Lay off Concentrix?
In the last few days the gathering storm over HMRC out-sourcing some of their compliance activities to Concentrix has finally broken, apparently prompted by this screaming headline from the BBC’s Victoria Derbyshire programme
While such exposures and complaints are by no means new, this one seems to have led to a very sudden announcement by HMRC that they won’t be renewing the £75m contract when it expires next year. Concentrix are outraged to have discovered this only from news reports and doubtless it will come as a nasty shock for their hundreds of workers in Belfast who on the face of it will be out of a job. And the latest screamers are calling for a National Audit Office enquiry into the contract. I’ve been meaning to write something about the apparent (lack of) success of the contract for some time, so here it is ….
As part of its ambitious plans to run down its workforce, HMRC shut down one of its main sites in Northern Ireland. Staff there had carried out some key elements of operating Tax Credits. I’m not entirely sure how the timelines intertwine, but around the same time Concentrix was awarded a contract to carry out in Belfast similar work checking tax credits awards for irregularities, in particular looking for undeclared partners whose income should be taken into account when calculating what a claimant is entitled to. This is from the original tender document …
Having been awarded the contract, Concentrix then went around hiring staff to do the work. Here’s an advert for a Team Leader. The print might be a bit small for some screens, but it offers £17k to 20k to work an all-week shift pattern managing a team of Case Workers and Senior Case Workers.
At the time an Executive Officer II (lower paid than Executive Officer I) in the Northern Ireland Civil Service would get rather more than that. Typically this is the lowest level at which anyone would be leading any kind of a team. There are only two grades below EO: Administrative Officers and then Administrative Assistants. As Concentrix Team Leaders have both Senior Caseworkers and Caseworkers below them this means in terms of hierarchy and responsibility they cannot realistically be anything less than EO equivalents. But they’re getting paid about 25% less.
I do remember George Osborne saying that his policy of civil service cuts would result in the creation of new private sector jobs. I’m not sure whether this is the kind of thing he had in mind, as on the face of it public sector jobs were directly replaced by considerably lower-paid private sector jobs. Still, in the absence of those public sector jobs I suppose that like David Cameron said in March 2014 “Most important of all, [a job] gives you a sense of security and the peace of mind that you can support yourself and your family.” Whether that is true of any and every job I’m not sure either.
Concentrix don’t seem to have had too much difficulty recruiting people as they were ready to open for business on time. Unfortunately they had some teething troubles as their IT systems couldn’t initially connect to HMRC’s as was reported at the time in Accountancy Age. Notably this reports the average full-time salary of a mere £15,000. So presumably the Caseworkers, if not the Senior Caseworkers, weren’t getting paid a great deal more than the minimum wage, while working out of the very building abandoned by HMRC. Still, they were properly in business from December 2014.
So once up and running what actually happens when Concentrix find an undeclared partner? The contract tender document said that the contractor would then amend the award on HMRC systems as if they were HMRC staff. Well HMRC estimate that the average overpayment in an undeclared partner award is about £2,900 (see below). Amending the award really means essentially no more than recording an overpayment to the claimant of that amount. HMRC doesn’t actually get any money from Concentrix, just more debt to add to HMRC £7.1bn tax credits debt mountain! The National Audit office report that HMRC estimates that only £2.9bn of this is recoverable, implying that only 40% of identified overpayments will be recovered.
Further, most overpayments are recovered from ongoing awards or ‘time to pay’ arrangements, because most claimants are actually not well off and if they’ve been over-claiming the average ‘undisclosed partner’ amount they’ve been getting £55 a week too much over a year. I’ve looked at lots of such cases and that kind of money just gets spent in getting by with your kids on a very low income. How many could afford to pay back a lump sum of £2,900? Virtually none. So it’s not very meaningful when the Concentrix spokesman says “Through the term of the contract we are pleased to have saved the taxpayer nearly £300m”. They will instead have identified £300 million of overpayments and have only ‘saved the taxpayer’ whatever turns out to be recoverable by HMRC over time. All HMRC seem to have been paying for is the identification of overpayments, and even on payment by results I do wonder what they will actually end up paying for each £1 identified. Their repayment framework makes clear that some overpayments really do take quite a long time to recover.
So if £300m is identified, the normal rate of recoverability would take that down to £122m which with recovery mainly through ongoing awards and Time to Pay arrangements means only tens of millions a year. That’s very little bang for buck compared with what HMRC gets from investment in compliance activity it carries out itself. I know the contract was let on a payment by results basis but have no idea on exactly what basis and whether there was an underpinning minimum payment provided for in the contract. But certainly it doesn’t seem to have been worth anything like the headline £75m.
