Perverse Incentive

A perverse incentive is an incentive that has an unintended and undesirable effect, that is against the interest of the incentive makers. While perverse incentives are equally possible in government-run delivery of services, this article focuses on their application to delivery by third-party providers.

Cherry picking, Creaming and Parking

The most well known effects of perverse incentives are also the most ambiguous. Cherry picking, creaming and parking all rely on the idea that targeting help towards more employable customers is cheaper and more effective than trying to help less job ready ones. It is arguable as to whether this is a genuine perverse incentive - these practices are thought to maximise the number of job outcomes from a group of customers for a set amount of resources, which is presumably in line with the government's intentions:

  • Voluntary provisions which recruit their own customers or have a pre-referral interview allow providers to 'cherry pick' the customers most likely to move into employment with the least assistance.
  • Creaming means targeting delivery resources at the most employable customers. The idea is that more employable customers are cheaper to help and more likely to find work. Customers who are deemed unlikely to find work are 'parked' and given minimal support.

Criticisms - These practices effectively undermine the purpose of referral to provision for 'parked' customers, removing support for society's most deprived and marginalised groups. Additionally, the impact of only moving the most employable customers into work is that a 'hard core' of long-term unemployed customers is likely to emerge, reducing the effectiveness of delivery over time as these customers cycle through provision after provision.

Example - In the New Deal for Disabled People Third Synthesis report, the DWP identified evidence of cherry picking in NDDP recruitment, although it was not referred to in those terms. No judgement was made as to the desirability of this. As part of the recent recontracting for NDDP, the guideline payment for moving a person into work went down substantially, which is likely to increase the pressure on providers to take on job ready IB claimants rather than people in need of long-term, intensive support. It's worth noting that NDDP is one of a range of provisions targeting claimants with disabilities or health conditions, and may not be the most appropriate for people with specific needs or barriers in any case.

Stuffing

A more unambiguous example of perverse incentives occurs in provisions that do not limit registrations and have an absolute outcomes requirement. This opens the way for providers to register as many customers as possible, and rely on them moving into employment regardless of provider support. This is unhelpful as it has no real impact on unemployment rates. Virtually all DWP contracts are designed to avoid this trap, although there have been some that, by measuring job outcomes against leavers, have encouraged providers to register large numbers of people and leave them on the rolls until the end of contract, making the performance of a provision appear better than it actually is. Needless to say, this doesn't work in the long term.

Rumours abound that some FE colleges meet their LSC targets by holding off from registering starts who then drop out, a practice occasionally called an 'active withdrawals policy'. This would not work in DWP provisions, which generally require notification of customer start shortly thereafter.

Unsustainable job entry

Many current provisions measure and pay against job entries, rather than job sustainment. This makes it worthwhile for less scrupulous providers to place people in employment who they know will fall out again, or even to try and claim temporary assignments as job entries. Recent government figures show that around 50% of customers starting a job fall back onto benefits again shortly thereafter. The obverse of this is that paying on long term job sustainment requires providers to fund their operations for a long time before they get paid. Additionally, measuring sustainment can be expensive as it requires providers to keep chasing up (frequently unco-operative) customers for a long time after they enter work.

A similar issue arises with short-term sustainment measures. If a provider is paid for a customer who is still in work 16 weeks after they first start a job, it may pay providers to place them into a second job at the 15 week point, without any expectation that the second job would sustain either.

Discouraging job entry

This is fortunately not common in current DWP provision. It has occurred in LSC provision with job outcome requirements that were tied to specific job types, such that the provider is penalised if a customer takes up a job offer that doesn't match the LSC's job requirements. There were also issues with performance measures on some Jobcentre Plus provisions at one time, where a customer starting work during a provision was not counted as a completer and thus damaged the performance figures.

A related issue arises with the hours requirements on job entry payments. If a provider is paid when a customer starts work over 16 hours a week, then it may be rational for the provider to discourage customers from accepting work at less than those hours, even if it is in the customer's best interest.

Overpromising and Underdelivering

In certain circumstances it makes sense for providers to deliver a poorly funded, low quality service that does not meet contract targets. This depends on DWP willingness to tolerate underperformance, a lack of financial penalties to make underperformance unattractive, and a bid process that does not verify past performance.

The bid process by which providers win new contracts is also open to perverse incentives. Overpromising is common, as the bid winner is generally the one who convincingly promises the most added value (on delivery content, price or performance). Historically, since actual delivery content was set by the government at a national level, the management of contracts was often against the national requirements and promises of added value were not followed up on.

The impact of the new Commissioning Strategy

The new Commissioning Strategy should have the following impact on the listed behaviours:

  • Cherry picking, creaming and parking - The emphasis on outcome payments may encourage creaming and parking, although the 5-7 year contract length means that providers will likely face an increasing number of parked customers.
  • Unsustainable job entry - Possibly the area set to benefit most from the new approach to payments. Paying on 6 month sustainment or longer should end the practice of placing people into unsustainable jobs.
  • Overpromising and underdelivering - The opening up of performance figures and the centralisation of contract management should substantially increase accountability, and the increasing flexibility in delivery content will enable the focus to shift to the more important issues of performance and quality of service.

Disagree? Have more material to add? Then log in and edit it!