The distributional implications of universality (again)

Paul Johnson, the director of the IFS, has an article in the Times (also available on the IFS website without a paywall) about how different ideas about fairness are behind some of our political disagreements. There’s plenty to agree with in the piece. But when he touches on a couple of themes that regular readers of the Policy Sketchbook will be somewhere between familiar with and bored of – social care reform and redistribution – Paul says some things that I think are a bit misleading.

Here are the offending paragraphs:

“The Conservatives got into dreadful trouble over their manifesto proposals on social care funding. We have spent decades making no progress on how to reform the funding system, partly because of the way we think about fairness. Some think it unfair that anyone should have to use their own assets, including their house, to pay for care. Yet one of the reasons why proposals to cap the amount that anyone has to pay have not been implemented is because, compared with the system we have today, the winners would be the relatively well-off.

In fact, this is a fundamental disagreement about the role of the state as much as it is about fairness. If you think the state is there to provide a degree of social insurance, stepping in where private insurance markets don’t work to pay for those who are unlucky enough to need care, then you are likely to favour it paying all the costs above a certain level. That’s how we tend to think of the NHS. But if you think the state is there just to redistribute money from rich to poor then you might think it unfair.”

This is a familiar take: moving to universal benefits means less redistribution and benefits the well-off. But is it true? Well, that depends. The most important thing to remember when thinking about the distributional impact of changes in government spending is that the money doesn’t just appear out of thin air. The effect of a policy on the level of redistribution that government does depends on how it is paid for. By choosing different funding sources you can get pretty much any distributional effect you want, but some scenarios are more relevant and plausible than others.

Three different ways of paying for a universal benefit

So let’s take a look at the distributional consequences of moving from a means-tested benefit to a universal one, paid for in different ways. We’ll use a stylised example for clarity, but the conclusions generalise pretty well. Imagine we have a benefit worth £1000 per person, but means-tested so that only the bottom 30% of the income distribution get it. An independent commission recommends making this universal, so that everyone gets the benefit regardless of their income. There are broadly three ways we can pay for this and each has different distributional consequences.

Reducing means-tested benefits to pay for universal benefits is regressive

One way to pay for the new universal benefit would be to reprioritise some money that is currently spent on means-tested benefits. It’s trivial to see that this is going to be regressive, but let’s run the numbers anyway. Let’s say we take the money that is currently spent on our stylised means-tested benefit and use it to fund a universal version. Instead of the bottom 30% of the income distribution getting a benefit worth £1000, everyone now gets one worth £300. The bottom 30% lose £700 each and everyone else gains £300. This is nailed-on regressive and it would be a similar story if the money were reprioritised from some other area of means-tested spending.

Means-testing one universal benefit to pay for another is distributionally neutral

Another way of paying for the new universal benefit is to reduce universality in another area. This is the sort of thing that was discussed in the wake of the Dilnot Report: universal coverage for social care could be funded by means-testing some of the universal benefits that older people currently get, like winter fuel allowance or free bus passes. The distributional consequences of the switch would depend on the details, but it would be roughly neutral – we are taking money away from the same people who will get more from the new benefit.

Paying for a universal benefit through higher taxes hits the rich

The third way that we could pay for the new universal benefit is through higher taxes. We can design taxes with various distributional profiles, but we can get an idea of who would be hit by a “typical” tax rise if we look at the distribution of current UK taxes as a whole.

The chart below (based on ONS data) shows the distributional impact of extending our £1000 benefit to the whole population and funding it through an increase in “general taxation” – by which we mean a tax rise with the same distributional profile as the current UK tax system. The (positive) green bars show the additional benefits people in each income decile get and the (negative) red bars show the additional taxes they pay. The red and green bars sum to zero because, at risk of labouring the point, the money has not appeared out of thin air.

