Skip to content

Crime · Sweden · Updated 2026-10-26

How does registered crime differ by migration background in Sweden?

In Brå's study of everyone aged 15 and over living in Sweden at the end of 2014, people born abroad were registered as suspects of an offence committed in 2015-2018 2.51 times as often as people born in Sweden to two Swedish-born parents and people born in Sweden to two foreign-born parents 3.21 times as often; after Brå adjusted for age, sex, disposable income, education and municipality type the ratios were 1.76 and 1.68, and studies using other periods and definitions put the unadjusted ratio for people born abroad between 1.81 and 2.5.

Every figure in this summary is calculated from the sources listed on this page.

Born abroad: times as often registered as a suspect (unadjusted)

Share of people born abroad registered as suspected of at least one offence, divided by the same share among people born in Sweden to two Swedish-born parents, with no adjustment. Sweden, people aged 15 and over registered as resident on 31 December 2014, Offences committed 2015-2018. Unit: ratio (1 = same as the reference group).

2.51ratio (1 = same as the reference group), 2018

60Trust scoreModerate certainty

Score breakdown

Quality30.3 / 40

How well the sources are made

Agreement13.3 / 30

How close the sources are to each other

Independence12 / 20

How many separate ways the number was measured

Freshness4 / 10

How recent the newest data is

How much each source counts

  • Brå 202157%
  • Brå 20057%
  • Bäckman 202129%
  • Adamson 20207%

Sources

SourceValueDefinitionQualityWeight
Brå report 2021:9, Misstänkta för brott bland personer med inrikes respektive utrikes bakgrund, Table 1Brottsförebyggande rådet (Brå) · Police and justice recordsRetrieved 2026-10-262.512018
Share of people born abroad registered as at least reasonably suspected (skäligen misstänkt) of an offence committed in 2015-2018 (7.99%), divided by the share among people born in Sweden to two Swedish-born parents (3.18%).

Includes suspicions that were later dropped. Population: 8,066,363 residents aged 15 and over on 31 December 2014; people not registered as resident are excluded.

Table 1, p. 47. Report page: https://bra.se/rapporter/arkiv/2021-08-25-misstankta-for-brott-bland-personer-med-inrikes-respektive-utrikes-bakgrund. English summary: https://bra.se/download/18.45e4b8e192705389a34c2b/1729515966490/2021_9_Registered_offending_among_persons.pdf.

4 of 5 checks
  • Yes: Method and definitions published. The report's method chapter sets out the register, the degree of suspicion counted and SCB's standard background groups.
  • Yes: Official statistics or peer-reviewed. Analysis by Brå, which keeps Sweden's official register of suspected persons, linked to SCB population registers; two external professors read a draft.
  • Yes: Full count or representative sample. Covers every resident aged 15 and over on 31 December 2014, about eight million people, not a sample.
  • No: Raw data downloadable. The linked microdata are not public; the figures are only in the report's PDF tables.
  • Yes: No known break or quality problem. One fixed four-year window with no break; people who moved to Sweden after 2014 are outside the population by design.
57%
Brå report 2005:17, Brottslighet bland personer födda i Sverige och i utlandet, Table 7Brottsförebyggande rådet (Brå) · Police and justice recordsRetrieved 2026-10-262.52001
Same reference group and register of suspected persons, but the population is residents born 1945-1981 registered at the end of 1996, and the window is five years (1997-2001).

Published to one decimal.

Table 7, p. 34.

4 of 5 checks
  • Yes: Method and definitions published. The report describes the registers, the population and the background groups.
  • Yes: Official statistics or peer-reviewed. Analysis by Brå of the official register of suspected persons linked to SCB population registers.
  • Yes: Full count or representative sample. Covers every resident born 1945-1981 registered in Sweden at the end of 1996, not a sample.
  • No: Raw data downloadable. No public microdata; the figures are only in the report's PDF tables.
  • Yes: No known break or quality problem. No break reported within the five-year window.
7%
Bäckman, Estrada, Nilsson & Sivertsson (2021), Nordic Journal of Criminology, Table 1Stockholm University, Department of Criminology · Police and justice recordsRetrieved 2026-10-261.812017
Convictions, not suspicions: share with at least one conviction (court judgment, summary sanction order or waiver of prosecution) between ages 15 and 24, cohorts born 1990-1993 who lived in Sweden throughout those ages.

