To almost every economist, a simple tax system is a good tax system. But here in the UK we have a complex, cost-creating nightmare of a tax system. VAT (value added tax) is one of the worst offenders. 

VAT is a good idea in principle, and is certainly far superior to a US-style sales tax. But in practice our VAT system is absurd, with almost all the absurdities coming from some products being eligible and others not, often with unclear definitions. As well as being absurd, these tax distortions distort the real economy, and make our country poorer and less productive. 

Simplifying VAT could be a way to end this chaos, all while raising billions for the Treasury at a time when public finances are under enormous pressure and living standards for the poorest households continue to decline. Untangling the mess that is VAT could be an encouraging step towards a smarter, better tax system.

Let’s start with a question famous in both tax and legal circles: ‘Is a Jaffa Cake a cake?’. Legally they were cakes, then biscuits, and are now cakes again. This matters because biscuits have VAT and cakes do not. For that reason McVities, who make Jaffa Cakes, were willing to spend a lot of money to prove that these snacks are indeed a type of cake. In the end they won, but ultimately the entire case was a waste of time and money, caused by almost unworkable pieces of VAT law. 

This is not a unique case. Back in 2009 the courts decided that Pringles are crisps not cakes. Pringles, like McVities, wanted their product defined as a cake because this would excuse them from VAT. They argued that Pringles are not slices of potato that are fried, but rather a product made of dough, only part of which is made from potato. In the end they lost, but it’s hard to say anyone won. 

My favorite absurdity - although, by that I mean it is the one I dislike most - is the difference between a muesli bar and a flapjack. The former is liable for VAT, the latter is not. The legal definition of a flapjack is whether the item is “of a standard to be served to guests as a treat with afternoon tea”. You literally couldn’t make it up, and I didn’t. Basically, what the law is saying is that if your muesli bar/flapjack is a bit posh, it is not liable for VAT. But if it looks mass-produced, then VAT must be applied.

These examples are part of a wider issue: food is a VAT minefield. A gingerbread man is exempt from VAT unless it has a chocolate smile. But don’t imagine the tax authorities hate chocolate: chocolate milkshake powder is VAT exempt, but strawberry and vanilla milkshake powder are not. Marshmallows have VAT, unless they are large enough that the manufacturer can claim that they are designed for toasting over an open fire, because then they are a cooking ingredient, not a snack, and so do not attract VAT. Confused yet?

If we were starting from scratch, we would not keep this current approach to taxing food. The two sensible options are to put VAT on all food, or exempt all food. But unless we are going to excuse dining-out from VAT, exempting food creates complexity about different rates for hot and cold foods - as anyone who remembers the George Osborne ‘pasty tax’ fiasco will testify. 

So, option one it is then. Extending VAT to all food would raise £30.8 billion. That is a genuinely large sum - roughly on par with the annual transport budget. It might even cover the black hole that the government needs to fill this autumn. Raising this figure would also be sufficient to cut VAT on everything from 20% to 17%.

According to HMRC data for the 2024-25 tax year, VAT is currently the third largest source of income for the government. Applying a universal levy to food would bump VAT up to second place, leapfrogging the revenue raised through National Insurance contributions, all else being equal.

Applying VAT to all food would be regressive, but not as regressive as people might think. Lower-income households do spend a higher proportion of their income on food, but the difference is relatively small. On average, the poorest two deciles spend less than 1% extra proportionally compared to those on average incomes.

Put differently, if there was currently no VAT on any type of food, and we suddenly imposed it on all food, then the poorest tenth of the population would pay 35p more per week in VAT because they spend a higher proportion of their income on food. The average household spends 12.6% on food, compared with 13.2% for the poorest decile. That 0.6 percentage point differential amounts to £1.74 per week, so the disproportionate VAT burden would be 35p if VAT was levied on all foods. That estimate is of course an overstatement of the bad effects of the proposed change, because some food already has VAT on it. It really wouldn’t be difficult or expensive to ensure that everyone in the bottom tenth had their benefits raised by 35p a week - that is a rounding error for HM Treasury to find.

An important caveat here is that, as Nesta has shown, the amount spent on food is not a good predictor of the quality of food consumed.

No data are available about spending on food with and without VAT by different income groups. Even the detailed data are not very useful for this purpose as, for example, spending on cakes and biscuits are reported together. This shows, incidentally, the arbitrariness of VAT distinctions between food products. 

