Ways of taxing wealth: alternatives and interactions

In this paper, I examine the role of a wealth tax in the context of the UK’s existing taxes on wealth. First, I discuss several ways in which the UK could be said to tax wealth already, and I set out two possible directions for reforming these taxes, highlighting policies that are merited under either approach. Second, I consider whether and under what circumstances a broad-based tax on the ownership of wealth – a ‘wealth tax’ – could be justiﬁed instead of or in addition to these reforms. Third, I address how a wealth tax should interact with other taxes, focusing on concerns regarding ‘double taxation’ and (conversely) proposals for an alternative minimum tax based on wealth. I conclude that there is a large degree of consensus amongst existing proposals to reform our current taxes on wealth, and that most of these reforms would be required whether or not a wealth tax is introduced as well.

taxes on wealth as part of a review into the entire tax system. 1 But a range of institutions have made recent recommendations on specific taxes. Each of these proposals differs somewhat in its details, and to some extent its objectives -but there is also a large measure of agreement between the proposals.
Despite these recommendations, minimal progress has been made. There have been some positive developments: for example, an increase in dividend rates in 2016, and measures to limit business asset disposal relief (previously known as entrepreneurs' relief) for capital gains tax (CGT). But in the period since the 2008 financial crisis, there have been hardly any structural changes and, for some taxes, one must look much further back for the last major reform. 2 More active areas of legislation, such as CGT, have been characterised by as many steps backwards as forwards.
Given the flaws in the UK's existing approach to taxing wealth, could a new broad-based tax on the ownership of wealth -a 'wealth tax' -provide the solution? Two key issues must be considered here.
First, to what extent (if at all) is a wealth tax really an alternative to reforming existing taxes? Second, if a new wealth tax were to be introduced, how should it interact with the other taxes in our current system? In this paper, I provide a framework for answering both of these questions. My focus is on the UK context, although the many of the issues and principles also apply to other developed tax systems.
In Section 2, I outline what I mean by 'taxes on wealth' (distinct from a 'wealth tax'). In Section 3, I discuss two main directions for reforming the UK's existing taxes on wealth. The aim here is to synthesise existing proposals and to highlight policies that are merited whichever broader view is taken. Section 4 addresses two questions. First, could a wealth tax provide an alternative to any of the reforms canvassed in Section 3? Second, if these reforms were implemented in full, would there be any role left for a wealth tax to play? Finally, in Section 5, I consider the interactions between a wealth tax and other existing taxes on wealth, focusing on concerns about 'double taxation' and (conversely) proposals for an alternative minimum tax based on wealth.

WHAT IS A 'TAX ON WEALTH'?
A wealth tax, as I use the term in this paper, has two essential characteristics. First, it is based on the ownership of wealth; that is, the amount of tax that a person pays depends exclusively on how much wealth they own, rather than (for example) how they acquired the asset or what they do with it. Second, it is at least to some extent broad-based, in that it applies to more than one type of asset; this distinguishes a wealth tax from, for example, a narrower 'property tax'. 3 The UK currently has no taxes of this type. So, when commentators talk about reforming our existing taxes on wealth, which taxes do they have in mind? There is no universal definition or standard basket, so the answer to this question requires some second-guessing. At the core are surely the classic 'capital taxes': CGT, IHT and stamp duty land tax (SDLT). 4 To this can be added income tax on savings and investment income (including the investment element of pension income when taxed on withdrawal), 5 on the basis that -like CGT -it taxes the returns that flow from having wealth.
Beyond these core taxes on wealth, the choice of which taxes to include in the definition is more contestable. Value added tax (VAT) could be thought of as a tax on (some forms of) expenditure out of accumulated wealth. Corporation tax reduces the effective return on wealth held in the form of shares.
Council tax could be reformed to tax the ownership of wealth held in residential property, and the annual tax on enveloped dwellings (ATED) already does this in limited circumstances. All of these taxes potentially affect those with wealth, to varying degrees.
Consequently, to avoid excluding reforms that people might plausibly consider as alternatives to introducing a wealth tax, I adopt a broad notion of taxes on wealth for the purposes of this paper. This involves looking beyond the core capital taxes to include the multitude of other ways in which the tax system can affect people who own substantial wealth ('the wealthy'). 6 Below, I outline four main ways in which the UK tax system could be said to tax the wealthy already. 7 (1) Income tax and CGT tax the returns to wealth. Corporation tax also affects returns received by shareholders.
(2) IHT and SDLT tax transfers of wealth (although with SDLT there is typically no net transfer, merely an exchange).
(3) ATED taxes the ownership of wealth held in residential property, where the property is owned by a company and certain other conditions are satisfied. Owner-occupiers of residential property are liable to council tax. (4) VAT taxes some of the ways in which wealth may be spent.
For readers not already familiar with the UK system, the key features of each of these taxes are succinctly summarised by Pope and Waters (2016). 8 In combination, their effect on wealthy individuals depends on several factors, including the forms in which they hold their wealth, the extent of the financial returns (income and gains) that they receive from it, and the ways in which they spend or otherwise choose to use their wealth, for example by making gifts or bequests. Unlike under a wealth tax, there is no direct relationship between the amount of wealth that an individual owns and the amount of tax that they pay. I discuss whether there should be such a link in Section 4.2.

