Day: 9 June 2024

  • The Lib-Dem buyback tax won’t raise £1.4bn, and could raise nothing

    The Lib-Dem buyback tax won’t raise £1.4bn, and could raise nothing

    The Lib Dems are proposing a 4% tax on share buybacks that they say would raise £1.4bn/year. It’s based on a similar proposal in America. But circumstances in the US and UK are very different. This means that the rationale for the US tax isn’t relevant to the UK and, more importantly, that the Lib Dem proposal would raise much less than £1.4bn. It’s plausible it could raise almost nothing.

    UPDATED 9 June 2024 to reflect the latest Lib Dem proposal, which ups the main estimate to £2.2bn, but then knocks £800m off out of “caution”. There’s also a fair take on this from fullfact.org here.

    The US tax benefit of buybacks

    In 2022, the US imposed a 1% excise tax on share buybacks.

    Why?

    Primarily because two significant classes of investors in US shares receive a tax benefit from buybacks as opposed to dividends:

    • US retail investors directly hold about a third of the US equity market. Dividends they receive are taxed at up to 23.8% (plus State income taxes, where applicable). Capital growth from a buyback isn’t taxed immediately at all. Some investors will never sell and the gains will never be taxed; if they do sell, long term capital gains are taxed at 20% (plus any State capital gain taxes).
    • Foreign investors hold about 16% of the US equity market. The US imposes a withholding tax on dividends that ranges between 15% (for investors in countries with a favourable tax treaty with the US) and 30% (the worst case). But a buyback increases the value of an investor’s shares; any capital gain made on a subsequent sale by a foreign investor is not subject to US tax at all.

    We can therefore make a rough estimate that paying profit out as a buyback rather than a dividend means a reduction in overall US tax paid of somewhere between 4% and 14% of the amount of the buyback.

    It is therefore not that surprising that the 1% buyback tax did not noticeably reduce the volume of buybacks – the tax is significantly less than the tax benefit.

    How high would the buyback tax have to be to equalise the tax treatment? There is no simple answer, given the diversity of investors and their tax positions, but the simple calculations above suggest the answer is at least 4%.

    It is therefore probably not a coincidence that President Biden is now proposing to increase the tax to 4% (although we understand that this has little chance of becoming law in the current US political environment).

    The UK tax benefit of buybacks

    The differences between US and UK stock markets and tax rules mean that buybacks by listed companies have very little tax benefit in the UK.

    • UK individual investors who hold onto their shares have a big tax benefit from buybacks, as their eventual capital gain would be taxed at 20%, but dividends taxed at a top rate of 39.35%. However UK individual investors directly hold only about 4% of the UK equity market; another 7% is held through ISAs but, as ISAs aren’t taxable, these investors have no preference for buybacks vs dividends.
    • UK companies hold a small proportion of the equity market (1.4% in the 2020 figures). If they participate in a buyback then the position is the same as if they had received a dividend – it’s exempt. If they don’t, then buybacks provide them with a worse tax treatment: corporate capital gains are taxed at 25% but dividends are exempt.
    • Foreign investors hold almost 60% of UK listed shares, but the UK doesn’t tax them on either dividends or capital gains.

    We can therefore make a rough estimate that paying profit out as a buyback rather than a dividend means a reduction in overall UK tax paid of about 0.8% of the amount paid out in the buyback. This is pleasingly close to the existing 0.5% stamp duty charged on buybacks, leaving a surplus benefit of probably no more than 0.3%.

    It is therefore unsurprising that, whilst tax is often cited as a driver of US buybacks, it is not usually cited by market observers as the reason for UK buybacks.

    The consequences

    The lack of a material tax benefit from UK buybacks has two important consequences.

    First, it makes it hard to understand the rationale for a buyback tax. If the Lib Dems want to increase tax on companies, they could increase corporation tax (although I would be sceptical this is a good idea right now).

    Second, it means that the Lib Dems’ revenue projection is wrong.

    The Lib Dems say their 4% tax would raise £1.4bn annually. They haven’t published their methodology – they just say this:

    It’s reasonably clear all they’ve done is multiply 4% by the approximately £50bn volume of buybacks in 2022 and 2023, to come up with a static estimateo f £2.2bn. They then “take a cautious approach to account for potential changes in company behaviour”, and reduce this to their claimed £1.4bn.

