VAT and the Sausage Roll
VAT, sausage rolls, and why the number on the menu isn't what you think it is.
This is the first of two parts. Part one covers the mechanics—what VAT actually is, why it hits hospitality differently than almost any other sector, and why the tax system has strong opinions about sausage rolls. Part two, coming later this week, makes the argument: why 20% is the wrong number, what it’s actually costing the Treasury, and why the countries around us have already worked this out.
If you would like part two delivered straight to your inbox, subscribe below. It’s free, and I would love to have you along. If this topic interests you, I would love to hear your thoughts in the comments—and if you are more economically inclined than I am, your perspective would be genuinely welcome. I am making the argument as an operator, not an economist, and I am happy to be challenged.
A schnitzel sits on the menu for £30. It’s printed clearly, without footnotes or addendums. The guest sees £30. What they don’t see—and what almost nobody thinks about—is that £5 of that was never mine.
Before a single penny goes to pay for the working parts—wages, ingredients, utilities, rent—£5 belongs to HMRC. Your food arrives at the table, seasoned generously with public policy.
What the hell is VAT?
Value Added Tax—VAT for short—is charged at 20% on most hospitality food and drink in the UK. Businesses collect it on the government’s behalf, hold it, and remit it quarterly. The mechanics are straightforward. Every pound of gross revenue is one-sixth tax. Not one-fifth, as is commonly touted. A £120 bill is £100 of revenue, and £20 of VAT.
The VAT registration threshold—the point at which a business must register and start collecting—sits at £90,000 of rolling 12-month taxable turnover. At that level, the UK has the joint-highest threshold in the OECD.1
VAT is supposed to be a tax on consumption—neutral, in theory, to the businesses that collect it, because they reclaim the VAT paid on their inputs and only remit the difference. For most businesses, this works reasonably well. A manufacturer buys materials, pays VAT on them, reclaims it, charges VAT on finished goods, remits the net. The tax flows through. The business is a conduit.
Hospitality breaks this model in two specific ways.
The first: its primary input is labour, and labour carries no VAT. There is nothing to reclaim on wages. In a sector where staff costs typically represent 35–40% of revenue, this alone limits the conduit model significantly.
The second: most food purchases are zero-rated. A restaurant buying ingredients from a supplier pays no VAT on them—which sounds like a benefit, but means there is also nothing to reclaim. The input credit that makes VAT theoretically neutral simply doesn’t exist for a hospitality business’s two largest costs. What remains reclaimable—VAT on energy, equipment, packaging, some beverages—represents a fraction of total outgoings.
The result is a business that collects 20% on every pound of food and drink sold, has almost nothing to offset against it, and remits the gap in full. The effective tax burden on value genuinely added is higher in hospitality than in almost any other sector. It is the structural reason a 20% rate lands differently here than elsewhere.
The Cliff Edge
Picture two cafés on the same street.
The first turns over £85,000 a year. At this level, VAT registration isn’t required. The pricing is simple. The cost plus the viable margin goes on the menu. There are no quarterly returns, no software configurations, no hidden arithmetic.
The second turns over £95,000. Now, they need to register. They have two choices; raise prices by 20% and risk losing customers, or absorb the tax and make a lower margin. Neither option is positive. The first feels punitive to guests, who now have to cough up 20% extra for the same coffee they were drinking a week ago, and the second can compress margins to an unsustainable level.
Some businesses do something else entirely. They restrict their turnover. They stay below £90,000 revenue. They reduce their opening hours, they close for a few weeks in the summer, they turn down opportunities. This isn’t simply hypothetical. HMRC’s own research has documented what economists call “bunching”—an abnormal clustering of businesses just below the registration threshold.2 When systems are designed, they incentivise behaviour, particularly when there’s a cliff edge. The businesses sitting at £88,000 of turnover aren’t failing. They’re being rational.
Who Wants To Be A Millionaire?
