Further proof arrived this morning of one of the pernicious effects of privatisation of state functions and particularly the use of payment-by-results. A report by the public accounts committee found the private companies running the work programme for Iain Duncan Smith's employment allowance project were parking hard cases and focusing instead on easier ones to maximise their profit.
This is a common pitfall of privatised systems. Once you introduce a profit motive, companies naturally ignore hard-to-reach cases and focus on easy ones.
That's why private firms fall over themselves to bid for knee operations in health services, while people suffering from hard-to-assess ailments are failed. It's why paying people to stop former prisoners reoffending is a bad idea: they tend to ignore the people with a long history of abuse, mental health problems, and low literacy levels and focus on the people who are more likely to make them their money back.
Very often, private firms underbid for massive public sector contracts, pricing out charities from competing. They then subcontract out the difficult, risky, non-profitable cases to charities. A report by the National Coalition for Independent Action found charities were being forced into imposing zero-hours contracts on staff, asking unpaid volunteers to take on extra responsibilities and prioritising their service around "a narrower spectrum of people in need".
Many charities used to be 'prime' organisations in getting the long-term unemployed into work. But under the coalition they have been muscled out by larger private firms and relegated to sub-contractor work with the toughest cases. They call it 'creaming and parking'.
Today's report found that IDS's policy of paying more when firms get hard-to-reach cases into work was not working.
Committee chairman Margaret Hodge said:
"Evidence shows that differential payments have not stopped contractors from focusing on easier-to-help individuals and parking harder-to-help claimants, often those with a range of disabilities including mental health challenges.
"Data from Work Programme providers shows that they are, on average, spending less than half what they originally promised on these harder-to-help groups.
"It is a scandal that some of those in greatest need of support are not getting the help they need to get them back to work and are instead being parked by providers because their case is deemed just too hard."
Providers aren't even made to collect and publish information on how much they're spending on each payment group. This is standard for privatised contracts, the odds are usually stacked in their favour.
The Department of Work and Pension's own interim evaluation concluded there is no clear relationship between payment groups and the nature of the support which participants receive. That suggests differential payments have not been effective. The hard-to-help claimants, those with a range of mental health problems or disability are being failed by the system.
IDS originally estimated that 22% of employment and support allowance claimants would find their way into a job, then he revised it down to 13%. It's currently 11%.
The department promises to do better. It told the committee it would focus its new performance management regime on the hardest-to-help, for instance by requiring contractors performing badly to specifically improve their results with that group.
Don't hold your breath. The reason privatisation fails hard-to-reach groups so badly is the same reason it fails complex health ailments. These cases do not make a profit and they are still unlikely to at the new level of payment required for the system.
Public money spent on such cases eventually pays off. Down the line a person in work contributes more to the Treasury and takes more weight off the benefits bill. But for a private firm that long-term economic benefit does not apply, so they lack the incentive to do the job.
This is a long running feature of privatised public projects. The failure of the welfare to work programme suggests this is unlikely to change.