Skip to main content

The dynamics of child poverty in Sweden


This paper studies the dynamics of child poverty in Sweden. We find that one out of every five children is disposable income poor at least once during childhood, while only 2% are chronically poor. Children in Sweden are protected economically from many serious events such as parental sickness and death. Family dissolution and parental unemployment do push some children into poverty. However, these poverty spells are mostly temporary. Single mothers, for example, are overrepresented among the poor but not among the chronically poor. Children with immigrant parents are strongly overrepresented among the chronically poor.

This is a preview of subscription content, access via your institution.


  1. Policy evaluators need this information too. Testing the causal impact of a particular policy on child poverty requires the evaluator to model child poverty dynamics properly in order to obtain a clean counterfactual.

  2. When inequality is large (i.e., families find themselves very far below the poverty line), then the distance needed to travel to escape poverty is larger. Thus, even when poor families in most countries travel the same distance from 1 year to the next, those living in high inequality societies do not travel far enough to escape poverty.

  3. In related work, Fritzell and Henz (2001) use register income data from 1971–1991 matched to the Swedish Level of Living Survey to study household poverty dynamics. Key references concerning the dynamics of child poverty include the book edited by Bradbury et al. (2001a) and studies by Jenkins and Schluter (2003), Crossley and Curtis (2006), Fertig and Tamm (forthcoming), and Corak et al. (2008).

  4. For an excellent review of poverty in Sweden during the 1990s, see Gustafsson et al. (2007).

  5. There are also a number of methodological differences between their paper and ours. Hansen and Wahlberg (2009) adopt a multiple spell hazard model, which allows them to estimate the duration of poverty spells and to test for poverty spell dependency despite the fact that the poverty histories that they study are censored (i.e., their data do not allow them to follow the same individuals from age 18 years to death). The fact that our data are largely uncensored allows us to study poverty dynamics using more traditional panel methods. Fertig and Tamm (forthcoming) apply Hansen and Wahlberg’s (2009) methodology to a panel of children from the German Socio-Economic Panel.

  6. Key references concerning poverty dynamics in other countries include Bane and Ellwood (1986) and Stevens (1994, 1999) for the USA, Cappellari and Jenkins (2004) for Britain, Finnie and Sweetman (2003) for Canada, and Devicienti and Gualtieri (2007) for Italy.

  7. Our measures are also unidimensional. Multidimensional measures that included access to services such as health care and education would most likely lower our poverty figures, since these services are provided by the government to all children free of cost.

  8. Having this large sample is important for what we want to accomplish in this paper. It allows us to precisely measure the effects associated with important but less frequent events, such as parental death or disability. This is simply not possible in smaller samples.

  9. In 1986, Statistics Sweden began taking in income information from both the tax registers and from employers. This eliminates the left censoring of reported incomes found in the Swedish tax registers, since employers are required to report even small amounts of income paid out.

  10. Also, if you are either severely poor or have been poor for a long time, then you may have already pushed your savings down to zero and may not be able to borrow more money. If this is true, then current income and consumption expenditures will be equalized.

  11. One example is the recent study published by Statistics Sweden (2005) concerning the living conditions of children in Sweden. This study (a joint project with the Swedish Institute for Social Research) interviewed children about different aspects of their lives, including; health, well being, housing, own finances, leisure activities, school activities, family structure, familial relations, and other important topics. Several topics were even suggested by the children themselves, such as the importance of having pets. Parents were also interviewed (separately) about the living conditions of their children.

  12. In their study on child poverty in Canada, Crossley and Curtis (2006) compare an income-based measure of poverty with several measures based on consumption expenditures. They find that poverty among children is lower when measured using consumption expenditure than when measured using income.

  13. The database is maintained by Statistics Sweden; see See also Edin and Fredriksson (2000) for more information on LINDA.

  14. A similar problem arises when studying child poverty using US Census of Population data. The extent of this problem is analyzed in Carlson and Danziger (1999).

  15. We started out with 19,376 children: 136 children die, 707 emigrate from Sweden, and 20 are lost for unknown reasons. In total, we lose 4.45% of the original probability sample.

  16. For a description of job related benefits see Sjögren Lindquist and Wadensjö (2006).

  17. The assumption of equal sharing of resources within the household has not received much support from the relevant research. Some authors argue that poor parents cushion their children from poverty by consuming less themselves (Middleton et al. 1997), while others stress that mothers and fathers spend different amounts on different types of consumption – with the underlying hypothesis being that mothers will spend more resources on children than fathers (see, e.g., Lundberg et al. 1997). This literature also discusses the importance of the within-household bargaining power derived from a spouse’s own earnings and income, which, in turn, affects the outcome of a bargaining game that determines the composition of family expenditures (see, e.g., Phipps and Burton 1998). Unfortunately, this literature can not provide us with a better sharing rule than the one we have adopted. We can only hope that the equal sharing rule reflects some kind of average among different families.