The winners from this shift to universal benefits are not the poorest, who already got the benefit, but it would be misleading to say they are the “relatively well-off”. The winners are the people in the middle, especially those just above the current means-test threshold. The biggest losers by far (in cash terms) are the richest 10%. This is a key point: the main distributional effect of a tax-funded change from means-tested to universal benefits is to move money from the people at the top of the income distribution to the people in the middle.

But you may have noticed something else in the chart. The bottom 30% of the income distribution also lose out, because they pay more tax and don’t get any additional benefit. They don’t pay much in absolute terms, but relative to their income it’s quite a lot. The chart below shows the same figures as a proportion of income: the richest still pay the most on this measure, but now the poorest are not far behind.

Now, unless you think that the bottom 30% do rather too well out of government as things stand, this doesn’t seem particularly fair. Luckily, it’s cheap to fix, since the bottom 30% only pays 7% of the cost of this policy if the money is raised through general taxation. If we were to exempt them from these tax rises, we’d still have enough to fund a universal benefit worth £970 for the rest of the population. But the point is worth noting: if universal benefits are to be funded through tax rises, the additional taxes shouldn’t hit the poorest.

The size of the state

So we’ve seen three different ways of paying for a new universal benefit with three different distributional effects: regressive, neutral and progressive. It should be obvious by now that unless you are clear which one of these you think will happen, you shouldn’t be saying anything about the distributional consequences.

How we assume the additional spending will be balanced rather depends on our assumptions about the size of the state. If we believe that tax rises are impossible or highly undesirable, then we are going to want a new universal benefit to be paid for by reprioritising spending. Up to a point we can reprioritise from other universal benefits and get something distributionally neutral – but so much of the British state is already means-tested that (unless you are willing to go for pensions or the NHS) there is limited capacity for such reprioritisation. With a fixed public spending envelope and tax profile there is only so much universality you can afford if you also want to do a certain amount of redistribution and under these assumptions, Paul’s claim that “the winners would be the relatively well-off” is just about defensible – although for a small spending item like the Dilnot reforms it would be quite easy to reprioritise spending in a neutral or progressive way.

But if we assume that more universal benefits will mean higher taxes then the picture looks very different. As shown above, the biggest losers from this would be the richest and the winners would be the people in the middle. And there are a couple of reasons to believe that this is ultimately the more reasonable assumption.

The first is that the people who are in favour of universal benefits tend to be the people who are in favour of higher taxes. Elect a government that does one, they are likely to do the other. We can see this in the recent Labour manifesto, where universal free university tuition was proposed, funded by an increase in income tax. Although the Dilnot Commission didn’t say much about how its proposals should be funded, the previous abortive attempt at social care reform (Labour’s National Care Service) was to be funded by an increase in inheritance tax.

The second is that countries with more universal benefits tend to have higher tax rates. There’s more going on in these figures than just universality versus means-testing, but in general the countries with the most universal benefits, such as the Nordics, France and Belgium, raise the most in taxes. Internationally, means-testing goes with low taxes and universality goes with high taxes.

Some conclusions

To be fair to Paul, he is far from the only person to go around saying that universal benefits are regressive. It’s even true under certain assumptions – specifically that the UK must remain a low tax country. The problem is that no one who writes opinion pieces saying that universal benefits are regressive ever seems to find space to clarify that this is their assumption. If they did, readers would probably notice that this isn’t what the proponents of universality are usually proposing, and it isn’t what countries with more universal benefits do. Moving towards a European model with more universal benefits and higher taxes would most likely amount to a transfer from the rich to the people in the middle of the income distribution. I doubt that fans of redistribution would, as Paul suggests, think this is unfair.

The distributional impact of scrapping tuition fees

Since Labour decided to make the abolition of tuition fees a central piece of their election manifesto, we have been treated to a healthy stream of articles arguing for and against the proposal. There are lots of different arguments to be made here, but this post is just going to focus on one: whether abolishing tuition fees would be regressive.

