New Caledonia Market Size After The 2024 Crisis

Someone asks how big the New Caledonia market is. The honest answer starts with a question back: big for what. The number that sizes a retail catchment is not the number that sizes a hotel's addressable demand, and neither is the number that tells you whether a business-to-business software vendor should bother. After May 2024 those three answers diverged sharply, and a lot of planning still runs on figures from before they did.

This article walks through the numbers that actually answer the question, names the institution and date behind each one, and flags the ones that mislead. Several figures in circulation for this territory are wrong in specific, traceable ways. One of them is a population count that is not a population count. Another is a jobs figure that appears in at least five different versions depending on which report you happened to read.

We work this market from Indonesia, so we have a commercial interest in it looking attractive. Read what follows with that in mind, and note that we have not made it look attractive.

The population, and the number that is not one

ISEE announced first results of the 2025 census in Noumea on 29 July 2025. Enumeration ran from 22 April to 28 May. The count came to 264,596 residents.

That is down 6,811 from the 271,407 recorded in 2019, an annual rate of about -0.45%. The composition matters more than the total. Births exceeded deaths by roughly 11,000 over the period, which on its own would have produced growth of about 0.7% a year. A migration deficit of roughly 18,000 people cancelled that and more.

Put the sentence plainly: the territory is still having children, and it is losing adults faster.

One more piece of context makes this land properly. This is the first time the population of New Caledonia has fallen between two censuses since 1946. Not since the 2019 count, not since the pandemic, since 1946. That is the framing the census mission itself used, and it is the right one, because it separates a demographic turning point from an ordinary bad year.

Now the number to discard. You will see 296,000 quoted as the population of New Caledonia, sometimes in otherwise careful documents. It is not a census figure and it is not a population estimate. It is arithmetic: 243,000 internet users divided by 82.0% penetration equals roughly 296,000. That is the denominator a connectivity dataset used inside its own model, reverse-engineered and then promoted to a population count by people reading the report rather than the methodology.

The consequence is not academic. A market sized at 296,000 is 12% larger than a market sized at 264,596, and every audience estimate, share calculation and revenue projection built on it inherits that error. We caught this figure propagating through two full research documents during preparation for this market, where it had been used to build an argument about market scale and then cited by a companion document as established.

Use 264,596. Cite ISEE and the July 2025 announcement. If someone hands you 296,000, ask which census produced it.

The economy, in the order the institutions published it

Real GDP contracted 13.5% in 2024. That figure comes from CEROM, the joint statistical partnership of ISEE, AFD and IEOM, in its rapid economic accounts for 2024, published in September 2025. Of every figure in this article, it is the one with the widest agreement across sources and the least ambiguity about what it measures.

For 2025 the picture is less settled, and we will present it as the range it is rather than picking a side. One published projection put the further contraction at around 8%. Another, from the AFD observatory, put it nearer 13%, which would make the cumulative fall against 2023 approximately 28%. Both are attributed, both are recent, and averaging them would produce a number that nobody published. So: somewhere between a further 8% and a further 13%, on top of 13.5%.

Household finances tell the same story from a different angle. Over-indebtedness filings rose 159% year on year, per IEOM's 2025 annual report published 23 June 2026. That deserves a second sentence, because the headline invites a wrong conclusion about today: New Caledonia accounted for 52% of all filings handled across the New Caledonia and French Polynesia commissions, roughly 264 cases, against 18% of filings back in 2019. IEOM attributes the surge to the aftermath of May 2024, the winding down of post-crisis support measures, the resumption of loan repayments, and an employment situation that did not improve during the year. And the rate has since slowed sharply. In the June 2026 conjunctural note, filings were up 23.7% year on year and credit repayment incidents 19.8%, so household distress is still deepening but at roughly a seventh of the 2025 pace. Government fiscal revenues fell 26% in 2025 against an initial forecast of a 20% decline, which means the shortfall was worse than the pessimistic planning case.

State support has been doing a great deal of load-bearing. In 2026 the French state committed 9 billion francs to finance 64 structural projects under the pacte de refondation. Read alongside the activity indicators, that is the context in which any talk of recovery has to sit.