DIRECTIONS FOR REFORM
Proposals for reforming the UK's existing taxes on wealth are not in short supply. Many of these have a long history, traceable at least as far back as the Meade Report, published by the Institute for Fiscal Studies in 1978. 9 New recommendations have proliferated in recent years, as taxes on wealth have moved up the political agenda and attracted more attention from policymakers. The abundance of proposals can make it more challenging to identify coherent directions for reform. However, most proposals turn out to reflect one of two possible directions. One package of reforms works in the direction of a comprehensive consumption tax (CCT). This approach reflects the premise that an individual's welfare (or ability to pay) 10 is best reflected by their lifetime consumption (i.e. the total amount that they consume over their whole life). In order to raise revenue and redistribute efficiently, the tax system should therefore rely on consumption as the main tax base, but it must avoid distorting individual's choices over when they consume (i.e. whether to spend their savings now or later). In the UK, the Mirrlees Review provides the flagship example of this approach. 11 However, the review recognised that consumption is difficult to measure directly and that our current tax system relies heavily on income taxes. Consequently, the package of reforms proposed 6 The amount of wealth required to be considered wealthy is of course highly contested and I will not attempt to specify a threshold. For a recent qualitative study of lay perceptions on this issue, based on focus groups held in London, see Davis et al. (2020). 7 This approach reflects the tripartite scheme adopted in Hills et al. (2013), but with the addition of taxes on spending wealth. 8 Boadway et al. (2010) and Summers (2019) provide further detail and evaluation. 9 Meade et al., 1978. 10 I return to the distinction between welfare and ability to pay in Section 3.1.3 below. 11 Mirrlees et al., 2011. by Mirrlees et al. sought to achieve consumption tax treatment indirectly, through a combination of reforms to taxes on expenditure and taxes on income/gains, where the latter were designed effectively to serve as advance taxes on future consumption.
Another package of reforms instead follows a comprehensive income tax (CIT) approach. This approach is founded on the view than an individual's ability to pay depends on their control over economic resources, measured by their income. Although there are different ways in which income could be defined for these purposes, 12 the traditional conception of comprehensive income is based on the amount that an individual could consume over a given period without reducing their net wealth. 13 In policy terms, the amount of tax an individual pays depends on how much income they receive, but income is construed broadly to encompass all forms of accretion to wealth. In addition to the types of income currently charged to income tax, this includes accrued capital gains and any other receipts such as inheritances or gifts.
There are two main conceptual differences between the Mirrlees approach and the CIT approach in relation to taxes on wealth.
The first concerns the tax treatment of returns on wealth (i.e. savings and investment income) and capital gains. Both the Mirrlees approach and the CIT approach seek to align tax rates on returns on wealth with the rates applying to income from work, and to avoid distortions to investment decisions across different types of asset. Both approaches favour an allowance -taking effect as a deduction from the tax base -applied to the nominal returns on wealth. However, the nature and purpose of this allowance differs. Mirrlees favours an allowance for the normal rate of return, reflecting the idea that the tax system should leave individuals indifferent as to whether they consume now, or save to consume later. 14 By contrast, the CIT approach favours an allowance for inflation, reflecting the idea that ability to pay depends on how much an individual's real resources have increased.
The second main conceptual difference concerns the tax treatment of transfers of wealth. Under a CIT, all inheritances and gifts should be taxed as income of the donee in essentially the same way as any other income (e.g. earnings), albeit with some provision to smooth the relative lumpiness of such receipts. Under a CCT, there would instead be a strong case for taxing bequests and lifetime gifts as consumption of the donor, and additionally (under a Mirrlees framework) levying income tax as advance tax on the consumption of the donee. However, the Mirrlees Review did not adopt this approach. Instead, it recommended a donee-based tax known as a lifetime receipts tax (described in more detail below) with no tax on the donor.
It is important to note that taxes on the ownership of wealth (such as a wealth tax) are not integral to either approach. The CIT approach relies on income as the tax base; the Mirrlees Review aims to approximate a consumption tax base. Neither includes wealth as a distinct tax base. Nevertheless, it remains an open question whether taxes on the ownership of wealth could be motivated on other grounds. The contributors to the Mirrlees Review explicitly rejected any reliance on wealth as a tax base. 15 However, the earlier Meade Report concluded in favour of including wealth as a distinct tax base in addition to either consumption or income. I explore the reasons for this divergence in Section 4.2.3. 16 Having set out this high-level comparison between the Mirrlees and CIT approaches, the following two subsections outline the main proposals for reform entailed by each approach, linking these with other proposals that reflect a similar direction of travel. 12 Meade et al, 1978, Chapter 3. 13 This is known as the 'Haig-Simons' comprehensive income definition. More formally, comprehensive income is equal to consumption plus change in net wealth (over a given period), where capital expenditure is amortised rather than immediately expensed. 14 For further explanation, see the first paragraph of Section 3.1. 15 Banks and Diamond, 2010;Boadway et al., 2010. 16 See also Adam and Miller, 2021.