    That is, however, not a realistic basis for estimating the revenues for a tax. You have to properly take into account the taxpayer response – the tax elasticity. If we impose a £10,000 tax on men with beards then we cannot calculate the revenue as (£10,000 x 15 million men with beards). There would be an obvious taxpayer response (shaving), and the actual revenue would be close to zero.

    In the case of buyback taxes there is an equally obvious taxpayer response – paying a dividend instead of buying back shares.

    The Lib Dems cite an IPPR paper from 2022 which proposed a 1% tax on buybacks. The IPPR said their proposal would raise £225m, using the same simple methodology now adopted by the Lib Dems, but with an important caveat:

    • A 1 per cent tax on the share buyback schemes of FTSE-listed companies could raise an additional £225 million in a single year at current levels of buybacks. In the US, the Biden administration is seeking to rein in rising levels of share buybacks with a 1 per cent tax on stock repurchasing schemes. The tax will raise revenues to reduce the deficit and invest in renewable energy.

    And the IPPR explicitly warned about the risk of a higher tax:

    Taxing shareholder payments at a higher level would raise more revenue while encouraging companies to reinvest their profits in the productive economy. However, levying a higher tax on buybacks is likely to discourage companies from repurchasing their own stock. It is therefore fair to assume that the revenues raised from a buyback tax will change in line with the level at which they are taxed.

    The current US buyback tax at 1% is considerably lower than the overall tax 4% to 14% tax benefit from buybacks. Biden’s new proposal at 4% approaches the bottom-end estimate, but does not exceed it, and that is surely deliberate. So it would be rational to expect the 4% tax to somewhat reduce the volume of buybacks, but only to a degree.

    The Lib Dem tax is very different, because it is at least four times greater than the tax benefits of buybacks (particularly once we take account of the existing 0.5% stamp duty). There are other benefits of buybacks; they can be more flexible, and they send out price signals (inflating EPS but without a “real” economic effect). It is not at all obvious that these, rather ephemeral, benefits are worth 3% of the value of a buyback.

    The natural conclusion is that a 4% buyback tax will simply result in companies switching from buybacks to dividends. And because 95% of investors receive no tax benefit at all from buybacks, but would bear the cost of the 4% buyback tax, there would likely be significant shareholder pressure to drop buybacks entirely.

    The other justification provided for the tax is that it would increase investment. This doesn’t make any sense. If a company has decided to return cash to investors then a buyback tax may incentivise it to move to a dividend; it’s unclear why it would incentivise it to retain the cash. It also seems simplistic to regard cash retained by a company as investment, but cash returned to shareholders as simply disappearing.

    So the cautious estimate is not £1.4bn – it’s nothing.

    We agree with Stuart Adam from the Institute of Fiscal Studies:

    That is, however, not the end of the analysis, because there are second order effects:

    • Buybacks are currently subject to 0.5% stamp duty/stamp duty reserve tax. So an end to buybacks would mean a loss of c£250m of stamp duty revenue.
    • An end to buybacks means more dividends, so the c4% of UK individuals directly holding shares would pay more income tax. On the basis of our top-end estimate above, this amounts to somewhere less than 0.8% of buyback values i.e. £400m of additional tax revenue.
    • Then there are the costs to Government/HMRC of creating the tax, and the cost to business of complying with it.

    We don’t have enough data to properly estimate the net result of these effects. They would probably be small, but the direct revenues from the tax would also probably be small.

    There are many historical examples of people taxing shares without thinking through how people would respond. These usually ended badly – the Swedish financial transaction tax and US interest equalisation tax are the most notorious examples.

    The general rule remains that your motive for introducing a tax is irrelevant. The key questions are: what will happen in practice? What incentives are you creating? How will people respond?

    It’s all very tedious. It’s also necessary.