Two more indicators from the same annual report belong in any sizing exercise, because they describe capacity to pay rather than willingness to spend. The ratio of doubtful debts to total credit extended in New Caledonia reached 12.4% in 2025, up from 11% a year earlier, which is close to one franc in every eight carrying repayment risk. And at the end of 2025, 29% of firms responding to the IEOM survey still feared their own failure, against a pre-crisis level estimated near 18% in 2023. In the fourth quarter of 2025, a third of surveyed firms reported activity below half of what it had been before the crisis.

A counter-signal sits alongside those, and leaving it out would be selective. Actual business failures have been falling, down 30% since 2022 on a trend IEOM describes as running for three years. Fear of failure is elevated while realised failure is declining. Both are true, and the gap between them is roughly what you would expect in an economy held up by public support.

Employment, and the five different numbers

If you research this market you will encounter at least five figures for private-sector job losses: more than 10,000, between 11,000 and 12,000, between 11,000 and 14,000, between 12,500 and 13,600, and a flat 13,600. They are not all wrong. They are measuring different windows, and most of the sources quoting them do not say which.

The IEOM series resolves it, because it publishes the same measure repeatedly against a stated base.

Measurement point	Figure	Reported in
Total private salaried jobs, first quarter 2024	66,581	IEOM Tendances conjoncturelles, second quarter 2025
Jobs lost between first quarter 2024 and first quarter 2025	more than 10,000	IEOM Tendances conjoncturelles, second quarter 2025
Total private salaried jobs, second quarter 2025	60,550	IEOM Tendances conjoncturelles, fourth quarter 2025
Jobs lost between second quarter 2024 and second quarter 2025	nearly 4,300	IEOM Tendances conjoncturelles, fourth quarter 2025
Year-on-year change through 2025, by quarter	-15.4%, then -7.1%, -2%, -0.9%	IEOM Tendances conjoncturelles, first quarter 2026
Status, first quarter 2026	private employment stabilising	IEOM Tendances conjoncturelles, first quarter 2026

Read down that quarterly sequence and the shape becomes obvious. The collapse was front-loaded into early 2025, and the rate of decline fell by more than fourteen percentage points over four quarters. By the fourth quarter the year-on-year loss was under one percent, which is close to flat.

This is the most important distinction in the whole article. A market that lost more than ten thousand jobs and is still losing them at the same rate is a market you avoid. A market that lost more than ten thousand jobs and has flattened is a market you enter carefully, with realistic expectations about wallet size. The employment data says the second thing. Anyone quoting only the headline loss figure, with no window and no trajectory, is describing a situation that ended in early 2025.

Business confidence, and the caveat the institution attached to it

IEOM publishes a business climate indicator, computed from the responses of roughly 200 business leaders, on a scale where 100 represents the long-run average. The series through the crisis reads as follows.

The historic low was 65.9, in the second quarter of 2024, immediately after the May unrest. A year later it stood at 88.2. In the first quarter of 2026 it reached 99.3, up 2.7 points on the preceding quarter, which puts it just under the long-run average and still below the levels of 2021 to 2023.

Thirty-three points of recovery in under two years looks like a strong story. IEOM does not tell it that way, and neither should anyone quoting the number. The institution's own framing is that the movement reflects stabilisation at a heavily degraded level, in a context of substantial public support, rather than economic recovery. Both halves of that sentence have to travel with the figure.

There is also a figure of 91.0 for end-2025 circulating in some research summaries. It does not fit. If the indicator reached 99.3 in the first quarter of 2026 after rising 2.7 points, the preceding quarter sat near 96.6, and 91.0 also sits awkwardly against the 88.2 reading from mid-2025. We could not reconcile it with the published IEOM series and do not use it.

The sector that decides whether stabilisation holds

Everything above describes symptoms. Nickel is the underlying condition, and no honest sizing of this market can leave it out, because the territory's fiscal base, its export earnings and a large share of its industrial employment all rest on it.

The recent readings are genuinely mixed, which is why they get quoted selectively in both directions.

On the positive side, ore extraction rose about 24% in 2025. Eramet reported mine output at SLN up 10% in the first half of 2026, published 29 July 2026. Those are real production numbers and they point up.

On the other side, only two operators remain active at all, SLN and Prony Resources, and IEOM notes both still face significant structural difficulties aggravated by the fall in international prices. KNS has been in care-and-maintenance since September 2024. That is not a temporary slowdown in output, it is a plant not running, and it removes both direct employment and the supplier activity that sits underneath it.

