Half the British public doesn’t understand income tax – new data

Updated polling evidence from Tax Policy Associates and WeThink shows that half the public doesn’t understand a basic principle of income tax: the way that tax rates apply to income above a threshold. Half of voters believe that, once you hit the higher rate threshold, the 40% higher rate applies to all your earnings.

The full data polling data and our analysis spreadsheet is available here.

In the 1990s, the Liberal Democrats had a policy to “put a penny” on the basic rate of income tax to increase funding for schools. Daniel Finkelstein was then the Conservative Party’s head of research. He and their head of polling, Andrew Cooper, commissioned ICM to look into how popular this policy was. They found the policy was very popular, but that many of those supporting it thought it would cost them one penny. Not 1%, but one shiny copper penny.¹This story has been going round for years, and is often regarded as an urban myth, but Baron Finkelstein kindly confirmed it to me earlier in the week; prompting me to finally update this article.

There’s been research in the US about another tax misunderstanding – that if you go into a higher tax bracket, it means all your income is taxed at a higher rate. You earn one cent more, but pay thousands in tax. Some polls have shown a majority of Americans believe this; others that a third do; YouGov found 52% getting it right.²The difference presumably down to the wording of the poll. The first poll had “You pay your marginal tax rate on all of your income” vs “You pay the same rate as others on income up to a certain amount, then a higher rate on every dollar up to the next threshold”. This seems a little technical and long-winded, particularly given there is no “don’t know” option. The second is much simpler, but refers to “tax brackets”, which rather presupposes their existence is understood. I favour the YouGov approach The misunderstanding is sufficiently common amongst ordinary members of the public that the Tax Foundation has a page dedicated to correcting it.

Which made us wonder whether there is a similar confusion in the UK. When we poll people about different tax rates, do people understand that (for example) the higher rate of tax only applies to the part of your income that falls in the higher rate band?

All things being equal, Americans should understand more about their tax system than we do, given that they are required to file tax in a much more detailed and laborious manner than us.³The amazing reason why the US doesn’t have a UK-style self-assessment system is that Intuit and the tax preparation industry lobbied to prevent the IRS creating a free automated filing system. This sounds like a conspiracy theory but is well established.

How tax bands work

For example: if you live in England⁴or Wales or Northern Ireland, earn £50,269, and receive a £1 pay rise, you’ll pay an additional 28p in tax – 20% income tax and 8% national insurance.

The £50,270 you’re now earning puts you at the top of the basic rate tax band. Get another £1 pay rise and you’re now in the higher rate tax band – and you’ll pay an additional 42p in tax – 40% income tax and 2% national insurance.⁵It used to be slightly more because of the High Income Child Benefit charge, but that’s now moved from £50k to £60k. However that doesn’t change the point of this article – the additional tax is still less than the £1 of additional earnings

The important thing is that the higher 40% rate applies only to your income above the £50,270 40% threshold. Hitting that threshold doesn’t mean all of your income becomes subject to tax at 40%, so that the £1 pay raise results in thousands of pounds of additional tax.⁶In other words, a marginal tax rate of over 100%. There are some points in our income tax system where that actually happens, but not here (with the exception of the savings allowance, of which more later). And stamp duty land tax used to work in this way.

The polling evidence

WeThink kindly included this issue as part of their regular opinion polling back in April. They polled 1,164 people, before weighting (a pretty typical number for this kind of poll). I wrote about it at the time, but the answers were so surprising that WeThink, even more generously, carried out some more polling to obtain an unusually large sample (3,312). That lets us at the subgroups and try to figure out what is going on.

Our question was:

“Suppose that you earn £50,270, the highest amount in the basic rate 20% income tax band. You get a £1 a year pay rise, and are now in the 40% higher rate tax band. How much additional tax do you think you will pay?”

The options we gave were modelled on US polling from YouGov:

• “a small amount of extra tax”, or
• “a substantial amount of extra tax”⁷Note that we forced people to make a choice, and didn’t provide a “not sure” option. The question of whether “forced choice” is the best approach has a long history… I have no expertise in this, and was happy to be guided by the experts at WeThink.

I think it’s clear that the correct answer is “a small amount”.⁸Depending on how you read the question, the additional tax might be 14p, 20p, 40p, or 42p – but I don’t think that matters.. the amount is “small” in each case. The only exception is if you have significant savings, due to the loss of £500 personal savings allowance you’re a higher rate payer – more on that later. But 49% of the public, and 54% of women, don’t agree:

This is consistent with YouGov’s polling of an essentially identical question in the US.

We have large enough subgroups that we can test if there is a statistically significant difference between supporters of different parties – there isn’t:⁹chi-squared test for independence, p-value=0.34, all calculations in spreadsheet linked at the top of this article.

This suggests that whatever is driving the difference we’re seeing, it isn’t ideological.¹⁰Our initial polling had a very high understanding of the “correct” position amongst Green Party supporters – at 66% it was the highest of all subgroups. However the small numbers in that subgroup suggested it wasn’t statistically significant, and the updated polling with larger subgroups has confirmed that, with the high Green figure disappearing.

That is very different from YouGov’s US polling, where there was a very large difference between the parties, perhaps reflecting the very partisan nature of US politics and peoples’ views of the US tax system (i.e. the “substantial” answer from Republications signally a hostility to tax and the US Government rather than an assessment of the arithmetic):

An obvious question is whether understanding of the tax system changes with age.

It does not – again, no statistically significant difference here:¹¹p=0.45

We do, however, see a highly statistically significant difference if we look at income levels:¹²p=0.0020

Scotland

The Scottish rates are different. There is a 21% “intermediate rate” income tax band for earnings up to £43,662, and then a 42% “higher rate” tax band.