    Photo of Ed Davey by Dave Radcliffe, licensed under Attribution-NoDerivs (CC BY-ND 2.0)

    Footnotes

    1. “directly” meaning this is excluding holdings through ETFs, mutual funds and pensions, which have different tax treatment ↩︎

    2. i.e. because the minimum saving will be (34% x 3.8% + 16% x 15%) and the maximum saving will be (34% x 23.8% + 16% x 30%). This ignores State taxes and a large number of other complications, so should be regarded as no more than a very rough approximation ↩︎

    3. In principle one might say that there should be a different result, because the majority of investors in US equities obtain no tax benefit from buybacks, but now suffer the cost of the excise tax, and they could be expected to agitate against buybacks. A plausible answer is that retail investors have an outsize influence. ↩︎

    4. After writing the first draft of this piece, we found this analysis by the left-leaning Tax Policy Center, which uses different data and a slightly different approach but also concludes the answer is around 4%. ↩︎

    5. There is a separate question about unlisted/private companies engineering a return of capital rather than a dividend to obtain a tax advantage, i.e. because of the large differential between the 39.35% top rate of income tax on dividends and the 20% capital gains tax rate. The Lib Dems aren’t proposing to tax private companies but, even if they were, a buyback tax would not come close to reversing this benefit. The more effective and simpler answer would be a specific anti-avoidance rule. ↩︎

    6. Investors whose shares are bought back are mostly taxed on the buyback as income, as generally only the nominal value of the share is treated as a capital gain. Hence a rational UK individual investor will not take-up a buyback; the tax treatment is much worse than simply selling their shares in the market. ↩︎

    7. UK individual investors hold about 11% of the UK listed market, equating to about £250bn. However ISA investors hold about £400m of stocks/shares and have a 37% weighting towards the UK, implying they hold about £150bn of UK equities. Thus only 40% of UK individual investors’ holdings in UK listed equities are held directly. ↩︎

    8. There are exceptions to both rules, but for listed companies the exceptions generally won’t apply. ↩︎

    9. i.e. because 4% x 19.35% = 0.8%, but that’s a top-end estimate because many investors won’t pay the additional rate, and any investors actually participating in the buyback pay more tax as a result. ↩︎

    10. As an aside, whilst it’s a very good idea to benchmark tax policy proposals against other countries’ experiences, it’s dangerous to assume that a tax policy that’s successful in one country will also be successful in another. There are a myriad of tax, legal and societal reasons why that is often not the case. In this instance it’s the difference between the US and UK markets plus the difference in the tax treatment of foreign investors – both are sizeable differences, and together they make the buyback landscape in the US markedly different from the UK ↩︎

    11. There has been research into the impact of stamp duty on share trading, and the elasticity of share prices with respect to transaction costs. In principle similar research could look at the impact of existing stamp duty on buybacks (by reviewing data from when the rate changed from 1% to 0.5% in 1986. However I’m not aware of anyone doing this; possibly the volume of buybacks around 1986 was too small to make this feasible. ↩︎

    12. There are other problems with the estimate. The US buyback tax exempts certain types of mutual funds because they engage in buybacks to minimise their share price discount vs their NAV. Realistically a UK buyback tax would have to exempt investment trusts, and probably create other exemptions too. So the £2.2bn estimate is wrong even if we ignore elasticity/taxpayer response, but elasticity is by far the most important effect. ↩︎

    13. Not quite zero, because some people would make a mistake; some people would try and fail to avoid the tax (with complex boundaries between moustaches and beards, and difficult caselaw around false beards). And having a beard would become a signal of enormous wealth ↩︎

    14. It’s sometimes suggested that buybacks are used by executives to manipulate their own remuneration targets. This would be possible in theory if executive remuneration packages are not designed and implemented carefully. A detailed study looked at the FTSE 350 to see if there was evidence of buybacks inflating executive pay – it found that there was not. ↩︎

    15. Warren Buffet said “When you are told that all repurchases are harmful to shareholders or to the country, or particularly beneficial to CEOs, you are listening to either an economic illiterate or a silver-tongued demagogue (characters that are not mutually exclusive).” ↩︎

  • Stamp duty is a terrible tax. We should abolish it – but there’s a price.

    Stamp duty is a terrible tax. We should abolish it – but there’s a price.