Mirrlees Review
The Mirrlees Review recommends reforming taxes on the returns from wealth by aligning tax rates across all sources of income and capital gains, but with a rate of return allowance (RRA) set by reference to the normal rate of return. The theory behind this approach follows from the aim of achieving comprehensive consumption tax treatment. The idea is to compensate individuals for their choice to save rather than spend immediately, such that the tax system does not influence their decision whether to consume now or later. The RRA does this by providing an allowance equal to the normal or risk-free rate of return, which in simple terms is the interest rate that someone could expect to receive without putting their capital at risk. The Mirrlees Review suggested that this rate should be measured by the yield available on medium-maturity government bonds (gilts). 17 In implementation, the RRA would operate by allowing individuals to claim a deduction equal to the base cost of their investment multiplied by the normal rate of return over the period when the returns accrued. 18 Where an asset yields both income and gains, individuals should receive only one allowance set against both forms of return. 19 If the actual yield after combining income and (realised) gains turned out to have been less than the normal rate of return over the relevant period, the individual would be treated as having made a loss to that extent and this could be offset against other income and gains. It is important to note that alignment with the tax rates on income from work should include national insurance contributions (NICs) as well as income tax; in the case of dividends, alignment should also take account of corporation tax. 20 In application to income tax, the implementation of this approach varies depending on the type of investment, essentially for reasons of practical administration. For example, given that the interest available from savings held in bank accounts typically does not exceed the normal rate of return, these returns can simply be exempted from tax, equivalent to removing the limit on cash ISAs. For pension savings, the Mirrlees Review recommends retaining the current basic structure, but abolishing the 25 per cent tax-free lump sum and resolving the anomaly whereby NICs are not charged on either contributions or withdrawals. For most other sources, the standard approach of taxing at aligned rates, but subject to an RRA, would apply.
In application to CGT, the Mirrlees approach involves aligning rates with the combined rates of income tax and NICs, but again applying an RRA. This implies a very large increase in headline CGT rates. It is also important to note that, even accounting for the RRA, these reforms would also significantly increase effective rates on capital gains for many individuals. Given that the base cost of investments for many types of gain will be close to zero -for example, on carried interest or many of the gains realised by owner-managers currently qualifying for business asset disposal relief -the RRA available on these investments would also be close to zero. Consequently, a faithful implementation of the Mirrlees approach would -given current tax rates -result in effective top tax rates on gains by private equity managers and (some) business owners exceeding 60 per cent. 21 In relation to taxes on transfers of wealth, the Mirrlees Review was strongly critical of both SDLT and IHT. They recommended abolishing SDLT altogether, though this was linked with other proposals that would increase council tax on the most expensive properties (see further below). On IHT, Boadway et al. (2010) and Mirrlees et al. (2011, Chapter 15) canvassed two main options for reform. Their first preference -albeit with some reservations about administrative feasibility -was to replace IHT with a lifetime receipts tax (described below). 22 Alternatively, a second-best option would be to overhaul the existing IHT to reduce opportunities for tax planning by expanding the tax base and eliminating reliefs. Neither of these proposals follows from the comprehensive consumption tax approach that forms the centrepiece of the Mirrlees Review. Instead these reforms were motivated mainly by a distinct concern for equality of opportunity. The implications of this objective, in the context of a wealth tax, are discussed further in Section 4.1.2.
A lifetime receipts tax would be based on the cumulative amount received by the donee over their lifetime, rather than the amount given away by the donor as under the existing IHT. By focusing on the windfall to the donee, this form of tax more aptly emphasises the equality of opportunity motivation. 23 Unlike the UK's existing IHT, which only taxes lifetime gifts if made within seven years prior to death, a lifetime receipts tax would draw no distinction between inheritances and gifts regardless of when they were made. 24 However, the tax base for these receipts would be distinct from other taxable income and gains and subject to a separate lifetime allowance. In Section 3.2, I discuss separately the possibility of incorporating inheritances and gifts received within the income tax base of the donee.
The Mirrlees Review also recommended extensive reforms to council tax, although these were not conceived in terms of a tax on ownership of wealth. Instead, they focused on two other aspects of residential property. First, houses provide a home to live in: the occupier consumes housing services. Mirrlees et al. (2011, Chapter 16) recommended that this consumption value be taxed at a flat annual rate equivalent to VAT based on the property's rental value (known as housing services tax). Second, houses provide an investment: owner-occupiers could alternatively rent their house to someone else, and the house may gain in value. This calls for taxing housing as a return on wealth received by the owners, based either on the actual rental income (if the property was let) or an imputed rent (if owneroccupied), 25 plus any capital gain on sale. However, these returns would again be subject to an RRA equivalent to the normal rate of return on the initial purchase price.
In combination, these two new forms of tax on residential property would imply a huge increase in tax on owner-occupiers compared with the status quo under council tax. For example, consider a property in the London Borough of Westminster worth £1 million, with a rental yield of 4 per cent (£40,000). The occupier currently pays around £1,500 per year in council tax. Under a housing services tax set equivalent to VAT (at 20 per cent) this would rise to £8,000 (i.e. 0.8 per cent of the value of the property). On top of this, the owner would pay tax on the actual or imputed rental value of the property (£40,000), added to their taxable income. Because the normal rate of return is currently close to zero, the deduction of the RRA would be approximately nil. Consequently, if the owner was a higher-rate taxpayer, they would pay income tax of £16,000 (at the margin). For an owner-occupier, who would pay both the housing services tax and income tax on the imputed rent, their total tax bill would therefore rise to £24,000 in respect of this property. 26 Recognising the step change in tax liabilities that this approach would entail, Mirrlees et al. (2011, Chapter 16) suggested that housing services tax initially be set at a rate somewhat less than the equivalent rate of VAT (a rate of 12.5 per cent was estimated to achieve revenue-neutrality with the existing council tax). They also did not recommend the actual implementation of an additional tax on 22 This policy has also been proposed on a number of other occasions, including by the Resolution Foundation (Corlett, 2018) and the Institute for Public Policy Research (Dolphin, 2010). 23 Sandford, 1987;Bird-Pollan, 2017. 24 However, there is nothing inherent in a donor-based tax that requires it to leave most lifetime gifts tax-free. In the UK, the donor-based capital transfer tax that operated from 1974 until 1986 was charged on all lifetime gifts and bequests above a lifetime allowance. the imputed rent of owner-occupiers. However, these concessions were explicitly born out of political expedience rather than a principled position. One must acknowledge that the logic of the Mirrlees approach is quite radical in its application to housing. This mostly reflects the fact that owner-occupied housing is grossly undertaxed in our current system. The Mirrlees Review also recommended a significant expansion of the VAT base. All goods and services that currently receive a reduced or zero rate, or are exempt, would instead be charged at the main rate of 20 per cent. Consequently, VAT or an economic equivalent would be paid in full on (amongst other items): financial services, private healthcare and private education. These listed expenditures are likely to be disproportionately concentrated amongst the wealthy, so this element of reform could play a role akin to an additional tax on spending wealth. However, the expansion of the VAT base would also entail application of the 20 per cent rate to (amongst other items) food and children's clothing, which it would be more difficult to characterise in these terms. The Mirrlees Review did not make any recommendations to increase taxes on the consumption of durable assets other than housing, leaving a gap where these assets were undertaxed (or not taxed at all) when new.

