Risk factors of long-term social assistance recipiency among lone mothers. The case of Sweden

Abstract

For decades, lone mothers have been vastly overrepresented among Swedish social assistance recipients. In Sweden, social assistance is administered in the personal social services. The study aims to explore and analyse factors associated with long-term recipiency among lone mothers. The empirical material consists of micro data on 875 randomly selected lone mothers receiving social assistance in any of the three major cities of Sweden in 2007. In the analysis, a distinction is made between household demographics (e.g. ethnicity) and individual factors (subdivided into individual impairments (e.g. psychosocial conditions) and financial vulnerability (e.g. unemployment)). The main findings are: (1) even though household demographics alone show low explanatory value for long-term social assistance recipiency, analysed factors remain significant when controlling for individual factors; (2) among individual factors, only variables indicating financial vulnerability are significant. A main conclusion in the study is that social workers should be attentative that subgroups within the larger population of lone mothers have higher risk of long-term recipiency (e.g. non-native origin). Furthermore, policy revisions should be considered for this group, since personal social services extensively address individual impairments whereas long-term recipiency is strongly associated with structural factors.

Under flera decennier har ensamstående mödrar varit kraftigt överrepresenterade bland socialbidragshushållen. I studien undersöks faktorer kopplade till långvarigt bidragstagande (minst 10 månader under ett kalenderår) inom gruppen. Studien bygger på individdata kring 875 slumpmässigt utvalda ensamstående mödrar som uppbar socialbidrag någon gång i Stockholm, Göteborg eller Malmö under 2007. I analysen görs en distinktion mellan faktorer på hushållsdemografisk och individuell nivå, där de senare delas in i individuella hinder (t ex psykosociala problem) och ekonomisk utsatthet (t ex arbetslöshet). Medan hushållsdemografiska faktorer är mer eller mindre oföränderliga, kan individuella faktorer ses som principiellt åtkomliga via sociala insatser på individnivå (främst individuella hinder) eller fördelningspolitik på gruppnivå (främst ekonomisk utsatthet). Studiens huvudsakliga resultat är (1) trots att hushållsdemografiska faktorer (ålder, etnicitet, familjestorlek och utbildning) har lågt självständigt förklaringsvärde, kvarstår de som signifikanta vid kontroll för individuella faktorer; (2) bland individuella faktorer är enbart variabler som mäter ekonomisk utsatthet (arbetslöshet, avsaknad av ersättning från socialförsäkring samt avsaknad av löneinkomster) av betydelse. En slutsats som dras av studien är att socialbidragshandläggare bör vara uppmärksamma på att vissa undergrupper (t ex invandrare) bland ensamstående mödrar löper större risk att hamna i ett långvarigt bidragstagande. En annan slutsats är att man bör överväga policyrevisioner på samhällsnivå för klientgruppen. För närvarande syftar socialbidragshandläggning i stor utsträckning till att hitta förklaringar och lösningar på individers ekonomiska problem hos dem själva, samtidigt som det långvariga bidragstagandet inom gruppen i avgörande utsträckning har kopplingar till strukturella faktorer.

Keywords:

• Income Support/Social Assistance
• Social Policy
• Gender/Feminism
• Lone Mothers

Nyckelord:

• Ensamstående mödrar
• ekonomisk utsatthet
• psykosociala problem
• långvarigt socialbidragstagande

Notes

1. Long-term recipiency is defined as ≥10 month's uptake annually.

2. Immigrants receiving introductory allowance are excluded from the population.

3. Multicollinearity has been controlled using Pearson's r. The strongest correlation between independent variables in the forthcoming models are employment vs. earned incomes (r=0.456) and age vs. poor somatic health (r=0.227). The boundary for multicollinearity is often set to about >0.700 (for further discussions on the subject, see e.g. Tabachnick & Fidell, 2001).

4. Regional differences in this aspect are not reported in national statistics, but traditionally Stockholm, Gothenburg and Malmo have had a high representation of long-term recipients (NBHW [Socialstyrelsen], 1999). However, Stockholm reports 37% long-term recipients in the corresponding year (USK, 2008). There were no gender differences at the national level in the proportion of long-term recipients at year of measurement (NBHW [Socialstyrelsen], 2008).

5. During analytical work, separate logistic regression models were carried out for natives and non-natives within the group, since previous research has shown that there are substantial differences between these groups (Stranz & Wiklund, 2011). However, these models did not deviate in any notable respect compared with use of ethnicity as an independent variable.

6. About 43% of the European immigrants in the sample originate from the former Yugoslavia. Non-European immigrants originate predominantly from countries in the Middle East (e.g. Iraq) or in East Africa (e.g. Somalia). The large wave of refugees from the Balkans came in the early 1990s, whereas a large proportion of Iraqi and Somali refugees mainly came to Sweden during the 2000s.

7. During analytical work, a separate regression model was carried out including household demographics and individual impairments. Neither ‘mental health impairments’ nor ‘psychosocial conditions’ showed significant associations in this model.

8. The ≥10 months per year inclusion criterion is based on uptake in one specific agency. The criterion is consistent with most definitions used in Swedish official reports and research.

9. During analytical work, a regression model including both lack of social transfers and lack of wage earnings as categorical variables was carried out; clients with incomes/transfers below or equal to the sample median (1) (social transfers=about 490 euro, wage earnings=about 550 euro) were separated from those with incomes above the sample median (2). Analysis showed a significant and gradual decrease in odds ratios regarding social transfers (1=0.47 and 2=0.34). Odds ratios regarding wage earnings were similar (1=0.54 and 2=0.46), but partly non-significant.