"Immigrant" = born abroad. The methods section defines natives as born in Sweden with at least one Swedish-born parent (figure captions say Swedish-born parents). Men and women combined by RealStats from the published percentages and group sizes.

Table 1, cohort 1990-93, row ">= 1 conv." (https://www.tandfonline.com/action/downloadTable?id=t0001&doi=10.1080%2F2578983X.2021.1900038&downloadType=CSV).

4 of 5 checks
  • Yes: Method and definitions published. The article describes the registers, the cohort design and the group definitions.
  • Yes: Official statistics or peer-reviewed. Peer-reviewed article in the Nordic Journal of Criminology (open access).
  • Yes: Full count or representative sample. Whole birth cohorts from Sweden's population and conviction registers, not a sample.
  • No: Raw data downloadable. Register microdata are not public; the article publishes percentage tables.
  • Yes: No known break or quality problem. No break in the conviction register over the period reported by the authors.
29%
Adamson (2020), Migrants and Crime in Sweden in the Twenty-First Century, Society, Table 5Göran Adamson (University West), with Tino Sanandaji · Police and justice recordsRetrieved 2026-10-262.12017
Share of foreign-born people suspected on reasonable grounds of at least one offence in 2013-2017, relative to people born in Sweden to two Swedish-born parents, from Brå's suspect register and SCB's population register (a separate extraction from the same register Brå uses).

Age range not stated in the article. Not adjusted: the author chose not to adjust for socio-economic factors.

Table 5, "Total crime", row 2013-2017, column "Of which foreign-born".

1 of 5 checks
  • No: Method and definitions published. Groups are defined, but the article does not state the age range or exact population used.
  • No: Official statistics or peer-reviewed. Published as a feature article in the journal Society; peer review is not stated for it.
  • Yes: Full count or representative sample. Based on the national suspect and population registers rather than a sample.
  • No: Raw data downloadable. No public data; the figures are only in the article's tables.
  • No: No known break or quality problem. The author notes a backlog in the justice process may affect the most recent period, and that age groups and methods differ from the earlier Brå studies.
7%

Born abroad: times as often registered as a suspect (adjusted)

Share of people born abroad registered as suspected of at least one offence, divided by the same share among people born in Sweden to two Swedish-born parents, after standardising for age, sex, disposable income, education and municipality type. Sweden, people aged 15 and over registered as resident on 31 December 2014, Offences committed 2015-2018. Unit: ratio (1 = same as the reference group).

1.76ratio (1 = same as the reference group), 2018

48Trust scoreLow certainty

Score breakdown

Quality32 / 40

How well the sources are made

Agreement0 / 30

How close the sources are to each other

Independence12 / 20

How many separate ways the number was measured

Freshness4 / 10

How recent the newest data is

How much each source counts

  • Brå 202162%
  • Brå 20058%
  • Bäckman 202131%

Sources

SourceValueDefinitionQualityWeight
Brå report 2021:9, Misstänkta för brott bland personer med inrikes respektive utrikes bakgrund, Figure 7Brottsförebyggande rådet (Brå) · Police and justice recordsRetrieved 2026-10-261.762018
Same register and population as the unadjusted figure.

Brå standardised the group of people born abroad to the age and sex distribution, the shares with low disposable family income and low education, and the municipality types of the reference group. Standardised share suspected 5.61% against 3.18% for the reference group.

Report text p. 50 and Figure 7 p. 51; standardised shares in Tabellbilaga 1, Table B7, p. 11 (https://bra.se/download/18.366598ab19270941a1c3955/1729256818286/2021_Tabellbilaga%201%20delstudie%201.pdf). Table B24 (p. 28) gives 1.77 for the same row.