This seems to be a popular line of criticism among centre-left commentators and Tory politicians alike, and they appear to have some heavyweight backing in the form of the IFS. However, since “regressive” is a poorly defined term and routinely abused in public policy debates, it’s generally a good idea to be suspicious this sort of claim. So let’s take a closer look at how tuition fees (and the idea of scrapping them) hold up against different definitions of “progressive” and “regressive”.

Tuition fee repayments raise more money from richer people

The IFS has done some modelling of tuition fees to work out how much graduates with different levels of lifetime earnings will repay. They find that higher fees mean that most people will never repay all of the money they borrow, so graduates who go on to earn more end up paying more for their tuition. Here’s a chart they published a couple of months ago comparing lifetime repayments under the current system with Labour’s proposal to scrap tuition fees and reintroduce maintenance grants:

It’s fairly clear from this chart that tuition fees are, under a certain definition of the term, progressive. People who go on to be richer pay more into the system while those who go on to earn the least pay no more than they would under Labour’s proposals. So if we define the term progressive to mean that richer people pay more then this is clearly a progressive way to raise money to fund our universities.

As a proportion of income, the richest 10% of graduates pay less than the next 40%

It’s no great surprise that rich people pay more, since repayments are proportional to earnings. In fact, the IFS notes that “for [the] majority of individuals, student loans are almost indistinguishable from an additional 9% graduate tax on their earnings” (or more precisely, their earnings above £21,000). But if we are going to think about student loan repayments as a kind of income tax then this suggests another definition of progressivity. Taxes are often called progressive if richer people pay more as a proportion of their income, not just in absolute terms. This definition sets the bar a bit higher – so do tuition fee repayments clear it?

The IFS doesn’t show these figures as a proportion of lifetime income, but I was able to cobble together a rough version based on the charts in their latest briefing note on the subject. The charts below show the difference in repayments between the current system and Labour’s proposal (i.e. the distance between the top and bottom lines in the IFS chart above) in absolute terms and as a proportion of lifetime income.

Now the picture looks a bit more complicated. Student loan repayments remain progressive at lower end of the income distribution. This is because repayments aren’t taken from the first £21,000 of your income and many graduates with the lowest lifetime incomes never earn much above this threshold. However, repayments now begin to look faily regressive at the top end of the income distribution. The richest 10% of gradutes contribute a smaller share of their income than the next richest 40%. This is because if repayments are a graduate tax, they are a tax with capped lifetime repayments. Once the full loan has been paid off, you stop paying tax. A significant proportion of the top 10% of earners pay off the full loan well before the 30-year limit (when it is written off), so they pay this tax for a shorter amount of time than everyone else.

So student loan repayments are progressive at the bottom of the income distribution, but under certain definitions they are regressive at the top end. But of course, that’s not really the question we are trying to answer here. We want to know whether abolishing tuition fees would be regressive. The IFS seem to think they have answered this question with the chart above, saying that “as high-earning graduates repay the largest share of their student loans, they benefit the most from the removal of tuition fees”. This, however, is nonsense. Unless they believe that we are going to replace the funding that universities lose with money that we have conjured out of thin air, they simply haven’t done the analysis that is required to prove or disprove this claim.

There are other progressive ways to fund university tuition, such as taxes

To work out the impact of abolishing student loans we need to have an idea of what will replace them as a funding stream for universities. The answer is of course that universities will get more direct funding from government, which will in turn be funded by higher taxes now or in the future, or by reductions in other areas of government spending. If the net effect of abolishing tuition and raising whatever taxes (or cutting whatever services) will pay for it leads to rich people doing better and poor people doing worse, then it seems reasonable to call the change regressive.

There are countless different ways to increase taxes, but since we are equating loan repayments to a sort of income tax, let’s focus on that. The chart below shows how the distributional effects of tuition fees compare with income tax.

