What happens to a criminal charge that never gets tested in court? Sometimes it's withdrawn: the prosecution pulls it before a verdict. In 2025, the most recent year in the figures, New Zealand courts recorded 42,311 withdrawn charges across all regions and all offence types (Source: Stats NZ, charges-by-offence-type).

That number has been climbing hard. In 2021 the count was 30,120. Four years later it's 42,311. That's 12,191 more withdrawn charges, a rise of 40.5 percent in four years.

Here's the part that's easy to miss. The climb has almost run out of puff. Withdrawals rose 9.8 percent in 2022, then 13.0 percent in 2023, then 11.5 percent in 2024. Then 2025 added just 632 charges, an increase of 1.5 percent. Three years of double-digit-ish growth, followed by something close to flat.

Now zoom out, because the long series changes how the current figure reads. Stats NZ has this data back to 1986, forty years of it. Courts withdrew more charges than 42,311 in every single year from 1992 through 2010, and in 1986 through 1989 as well. Only 1990 and 1991 came in lower than today's figure. The peak was 58,396 in 2009, about 38 percent above the 2025 count.

Then came a collapse. From that 2009 high, withdrawals fell for six straight years to 28,120 in 2015, the lowest figure anywhere in the series. That's a drop of more than half in six years. The following decade has been a slow, uneven walk back up: a bump in 2017, a dip in 2021, then the four-year run that gets us here.

So which is the real story? The most likely reading is that both are. A 40 percent rise in four years is a genuine shift in how charges are resolving. But the level we've arrived at is not historically unusual. It's roughly where the courts sat in the late 1980s, and comfortably below the 50,000-plus years that were routine through the 2000s. A number can be the highest since 2010 and still be ordinary by the standards of the twenty years before that.

The honest caveat: these are counts, not rates. This dataset shows how many charges were withdrawn, not what share of all charges that represents. If the total number of charges going through the courts has grown over the same period, a rising withdrawal count could mean the withdrawal rate is flat or even falling. The 2009 to 2015 collapse also overlaps with major changes in how low-level matters were dealt with, so some of that fall may reflect charges never being laid rather than charges being dropped. And recent years in official series are the ones most likely to be revised.

My own view: the single most useful number here isn't the headline 42,311. It's the 632. If withdrawals have genuinely levelled off after three years of steep increases, that's the turning point worth watching, and it will only be confirmed or killed by next year's release. Until then, anyone reaching for the 2025 total as proof that something has broken should look at what 1998 and 2002 and 2009 looked like first.