Comprehensive income tax
Under a CIT approach, all income from wealth and (accrued) gains would be taxed at the same headline rates as income from work, but with an allowance for inflation. Just as with the Mirrlees approach, alignment should take account of NICs as well as income tax, and corporation tax in the case of dividends. The key difference is that under a CIT the allowance would be for inflation instead of the normal rate of return. Again, to implement the allowance, individuals would claim a deduction equal to the base cost of their investment multiplied by the rate of inflation over the period during which income and/or gains accrued. In the case of CGT, this was the approach used previously in the UK under indexation allowance, which applied (in full) from 1988 to 1998. 27 A return to this approach was recently proposed by the Institute for Public Policy Research. 28 In relation to taxes on transfers of wealth, a CIT approach implies that inheritances and gifts should be taxed on the donee in the same way as other forms of income (i.e. included in the income tax base). 29 This gives rise to some challenges resulting from the lumpiness of these types of receipt, but Meade et al. (1978, Chapter 7) identify several possible solutions using different approaches to averaging; the policy would be otherwise relatively straightforward to implement from an administrative -if not political -perspective. 30 Although proposals have been made along these lines in the past, 31 recently the option of taxing inheritances and gifts as income has not received as many backers as the lifetime receipts tax. 32 It is difficult to discern the reasons for this, but a relative lack of international examples -compared with a lifetime receipts tax -may be one explanation.
A CIT approach also implies that council tax should be replaced with a tax on the imputed rent of owner-occupied properties, although allowing a deduction for mortgage interest as an expense incurred in obtaining this benefit. The net rent would be aggregated with the individual's other sources of income and gains and taxed at their standard income tax rates. This approach would effectively 27 However, during this period, although CGT rates were aligned with income tax rates, this alignment did not take into account NICs. 28 Nanda and Parkes, 2019. 29 The Meade Report highlighted several options for the tax treatment of the donor in these circumstances (Meade et al. 1978, Chapter 3); however, typically it is assumed that the donation would be treated as an expenditure rather than a subtraction from the donor's income, so would have no tax consequences for the donor under CIT treatment. 30 Challenges would remain in relation to trusts where the wealth could be held indefinitely in that vehicle but accessed by beneficiaries. Issues such as how gifts between spouses and gifts to non-residents should be taxed would also need consideration. 31 Goodhart, 1988;Robinson, 1989. 32 The Resolution Foundation modelled the revenue that such a policy could raise but did not adopt it as their central recommendation (Corlett, 2018). mark a return to the old 'Schedule A' charge to income tax, although with an additional allowance for inflation. 33 Schedule A was abolished in 1963, although the associated mortgage deduction was only finally phased out in April 2000. Perhaps scarred by this experience, there do not appear to have been any recent proposals for a return to this approach, although it has some affinity with the tax on imputed rent canvassed (but not recommended) in the Mirrlees Review.

Reforms merited under either approach
It is important not to overstate the differences between the consumption tax treatment recommended by the Mirrlees Review, and a CIT approach. Below, I outline a series of reforms to the UK's existing tax system that would be merited under either approach. These reforms have also been proposed or endorsed by a range of other authors and institutions in addition to the Institute for Fiscal Studies (which was responsible for the Mirrlees Review); some prominent examples are cited below. From a policy perspective at least, these reforms are the 'easy wins' in the sense that they correct anomalies in our current approach to taxing wealth that cannot be justified on any coherent basis. Both the Mirrlees and CIT approaches recommend full alignment of the effective tax rates on returns from wealth (income and gains) with the rates applicable to earning from work. In both cases, this includes levying an equivalent of NICs on investment income and capital gains, accounting for both the employee and employer components. The main difference between the two approaches is whether to give an allowance for the normal rate of return or for inflation. However, this distinction has been relatively unimportant over the past decade because both rates are very low by historic standards. Figure 1 shows a comparison since 1970. To be sure, the difference between these rates has been substantial at some points in the past, and could well be again, but at the time of writing the issue of what type of allowance to give remains of second-order importance compared with the disparities in headline rates across different forms of income and gain.
Consequently, both the Mirrlees and CIT approaches would result in significantly higher effective tax rates on returns on wealth than at present. An exception is for assets that yield a low nominal return (less than or on a par with inflation or the normal rate of return) such as savings held in bank accounts, for which the tax rate would decrease; however, for most people these returns are already tax-exempt if held in ISAs. In some cases, the tax increase resulting from implementation of either the Mirrlees or CIT approach would be very large, particularly on dividends received by owner-managers and associated gains, where the low base cost of the original capital investment means that only a small allowance would be warranted, irrespective of which approach was adopted. Of course, this reflects the fact that such returns are typically not really returns on wealth at all, but rather returns to the owner-manager's own labour inputs. 34 The effect on returns on genuine capital investments would be less marked although still substantial (see the online Appendix for an example).