4 of 5 checks
  • Yes: Method and definitions published. The report's method chapter sets out the register, the degree of suspicion counted and SCB's standard background groups.
  • Yes: Official statistics or peer-reviewed. Analysis by Brå, which keeps Sweden's official register of suspected persons, linked to SCB population registers; two external professors read a draft.
  • Yes: Full count or representative sample. Covers every resident aged 15 and over on 31 December 2014, about eight million people, not a sample.
  • No: Raw data downloadable. The linked microdata are not public; the figures are only in the report's PDF tables.
  • Yes: No known break or quality problem. One fixed four-year window with no break; people who moved to Sweden after 2014 are outside the population by design.
62%
Brå report 2005:17, Brottslighet bland personer födda i Sverige och i utlandet, Table 10Brottsförebyggande rådet (Brå) · Police and justice recordsRetrieved 2026-10-262.12001
Residents born 1945-1981, suspected at least once in 1997-2001 (five years).

Standardised for sex, age, education and earned income (förvärvsinkomst); not for disposable income or municipality type. Published to one decimal.

Table 10, p. 40.

4 of 5 checks
  • Yes: Method and definitions published. The report describes the registers, the population and the background groups.
  • Yes: Official statistics or peer-reviewed. Analysis by Brå of the official register of suspected persons linked to SCB population registers.
  • Yes: Full count or representative sample. Covers every resident born 1945-1981 registered in Sweden at the end of 1996, not a sample.
  • No: Raw data downloadable. No public microdata; the figures are only in the report's PDF tables.
  • Yes: No known break or quality problem. No break reported within the five-year window.
8%
Bäckman, Estrada, Nilsson & Sivertsson (2021), Nordic Journal of Criminology, Table 2Stockholm University, Department of Criminology · Police and justice recordsRetrieved 2026-10-261.162017
Convictions, not suspicions: share with at least one conviction (court judgment, summary sanction order or waiver of prosecution) between ages 15 and 24, cohorts born 1990-1993, restricted to people whose parents had the lowest decile of work income in two of the three years at ages 15-17.

Holds childhood family income roughly constant; no adjustment for education or municipality type. The authors note this low-income selection is probably stronger among natives. Men and women combined by RealStats from the published percentages and group sizes.

Table 2, cohort 1990-93, row ">= 1 conv." (https://www.tandfonline.com/action/downloadTable?id=t0002&doi=10.1080%2F2578983X.2021.1900038&downloadType=CSV).

4 of 5 checks
  • Yes: Method and definitions published. The article describes the registers, the cohort design and the group definitions.
  • Yes: Official statistics or peer-reviewed. Peer-reviewed article in the Nordic Journal of Criminology (open access).
  • Yes: Full count or representative sample. Whole birth cohorts from Sweden's population and conviction registers, not a sample.
  • No: Raw data downloadable. Register microdata are not public; the article publishes percentage tables.
  • Yes: No known break or quality problem. No break in the conviction register over the period reported by the authors.
31%

Born in Sweden to two foreign-born parents: times as often registered as a suspect (unadjusted)

Share of people born in Sweden to two foreign-born parents registered as suspected of at least one offence, divided by the same share among people born in Sweden to two Swedish-born parents, with no adjustment. Sweden, people aged 15 and over registered as resident on 31 December 2014, Offences committed 2015-2018. Unit: ratio (1 = same as the reference group).