We cannot calculate the exact consequences of extending VAT to all food on the poorest. We know that it is less than 35p a week, and common-sense suggests that the answer is probably between 20p and 30p per household per week. But with £30.8 billion of additional revenue to play with, the government could compensate the relevant groups (via small rises in pension and Universal Credit, for example).

Food is not the only VAT nightmare. The government has a 12-part flow chart for the VAT treatment of clothes made from animal skins. Gloves are always exempt, no matter the fur or skin used. So is anything made from rabbit or sheep skin. But tanned cow hide always has VAT (unless, of course, it’s made into gloves). Whether products made from dog skins have VAT or not depends on how the skin is treated, while goat skin is always exempt, unless it comes from Mongolia, Yemen or Tibet. If it does come from any one of those three countries VAT must be charged - unless the skin is made into gloves, or headgear, or belts, in which case VAT is not imposed. 

The mind boggles. Perhaps we once had a conference to decide these things. Perhaps these choices were the result of haggling in trade negotiations. Perhaps the glove-makers association was especially powerful and gained a complete exemption. I don’t know, I doubt anyone knows. 

The big VAT exemption in clothing is not gloves or dog skins treated in particular ways, but rather that all clothes and shoes designed for children are exempt from VAT. 

The definition of children’s clothes has many, many odd aspects. A child - for children’s clothing VAT purposes at least - is someone who is up to the deemed average size of a child aged 13 years and 51 weeks. This means that clothes designed for children aged 14 and 15 (about 12% of the country’s kids) attract VAT.

The size-based policy causes further issues if we compare boys and girls. Boys' clothes are exempt from VAT if they are designed for someone up to 163 cm tall. That is indeed the height of the median boy on their 14th birthday, rounded up to the nearest centimetre. But girls’ clothes are exempt from VAT if they are designed for someone whose height does not exceed 161cm. The median 14 year old girl is only 160cm tall, again rounded up to the nearest centimetre. This means the average girl escapes paying VAT for over four months after her 14th birthday. This centimetre is surely a gender-based judicial review case in waiting.

The VAT criteria for shoes also favour women. Men’s shoes are exempt from VAT only up to size 6.5. That means that only men whose shoe size is a long way below the average (size 10) can get away without paying VAT. But women do much better - women’s shows are VAT free up to size 5.5, which is only half a size smaller than the average (size 6). Virtually no man, but almost half of women, can buy shoes that are VAT exempt. 

Of course, this being VAT the system is more complex. The previous paragraph is correct, but not complete, since the rules for women’s shoes include different VAT exemption requirements for high heels. These attract VAT if they are size 3 or greater. This means that the government has to define high heels. If you are interested, high heels are defined as a “heel height does not exceed the sole depth by more than 4cm”. Get your rulers out, bargain hunters.

It doesn’t end there. Shoes that are adapted for people who have a physical or mental impairment which has a long-term and substantial adverse effect upon their ability to carry out everyday activities are also exempt from VAT, but only if the person buying them signs a form or letter to the retailer saying that they meet this criteria. They do not have to provide any proof. If a care home or similar buys such shoes for the use of residents they will have to pay VAT, because they will be used by more than one person. But if you buy a pair for a friend or relative they are VAT exempt. 

No part of a wardrobe or cloakroom is spared from the absurdity of VAT. Following a safety campaign, all bicycle helmets are currently exempt from VAT. But the rules are different for horse riders. Their hats are exempt up to a particular size. The rules are different again for helmets for skate boarding and ice hockey and so on, where the only helmets that are exempt from VAT are those that are both small and marketed exclusively to children. 

So, adults can always protect their head without paying VAT if they ride a bike, can protect their head if they ride a horse only if they have a small head, and can protect their head if they are a skateboarder or ice hockey player only if they have a small head and are willing to wear a product demonstrably designed for a child. If the skateboarding helmet has dinosaur stickers on it, it is more likely to qualify as VAT exempt.

Putting VAT on children’s clothes could be a sensible option. It’s hard to tell if this is regressive or not. This is partly because the spending data are reported for all families, with and without children. As a proportion, people in the middle of the income distribution spend the most on children’s clothes. 

Children’s clothing is certainly available at all price points, and serves as a ‘status good’ to at least some people. For example, Asda sells official Disney babygrows for £1. A more affluent parent can instead pay £16 in JoJo Maman Bebe, or £79 for Ralph Lauren. The super rich can then pay over £500 for Gucci, Hermes, and Dolce et Gabbana. And if that’s not enough, Dior has a baby couture department - prices, I guess, on application. 

