Abstract:

District Services Improvement Program (DSIP) funds are constituency development grants which members of parliament in Papua New Guinea (PNG) exert considerable control over. Similar funds exist in a number of other countries, but on a per capita basis PNG’s funds are some of the highest globally. The funds are also controversial, and often said to be tools of patronage. In this paper, we study whether the funds have any discernible impact on MP re-election rates. We do this by taking advantage of the natural experiment afforded by malapportionment in PNG. DSIP funds are given in equal amounts to all constituencies. Owing to substantial malapportionment, larger electorates receive considerably less funding per capita. We test whether incumbents are more likely to win re-election in smaller electorates where DSIP funding is higher on a per capita basis. Despite including (and excluding) a large range of controls, as well as different variants of variables, we find only very limited evidence of a difference in re-election rates. We also show that this is not simply because it is inherently harder to be re-elected in smaller electorates. We do find some evidence of DSIP money being associated with higher incumbent vote shares. However, even this evidence is limited. The balance of available evidence strongly suggests that having more DSIP money does not help MPs win re-election on average.