Sweden: From Disability Pension to Work-First — A Policy Experiment in Progress
The Swedish Reform Experiment
In 2008, Sweden enacted a far-reaching reform of its disability and sickness benefits. Time-limited sickness compensation (tidsbegränsad sjukersättning) was abolished and the eligibility criteria for permanent sjukersättning (sickness compensation, which replaced the former disability pension in 2003) were tightened sharply: work capacity must now be judged permanently reduced, for the foreseeable future and in all types of work. A time limit of 914 days was placed on sjukpenning (sickness benefit), and sickness-benefit claims are tested under a "rehabilitation chain" (rehabiliteringskedja) with assessments at fixed points — after day 180, a person keeps sickness benefit only if they cannot do any work that normally occurs on the labour market.
The caseload results were dramatic and sustained. In December 2024 about 220,000 people received sickness compensation — 3.5% of the population aged 19–65, compared with 9.6% twenty years earlier. Försäkringskassan attributes the decline mainly to the 2008 changes: the number of recipients fell by 58% between 2004 and 2024.
The story is more complicated than the caseload figures suggest.
Where Did the Caseload Go?
Two official evaluations have examined what happens to people whose sickness-benefit claims are rejected.
Riksrevisionen (Swedish National Audit Office), RiR 2020:12, "Vägen till arbete efter nekad sjukpenning" ("The road to work after a denied sickness benefit claim"): this 2020 audit examined 556 sickness-benefit cases closed between October 2018 and March 2019. It found that just under half of those denied sickness benefit were back in their ordinary work, at their ordinary hours, 6–12 months after the decision — and that many went on to a new period of sick leave.
ISF Rapport 2023:4, "Utan inkomst efter avslag" ("Without income after denial"), from the Swedish Social Insurance Inspectorate, followed everyone denied sickness benefit between December 2018 and December 2019. Of 51,119 people, 7% (3,690) had no income — neither from work nor from benefits such as unemployment insurance — in the months after the denial; living on social assistance alone counted as having no income. Most had income from work: 68% of those denied after a period on sickness benefit and 62% of those denied at their first application. After two years, roughly half of the no-income group still had no income. The risk was highest for people with weak labour-market attachment and for men; among those denied after a period on sickness benefit, the strongest risk factors were having been on sickness benefit for several years and having previously received activity compensation.
The Social Insurance Assessment Controversy
Rejections at the day-180 assessment rose until 2020, when close to 30% of continued sickness-benefit claims were refused after day 180. They then fell markedly in 2021, to just under 7% among women and just over 8% among men. Försäkringskassan links the fall partly to a rule change on 15 March 2021 that eased the exception for people expected to return to their employer within a year, and to the government's decision to temporarily halt assessments against "normally occurring work" for people in days 181–365 of the rehabilitation chain.
Regional variation in outcomes was substantial. Försäkringskassan's own Socialförsäkringsrapport 2022:1, covering January 2018 to June 2021, documented a gap of 25 percentage points between the county with the highest rejection rate after day 180 — Jämtland, at 34.3% — and the county with the lowest, Östergötland, at 9.1%. The report notes that such differences may reflect how the agency applies the law, but also how regional health care handles sickness certification. Either way, the outcome of an assessment depended partly on where in Sweden an applicant happened to live.
Sickness compensation itself also became harder to obtain. Försäkringskassan's own review describes the number of new grants as historically low in 2018–2021: inflow reached historically low levels in 2018 and stayed at around 9,500 people a year through 2021. It links the lower inflow after 2015 to the agency's work, begun in 2015, to apply the rules more correctly, and to the removal of the sickness-benefit time limit in 2016, which took away a natural point to test whether all rehabilitation options had been exhausted.
What Works in Sweden: The Evidence Base
Swedish labour market policy has generated evidence on a handful of intervention types, though the reliable evidence base is thinner than is often assumed.
Samordningsförbund (Coordination Associations): multi-agency coordination bodies bringing together Försäkringskassan, Arbetsförmedlingen (Employment Service), municipalities, and healthcare. ISF's evaluation "Gör samordningsförbund någon skillnad?" ("Do coordination associations make a difference?", ISF Rapport 2019:4, published 15 March 2019), which studied the years 2005–2010, found mixed and gender-specific effects: in member municipalities the share of people on sickness benefit fell while the share on sickness and activity compensation rose, and more unemployed men found work than in non-member municipalities — but no such effect was found for women.
Supported Employment (IPS-model): the Swedish randomised evidence on Individual Placement and Support comes from Bejerholm et al. at Lund University, whose trial in Malmö ran from 2008 to 2011 with 120 participants with severe mental illness. At 18-month follow-up, 46% of participants receiving IPS had gained competitive employment, compared with 11% in traditional vocational rehabilitation — a substantial difference.
The Current Picture
In 2024, Sweden had:
- about 220,000 people on sjukersättning (sickness compensation, December 2024)
- 237,082 people who received a sjukersättning payment at some point during 2024
- 40,144 people who received aktivitetsersättning (activity compensation, for ages 19–29) during 2024
InkludX estimates spending on disability and sickness cash benefits at approximately €8.4 billion in 2024, or 1.5% of GDP — €4.0 billion on sickness compensation and €4.4 billion on sickness benefit. This covers public cash benefits only; it excludes employer sick pay, lost tax revenue and healthcare costs.
The Political Balance Sheet
Sweden's 2008 reform demonstrated that aggressive tightening of benefit criteria can reduce recorded caseloads rapidly and durably. It also demonstrated the limits of a "work-first" approach applied without adequate labour demand, employer support, or workplace accommodation infrastructure.
In InkludX's assessment: activation without accommodation produces churn, not employment. The people who most need labour market support — those with weak labour-market attachment and long benefit histories — are precisely those for whom a supply-side shock to benefit conditionality is least effective, and precisely the group most likely to end up with no income at all rather than a job.
Sweden's policy challenge for the 2020s is to build the demand-side infrastructure — supported employment at scale, employer incentive reform, workplace health services — that the 2008 reform assumed would follow automatically but did not.
Sources: Försäkringskassan, Korta analyser 2025:2, "Sjukersättningens utveckling över tid" (2025); Försäkringskassan, Socialförsäkringsrapport 2025:9, "Sjukersättning 2003–2023" (2025); Försäkringskassans årsredovisning 2024; Försäkringskassan, Socialförsäkringsrapport 2022:1 (2022); Riksrevisionen RiR 2020:12 (2020); ISF Rapport 2023:4 (2023); ISF Rapport 2019:4, "Gör samordningsförbund någon skillnad?" (2019); Bejerholm et al., "Individual Placement and Support in Sweden — A randomized controlled trial", Nordic Journal of Psychiatry (2015); InkludX, Fiscal cost of disability and sickness benefits — Sweden (2026).