We expect employers to reward effort even if the employers knew output was determined by luck; when effort is unobservable, we work harder if the employer doesn't know earnings are determined by luck

Effort Provision in a Game of Luck. Mads Nordmo Arnestad, Kristoffer W. Eriksen, Ola Kvaløy and Bjørnar Laurila. Front. Psychol., May 20 2021. https://doi.org/10.3389/fpsyg.2021.637339

Abstract: In some jobs, the correlation between effort and output is almost zero. For instance, money managers are primarily paid for luck. Using a controlled lab experiment, we examined under which conditions workers are willing to put in effort even if the output (and thus their employer’s earnings) is determined by pure luck. We varied whether the employer could observe the workers’ effort, as well as whether the employer knows that earnings were determined by luck. We find that, workers believed that the employer will reward their effort even if their effort does not affect earnings. Consequently, workers work harder if the employer could observe their (unproductive) effort. Moreover, even when the employer only saw earnings and not effort, workers labored harder if the employer did not know that earnings were determined by luck.

Discussion

Our experimental results provide support for all four hypotheses:

1. Most subjects exerted positive effort even when effort was unproductive.

2. They exerted more effort when effort was observable.

3. They expected employers to reward effort even if the employers knew output was determined by luck.

4. In the case where effort was unobservable, subjects worked harder if the employer did not know earnings were determined by luck.

The latter results were driven by female workers, reflecting past research suggesting that females place an overall higher personal value on effort (McCrea et al., 2008). It is important to note that we did not expect a gender difference at the outset of the experiment. As such, there is a relevant chance that the observed relationship reflects a random effect. However, we find that the result ties in with a greater stream of research indicating that female research participants demonstrate a stronger general tendency to portray themselves in a socially desirable manner (see Dalton and Ortegren, 2011).

To the best of our knowledge, these results are novel. The effect of noise on effort provision has been explored before, but no past studies have looked at effort provision in a setting where the correlation between effort and outcome is zero. Similarly, the relationship between observable effort and judgments of character has been explored numerous times but never in a setting where the futility of effort is common knowledge. Even in cases where effort was completely unrelated to outcomes, participants in this study tended to obey a work ethic heuristic. This was especially true when effort was observable, suggesting the work ethic heuristic has less to do with outcomes and more to do with social signaling. Our participants also expected to be rewarded for effort, even if the lack of relationship between effort and outcomes was common knowledge. This implies our participants expected that the work ethic heuristic was shared among their peers and that those who followed it would be rewarded for doing so, regardless of the outcome. While all participants exerted effort as an outward social signal when effort was observable, female participants also exerted effort as an inward social signal by working hard even when effort was unobservable.

There are some other possible reasons why the research participants chose to exert unproductive effort. Experimenter demand-effect may have prompted some of the participants to work. Similarly, boredom could be a motivating factor. While we cannot rule out these factors completely, we nevertheless believe that their role in the observed relationships is limited. Firstly, the demand effect or boredom effect would have been equal across treatments. Secondly, the participants were told that they were allowed to use their phones when they had finished working. As such, they would most likely have found alleviation from boredom more effectively by surfing the web rather than working at a mindless task which was explicitly unrelated to outcomes.

We instead interpret our results in the light of a work-ethic heuristic; the simplified view that effort is always preferable to less effort. As a general rule in life, people will observe that effort is related to outcomes, and outcomes are related to rewards. As such, most adults will approach any novel task with an implicit understanding that their performance can be improved with effort, and that good performances will be rewarded. This relationship is further cemented by cultural norms and practices that elevate the moral value of hard work, and condemn the sin of sloth and inactivity. The combined effects of cultural norms and intra-personal learning makes people behave in a way that is consistent with a work-ethic heuristic. In our experiment, however, effort was unrelated to performance. This demonstrates that the work-ethic heuristic, like most heuristics, is useful and adaptive in the normal set of circumstances, but lead to unproductive behaviors in different circumstances. As a general rule, reliance on the heuristic is beneficial at both the individual, organizational and societal level. However, in the few but notable cases where effort is unrelated to outcomes, the consequence of continued reliance on the work-ethic heuristic depends on the perceived cost of effort. If the workers experienced cost of effort is negative, reliance on the work-ethic heuristic will still produce a favorable outcome. However, if the experienced cost of effort is positive, as we argue it was in our experiment, continued reliance on the work-ethic heuristic leads to waste of resources.

Our experimental design is rather stylized. In the real world, neither workers nor employers will have full knowledge about the relationship between effort and output, and they will typically hold beliefs that effort—to some extent or in some cases—leads to higher performance. However, these lab experiments offered the advantage of an environment where only luck mattered and where we could control whether and to whom this information was available. This helps rule out confounding factors that may matter in real world environments where luck is important but not definitively. Additionally, it allows us to rule out standard economic theory as potential explanations for the results we achieved.