Royal Statistical Society
Edinburgh Section – 11 October 2016
Britain’s Misleading
Benefit Sanctions Statistics
David Webster
Urban Studies
University of Glasgow
Outline
• What are benefit sanctions?
• Why current concern – the great sanctions drive
2010-16 & extension to part-timers under UC
• Deficiencies of the official statistics
– Misrepresentations
– Missing information
• Complaint 2015 to the UK Statistics Authority &
current state of play
• Statistical difficulties in creating a historical
account
Disqualifications & sanctions
• Both mean loss of benefit for a defined period
• Disqualifications - passive conditionality
– availability, signing-on/interviews, vol. leaving, refusal
of job
• Sanctions – active conditionality
– ‘actively seeking work’, training schemes, workfare,
Jobseekers Directions, ‘claimant commitment’
• National Insurance scheme in 1911 had only
disqualifications
• Growth of sanctions since 1986 under influence
of ‘active labour market policy’
The JSA sanctions campaign 2010-16
•
•
•
•
JSA sanctions/disqualifications doubled after 2010
Peak 1,041,000 2013; 310,000 (inc. UC) 2015/16
Now back to pre-2010 level
22% of all JSA claimants were sanctioned (after
challenges) over 5 years 2009/10 to 2013/14 (FoI) – one
quarter before challenges
• 12.9% of all JSA claimants sanctioned (after challenges) in
2014/15; of these, a quarter sanctioned more than once
(FoI)
• All the most frequently occurring JSA sanctions
lengthened from 22 Oct 2012 – min. now 4 weeks, harsh
penalties for repeats with max. 3 years
• Loss of all benefit, have to apply for 60%/80% ‘hardship
payments’ – designed to clean claimant out of resources
JSA sanctions per month as % of claimants, before
and after review/reconsideration or appeal, last 12
months
Coalition
8
7
JSA originally adverse decisions as % of claimants, last 12 months
6
JSA adverse decisions after review/appeal as % of claimants, last
12 months
5
John Hutton Sec of State
4
3
2
Note: All pre-challenge
figures are approximate
estimates – see later
1
45J/24
Dec-16
Jul-16
Feb-16
Apr-15
Sep-15
Nov-14
Jun-14
Jan-14
Aug-13
Oct-12
Mar-13
May-12
Dec-11
Jul-11
Feb-11
Sep-10
Apr-10
Nov-09
Jun-09
Jan-09
Aug-08
Mar-08
Oct-07
May-07
Dec-06
Jul-06
Feb-06
Sep-05
Apr-05
Jun-04
Nov-04
Jan-04
Aug-03
Mar-03
Oct-02
Dec-01
May-02
Jul-01
Feb-01
Sep-00
Apr-00
0
Changing reasons for JSA sanctions (post-challenge)
(no. of sanctions for each reason as a percentage of total
sanctions)
60.00
1997
2003
50.00
2009
03revK/05
2013
40.00
2014
2015
30.00
20.00
10.00
0.00
Vol./Mis.
Ref. empl.
Tr&Empl sch. MWA & Wk Exp. JS Direction
FTA interview
ASW
Availability
Other
JSA sanctions (before challenges)
and hardship payment awards, 2011-15
120000
JSA sanctions before challenges
(est.)
100000
Harsher regime
80000
60000
40000
57/03
20000
0
JSA hardship payment awards
ESA sanctions
• Escalated ESA (WRAG) sanctions since 2011 –
2.9% of all claimants in 2014/15
• More repeat sanctions than for JSA
• Peak 49,400 in 2013/14, 20,200 in 2015/16
• ESA sanctions also much harsher from 3 Dec 2012
– now lose all of personal allowance £73.10 p.w.
