ECONOMIE
Operation ATOM: What ANAF Is Testing on the Second-Hand Car Market
The General Directorate for Fiscal Anti-Fraud launched, on Thursday, October 1, 2026, Action ATOM, which in its first stage targets 35 second-hand car dealers with a stock valued at over 38 million euros.
The stakes go beyond the 35 companies under control. Romania's second-hand car market is one of the largest in the European Union by the number of registrations of imported used vehicles, and transactions in this sector have repeatedly been associated, in fiscal analyses, with risks of VAT evasion and fictitious invoicing.
The thesis of this article is verifiable: the operation will be a test of ANAF's capacity to turn large-scale declarative controls into actual revenue recoveries, and the result will be visible in the institution's press releases and reports in 2026–2027. The dominant scenario, with an estimated probability in the 30–45% range, is the continuation of the operation with additional phases; the alternative, stopping at the first stage, has a band of 25–40%.
1. What happened: the timeline and scope of the operation
ANAF announced on October 1, 2026 the launch of "Action ATOM – Analysis and Traceability of Car Operations," according to a press release picked up by Juridice.ro. The controls are being carried out nationwide and are executed by the General Directorate for Fiscal Anti-Fraud (DGAF), ANAF's specialized unit in charge of tax fraud and customs fraud. The targeted scope is explicit: economic operators that sell or intermediate the sale of second-hand cars.
The selection of companies in the first stage was made on the basis of risk analysis, according to the same press release. Antena 3 CNN and Digi24 reported on the same day, October 1, that 35 companies involved in the second-hand car trade are targeted. Economedia added the stock detail: cars valued at over 38 million euros.
The ANAF press release, as reported by the press, does not yet specify the legal classification of potential violations, the names of the companies under control, or the estimated value of the fiscal damage. These elements typically appear only after the controls are completed, through tax assessment decisions or criminal referrals. Until then, the public figures remain those of the scope: 35 entities and a stock of over 38 million euros.
It is important what the operation is not. ATOM is not a police raid and does not involve, in the form described by the press release, searches or detentions. It is a fiscal control and operations-traceability action, meaning the verification of the documents through which the cars entered the country, were purchased, sold, and invoiced. The difference matters for what can result from it: fiscal sanctions, not automatically criminal files.
2. The mechanism: why the second-hand trade attracts anti-fraud controls
The causal chain that explains the operation has three links: the structure of the market, its fiscal vulnerability, and the budgetary pressure on the revenue collector. The first link is the market structure. Romania has for years imported a high volume of used cars from Western Europe, and the second-hand trade operates through a network of low-capital companies, quickly set up and dissolved, which buy lots of cars at auction and resell them individually. The second link is the fiscal vulnerability of the scheme.
The classic risk in this sector is VAT carousel fraud: a company imports cars, sells them with VAT included, but does not remit the tax to the budget, then disappears. A more discreet variant is fictitious invoicing, whereby non-existent expenses or sales are documented on paper to reduce the taxable base. DGAF exists precisely for such schemes, and the traceability of operations — where the car came from, who paid for it, who invoiced it — is the standard verification tool. The third link is budgetary pressure.
Romania operates a high budget deficit relative to European requirements, and revenue collection is one of the levers the Government can pull without new legislation. The three links connect as follows: a market structured around volatile companies, a fiscal scheme exposed to carousel fraud, and a collector under pressure to increase revenues — the combination produces an operation of the ATOM type. The counter-hypothesis must be weighed just as seriously.
It is possible that ATOM is primarily a routine compliance action, with no significant fiscal damage discovered: risk-based selection also produces false positives, and a substantial percentage of the controlled companies may come out without major findings. The available data do not yet allow establishing which of the two explanations is correct; the press release describes the scope of the action, not its results. Both hypotheses are compatible with the facts known as of October 2, 2026.
3. The actors: who acts and who pays
The central actor is the National Agency for Fiscal Administration (ANAF), the institution that collects the state's revenues and which officially announced the operation. The executive role belongs to the General Directorate for Fiscal Anti-Fraud (DGAF), the structure under ANAF specialized in combating tax and customs fraud, which is carrying out the controls nationwide. The two institutions appear in all the reports on the operation, from Juridice.ro to Profit.ro.
On the other side are the 35 controlled entities: second-hand car dealers and intermediaries. The available sources do not name the companies and do not specify their individual size; the only aggregate figure is the stock of over 38 million euros, i.e., an average of approximately 1.1 million euros of stock per entity — a simple calculation that says nothing about the actual distribution among companies. For these companies, the immediate costs are administrative: management time consumed with the controls, the temporary blocking of some stocks pending clarification of documents, and possibly taxes and penalties if findings emerge.
The third circle of actors consists of end buyers and the labor market. If the checks increase dealers' compliance costs, part of the cost may be passed on to the price of second-hand cars, and smaller companies with limited financial reserves may exit the market. This effect is inference, not finding: the available data do not yet show any price movement or sector consolidation.
The sources do not say who may have filed denunciations or referrals that would have underpinned the selection of the companies; the press release describes the selection as made on the basis of internal risk analysis. It is also not known who, at this moment, is the injured party: until a finding of damage, the operation is a verification, not a case file.
