With the very kind assistance of APSCo and two members of HMRCs Employment Status and Intermediaries Policy team, we recently held two client workshops on the practical issues arising to end-users and recruiters from “new” IR35.

Our guests made it clear that pending the implementing legislation and the accompanying Guidance (current ETA for both, July), they were expressing views only, rather than settled policy. However, while none of this is therefore gospel, there were still some very useful snippets:

  • How do you tax a VAT invoice? How does the end-user even know the rate of tax to be deducted? If as a PSC, I submit a bill for £5,000 plus VAT, how much do I get back? Current indications are that the VAT will be payable as normal, i.e. £1,000, and that PAYE income tax (probably basic rate) is then deducted from the £5,000 (so you are in effect paying VAT on tax, surely the Treasury’s ultimate dream). Therefore I am going to get something under £4,000 back of my £5,000 bill. But my contract provides for £5,000, so why could/would I not now sue the end-user for the difference, especially if I don’t agree that IR35 should apply to me? HMRC said that the issue has not come up in IR35’s application to the public sector, but that doesn’t mean it won’t. So a drafting tip – if you are entering a new contract with a PSC between now and April, it may be sensible to state expressly any fee payable after that date will be less such tax, if any, as you are obliged to deduct.
  • HMRC told us (subject to the caveat in the opening paragraphs) it would probably not look backwards from April 2020 to see if the contract terms being applied, especially as might be implied from the parties’ behaviours, shed any light on the question of hypothetical employee status. It will only look at the contract and working practices as they stand from April, it said. We shall see if that commitment makes the Guidance, bearing in mind at the same time there was a clear refusal to provide any form of amnesty for the past and that such an approach would effectively deny HMRC the ability to rely on the parties’ conduct for many months (long enough to establish whether it shows the written terms not to represent the reality) after April. HMRC fully expects the vast majority of PSC contracts to be re-written between now and next April, but emphasised that dolling up the paperwork will not help if employee-type working practices carry over unaltered.
  • Contracts for the supply of services, as opposed to labour, can side-step IR35. HMRC was at pains to say that simply describing your consultancy contract with the PSC as being for services will not be enough to prevent an adverse finding if the reality is otherwise.
  • It is unclear how far the existence of other clients of the PSC would prevent the application of IR35 to its supply of the contractor if, on that particular assignment, he/she is doing something which could otherwise be an employed role. The question is integral to the issue of worker status, since that excludes people in business on their own account, so it would be bizarre if it were not also relevant to the IR35 assessment. It is to be hoped that HMRC does not go down the “economic dependency” path trodden by recent ET decisions, since that is far too vague an indicator on which to base decisions of such potential significance.
  • Much was made of the virtues of reliance on HMRC’s CEST (Check Employment Status for Tax), an online application for getting at least a preliminary indication of which side of the line you fall. It has a number of well-published shortcomings so far (entering the facts of decided cases produces a material number of “wrong” answers, and in about 15% of cases – presumably the difficult ones – it simply can’t decide). Further work is being done on it but, nonetheless, HMRC said that it would respect the outcomes of the CEST provided that the questions in it are correctly answered. The problem with this is that CEST currently seeks to produce an answer out of just 16 questions on what can be very complex and nuanced indicators. It is not hard to see end-users doing their level best to apply what is still a fairly blunt instrument to a factually very delicate situation, and then relying on CESTs answer to their own detriment.