Every county but one, from Tulcea to Timiș, registered fewer new businesses this August than last. The national headline hides nothing — because there is nowhere left to hide.

The comfortable theory going around analysis desks this month was that Romania’s August business contraction must be a Bucharest-and-friends affair — a capital-city hangover dressed up as a national trend. Poke the data, and the theory collapses on contact.

Of the 42 counties that report registrations, 41 posted a year-over-year decline in new business formation in August 2026. The lone exception is Vaslui, which grew from 80 to 84 — four extra businesses. Not “some.” Not “most.” Forty-one of forty-two. The shallowest decline belonged to Teleorman, down 11.11%, with Bucharest itself a hair behind at 11.13%. The deepest belonged to Caraș-Severin, down 68.6% — from 121 new registrations in August 2025 to just 38.

This is the finding that matters: the working hypothesis — that the national net loss is geographically concentrated — is false, and demonstrably so. The damage is not concentrated. It is total.

The national figure, briefly

Romania recorded 8,103 new business registrations in August 2026. Against 8,744 exits — 1,015 suspensions, 3,198 dissolutions and 4,531 deregistrations — that produced a net loss of 641 businesses and a churn rate of 107.91 exit events per 100 registrations.

August is Romania’s third-weakest calendar month by historical mean, averaging 10,681 registrations, so some softness is baked in. But this August came in roughly 25% below even that weak baseline. Seasonality does not explain a 3,332-registration year-over-year drop. It does not explain Caraș-Severin losing two-thirds of its formation activity.

The trailing 12-month average of registrations sits at 12,627 — more than half again the August figure itself, a gap that shows how far this month fell below the recent norm.

Where the losses actually fall — and why “concentration” is the wrong frame

Here is the crucial nuance that both the alarmists and the minimizers will get wrong. The absolute losses are indeed concentrated: Bucharest alone accounted for 980 deregistrations and 674 dissolutions, the single largest block of exits in the country. The top five counties by volume — Bucharest, Ilfov, Cluj, Timiș, Iași — are exactly the five largest economic centers, and they are exactly where the raw exit numbers pile up.

But raw counts are a function of size, not of distress. The correct measure is the churn rate — exits per 100 registrations — and there the geography inverts completely.

Bucharest’s churn rate was 78.38 — the lowest in the country. The capital, in other words, is the least distressed large market in Romania, precisely because it still generates enough new formations to absorb its closures. The counties bleeding proportionally worst are small: Tulcea at 251.52, Bistrița-Năsăud at 210.96, Harghita at 178.43, Caraș-Severin at 168.42, Alba at 167.97.

So the national net loss of 641 businesses is not the sum of a few big local disasters. It is the aggregate of forty-plus simultaneous, individually modest local deteriorations — a thousand small cuts rather than one large wound. That is a worse diagnosis than concentration would have been, because there is no single lever, no single county authority, and no single policy fix that addresses it.

The structural driver is hiding in the formation mix

The reason the breadth matters is that it points to a cause that operates everywhere at once. Nationally, SRL (limited liability company) registrations collapsed 41.95% year-over-year, from 7,613 to 4,419, while PFA (sole trader) registrations fell only 1.9%. The incorporated-company engine that drives real employment and capital formation is failing far faster than the sole-trader form — and that is a national phenomenon, not a regional one, which is exactly why it shows up in every county.

The pattern is visible even in Bucharest, where SRLs (1,185) only narrowly outnumbered PFAs (1,083) in August. In the capital’s most economically dynamic market, the incorporated sector is nearly at parity with informal self-employment. Everywhere else, the margin is almost certainly gone.

The one number that should worry the reader most

August’s net loss of 641 is the ninth consecutive net-negative month, a run that began in December 2025 and has now cumulatively destroyed 16,756 net businesses.

Note carefully what that streak is and is not: it is measured only from July 2024, the earliest month with trustworthy exit data, and it cannot be described as an all-time record. But there is a real bright spot worth stating plainly — August’s net loss of 641 is the second-smallest of the nine (only February’s 119 was smaller), improving from July’s 1,030 and from April’s brutal 3,018.

What this means

The temptation this month is to read the improving monthly sequence — from -3,018 in April to -641 in August — as a recovery. The geographic data says otherwise. A shrinking national deficit driven by simultaneous weakness in 41 of 42 reporting units is not healing; it is a broad, shallow hemorrhage that happens to be slowing. The prior-year comparison makes the point sharper: August 2025 registered 11,435 businesses and still managed far less churn than this August’s 8,103 against 8,744 exits.

Three facts should anchor any honest reading of this month:

1. Breadth, not depth. The loss is not concentrated. Every county but one is down year-over-year; the “national average” is a genuine description of almost every Romanian’s local reality, which is the opposite of the assumption this desk started from.
2. The big economies are the resilient ones proportionally. Bucharest, Ilfov and the large university cities post the lowest churn rates. The distress is a small-county phenomenon hidden inside a national number.
3. The structural engine is failing fastest. SRL formation is down 42% year-over-year. Sole-trader registrations barely moved. That is the pattern that reproduces in every county and every sector at once.

The warning is not that Romania’s business ecosystem is dying in one place. It is that it is weakening everywhere, and the aggregate headline is the only place where any part of it still looks manageable.