Uber’s latest financial reports, filed yesterday, reveal that HMRC is challenging its VAT position and claiming £386m. Uber is appealing but has had to pay the disputed VAT up-front.
Uber historically charged no VAT. It said it was just an app, with its drivers supplying the service to customers. As most drivers‘ income fell below the VAT threshold of £85,000, that meant Uber rides were free from VAT. That was challenged by HMRC in 2022, with Uber paying £615m. Many people expected Uber to then charge 20% VAT on its fares, but it instead used the Tour Operators Margin Scheme to pay a low effective rate of VAT HMRC is now challenging that.
UPDATE: Uber just got in touch to confirm that I’m correct that the new dispute is about TOMS; but I was wrong to think the 2022 settlement was on the basis of TOMS – it wasn’t.
The new disclosure
It’s in Uber’s Q2 2023 results:
Most businesses charge VAT at 20% on their goods and services. But the UK has an unusually high VAT threshold of £85,000 and, as most taxi drivers earn less than that, most taxis don’t charge VAT.
That was historically the case for black cabs, and for most private cabs. They might be coordinated by booking service, but realistically the drivers are independent businesses and the booking service is just their agent. Each driver pays a fee (say 10%) to the booking service, and that would be subject to VAT, but the fee the driver charged the customer is the fee for his services, and so no VAT applies.
Uber’s position was more interesting:
- Uber said it was the same as a normal taxi booking service, and was just an agent for the drivers – so the fares were not subject to VAT1 in principle, they should’ve been if the driver earned over £85,000; I’m not sure how this was dealt with in practice, and it may never have happened.
- But in tax terms, Uber was pushing their luck. Uber controls many aspects of their drivers’ business; it’s more than a mere agent, and so it always seemed more plausible that Uber was the “principal”.2The original version of this paragraph said that Uber was an outlier compared with other taxi firms; StuartW in the comments below gives good reasons to doubt that was correct. That meant they should charge VAT.
- Uber’s position became hard to defend after, in a March 2022 employment law judgment, the High Court confirmed that Uber was indeed the principal, following a Supreme Court decision that Uber drivers were “workers” and not self-employed.
- HMRC eventually agreed – they asserted VAT was due and Uber caved and paid up £615m in an agreed settlement with HMRC.3The first draft of its article This followed legal action by Jolyon Maugham KC4HMRC say they were always going to challenge Uber anyway and the timing was a coincidence
Uber saw this as putting it at a competitive disadvantage against other taxi firms, and so it looks like it has started to litigate to ensure VAT is also applied to its competitors.5It did seem to me that Uber is different, but the High Court in the Sefton case didn’t agree; see also the comment from StuartW below.
The new Uber model
A normal person, or even a normal tax advisor, might think that Uber having to charge VAT meant a 20% increase on all fares. Uber would be able to recover VAT on its expenses (“input VAT”), but this won’t be much – Uber’s main expense is paying drivers, and they mostly/all earn too little to charge VAT.
But there is something called the “Tour Operators Margin Scheme” (TOMS).
Say I am a travel agent. I put together a package holiday involving a whole bundle of different services: hotel, flights, coach services, restaurants, train tickets, etc. Potentially across multiple different countries. In theory, I should be charging my client 20% VAT on the package holiday, and recovering input VAT on those costs are incurred that were subject to VAT. But all the different countries mean my VAT position would be an unholy mess. TOMS says: yeah, it’s all too difficult. Let’s not bother with the usual VAT accounting. Instead, I’ll just account for 20% VAT on my profit margin. That’s TOMS.
