Local disparities in paths through social assistance in Belgium Sarah Carpentier Herman Deleeck Centre for Social Policy Antwerp University Aspen-Etui.

Local disparities in paths through social assistance in BelgiumSarah CarpentierHerman Deleeck Centre for Social Policy Antwerp UniversityAspen-Etui Conference 2009 Brno (20-21 March)

2

1. Background2. Characteristics of the Belgian social

assistance scheme3. Towards an individualised work based social

assistance scheme 4. Geographical disparities in the incidence of

social assistance claimants5. Evidence on the duration of the first spell 6. Explanatory mechanisms for local variation in

time on welfare7. Conclusion

Outline

3

• Transition to an active welfare state is well documented

• But local variation is largely uncharted territory (Finn, 2000; Meyers, 1998)

• Yet, pertinent for legal security and equal access to social rights (Cox, 1998)

• Literature on social assistance dynamics- Covers mainly the dependency versus

heterogeneity hypotheses - Institutional impact and contextual effects are

rarely documented (except ESOPO-project)

1. Background

4

• Low percentage of social assistance claimants(0.8% of adult population in 2004)due to - Corporatist welfare regime (Esping-

Andersen, 1990)- Existing social security schemes

school leavers: flat-rate unemployment benefitunemployment benefit in principle unlimited in time

2. Characteristics of the Belgian social assistance scheme

5

• Decentralised organisation (since 1974)- National statutory framework - Local welfare agencies (OCMW’s/CPAS)

= Autonomous organisations under public law governed by politically appointed

councils• Administer federal law with broad

autonomy & margin of appreciation • Task to take supplementary initiatives

according to own insight & to local needs to ensure human dignityDevolution

2. Characteristics of the Belgian social assistance scheme

6

• Entitlement is conditioned on a family-based means test

• Entitlement is unlimited in timeAdministratively, subject for revision after one year

• To international standards Belgium has a low benefit level (Van Mechelen & De Maesschalck, 2009) and supplementary benefits are fully subject to local discretion

2. Characteristics of the Belgian social assistance scheme

7

• Spring Programme (2000)aims at increased labour market integration by - Enhanced subsidy for employment

programmes- Diversification of general & specific social

employment programmes

3. Towards an individualised work based social assistance scheme

8

• Social Integration Act (2002)= Introduction of work based &individualised social assistance scheme- Explicit aim of societal integration and self-

sufficiency through labour market integration - Strict interpretation of availability for

work - Focus on rights and duties - Strengthening of tailor-made approach - Reinforcement of recipients as active

participants

3. Towards an individualised work based social assistance scheme

9

• Implementation of Social Integration Act (2002)- Generalisation of personal social integration

contract - No methodologies to assess availability for

work and exemption from availability for work - No instruments to allocate recipients to

training and employment programmes - Crucial notions are not (clearly) defined

Enhanced local discretionary power

3. Towards an individualised work based social assistance scheme

4. Geographical disparities in percentage of adults claiming social assistance (2004)

Source: Data Warehouse Labour Market & Social Security, 2004, own calculations

N (municipalities) %

Region

Brussels region 19 3.1

Flemish region 308 0.3

Walloon region 262 0.9

Number of inhabitants

≤ 9999 inhabitants 253 0.5

10000 – 19999 inhabitants 201 0.6

20000 – 39999 inhabitants 100 0.8

40000 – 99999 inhabitants 27 2.1

≥ 100 000 inhabitants 8 2.8

Minimum for Belgium 0.0

Maximum for Belgium 7.9

Mean percentage of adults claiming social assistance by groups of municipalities

N (municipalities) Mean % of adults

Socio-economic clusters in Brussels

Living municipalities 10 1.8

Urbanized central municipalities 9 4.7

Socio-economic clusters in Flanders

Living municipalities 83 0.3

Rural municipalities 97 0.2

With a concentration of economic activity 40 0.2

Semi urban municipalities 42 0.4

Centre municipalities 38 0.7

Tourist municipalities 8 0.6

Socio-economic clusters in Wallonia

Living municipalities 77 0.6

Rural municipalities 68 0.7

With a concentration of economic activity 39 1.0

Semi urban municipalities 25 1.4

Centre municipalities 53 1.4

13

• Data - Administrative records from the Data Warehouse

Labour Market and Social Security - Advances on social security benefits are excluded- 1/3 proportionally stratified sample of 2004

cohort entrants (N=14503) aged between 18 and 65

- Observation window of 36 months, monthly recorded data

5. Evidence on the duration of the first spell

14

• Methodology- Event history analysis in a continuous

time framework- No time-varying covariates

Municipality characteristics at entry time• Study of the duration of the first spell

