For the past decade or so, there’s been a movement to bring computers, telephones and other “information and communication technology” into developing nations to increase economic development and eliminate poverty. Those of us involved with this movement - colloquially called ICT4D (Information and communication technology for development) - have argued that information imbalances underly major problems in economic development. If farmers don’t know fair prices for their commodities in big cities, they’ll sell for too little money. If students can’t access textbooks or other resources, they’re doomed to a poor education.
There’s a strong critique of ICT4D that argues that the importance of information is overstated and that ICT4D proponents either overvalue information technology because they’re personally attached to the tools, or more sinisterly, because they’re looking to create developing world markets for these tools. Many supporters of ICT4D - myself included - will concede that there are lots of badly thought out and poorly executed projects that do little more than drop expensive technology in areas where it’s a scarce resource and likely to stay a scarce resource for a long time to come.
One bright light for the ICT4D field has been the rise of eGovernment, a movement that tries to get governments to deliver key services to citizens using digital technology. India has been the location for many eGovernment pilot projects, some of which have been very successful in delivering key information services to citizens. In many states, citizens can visit information centers where they can obtain driver’s licenses, business licenses, residency or birth certificates, and other critical documents.
Jennifer Bussell, a political scientist who recently completed a PhD at UC Berkeley, has spent a great deal of time studying these projects and asks a tricky and important question about eGovernment in India - why do some states adopt eGovernance more readily than others? Are there policy environments that we can put in place to make it more likely that eGovernment projects will succeeed and that they’ll affect the lives of citizens positively?
In a talk at the Berkman Center on Tuesday, she offered an interesting opening paradox. The state of Karnataka is comparatively wealthy and extremely engaged with information technology - its capital is Bangalore, the epicenter of India’s technology and outsourcing industries. Chhattisgarh is a new state, carved out of Madhya Pradesh in 2000, and is extremely poor and low-tech. We’d expect eGovernment services to catch on in Karnataka much more quickly than in Chhattisgarh… and we’d be wrong. eGovernment has caught on far more quickly in this young, poor state than in the technology giant, raising questions about what factors actually contribute to the success or failure of eGovernment projects.
To understand what’s going on in these two states - and indeed, across many of India’s states (Bussell developed her theories in seven Indian states and has tested them on nine additional states, analyzing 16 of India’s 28 states) - it’s important to understand corruption, and how eGovernment might affect corruption. Indian citizens pay a lot of money in bribes. It’s estimated that Indians pay $5 billion USD annually to bribe government officials. Sometimes this is wealthy citizens paying money to “jump the queue” and obtain services more quickly that average citizens. But extremely poor citizens pay bribes as well - Bussell references a study that suggests that citizens below the poverty line collectively paid $22 million in bribes to access essential and guaranteed government services.
Taking old, paper-based bureacracies and turning them into “e-government” services appears to squeeze some opportunities for corruption - “rent-seeking”, in the language of political economics - out of the system. It’s not entirely clear why this is - the service centers rolled out in Indian states don’t generally put computers in the hands of citizens and let them access services directly. There’s an opportunity for the operators of these new systems to seek bribes. But the digitalization of India’s massive railway system is a good example of what’s happened in some eGovernment systems. Before digitalization, it was difficult to purchase a ticket without knowing someone to bribe within the system. Now tickets can be purchased online, and transactions within railway stations are simple, efficient and bribe-free (even if you’re a clueless American looking for trains from Rajastan to Delhi, as happened to me not very long ago.)
Bussell argues that e-services tend to systematically reduce corruption, and that they therefore can be threatening to existing political elites. Elites have the power of transferring bureacrats, moving them from a job where it’s easy to seek bribes (the customs service) to one where it’s harder to do so. They exercise this power by demanding kickbacks from bureacrats, which they use as campaign finance. A politician whose political livelihood relies on control of bribes and rent-seeking officials is likely to be threatened by eGovernment efforts and might fight their introduction.
Bussell further theorizes that the removal of bribes could be a threat to political stability within coalition governments. A coalition can be thought of as a group of politicians all seeking a share of the benefits of being in control of a state’s government - part of this control includes control over offices with a high chance for gains through corruption. So she theorizes that we’ll see eGovernment projects succeed in areas where there’s lower corruption, and where there’s a single party in power.
She studies eGovenrment adoption by tracking how many services are available in a given state - some offer just a few, like driver’s licenses, while others offer dozens. Her models try to explain the adoption of eGovernment services in terms of several factors. Some turn out to be largely irrelevant. Technology infrastructure isn’t statistically significant in explaining why some states have aggresively embraced eGovernment. Nor is the time of adoption - states that started eGovernment earlier aren’t neccesarily ahead of the curve. And the level of economic development isn’t statistically significant either.
Corruption, on the other hand, is a strong factor - states with above average corruption (based on surveys by groups like Transparency International) have adopted 10.6 services on average, while those with below-average corruption average out at 20.1 services. Unitary government matters as well - single party governments with below average corruption adopt services more aggresively than coalition governments, even in below-average corruption states.
This is useful information for anyone attempting to build eGovernment systems and roll them out in developing nations, though it doesn’t offer much insight on what to do if you’re in a high-corruption, coalition-governed area. (Duck and cover, perhaps.) And there’s a intriguing larger question - how does the introduction of eGovernment affect corruption in the long term? Do states that adopt eGovernment systems become progressively less corrupt over time? Bussell’s intrigued by these questions and looking for ways to study them going forward, which is good news for anyone who cares about ICT4D and wants to make sure people are doing rigorous, careful evaluation of what works and what fails.
This post originally appeared on Ethan Zuckerman's blog, My heart's in Accra.