Day 3: Research Session – Political Economy, Accountability and Governance

The session was hosted by Professor Eliana La Ferrara (Bocconi University). In the first part of the session, the presenters elaborated new ways of fostering political engagement to find a way around clientilistic politics. The second part of the session took a different approach to the field by looking at the role of capital cities in civil conflict and regime change.

Prof. Leonard Wantchekon (Princeton University) discussed his work on reducing clientelism. Especially in low and middle income countries, clientelism is a destructive political dynamic that can lead to entrenched corruption. The role of power brokers is important in this as these brokers mobilise support and invest to get a candidate elected, but they then often demand a payback that is ultimately sourced from public resources.

In terms of election campaign approaches, clientelistic platforms essentially promise rewards in return for votes. These create reciprocity between politicians and voters, and they tend to be effective in achieving strong election results. Programmatic platforms where normative policy issues are discussed, on the other hand, can generate communication between voters and with voters and the candidate, but they’re not that useful in winning elections.

The paper is about two experiments that support a programmatic election approach. The first experiment was conducted in the Philippines. It only cost about $5,000 and the intervention was to organise town hall meetings where political programmes were discussed. These had a positive effect on both turnout and electoral support for candidates receiving the treatment. The second experiment was conducted in 90-120 minute meetings, without the candidates, in treatment villages. The control was again a business-as-usual campaigns. The aim was to isolate the effect of programmatic versus clientelistic campaigns.

The effects observed were several. Treatment raised election turnout, but not that much in aggregate. More significant was the impact on vote shares. In the Philippines parties gained about 2%, enough to get them into parliament. Importantly it was shown that the programmatic approach was as effective as the clientilistic one, but at a much lower cost. In sum, programmatic approaches alone don’t work, there needs to be dissemination. Deliberative campaigns are effective in boosting the success of programmatic platforms.

Rachel Glennerster (Director, IPA and IGC Sierra Leone Lead Academic) presented a complementary piece of research that focuses on improving the interaction between voters and candidates. The question was whether debates affect voting behaviour. While journalistic accounts suggest that they do, there has often been cynicism in the academic community due to the belief that clientilism limits any impact of such debates.

A local NGO, Common Ground, hosted and filmed debates between parliamentary candidates with a stardardised structure with various questions on politician background, policy goals and the like. These videos were shown to participants as it was not possible to hold similar debates in several places. There were treatment and control polling stations.

The results included that voters learned a lot about the candidates and also the political system and their rights as citizens. Seeing the debates increased voter knowledge along several dimensions, changed results to those politicians with better policies and increased voting across ethnic lines. Voters then chose candidates more attuned to their own policy preferences and politicians got 5% more votes if they won a debate.

The third and final presentation of the day was by Filipe Campante (Harvard Kennedy School) who sought to answer the question: why don’t politicians steal the entire budget they’re entrusted with? Constraints probably include elections, pressure from parliament, possible legal action and the media.

Even in superficially unconstrained government contexts, the risk of civil conflict and popular pressure leading to a change in government is present. In such civil strife, capital cities play the pivotal role. Recent events in Ukraine and Bangkok suggest that even elected governments may be affected by large scale demonstrations in the capital.

The project looked at the remoteness of capitals and effects of shifting capitals in a cross country setting. The data heavily draws on PIRIO-GRID data and includes controls such as luminosity, terrain and climate. Predictions were that A) conflict is more likely to emerge close to the capital as it’s cheaper to buy off those far away, B) conflict close to the capital is more likely to topple incumbent governments and C) isolated capitals are associated with mis-governance. In sum, it was thought a priori that isolation of capitals is associated with less accountability and worse governance.

Running the numbers, the first result is that these effects can generally be shown, but only in autocratic regimes. Democratic countries, by contrast, appear largely unresponsive to these effects. In terms of conflict prevalence, civil strife becomes more likely if the capital is moved closer. Related to this, regime change is more likely if conflict occurs close to the capital and regime change also becomes more likely if the capital is moved closer to areas of conflict. Overall, it is indeed the case that isolated capitals are associated with worse governance in autocracies. A general though emerging from this is that the spatial distribution of the population is important for understanding governance and institutions. A hypothesised reason is that information transmission functions better in densely populated areas.

By Tim Ohlenburg, Country Economist, IGC Uganda