DP96 Why do so many MPs lose their seats in PNG?

Multiple regressions are used to assess the relative significance of a number of political and social variables that might be associated with the likelihood of sitting MPs retaining or losing their seats in the 85 partially rural open electorates of Papua New Guinea (PNG). The tests reveal that the most significant of the variables for which we have reasonably accurate measures is the amount of ministerial power accrued by sitting members, either over the previous term of parliament or at the time when they stand for re-election. MPs who have spent more of the past parliamentary term in more powerful ministerial roles are less likely to lose their seats in the next election. Also significant, though less so, is a second political variable. This is the level of political competition in each electorate, as measured by the average number of candidates challenging the incumbents. Incumbents are more likely to lose their seats in electorates where more candidates have typically stood for election. Relationships between the rate of MP turnover and the social variables selected for inclusion in this analysis appear to be quite weak or entirely non-existent, but the evidence does indicate that the rate of turnover has little if anything to do with the level of ‘development’ in each electorate, and is certainly no higher in electorates with higher levels of linguistic diversity than it is in electorates with greater linguistic uniformity.