plus fixed penalty 1, 2 or 4 weeks (previously only
WRAG component £29.05 p.w., half for first 4
weeks then all)
• 85% of ESA sanctions are now for nonparticipation in ‘work related activity’
ESA sanctions per month as % of claimants, before
and after review/reconsideration or appeal, last 12
months
2.5
ESA originally adverse decisions as % of WRAG, last 12 months
2
ESA adverse decisions after review/appeal as % of WRAG, last 12
months
1.5
1
Note: ESA sanctions started in Oct 2008,
but the published data only permit this
chart to be shown for the months since
0.5
45J/07
Oct-16
Jul-16
Apr-16
Jan-16
Oct-15
Jul-15
Apr-15
Jan-15
Oct-14
Jul-14
Apr-14
Jan-14
Oct-13
Jul-13
Apr-13
Jan-13
Oct-12
Jul-12
Apr-12
Jan-12
Oct-11
Jul-11
Apr-11
Jan-11
Oct-10
Jul-10
Apr-10
Jan-10
Oct-09
Jul-09
Apr-09
Jan-09
Oct-08
0
Oct-16
Jul-16
Apr-16
Jan-16
47H/01
Oct-15
Coalition
Jul-15
Apr-15
Jan-15
Oct-14
Jul-14
Apr-14
Jan-14
Oct-13
Jul-13
Apr-13
Jan-13
Oct-12
Jul-12
Apr-12
Jan-12
Oct-11
Jul-11
2.5
Apr-11
3
Jan-11
Oct-10
Jul-10
Apr-10
Jan-10
Oct-09
Jul-09
Apr-09
3.5
Jan-09
Oct-08
4
Reasons for ESA sanctions (thou.),
monthly after challenges Conservative
govt
FTA mandatory
interview
FTP work related activity
2
1.5
1
0.5
0
6000
5000
ESA sanctions (before challenges)
and hardship payment awards, 2011-15
ESA sanctions before challenges
(est.)
ESA hardship payment awards
4000
Note: ESA figures include hardship
awards to disqualified claimants. See
3000
Harsher regime
2000
1000
0
Universal Credit sanctions
• Sanctions currently apply to 1.2m claimants on
JSA/UC/ESA and 0.42m lone parents on IS
• Universal Credit will add 1.3m part time workers
receiving what was WFTC
• Currently in-work sanctions apply only to 15,000
claimants in ‘In-Work Progression Randomised Control
Trial’, interim report 2017, final 2018
• Under Universal Credit hardship payments become
repayable, in effect sanctions 2½ times as long;
consecutive, not concurrent; must demonstrate
‘compliance’ for 7 days before application; must
reapply each 4-week period; 80% rate abolished
Negative impacts of sanctions
– on the claimant & family
• Damage to physical & mental health up to and
including death
• Hunger & resort to Food Banks (Loopstra et al.
2015)
• Family and friends suffer hardship
• Damage to family relationships
• Debt
• Homelessness
• Increased crime (incl. ‘survival theft’ )
• Longer term loss of resilience
Negative impacts of sanctions
– on the jobfinding process
• Claimants’ & employers’ time wasted through pointless
job applications
• Interference with people’s own strategies
• Employment services worse & people alienated
• Claimants fear to complain re bad treatment
• Many people drop out of benefit & employment
services altogether (only half of unemployed young
people claim JSA)
• Diversion of Jobcentre Plus & contractor resources
• Huge administrative waste (e.g. ‘cancelled’ referrals,
duplication of ASW disentitlement & sanction)
• Diversion of voluntary & other public sector resources
Secret penal system
• Referrals and decisions controlled by Sec of State –
independent adjudication abolished 1998
• Scale of penalties higher than in magistrates or Sheriff
courts
• Claimant’s payments stopped without hearing and before
any appeal
• No legal representation
• Prior to Coalition, 60% of JSA referrals resulted in sanction –
now 80% (98% for ‘not actively seeking work’)
• One quarter of JSA sanctions challenged, 19% overturned
(75% success rate)
• Half of ESA sanctions challenged, 25% overturned (50%
success rate)
• Mandatory Reconsideration (Oct 2013) has reduced JSA &
ESA challenges & all but killed off Tribunals
JSA sanctions for all reasons: Referrals, proportion of referrals
resulting in sanction, and sanctions
18.00
Coalition
Referrals for sanction as % of JSA claimant unemployed govt
16.00
14.00
Originally adverse decisions as % of referrals excluding reserved &
cancelled ÷ 10
Originally adverse decisions as % of JSA claimant unemployed
12.00
10.00
8.00
6.00
4.00
2.00
68/1
Oct-16
Apr-16
Oct-15
Apr-15
Oct-14
Apr-14
Oct-13
Apr-13
Oct-12
Apr-12
Oct-11
Apr-11
Oct-10
Apr-10
Oct-09
Apr-09
Oct-08
Apr-08
Oct-07
Apr-07
Oct-06
Apr-06
Oct-05
Apr-05
Oct-04
Apr-04
Oct-03
Apr-03
Oct-02
Apr-02
Oct-01
Apr-01
Oct-00
Apr-00
0.00
Do financial sanctions work?
