Getting down to brass tax
Helen Miller, Director of the Institute for Fiscal Studies, talks to Will Snell about how to build a better tax system
In this in-depth interview, I speak to the new Director of the Institute for Fiscal Studies, Helen Miller, about where the tax system is doing well – including on raising revenues – and where it needs to do better. Top of the wish list is a simpler tax regime that creates fewer economic distortions and treats people more consistently. The challenge is to find a way to design, implement and sell policy reforms that go deeper than the usual tinkering with tax rates to change the tax base (who and what is taxed). Done well, this can increase incentives for investment and wealth creation while taxing economic ‘rents’ and externalities more effectively. Redesign of the tax system needs to be anchored around a small number of key priorities that will enable officials and ministers alike to focus on building the foundations for effective long-term reform.
This is a long post, so if you prefer you can jump straight to a specific section:
Where is the tax system failing?
Balancing fairness in the tax system with other goals
The Budget and taxing wealth
Wealth creation versus wealth extraction
Automation and AI
Evidence gaps after the Deaton Review
Has economics focused too much on efficiency?
Where is the tax system failing?
Will:
If the main functions of taxation are to raise revenue, to redistribute income and wealth, and to correct market failures, on which of these is our current tax system most obviously failing?
Helen:
While I am going to go on to say lots of things about the tax system having problems and needing to be better, it’s important to remember we’re not looking at a car crash. The tax system raises a lot of revenue. It’s due to bring in well over a trillion pounds - close to 40 per cent of GDP. There are lots of countries in the world that would love to have the tax system we’ve got, because it brings in so much money at a relatively low administrative cost.
It also does do a lot of redistribution. Two things are worth pointing out. One is that it really is the tax and benefits system combined – as you should expect. We don’t think the tax system should be doing all of it. Second, there’s no single right answer for how much redistribution you want; people simply differ in their preferences, so some will always think there is too much and others too little.
Our tax system also does an increasing amount of trying to sort out market failures. Pollution and carbon emissions are the classic textbook case where you would want a tax: there is an externality, so you put a price on it and then let the market work. In practice, we are a long way from the textbook solution. Effective tax rates on carbon vary hugely between households and businesses and across different uses, and taxes and regulation interact in ways that government does not seem to have fully under control. There is a lot of scope for improvement here.
The big problem is not that the system completely fails on any of its core functions, but that it does more economic damage than it needs to while trying to achieve its goals. Once you accept that you want to raise revenue and redistribute, you inevitably create some distortions, particularly to work incentives, and that gives you the classic equity–efficiency trade‑off. The question is how to design the system so that, for any given amount of revenue and redistribution, you minimise the costs in terms of discouraged work, saving and investment. On that issue, the UK could do much better.
Balancing fairness in the tax system with other goals
Will:
Do you think fairness within the tax system – treating similar people similarly – is a more important goal than thinking about the impacts of the tax system on growth and on vertical inequalities in society? Or is that a false distinction?
Helen:
No, I don’t think it’s more important, and I do think it’s a bit of a false distinction. Raising revenue and redistributing are the key main reasons you have a tax system – if you didn’t want to do those, you wouldn’t have one. In that sense, they are the fundamental goals.
But you can have horizontal equity and achieve your other outcomes. It’s not either/or. Often, horizontal equity is actually good for the other things. A concrete example: take people within the top 1 per cent. They are taxed very, very differently. I think that’s problematic from a fairness point of view – if you and I are both millionaires, and one of us faces a much higher tax rate than the other, that’s just unfair.
It is also problematic because if I can switch into a lower‑tax form of income, it undermines the state’s ability to raise revenue and to achieve vertical redistribution. I think of horizontal equity as: once you’ve decided how much money you want to raise and how much you want to redistribute, you want to do it in a way that is horizontally equitable – both for fairness reasons and because that supports your other goals.
Will:
And there’s also the point that these things can be mutually supportive – for example, increasing taxpayer confidence in the system and therefore its revenue‑raising potential, which a lack of horizontal equity can undermine.
Helen:
Absolutely. Why is raising tax difficult? Often because people don’t want to pay it. Why are there thousands of studies on the design of taxes? Because when you tax people, they change their behaviour. That is at the heart of it.
Whenever you take two similar people and treat them differently, that opens up lots of opportunities for people to change their behaviour. That is where you get problems. Sometimes there may be a good reason you can’t treat similar people the same, but usually that’s not what’s going on - it’s just poor tax design.
I find it very hard to think of a case where horizontal inequity helps you raise more revenue or achieve vertical redistribution. I can think of tons of examples where it hurts. As a rule of thumb, whenever you see horizontal inequities, you’re seeing a problem in the tax system that needs to be got rid of.