    Stamp duty is a terrible tax. The Tories want to abolish it for most first time buyers. But the evidence shows that cutting stamp duty increases house prices, and that previous attempts to provide relief for first time buyers were ineffective.

    Council tax is also terrible tax – with Buckingham Palace paying less council tax than a semi in Blackpool.

    We can solve both problems together, and tax land in a way that encourages housebuilding and economic growth. But that requires smart thinking and brave politics.

    The problem with stamp duty

    Stamp duty land tax (SDLT) is a deeply hated tax.

    It reduces transactions, distorts the housing market, and often stops people moving when they want to. Stamp duty makes it harder to borrow from a bank (because the stamp duty is “lost value”). All of this means it reduces labour mobility, results in inefficient use of land, and plausibly holds back economic growth.

    And the rates are now so high that the top rates raise very little; HMRC figures suggest that increasing the top rate any further would actually result in less tax revenue.

    It also makes people miserable.

    Stamp duty only exists because, 300 years ago, requiring official documents to be stamped was one of the only ways governments of the time could collect tax. We have much more efficient ways to tax today – but stamp duty remains. Until four years ago HMRC even still used the Victorian stamping machine in the picture at the top of the page.

    The problem is that, like many bad taxes, politicians have become addicted to it. SDLT now raises £12bn each year – an amount that’s hard to ignore.

    And there’s an even worse problem: abolition would inflate property prices.

    The problem with abolishing stamp duty

    The link between stamp duty and prices is clear when we look at the impact on house prices of the stamp duty “holidays” in 2021:

    The spikes in June and September coincide with the ends of the “holidays”. A rush of people to take advantage of the discounted stamp duty.

    Of course the “holidays” were temporary – but the chart suggests that there was a permanent upwards adjustment in house prices (probably due to the “stickiness” of house prices).

    Previous stamp duty holidays had less dramatic effects. There’s good evidence that the 2008/9 stamp duty holiday did lead to lower net prices, but 40% of the benefit still went to sellers, not buyers. I’d speculate that the difference is explained by the much lower stamp duty rates at the time.

    A detailed Australian study looked at longer-term changes than the recent UK “holidays” – it found that all the incidence of stamp duty changes fell on sellers (and therefore prices). This is what we’d expect economically in a market that’s constrained by supply of houses.

    These effects mean that stamp duty cuts aimed at first time buyers may end up not actually helping first time buyers. An HMRC working paper found that the 2011 stamp duty relief for first time buyers had no measurable effect on the numbers of first time buyers.

    The problem with council tax

    Stamp duty isn’t our only broken property tax. Council tax is hopeless – working off 1991 valuations, and with a distributional curve that looks upside down.

    We can see the problem immediately from the Westminster council tax bands:

    The bands cap out at £320k – equivalent to about a £2m property today. So there are two bedroom apartments paying the same council tax as Buckingham Palace.

    And the top Band H rate – restricted by law to twice the Band D rate, is pathetically small compared to the value of many Westminster properties.

    The problem is then exacerbated by the fact that poorer areas tend to have higher council taxes. Here’s Blackpool:

    So Buckingham Palace pays less council tax than a semi in Blackpool.

    That’s why, if we plot property values vs council tax, we see a tax that hits lower-value properties the most:

    In a sane world, this curve would either be reasonably straight (with council tax a consistent % of the value of the property), or it would curve upwards (i.e. a progressive tax with the % increasing as the value increases). This curve is the wrong way up.

    The solution

    The solution is to fix council tax and stamp duty at the same time.

    Abolish stamp duty altogether, and change council tax to make it fairer… calibrating that change so that end of stamp duty doesn’t just send house prices soaring. This is not an original proposal – it was one of the recommendations of the Mirrlees Review in 2010. Paul Johnson of the Institute of Fiscal Studies has also written about it.

    But we can go further. The really courageous answer is to scrap council tax, business rates and stamp duty – that’s about £80bn altogether – and replace them all with “land value tax” (LVT). LVT is an annual tax on the unimproved value of land, residential and commercial – probably the rate would be somewhere between 0.5% and 1% of current market values. This excellent article by Martin Wolf makes the case better than I ever could.

    There are two amazing things about LVT.