Applying VAT to babygrows would increase prices by 20p in Asda, £3.20 in JoJo Maman Bebe and by more than £100 for designer brands. It’s unclear whether putting VAT on all children’s clothes would be regressive. But if it was, then the child element of Universal Credit would be a straightforward way to compensate people.

Taxing food and children’s clothes is absolutely normal across the world. Within Western Europe, Austria, Belgium, Finland, France, Germany, Italy, the Netherlands, Norway, Portugal, Spain and Sweden all have VAT on both (albeit with food at a reduced rate). To the best of my knowledge, Ireland is the only other Western European country without VAT on food. Beyond Europe, Australia, Japan, New Zealand and Singapore tax both food and children’s clothing. 

If we were starting from scratch, the most sensible approach would be to have a universal or near universal VAT system like most countries. If we levied VAT on everything, at the current standard 20% rate, we would raise an additional £180 billion. That would be enough to cut the VAT rate from 20% to around 10%. As an alternative, we could literally abolish the basic rate of income tax, so that no-one would pay income tax unless they earned at least £50,000 a year. That might be appealing to many.

The country that comes closest to universal VAT is New Zealand, where VAT (called general sales tax, or GST) applies to almost everything. The principal exceptions are financial transactions such as mortgages, and renting or buying a flat or house.

We can use New Zealand’s GST system as the basis for a hypothetical ‘reformed’ framework for the UK. This allows us to assess the distributional effects of moving from our current regime (in which VAT is levied on some goods at 20%, some goods at 5%, and not at all on other goods) to one where a blanket 10% rate is applied to all goods. This would leave the government’s overall tax take the same.

Comparing the two systems shows that VAT is more regressive in New Zealand than in the UK, but the differences are minor if we look at the middle of the distribution. 

Even at the extremes of the distribution the differences are not large. The poorest decile would be about £3.60 a week worse off. But given the rest of the graph, this figure is something of an outlier and it is possible that it is caused by sampling issues in either the UK or New Zealand. The sample sizes are around 5,300 in the UK and 3,500 in New Zealand, and we know that the poorest households are often hardest to survey. Nevertheless, let us take this number at face value. 

At first sight it may seem unbelievable that we can levy VAT on food, children’s clothes, toothpaste and so on, and make poor people only a fraction poorer. The reason is the money raised would be used to lower the rate of VAT from 20% to around 10%. That means that all sorts of products that are bought by people across the income distribution - like loo paper - would get cheaper. To a large extent the changes in the system are swings and roundabouts. 

But not quite: the poorest decile would be £3.60 worse off. The obvious approach to prevent this group of households becoming poorer would be to raise benefits by a relevant amount (£187.20 per year based on the calculation above). This might be funded via the additional revenue generated by simplifying VAT, which would be simpler to operate, and therefore save some money. The government should also consider other taxes or charges that disproportionately affect the poorest. 

One such example is the social costs currently levied on electricity. Nesta estimates that these levies amount to £5.2 billion. The poorest 10% of UK households spend about three times as much of their income on electricity as the richest. Moving these costs onto general taxation - funded by almost any tax - would be progressive. If the saving was used to reduce the standing charge, this alone would cover almost half the losses (about £1.70 per week) for the poorest decile, reducing the need for higher benefits to cover the remaining costs of reforming VAT.

Clearly the figures presented here are indicative rather than definitive. HM Treasury, the Office for Budget Responsibility, or indeed the Institute for Fiscal Studies (IFS) could construct more precise estimates. Nevertheless, they are accurate enough to serve as proof of concept: they show that the distributional effects of moving to a system of VAT on everything - levied at roughly half the current rate - are small and manageable. 

The new director of the IFS, Helen Miller, rightly describes our VAT system as “mad”. We should never have created it like this. We could change it and we should. Not only would this bring an end to the madness, but it could also raise billions in additional tax revenue. With the state of the UK public finances looking increasingly precarious, Rachel Reeves should consider bold ideas like untangling VAT. At the same time, any changes to the UK’s fiscal landscape must be carefully counterbalanced with measures to protect the most vulnerable. Better, fairer systems are possible. It’s time to take a serious look at them.

Charts and editing by Nesta’s data journalist Charlie Meyrick