In Papua New Guinea, District Services Improvement Program (DSIP) funding, given to 89 MPs representing open seats, is often talked of as a tool used to maintain government coalitions and increase an MP’s chances of being re-elected. In this post we investigate whether the funds really do help hold together governing coalitions and assist MPs in their quest for re-election.

It has been argued that DSIP funds were introduced as a tool for holding together political coalitions. Their popularity among MPs is thought to stem from their utility as a means of rewarding and expanding an MP’s support base.

As DSIP data is now available, it is possible to test if these beliefs are actually correct. Is there a relationship between DSIP funds and government tenure? And between DSIP funds and the overall election rate of open MPs?

The following charts show these relationships.

At first glance, the first chart appears to show a positive correlation between DSIP volumes (the sum of real DSIP funds disbursed in each year they were PM) and prime ministers’ tenures. Governments have lasted longer on average when real (inflation-adjusted) DSIP volumes have been higher. However, this relationship is driven entirely by two outliers: Sir Michael Somare’s tenure from 2002 to 2011, and Peter O’Neill’s tenure from 2011 to 2019. If these outliers are removed, there is no clear relationship between real DSIP funds and a PM’s longevity (r = –0.2).

Somare’s longevity does not seem to be a direct product of DSIP funds, rather it is best explained by the restrictions imposed by the Organic Law on the Integrity of Political Parties and Candidates 2003 (OLIPPAC). Among other things, OLIPPAC prevented the movement of MPs from one party to another, and required MPs to support the PM they voted for after national elections, if there was a vote of no confidence. (These restrictions were declared unconstitutional in 2010.)

Meanwhile, O’Neill’s long tenure was foremost the product of a slew of constitutional amendments: lengthening the grace period in which no-confidence motions are allowed, reducing the number of parliament sitting days, increasing the number of MPs required to lodge a vote of no-confidence motion, and lengthening the time in advance required to lodge a vote of no-confidence motion. During O’Neill’s tenure, these amendments were eventually found to be unconstitutional, but they helped O’Neill prior to that ruling. He later made use of reduced sitting days and longer grace periods to avoid a vote of no confidence. O’Neill also adjourned parliament several times to avoid votes of no confidence. It is true that O’Neill’s long tenure appears to have been aided by his willingness to withhold DSIP funds from opposition MPs, but this wasn’t the sole source of his long stay in power.

There is also no clear relationship (r = –0.1) between real DSIP funds (in the year prior to election) and open MP win rates in elections, as the second chart above shows. This would suggest that the amount of DSIP funds disbursed does not provide MPs much by way of electoral advantage.

In summary, DSIP funds may have helped hold together government coalitions in particular instances, but they haven’t been universally helpful. Increasing DSIP funds also do not appear on average to have helped MPs win re-election.

Sadly, we don’t expect these findings to diminish the enthusiasm of PNG’s politicians for the DSIP. Electorate funds have been popular with every prime minister since 1984. Even Sir Mekere, PNG’s foremost reformer, feared the political implications of ceasing MP funding when he became PM in 1999.

Ultimately, a decision on DSIP funding should be made on developmental not political grounds. Nevertheless, the fact that incumbency rates have not increased with DSIP funding is surprising, and calls for more research on how DSIP funds are spent and what the attitudes are of voters to them.