Under both the Mirrlees and CIT approaches, there are several reforms that should be made to the tax treatment of pensions. In broad terms, pensions are currently exempt from income tax 'on the way in' (i.e. when contributions are made), and on returns on assets held within a registered pension scheme, but pension income is then taxed 'on the way out' (i.e. when it is withdrawn). However, the following features should be noted.
(1) Individuals can withdraw up to 25 per cent of their pension pot as a tax-free lump sum. This exemption can only be justified to the extent that it provides a necessary and effective incentive to save in a pension; there are likely better ways to achieve this objective.
(2) There are no NICs on withdrawals, and only some forms of contributions are subject to NICs.
This means that none of the returns to pensions saving are subject to NICs. The Mirrlees and CIT 33 For a full history of the Schedule A charge to income tax, see Chick (2020). 34 Advani and Summers, 2020a.
approaches differ in whether pension savings should be subject to NICs 'on the way in' or 'on the way out', 35 but neither approach justifies the current treatment.
(3) Any funds that have not been withdrawn at death can be passed on without IHT and (if the deceased was under 75) without the beneficiary being liable to income tax when they subsequently withdraw the funds. Again, neither approach justifies these choices.
Correcting these features would ensure that the investment element of pension savings (as well as the earnings element) was appropriately taxed. F I G U R E 1 (a) Normal rate of return and inflation rate 1970-2020, and (b) the difference between them.
Note: The date axis in both panels refers to the end of the tax year (e.g. 2020 is the tax year 2019-20). Inflation is measured by growth in the consumer prices index (CPI) over the 12 months to the end of the tax year. From 1989 onwards, this is based on the Office for National Statistics (ONS) official index; for earlier years, it is based on a modelled ONS historical CPI estimate. The normal rate of return is measured by the interest rate on medium-maturity government bonds available at the start of the tax year, following Mirrlees et al. (2011, Chapter 13). Source: Author's calculations using data from the ONS and the Bank of England.
Several details of our existing CGT also cannot be justified on any principled basis. Every paper and report that has considered the point recommends that there should be no forgiveness of CGT on death. 36 There is not complete consensus on how this reform should be implemented: transfers on death could be treated as a disposal event, such that CGT would be payable immediately; alternatively, transfers to heirs could be on a 'no gain, no loss' basis, as applies for spousal transfers at present. 37 But it is clear that the gain accruing before death should be brought within CGT one way or other. Another anomalous aspect of CGT is the separate tax-free allowance given in addition to the personal allowance for income tax; it follows from the aim of aligning tax rates on income and gains that these allowances should be merged, or at least that any separate allowance should be restricted to an administrative de minimis. 38 The Mirrlees Review diverges from the CIT approach in its first-best proposal for taxing inheritances and gifts. Under a CIT approach these receipts would be aggregated with the other income of the donee, whereas the Mirrlees approach provisionally recommends a separate lifetime receipts tax. However, neither approach provides any justification for the anomalies under the existing IHT. It is consistent with both approaches that if IHT is to be retained (for example, because donee-based alternatives are regarded as administratively or politically infeasible), the tax base should be expanded significantly by abolishing reliefs for agricultural and business property and imposing the tax on all lifetime gifts over a limited annual allowance. The seven-year rule would therefore be abolished. 39 These and other related reforms were recently recommended in a report by the All-Party Parliamentary Group for Inheritance & Intergenerational Fairness (2020); 40 they are also reflected in the second-best option recommended by Mirrlees et al. (2011, Chapter 15).
The Mirrlees Review firmly recommended the abolition of SDLT and other stamp duties. 41 There are no grounds for retaining these taxes under a CIT approach either. Indeed, it is hard to think of any economic or philosophical rationale for using arms-length transactions as a tax base. However, there are some practical and distributional reasons why it would be unwise to abolish SDLT without making associated reforms to the taxation of residential property. The existing tax does at least have the merit of being easy and cheap to administer. The design of SDLT also ensures that it has a very low 'tax gap' 42 and there are limited opportunities for planning, although there remains a strong incentive to avoid SDLT by transferring ownership of rental properties via shares in a holding company.