3.21ratio (1 = same as the reference group), 2018

49Trust scoreLow certainty

Score breakdown

Quality32 / 40

How well the sources are made

Agreement1.2 / 30

How close the sources are to each other

Independence12 / 20

How many separate ways the number was measured

Freshness4 / 10

How recent the newest data is

How much each source counts

  • Brå 202162%
  • Brå 20058%
  • Bäckman 202131%

Sources

SourceValueDefinitionQualityWeight
Brå report 2021:9, Misstänkta för brott bland personer med inrikes respektive utrikes bakgrund, Table 1Brottsförebyggande rådet (Brå) · Police and justice recordsRetrieved 2026-10-263.212018
Share of people born in Sweden to two foreign-born parents registered as at least reasonably suspected (skäligen misstänkt) of an offence committed in 2015-2018 (10.22%), divided by the share among people born in Sweden to two Swedish-born parents (3.18%).

Includes suspicions that were later dropped. Population: 8,066,363 residents aged 15 and over on 31 December 2014; people not registered as resident are excluded.

Table 1, p. 47. Report page: https://bra.se/rapporter/arkiv/2021-08-25-misstankta-for-brott-bland-personer-med-inrikes-respektive-utrikes-bakgrund. English summary: https://bra.se/download/18.45e4b8e192705389a34c2b/1729515966490/2021_9_Registered_offending_among_persons.pdf.

4 of 5 checks
  • Yes: Method and definitions published. The report's method chapter sets out the register, the degree of suspicion counted and SCB's standard background groups.
  • Yes: Official statistics or peer-reviewed. Analysis by Brå, which keeps Sweden's official register of suspected persons, linked to SCB population registers; two external professors read a draft.
  • Yes: Full count or representative sample. Covers every resident aged 15 and over on 31 December 2014, about eight million people, not a sample.
  • No: Raw data downloadable. The linked microdata are not public; the figures are only in the report's PDF tables.
  • Yes: No known break or quality problem. One fixed four-year window with no break; people who moved to Sweden after 2014 are outside the population by design.
62%
Brå report 2005:17, Brottslighet bland personer födda i Sverige och i utlandet, Table 7Brottsförebyggande rådet (Brå) · Police and justice recordsRetrieved 2026-10-2622001
Same reference group and register of suspected persons, but the population is residents born 1945-1981 registered at the end of 1996, and the window is five years (1997-2001).

Published to one decimal.

Table 7, p. 34.

4 of 5 checks
  • Yes: Method and definitions published. The report describes the registers, the population and the background groups.
  • Yes: Official statistics or peer-reviewed. Analysis by Brå of the official register of suspected persons linked to SCB population registers.
  • Yes: Full count or representative sample. Covers every resident born 1945-1981 registered in Sweden at the end of 1996, not a sample.
  • No: Raw data downloadable. No public microdata; the figures are only in the report's PDF tables.
  • Yes: No known break or quality problem. No break reported within the five-year window.
8%
Bäckman, Estrada, Nilsson & Sivertsson (2021), Nordic Journal of Criminology, Table 1Stockholm University, Department of Criminology · Police and justice recordsRetrieved 2026-10-261.672017
Convictions, not suspicions: share with at least one conviction (court judgment, summary sanction order or waiver of prosecution) between ages 15 and 24, cohorts born 1990-1993 who lived in Sweden throughout those ages.

"Second generation" = born in Sweden to foreign-born parents. The methods section defines natives as born in Sweden with at least one Swedish-born parent (figure captions say Swedish-born parents). Men and women combined by RealStats from the published percentages and group sizes.

Table 1, cohort 1990-93, row ">= 1 conv." (https://www.tandfonline.com/action/downloadTable?id=t0001&doi=10.1080%2F2578983X.2021.1900038&downloadType=CSV).

4 of 5 checks
  • Yes: Method and definitions published. The article describes the registers, the cohort design and the group definitions.
  • Yes: Official statistics or peer-reviewed. Peer-reviewed article in the Nordic Journal of Criminology (open access).
  • Yes: Full count or representative sample. Whole birth cohorts from Sweden's population and conviction registers, not a sample.
  • No: Raw data downloadable. Register microdata are not public; the article publishes percentage tables.
  • Yes: No known break or quality problem. No break in the conviction register over the period reported by the authors.
31%

Born in Sweden to two foreign-born parents: times as often registered as a suspect (adjusted)

Share of people born in Sweden to two foreign-born parents registered as suspected of at least one offence, divided by the same share among people born in Sweden to two Swedish-born parents, after standardising for age, sex, disposable income, education and municipality type. Sweden, people aged 15 and over registered as resident on 31 December 2014, Offences committed 2015-2018. Unit: ratio (1 = same as the reference group).

