The president of the Seine-Saint-Denis departmental council, Stéphane Troussel, contacted the Defender of Rights about the possibly discriminatory nature of an algorithm used by the National Family Allowance Fund (Cnaf) to target its controls, according to a letter consulted Wednesday by AFP.
“It turns out that the criteria used by the algorithm have an obvious discriminatory impact towards the most vulnerable populations and that this is all the more important in a territory which combines socio-economic difficulties such as the Seine-Saint-Denis,” writes the elected socialist in a letter to Claire Hédon, dated Tuesday.
Used since 2011
Used by the Cnaf since 2011, this statistical tool aims to identify among the 13.5 million beneficiaries most likely to make errors in their declaration, the director general of the Cnaf, Nicolas Grivel, explained to AFP at the end of November. .
Beneficiaries of certain social minimums – such as the RSA or the activity bonus –, whose income often varies, must complete quarterly income tax returns with complex forms, and are more at risk of making errors, Cnaf indicated. . According to her, this algorithm aims to identify these beneficiaries to carry out checks quickly and rectify errors.
“Suspicion score”
The La Quadrature du net association denounced in November a higher “suspicion score” for people with low incomes, the unemployed, recipients of minimum social benefits or residents of “disadvantaged” neighborhoods.
According to this association, the risk of being checked following an event such as a move would be greater for a mother without income than for a wealthier couple.
“It is clear that criteria that are nevertheless protected, such as age, disability or family situation, are used in the algorithm, and that this results in more systematic targeting of the most vulnerable populations,” writes Stéphane Troussel in his letter. He wants to know the number of checks carried out in Seine-Saint-Denis, and wonders about “excessive targeting”.
“Essential” controls, says the government
The Cafs “do not use an algorithm to monitor recipients but rather to identify files that may present a risk of error and subsequently involve difficulties for the recipients concerned,” declared the minister responsible for disabled people on Wednesday. during questions to the government in the Assembly. “For Cnaf, it is essential to carry out checks, especially since certain services are subject to numerous reporting errors,” added Fadila Khattabi.
“By nature, the Cafs pay a greater number of aid to the poorest people, those most in difficulty, and therefore it is logical that these people are over-represented among the risks of error,” she added.
Social assistance fraud represents a shortfall of several billion euros. For social benefits alone, it is estimated between 6 and 8 billion per year according to the Court of Auditors. The government unveiled a plan in May aimed at making savings and doubling adjustments by 2027.