Questions and limitations

Yes. Ruist's 2018 report reports refugees including their family members as a single group. The same applies to the National Institute of Economic Research (NIER) refugee category in 2025. Family members of labour migrants are generally closer to other foreign-born groups in net contribution.

No, not in any substantial way. Neither Ruist 2018 nor NIER 2025 includes a separate item for police, the justice system, corrections or indirect costs of crime linked to origin. The direct costs of the justice system are part of general public expenditure but are not allocated specifically between groups.

Yes, direct public expenditure is largely included — introduction benefit, social assistance, housing allowance, healthcare, schooling, Swedish-language tuition, care services, pensions and more. The items are allocated at individual level using register data. This is what makes the refugee group clearly negative in both studies.

Many of the heaviest costs (social assistance, schooling, social services, reception) fall on the municipalities, while large revenues (central income tax, VAT, employer contributions) go to the state. The national aggregate hides this imbalance — a municipality with a high share of newly arrived residents can have a strongly negative net position even when the country as a whole looks more balanced.

Ruist 2018 focuses specifically on the refugee group and calculates over a lifetime horizon (SEK 74,000 per year in average net cost per refugee, range SEK 53,000–94,000). NIER 2025 covers all foreign-born residents from 1983 to 2022 and reports both annual cross-sections and lifecycle NPV. The studies use different boundaries and different base years — they do not contradict each other, they complement each other.

A lifetime average is more stable than a single year's cross-section. NIER's +SEK 6bn for 2022 is affected by the fact that refugee inflows were historically low that year and the labour market strong — an unusually favourable year. Ruist's SEK 74,000 per year describes the pattern across an entire life, from settlement to retirement.

Indirect effects on wages and employment among the native-born, crime, the second generation, dynamic effects on the labour and housing markets, and long-term impact on public investment. See the summary below.

All figures come from public primary sources: government agencies (Statistics Sweden, the National Council for Crime Prevention, the Migration Agency, the National Agency for Education, the National Board of Health and Welfare, the Police, the Election Authority, the National Institute of Economic Research, ESO, the National Audit Office), research institutes and international bodies (OECD, Eurostat, IMF, World Bank). Every figure links directly to its source. See /en/sources for the full list.

Every page shows 'Last updated' in its header. Data is updated when the primary source publishes new figures — some series monthly (reported crime), others annually (Statistics Sweden register data), a few with longer lags (lifetime studies). The full history is available at /en/changelog.

Different agencies measure different things using different definitions and time windows. 'Foreign-born' at Statistics Sweden is not the same as 'foreign background' at the National Council for Crime Prevention. 'Reported offences' is not the same as 'persons convicted'. We always state which definition applies to each figure. See /en/definitions.

When we combine several primary sources, run cohort simulations or projections, we do the calculation ourselves. Such figures are always labelled 'own calculation' with the formula, assumptions and sensitivity analysis stated — and the underlying raw data always comes from the primary source.

Some figures are not measured at all in official statistics (for example political affiliation by immigrant group), others are confidential at individual level, and some series have been discontinued (Statistics Sweden stopped publishing certain breakdowns in 2004). When data is missing, we say so.

A model is a simplification of reality built on assumptions. When we use a model (for example Ruist's lifetime net position, the NIER macro model) we show which assumptions apply, the range the outcome produces and what happens if the assumptions change. Models do not deliver 'the truth' — they deliver scenarios.

A projection extrapolates today's data given certain assumptions about the future (fertility, migration, employment). Statistics Sweden's population projection is the standard example. We always use official projections where they exist and state clearly when something is extrapolated.

Via /en/contact. All external contact goes through a public AnonAddy alias. State the page, the exact wording and the primary source you believe is more accurate. Confirmed errors are corrected openly and logged in /en/changelog.

We have a dedicated page: /en/cite. APA, Harvard, simple web citation and BibTeX. Citing, referencing and linking are always welcome. Sverigefakta's own texts and charts are not offered under a general Creative Commons licence — larger republication or commercial use requires written permission. Underlying source data remains subject to each source's own terms.

Yes, in editorial and academic contexts to the extent permitted by law. Use the caption: 'Source: sverigefakta.com, based on [primary source].' If the chart shows raw primary data without processing, credit the primary source first. Commercial use requires written permission — please contact us.