1.68ratio (1 = same as the reference group), 2018

62Trust scoreModerate certainty

Score breakdown

Quality32 / 40

How well the sources are made

Agreement14.3 / 30

How close the sources are to each other

Independence12 / 20

How many separate ways the number was measured

Freshness4 / 10

How recent the newest data is

How much each source counts

  • Brå 202162%
  • Brå 20058%
  • Bäckman 202131%

Sources

SourceValueDefinitionQualityWeight
Brå report 2021:9, Misstänkta för brott bland personer med inrikes respektive utrikes bakgrund, Figure 7Brottsförebyggande rådet (Brå) · Police and justice recordsRetrieved 2026-10-261.682018
Same register and population as the unadjusted figure.

Brå standardised the group of people born in Sweden to two foreign-born parents to the age and sex distribution, the shares with low disposable family income and low education, and the municipality types of the reference group. Standardised share suspected 5.33% against 3.18% for the reference group.

Report text p. 50 and Figure 7 p. 51; standardised shares in Tabellbilaga 1, Table B7, p. 11 (https://bra.se/download/18.366598ab19270941a1c3955/1729256818286/2021_Tabellbilaga%201%20delstudie%201.pdf). Table B24 (p. 28) gives 1.68 for the same row.

4 of 5 checks
  • Yes: Method and definitions published. The report's method chapter sets out the register, the degree of suspicion counted and SCB's standard background groups.
  • Yes: Official statistics or peer-reviewed. Analysis by Brå, which keeps Sweden's official register of suspected persons, linked to SCB population registers; two external professors read a draft.
  • Yes: Full count or representative sample. Covers every resident aged 15 and over on 31 December 2014, about eight million people, not a sample.
  • No: Raw data downloadable. The linked microdata are not public; the figures are only in the report's PDF tables.
  • Yes: No known break or quality problem. One fixed four-year window with no break; people who moved to Sweden after 2014 are outside the population by design.
62%
Brå report 2005:17, Brottslighet bland personer födda i Sverige och i utlandet, Table 10Brottsförebyggande rådet (Brå) · Police and justice recordsRetrieved 2026-10-261.52001
Residents born 1945-1981, suspected at least once in 1997-2001 (five years).

Standardised for sex, age, education and earned income (förvärvsinkomst); not for disposable income or municipality type. Published to one decimal.

Table 10, p. 40.

4 of 5 checks
  • Yes: Method and definitions published. The report describes the registers, the population and the background groups.
  • Yes: Official statistics or peer-reviewed. Analysis by Brå of the official register of suspected persons linked to SCB population registers.
  • Yes: Full count or representative sample. Covers every resident born 1945-1981 registered in Sweden at the end of 1996, not a sample.
  • No: Raw data downloadable. No public microdata; the figures are only in the report's PDF tables.
  • Yes: No known break or quality problem. No break reported within the five-year window.
8%
Bäckman, Estrada, Nilsson & Sivertsson (2021), Nordic Journal of Criminology, Table 2Stockholm University, Department of Criminology · Police and justice recordsRetrieved 2026-10-261.242017
Convictions, not suspicions: share with at least one conviction (court judgment, summary sanction order or waiver of prosecution) between ages 15 and 24, cohorts born 1990-1993, restricted to people whose parents had the lowest decile of work income in two of the three years at ages 15-17.

Holds childhood family income roughly constant; no adjustment for education or municipality type. The authors note this low-income selection is probably stronger among natives. Men and women combined by RealStats from the published percentages and group sizes.