A restaurant with £1,000,000 in gross annual turnover sounds substantial. It sounds like a business that’s made it.
Remove the VAT. That’s £833,333 in net revenue. Then subtract food cost—typically 28–32% of net revenue for a full-service restaurant. Call it 30%: £250,000 gone. Labour: another 35–40%. Call it £291,667. Then rent, business rates, energy, insurance, merchant fees. The numbers compress quickly. The average net profit margin across UK restaurants was 4.2% in 2024. For full-service independents—the kind most people picture when they think of a restaurant—it typically runs at 3–5%.3
The word “million” changes meaning once you’ve done the arithmetic. What sounds like abundance is, in practice, a business operating on extremely thin clearance. A £1,000,000 restaurant might retain £40,000–£50,000 net. It’s not a complaint—margins reflect the model, and hospitality has always operated this way. But it is worth understanding when you look at the menu price. Much of that number is load-bearing in ways that are invisible from the other side of the table.
A Tax on the High Streets, Peculiar in the Spreadsheets
VAT is not just a percentage. It’s also a collection of definitions. And definitions, in Britain, can get very strange indeed.
Take coffee. A takeaway hot coffee attracts 20% VAT. A takeaway iced coffee is zero-rated. The same espresso, the same machine, a different tax treatment depending on whether the milk is warm. For a business, this is not simply academic: it requires EPOS categorisation, menu structuring, periodic reconciliation.
Have you ever been into a bakery and bought a sausage roll? Have you ever asked them if they could heat it up, and they said no? Perhaps the last time you visited, it was fresh from the oven, and it was oh so much better, and now, these bastards are insisting you eat your sausage roll cold. The reason, my friends, is VAT.
Baked and left to cool: zero-rated. Kept warm in a heated cabinet: 20%. Reheated to order: 20%. The distinction hinges on the intention to serve hot.4 Equipment decisions—cabinet specifications, holding temperatures—are therefore, in part, tax decisions. A bakery choosing not to install a warming unit is making a VAT calculation alongside a commercial one.
And it gets better: take your cold sausage roll out the door, and no VAT applies at all. Sit down at one of their tables to eat it, and you’ve just triggered 20%. Same pastry. Same filling. Same bakery. Different chair.
Then, there’s my favourite example (sad that I have a favourite). The Jaffa Cake. Is it a cake, or is it a biscuit? Who actually cares? You know who really cares…HMRC.
In 1991, United Biscuits (McVitie’s) appeared before a VAT tribunal to contest HMRC’s reclassification of Jaffa Cakes from zero-rated cakes to standard-rated chocolate-covered biscuits. The tribunal found in McVitie’s favour. One of the deciding factors: a Jaffa Cake goes hard when stale, as cakes do, rather than soft, as biscuits do. The structural behaviour of a cooling sponge carried multimillion-pound tax implications. McVitie’s reportedly produced a 30cm giant Jaffa Cake as courtroom evidence.5
These aren’t just footnotes. The temperature of a pastry, the ambition of a warming cabinet, the staleness behaviour of a sponge—each is a margin decision dressed as a product decision. The complexity is real, the administrative burden is real, and the consequences of misclassification are real.
The Experiment
Between July 2020 and April 2022, the government ran what amounts to the closest thing we have to a live trial. VAT on hospitality was cut first to 5%, then stepped to 12.5%, before returning to 20%. The stated aim was survival: keep the sector breathing through the pandemic and its aftermath.
The total cost to the Exchequer was over £8 billion.6 What did that buy?
Some prices fell. ONS data recorded a 5.7% month-on-month drop in catering prices between July and August 2020, though this coincided with the Eat Out to Help Out scheme, making the VAT cut’s individual contribution difficult to separate. Academic research found pass-through of around 20–50%, peaking briefly then fading within months.7 The landmark Benzarti and Carloni study of France’s 2009 restaurant VAT cut found that only around 18% of the saving reached consumers in lower prices. The IFS has cited this in arguing that reduced VAT rates are an inefficient policy tool.
