Measure 11.01 — Capping some social benefits paid in cash: what would it really save?

Starting point: separate cash caps, in-kind support and actual budget savings

Measure 11.01 proposes to cap the share of some social benefits paid in cash and replace part of it with support earmarked for essential needs. The issue is sensitive because it directly affects the disposable income and autonomy of low-income households. It therefore requires more precision than a slogan such as “social card instead of cash”. Three different policy goals may be pursued: ensuring that a defined share of support is spent on essential needs; restricting certain uses or reducing some forms of abuse; and reducing public expenditure. Those three goals are not interchangeable. A reform can achieve the first without achieving the third.

Delta-Sierra’s audit identified a major internal inconsistency in historical versions of the Plan. The index attributed €5–8 billion of annual savings to measure 11.01, while the detailed section used a much lower range of €650 million to €1.3 billion before subtracting the value of in-kind support that would replace part of the cash. That discrepancy is not a minor drafting issue. It means that no amount from this measure should be added to the Plan’s consolidated savings until the benefits covered, the cash cap, the compensating entitlements and the implementation costs have been rebuilt on a consistent basis.

How the RSA already works: a means-tested benefit, not a flat cash amount for everyone

The French revenu de solidarité active (RSA) provides a concrete example. From 1 April 2026, the family-allowance system publishes a standard monthly amount of €651.69 for a single person without children. The amount changes with household composition and is reduced by other household resources and a housing allowance factor. In simplified terms, the RSA formula is the standard amount minus other resources and the housing forfait. It is therefore wrong to model reform as though every recipient automatically received the maximum amount in cash.

The Social Action and Families Code also provides that household resources are taken into account subject to the relevant exceptions, including certain benefits in kind. Converting part of a monetary benefit into earmarked or in-kind support therefore interacts with existing rules on resources, housing, household composition and benefit combinations. A badly designed reform could create thresholds, double counting or accidental losses of entitlement.

The scope of measure 11.01 consequently has to be defined benefit by benefit. RSA does not serve the same purpose as disability benefits, housing assistance, family benefits or local emergency support. Some payments compensate for a specific disability or cost and may be poorly suited to further spending restrictions. A serious law should therefore contain an explicit schedule of included and excluded schemes rather than referring vaguely to “all social benefits”.

What earmarked support could look like without turning welfare into constant surveillance

Technically, the State could place a defined share of a benefit into an account or payment card on which some categories of expenditure are allowed or restricted. Australia operates Income Management arrangements using a SmartCard for some participants: funds can be used for essential household needs while certain products and transactions are restricted. The Australian experience proves that such a payment mechanism can be built. It does not prove that it is socially desirable for France, that it saves money, or that outcomes would transfer from one country to another. It is a technical comparator, not evidence of French policy effectiveness.

If France chose such a mechanism, privacy and dignity would have to be designed in from the start. The card should not report a detailed list of purchases to government when category-level authorisation is sufficient. Merchants should not learn more about the person’s social status than technically necessary. Wrongly blocked payments must be contestable quickly. Emergency fallbacks would be needed for card loss, outages, domestic violence, urgent expenses and rural areas where compatible merchants are sparse.

Who decides, pays, implements and audits?

Who decides? National benefits are governed by legislation and regulation according to their design. A substantial change to payment conditions would require the appropriate legal text and review of fundamental-rights implications. Who pays? Benefit bodies continue to finance entitlements, while the State or administering bodies also pay for the earmarked-payment infrastructure. Who implements? Family-allowance funds, other benefit bodies, the payment operator and social services, depending on the chosen scope. Who audits? Parliament, the Cour des comptes, data-protection authorities, competent courts and appeal bodies.

The correct calculation starts with cash removed and subtracts everything recreated

Costing must be based on microsimulation because the outcome depends on millions of different household situations. For each household, the model needs the current benefit, the share that would be capped, the value of replacement entitlements, the technical cost of the card or account, changes in benefit take-up and transition expenses. National net savings are then the sum of household-level effects plus or minus central administrative costs.

A worked example — deliberately not a forecast for France

Suppose, only to explain the formula, that a reform reduces cash payments by €1 billion. Suppose it creates €700 million of in-kind entitlements, costs €80 million a year for payment infrastructure, control and support, and increases spending by €100 million because some households make fuller use of their entitlements. Net saving would be: 1,000 − 700 − 80 − 100 = €120 million. This does not estimate the French reform. It demonstrates why a multi-billion-euro saving cannot be inferred from the gross amount of cash that is capped.

Fraud has to be modelled separately

An earmarked card can restrict some purchases or cash withdrawals, but that does not mean it addresses all social-benefit fraud. Fraud concerning eligibility, income, residence, identity or household composition occurs before payment and requires data checks, investigations and due process. Mixing “control of how a benefit is spent” with “fraud in obtaining the benefit” would either double-count savings or promise an effect the card cannot deliver.

Four terms must also be kept separate: fraud detected, amount reassessed, expenditure prevented and money actually recovered. A debt notice may be challenged, cancelled or never collected. For a permanent reform to be financed honestly, only expenditure actually avoided or revenue actually collected, net of control costs, belongs in the fiscal calculation.

Social risks and unintended effects

The first risk is stigma. A visibly identifiable “welfare card” could reveal a recipient’s circumstances in public. The payment instrument should therefore look like an ordinary card and disclose as little as possible to merchants. The second risk is mismatch with real life. A recipient may need to pay for a small repair, a school activity, an emergency service or a supplier that does not accept the card. A meaningful unrestricted share and a fast exception process are therefore necessary.

The third risk is territorial inequality. A restriction that is easy to use in a large city may become punitive in a rural area with limited merchants and transport. The fourth risk is domestic control: an overly rigid system could reduce autonomy for victims of abuse. The fifth risk is administrative cost. If IT systems, calls, appeals and merchant support consume much of the gross saving, the reform becomes another layer of bureaucracy rather than a simplification.

Pilot first, nationwide rollout later

A reform of this kind should begin with a limited experiment, voluntary where appropriate or otherwise tightly framed in law, focusing on one benefit and several representative territories. The evaluation should measure the full operating cost, blocked-payment incidents, take-up, essential spending, hardship, user satisfaction and fraud effects that actually fall within the mechanism’s scope. A control group or equivalent evaluation method is needed to separate the effect of the programme from general changes in income and prices.

National rollout should occur only if three conditions are met: the social objective is demonstrably improved, rights are adequately protected and the net cost is acceptable. If the main effect is to guarantee essential spending while fiscal savings are small, government should say so explicitly. That would be a conditionality and protection reform, not a multi-billion-euro deficit-reduction measure.

What is known, hypothetical and contradictory today

Known: the basic RSA rules and 2026 amounts, and the technical existence of earmarked-payment mechanisms abroad. Hypothetical: the acceptable cash cap, the cost of a national system and behavioural responses. Contradictory in the historical Plan: a €5–8 billion range existing alongside a €650 million–€1.3 billion range before replacement support is deducted. That contradiction must be resolved before any consolidated saving is claimed.