(Since the IFS charts are based on the graduate income distribution – and graduates are richer than the rest of the population – they understate the progressivity of tuition fees in the population as a whole. I have roughly mapped the figures onto the population-wide income distribution so that we can make a better comparison with income tax figures. The mapping is very approximate, so please take the numbers with a pinch of salt.)

Compared to income tax, tuition fee repayments take less money from the people right at the bottom of the income distribution. This makes intuitive sense: income tax kicks in at £11,500 but tuition fees don’t kick in until £21,000 – plus there are more non-graduates at this end of the distribution who don’t make any repayments at all. At the other end of the distribution, tuition fees also take less money from the richest. The top decile pays around 40% of all income tax, but only around 30% of tuition fee repayments. Again it’s not hard to see why: marginal tax rates increase with income under income tax but not under tuition fees; and many of the richest graduates pay off their loans early and then stop paying their 9% “tax”. The people who do worse under tuition fees are those in the top 40% of the income distribution but outside the top 10%.

So what would  be the distributional impact of abolishing tuition fees and raising the same amount of money through income tax (in a distributionally identical way to current income tax take)? Well you would take some money from the richest 10% and some more from the poorest 60% and give it to those in between. Is that progressive or regressive? Well it’s kind of both: regressive with respect to poor people and progressive with respect to the rich. Either way, it’s quite different to the IFS claim that high-earning graduates would benefit the most from the abolition of fees. If the money was raised by increasing income tax equally for everyone then the richest 10% would be the biggest losers.

Of course, there are lots of other ways to pay for scrapping tuition fees and some clearly are regressive. The tax system as a whole is less progressive than income tax and if the policy was funded by a hike in VAT, for example, poor people would end up paying much more than under the current system and rich people much less. If other areas of government spending were cut you could generate just about any distributional profile you wanted, but since a large part of government spending in the UK is targeted to low-income groups, many options would be less progressive than tuition fees. The point though is that you can’t say whether scrapping one funding stream is progressive or regressive unless you consider its replacement.

Labour’s tax proposals are more progressive than tuition fee repayments

As it happens, the Labour manifesto – in which scrapping tuition fees was the biggest spending item – set out a number of tax rises to pay for their spending commitments, including an increase in income tax. However, the Labour policy wasn’t to increase income tax equally for everyone, but to increase it for the top 5% of earners only. We don’t need to draw any more charts to see that this is a much more progressive way to raise money than tuition fee repayments: 100% of the revenue is raised from the top decile. This would only cover a bit more than half of the cost of scrapping tuition fees, but the manifesto also included promises to reverse recent cuts in inheritance tax and capital gains tax, and charge VAT on private school fees. Scrapping tuition fees and paying for it by making these sorts of changes to the tax system would mean poor people paying less and rich people paying more. It would be progressive.

In conclusion

So what have we learned from all this? Well, tuition fee repayments are a progressive source of funding for higher education, insofar as rich people pay more for the same product. The system is great for the poorest, who may not end up paying anything back at all. However, if you look at how much people pay as a proportion of their income – which is how we usually assess the progressivity of a tax – the system starts to look quite regressive at the top end of the income distribution and the richest 10% seem to do rather well.

But just because tuition fees are fairly progressive, this doesn’t mean that scrapping them is regressive. This would be true if loans were scrapped and people had to pay current tuition fees out of their own pockets, but that is not what is being proposed. If taxes are increased to pay for the change then there are plenty of ways that the net effect could be progressive – and plenty of other ways that it could be regressive. But with the sort of tax increases that Labour set out in their manifesto, it seems clear that the net effect would be to increase the overall progressivity of government activity.

