The latest test data published by HMRC to ascertain the accuracy of its updated Check Employment Status for Tax (CEST) tool has been found wanting. HMRC recently published three pages of documentation summarising the test cases where it claims the tool had been applied, but without detail as to how CEST’s questionnaire was answered for each engagement.
The publication comes less than two years after the taxman responded to a Freedom of Information (FOI) request from ContractorCalculator for test data concerning the initial iteration of CEST, returning a similarly inadequate table of cases, tribunal outcomes and CEST’s alleged results.
“We’re still in ‘the dog ate my test data’ territory,” says ContractorCalculator CEO Dave Chaplin, who has over two decades’ software engineering experience, and has led teams of developers on projects for UK financial institutions and legal practices. “HMRC’s test data suggests that CEST is accurate when fed the correct information, but how are we to know when the taxman won’t publish the test input?
“From a software engineering perspective, the CEST test data is shamefully inadequate. It simply would not pass muster at a commercial organisation, so why HMRC believes it is fit for a tool used to determine the tax status of hundreds of thousands of UK contractors is anybody’s guess.”
How has HMRC tested the updated CEST?
Following recent amendments to CEST, HMRC claims to have tested the tool against the 24 test cases initially published in 2018, although not even the outcomes of these tests were revealed in the taxman’s testing documentation.
In addition to this, HMRC notes that CEST was tested against a further 13 cases that have reached the First Tier Tribunal (FTT) since the initial batch of testing was conducted. Notably, these cases were tested against the judge’s findings of fact from the published decisions, as well as HMRC’s view of the facts at the time of submission.
The documentation acknowledges that only tests against cases that are final are published, ruling out a number of active cases that HMRC has lost and is pursuing via the Upper Tribunal (UT). With test outcomes pending for a further two recent cases – Red, White & Green Ltd and Northern Lights Solutions Ltd – this leaves just five test cases against which HMRC has published CEST’s outcomes.
HMRC also observes that CEST’s outcomes were tested against a number of active and settled compliance interventions, detailing the number of times each status determination was reached.
No credibility to CEST’s three-page testing documentation
“It’s astonishing that, two years after we exposed HMRC’s failure to sufficiently evaluate its own tool, and having been subject to the scrutiny of the Public Accounts Committee (PAC) in the interim, the taxman returns with the same feeble formula,” says Chaplin.
“There is no detail and nothing within the documentation that can even conclusively prove that testing was genuinely undertaken. For a project of this scope, thousands of pages of testing documentation should have been produced. HMRC has provided just three.”
Chaplin adds: “HMRC needs to lift the lid on CEST and reveal the detail behind the supposed testing before the industry can even begin to consider granting the tool any sort of credibility.”
HMRC inspector failures exposed by CEST test
Notably, all five published test cases went in favour of the contractor at FTT, while HMRC’s testing documentation asserts the updated CEST to have reached the same ‘not caught’ decision when answered in accordance with the judge’s findings of fact. However, when answered in line with HMRC’s view of the facts, the taxman acknowledges that CEST returns a ‘caught’ outcome in all five test cases.
Attempting to clarify the discrepancies, HMRC states: ‘Where HMRC has lost a case at the FTT it has been because new evidence came to light or the judge weighed the facts differently. The judges have confirmed that the factors that HMRC take into consideration when determining status are correct.’
However, for Chaplin, the taxman’s terrible recent track record is inexcusable: “Everybody knows the employment status factors to consider. HMRC’s first problem is its interpretation is often contrary to the law. Its second is that many of the factors to be considered when determining status are completely ignored by CEST.
“While it’s impossible to ascertain the accuracy of CEST from this testing documentation, there are few doubts about the incompetence of HMRC’s inspectors. The five test cases published show that the taxman either consistently misinterprets the facts of the matter, fails to evaluate engagements in their entirety, or simply lets bias influence its decisions.
Chaplin adds: “HMRC has already shown it is not averse to challenging a CEST determination where it believes the questionnaire has been answered inaccurately. In such circumstances, the taxman is permitted to impose tax bills, penalties and interest upon the supposedly offending party. How is this justified when its own test data demonstrates that HMRC inspectors don’t know what they are doing?”
CEST assessments less conclusive after updates
HMRC has maintained that CEST returns a status determination in all but 15% of cases, with little regard for the accuracy of the decision. However, a summary of CEST’s enhancements published alongside the testing documentation suggests that the recent updates mean the tool is unable to fulfil even this promise.
Providing an overview of CEST’s usage data covering the period from 25 November 2019 to 31 January 2020, HMRC notes that the tool was used 271,615 times, and was unable to determine the status of the engagement on 56,026 occasions. This works out at a rate of almost 21%.
“We’ve always said that HMRC’s argument concerning the frequency that CEST returns an outcome is worthless because it is often the wrong outcome,” notes Chaplin. “A coin toss provides an answer 100% of the time and is just as accurate as CEST. The fact that CEST cannot reach a decision in more than a fifth of cases is merely the cherry on top.”