What ‘hidden’ success stories tell us about anti-corruption policy and practice
What ‘hidden’ success stories tell us about anti-corruption policy and practice
Heather Marquette and Caryn Peiffer
Identifying ‘Islands of Integrity’ in Uganda and South Africa
When Uganda’s Minister of Health literally went undercover in a burqa to investigate bribery, her televised exposé of the health workers who asked her for bribes was one of many recent high-profile anti-corruption interventions. Despite the criticism levelled at such interventions in the literature, these seem to have worked. But for how long?
The corruption literature abounds with titles such as ‘Why do so many anti-corruption efforts fail?’ or ‘Curbing corruption…an impossible dream?’ While there are interesting case studies, typically on single countries that have overcome systemic corruption – mostly developed countries – development policy makers have few sources to turn to when they want inspiration for programme design.
Our research project on ‘Islands of Integrity’ tested a novel three-stage methodology developed to identify potential ‘hidden’ positive outliers on bribery reduction in contexts where success is unexpected. These are cases where sector-specific bribery rates have fallen far more than expected, compared with other sectors in the same country.
We found that bribery for health services in Uganda halved between 2010 and 2015, while bribery in all other sectors went up. Our research suggests that the credit for this dramatic reduction belongs to the recently established Health Monitoring Unit (HMU). Among other activities, the HMU conducts unannounced—and sometimes televised—investigations in health facilities to catch health workers ‘red handed’, including the infamous example involving the disguised Minister of Health. The two arrests that resulted from that sting were just two of over 600 arrests that the unit can claim. Our research suggests that health workers have grown more cautious about asking for bribes out of fear of being publicly shamed and punished by the HMU.
In South Africa, the statistics focused our attention on the province of Limpopo. There, the bribery rate in the police fell by almost 15 percent between 2011 and 2015, another surprising reduction given bribery trends across other sectors and provinces. In this case, credit goes to an unrelated national government intervention. With the province facing bankruptcy, the national government took control of five provincial departments in Limpopo from 2011-2015, and carried out many corruption investigations; interestingly for our research, the intervention did not specifically target the police. At least 300 people were charged with corruption as a result, all processed in local police stations. Interviews with several officials suggested that even though the intervention did not target the police directly, it is likely to have reduced their willingness to engage in bribery. The police seem to have avoided asking for bribes out of fear of being caught and uncertainty about whether they were also a target of the investigation, even going so far as (allegedly) bugging an investigator’s accommodation to find out more.
Both cases are examples of the quintessential principal-agent reforms that recent literature tells us don’t work. Yet these did, and fast.
But what about sustainability and unintended consequences?
Sadly, the story doesn’t end there. Our research shows that the success we observed is unlikely to be sustained and could lead to counterproductive effects.
Firstly, according to principal-agent theory, corruption patterns can be disrupted by targeting opportunities to engage in corruption or by making the costs of doing so outweigh the perceived benefits. For the police in Limpopo and health workers in Uganda, the fear of being caught and punished increased and, for many, outweighed the benefits.
Secondly, both interventions highlight the need to think about potential unintended consequences of anti-corruption work. This seems to be a major gap in the literature, given that anti-corruption is – at its heart – about forcing those with the power to abuse their positions for their own private gain to give up this power and to be accountable for their actions. History has shown this rarely to be a peaceful or uncontested process. In anti-corruption, there is a high potential for “blowback”.
In Uganda, the HMU’s approach to ‘naming and shaming’ has led health workers across the country to go on an unprecedented strike lasting over a month, which was said to have brought the health system to its knees. Our research does not excuse health workers demanding bribes for services they should provide for free. But weakening staff morale resulting at least in part from the anti-bribery intervention is unlikely to lead to the massive improvements in health care needed for services to improve for patients. At least in the short-term, humiliated and fearful health workers and worsening patient care appear to be ‘blowback’ for the HMU’s aggressive approach.
In South Africa, the tale of unintended consequences is a very different one. Our research did not uncover an anti-police bribery intervention that led to unintended consequences; instead, the reduction in bribery itself is the unintended consequence, the sort of thing Perri 6 calls a ‘benign side effect’. And of course, you can’t make sustainable what you don’t realise has happened in the first place! It would be useful if we knew more about when and why such benign side effects have happened in other cases. This could lead to research to help predict when they may occur or the factors that could encourage them to emerge.
These were complex interventions with complex and, in some ways, contested outcomes: a simple narrative about success and failure cannot do justice to this complexity. While these cases are indeed statistically significant, they also show why we need to contextualise and question the ‘positive’ in ‘positive outliers’. Thinking about the functions that corruption fills; figuring out how to make disruptive approaches ‘stick’; better understanding unintended consequences – whether good or bad, our research suggests that all of these are important for developing more effective, sustainable anti-corruption interventions. Further research on such cases could help policy makers better anticipate when and why the serendipity of ‘benign side effects’ can occur as well as helping to avoid the potential disaster of ‘blowback’.
Heather Marquette is Professor of Development Politics at the University of Birmingham.
Caryn Peiffer is Lecturer in International Public Policy and Governance in the School for Policy Studies at the University of Bath.