• Overseas academic studies show some evidence of small
effects (of different regimes!) on employment
• But DWP does not record destinations of sanctioned
claimants who stop claiming (cf. Wk & Pensions Comm
recs) – off-flow is sole management target
• No evidence to support longer sanctions
• No evidence to support escalation of repeat sanctions
• No evidence to support financial sanctions as opposed to
other means of influence (e.g. simple job search monitoring
– McVicar 2010)
• Conversely, much evidence that sanctions make getting a
job more difficult for many claimants
• And that sanctions push people into worse, less well paid
and less stable jobs (Arni et al. 2012)
• UK’s ‘jobs miracle’ has meant dire productivity and low
wage growth
Statistical misrepresentations
• Monthly sanctions rate presented as if it is
‘ever sanctioned’ rate
• All data are for post-challenge sanctions - this
understates stoppages of money by about
20% for JSA, 25% for ESA
• Misquotation of research
How likely is a JSA claimant to be sanctioned? - different stories
25.00
Monthly average JSA sanctions as % of claimants (before appeals) (approx.)
Monthly average JSA sanctions as % of claimants (after appeals)
20.00
The claimants'
experience
% of JSA claimants sanctioned in the year (before appeals) (approx.)
% of JSA claimants sanctioned in the year (after appeals)
15.00
10.00
65/01
5.00
Figures publicised by DWP
0.00
2007/08
2008/09
2009/10
2010/11
2011/12
2012/13
2013/14
2014/15 (Jul-Jun)
Misunderstanding actively promoted
by DWP
• DWP standard statement ‘Sanctions are only used as a last
resort for a tiny minority who don’t follow the rules .... they
are only applied as a last resort’
• Frequently repeated by Conservative MPs
• BUT one quarter of claimants is not a tiny minority
• There are no procedures to make sanctions a last resort
• Scottish Parliament 5 Oct 2016 – Dean Lockhart MSP (Con.)
‘sanctions affect only a small number of claimants. Fewer
than 2.5 per cent of JSA claimants and only 0.26 per cent of
ESA claimants are sanctioned. Perspective is important.’
The case for a monthly rate
• Monthly claimant count denominator is:
– simple
– easily understandable
– familiar and much used elsewhere
– readily available for different areas & categories of
claimant
• Reveals trends quickly
• BUT typical JSA claim duration is 5 months
• So does not correspond to commonsense idea
of how likely is a claimant to be sanctioned
Monthly rate (2)
• DWP Sanctions Statistics Publication Strategy:
‘with the data available to our statisticians,
the best estimate ... is a monthly figure of the
number of individuals sanctioned in a month
by the number of claimants in that month.’
• Lone parent IS sanctions to be changed from
annual to monthly basis
Why post-challenge data?
• Current statistical system adopted April 2000,
with end of independent adjudication
• Previous Adjudication Officers’ data showed
initial decisions, reviews & appeals separately
• April 2000 system adopted ‘latest status’ basis –
reasons not known (to me, anyway)
• DWP appear to like the resulting understatement
• But true, temporally correct picture on original
decisions, reviews & appeals cannot be recovered
from the published data
• Does the system actually discard data on earlier
statuses, or could it be reprogrammed to produce
the full story?
Gaps in the statistics
• Escalated penalties depend on further ‘failures’ within
a year
• BUT DWP does not record date of ‘failure’, hence no
data on actual length of sanctions, in particular no data
on 3-year sanctions
• No regular publication of data on hardship payments
• No data on Universal Credit sanctions & no date for
them
• Excessive reliance on Freedom of Information requests
or Parliamentary Questions
– Inefficient
– Lack of comparability
– Hard to find
Misquotation of research
• Key examples:
– ‘72% say more likely to follow rules due to
sanctions’ – statement of the obvious, begs
question whether following the rules does any
good, answer: it doesn’t
– ‘the vast majority of claimants thought that some
or all of the actions they accepted as part of their
Claimant Commitment would increase their
chances of finding work, took account of their
personal circumstances, and were achievable’ –
but they get sanctioned if they don’t do any one
of the actions, and for each of the three about
half thought the statement was not true
Complaint to the UK Statistics Authority
27 July 2015
• 5 heads of complaint:
– Systematic understatement of the number of sanctions
and of the number of claimants affected.
– Gross and systematic misrepresentation of sanctions as
affecting only a small minority of claimants when the data
show that they affect a large minority (this was Jonathan
Portes’ complaint but the problems go much wider than he
indicated).
