Women leaders and deceptive behaviour

Are women in leadership positions more dishonest than men? Based on an artefactual field experiment in rural Bihar, this column finds that women in leadership positions deceive more than men, especially in villages that have previously experienced a female village chief. It suggests that simply reserving seats in village councils for women may not necessarily lead to better outcomes for villagers or women.

Women play an important role in generating sustainable economic development, but in many parts of the world, their potential and participation in leadership is limited. In recent years, the Indian government has taken a number of steps to redress this injustice, with the 73rd Amendment to the Constitution of India reserving one-third of all positions in village councils (including that of the village chief or mukhiya) for women. In many states in India, this proportion has increased to half. Major international organisations often place gender at the heart of development programmes. For example in 2012, 83% of all World Bank loans require gender to be a key component of project outcomes (World Development Report (WDR) 2012).

Extensive research in economics and other fields examines the effects of female leaders on economic and social outcomes. In the context of village governance in India, there is evidence that the presence of female leaders can increase girls´ aspirations and educational attainment (Beaman et al. 2012), improve distribution of public goods towards female-preferred goods such as water and sanitation (Chattopadhyay and Duflo 2004), and results in a decrease in reported crimes against women (Iyer et al. 2012). However, these positive effects might take time to eventuate, partly because good governance takes time to learn, and often women become village heads with less governance experience than men (Afridi et al. 2013). Also, there may be backlash from males (Gangadharan et al. 2014) since women in positions of leadership may be perceived as violating existing social norms.

While gender equality and the greater representation of women in public life is a desirable goal and has resulted in some tangible policy changes, increasing representation of women through quotas may not necessarily have the desired economic or social impact. This is particularly important when social norms do not necessarily support the appointment of women to such positions of power. An excellent example comes from a policy change in Norway, which mandated that publicly-listed companies should allot at least 40% of board seats to women. The outcomes of this policy have not been particularly encouraging. There has been an exodus of firms that were listed in Norway’s stock exchange, from 563 in 2003 to 179 in 2008. Further, despite the affirmative action policy, The Economist recently reported that only 6% of listed Norwegian firms had a female chief executive, which is not very different from Fortune 500 American companies (Economist 2014).

Do female leaders behave differently from male leaders?

This question is typically difficult to answer using observational data because the actions of the leader (of either gender) cannot be separated from the social environment in which they occur. In addition, differences in actions by men and women can simply be driven by differences in experience, preferences or constraints they face in policymaking.

To overcome this challenge, we use an artefactual field experiment (see Harrison and List 2004 for a definition) in which all subjects have the same level of experience so that we can isolate the actions of the leader from those of citizens and others around the leader. This provides a uniform and straightforward measure of the actions of male and female leaders. Specifically, we examine, within the context of the field experiment, whether women leaders are more deceptive than male leaders. This is an important question because dishonest and deceptive behaviour can lead to a breakdown in cooperative relationships, contracts and trust and as a result, hinder economic growth. Additionally, greater deception by women may reinforce gender-based discrimination such as male backlash leading to deterioration in trust and threatening women leaders’ ability to govern.

Using data from the 2011 Census of India, we randomly selected 40 villages in three districts of Bihar (Madhubani, Gaya and Khagaria), ensuring that villages are observationally similar in terms of village demographics and infrastructure, as well as equidistant from Patna, the capital of Bihar. An important difference within the sample was that some of these villages had experienced a female village mukhiya while others had not. Whether or not a village council had had a female mukhiya was randomly determined as reserved village councils were decided according to a rotating schedule.

955 male and female residents of the 40 villages, from different socioeconomic backgrounds, participated in the experiment. Within each village, participants were randomly allocated into groups, each comprising two men and two women. Each participant received an initial endowment of Rs. 200 and decided how much of this endowment to contribute to a group account, with the remaining to be allocated to a private account. Participants earned Re. 1 for themselves for each rupee placed in the private account. Contributions to the group account by the group members were aggregated, doubled and then divided equally among all group members.