  18. Note that one could have chosen the median income of all children as the benchmark. We choose the median income of all persons in the population.

  19. The Swedish Government has frequently used 50% of the median income as an unofficial poverty line, for example, in its 2001 National Action Plan against Poverty and Social Exclusion written for the European Union. The United Nations and the OECD also use this measure. Another reason for choosing this poverty line is that we want to be able to compare our results with those produced in Duncan et al. (1993), Gustafsson (2000), Oxley et al. (2000), and Bradbury and Jäntti (2001a, b), all of which used 50% of the median income as their poverty line.

  20. The participation rate for women fluctuates between 70% and 80% during the time period studied.

  21. The parental leave insurance system provides families with newborn and small children with 13 months of taxable benefits equal to 80% of the at-home parent’s current salary up to a maximum monthly salary of approximately 33,000 SEK (≈US $4714/month). Two of these months must be taken by the father or they are lost. After these 13 months are used up, parents have an additional 90 days of benefits paid out at a fixed daily rate of 180 SEK (≈$26)/day. Many parents also have extra parental leave insurance provided for them through their collective labor agreements. All central government employees, for example, receive 90% of their salary and the monthly salary ceiling is removed entirely.

  22. Bradbury and Jäntti (2001a) report 39% market poor children for Sweden. They report 31% for the USA and an average across 23 countries of 31.7%. Their measure of market poor used for Sweden differs from ours, since it is based on the older, tax unit definition of households. For the other countries, their measure is based on a household definition similar to our RTB definition.

  23. Using cross-sectional data on disposable income from the Luxembourg Income Study and applying the same poverty definition as we use, Bradbury and Jäntti (2001a) calculated that the share of disposable income poor in Sweden in 1992 to be 3.7%. Our cross-sectional measure for 1992 is 5.8% (see Table 1). Using similar methods, the Swedish poverty rate in 1995 was estimated at 2.6% and 2.7%, respectively, by UNICEF (2000) and Förster and Pellizzari (2000). Our cross-sectional estimate for 1995 is 4.1% (see Table 1). Thus, our low measure of poverty actually lies above the level of poverty reported in previous studies. This difference is most likely due to the fact that previous studies have been based on “tax units” as opposed to households. Our study uses the new RTB family definition, which is more similar to the definition of a household used in most other countries.

  24. We have two reasons for not attempting to adjust for age in our baseline calculations and estimations. First, we want to use the same equivalency scale as previous authors in order to make our results comparable with theirs. Second, we do not actually know how old all of our cohort’s siblings are. That is, we know our cohort members’ ages, and we know how many siblings they have. However, we do not know the ages of their siblings.

  25. Existing definitions of chronic poverty have been largely determined by the characteristics of the data used. Our definition is similar to the one used by Walker and Ashworth (1994). Using data from the Panel Study of Income Dynamics, they found that 12% of American children were poor for 7 years or more during their childhood. Hill and Jenkins (2001) used 6 years of the British Household Panel Survey and defined chronic poverty in two different ways. They first counted the number of children who had been poor every year for all 6 years. They found that 2.4% of children aged 0–5 years, 1.5% of children aged 6–11 years, and 0% of children aged 12–17 years had experienced chronic poverty. Then, they gave each child his or her average income (i.e., they smoothed incomes over the 6-year period) and then calculated the share of children who fell below the poverty line in all 6 years. This resulted in 14%, 8.2%, and 2.9% of chronically poor children in each respective age group.

  26. Please note that we are not looking for the “causes” of poverty in this exercise. We are merely trying to identify who the poor children are or, more precisely, to look for observable characteristics shared by poor children.

  27. We do not feel that it makes much sense to ask the question: How does the poverty risk change if we increase the share of adults in the household that are born abroad by 1%? The immigrant share variable is not a continuous one. In practice, the most common values of Immigrant Share are 0.25, 0.33, 0.5, 0.66, 0.75, and 1. So instead, we ask: How does the poverty risk change if you have one parent born abroad or two parents born abroad?

  28. In practice, this means that (for some children in single parent households) we will be underestimating the available resources within the household.

  29. KAS was removed in 1999, so we only have data on minimum employment benefits up until 1998.

  30. Currently, the replacement rate is 80% for the first 200 days up to a ceiling of 20,075 SEK/month. After that, the replacement rate falls to 70%. Unemployed workers are eligible for unemployment insurance for a total of 300 days. Many workers have additional insurance provided through their employment contracts (see Sjögren Lindquist and Wadensjö 2006).

  31. Keep in mind that we may be underestimating unemployment among poor children since we do not have a proper measure of unemployment spells, underemployment, discouraged workers, or job search. What we have are payments made from the unemployment insurance system and from labor market programs. We can only use these as partial indicators of unemployment. Some parents may be unemployed and find themselves outside of these two systems.