Then there is the price signal, which requires care. Nickel prices on the London Metal Exchange rose about 11.5% in the first quarter of 2026. Read alone, that reads as recovery. Read alongside the fact that exchange stocks rose 50.9% year on year over the same period, it reads differently. Rising prices with rising inventories is not the signature of demand outrunning supply. It is a market where metal is accumulating in warehouses while the price moves for other reasons. Anyone quoting the price increase without the stock figure is showing you half a chart.

Why this matters for a commercial plan rather than a mining newsletter: nickel is the mechanism by which the territory either grows out of the current position or stays dependent on transfers. The 9 billion francs committed for 64 structural projects in 2026 is bridging finance. Bridging finance ends. If nickel activity recovers to something like its former contribution, household incomes and business confidence follow it up with a lag of quarters, not years. If it does not, the stabilisation visible in the employment and confidence data is a floor rather than a launch pad.

We do not know which. Nobody publishing on this territory knows which. What we can say is that the current indicators are consistent with either outcome, and that any agency presenting you with a three-year growth projection for this market is projecting through a nickel assumption it has not stated.

The 2024 shock, sized honestly

One more figure gets quoted a lot, and it deserves the same treatment as the job losses: a range, with the disagreement shown.

Material damage from the May 2024 unrest has been put at approximately 2.2 billion euros by one widely cited assessment, at 3 billion by another, and at around 2 billion by a third. That spread is wide, and the differences come from what each assessment counted, whether it included business interruption alongside physical destruction, and when the estimate was produced.

The lived consequence shows up in a cleaner statistic. Household consumption fell 7.2% in volume terms during 2024, per the CEROM rapid accounts. That is not property damage, that is people buying less, and it flows directly into the addressable spending of every consumer-facing business in the territory.

For planning purposes, the damage estimate is context and the consumption figure is input. Use the second one.

What stayed intact

Every number so far points the same way. One does not, and it is the one that decides whether any of this is a market at all.

Internet penetration stands at 82.0%, roughly 243,000 users, as of October 2025. Connectivity did not follow the economy down. The population shrank, employment fell, household finances deteriorated, and the share of people online stayed where it was.

That single fact separates New Caledonia from the mental model most people reach for when they hear that an economy contracted by more than a quarter across two years. The audience is smaller than it was, and poorer, and more cautious. It is not offline, it is not unreachable, and it has not stopped searching.

One important handling note. Do not divide those 243,000 users by the 264,596 census figure and publish the result as a penetration rate. The connectivity dataset uses its own population denominator, which is where the discredited 296,000 came from in the first place. Quote the two figures separately, each with its own source, and resist the urge to make them agree.

So how big is it, for what

Three different businesses, three different answers, all from the same underlying data.

Selling to residents. The ceiling is 264,596 people, falling slowly, with a migration deficit that removes working-age adults specifically. Household purchasing power is under pressure, with over-indebtedness filings up 159%. This is a share-capture market, not a growth market. The case for investing in search here rests on taking a larger slice of a smaller pool from competitors who have not adapted, and it is a real case, because most local competitors have not adapted. It is not a case that survives being dressed up as expansion.

Selling to visitors. Your addressable demand sits almost entirely outside the territory, which means none of the domestic contraction caps it directly. What caps it is arrivals, and those are recovering from a low base on an improving monthly trend. The relevant planning numbers are in the tourism data, not the census, and the source-market mix has changed enough since 2023 that pre-crisis targeting is actively misleading.

Selling to businesses. The base is roughly 55,000 private salaried jobs across an economy that has stabilised rather than recovered, with public support carrying an unusual share of activity. Deal sizes will be smaller than a metropolitan comparison suggests and sales cycles longer, because buyers are cautious in a way that survey data captures poorly. Against that, competitive intensity in specialist digital services is very low, and the practical alternative most Caledonian firms face is doing nothing rather than choosing between vendors.

None of those three is a bad market. All three are markets where the honest number is smaller than the number someone else will quote you.

Tessar Napitupulu writes about sizing and measuring markets where the standard commercial datasets do not reach in Found Before They Search, available as a free gated edition, with retailer editions on Amazon, Google Play and Apple Books. The sourcing discipline described here is what our SEO service for New Caledonia is built on, and it is why we will not put a traffic forecast in a first proposal.

---

Frequently Asked Questions

What is the current population of New Caledonia?