4. What we cannot know yet and what should make us skeptical
The number of 35 entities and the stock of 38 million euros describe the first stage, not the market. Romania has thousands of economic operators active in the second-hand car trade, and an action targeting 35 of them covers a small fraction of the sector. There is also a risk of overweighting visibility. A publicly announced operation, with a press release and reports from several newsrooms, produces a signaling effect that can be confused with a collection effect.
The signal is real and can change the behavior of some operators; actual collection, however, depends on what the controllers find in the files of the 35 companies and on ANAF's capacity to turn findings into sums collected. The two are not the same thing, and the recent history of large-scale operations announced with fanfare, with results published late and partially, recommends caution. A third point of skepticism concerns selection.
Risk analysis is a legitimate method, but it is not neutral: the parameters it relies on — differences between VAT collected and remitted, transaction volume, company age — can flag both real evaders and companies with unconventional but legal accounting. Without the subsequent publication of the criteria and of the results by segment, evaluating the quality of the selection remains impossible from the outside.
5. Scenarios: what follows after the first stage
The dominant scenario, with an estimated probability in the 30–45% range, is escalation through additional phases. The driver is the proven operational capacity of the Anti-Fraud directorate in previous national operations, plus the budgetary pressure on collection. The constraint is DGAF's human resources, which limit the number of companies checked simultaneously. The consequence: hundreds of dealers checked in 2026–2027, with a corresponding increase in compliance in the sector. The indicator to watch: ANAF press releases announcing a second stage or an expansion of the scope.
The plausible scenario, in the 25–40% band, is stabilization at targeted control: ATOM ends with moderate results on the 35 entities, without expansion. The driver is the reallocation of the Anti-Fraud directorate's priorities to other sectors; the constraint, execution capacity. The consequence: a limited effect on the market, but a localized compliance effect on the controlled companies. The indicator: the absence of expansion press releases in the first two quarters of 2027.
The tail scenario, below one in four, is the legal blockage and the legal challenges: control measures challenged in court, stocks unblocked through court rulings, delayed results. The driver is the legal aggressiveness of the targeted companies; the constraint, the evidentiary quality of the control files. The indicator: case files with ANAF as a party on the portal.just.ro portal, filterable by the subject of fiscal challenges.
The improbable scenario, below one in ten, is a regulatory reform of the second-hand car market triggered by the operation — for instance, stricter rules for intermediaries or extended mandatory traceability. This would require legislative intervention and political, not merely administrative, will, and the signal would be a bill submitted to Parliament in 2027.
6. Verifiable predictions
| Prediction | Horizon | Probability | How to verify | Verification source |
|---|---|---|---|---|
| ANAF announces a second stage of Action ATOM, with an expanded scope | 31.12.2027 | 30–45% | Existence of an official expansion press release on ANAF's website | anaf.ro, press releases section |
| ANAF publishes aggregate results of the ATOM controls (additional taxes, fines) | 30.06.2027 | 40–60% | Press release or activity report with figures on ATOM | anaf.ro |
| Fiscal challenges by the controlled companies reach the courts | 31.12.2027 | 10–25% | Case files with ANAF as a party, subject: fiscal challenges | portal.just.ro |
| At least one of the 35 companies becomes fiscally inactive or enters insolvency as a result of the controls | 31.12.2027 | 15–30% | Companies' fiscal status; insolvency files | anaf.ro, portal.just.ro |
| Legislative bill on regulating the second-hand trade | 31.12.2027 | 5–15% | Monitorul Oficial; legislative initiatives | cdep.ro |
Method, definitions and figures
The figures used in this article come exclusively from the journalistic sources listed: the ANAF press release as reported by Juridice.ro, Antena 3 CNN, Digi24, Economedia, and Profit.ro. The figure of 35 entities designates the economic operators selected in the first stage of Action ATOM, and the figure of over 38 million euros designates the value of these entities' car stock, as assessed by ANAF and reported by the press.
The average of approximately 1.1 million euros per entity is our own calculation (38 million divided by 35) and describes an arithmetic mean, not the actual distribution of the stocks. The probabilities in the scenarios section and in the predictions table are editorial calibration intervals, not measurements. The terms "dominant scenario," "plausible," "tail," and "improbable" are verbal confidence categories, not severity labels.
They are not comparable to each other: the value of the stock (an inventory assessment) and possible future additional taxes (a fiscal recovery) measure different things. Also, the press data are reiterations of the ANAF press release, not independent verifications of the figures; any official revision of the figures would take precedence.
Limitations of the analysis
The analysis cannot establish, at the time of publication, whether the controlled companies committed fiscal violations, what the magnitude of any damage might be, or whether the selection of the 35 entities was accurate. The primary data are missing: the full ANAF press release with the selection criteria, the results of the controls, and the names of the companies. Everything described here as mechanism is plausible inference, not proven finding.
The diagnosis can be invalidated in two ways: if ANAF publishes neither any additional stage nor any aggregate result by mid-2027, the escalation thesis falls; if press releases with substantial fiscal results appear, the counter-hypothesis of a routine control weakens significantly. A re-evaluation should be triggered at the first ANAF press release about ATOM after October 1, 2026, or upon the appearance of case files on portal.just.ro with ANAF as a party.
The most fragile hypothesis of the analysis is the link between budgetary pressure and the timing of the operation: it is coherent with the context, but it is not supported by an official document stating it. Asserting this link depends on interpretation, not on a primary source.
