Corrections has ended contracts with iwi organisations providing bail beds and wrap-around support, leaving only 10 funded bail beds, even as the remand prison population grows, as reported by RNZ. If you want to understand why that decision lands hard, look at one offence category in the remand data.

Burglary. Unlawful entry with intent, break and enter. In 2025 the count of remand prisoners in that category hit 3,207 (Source: Stats NZ, remand-prisoners). That is the highest figure in the entire 27-year series, going back to 1998.

Here is the part that should make you sit up. The previous high was 2,658, set in 2019. Nothing in the record before that came close to 3,207. From 2004 onwards the number sat in a fairly stable band, bouncing between roughly 1,970 and 2,658 for two solid decades. Governments changed. Bail laws changed. Sentencing rhetoric changed. The burglary remand count stayed inside that band year after year.

Then 2025 broke it. The count went from 2,202 in 2024 to 3,207, an increase of 1,005 people in twelve months. That is a 46 percent jump in a single year, and it clears the 2019 peak by about 21 percent.

For scale: the 2025 figure is nearly five times the 660 recorded in 1998, the first year of the series. It is also more than 800 above the average of the last decade's readings, which never once cracked 2,700.

The most likely reading is not that New Zealanders suddenly started breaking into houses at half again the usual rate. Remand counts measure people held before their case is resolved, not offences committed. A number like this moves when bail gets harder to get, when court backlogs push cases out longer, or when fewer supported places exist for someone to be bailed to. All three of those levers have been in play. The bail bed cuts RNZ reported sit directly on the third one.

Now the honest caveat, because this is where a single year of data can mislead. A 27-year series with a one-year spike of this size deserves suspicion before it deserves headlines. If 2025 is a partial or still-settling year, or if counting rules for "Total Remands" duration changed, some of that 1,005 could be an artefact rather than a real increase in people behind bars. This series also cannot tell you whether burglary itself is rising. It counts people awaiting court, and nothing else.

Even allowing for all that, the shape of the chart is hard to argue with. Twenty years of stability, then a vertical move.

My view: if you are going to hold a thousand more people on remand for burglary than you did the year before, cutting the small number of beds that let some of them wait in the community is the wrong sequence. Remand is the most expensive way to house someone who has not been convicted of anything yet.

The 2025 count is the most recent available. Watch 2026. If it settles back toward 2,600, this was a blip. If it holds above 3,000, the band that held for twenty years is gone.