Consequently, if SDLT was abolished, there would be a very strong case for replacing it with a progressive annual property tax. The obstacles to reforming council tax to serve this purpose are purely political. It is impossible to mount any plausible defence of the reliance on 1991 valuations. There is broad consensus that the rate structure should be overhauled so that it is no longer regressive. Mirrlees et al. (2011, Chapter 16) recommended a single flat rate, initially set at 0.6 per cent of the property value (which is lower than would be required to achieve equivalence with VAT). The Resolution Foundation, amongst others, has proposed that rates should rise progressively. 43  37 A disadvantage of the latter approach is that it would result in a distortion between lifetime gifts, which are generally charged to CGT immediately on disposal, unless gift holdover relief is extended. 38 The Office of Tax Simplification recently recommended reducing the annual exempt amount for CGT from its current level of £12,300 to between £2,000 and £4,000 (Office of Tax Simplification, 2020). 39 The comprehensive taxation of lifetime gifts on the donor would mark a return to the capital transfer tax that operated from 1974 to 1986; see further Tiley (2007). 40 These proposals included the abolition of the nil rate band on lifetime gifts, and abolition of almost all reliefs other than the spouse exemption and charitable exemption. 41 Mirrlees et al., 2011, Chapter 15. 42 Troup, Barnett and Bullock, 2020. 43 Corlett and Gardner, 2018. on the most expensive properties, 44 in practice there is not much divergence here. The implementation of these reforms could now also look to ATED as a suitable legislative and administrative model. 45 In summary, many of the same reforms to existing taxes on wealth are merited under both the Mirrlees and CIT approaches, because our current system contains a raft of anomalies that have no principled justification. In other words, despite their differences, the two reform packages outlined above have a lot more in common with one another than they do with the design of our existing taxes on wealth. 46 For an illustration, see the online Appendix, which compares all three approaches in application to a simple example. It is hard to avoid the conclusion that the reasons for the present failure to reform existing taxes on wealth have more to do with politics than with economic principle. 47

THE ROLE OF A WEALTH TAX
A new broad-based tax on the ownership of wealth could operate either instead of or as well as the various reforms canvassed in the preceding section. Below, I evaluate each of these options in turn. In doing so, it is important to keep in mind two key design choices affecting the role that a wealth tax could play: A wealth tax could be one-off or recurrent. Typically, recurrent wealth taxes have been levied on an annual basis, although in principle they could recur over a longer or shorter interval. It is important to emphasise that the distinction between a one-off wealth tax and an annual wealth tax concerns the recurrence (or otherwise) of assessment, rather than the process of administration and collection. It would be possible to have a one-off assessment of wealth where the resulting tax liabilities were collected in instalments over several years and may still be due on an annual basis. The point is that the amount of tax that an individual would pay would depend on their wealth at the time of the initial assessment rather than being re-assessed each year. 48 The West German capital levy that operated after World War II was of this type. 49 A wealth tax could also be comprehensive or partial. This distinction refers to the scope of the tax base and the rates applied to different types of asset. A comprehensive wealth tax is one that captures all forms of wealth, and taxes them at a common rate. The Wealth Tax Commission recommended that 'wealth' should be defined to include all forms of private property, net of all private debts. 50 This would encompass, for example, main homes, pension rights and businesses, as well as financial assets, second homes and other land, intellectual property, legal claims and other intangibles.
There are two respects in which a wealth tax may be partial. First, some types of asset may be exempted from the tax base altogether; second, different types of chargeable asset may be charged at different rates. Such differential rates could stem either from differences in the headline rates applied to different types of asset, or (more likely) from differences in effective rates that result from valuation discounts or other reliefs. The Wealth Tax Commission argued strongly against a partial tax base on the grounds that it could introduce significant scope for avoidance and would undermine the horizontal 44 Boadway et al., 2010. 45 ATED could relatively easily be converted into a progressive replacement for council tax, simply by removing the restriction to properties owned via companies and relief for rented properties. 46 This is especially true under current economic conditions, where both the inflation rate and normal rate of return are low. See further the online Appendix, which compares the three approaches in application to a simple example. 47 For further discussion of the politics of taxing wealth, see Clark et al. (2020) and Perrett (2021). 48 There could also be an intermediate design involving a one-off initial assessment but an asymmetrical adjustment, operated as a relief, where an individual's wealth subsequently fell as the result of a decline in the value of their assets. This approach may avoid some of the strong objections to one-off assessment with annual collection on grounds of ability to pay. 49 O'Donovan, 2021. 50 Advani, Chamberlain and Summers, 2020. and vertical equity of the tax. 51 Nevertheless, it must be acknowledged that no actually implemented wealth tax has ever achieved a fully comprehensive tax base.

Instead of other reforms
Suppose that a government wished to implement the reforms to existing taxes on wealth recommended by the Mirrlees Review or a CIT approach, but was not able to do so -perhaps because those reforms appeared infeasible, either administratively or politically. Could a wealth tax provide an alternative? Three steps are needed to answer this question. First, we must clarify the objectives of the desired reforms to existing taxes on wealth. Second, could a wealth tax be designed in a way that would achieve those objectives? 52 If the answer is no -or even worse, if it would be antithetical to those objectives -then a wealth tax is not really an alternative at all, but rather would be doing something different. Third, to the extent that a wealth tax could advance similar objectives, would introducing this new tax be any more feasible -administratively and/or politically -than the reforms to existing taxes that it would stand in for? 53 4.1.1 Taxes on returns on wealth Reforms to existing taxes on returns on wealth (income tax and CGT) should aim at two key objectives.
The first is to achieve neutrality of tax treatment across asset classes and compared with labour income (source neutrality). This objective is shared by both the Mirrlees and CIT approaches. The second objective is to achieve neutrality in the timing of consumption, or in other words to ensure that the tax system does not distort individual's choices about whether to consume now or later (timing neutrality). This is a key objective under the Mirrlees approach only. A wealth tax could go some way to improving source neutrality with respect to labour income by reducing the net-of-tax return on wealth; however, it would not do this as effectively as alignment of income tax and CGT rates on all sources of income and gains, because it would still leave some excess returns on wealth relatively undertaxed. A comprehensive wealth tax would not assist source neutrality across asset classes; it would merely add another layer of taxation on top of existing distortions. A partial wealth tax that focused on assets that are currently undertaxed (such as housing) could potentially rebalance tax treatment across asset classes. However, it is unclear why one would want to use a wealth tax for this purpose; for example, the administrative and political hurdles of implementing a tax on net housing wealth do not appear any easier than imposing CGT (and/or a charge on imputed rent) on main homes.