    The first is that it has support from economists and think tanks right across the political spectrum. How many other ideas are backed by the Institute of Economic Affairs, the Adam Smith Institute, the Institute for Fiscal Studies, the New Economics Foundation, the Resolution Foundation, the Fabian Society, the Centre for Economic Policy Research, and the chief economics correspondent at the FT?

    The second is that everyone says it’s politically impossible.

    I wonder how true that is.

    So let’s definitely not do land value tax. Let’s instead abolish stamp duty and fund it by adding some bands to council tax, so it more closely tracks valuations. Most people will pay a bit more tax, but not much more – and it’s worth it to get rid of the hated stamp duty. Whilst we’re at it, let’s update valuations more regularly, so it’s fairer. And why not make it apply to the unimproved value of land, so people aren’t punished for improving their property?

    Everyone agrees business rates need reform – so let’s make similar changes to business rates.

    What we end up with won’t be called “land value tax”, and won’t exactly be a land value tax. But it’s getting awfully close.

    The price

    That’s the price of abolishing stamp duty: some of us have to pay a bit more council tax (or, in my fantasy world, land value tax). That’s worth doing for a saner housing market that doesn’t hold back growth. And a land value tax should encourage house-building and actually boost growth.

    But if all we do is abolish stamp duty, most or all of the tax saved by buyers will be eaten up in higher property values. It becomes a £12bn government handout to sellers.

    There’s no free lunch. But there is an opportunity for a big pro-growth tax reform. It might even be popular.

    Photo of original stamping machine is Crown copyright, and reused here under the Open Government Licence

    Many thanks K for assistance with the economic aspects of this article.

    Footnotes

    1. Apologies to all tax professionals, but I’m going to call SDLT “stamp duty” throughout this article. ↩︎

    2. The elasticities found in HMRC research are incredible; a 1% change in the effective tax rate results in almost a 12% change in the number of commercial transactions and a 5-7% change in the number of residential transactions. (Strictly semi-elasticities because they are by reference to absolute % changes in the tax rate, not percentage changes in the % tax rate). ↩︎

    3. Hello tax professionals. Yes, I know stamp duty and SDLT parted ways in 2003… but the point about the antiquated nature of stamp taxes remains valid. And I like the picture. ↩︎

    4. There’s some published research on the 2021 holiday, but it’s qualitative as it was completed too soon to catch the September heart attack. I’m not aware of anything more recent, which is a shame – 2021 was a brilliant double natural experiment. ↩︎

    5. i.e. because tax incidence theory says that where supply is inelastic and demand is elastic, the seller bears the incidence. ↩︎

    6. Meaning the Royal Residence at Buckingham Palace – most of the rest of the complex isn’t a dwelling, and pays business rates not council tax. I haven’t seen any data on the value of the Royal Residence, but safe to assume it is very high indeed. ↩︎

    7. i.e. because economically we can expect the present value of future council tax payments to be priced into house prices, and if we increase council tax slightly at the low end and significantly at the high end, we should be able to undo the price effects of abolishing stamp duty. ↩︎

    8. A quick health warning: many of the people and websites promoting land value tax are eccentric. I once had a lovely discussion with someone from a land value tax campaign. After a while I asked what kind of rate he expected – 1% or 2% perhaps? His answer was 100%. Land value tax’s supporters remain one of the biggest obstacles to its adoption. They often suggest income tax/NICs, VAT and corporation tax could all be replaced with LVT – a look at the numbers suggests this is wildly implausible. ↩︎

    9. i.e. as if there was nothing built on it. ↩︎

    10. Meaning a higher % of the unimproved value; but it’s the % of market value that people will care about when the tax is introduced. ↩︎

    11. That would be quite unfair on someone who has just paid a large SDLT bill to buy an expensive property – they get punished under the new rules and the old. It would make sense to give some form of relief for recent SDLT… for example allowing SDLT to be written off over ten years worth of neo-council tax/LVT. So for example someone who paid SDLT nine years ago would get 1/10th of that credited against the new tax for one year. Someone who paid SDLT yesterday would get 1/10th of that credited for each of the next ten years. But this is one of many ways it could work. ↩︎