- Unspecified exit - Breaks of maximum 2 months are allowed- (Complementary) participation in a social

employment programme is considered as an exit

5. Evidence on the duration of the first spell

15

• An exit is not necessarily a meaningful improvement of the living standard (Kazepov,1999; Hansen, 2008)Exit covers several realities: - Participation in labour market programme- Regular employment - Entitlement to social security benefit or other

social assistance benefit- Change in household composition or labour

market situation (no longer entitled)- Administrative sanction - Suspension due to stay in foreign country or

admission in institution covered by the state

5. Evidence on the duration of the first spell

16

• Duration of first spell may be highly misleading (due to cycling)

• A short time on social assistance is not necessarily an indication of a good welfare policy (Kazepov, 1999; Saraceno, 2002)

5. Evidence on the duration of the first spell

Huge local variation in median duration & a high turnover to international standards

Belgium(N=14503)

Minimum by municipality(N=11615)*

Maximum by municipality (N=11615)*

Median duration 8.0 2.5 27.0

Source: Data Warehouse Labour Market & Social Security, own calculations*= calculated for the 159 municipalities with at least 20 recipients in our sample

Source: Data Warehouse Labour Market & Social Security, own calculations

N Median duration

Region

Brussels region 3091 10

Flemish region 5628 6

Walloon region 5784 10

Number of inhabitants

≤ 9999 inhabitants 1365 12

10000 – 19999 inhabitants 2734 9

20000 – 39999 inhabitants 3173 8

40000 – 99999 inhabitants 3485 8

≥ 100 000 inhabitants 3743 7

Median duration of 1st spell by region and by number of inhabitants in the community

Source: Data Warehouse Labour Market & Social security, own calculations

0.0

00

.25

0.5

00

.75

1.0

0

0 10 20 30 40analysis time

Brussels WalloniaFlanders

Kaplan-Meier survival estimates, by region

Source: Data Warehouse Labour Market & Social security, own calculations

0.0

00

.25

0.5

00

.75

1.0

0

0 10 20 30 40analysis time

<10000 10000-1900020000-39000 40000-99000>=100000

Kaplan-Meier survival estimates, by magnitude

N Median duration

Socio-economic clusters in Brussels

Living municipalities 715 12

Urbanized central municipalities 2376 10

Socio-economic clusters in Flanders

Living municipalities 859 9

Rural municipalities 536 8

With a concentration of economic activity 361 7

Semi urban municipalities 694 6

Centre municipalities 3018 5

Tourist municipalities 160 4

Socio-economic clusters in Wallonia

Living municipalities 619 11

Rural municipalities 404 13

With a concentration of economic activity 803 11

Semi urban municipalities 768 12

Centre municipalities 3190 9

22

• Population composition • Local labour market conditions• Institutional design and implementation of social

assistance- Generosity - Allocation to training and labour market programmes

- Sanctions • Local welfare mix • Networks and resources • Efficacy of training and labour market programmes

6. Explanatory mechanisms for local variation in time on welfare

23

7. Conclusion

• Substantial local variation in the percentage of adults claiming social assistance by municipality (0 – 8%)A clear regional pattern with regard to the incidence, and a concentration in the cities

• High turnover in Belgium to international standards (median duration of 8 months)

• Huge disparities in the median duration of the first spell by municipality: from 3 till 27 months

24

7. Conclusion

• The Flemish region has the highest turnover (median duration is 6 months)

• Indication of a negative relationship between the incidence of social assistance and the duration of the first spell and for the number of inhabitants in a municipality and for the socio-economic clusters

• Further exploration of determinants of local disparities is needed to gain insight in the role of local welfare agency policy

Thank you for your attention!

Comments are also welcome on [email protected]

Percentage of net disposable social assistance income for different household types to the 60%

poverty threshold

Source: Stasim, 2004, own calculations