By Mirco Tonin and Michael Vlassopoulos, lecturers in economics at the University of Southampton; cross posted from VoxEU
Aside by Yves; I have very mixed feelings about publishing this article. First, any study that reaches men v. women generalizations has to be viewed with a lot of skepticism. For groups as large as men and women, the variations in behavior within each class are going to be greater than the variations across classes. So I wonder whether this sort of thing merely reinforces and legitimates the propensity to stereotype. Second, the authors cite pre-existing research, and their study findings were consistent with that research. This study was not double blind. Query whether researcher bias influenced the results.
To the article:
What motivates workers in their job? This column presents evidence from a recent field experiment suggesting that women are motivated by concern about the social cause pursued by their employer, while men are not. This may provide new insight into the gender earnings gap.
What motivates workers? The canonical view in economics is that workers respond to monetary incentives. In line with this perspective, a large literature has been devoted to analysing how compensation contracts should be designed in order to induce workers to be more productive (Prendergast 1999).
Recently, economists are becoming increasingly aware that this may not be enough for certain types of workers. Besides extrinsic rewards such as bonuses or promotions, an important driver is a concern about the social cause pursued by the organisation for which they work or a sense of altruism towards the welfare of a third party who is the recipient of the good or service being produced in their workplace.
A worker with this type of motivation will provide more effort and require less monetary compensation compared to a worker solely motivated by monetary rewards. This type of altruistic motivations is potentially very important in organisations engaging in the provision of education, health care, childcare, and social services – as well as in charities and non-government organisations.
“Warm glow” or “pure altruist”?
Research on altruistic behaviour distinguishes between two types of pro-social motivation:
A worker may enjoy contributing to a cause he or she cares about. This joy that arises from the act of contributing is referred to as “warm glow” (Andreoni 1990). For example, a nurse may enjoy to be actively involved improving the lives of sick people.
A worker may be directly concerned about the well-being of others. Such individuals are referred to as “pure altruists”. For example, a nurse may be interested in the fact that patients’ health improves, regardless of whether he or she has a personal role in this.
The distinction between these two perspectives of pro-social motivation – which are of course not mutually exclusive – has significant implications. For example, if workers are pure altruists, then non-profit organisations may have an advantage in eliciting their effort compared with for-profit organisations (Francois and Vlassopoulos 2008). No such difference arises if workers are solely motivated by warm-glow altruism. Thus understanding the sources of workers’ pro-social motivation has implications for the policy debate on whether public services should be provided through the public sector, the for-profit private sector, or non-profit organisations. For example, drawing on the nurse example from above, one implication of privatising a health facility is that a pro-socially oriented and purely altruistic nurse may decrease or withdraw altogether the amount of donated labour he or she provides.
Despite the important policy implications, no attempt to quantify and discriminate the importance of the two views of altruistic motivation using non-experimental data has been made, in part because appropriate field data that would allow for sound econometric analysis are difficult to come by. In recent research (Tonin and Vlassopoulos 2009), we have performed a field experiment that aims to identify and quantify these sources of pro-social motivation in the workplace.
A field experiment with student-workers
Our research uses experimental methods in which we randomly assign workers to environments that are designed to elicit the different motivational aspects described above. We then measure the impact on productivity. An important aspect of our study is that we observe subjects providing real effort in a natural work environment, as opposed to the artificial environment of the lab where behaviour may not be representative of what happens in the field (Levitt and List 2007).
In the autumn term of 2009, we hired university students through email announcements to perform a short-term computer data entry job. In particular, workers were employed on two separate occasions lasting one hour each and taking place two weeks apart. For each session workers received £10 plus a performance bonus based on their productivity.
On the second occasion, we randomised students into three different groups. For one of the groups the second occasion was identical to the first one. This baseline condition acts as our control, as it accounts for any change in productivity due to experience, learning and so on. For the two other groups, we implemented two treatments aimed at eliciting, respectively, warm-glow effort and effort induced by both types of altruistic preferences. In both cases, personal compensation was identical to the one received in the first session and, moreover, a donation to a charity was made on the worker’s behalf based on his or her productivity. For one group the amount received by the charity was held constant, as the donation done on the worker’s behalf crowded out one-for-one a donation done on the employers’ behalf. For the second treatment group, there was no crowding out, so that the more work, the higher the total amount received by the charity.
Comparison of productivity across the three groups allows us to assess the relative strength of the two sources of pro-social motivation. In particular, comparing the changes in productivity between the treatment with crowding out and the control group, allows us to detect any effort due to warm-glow altruism, while comparing the changes in productivity between the treatment without crowding out and the treatment with crowding out, allows us to detect any effort due to pure altruism.
New insight: Gender differences in pro-social behaviour
We find that women’s effort is positively affected by an environment that induces warm-glow altruism, while there is no additional impact due to pure altruism. In particular, in the treatment condition eliciting warm-glow altruism, women increase their productivity between the two sessions by an additional 10% compared to women in the control group.
On the other hand, we find no statistically significant differences in productivity changes between the control and any of the treatment groups for male subjects. This unresponsiveness suggests that pro-social preferences are less relevant for men than for female workers in our sample. This finding is consistent with other research on gender differences in social preferences (Croson and Gneezy 2009).
The finding of a gender difference in pro-social behaviour in a workplace setting may have important implications for understanding women’s economic outcomes. If women are indeed motivated by a concern for the social cause pursued by the organisation they work for, then they will be more likely to enter occupations and sectors with characteristics that engender pro-social behaviour (such as health, education and social care) and will require less monetary compensation. Gender differences in pro-social motivation might therefore help explain the observed occupational segregation by gender that accounts for a substantial portion of the overall gender earnings gap.
While our results provide a new insight, it would be inappropriate to draw firm implications for the labour market as a whole from a single field experiment using student workers who exert effort on a short-term job. Further empirical studies are needed to evaluate whether our findings are robust in other labour market settings and populations of workers.