It’s a reasonable observation. It’s also, as we will see, only half the picture.
As for closures—formal insolvency statistics were mechanically suppressed by government restrictions on winding-up petitions until March 2022, making it impossible to cleanly isolate the VAT cut’s effect. What we know is that 9,930 licensed premises closed permanently in 2020 alone, and that once legal protections ended, insolvencies surged sharply. It’s ambiguous, at best, but it doesn’t necessarily resolve in the Treasury’s favour.
The Comparison
Let’s look at what Britain’s neighbours have decided.
France has charged 10% VAT on restaurant food since 2014, having first cut to 5.5% in 2009 as an explicit jobs policy. Spain and Italy both apply 10% to restaurant services. Ireland—which cut to 9% in 2011, removed it in 2019, restored it temporarily during Covid—passed legislation in late 2025 reinstating a permanent 9% rate from July 2026. Germany, which let its temporary 7% pandemic rate expire in January 2024 and reverted to 19%, has since legislated a permanent 7% rate on restaurant food effective January 2026.8
Denmark charges 25% with no reduction, and has a thriving restaurant culture. It also has average wages and a cost structure that make the comparison fairly academic.
The direction of travel is clear. Most of Europe levies hospitality VAT at 7–13%. Britain, at 20% flat, will shortly find itself standing almost alone. Different countries have made different decisions about how to treat hospitality within their tax systems. Britain has made one too. The question worth asking is whether ours is still the right one.
The 20% Sitting at Every Table — Part Two
In part one, I covered the mechanics — what VAT actually is, why it hits hospitality differently than almost any other sector, and why the tax system has strong opinions about sausage rolls. This part makes the case for changing it.
OECD Tax Policy Analysis, 2024. UK threshold joint-highest with Switzerland; more than double the EU average.
HMRC Research Report 761. 2022 documents statistically significant bunching below the registration threshold, most pronounced in hospitality and construction.
UKHospitality research, 2024. Full-service independent restaurants typically 3–5%; sector average 4.2% (UKHospitality / Opsyte Benchmarking Report 2024).
HMRC VAT Notice 701/14, section 4.4: food held above ambient air temperature at point of supply is standard-rated.
United Biscuits (LON/91/0160), VTD 6344, 1991. HMRC manual VFOOD6260 confirms the “goes hard when stale” test among deciding factors.
HM Treasury confirmed total cost exceeding £8 billion. OBR Economic and Fiscal Outlook, March 2021.
Benzarti, Y. and Carloni, D. (2019): "Who Really Benefits from Consumption Tax Cuts? Evidence from a Large VAT Reform in France," American Economic Journal: Economic Policy, 11(1): 38–63. UK pandemic pass-through: Piga, Onnis, Conti and Bottasso (2022), estimated 20–50%.
Ireland: Finance Act 2025, permanent 9% rate for food and catering from 1 July 2026. Germany: Jahressteuergesetz 2024, permanent 7% rate on restaurant food from 1 January 2026.
So if the money was never yours in the first place, is the real issue here that you can't charge the correct amount to make a decent profit because of perceived value?
This is one of the clearest operator explanations of VAT I’ve read.
The key point is that hospitality isn’t a neutral conduit. When your two biggest costs are labour and largely zero-rated food, there’s very little to offset. Twenty percent isn’t just a consumption tax in this sector. It compresses margin directly.
The £1m example makes it real. Strip out VAT, then 30% food, then 35–40% labour, then fixed costs. What sounds like scale quickly becomes fragility.
And the threshold cliff is telling. When rational operators cap turnover to avoid registration, that isn’t poor ambition. It’s the system shaping behaviour.
Whether 20% is right or wrong is political. But structurally, it clearly lands harder here than in sectors with reclaimable inputs and fatter margins.
Looking forward to part two !!!