Servizio Comunitario di Informazione in materia di Ricerca e Sviluppo - CORDIS


GOVERN Sintesi della relazione

Project ID: 283837
Finanziato nell'ambito di: FP7-IDEAS-ERC
Paese: United Kingdom

Final Report Summary - GOVERN (Local Governance and Dynamic Conflict in Developing Countries)

This research project follows two strands of enquiry motivated by the political economy of developing countries.
The first line of enquiry aims to use the introduction of local elections in rural China through the 1980s and 1990s to understand the benefits of democracy in relatively small communities and in a context of autocracy. These elections were introduced due to the perceived generalized failure of exiting local officials, who were appointed and controlled via the traditional vertically organized bureaucratic system. These local officials are very important to village life, because they organize and collect the funds necessary for providing public goods (schooling, digging ditches, paving roads) and are also responsible for land allocation to families and for the smooth running of village enterprises.
The reform that was implemented was stark and simple. All the characteristics of these official posts were kept untouched except one: where they were formally appointed by higher levels of government, they would not be elected by villagers.
Of course, elections in autocracies are often used for window-dressing as no real room is given to the voice of the people. In this particular case, the local Communist Party cell and the upper levels of government stay in place and hence could potentially undo any positive effect of elections.
Together with Qian and Yao we collected a novel dataset that covers the period 1986-2005 for more than 200 rural villages. This dataset combines existing data from the Ministry of Agriculture with a newly run retrospective survey and contains the economic and political history of these villages for the period.
We use the staggered timing of introduction of elections to show that elections indeed have a very large effect on public goods provision, it increases by about 50%. In addition we find that proxies of corruption such as land allocated to enterprises goes down and also does within-village inequality. These results all suggest that the system of local elections, which leverages villagers’ knowledge, is better at controlling local officials than bureaucratic oversight.
We also explore which villages benefit the most out the introduction of elections. We find that, in line with existing literature, villages where citizens are fractionalized in several social groups benefit less from elections. The surprising finding is that the dimension of fractionalization along which this happens is religious affiliation, which has been traditionally downplayed as a relevant social cleavage in rural China. Similarly, we find that villages with a good tradition of civic capital benefit much more from the introduction of elections. Formal institutions (such as elections) turn out to be complementary to informal institutions (civic capital) in this context.
Finally, we formalize the trade-off that introducing elections causes for an autocrat. On the one hand, elections are beneficial because local villagers have better information than distant bureaucrats on the performance and type of local officials. On the other hand, local villager preferences are not completely aligned with the autocrat. The model predicts that elections will have different effects on policies as a function of their alignment. We show that indeed, well aligned policies such as public goods provision increase when elections are introduced. But policies such as the One Child Policy, in which interests strongly diverge, become worse implemented. We discuss the implications of this theory for the motivation of introducing elections in the early 80s and weakening them from 2004 onwards.
The second line of enquiry focuses on corruption.
In a first article with Chassang I analyse the strategies that a principal with limited instruments can follow in order to maximize the information that he receives from potential whistle-blowers. We show that by altering his audit policy the principal can complicate the ability of corrupt agents to threaten whistle-blowers and thus increase the flow of information. We also show that even if reports are non-verifiable and the principal has no clear priors on underlying corruption, a set of policy experiments can provide guidance on the prevalence of the problem.
In an article with Burgess, Jedwab, Miguel and Morjaria I show that there has been very high ethnic bias in the allocation of road funds in Kenya since independence. In particular, regions inhabited by the co-ethnics of the President have received more than double expenditure per capita than other regions. Interestingly, this dynamic political corruption disappears during democratic years. We interpret this using a model of constraints on executive action.

Reported by

United Kingdom