  32. We also ran the following alternative experiments. First, we used family characteristics in 1991 in an attempt to predict who would become chronically poor. Parents’ immigrant status and unusually low education were the only two predictors of chronic poverty. Being born to a single mother did not predict chronic poverty. Second, we ran simple univariate probit regressions of chronic poverty on the number of years spent as a single mother. In one specification, years spent as a single mother entered the regression as a continuous variable, and in an alternative specification, the number of years entered as dummies. Neither of these regressions gave any indication that single mothers were overrepresented among the chronically poor.

  33. Once again, our measures of unemployment may lead us to understate the importance of unemployment, particularly among the chronically poor, since unemployment is highly correlated with both low education and immigrant status.

  34. Actually, we only know the year within which an event occurs. Also, we say remarriage, but this includes cohabitation if the two partners have at least one child together.


  • Bane MJ, Ellwood DT (1986) Slipping into and out of poverty: the dynamics of spells. J Hum Resour 21(1):1–23

    Article  Google Scholar 

  • Bradbury B, Jäntti M (2001a) Child poverty across twenty-five countries. In: Bradbury B, Jenkins SP, Mickelwright J (eds) The dynamics of child poverty in industrialised countries. Cambridge University Press, Cambridge, pp 62–91

    Chapter  Google Scholar 

  • Bradbury B, Jäntti M (2001b) Child poverty across the industrialised world: evidence from the Luxembourg Income Study. In: Vleminckx K, Smeeding TM (eds) Child well-being, child poverty and child policy in modern nations: what do we know? The Policy Press, Bristol, pp 11–32

    Google Scholar 

  • Bradbury B, Jenkins SP, Mickelwright J (2001a) The dynamics of child poverty in industrialised countries. Cambridge University Press, Cambridge

    Book  Google Scholar 

  • Bradbury B, Jenkins SP, Mickelwright J (2001b) The dynamics of child poverty in seven industrialised nations. In: Bradbury B, Jenkins SP, Mickelwright J (eds) The dynamics of child poverty in industrialised countries. Cambridge University Press, Cambridge, pp 92–132

    Chapter  Google Scholar 

  • Cappellari L, Jenkins SP (2004) Modelling low income transitions. J Appl Econ 19(5):593–610

    Article  Google Scholar 

  • Carlson M, Danziger S (1999) Cohabitation and the measure of child poverty. Rev Income Wealth 45(2):179–191

    Article  Google Scholar 

  • Corak M (2006) Principals and practicalities for measuring child poverty in the rich countries. Int Soc Secur Rev 59(2):3–36

    Article  Google Scholar 

  • Corak M, Fertig M, Tamm M (2008) A portrait of child poverty in Germany. Rev Income Wealth 54(4):547–571

    Article  Google Scholar 

  • Crossley TF, Curtis LJ (2006) Child poverty in Canada. Rev Income Wealth 52(2):237–260

    Article  Google Scholar 

  • Departementsserien (DS) (2004) Ekonomiskt utsatta barn (Economic hardship among children). Socialdepartementet DS 2004:41, Stockholm

  • Devicienti F, Gualtieri V (2007) The dynamics and persistence of poverty: evidence from Italy. Centre for Employment Studies Working Paper No. 63

  • Duncan GJ, Brooks-Gunn J (eds) (1997) The consequences of growing up poor. Russel Sage Foundation, New York

    Google Scholar 

  • Duncan GJ, Gustafsson B, Hauser R, Schmauss G, Messinger H, Muffels R, Nolan B, Ray J-C (1993) Poverty dynamics in eight countries. J Popul Econ 6(3):215–234

    Article  Google Scholar 

  • Edin P-A, Fredriksson P (2000) LINDA—Longitudinal Individual Data for Sweden Working Paper 2000:19. Department of Economics, Uppsala University

  • Fertig M, Tamm M (forthcoming) Always poor or never poor and nothing in between? Duration of child poverty in Germany. German Economic Review

  • Finnie R, Sweetman A (2003) Poverty dynamics: empirical evidence for Canada. Can J Econ 36(2):291–325

    Article  Google Scholar 

  • Fritzell J, Henz U (2001) Household income dynamics: mobility out of and into low income over the life-course. In: Jonsson JO, Mills C (eds) Cradle to grave: life-course change in modern Sweden. Sociology Press, York, pp 184–210

    Google Scholar 

  • Förster MF, Pellizzari M (2000) Trends and driving factors in income distribution and poverty in the OECD area. OECD Labour Market and Social Policy, Occasional Papers No. 42

  • Galloway TA, Gustafsson B, Pederson PJ, Österberg T (2009) Immigrant child poverty in Scandinavia: a panel data study. IZA Discussion Paper No. 4232

