How France helped the unemployed start businesses that survived

In 1977, France began exempting unemployed people from the social security contributions that fall due the moment someone becomes self-employed, long before a new business earns anything. Four decades later the scheme, by then renamed ACRE, reached a peak of 639,400 people in a single year. The lesson was in the sequence. It removed the standing cost of entry before it offered anything else.

Take away the standing cost before you add anything else

The most transferable insight from France's programme is that the first barrier facing an unemployed founder is rarely capital or confidence. It is a fixed, recurring cost. The social charges arrive on day one of self-employment, whether or not a single invoice has been raised. The Aid for the Unemployed Creating or Taking over a Business (Aide aux chômeurs créateurs et repreneurs d'entreprise, ACCRE) cancelled that cost first, then layered cash and advice on top. Programmes that reverse the order push people into businesses that lose money from the first month.

Unemployment was treated as a gap to plug, not a base to build from

When ACCRE was introduced in 1977, the stated aim was to encourage unemployed people to 'create jobs for themselves'. The problem was specific. An unemployed person who wanted to start or take over a business faced the same social security regime as an established firm, with none of the revenue to carry it. The cost of becoming self-employed landed before the income did, and for someone already without a wage that gap was often enough to stop the idea before it started. France's answer was not to hand the founder money. It was to remove the charge that made the early months unaffordable.

What the programme did, and why it worked

The mechanism was a temporary exemption from the social security contributions an entrepreneur would normally pay. Eligibility was kept narrow and deliberately checkable. A person had to create or take over a business individually or as its principal owner, and could not have drawn on the scheme in the previous three years. Those using the simpler micro-entrepreneur legal form qualified only if they were already receiving unemployment benefit, or had been registered as a jobseeker for more than six months out of the previous eighteen.

Exempt on a taper, not a cliff

The exemption was means tested, but gently. A beneficiary whose professional income stayed below 75% of the annual social security ceiling, EUR 30,852 in 2022, received a total exemption. Income between 75% and 100% of that ceiling, up to EUR 41,136, drew a partial exemption that fell away in proportion rather than all at once. The relief ran for 12 months, and up to 24 months for micro-entrepreneurs who met certain conditions. A hard income limit would have penalised the founders who grew fastest in their first year, which is the opposite of what the scheme set out to reward.

Let the supports stack

The single most important design choice was that ACRE was built to combine with other measures, not to stand alone. A founder could hold the contribution exemption while also drawing on the New Support for the Creation or Take-over of a Business (NACRE), which gave personalised guidance before, during and after launch, and on the Aid for the Take-over or Creation of a Business (ARCE), which released 45% of a person's remaining unemployment entitlement as a lump sum to put into the business. People on solidarity welfare payments could keep that income during a transition of up to 12 months. The combination, not any single scheme, was the lever.

In 2019 the government opened the contribution exemption to the whole population and renamed it Aid to Business Creators and Buyers (Aide aux créateurs et repreneurs d'entreprise, ACRE). Uptake nearly doubled, peaking at 639,400 beneficiaries that December before settling to 373,200 by March 2021. The headline figure was never the point. What the evaluations measured was whether the businesses lasted.

Transferable lessons

Four lessons travel well from France to other employment and enterprise support programmes.

• Remove the recurring cost of entry before you offer cash or advice. The charge that lands on day one does more to deter an unemployed founder than the absence of a grant. Take it away first.
• Taper the means test instead of cutting it off at a hard line. A hard limit punishes the founders who grow fastest in year one. A proportional withdrawal keeps the incentive pointing the right way.
• Design supports to stack, and tell founders they can. France's gains in access to finance and survival came from people combining the exemption, the lump sum and the guidance, not from any one of them on its own.
• Measure survival and jobs, not enrolment. Beneficiary counts move with eligibility rules. What tells you the programme worked is how many businesses were still trading three years on, and how many people they employed.

What to watch out for

The approach has limits. The exemption works because France funds self-employment through social contributions in the first place. In a system where the early cost of starting sits somewhere else, in licensing, tax or insurance, the same relief would need to point at a different charge to have the same effect. The evidence also carries a quiet warning. Supported founders were more likely to report obstacles than unsupported ones, probably because guidance made them aware of skills gaps they would otherwise have missed. A programme judged on founder satisfaction rather than business survival could read that the wrong way and pull back the very support that was working.

Conclusion

What France's scheme shows is that the cheapest intervention, waiving a charge, can outperform the expensive ones when it comes first in the sequence. Of the unemployed people who started businesses with this support, 74% were still trading after three years, and each had created an average of 1.8 jobs. Designing that sequence, and proving it worked, is the part most teams find hardest to resource on their own. That is where outside support tends to pay back.