Table 2, cohort 1990-93, row ">= 1 conv." (https://www.tandfonline.com/action/downloadTable?id=t0002&doi=10.1080%2F2578983X.2021.1900038&downloadType=CSV).

4 of 5 checks
  • Yes: Method and definitions published. The article describes the registers, the cohort design and the group definitions.
  • Yes: Official statistics or peer-reviewed. Peer-reviewed article in the Nordic Journal of Criminology (open access).
  • Yes: Full count or representative sample. Whole birth cohorts from Sweden's population and conviction registers, not a sample.
  • No: Raw data downloadable. Register microdata are not public; the article publishes percentage tables.
  • Yes: No known break or quality problem. No break in the conviction register over the period reported by the authors.
31%

Why the numbers differ

  • "Suspected" is not "convicted". Brå counts everyone registered as at least reasonably suspected, including suspicions later dropped. Counting only suspects whose case ended in a prosecution decision, Brå's ratios are 2.21 for people born abroad and 3.5 for people born in Sweden to two foreign-born parents, a pattern Brå judges close to that for all suspects (Tabellbilaga 1, Table B50; report pp. 145-146). Bäckman et al. count convictions only. (source, retrieved 2026-10-26)
  • Age and income structure. The groups differ in age, and young people are suspected more often. In Brå's analysis, standardising for age and sex alone moves the ratio from 2.51 to 2.25 for people born abroad and from 3.21 to 1.96 for people born in Sweden to two foreign-born parents; adding disposable income, education and municipality type gives 1.76 and 1.68 (report Figure 7, p. 51). (source, retrieved 2026-10-26)
  • Reporting behaviour and police focus. Brå describes two possible selection routes: crimes may be reported more often when the presumed offender has a foreign background (a Swedish survey found this for violent crime), and minorities are over-represented in police stops, which raises registration for drug and traffic offences and the chance that other offences are found. Brå concludes this could explain part of the differences but hardly to a decisive degree (report pp. 112-113). (source, retrieved 2026-10-26)
  • Brå's official suspect statistics add that a suspect can only be registered if the crime is discovered and reported, that police and prosecutors' priorities can vary between groups of offenders and over time, and that control offences such as drug and traffic crimes almost always have a suspect when recorded (report 2026:13, p. 6). Bäckman et al. note that drug convictions depend heavily on police stops. (source, retrieved 2026-10-26)
  • Different designs. Brå 2005 used a five-year window and people born 1945-1981; restricting the 2015-2018 data to ages 15-51 to match, Brå gets 2.2 for people born abroad and 2.4 for two foreign-born parents, against 2.5 and 2.0 in 1997-2001 (report Table 2, p. 54). Bäckman et al. follow birth cohorts to age 24; Adamson uses 2013-2017 without stating the age range. (source, retrieved 2026-10-26)
  • Who is counted. Brå's population is people registered as resident at the end of 2014, so later arrivals, asylum seekers without residence, visitors and other non-residents are outside every ratio. In 2018, 9.0% of registered offences with a suspect were linked to non-residents (English summary, p. 13). (source, retrieved 2026-10-26)

Common mistakes

  • Reading the ratio as a share of the group. In 2015-2018, 7.99% of people born abroad and 3.18% of people born in Sweden to two Swedish-born parents were registered as suspects at least once; most people in every group were not. (source, retrieved 2026-10-26)
  • Confusing share of suspects with risk. People born abroad were 18.37% of the population studied and 32.9% of suspects; a group's share of suspects depends on its size as well as its rate (report Table 1, p. 47). (source, retrieved 2026-10-26)
  • "Foreign background" in Swedish statistics covers both people born abroad and people born in Sweden to two foreign-born parents. Their ratios differ and change differently after adjustment, so they are shown separately here.
  • Reading the adjusted ratio as the effect of migration background. Standardisation only removes differences in the factors Brå measured; Brå states the study does not allow firm conclusions about causes (report p. 110). (source, retrieved 2026-10-26)