Longitudinal Analysis of Benefits Data for Workless People

1
Longitudinal Analysis of Benefits Data for
Workless People

• Chelsie Anttila
• Social Disadvantage Research Centre (SDRC)
• University of Oxford

2
Overview of the work

• Changing patterns of worklessness in Wales
• Follow on from earlier work, e.g. Claiming
  Matters, 2004
• Worklessness defined as involuntary exclusion
  from labour market. This definition includes
• Unemployed people claiming Job Seekers Allowance
  (JSA)
• Sick people claiming Incapacity Benefit/Severe
  Disablement Allowance (IB/SDA)

3
Data sources

• DWPs Tabulation Tool
• (http//www.dwp.gov.uk/asd/tabtool.asp)
• DWP HMRC WPLS

4
WPLS

• Work and Pensions Longitudinal Study
• Changes to data protection law in 2002 allowed
  the DWP to
• Receive data on employment from Her Majestys
  Revenues and Customs and
• Link it to the DWPs benefit database.
• Result was the WPLS database

5
Uses of WPLS

• December 2003, Parliament stated that
• WPLS will be used for statistics and research
• Aid in the investigation of benefit fraud
• Improve the way initiatives are targeted
• Modernise DWP statistics (replaced 5 sample Oct.
  2005)
• WPLS database was introduced in January 2004 but
  contains data back to 1999

6
Advantages of WPLS over other administrative data

• Unprecedented capacity to link benefit and
  employment claims
• Unlike other admin data, WPLS is a single
  database including all benefits and spell dates
• Able to detect short claims, WPLS takes extracts
  in shorter intervals (2, 4, or 6 weeks)

7
WPLS Overview

• Three parts
• DWP claimant records
• DWP HMRC records
• HMRC records
• We use parts 1 2
• Focus on those of working age
• About 10 million records at individual level,
  most with claimants home postcode

8
Limitations of WPLS

• Not all employment spells captured
• Taxable benefit spells may appear as employment
• Subject to revision
• Good news because data is constantly improved
• Bad news because work/information can become out
  of date
• Large dataset cumbersome time consuming to use
• Remains the best data source available

9
Overview of worklessness in Wales, 1999-2005
10
Worklessness in Wales England, 1999-2005
11
Welsh JSA caseload by age group, 1999 2005
12
Welsh IB/SDA caseload by age group, 1999 2005
13
What happens to the workless population over time?

• Matching the claimant population on one date
  with the claimant population on a subsequent date
  to study transitions/trajectories
• Three main destinations
• Claiming the same workless benefit as at previous
  time point
• Claiming another workless benefit (i.e. switch
  from JSA to IB/SDA or vice versa)
• Not claiming any workless benefit

14
2005 destinations of 1999 JSA claimants
15
Of course, its not quite that straightforward
16
Trajectories of Welsh 1999 JSA claimants
17
More detail on JSA exiters
18
2005 destinations of 1999 IB/SDA claimants
19
More detail on IB/SDA exiters
20
Worklessness rates in Welsh and English local
authorities in 2005, ranked in order
21
Changes in worklessness rates in Welsh local
authorities, 1999-2001, 2001-2003 and 2003-2005
22
Changes in unemployment rates in Welsh local
authorities, 1999-2001, 2001-2003 and 2003-2005
23
Changes in work-limiting illness rates in Welsh
local authorities, 1999-2001, 2001-2003 and
2003-2005
24
2005 Destinations of 1999 JSA claimants by LA
25
2005 Destinations of 1999 JSA claimants in Wales
26
2005 Destinations of 1999 IB/SDA claimants in
Wales
27
Separating LSOAs into deciles by percentage point
changes in unemployment rates
28
Separating LSOAs into deciles by percentage point
changes in work-limiting illness rates
29
Swansea as case study

• Among the 50 most work deprived local
  authorities in Wales (ranked 8 out of 22, where 1
  is the most deprived, in both 1999 and 2005)
• Home to both the most and the least deprived
  LSOAs in all Wales, both in terms of overall
  worklessness and work-limiting illness/disability
• 147 LSOAs in Swansea, rates of worklessness
  ranged from 1.0 to 53.5 in 1999 and from 0.9
  to 55.3 in 2005
• Highest worklessness rates in 025A, Castle ward
• 3rd highest worklessness rate in England and
  Wales in 2005
• Home to Swansea Castle, rail station, possibly
  BBs, temporary accommodation
• large proportion (about 63) of the working age
  population are men
• more than 63 of men in the area were workless in
  1999 as compared to about 40 of women
• Lowest worklessness rates in 023A, Killay North
  ward
• Home to a nature reserve, student housing and
  academic departments of University College
  Swansea
• 28th lowest worklessness rate in England and
  Wales in 2005

30
Example of Swansea 001B

• Pontardulais ward, north west boundary of LA
• Moved from the most deprived decile in 1999 into
  the 30 least deprived of LSOAs in Wales in 2005

31
Conclusion

• Worklessness rates in an area can decrease while
  people in that area remain workless
• Importance of studying workless people in
  addition to workless places
• E.g. Rates of unemployment in Communities First
  areas may be decreasing, but experimental models
  show no positive significant impact on likelihood
  of unemployed individual moving into work

32
Conclusion

• While person and place-based initiatives can
  both be effective, their relative effectiveness
  cannot yet be adequately addressed. However, the
  development of datasets that link
  individual-level information over a prolonged
  period of time may soon make this possible, if
  only retrospectively. If more is to be learned
  from the next generation of evaluations, it must
  always be clear how policies are expected to
  work, and these theories of change must inform
  the design of impact and process evaluations and
  the interpretation of the results.