the observed subjects by using financial incentives.

In summary, a comprehensive trend of improved performance as a result of the use of financial incentives can be determined throughout those analyses. Individual performance of workers increased compared to the absence of financial incentives. Furthermore, the study of Chang shows the implication of individual reference points beyond which personal effort on further earnings decreases.