The Budget and taxing wealth
Will:
November’s Budget arguably represented a missed opportunity to address the undertaxation of wealth relative to work. What should the Chancellor be thinking about doing next to tackle this problem? And how can we persuade or incentivise our politicians to grip these big long-term issues in the face of a range of political and systemic incentives to do precisely the opposite? Is institutional reform part of the answer?
Helen:
A great question. First, it’s worth pointing out that the Chancellor did do some things in the Budget to tax wealth. She increased taxes on interest, dividends and property income, and there was a previous capital gains tax rise.
From my point of view – and I’ve been working on this for most of my career, so I feel strongly about it – step one is to make sure everyone understands the problem. For me there are two parts.
Everyone can see the rate differentials – it’s obvious that some things are taxed higher or lower than others. But coming back to horizontal equity, I think what’s missed is this: imagine I’m an employee and you’re a business owner, and we both make £50,000. We might look similar, and it’s tempting to say we should be taxed the same. But if you’ve put lots of your own money into your business, taken lots of risks, invested in capital, then part of your return is a return to capital, not labour. Horizontal equity doesn’t mean just treating us the same. It means treating the labour income the same, but treating the capital income differently. I think there is a lack of understanding about that.
People often think what makes workers and entrepreneurs different is that entrepreneurs need lower tax rates to boost their labour supply. I don’t buy that. What makes them different is capital – their own money at risk. If you just raise the rates, you get a trade‑off. On the one hand, you fix horizontal equity on labour incomes. On the other, you discourage saving and investment. You can choose that trade‑off: accept less saving and investment and less growth in exchange for more horizontally equal treatment of labour. But you don’t have to. You can fix the tax base and sort of have your cake and eat it.
In my mind, what I’d say to the Chancellor is: you have to go for that two‑part solution. First, fix the base so you’re not discouraging saving and investment; second, align the rates on what’s left. I think that is just better – you can have horizontal equity on labour income without reducing saving incentives.
Even politically it is more stable. At the moment, you get some people arguing for higher rates, rates go up, and then later another government comes along saying they’re bad for investment incentives and cuts them again. Up, down, up, down. If you had a system where changing rates didn’t affect investment incentives so much, you could leave them alone more.
So the first step is understanding. And that relates to your second question about incentives. If a Chancellor just “whacks up” tax rates, then groups like the CBI will understandably object, and they’ll often be right to say you’ve damaged investment incentives. You can imagine a system where you fix the base and the investment‑incentive issues first, and then raise rates. They still won’t love it, but they’d probably be much less upset.
All the time we pretend there isn’t a trade‑off and act as if you can just change the rates, we’re in an unstable position. But this is hard politically, because tax rates are salient and easy to understand, whereas the tax base is much more complex. Part of the challenge is finding ways to make the base issue more salient.
Will:
So you’re arguing for improving the quality of the debate, rather than deep structural changes to, say, short‑term political incentives, Treasury reform and so on?
Helen:
Those bigger reforms would be great too. I regularly complain about short‑termism in politics. But it’s worth remembering that sometimes politicians do take long‑term decisions. Rachel Reeves has put forward planning reforms whose benefits will mostly be outside the immediate political window. Jeremy Hunt did full expensing for investment; the biggest effects won’t be immediate. So yes, more long‑termism would be good, but we shouldn’t say it never happens. I think first you have to win the argument: make politicians realise this is worth doing, and get a broader set of stakeholders to think it’s worth doing. Why did planning reform get through? Lots of people thought it was a good idea.
If there were an equivalent level of support for fixing the structure of the tax system – independently of people’s views about how much redistribution they want – that might make it easier to take a longer‑term punt on it. Of course, there are still issues about cost, timing of revenues, and things like how the OBR scores policies, which can create funny incentives. Reform is always easier when you’ve got revenue to give away.
I also think there is a specific political problem in the UK, which I suspect has got worse. We have a tendency to bolt something else on. We have a really, really complex tax system. Politicians are much more excited about saying “I’ve got something new” than “I’ve taken things away.”
I think the biggest problem the tax system has is trying to do too much. Every time I hear someone say, “We need a new incentive,” I think: no, first think about five incentives you can remove. We rarely appreciate how all these things layer together to create odd incentives.
There’s also something about the structure of the budget cycle. If you have lots of fiscal events, you’re standing up in Parliament wanting to announce something to lots of different groups, and that tends towards newness. That drives the proliferation of measures. If this Chancellor really sticks to one fiscal event per year, that might help a bit.
We need to get better at stopping disincentivising things – getting out of the way – rather than “incentivising” things. There are a few things government should positively incentivise through the tax system, but they are few and far between. Mostly, tax should be trying to get out of the way.
Will:
Do you have a view on why significant capital gains tax reform didn’t make it into the Budget?
Helen:
I don’t know the answer. I think there are a whole bunch of things going on. It is worth saying that this stuff is genuinely hard. I don’t mean technically hard in the sense that you couldn’t design something – I reckon I could write down what we should do, and I don’t think it would be that hard to administer. But the transition from here to there is genuinely hard.
