PRESS RELEASE
15/11/2023
Citizen-driven report
DETECTION OF TAX FRAUD
BY INDIVIDUALS
In 2022, the Court of Accounts held its first public consultation to gather proposals for
investigations to enrich its work programme.
Following nearly 9,000 registrations and 333 online
contributions, six topics were investigated in 2023.
The detection of personal tax fraud is the
fourth citizen-driven report to be published.
The success of the second campaign in 2023 (over
20,000 participants and 622 contributions) demonstrated the public’s interest in this tool.
The new
themes selected by the financial jurisdictions will be presented on 3 January. The citizens’
contributions at the origin of this survey justified this proposal by "the need to check that
the administration has effective means of detecting tax fraud by individuals", a sign of the
French people’s attachment to tax civic-mindedness and the principle of equality before the
tax. In accordance with the request made to the Court, this report deals only with taxes paid
directly by individuals. In 2022, this amounted to more than €160 billion. Effectively
combating tax fraud is not only one of the conditions for tax compliance, but also a
budgetary issue, given that the total amount recovered by the tax authorities after an audit
amounted to €14.6 billion in 2022, of which just over a third concerned private individuals.
Implementing a strategy to detect tax irregularities
While it was largely based on the systematic three-yearly audit of the largest taxpayers in terms
of income and assets, the scheduling of tax audits - the only way of identifying and punishing
fraud - has undergone a ‘technological revolution’ over the past decade. The digitisation of
virtually all tax processes and the increase in the volume of data available to the tax authorities
have led them to develop powerful automatic processing tools to identify inconsistencies and
anomalies in tax returns. Nearly a third of tax audits are now based on risk analyses fed by mass
data processing (data mining). Other experiments, such as the innovative land management
system that relates elements declared for property tax purposes to public collections of aerial
views, bear witness to the growing use of artificial intelligence technologies. The DGFiP is also
committed to stepping up the collection of information, for example from "tax informants", to
identify sophisticated fraud and new fraud schemes. As technologies designed to detect tax
irregularities in individuals are potentially intrusive, they have resulted in new provisions to
protect taxpayers’ rights, as the Constitutional Council has tightened the scope of the tax
authorities’ powers. The cross-referencing of mass data can be used as an inspection aid, but
can never be used to automatically identify fraud.
A technological strategy the effectiveness of which remains difficult to measure
Despite the successful deployment of these new techniques, their effectiveness remains
difficult to measure. Unlike many countries, France does not calculate a rigorous statistical
estimate of tax fraud, with which the volume of irregularities detected and the amount of taxes
recovered can be reconciled (€14.6 billion in 2022, including €2.6 billion in penalties). It is
therefore impossible to assess what proportion of fraud is actually detected and punished, and
whether this proportion has increased with the use of digital technologies. This is a persistent
and regrettable shortcoming, which the Court has already criticised. The tax authorities must
also catch up with their foreign counterparts by estimating the amount of fraud using tried and
tested statistical methods. The DGFiP plans to complete such estimates by 2027 for business
taxes, but there is not yet a programme of work for personal taxes. The proportion of tax audits
of individuals that lead to a tax adjustment, a measure of the relevance of the audit programme,
currently stands at 55
%. This is a high rate, but one that has barely improved since 2018, which
also excludes the possibility of concluding that there has been a qualitative leap as a result of
the ‘technological revolution’ of 2013-2022. Furthermore, the shortcomings of the tax
authorities’ IT systems make it impossible to establish a link between the reasons for scheduling
audits and the adjustments actually made.
A more structured and transparent strategy for detecting tax irregularities to be
implemented in conjunction with the national anti-fraud plan for 2023
There are several possible approaches to detecting tax irregularities. They may systematically
target large taxpayers or focus on randomly selected cases, at the risk of devoting resources to
cases where there are no irregularities; they may also be based on risk analysis, in order to
maximise the yield from the audit. It is this second approach that is implemented, leaving a
large degree of autonomy to the local departments, although the principles and procedures are
not set out. Formalising a more structured and transparent strategy for detecting tax
irregularities would make it possible to allocate the resources assigned to audits on the basis of
clearly formulated objectives. The national anti-fraud plan presented by the French government
in June 2023 is an opportunity to formalise such a strategy. The Court has made six
recommendations in this regard, relating to the estimation of fraud, the scheduling of audits,
improved monitoring tools, greater mobilisation of tax intelligence, the introduction of a
proactive approach to prevention and deterrence, and better management of the professional
skills needed to combat tax fraud.
Read the report
PRESS CONTACT:
Julie Poissier
◼
Head of Press Relations
◼
T
+33 (0)6 87 36 52 21
◼
julie.poissier@ccomptes.fr
@Courdescomptes
ccomptes
Cour des Comptes
Cour des Comptes