A wealth tax would worsen timing neutrality. The existing tax treatment of savings and investment income and gains already incentivises individuals to consume now rather than later and hence to undersave, by levying income tax and CGT on the normal 'risk-free' rate of return. As Adam and Miller (2021) point out, a recurring wealth tax would make this problem worse. A wealth tax is insensitive to the actual rate of return, and so effectively (and broadly speaking) taxes the normal rate of return but not the excess returns above that, thereby further distorting the decision whether to consume now or later. In this respect, a recurring wealth tax is antithetical to the objectives of the Mirrlees reforms to existing taxes on wealth.
It is also very doubtful whether implementing a new wealth tax would be any easieradministratively or politically -than reforming income tax and CGT. Unlike a wealth tax, the 51  reforms recommended by the Mirrlees Review could be implemented without requiring any major new administrative processes. There is widespread public support for closer alignment of the tax rates on returns from wealth compared with income from work, 54 and -unlike a wealth tax -reforms in this direction are already canvassed by politicians across the political spectrum. It is hard to imagine conditions under which a wealth tax would be politically feasible but raising tax rates on capital gains -and even dividends -would not be.

Taxes on transfers of wealth
As the Meade Report identified, 55 a wealth tax has some affinity with IHT: 'a tax on the transfer of inherited wealth as it passes down from parent to child is in effect a periodic charge, levied once a generation, on the holding of such wealth. In so far as this is the case, it may be regarded […] like an annual wealth tax, but one which is levied once a generation instead of once a year.' The notion of IHT as a 'once-per-generation' wealth tax may now be out of date as there is also a good deal of generation-skipping. 56 But even as a tax once per generation, IHT is extremely leaky: it grants reliefs over large parts of the tax base; many lifetime gifts are not covered; and the wealthiest also benefit from other tax planning opportunities. The reforms to IHT proposed by the Mirrlees Review 57 sought to advance equality of opportunity and achieve redistribution at relatively low efficiency cost. 58 A wealth tax seems a poor substitute for IHT reform in terms of equality of opportunity, because unlike a tax on the transfer of wealth it does not discriminate based on source: 59 a wealth tax falls equally on someone whose wealth results from their own effort and skill (e.g. the archetypal rags-to-riches entrepreneur) as it does on someone whose wealth is entirely derived from inheritances. At the same overall level of redistribution across the entire population, a wealth tax is less well targeted at redistributing undeserved wealth than an inheritance or gift tax.
It is difficult to assess whether or not a wealth tax could redistribute more efficiently than a reformed IHT. Under the current IHT, despite a headline rate of 40 per cent, the effective average rate peaks (on average) at around 20 per cent and actually declines to 10 per cent for estates valued at over £10 million. 60 This is largely due to the use of agricultural and business property reliefs, which disproportionately benefit the wealthiest. A wealth tax may therefore be able to do a better job of redistribution than the current IHT, but only if it manages to avoid the political pressures for exemptions and reliefs that have rendered IHT ineffective. Is there any reason to think that a wealth tax would fare better in this regard?
The annual wealth taxes in France, Spain and Sweden all exempted business assets altogether (provided certain conditions were met), and several other countries have granted tax preferences to businesses in the form of special valuation rules or part-exclusions. 61 Perret (2021) reports that artwork and antiques have also commonly been exempted. The UK's last attempt to introduce a wealth tax, in 1974, also fell foul of vested interests. Glennerster (2012) reports that Treasury officials got cold feet shortly after receiving representations from the National Farmers' Union, various bodies representing 54 Savanta ComRes, 2021. 55 Meade et al., 1978, p. 317. 56 Bourquin et al., 2020. 57 Mirrlees et al. 2011, Chapter 15. 58 This invocation of equality of opportunity represents an ad hoc departure from the optimal tax theory approach that underlies the other key recommendations of the Mirrlees Review. Furthermore, strict adherence to the aim of equality of opportunity would tend to imply a 100 per cent IHT (Haslett, 1986). 59 For the same reason, a wealth tax would not achieve horizontal equity across different forms of receipt, as aimed at under a CIT approach. 60 Office of Tax Simplification, 2018. 61 For example, Germany, Norway, Luxembourg and Ireland, see further Perret (2021). the owners of country houses, small businesses, and the City of London. These experiences cast some doubt on the notion that introducing an effective wealth tax would be any easier than reforming our existing IHT.

As well as other reforms
Suppose that the reforms to existing taxes on wealth recommended either by the Mirrlees Review or under a CIT were implemented in full. Would there be anything left for a wealth tax to do? Put another way, is there any justification for using wealth as a tax base, as distinct from taxing the income from wealth, the consumption that can be obtained from spending and enjoying the use of wealth, and transfers of wealth across generations? Here I focus exclusively on the 'in principle' case -setting aside questions of political feasibility -because, for present purposes, we are already positing a rather unrealistic world in which our existing taxes on wealth have been perfectly reformed. In reality, of course, there is often limited political and administrative bandwidth for major tax reforms, in which case pursuing a wealth tax could come at the expense of other objectives.