264,596, from the ISEE census enumerated between 22 April and 28 May 2025, with first results announced in Noumea on 29 July 2025. That is down 6,811 from the 271,407 counted in 2019, an annual rate of about -0.45%. Natural increase of roughly 11,000 was more than cancelled by a migration deficit of roughly 18,000. It is the first population decline recorded between two censuses since 1946.

Why do some sources say the population is 296,000?

Because a modelling denominator got mistaken for a count. A connectivity dataset reported 243,000 internet users at 82.0% penetration, and 243,000 divided by 0.82 is roughly 296,000. That is the population figure the model used internally, not a measured one. It matters because 296,000 is about 12% above the census figure, so every audience estimate and revenue projection built on it inherits the error. Use 264,596 and cite ISEE.

How far did the economy actually fall?

Real GDP contracted 13.5% in 2024, per CEROM, the joint ISEE, AFD and IEOM partnership, in its rapid economic accounts published September 2025. For 2025 the published projections diverge, from a further fall of around 8% to around 13%, the latter implying a cumulative decline near 28% against 2023. Both are attributed and recent, so the honest presentation is the range rather than an average nobody published.

Which private-sector job loss figure is correct?

They mostly measure different windows. The IEOM series against a stated base resolves it: 66,581 private salaried jobs in the first quarter of 2024, more than ten thousand lost over the following year, 60,550 remaining by the second quarter of 2025. The critical detail is the trajectory, with year-on-year decline easing through 2025 from 15.4% to 7.1% to 2% and finally 0.9%, and IEOM reporting stabilisation in the first quarter of 2026. Any figure quoted without its window describes a collapse that had already flattened.

Is business confidence recovering?

The indicator has risen substantially and the institution publishing it declines to call that recovery. IEOM's business climate indicator, computed from roughly 200 business leaders against a long-run average of 100, hit a historic low of 65.9 in the second quarter of 2024, reached 88.2 a year later, and stood at 99.3 in the first quarter of 2026. IEOM characterises the movement as stabilisation at a heavily degraded level in a context of substantial public support. Quote the caveat with the number.

Did internet use fall along with the economy?

No, and this is the most consequential exception. Penetration stood at 82.0% with roughly 243,000 users as of October 2025. The audience is smaller and under more financial pressure than in 2023, but it did not go offline. Handle the two figures carefully though: do not divide 243,000 by the 264,596 census count to compute a penetration rate, because the connectivity dataset uses its own denominator and that division is exactly how the discredited 296,000 figure entered circulation.

Is this a viable market for a foreign digital agency?

It depends entirely on which of three markets you mean. Selling to residents is a share-capture proposition against a ceiling of 264,596 people with pressured household finances, viable because most local competitors have not adapted, but not honestly describable as growth. Selling to visitors puts your addressable demand outside the territory, so domestic contraction does not cap it and tourism arrival data replaces the census as your planning source. Selling to businesses means a base near 55,000 private salaried jobs with smaller deals and longer cycles, offset by very low competitive intensity in specialist digital services.

How much does nickel still decide the outlook?

Enough that a growth projection for this territory is really a nickel assumption in disguise. The readings are mixed: ore extraction rose about 24% in 2025 and Eramet reported SLN mine output up 10% in the first half of 2026, but KNS has been in care-and-maintenance since September 2024. The price signal needs care too, since London Metal Exchange nickel rose about 11.5% in the first quarter of 2026 while exchange stocks rose 50.9% year on year. Rising prices alongside rising inventories is not demand outrunning supply. Nickel is the mechanism by which the territory either grows out of its position or stays dependent on transfers, and the 9 billion francs committed for 64 projects in 2026 is bridging finance rather than a permanent base.

Are household finances still deteriorating at the rate the headlines suggest?

No, and the most quoted figure is the one most likely to mislead about today. Over-indebtedness filings did rise 159% year on year in IEOM's 2025 annual report published 23 June 2026, with New Caledonia accounting for 52% of filings across the New Caledonia and French Polynesia commissions, roughly 264 cases, against 18% in 2019. But by the June 2026 conjunctural note the increase had slowed to 23.7% year on year, with credit repayment incidents up 19.8%. Distress is still deepening, at roughly a seventh of the 2025 pace. Two related figures are worth carrying: doubtful debts reached 12.4% of total credit extended in 2025 against 11% a year earlier, and 29% of surveyed firms still feared their own failure at the end of 2025 against about 18% pre-crisis, even as realised business failures fell 30% from 2022.

---