A change to Curia’s assumptions on electorate seats

This is a heads up that from next month, I am going to make a change to assumptions about electorate seats for any NZ political polls by Curia. This post explains why.

Up until the end of 2024, Curia assumed (as most pollsters do) that there will be no change in electorate seats held, when projecting what Parliament would look like on one of our polls. This is the safest or default position to take.

But in January 2025 National dropped to 29.6% in the Taxpayers’ Union poll which meant they would only get 38 seats based on their party vote. They currently hold 43 seats, so the status quo assumption would mean projecting they get 43 seats – an overhang of five seats. This would mean instead of projecting a CR:CL seat distribution of 62:58, it would be 67:58.

I was uncomfortable with this. I thought carrying on with this assumption would be overly favourable to National as it would be saying they will win 43 electorate seats no matter how low their party vote drops. Based on 25 years of experience I knew that if National really did have an 8% drop in their party vote, they would then lose some electorate seats. I don’t know which ones, but that it would be some.

So since January 2025, the TU-Curia poll has not assumed any overhang seats for National. This assumption is explicitly mentioned in the report.

However I have assumed that Te Pati Maori will win all its existing six Maori seats. And for seven of the 11 polls this year, this has had them having a one or two seat overhang. I justified this on the basis there was no reason to think TPM would lose any of their current seats as their party vote was slightly up from the election.

So this year the TU-Curia polls have had assumptions that actually favour the centre-left parties as we have assumed National would not have overhang seats but TPM would. I will show later how our seat projections would have changed with different assumptions.

It is worth noting that the 1 News – Verian poll makes the same assumptions – no overhang seats for National, but TPM will retain six electorate seats.

The differing assumptions were justified based on the political environment of the time. But with the expulsion of two MPs from Te Pati Maori, I do not believe it is prudent to assume they will keep all six current seats. If the purpose of the poll is to give a snapshot of what a likely Parliament is, then that assumption can’t be justified any longer.

Rather than have assumptions that change over time based on events, my new policy is that Curia seat projections will no longer make any assumptions on overhang seats. The seat projections will be based purely on the party vote. There will still be an assumption that parties currently with electorate seats will retain at least one electorate seat in terms of whether they meet the MMP threshold of eligibility for List MPs (5% party vote or one electorate seat).

	Status Quo	No TPM overhang	National overhang
Jan-25	CR+4	CR+4	CR+9
Feb-25	CL+2	CL+2	CR+2
Mar-25	CL+4	CL+4	CL+3
Apr-25	CR+7	CR+8	CR+8
May-25	CR+5	CR+6	CR+6
Jun-25	CR+2	CR+4	CR+3
Jul-25	CR+8	CR+10	CR+9
Aug-25	Hung	CR+2	CR+3
Sep-25	CL+1	Hung	Hung
Oct-25	CL+2	CL+2	CR+3
Nov-25	CR+2	CR+4	CR+6

This shows the seat projections under three scenarios. The first column is the status quo of assume TPM has overhang but National does not. The second shows what would happen if you did not assume an overhang for TPM. It would only affect the outcome in August (from hung to CR) and September (from CL to hung).

The third shows what it would have looked like if one assumed an overhang for both TPM and National. This would have changed the projected outcome in February, August, September and October.

So again to make it clear – so far this year we have assumed no overhang for National, but an overhang for TPM. In future we will assume no overhang for either party.