Uber takes the position that TOMS applies to it. It is buying the drivers’ services, and no doubt other ancillary services as well, and then supplying a bundle to the customer in the form of their ride. You might well say; hang on, Uber is not a tour operator. However, TOMS does more than it says on the tin, and even the HMRC guidance goes out of its way to say that it can apply in cases where outside a classic tour operator.6The technical background is well explained here, but it’s subscription only
The consequence is that it is only Uber’s profit from each ride which is subject to 20% VAT. That is probably why Uber paid only £615m, and why Uber’s fares did not noticeably increase after it started charging VAT.7Uber has been stupendously loss-making for most of its existence, but from a VAT perspective that doesn’t matter, because VAT looks only at the income and costs attributable to a particular ride. Other elements responsible for Uber’s losses (such as all those coders and their expensive coffee, servers, marketing, cost of capital, etc) won’t be relevant here
The other consequence is that a business hiring an Uber taxi cannot recover VAT on the fare. That stands to reason because VAT is not the normal 20%, but some smaller amount which the business cannot know.
The new £386m claim
The disclosure doesn’t say, but my understanding from a well-informed source is that HMRC is claiming that Uber can’t use TOMS, so that Uber has to pay an additional £386m. This is very plausibly the difference between TOMS and standard VAT on one year of Uber’s revenue.8To ballpark this: Uber’s fee income is c£800m and its gross income is c£2.6bn. So the difference between TOMS and normal VAT for one year will be around 20% x (£2.6bn – £800m) = £360m.
The conditions are pretty simple:
- Uber must be a “tour operator”
- it must buy in supplies from another person (the drivers)
- those supplies must be “resupplied without material alteration or further processing”
- and Uber must supply them from an establishment in the UK, for the direct benefit of a traveller
All of these are easy, except the third. The most likely HMRC challenge is that Uber is not at all like a travel agent, because it shapes every aspect of the taxi driver’s services, so that there is “further processing” or “alteration” of those services. Importantly, the driver doesn’t control his or her pricing, and may not even know how much Uber charges.
HMRC recently made a challenge on a similar basis to a business that leased apartments and then let them to tourists. The “alteration” here was that the taxpayer painted the apartments and provided furniture. The taxpayer won; these alterations were found to be “superficial and cosmetic”. The case is Sonder Europe, and the judgment is here.
There’s another more ambitious and fundamental way in which HMRC could challenge Uber’s use of TOMS. It could argue that the drivers are in fact employees of Uber. In that case, there are no supplies bought-in, and TOMS cannot apply. The whole taxi fare becomes subject to VAT.
If I was HMRC I would run both arguments.
Is this tax avoidance?
There is no single legal definition of tax avoidance. I’ve written more about that here. But, for what it’s worth, I think Uber’s original approach was tax avoidance, because they were artificially taking a position that VAT was not due, when on the face of it, it was. The fact Uber backed down without a fight adds strength to this.
However, the question of whether TOMS applies is a dry technical question which does not relate to any particular structuring or act of Uber. Uber’s approach seems legitimate to me, whether or not it ultimately turns out to be correct. If we don’t like TOMS being used in this way, we can change the law.
Who will win?
That is hard to say. This is a difficult area with little relevant authority, and neither I nor our regular team of experts have sufficient experience to be able to call it. Both Uber and HMRC may think they have good arguments.
Given that, and the amounts involved, we can probably expect it to be appealed at least once, and possibly several times. It’s not unusual for this kind of dispute to take ten years. But, with VAT, you “pay to play”. Uber had to pay the £386m upfront. As time goes on, without the dispute being resolved, Uber will have to pay additional amounts upfront, reflecting the position that TOMS does not apply. Only if Uber eventually wins does it get its VAT back.9 Corporation tax and income tax don’t work this way at all. You almost always get to keep the tax while you dispute it, and only have to pay it to HMRC (plus interest) if you lose your final appeal. There is, needless to say, no rational basis for the distinction.
An Uber spokesperson told me “Uber is seeking clarity for the whole industry in order to protect drivers and passengers.”
Update 3 August: I belatedly remembered the way US accounting works for indirect tax disputes. We can deduce from the way the Uber disclosure is phrased that Uber has been advised that it is “more likely than not” to win the dispute, and recover the VAT back. It has therefore booked a receivable in its accounts to reflect that eventual recovery – that’s why the statement says there is a cash impact, but no impact on the income statement (because the £386m debit for the VAT is cancelled by a £386m credit for the receivable). Uber has confirmed to me that this is indeed the position.