It looks on the surface as if there’s better understanding in Scotland; but the Scottish and Welsh samples are both too small to be able to rule out a fluke – this difference is not statistically significant.¹³p=0.24

There’s a similar story if we look at the other subgroups; small differences, but mostly not statistically significant (and the large number of subgroups mean we should be careful not to cherry-pick or “p-hack“).

Did we ask the wrong question?

When we published our initial results, there were two ways in which some people said our question could be misinterpreted.

• The first is that people could think that paying 40% tax on your next £1 of income is a significant amount. So when we thought we were getting an arithmetic answer, we were getting a political one. I’m unconvinced this is a natural reading of the question. But, more importantly, if the answers were driven by politics more than misunderstanding, then we’d see differences between e.g. Labour and Conservative supporters. We don’t.
• The second is that people are accurately understanding the effect of the loss of the savings allowance. This is the £1,000 of interest you can earn tax-free until you become a higher rate payer – it’s then reduced to £500. For someone with around £12,000 of savings outside an ISA that could mean the £1 tax increase costs them £200 (£500 x 40%). For such a person “a substantial amount of tax” might be a correct answer. Intuitively it seems unlikely many people are aware about this effect. But if it was driving the answers, then we’d expect higher awareness of it, and more “a substantial amount” answers, as we go up the income levels. We see the opposite – which is consistent with people understanding the question as expected.

I’m grateful to the people who made these observations; it’s largely thanks to them that we went back to obtain more polling data.

Another possible criticism is that many people didn’t understand the question, and that plus the “forced choice” means that we’re seeing a lot of random noise. We can’t exclude that possibility.

So, if WeThink is kind enough to give us the opportunity, it would be interesting to re-run this polling with a slightly different question. We’d welcome suggestions.

But the consistency with the US polling results suggests that the effect we are seeing is a real one.

Why is this important?

It used to be that only a small number of people paid higher rate tax – that is no longer the case. By 2027/28, the IFS has estimated 14% of adults will be in this tax band, which equates to about a quarter of all individual income taxpayers. It’s reasonable to expect that about half of all households will include someone paying higher rate tax at some point in their lifetime. So the higher rate is more important than it ever was.

Our poll findings shouldn’t cause concern that people are paying the wrong amount of tax. Employees are paid by PAYE, which deducts the correct tax automatically. The self employed either use HMRC’s self assessment system (which calculates the tax due) or use an accountant.

There are, however, other potential consequences of a widespread misunderstanding as to how income tax works.

First, and perhaps most importantly, there are anecdotal reports of people turning away work because they believe entering the higher rate tax band will cost them large amounts of money. If that’s true for even a small percentage of the population, then it’s a problem for the individuals in question and the country as a whole. We’d suggest it’s something that HMRC and policymakers should investigate. (We are unlikely to be able to look into this ourselves with further polling, due to the statistical limits of polling samples.)

Second, it would be sensible to assume that other basic tax concepts are equally misunderstood. Policymakers, media and others communicating about tax (including Tax Policy Associates) should try to bear this in mind.

Finally, it means that we should be careful when looking at opinion polling on tax questions: the responses may be based upon fundamental misunderstandings of the question.

Many thanks to Brian Cooper and Mike Gray at WeThink/Omnisis for their generosity in running polls for us pro bono. They ask for nothing in return, and don’t even ask to be credited.

Many thanks to Daniel Finkelstein for the top quality anecdote.

Photo by William Warby on Unsplash

• 1
  This story has been going round for years, and is often regarded as an urban myth, but Baron Finkelstein kindly confirmed it to me earlier in the week; prompting me to finally update this article.
• 2
  The difference presumably down to the wording of the poll. The first poll had “You pay your marginal tax rate on all of your income” vs “You pay the same rate as others on income up to a certain amount, then a higher rate on every dollar up to the next threshold”. This seems a little technical and long-winded, particularly given there is no “don’t know” option. The second is much simpler, but refers to “tax brackets”, which rather presupposes their existence is understood. I favour the YouGov approach
• 3
  The amazing reason why the US doesn’t have a UK-style self-assessment system is that Intuit and the tax preparation industry lobbied to prevent the IRS creating a free automated filing system. This sounds like a conspiracy theory but is well established.
• 4
  or Wales or Northern Ireland
• 5
  It used to be slightly more because of the High Income Child Benefit charge, but that’s now moved from £50k to £60k. However that doesn’t change the point of this article – the additional tax is still less than the £1 of additional earnings
• 6
  In other words, a marginal tax rate of over 100%. There are some points in our income tax system where that actually happens, but not here (with the exception of the savings allowance, of which more later). And stamp duty land tax used to work in this way.
• 7
  Note that we forced people to make a choice, and didn’t provide a “not sure” option. The question of whether “forced choice” is the best approach has a long history… I have no expertise in this, and was happy to be guided by the experts at WeThink.
• 8
  Depending on how you read the question, the additional tax might be 14p, 20p, 40p, or 42p – but I don’t think that matters.. the amount is “small” in each case. The only exception is if you have significant savings, due to the loss of £500 personal savings allowance you’re a higher rate payer – more on that later.
• 9
  chi-squared test for independence, p-value=0.34, all calculations in spreadsheet linked at the top of this article.
• 10
  Our initial polling had a very high understanding of the “correct” position amongst Green Party supporters – at 66% it was the highest of all subgroups. However the small numbers in that subgroup suggested it wasn’t statistically significant, and the updated polling with larger subgroups has confirmed that, with the high Green figure disappearing.
• 11
  p=0.45
• 12
  p=0.0020
• 13
  p=0.24