Capital gains tax reform is particularly hard because it’s not an issue you can consult on in advance – people will change their behaviour. You can’t just say, “Here’s my five‑year plan.” And if you haven’t got your ducks in a row well before the Budget, it’s too late to work out all the mechanisms and staging of changes. Even something like introducing a new allowance to strip out inflation – we’ve written about this – raises questions. Do you apply it only to future gains? Some past gains? How do you phase it in? How do you deal with assets where you don’t have a base price?
None of that is impossible, but it does need detailed preparation. If you wait until the point where a Chancellor is suddenly interested in it, it’s too late for that fiscal event.
If I were the Chancellor, I’d focus on a handful of things – capital gains would be one – and say, “I want Treasury and HMRC to go away and work up properly how you would do this.” Then, by the time the Budget comes, you can look at it and say, “Yes, we’re going to do that.”
I think there’s also the “budget hamster wheel” problem. People in the Treasury and HMRC get through one Budget and then they’re into the next one or a Spending Review. There’s no time for reflection or planning. That means we have a bias towards things that are easy to change – like rates – compared with changing the base or indexing assets.
I also think we lack a clear vision. You need a guiding North Star – “This is where I want the tax system to get to.” The Chancellor can’t do the nitty‑gritty, she’s far too busy, but she has to set that clear direction so everyone is pushing towards it, and you can judge whether each step moves you closer. I have no sense that there is that clear guiding principle at the moment.
Wealth creation versus wealth extraction
Will:
Is the distinction between wealth creation and wealth extraction in our economy a useful one? If so, where would you draw the line between the two, and what role do you think the tax system has in disincentivising wealth extraction and incentivising wealth creation, as opposed to regulation, investment and other levers?
Helen:
I recognise what you mean by wealth creation and extraction – I intuitively get the difference – but I’m not entirely sure where the dividing line is.
Partly that’s because all wealth has to be created first. There is no extraction before there is creation. I started thinking that maybe it’s more useful to say there is wealth creation, and that can be of different types: productive innovation that creates something new for the world, or something more like digging oil out of the ground. Different types of creation, but in all cases wealth is created somehow.
Then, rather than “extraction”, I’m inclined to think about who is getting the proceeds of that wealth, and how that relates to the people who did the creating. Are they getting it? Is somebody else getting it? How fairly is it shared? Who is capturing it?
A lot of “extraction” ideas are basically about economic rents – people getting more in return than you needed to give them to get them to supply that input.
If you can tax those rents without changing behaviours, you can tax them at 100 per cent. The issues are: can you work out where they are, can you ring‑fence them, and are they mobile?
North Sea oil is the classic example. There are clearly rents. You can draw a circle around it and ring‑fence it. You can’t move it. When you can find those extractive industries, I think taxing them at high rates – with deductions for investment – is a really good thing.
The other issue is whether you can stop rents arising in the first place. That is hard, and that’s where tax is definitely not the only instrument. Competition policy, regulation and other levers are at least as important, if not more.
Some economists argue that higher personal taxes at the top can discourage rent‑seeking behaviour – the idea being that if you’re going to be taxed at a high rate, what’s the point of trying to seek rents in the first place? I see the theory, but we have to be very careful putting it into practice, because those high rates also discourage genuine innovation and activity.
My take is that you first want a government thinking really hard about why we have markets in which rents arise, and trying to stop them arising – recognising that “free markets” don’t exist. Governments are always shaping markets. Then, where rents still arise, you look for ways tax can capture them.
Innovation is where the rents story gets genuinely difficult. Take a company that does lots of investment over a long time, say on a medical innovation, makes losses for years, and then starts to make big profits in ways that looks like a rent, but it might just be that the return is concentrated in a particular period. If we taxed that at 100 per cent, we’d discourage innovation.
Patent policy is a good example: you want people to capture enough of the return to make the risky investment worthwhile, but at some point they are just getting rents, and you want to stop it. We need constantly to evaluate those policies. With superstar firms like Google, it’s been good in some sense that they could capture returns and we all benefited, but it’s also gone too far in some ways, with too many rents.
When it comes to the role of the tax system, the simple solution is to tax marginal returns at labour income rates. For rents you might want to tax them higher, but at least you won’t be undertaxing them if you tax them like labour. Look for ways to extract rents directly where you can, but otherwise tax everything like labour income, because that’s always a safe option. Often you can’t distinguish them, especially because things that look like rents can be returns to investment, like with patents. Maybe in future we’ll do better, but at the moment I just can’t see how we can be more precise than that.
Automation and AI
Will:
How should we be thinking about rewiring the tax system to prepare for the possibility of large-scale job losses due to automation and AI, and is it time to think about some form of a universal basic income (UBI)?
