Sharing Common Resources in Patriarchal and Status-Based Societies: Evidence from Tanzania

ABSTRACT

In rural African societies, socioeconomic differentiation linked to gender and social status exerts an important influence on the distribution of common-pool resources. Through a behavioral experiment conducted in 2008 in rural Tanzania, this contribution examines the influence of gender and social status on distribution behavior of users of self-governed common watersheds. It finds that men and women with low social status distribute water equally when water is abundant but keep larger shares when water is scarce, although low-status women try to be as fair as possible at the expense of their returns from irrigated agriculture. Men of high social status keep more than half of the available water for themselves, both in abundance and scarcity, and deprive others from sizeable returns from irrigated agriculture. Women of high social status share altruistically when water is abundant and equally when water is scarce, giving up on returns from irrigated agriculture.

KEYWORDS:

• Gender
• social status
• sharing
• common-pool resources
• rural Sub-Saharan Africa

JEL Codes:

• B54
• Q25

NOTES ON CONTRIBUTORS

Els Lecoutere holds a PhD in political science and is associated to the Conflict Research Group at Ghent University. Her interest is development research with a political economic perspective. She adopts a gender-specific research approach, seeking evidence of the impact of gendered social relations, norms, attitudes and power dynamics on decision-making processes and outcomes thereof. Her doctoral research focused on local governance of common water sheds in Tanzania. Currently, she is involved in research for empowerment and development in East Africa. She studies inclusiveness of value chains, land governance, and gender constraints to enhancing agricultural productivity.

Ben D’Exelle is Senior Lecturer in Economics at the School of International Development (University of East Anglia). Most of his research can be located in the fields of behavioral and experimental economics and impact evaluation in low-income countries. The research themes he has pursued and published on include: social networks and resource sharing, the functioning of rural markets, equity-efficiency trade-offs in irrigation systems, reproductive health, repayment and impact of microfinance, the micro-politics behind community-based development and the behavioral economics of rural investment. He has conducted most of his fieldwork in Nicaragua, Tanzania, and Uganda.

Bjorn Van Campenhout is a research fellow at the International Food Policy Research Institute based at the IFPRI-Kampala office in Uganda. His area of expertise are staple markets in East Africa. His research focuses on smallholder market participation, spatial and intertemporal price arbitrage, and the use of ICT for information and extension delivery to isolated areas.

ACKNOWLEDGMENTS

We acknowledge financial support from Ghent University, IDPM (University of Antwerp), FWO, and MICROCON. This research received research clearance from the Tanzania Commission for Science and Technology (COSTECH) and the Mufindi District Council. We are grateful for support from Charles Kyando and Incomet 2001. We thank Nathalie Holvoet, Anne Walraet, and participants of the Water in Africa (Porto) and ISNIE 2009 (Berkeley) conferences for useful comments on earlier drafts.

Notes

1 CPR are natural or man-made resource systems that are sufficiently large to make it difficult and costly to exclude potential beneficiaries (Ostrom 1990). The rival nature of the resources distinguishes CPR from public goods (Elinor Ostrom, Roy Gardner, and James Walker 1994).

2 Participants were neither expected to rank themselves nor irrigators they were not well acquainted with. Rungs could be removed/added if participants distinguished less/more than four categories.

3 The table and every other tool used during the experiment presented both figures and symbols that were easily accessible to all participants.

4 Payoffs were measured in TSH, with 1 US$ = 1,200 TSH at the time of data collection. A household budget survey conducted in 2007 estimates average monthly income at 39,362 TSH for mainland Tanzania and 28,418 TSH for rural areas (National Bureau of Statistics 2009).

5 Being upstream or downstream user in the experiment was independent of irrigators’ actual position in the irrigation schemes. We assumed that, in reality, most irrigators have a neighbor located upstream and downstream and are therefore familiar with decisions both users would have to make in the experiment.

6 One could argue that social distance effects and relational aspects of gender and social status may have had stronger effects on behavior if gender and social status of the paired participant were known. Yet, providing this information could have influenced behavior in uncontrollable ways, introduced experimenter demand effects, and jeopardized anonymity.

7 We use the survey sample rather than the experiment sample for checking correlations because of the bigger sample size.

8 Livestock value is the total value of big and small livestock owned by the respondent's household calculated using market values at the time of the survey. Calculations were based on http://www.lmistz.net/Pages/Public/chart.aspx, and market prices recorded in Mufindi district, July 2008. Percentage of food/cash crops is the percentage of the irrigated plot allocated to food crops (maize and beans) and to cash crops (tomatoes, carrots, cabbage, and African eggplant). Cash crops defined on the basis of the survey are crops of which more than 70 percent of the harvest is sold.

9 Figure 1 does not show frequencies of distribution decisions by the 20 percent of men and women upstream users with medium social status.

10 The inclusion of social status and interaction effects in Model B does not reduce the significance of gender as compared to Model A with only gender and scarcity as explanatory variables. Even if gender and social status are correlated, the two variables clearly have discernible effects. In the supplemental files, available online on the publisher's website, we investigate the extent of a possible specification error linked to the omission of the downstream user's reaction as an explanatory variable in Models A and B.

11 The feedback sessions were less useful for learning low-status people's explanations for their water-distribution behavior in the experiment, as they did not acknowledge selfish sharing during scarcity as being their own behavior. Possibly they did not admit to unfair distribution behavior when in a group, as it is ill accepted in their community.

12 While feedback participants agree that “native” village people are more likely to stick to water-sharing rules than “nonnative” people, they did not refer to women from minority ethnic groups as more likely rule breakers but rather to people from the nearby town renting irrigation plots for commercial purposes.

13 The differences, however, are not significant, most likely due to small sample sizes.