– Systematically biased commentary on the question
whether sanctions have been increasing or reducing.
– Omission of key data on Universal Credit sanctions,
‘hardship payments’ and repeated sanctioning of the same
individuals from the suite of publicly available statistics.
– Misleading use of the phrase ‘actively seeking work’.
UKSA response & state of play (1)
• Provide users with benefit sanction statistics
based on the actual number of sanctions
applied, making clear the numbers of reviews,
reconsiderations and appeals
• This would require changing the fundamental
architecture of the DWP’s system – couldn’t
be immediate
• But – no action at all!
UKSA response & state of play (2)
• Make clear the limitations associated with the
statistics
• Some minor improvements to metadata but
nothing that would stop the major
misrepresentations
• Metadata remain extremely poor
UKSA response & state of play (3)
• Include in the quarterly benefit statistics bulletin a
statement of the proportion of JSA claims subject to a
sanction, as well as the proportions of claimants who
have been sanctioned during the most recent one-year
and five-year periods, and the numbers on which these
proportions are based
• This would be easy for DWP to provide – algorithm
already exists. But no action.
• Separate IS Lone Parent Regime sanctions data to be
published on monthly rather than annual basis
• DWP ‘are currently considering how information could
be published on the likelihood of a claimant being
sanctioned’ (Benefit Sanction Statistics Publication
Strategy, 19 April 2016)
UKSA response & state of play (4)
• Ensure all statements made using the official
statistics are objective and impartial and
appropriately apply the definitions of the
variables underpinning the data, including
‘actively seeking work’
• No improvement
UKSA response & state of play (5)
• Extend the range of benefit sanction data
available by addressing the gaps in information
on repeat sanctions and hardship payments,
alongside the development of sanction data from
the Universal Credit system
• Repeats requires system change – couldn’t be
immediate – but no action
• Regular publication on hardship payments would
be easy – but no action
• UC sanction data said to be under development
but no commitment to date or content
UKSA complaint – where now?
• UKSA’s only formal sanction is withdrawal of
‘National Statistics’ badging
• Unfortunately this would push the statistics
even further from public scrutiny
• Also problem of limited public interest –
consultation 2012 drew only 5 responses
• Best strategy is probably continuing adverse
publicity
Problems of historical comparability
• Key question is how does contemporary rate
of sanctions compare with the past and why –
not a merely theoretical issue
• Key difficulties:
– Two separate systems prior to 1996:
unemployment benefit (UB) & unemployment
assistance (UA)
– Regular stats never published on UA
sanctions/disqualifications – odd mentions in PQs,
SBC annual reports etc.
– Long intervals when no UB stats published either
Historical comparability (2)
• Key problems (cont.):
– Numerous years/quarters of published UB stats 19862000 missing from all UK official/academic/public
libraries inc. DWP – last hope is HofC Library/NAO
– Denominators for UA and UB sanction/
disqualification rates often hard to obtain
– Interpretation of stats depends on changing statutory
provisions
– Particularly frequent & complex changes to UB rules
1918-35 – around 40 UB Acts in this period
– Prior to 1935 UA was the Poor Law – very complex
with big local variations
Historical comparability (3)
• However for UB it is clear that only one short
previous period stands any comparison with the
great sanctions drive 2010-16, namely the 192730 ‘genuinely seeking work’ campaign (cf. Deacon
1976)
• Next most persecutory period for UB was 198996 – should be possible to make precise
comparisons when missing vols available
• Statistical analysis shows that the scale of the
2010-16 drive resulted from (a) ministerial
pressure on staff (b) Work Programme rule that
any breach of requirements must start sanction
process
More information
• My Child Poverty Action Group webpage contains most of my statistical
analyses http://www.cpag.org.uk/david-webster
• My Sanctions project webpage describes the project & has a few refs
http://www.gla.ac.uk/schools/socialpolitical/research/urbanstudies/projects/ukbenefit
disallowances/
• Guest blog on the statistical misrepresentations:
•
•
ESRC Welfare Conditionality project
http://www.welfareconditionality.ac.uk/2016/04/tackling-britains-misleadingbenefit-sanctions-statistics/
• Complaint to UK Statistics Authority & their response
•
https://www.statisticsauthority.gov.uk/correspondence-list/, 27 July & 6 August
2015
• CPAG webpage with voluntary sector reports on impact of sanctions
http://www.cpag.org.uk/content/sanctions
• Email me at: [email protected] – ask to be put on my mailing list