So does this mean that Concentrix just aren’t any good? Their spokesperson certainly seemed to think this is what HMRC are implying as they fiercely reacted to the announcement of the contract termination …
The HMRC statement not to renew the contract attacks our professional credibility
And he or she continued
In addition, throughout the contract, Concentrix has employed good hard-working people within the UK, at Concentrix expense, in order to staff phone lines and handle customer calls which were agreed by HMRC and were based on HMRC assumptions
Ah — HMRC assumptions: this is where it gets interesting.
When the contract to carry out these checks was put out to tender on a payment by results basis, any potential bidder would obviously have asked themselves, and indeed HMRC, what results were possible. HMRC estimate losses through tax credits error and fraud each year through their Tax Credits Error and Fraud Analytical Programme (EFAP). This takes a sample of 5,000 finalised cases and gives them to specialist compliance officers to go through with a finetooth comb. HMRC publish the results annually, as one of their key objectives is reducing the amount of losses to error and fraud. The most recent publication estimates 140,000 undeclared partners in 2014/15 costing £405m. If there is that rate of overpayment annually, it is in line with hoping to save £1bn over three years if Concentrix identified nearly all of the undeclared partners every year.
Now if HMRC told potential bidders that there was around £400m a year being overpaid to claimants with undeclared partners then clearly bidders would build up their business case on assumptions about how many of them could be identified with a given amount of investment in staff and systems. Obviously the forecasts for savings seem to have assumed that a check of all the cases that might possibly involve an undeclared partner would uncover all (or almost all) of the undeclared partners. But it looks as though Concentrix came nowhere near that. In fact, in nearly a year and a half they seem to have found only 19,000 of them.
That was the result of investigating nearly a quarter of a million cases, and while there were still 95,000 outstanding, the ostensible hit rate of only 1 amended award in every 12 completed cases seems very low indeed.
I’m sure that Concentrix applied a bit of special algorithmic technology to their selection of cases and had access too to the enormous store of information HMRC get from Experian and others to spot people with financial and accommodation links. But as there are only about a million singles claiming Working Tax Credit then if EFAP is correct to estimate around 140,000 undeclared partners then Concentrix has done rather worse than if they had simply selected cases at random.
So there seem to be two broad possibilities: Concentrix were really bad at identifying cases with potential undeclared partners or they were sold a pup because there weren’t actually 140,000 undeclared partners in the first place and nothing like £400m a year was actually being overpaid. I’m sure that under ‘Contingencies and Risks’ in Concentrix’s business case someone will have included that HMRC might have overestimated the number of fish in the pool, but perhaps he or she thought that it couldn’t possibly be by an order of magnitude.
Certainly Concentrix seem to have done all they can. Indeed, it’s precisely the very wide cast of their net that has caused all sorts of upset to innocent people. Assuming they concentrated on singles claiming Working Tax Credit, then up to a quarter of them seem to have been scrutinised so it’s very hard to see that they will have cast the net in completely the wrong place.
This is something that both HMRC and the NAO need to think about. If the EFAP pointed to there being 140,000 undeclared partners but essentially a 100% check of claims where this was at all likely found nothing like that number then I’d put my money on the extrapolations from EFAP being wrong. This is clearly very significant to assessments of HMRC’s performance (better than thought as they weren’t losing as much as they thought) and to where resources can be best allocated (undeclared partners being less of a problem than thought). One might also wonder that if the EFAP is wrong on undeclared partners, whether it is correct on its estimates for other kinds of error and fraud.
In conclusion it seems to me that calls for an NAO enquiry into some perceived failure of Concentrix to deliver might be rather wide of the mark…
Update/addendum
Just on the point of the intrusiveness of Concentrix enquiries and the burden of getting a year’s worth of bank statements, bills etc together, Concentrix appear to ask for exactly the same body of evidence of not having an undeclared partner as HMRC. The greater visibility of their activities is just because, with their ‘payment by results’ resources, they can afford to check out near enough everyone who the credit-checkers and algorithmic systems throw up as a possibility for investigation, whereas HMRC only looked at a percentage of them, and presumably the most likely. It would be quite instructive of course to know how many undeclared partners HMRC typically identified themselves in previous years, but I don’t think that information is in the public domain.
