Published on Let's Talk Development

Social norms and incentives: Homo economicus is dead, long live bounded rationality, social interdependence and culture!

This page in:

The following post is a part of a series that discusses 'mind and culture,' the theme of the World Bank’s upcoming World Development Report 2015.

All public policy is based on assumptions, whether implicit or explicit, about human behaviour. These assumptions, particularly those about what motivates people, are often incomplete. A better understanding of human motivation, one that draws on a range of disciplines, offers a chance to improve the effectiveness of development policy.

This argument provides a foundation for the forthcoming World Development Report 2015 on Mind and Culture. It was also the point of departure for a very interesting session linking this welcome reflection on disciplinary diversification at the World Bank to evolving thinking on service delivery at the recent Conference on the 10th anniversary of the 2004 World Development Report, Making Services Work for Poor People.

The opening presentation by Varun Gauri suggested that social norms may have important impacts on each of the three points of the famous WDR 2004 ‘trio’ of accountability relationships linking citizens/service users, providers and policy-makers.

The evidence presented in the session indicated the time and effort that has been put into thinking about innovative ways to leverage social norms to shift behaviour in communities (e.g. in community-led total sanitation schemes). There are also clear lessons to be learned from a variety of disciplines about how to coax better behaviour from providers and thereby achieve better development outcomes. Of particular interest here was the discussion on establishing internalised professional norms among providers, with some very nice thinking from Pascaline Dupas and Karthik Muralidharan drawing on a growing body of research work.

So, how do we get teachers to turn up in classrooms and teach? How do we get nurses and doctors in clinics and hospitals to act in the best interests of their patients, despite particularly acute information asymmetries? There appears to be a role for both incentives and social norms. While much of Muralidharan’s work provides a case for changes in hiring and payment practices in the education sector to reflect what we’re learning about how to incentivize improved performance, Dupas pointed to emerging evidence from Ghana that even in the apparent absence of material incentives or sanctions, nurses tend to behave more professionally than expected. These are by no means contradictory conclusions and indeed some of the most exciting discussion broached potential complementarities and contradictions, blurred only slightly by some debate over the way in which social norms were defined.

Yet I cannot help wonder whether the emphasis on new approaches to changing behaviour was somewhat skewed. That is, do we think enough about whether this sort of analysis is, or ought to be, applicable to efforts to get more developmental behaviour out of the third corner of the WDR 2004 triangle – the policy maker? What lessons from attempts to establish professional norms among service providers could be applied to the way in which we select and manage politicians? The implicit assumptions that are so often challenged elsewhere seem to be left alone in terms of a general faith in the ability of the democratic system to select and provide effective incentives for politicians.

In this vein, one potential implication is that we need to further explore potential trade-offs between activities that appeal to international norms and their effect on local social norms and expectations. For example, indices that measure perceptions of the prevalence and severity of corruption suggest a belief in the potential to incentivise improved performance (reduced corruption) by providing information on the perceived level of corruption in a particular country. Where countries are performing worse than their peers, this is certainly a form of condemnation, but what can we say about the impact of this shaming effect compared with the potential for such indices to reinforce the belief that corruption is rife and the associated negative norms and expectations of behaviour? In other words (to paraphrase an argument Rothstein and Tegnhammar extend from Myrdal), if I, as a politician, am bombarded with evidence that everybody else is (or is seen to be) taking bribes, why shouldn’t I as well? These perverse effects have been documented elsewhere with respect to tax compliance and deserve further exploration.

Similar questions could be asked about norms related to narrowly focused patronage relationships and other aspects of governance and service delivery. Incentives, electoral and otherwise, clearly matter, but the premise of the forthcoming WDR 2015 – that social norms can either support or constrain material incentives – seems to demand closer scrutiny of the potential to leverage these norms to shift political behaviour.

Might we actually be better off with strategic misinformation campaigns that encourage the development of more positive norms? We may well decide that the moral implications of such duplicity are too severe, but there may be a more practical middle ground in some cases, such as the form of selective or purposeful reporting of positive behaviour (though this may run counter to journalistic instincts and incentives that favour a juicy corruption story) or targeted messaging (though this raises questions about traditional media partners).

In the end, this session raised far more questions for me than answers. The policy implications are far from clear, providing good reason to stay tuned as the WDR 2015 tries to synthesise the wide-ranging set of studies within its remit (some with questionable external validity as a result of the high context specificity of social norms). Whatever the result next year, this research agenda clearly raises questions that ought to challenge our preconceptions about the mechanisms we use to induce better behaviour by politicians, as well as providers and ourselves.


Daniel Harris

Research Fellow, Politics and Governance, ODI

Join the Conversation

The content of this field is kept private and will not be shown publicly
Remaining characters: 1000