  • Gustafsson B (2000) Poverty in Sweden: changes 1975–1995, profile and dynamics. In: Gustafsson B, Pedersen PJ (eds) Poverty and low income in the Nordic countries. Ashgate Publishing Ltd., Aldershot, pp 169–206

    Google Scholar 

  • Gustafsson B, Zaidi A, Franzén E (2007) Financial poverty. Int J Soc Welf 16(July supplement):67–90

    Article  Google Scholar 

  • Hansen J, Wahlberg R (2009) Poverty persistence in Sweden. Rev Economics Household 7(2):105–132

    Article  Google Scholar 

  • Hill M, Jenkins SP (2001) Poverty among British children: chronic or transitory? In: Bradbury B, Jenkins SP, Mickelwright J (eds) The dynamics of child poverty in industrialised countries. Cambridge University Press, Cambridge, pp 174–95

    Chapter  Google Scholar 

  • Jäntti M, Danziger S (1994) Child poverty in Sweden and the United States: the effect of social transfers and parental labor force participation. Ind Labor Relat Rev 48(1):48–64

    Article  Google Scholar 

  • Jenkins SP, Schluter C (2003) Why Are Child Poverty Rates Higher in Britain than in Germany? J Hum Resour 38(2):441–465

    Article  Google Scholar 

  • Lichter DT (1997) Poverty and inequality among children. Annu Rev Sociol 23:121–145

    Article  Google Scholar 

  • Lundberg SJ, Pollak RA, Wales TJ (1997) Do husbands and wives pool their resources? J Hum Resour 32(3):463–480

    Article  Google Scholar 

  • Middleton S, Ashworth K, Braithwaite I (1997) Small fortunes. spending on children, childhood poverty and parental sacrifice. Joseph Rowntree Foundation, York

    Google Scholar 

  • Oxley H, Dang TT, Antolín P (2000) Poverty dynamics in six OECD countries. OECD Economic Studies No. 30

  • Phipps SA, Burton PS (1998) What’s mine is yours? The influence of male and female incomes on patterns of household expenditure. Economica 65(260):599–613

    Article  Google Scholar 

  • Sjögren Lindquist G, Wadensjö E (2006) National social insurance—not the whole picture. Report for the Expert Group on Economic Studies (ESS), 2006:5

    Google Scholar 

  • Statens Offentliga Utredningar (SOU) (2001) Ur fattigdomsfällan (Exiting poverty), Slutbetänkande av Familjeutredningen SOU 2001:24, Stockholm

  • Statistics Sweden (SCB) (2002) Barnens del av kakan: Välstånd och Fattigdom bland Barn 1991–1999 (Childrens’ share of the cake: welfare and poverty among children 1991–1999), Demografiska Rapporter 2002:1

  • Statistics Sweden (SCB) (2005) Barns Villkor (Children’s living conditions), Levnadsförhållanden Rapport 110

  • Stevens AH (1994) The dynamics of poverty spells: updating Bane and Ellwood. Am Econ Rev 84(2):34–37

    Google Scholar 

  • Stevens AH (1999) Climbing out of poverty, falling back in. J Hum Resour 34(3):557–588

    Article  Google Scholar 

  • UNICEF (2000) A league table of poverty in rich nations. UNICEF Innocenti Research Centre, Innocenti Report Card No. 1

  • Vleminckx K, Smeeding TM (2001) Child well-being, child poverty and child policy in modern nations: what do we know? The Policy Press, Bristol

    Google Scholar 

  • Walker R, Ashworth K (1994) Poverty dynamics: issues and examples. Averbury, Aldershot

    Google Scholar 

Download references

Author information

Authors and Affiliations


Corresponding author

Correspondence to Matthew J. Lindquist.

Additional information

Responsible editor: Junsen Zhang

We are especially grateful to Johan Fritzell, Björn Gustafsson, Mats Johansson, Carina Mood, and to an anonymous referee for their valuable comments and suggestions. We would also like to thank seminar participants at Stockholm University, The Swedish Institute for Social Research, The Swedish National Social Insurance Board (Försäkringskassan), Uppsala University, and participants at the Centre for European Economic Research conference on Inequality and Poverty in the Global Economy. Financial support from The Swedish National Social Insurance Board is gratefully acknowledged.

Rights and permissions

Reprints and Permissions

About this article

Cite this article

Lindquist, M.J., Sjögren Lindquist, G. The dynamics of child poverty in Sweden. J Popul Econ 25, 1423–1450 (2012).

Download citation

  • Received:

  • Accepted:

  • Published:

  • Issue Date:

  • DOI:


  • Child poverty
  • Chronic poverty
  • Poverty dynamics

JEL Classification

  • I32
  • J13