In my experience, US corporations take these issues very seriously, and require clear advice before their accounting personnel and auditors permit such a receivable to be booked. Of course that doesn’t necessarily mean Uber will prevail – HMRC may also be confident of its position. But it does mean this is much more than a try-on. Asking around, there’s no clear view amongst VAT experts on who will win here.
What’s the consequence for the taxi industry, and our fares?
The outcome of the Sefton case means that the entire private taxi industry (i.e. not black cabs) will be affected by the result.
If Uber win, then we will continue with most of the taxi fare being outside VAT, and only the taxi firms profit margin being subject to VAT at 20%.
if Uber lose, then, unless I am missing something, the entire fair charged by taxi firms will be subject to 20% VAT. But, if you book a private car directly with the driver, there won’t be VAT. That seems a very distortive result, that would drive economic inefficiency. We all benefit from being able to book taxis in a centralised way, i.e. over the phone or on apps, and it seems crazy for tax to push in the other direction.
What about black cabs?
individual black cab riders usually earn less than £85,000, and so don’t charge VAT.
The centralised black cab booking services appear to use the same structure as Uber used before 2022, with the booking service, saying it is merely the agent for the driver. The judgements that stopped Uber and other private taxi services from operating in this way don’t apply to black cabs.
So, when I booked a taxi with FreeNow yesterday, I received two receipts.
First, the platform’s fee, with VAT.
Second, the driver’s fee, with no VAT:
The paragraph at the bottom gives away that the agency structure is being used.
Should the Government change the law?
This all seems very odd. There’s a plausible case that the TOMS rules should be amended, so that TOMS only applies in cross-border cases where there is a real need for it to apply (and, post-Brexit, the UK can easily make changes of this kind to our VAT system). That wouldn’t change Uber’s historic position, but would ensure full VAT applied going forward. Changing the law would be very easy; it looks like HMRC can just designate taxi firms as outside TOMS. There’s even an argument for abolishing TOMS – in the modern world it looks less like a simplification, and more like a hand-out.
But that would just formalise the distortive economic result that booking a taxi directly results in no VAT, and booking through an app means 20% VAT.
But the real problem here is the high VAT threshold, which (if Uber lose) will be responsible for a 20% cost difference between booking a taxi directly and booking one through an app/phoneline. That is, in my view, unjustifiable – it’s economically distortive and drives tax, evasion, tax avoidance, and uncertainty/tax disputes. The threshold should come down (enabling a VAT rate cut at the same time).
So I have to say that, from a policy standpoint, I hope Uber wins (so that there’s mostly no VAT on any taxi fare), or that the threshold is cut (so there’s VAT on all taxi fares).
Let’s just hope we don’t end up with other fudges, like a 0% rate for taxis, which would further erode the tax base and create opportunities for avoidance (e.g. by combining other VATable services with a taxi ride)
Thanks to B for the tip-off and expert input on this.
- 1in principle, they should’ve been if the driver earned over £85,000; I’m not sure how this was dealt with in practice, and it may never have happened.
- 2The original version of this paragraph said that Uber was an outlier compared with other taxi firms; StuartW in the comments below gives good reasons to doubt that was correct.
- 3The first draft of its article
- 4HMRC say they were always going to challenge Uber anyway and the timing was a coincidence
- 5It did seem to me that Uber is different, but the High Court in the Sefton case didn’t agree; see also the comment from StuartW below.
- 6The technical background is well explained here, but it’s subscription only
- 7Uber has been stupendously loss-making for most of its existence, but from a VAT perspective that doesn’t matter, because VAT looks only at the income and costs attributable to a particular ride. Other elements responsible for Uber’s losses (such as all those coders and their expensive coffee, servers, marketing, cost of capital, etc) won’t be relevant here
- 9Corporation tax and income tax don’t work this way at all. You almost always get to keep the tax while you dispute it, and only have to pay it to HMRC (plus interest) if you lose your final appeal. There is, needless to say, no rational basis for the distinction.