Authors:Jagna Olejniczak
Created:2023-04-28
Last updated:2023-09-26
No data, no problem: the DWP’s culture of secrecy
.
.
.
Marc Bloomfield
Description: PLP
On 6 April 2023, the Department for Work and Pensions (DWP) reluctantly released The impact of benefit sanctions on employment outcomes: draft report, which found that sanctions ‘decrease the rate of exit into higher paid work’ (page 3). It long refused1Patrick Butler, ‘Report on effectiveness of benefit sanctions blocked by DWP’, Guardian, 27 January 2022. to publish its findings and did so only after a ruling of the information commissioner2Freedom of Information Act 2000 (FOIA) Decision Notice IC-162699-X6D9, 6 March 2023. compelled it to do so. In the meantime, the use of sanctions has been growing3Benefit sanctions statistics to October 2022 (experimental), DWP, 14 February 2023. despite existing evidence that they do not work.
This is not an isolated example of the DWP refusing to put information in the public domain. As explored below, it often seeks to justify such refusal with one of two reasons: either the requested information is not readily available and making it available would cost too much; or the information exists but providing it would not serve the public interest. These practices push the burden of discovering centrally held information onto civil society and hinder the ability to scrutinise the effectiveness of DWP policies and their impact on groups such as people with protected characteristics.
‘Disproportionate cost’
Some 246 of the DWP’s answers to parliamentary questions from March 2022 to March 2023 cited ‘disproportionate cost’ as a reason for not collecting and disseminating data. For example, the number of children affected by benefit sanctions is not provided because sourcing it would ‘incur disproportionate cost’.4House of Lords Written Question UIN HL5909, 27 February 2023; answered 14 March 2023. This is despite such information being vital to assess the impact that the policy has on children.
Interestingly, this justification is also applied to data that used to be published in the past. Sanctions appeals statistics (last released in 2018) showed a minuscule (0.3 per cent) number of sanctions appeals, but a very high (81 per cent) success rate.5Caroline Selman, Benefit sanctions: a presumption of guilt, PLP, July 2022. This indicated that sanctions were often wrongly applied and people refrained from appealing them for various reasons. These figures were impossible to monitor from the following years as the department suddenly concluded that ‘to provide [them] would incur disproportionate cost’.6House of Commons Written Question UIN 45954, 2 September 2022; answered 27 September 2022.
Similarly, statistics on the length of universal credit sanctions (last updated in 20217Benefit sanctions statistics to January 2020, DWP, last updated 10 June 2021.) were withheld because of costs implications.8House of Commons Written Question UIN 74765, 31 October 2022; answered 8 November 2022. This made it easier for the DWP to downplay the consequences of sanctioned people being left out of cost of living payments, which serve as a crucial lifeline for those on low-income benefits in the face of spiralling price rises.9Jagna Olejniczak, ‘Cost of living payment – sanctioned UC claimants face double punishment’, PLP blog post, 8 March 2023. Dr David Webster estimated that the average length of sanctions has risen to around 11–12 weeks,10Dr David Webster, Briefing: benefit sanction statistics February 2023, Child Poverty Action Group, 2 March 2023, page 6. which suggests that more claimants will not get the payment because of receiving a £0 universal credit payment in a given month.11See further Jagna Olejniczak, ‘Cost of living payment – sanctioned UC claimants face double punishment’, PLP blog post, 8 March 2023. If data on average length were available, it would be easier to forecast the scale of the most vulnerable missing out.
‘Public interest’ and ‘gaming’ the system
Withholding information is also justified by public interest. The report on the effectiveness of sanctions remained undisclosed because it included ‘details of a sensitive nature’ and was ‘likely to prejudice the effective conduct of public affairs’.12FOIA Decision Notice IC-162699-X6D9, 6 March 2023, para 8, page 2. Little explanation was given about how these concerns could materialise apart from stating that limited evidence in the report would hinder comprehensive policy assessment. The information commissioner suggested that the decision came close to being ‘irrational or absurd’.13FOIA Decision Notice IC-162699-X6D9, 6 March 2023, para 20, page 4.
The lack of transparency around automated decision-making (ADM) used by the DWP to detect benefit fraud is a particularly concerning practice. For example, claimants suspected of fraud have their benefit payments suspended for up to 11 months but are not told that the decision is made by the algorithm.14Hansard HC Debates vol 707 col 392WH, 26 January 2022. There is also an emerging pattern of some groups being disproportionately impacted (eg, Bulgarian nationals). The DWP dismissed the request to publish the impact assessment of using ADM because it would ‘potentially enable third parties to work around our models ... to commit crime’.15DWP response to FOIA request from Caroline Selman, 29 September 2022. Such blanket refusal is an excessive response to purported gaming threats,16Alexandra Sinclair, ‘Algorithmic transparency is not a “game”’, Law Society Gazette, 13 January 2023. further undermining the accountability of the systems used.
The failure to publish certain documents can go beyond bad governance and may be found to be unlawful by the courts. In R (K) v Secretary of State for Work and Pensions [2023] EWHC 233 (Admin), keeping the internal guidance on the benefit overpayments waiver undisclosed prevented claimants from fully understanding the policy or making representations (para 116).
Conclusion
The above examples suggest that the DWP refuses or ceases to provide information when it does not support its policies or when it does not want them to be subject to effective scrutiny, and uses the all-encompassing ‘public interest’ and ‘disproportionate cost’ reasons to justify it.
This practice generates a twofold danger: first, it prevents decision-makers from meaningfully assessing the impact of their proposals; and second, it forces civil society to rely on speculations and their own findings, which should complement official data rather than act as a substitute for it. The ill-reasoned refusal to publish crucial information leads to a lack of engagement in evidence-based dialogue and, instead, to designing policies around ideological whim. Announcing further expansion of conditionality in the 2023 spring budget17Spring budget 2023, HC 1183, 15 March 2023. based on the unsupported conviction that it is ‘fundamentally the right system’18Oral evidence: the work of the secretary of state for work and pensions, HC 549, 29 March 2023, Q135. is a prime example of such practice.
The DWP decided not to appeal the information commissioner’s order and released the problematic sanctions report. This, however, will be a drop in the ocean of data that needs to be made available for public scrutiny.
 
1     Patrick Butler, ‘Report on effectiveness of benefit sanctions blocked by DWP’, Guardian, 27 January 2022. »
4     House of Lords Written Question UIN HL5909, 27 February 2023; answered 14 March 2023. »
5     Caroline Selman, Benefit sanctions: a presumption of guilt, PLP, July 2022. »
6     House of Commons Written Question UIN 45954, 2 September 2022; answered 27 September 2022. »
7     Benefit sanctions statistics to January 2020, DWP, last updated 10 June 2021. »
8     House of Commons Written Question UIN 74765, 31 October 2022; answered 8 November 2022. »
9     Jagna Olejniczak, ‘Cost of living payment – sanctioned UC claimants face double punishment’, PLP blog post, 8 March 2023. »
10     Dr David Webster, Briefing: benefit sanction statistics February 2023, Child Poverty Action Group, 2 March 2023, page 6. »
11     See further Jagna Olejniczak, ‘Cost of living payment – sanctioned UC claimants face double punishment’, PLP blog post, 8 March 2023. »
15     DWP response to FOIA request from Caroline Selman, 29 September 2022. »
16     Alexandra Sinclair, ‘Algorithmic transparency is not a “game”’, Law Society Gazette, 13 January 2023. »
17     Spring budget 2023, HC 1183, 15 March 2023. »