Now I don’t mean to suggest that we should scrap tuition fees just because we can find a more progressive way to fund university education. Distributional effects are not the only criteria by which we need to assess policies. Moreover, unless we think that the state currently does too little redistribution, it’s not clear that we would want every policy change to be progressive – and if we do think that, shouldn’t we be rectifying it by reforming the tax and benefit system? The aim of the university system is to educate, not redistribute, and any effects on access for people from poorer backgrounds seem much more important to reducing inequalities than the direct distributional consequences of choosing between tax-based and loan-based funding. So let’s talk about whether higher levels of debt will put some groups off applying to university, or whether a move back towards tax-based funding would lead to caps on student numbers and exclude poorer applicants. Let’s ask whether scrapping fees would lead to lower per capita funding for our universities, and whether this is a good or bad thing. There are many valid arguments to be made for and against tuition fees, but the claim that scrapping them would be regressive is not one of them.

Disclaimer

The analysis in this post is very rough. I don’t have access to the IFS model so I’ve read numbers off charts in their reports, mapped graduates very approximately onto the population-wide income distribution and conflated lifetime and in-year repayments. It wouldn’t be hard for someone at the IFS to do this analysis properly – and I would argue they should have already done it if they are making claims about who benefits from scrapping tuition fees. I’m pretty sure that the results would be similar to mine, but I would be delighted to be proved wrong. The point of this post isn’t to defend scrapping tuition fees. It is to insist that before we use words like “regressive” to describe a policy, we need to do the distributional analysis properly.

Three golden rules for discussing progressivity

What does it mean for a policy to be progressive? The way this question is addressed by the media (and often by government) can be infuriating. It’s got to the point where I am tempted to say the term should be banned, but instead I am going to make one last attempt to clarify it by proposing three golden rules.

I was reminded of this issue when reading Jo Maugham’s analysis* of the impact of the new “social care precept”. This is essentially a £2bn rise in council tax, a significant proportion of which will be paid by poorer households. Here are the figures that Jo gives:

So is this tax progressive or not? Well, rich people pay more, and for some that’s good enough. Fraser Nelson, for example, likes to point out that “the top 3,000 taxpayers in Britain stump up more income tax than the lowest-paid 9 million”. It is more common to look at how paying tax affects the living standards of different groups by comparing tax paid as a proportion of income. By this measure, council tax is regressive.

This is as far as the discussion usually goes. But both of these comparisons have no basis in reality – unless that reality involves collecting these taxes and throwing the money in the sea. The fact is that this money will show up somewhere else, either as increased spending or lower taxes. Which brings me to my first golden rule: the distributional effect of a policy change can’t be assessed without looking at both sides of the equation.

Since this is called the “social care precept”, we might think that it will lead to increased spending on social care. It’s not easy to find numbers on how social care spending is split between income groups, but modelling done in 2011 for the Dilnot Commission (see figure 11 here) made some estimates. Reading the numbers off the chart, I get something like this.

Social care spending is more heavily weighted towards poor people than council tax collection, so lower income groups make a net gain from this policy. That is, the introduction of this policy increases the total amount of redistribution that the government does, which is the only sensible definition I can think of for the word “progressive”, with reference to a change in policy.

But is higher spending on social care really the effect of this policy? That is, if it weren’t for the social care precept, would we see lower social care spending? You could argue that social care spending is going to rise either way, since we’ve got more old people than ever. If it’s not paid for by council tax rises it will be paid for by higher taxes elsewhere, higher borrowing, or cuts to other services.

So this is my second golden rule: identify a realistic counterfactual. We need to know whether the policy leads to more redistribution than what would otherwise have happened. That “what” can have a huge impact on how we view the distributional consequences.

Let’s say we believe that without the social care precept higher social care spending would have to be funded through an increase in income tax – or perhaps higher borrowing now, funded by future increases in council tax. As Jo points out, income tax is much more targeted on rich people than council tax. Here’s the net effect of raising £2bn through council tax instead of income tax.

This policy change would give money to the richest 20% at the expense of everyone else. I think we can all agree that’s regressive. So depending on the counterfactual, the social care precept is either highly progressive or highly regressive. Take your pick. We need to decide which counterfactual is more realistic. In this case, the first one is probably closer to the truth (raising income tax and borrowing more are not top of this government’s agenda) so I’d argue that the policy is probably progressive.

