CEST being right in 22 out of 24 cases shows it’s working. Right?

On August 22nd 2018, HMRC published a response to a request under the Freedom of Information (FoI) Act, for details of status cases that have been tested against its Check Employment Status for Tax (‘CEST’) tool, writes Matt Boddington of IR35 advisory Chartergates.

There is little HMRC commentary to accompany or explain the document, other than as follows:

The Check Employment Status for Tax (CEST) rules were developed in a workshop. The information provided here is a list of test results from tax cases produced afterwards.

The document itself is a list of 24 historic employment status and IR35 cases decided by the tax tribunals, with a note of what employment status outcome the CEST tool predicted, as well as the outcome predicted by its predecessor, the old Employment Status Indicator (‘ESI’).  The full list of cases is reproduced for reference at the end of this piece, with a couple of typos corrected.

In terms of IR35, there are a few cases missing from the list – notably the 2004 IR35 case of Ansell Computer Services v Richardson which was lost by HMRC; the JLJ Services case, and also the three IR35 judgments issued so far this year.  However, the JLJ appeal was part-allowed so it may be difficult to apply CEST to a split outcome, and we don’t know when this FoI request was made. or when CEST was tested, so let’s give HMRC the benefit of the doubt on the recent cases. 

This does however raise the issue that we don’t know which version of the constantly evolving CEST tool these tests apply to. Having said that, one immediately apparent characteristic of this table is that there is apparently no difference between the old ESI online tool, and the new and apparently improved CEST tool in terms of employment status outcome, so maybe it’s irrelevant anyway.

Is this table evidence that CEST is accurate?

At first glance of the HMRC-produced table, the CEST tool appears to fare pretty well. Out of the 24 cases the correct outcome was predicted by the CEST tool 22 times, making it 92% accurate over this relatively small sample of cases.  Not perfect – extrapolated, this is still tens of thousands of incorrectly assessed contractors – but not bad.

However, this is far from the entire picture. We still know that the tribunals give much greater significance to the mutuality of obligation (“MOO”) test than the CEST tool, which gives none at all. HMRC’s tool has dodged a bullet here because IR35 cases like ECR Consulting, Marlen and MBF Designs which were all decided on MOO happen to pass the test anyway on other factors. But the judgments are clear that MOO is a major consideration. 

It remains a fact that borderline cases where a lack of MOO could be the decisive factor will go the wrong way under the CEST tool.

This problem becomes vastly exacerbated under the IR35 reforms. Almost every IR35 case litigated by HMRC involves a relatively long-term contractor, typically engaged under one or more six-month contracts. Since the April 2017 reforms, we have seen public sector bodies using the CEST tool in relation to very short, sometimes single-shift agency engagements, including those that can be cancelled without recompense and at no notice.

The argument for a lack of MOO in these types of arrangements is clear, but because the CEST tool does not recognise the concept there is a risk that many agency workers are being incorrectly classified. Thousands more, including many ‘gig economy’ workers, are set to be taxed incorrectly if the reforms are extended into the private sector and end-clients rely on the CEST tool to make their assessments. So here’s our top tip: don’t rely on CEST if it tells you a contract is inside IR35, because CEST has not taken into account MOO.

HMRC’s bias exposed

There are two further themes highlighted by this list. The first is to shine a spotlight on just how few employment status cases are litigated by HMRC: out of the list of 24 cases, only 6 are ‘traditional’ employment status disputes and the rest relate to IR35. We expect the number of IR35 investigations to rise considerably if the private sector reforms are enacted, as HMRC will have a much more substantial target to pursue with their new ‘scattergun’ approach.

However, the most startling fact revealed by this list is that in 13 out of the 24 cases, HMRC argued unsuccessfully for employment when even their own CEST tool would have found the worker to be self-employed. This begs the question: why did HMRC bother to litigate these cases? The ESI has been available for many years and reflects HMRC’s even longer standing position on employment status.

There are two possibilities, and neither reflects well on HMRC. The first is that HMRC had all the facts, knew that the workers were self-employed and prosecuted the cases anyway.

The alternative possibility, given that HMRC presumably ran the facts from the tribunal judgments through the CEST tool quite recently, is that at the time of bringing these cases, HMRC simply refused to accept the facts that were ultimately found by the tribunals.  In short, over half of these cases were unnecessarily litigated due to intransigent HMRC officers who just did not accept the taxpayer’s description of their engagements. In one of the cases, Larkstar, HMRC even appealed the judgment to the High Court.

This is a major concern given that the IR35 compliance cost burden falls disproportionately heavily on small Personal Service Companies. All these cases ended up in the tribunal and rumbled on for years. It should not have been necessary in such a high proportion of cases to incur the costs and inconvenience of a tribunal hearing when the parties were all in agreement on the application of the law. This is an indictment of both the difficulties of applying the IR35 ‘hypothetical contract’ test and HMRC’s bias in employment status cases.

As the tribunal judge in one of the cases from the list, MAL Scaffolding, observed:

“[HMRC] appear to have approached their investigations on the basis that there must be an employment relationship between MAL Scaffolding and the workers there if one looks hard enough. Officers then went looking on that basis and persuaded themselves that they had found that for which they went looking. They have totally failed to persuade me.”

However effective HMRC claim the CEST tool to be (and to be clear – our position remains that the CEST tool is woefully deficient), it will be effectively useless unless employment status investigations can be approached impartially by HMRC.