Helen:
There is a lot we don’t know about AI. It might be our generation’s big challenge, which means we have to constantly watch how it is changing things, and constantly ask what’s new. We won’t know all at once what needs to happen – we need to keep monitoring.
But there are things we can already do. We can start taxing rents where we can identify them. We can tax returns to capital properly, so we don’t distort choices over capital inputs, and so that if we do need to redistribute more from capital owners in future, at least we are taxing their incomes properly now. We can tackle some of the existing inequalities – including from undertaxed land values.
We should stop incentivising self‑employment, which is a more precarious form of work. I’m not against self‑employment, but you shouldn’t push people towards it. Employment relationships are valuable to people; we shouldn’t be making them relatively less attractive at a time when they may already be under pressure.
We shouldn’t forget the externalities related to things like data centres. These things are hugely power‑hungry. If we don’t tax the externalities properly, we make them look cheaper than they actually are, relative to human beings. So we should be on top of that.
And, although it’s not my area, if I were the government I’d be saying we need to regulate AI to make sure it works for us as humans. It should be a tool we use, not a substitute for us. We should try to have some control over that.
On UBI, I think we should absolutely be thinking about whether we want a different kind of social support system in future – a different social security net to provide different forms of insurance or support. But in my mind, UBI either ends up being extremely expensive or too low to live on. In some sense, I think UBI is a bit of a red herring. The question is always: how do we redistribute, and from whom to whom? You have to work out who has income and how to get it to other people. That should be the question, rather than whether a UBI per se is the right mechanism.
The other thing worth highlighting is that people don’t just want income. They want good jobs and broader engagement with their fellow citizens. Even if we had a world where AI created lots of rents and we taxed them and redistributed the income, I don’t think that’s a world people would want if they didn’t also have good work and social participation.
Evidence gaps after the Deaton Review
Will:
A couple of years ago, the IFS published the Deaton Review of Inequalities, a very wide-ranging and comprehensive exercise. Where do you think are the biggest remaining gaps in the evidence base on the nature, causes and consequences of various forms of inequality, and on the solutions to them, and how can these gaps best be filled?
Helen:
There are a few things. One is that inequalities are not static – they change over time. So there’s an ongoing need just to measure them well, including different dimensions of inequality such as wealth, and to track them over time and across cohorts.
Another set of gaps is about interactions and causality between inequalities. Health and income is a classic example: poor health can lead to low income, and low income can lead to poor health. Similar stories apply to education, geography and other factors. Understanding those feedback loops is crucial for designing policy.
We also need more work on dynamic, long‑run effects. Something that looks OK or even good in a short‑run, partial‑equilibrium sense can look very different over 20 or 50 years once you account for how people and firms adapt. Immigration is one example where short‑term analysis can miss longer‑term structural change.
Then there’s the policy toolkit. Policymakers face taxes and benefits, regulation, competition policy, education, labour market policy and more. These instruments have to be combined to achieve a range of objectives – efficiency, growth, redistribution, resilience. Framing the trade‑offs in a way that is analytically sound but practically useful is very challenging and something we’re still working on.
In terms of themes, the Deaton Review highlighted areas like regional inequality and how to foster sustainable agglomerations beyond London and the South East; housing and its interaction with wealth inequality; health inequalities; and, increasingly, the distributional implications of technology such as AI. Those are areas where better data, better models and more evaluation of specific policy experiments would be especially valuable.
Has economics focused too much on efficiency?
Will:
Angus Deaton himself recently argued that economics as a profession has been overly focused on efficiency (and individual liberty) at the expense of social justice, and on ‘average’ indicators of economic success rather than questions of distribution. Do you agree with this analysis?
Helen:
There are areas – trade policy is a classic one – where economists historically focused heavily on aggregate gains and didn’t pay enough attention to who gained and who lost. That criticism has real force.
But I think there is a danger of caricaturing economics as only caring about efficiency. In practice, lots of economists, including at the IFS, have spent decades studying distribution. Equity–efficiency trade‑offs are built into the standard models.
Part of why efficiency can seem more prominent is communicative rather than substantive. On efficiency, economists often feel able to say, “This is the efficient outcome.” On distribution, there’s always an element of value judgement – “If you care about this distributional objective, then you should do X.” That makes efficiency results look crisper and more authoritative, even if distribution is being given equal analytical weight.
It’s important not to respond to past neglect by swinging too far in the other direction. Distribution matters enormously, but so does growth. Low‑growth economies are often terrible for equality in the long run, because there is less scope to improve living standards and politics becomes more zero‑sum. Policies that improve distribution today at the cost of significantly lower growth can leave future generations worse off, including those at the bottom.
The task is to take both efficiency and equity seriously, be explicit about the trade‑offs, and improve the tools for handling them – not to pretend that only one side of the trade‑off matters.