But there’s another more fundamental question here: how much redistribution do we want? Requiring all policy changes to be “progressive” implies that we think we don’t currently have enough. But at some point, if we were to go on increasing redistribution, we’d have too much. Views will differ wildly as to what the optimal level is, but in principle there must be one. I imagine few people think that there should be no redistribution and few think that redistribution should fully equalise living standards.

And even if we want more redistribution, we might not care if a policy is regressive if its effect on the overall level of redistribution is small and it achieves some other aims. The state does not exist solely for the purpose of moving money from the rich to the poor. My third golden rule is therefore this: the overall level of government redistribution is what matters. We need to know whether we want to increase or decrease it, and how much we care about small changes relative to other policy aims.

These rules really are essential. It is impossible to say anything about whether a policy change is progressive without considering both sides of the equation and being clear about what you think the alternative scenario is. It is impossible to know whether you actually want the policy to be progressive, or whether you care, without considering the overall level of redistribution.

These are not new insights and many organisations (the IFS, the OBR, sometimes even the Treasury) are quite diligent about doing distributional analysis properly. But much of the public discussion of “progressivity” fails to follow a single one of these rules, giving us a rather poor standard of debate on the state’s role in reducing inequalities.

---

* I’m not picking on Jo’s analysis because it is a particularly egregious example. On the contrary, I am picking on it because is one of the better examples of someone tackling this question, so Jo has done most of my work for me.

Means-testing and the progressive state

In Sweden, the amount you get from the state pension is linked to the amount you earned while you were working. That is, the richer you are, the more you get.

I can already hear the howls of “regressive” if this was suggested in the UK. Broadly speaking, our public pension pays out a flat rate (provided you’ve paid national insurance for enough years) which is just about enough for someone to live on. It seems more likely that we would take that away from wealthy people than give them an enhanced rate. Just look at the debate around the winter fuel allowance, which is effectively just a cold-weather top-up to the state pension.

This raises a paradox. Any Guardian reader knows that Sweden is a socialist utopia with high taxes, strong public services and lots of redistribution. How can they be doing something so regressive?

Of course, richer people in the UK also get bigger pensions than poorer people. The difference is that this is done privately, rather than by the government. In the UK, many people pay money into private sector pensions and when they retire they can convert this to an annuity (or at least they could until the government abolished pensions). The Swedish public system mirrors this setup by counting up how much you pay in and converting that to an annuity when you retire. The more you put in, the more you get out.

This is just not how we do things in the UK. State support (the NHS aside) is usually seen as a last resort for the destitute, to prevent poverty and starvation. Means-testing is used to target government resources on the poor and prevent them going to the rich. This is our idea of progressive government.

But our progressive European neighbours like the Nordics actually do a lot less means-testing than we do, while managing to be highly redistributive and have lower inequality than us. How do they manage that? By having higher taxes and a larger public sector. In general, the most progressive countries in the world are characterised not by their efforts to target spending on the poor, but by universal benefits and relatively high, progressive taxes.

Now it is possible to argue that this model is not optimal and that a smaller state would be good for economic growth. In that case means-testing makes sense, since it allows you to achieve more redistribution if you are constrained by the need to keep government small. Alternatively, it could allow you to shrink the size of government if you are constrained by the need to redistribute. Limiting benefits to the poor also supports the small state cause in a more insidious way, by stigmatising claimants and undermining public support for state spending.

I am yet to hear a UK politician make the argument that means-testing is the corollary of small government and low taxes. Instead, Nick Clegg tells us that he “doesn’t see why someone like Alan Sugar should be entitled to a winter fuel payment” since he’s “got a bob or two”. I’m sure that Nick Clegg understands the link between means-testing and the size of the state, but libertarians and state shrinkers clearly think that it is easier to get their policies through by having a pop at the rich than to try to convince voters of the merits of small government. Which is a shame, since there is an interesting debate to be had.