However, before subjects made their contributions, a leader was randomly selected from each group, with men leading half the groups, and women leading the other half. By selecting the group leader at random from the group, we avoided self-selection into leadership based on experience, ambition or other characteristics. The leader had to propose a non-binding contribution (they were free to change the amount later) between Rs. 0 and 200 towards the group account. This proposed contribution was revealed to the other group members (called citizens). Then all members of the group, including the leader, simultaneously decided on their actual contribution to the group account. These actual contributions were not revealed to the group members at any stage.

We define leaders to be dishonest or deceptive if their actual contribution towards the group account was less than what they themselves proposed. Since the proposed contribution by the leader was non-binding, akin to cheap talk, standard economic theory suggests that the proposal stage should have had no impact on citizens’ contribution decisions. Leaders also knew that group members may not follow their proposal, and therefore had little incentive to follow it as well. Recent experimental evidence (Levy et al. 2011) however suggests, that leaders’ suggestions, even though non-binding, effectively result in higher contributions from citizens. Anticipating this, income maximising leaders would contribute less than what they proposed.

Since leaders were randomly assigned to groups, we can compare the behaviour of male and female leaders. The results of our experiment show that women are significantly more deceptive leaders – deceiving (contributing an amount that is less than that proposed) in 56% of cases as opposed to 43% for men. The difference in the extent of deception (difference between the amount proposed and actual contribution) between male and female leaders is, however, not statistically significant. The results are not driven by differences in characteristics or experiences as individuals are randomly allocated to be a leader. Previous research (Swamy et al. 2001, Alatas et al. 2009) has shown that in several developing countries, there is little evidence to suggest that women are less corrupt than men. This is also consistent with evidence from the World Values Survey data, which asks respondents whether someone accepting a bribe is acceptable. Data from India (in addition to several other developing countries such as Bangladesh, Vietnam, Philippines, China etc.) show that women are as likely as men to report that accepting a bribe is acceptable.

Is this behaviour driven by the environment in which the field experiment was conducted? To investigate this, we examine whether and to what extent the behaviour of leaders in our experiment is influenced by perceptions or exposure to female mukhiyas. We do so by comparing the behaviour of female and male group leaders in villages with and without at least one female mukhiya in the last three village council elections.

Our experimental results show that female group leaders are deceptive more often and in a greater magnitude, contributing 26% less than proposed, in villages that have had at least one female mukhiya in the last three village council elections. In villages that have never been exposed to a female mukhiya, there is no difference in the likelihood of deception of male and female group leaders. In other words, women who have experienced a female mukhiya show a tendency to appropriate resources when acting as a leader in our experimental setting.

Policy implications

Increased representation of women in governance and business can potentially improve both gender equality and the quality of governance and state capacity. However, our results show that simply mandating female leadership does not necessarily lead to better outcomes for villagers or women, and instead female leaders can engage in more dishonest behaviour as compared to male leaders.

One possible explanation is that female leaders expect to be treated poorly, which leads to a self-fulfilling prophecy where female leaders behave in a negative manner, contributing less than proposed. Alternatively, in environments with few opportunities for leadership, women act myopically and take one-off decisions when presented with an opportunity as they do not expect to be re-elected. Indeed, for any individual woman, the likelihood of continuing to be in a position of village mukhiya is fairly low. For example, a Government of India survey in 2008 found that 89% of female village mukhiyas who were interviewed did not contest a second election and everyone who did lost. The erstwhile mukhiyas reported that while in the first instance, spouses and other influential people in the village encouraged them to contest the election, the same relatives and spouses later discouraged them from re-contesting. In contrast, men with greater experience of leadership may take a more long-term view, and seek to build and sustain a norm of good governance.

Improving the effectiveness of female leaders is of paramount importance for economic growth. Placing women in leadership positions through policies such as gender quotas is however, only the first step in ensuring that women are equally represented and have the opportunity to improve the quality of governance. Our research suggests that social norms also need to change in a way that changes women’s expectations of being treated poorly. Unless women expect to be re-elected based on their ability, their leadership potential may be impeded. This can have adverse implications on the community.