The rules of pork? The impact of electoral systems on legislators’ behaviour in Honduras
Muñoz-Portillo, Juan
(2013)
The rules of pork? The impact of electoral systems on legislators’ behaviour in Honduras.
PhD thesis, Dublin City University.
Carey and Shugart (1995) suggest that under closed-list proportional representation (CLPR) legislators are more likely to seek votes based on their personal attributes as
the number of seats per constituency (known as district magnitude or simply as M) decreases. By contrast, in open-list proportional representation (OLPR) the incentives for legislators to seek votes based on their personal reputation should increase as M increases. I test these scholars’ predictions and argue that regardless of the electoral system, legislators will focus on the provision of local public goods, known as pork-barrel politics, the smaller the M, because then it is easier to claim personal credit for projects for their constituencies. In OLPR the importance of
pork-barrelling increases compared to CLPR; however, because the necessity to differentiate themselves from co-partisans increases as M grows, depending on their personal attributes legislators will choose to provide non-targetable goods or private goods instead.
I study the case of Honduras to test this theory. This country used a CLPR system between 1989 and 2001. In 2004 the system changed to OLPR, keeping M fixed. The outcomes of the research are as follows. First, the results do not support the claims posed by Carey and Shugart (1995). Second, using different proxies of legislators’ behaviour, I found confirmation that pork-barrelling is more important
the smaller the M. Third, the evidence regarding the provision of private goods does not support the claims on its own. However, when private goods and local goods
were combined, some evidence was found that might indicate that, as expected, the change from CLPR to OLPR increases the likelihood that legislators will provide targetable goods the larger the M. By contrast, in CLPR the effect is the opposite.
Finally, the results suggest that electoral systems interact with the personal attributes of legislators in ways that can be anticipated.