Correct the past
The Mirrlees Review sought to design an optimal tax system starting now (i.e. assuming that we are happy with current distribution of wealth, which will represent individuals' starting points under the new system). The policies recommended by Mirrlees et al. (2011, Chapter 20) do not make any attempt to correct for past distortions in the distribution of wealth created by the old tax system. Some individuals may currently own more wealth than they 'should' have (under an optimal tax system) because -for example -their capital gains were undertaxed or they escaped paying any tax on inheritances or gifts received in the past. Moreover, proposals for reforming existing taxes on wealth also typically leave unaddressed other concerns about the processes by which past wealth has been accumulated (for example, as the result of market failures or historic injustices). Consequently, even once we have managed to design and implement the optimal tax system going forwards, we might still want to implement measures to correct the past. This would be like pushing the reset button before booting up the new system. There are undoubtedly a multitude of things that have gone wrong in the past, which we may wish to correct for. Ideally, we would rectify each of these missteps directly, but that is clearly infeasible in many cases. Even if we could agree in principle about which missteps should be corrected, we lack the information required to make such corrections at an individual level. In these circumstances, the tax system might provide the most appropriate means for rectifying past wrongs. Indeed, even the libertarian philosopher Robert Nozick recognised an inprinciple case for using the tax system in this way. 62 Whether a wealth tax is an appropriate tax policy tool for effecting corrections to the past depends primarily on whether it is a fair assumption that those with currently high levels of wealth are more likely to have been the beneficiaries of past inefficiencies or injustices than those with currently low levels of wealth. One would also need to be assured that current wealth provides a sufficiently strong predictor of these inefficiencies and injustices that this approach would be justified, despite the fact that it would also inappropriately tax individuals whose wealth had been accumulated in a manner that did not warrant such a correction (for example, the archetypal rags-to-riches entrepreneur).
If one wished to use a wealth tax to correct the past (at least in part), then this would have various implications for how the tax should be designed. First, if one was only concerned with past inefficiencies and injustices, a one-off wealth tax would be more appropriate than a recurrent levy, because it exclusively affects 'old capital'. In other words, 100 per cent of the tax base would be wealth accumulated under the old system. An annual wealth tax would be less well targeted for this purpose, as over time it would capture more and more new capital (for example, savings out of current labour income) that had been accumulated under the newly reformed system. Second, as a tool for correcting the past, a wealth tax ought to have a comprehensive base, unless there are good reasons to think that wealth held in particular forms is more likely to have benefited from the missteps that the tax is seeking to correct. It is true that some past problems of undertaxation have been focused on particular types of asset (for example, main residences that have been exempt from CGT). However, for the purpose of correcting the past, it would not follow that a wealth tax should target specifically these assets in the hands of their current owners, because the proceeds of previously undertaxed gains (or other tax inefficiencies) may have since been invested into other forms. 63 Consequently, it would not be possible to target past inefficiencies by restricting a wealth tax only to the asset types that have been undertaxed in the past.

An ongoing corrective
What if it is the case not only that wealth has become too unequally distributed as a result of past inefficiencies and injustices (as outlined above), but also that these or other unjustified processes are continuing to drive an overconcentration of wealth? To the extent that such processes result from imperfections in our existing taxes on wealth, these may be corrected by implementing the reforms recommended under the Mirrlees Review or a CIT approach (i.e. they do not call for the additional introduction of a wealth tax). However, even if such reforms were implemented in full, there may still be a residual concern that other processes will continue to operate -across various aspects of our economy and society -which tend towards unjustified sources of wealth accumulation. Could a wealth tax play a role in 'actively managing' such accumulations? Various specific concerns may be raised, including: market failures arising from excessive and uncontrolled market power; the impact of expansionary monetary policy (low interest rates and quantitative easing) on asset prices; superstar effects and impact of automation on the distribution of incomes; rent extraction through political influence on policymaking; or dynastic wealth accumulation through inheritances and gifts (which would not be reduced to zero even under proposed reforms to IHT). All of these examples are contestable, in terms of both the magnitude of their impact on the distribution of wealth and the extent to which wealth that is accumulated as a result of these processes is deserved or otherwise justified. Others may cite different processes giving cause for concern.
For present purposes, the question is whether, if these are valid concerns, a wealth tax could provide a suitable ongoing corrective. The most obvious point to make in this regard is that a wealth tax is, by design, insensitive to the source of wealth. For example, it treats wealthy monopolists and heirs to dynastic fortunes the same as competitive entrepreneurs. Ideally, one would use more targeted public policies to address ongoing concerns about the processes of wealth accumulation. Nevertheless, some people hold the view that in cases of extreme wealth it is reasonable to presume that something must have gone wrong in the processes by which it was accumulated (i.e. that extreme wealth is itself indicative of a failure of market mechanisms or government intervention). 64 This line of argument appears to lie behind the preference of some economists for a wealth tax that would target the very wealthiest individuals in particular. 65 The Wealth Tax Commission did not recommend an annual wealth tax applying to ordinary levels of wealth but left open that a tax with a very high threshold might be justified if the aim was to reduce extreme wealth inequality. 66 As Advani and Summers (2021) highlight, a wealth tax can have a more immediate effect on the distribution of wealth than other reforms because it does not rely on waiting for realisations of assets (as with CGT) or deaths (as with IHT) to occur. Moreover, unlike the taxation of capital income and gains, a wealth tax can, in principle, be set at a rate exceeding the rate of return on wealth, so can actively erode accumulations of wealth if desired. Both of these features make a wealth tax potentially a more powerful tool for addressing extreme wealth inequality than reforms to existing taxes on wealth.

4.2.3
Wealth affects welfare The amount of wealth that an individual has might provide information about their welfare or their ability to pay tax that cannot be gleaned simply from looking at how much income they receive, or how much they consume, even over a lifetime. 67 The notion of ability to pay has a strong pedigree in tax policymaking; it was the first of Adam Smith's four canons of taxation in The Wealth of Nations.
In the Meade Report, the idea of ability to pay was referred to as 'taxable capacity' and it formed the organising principle on which the report's policy recommendations were based. 68 The authors' view that wealth indicated taxable capacity distinct from consumption motivated their conclusion that the tax system ought to include both consumption and wealth as tax bases. In turn, this led the report to recommend the introduction of an annual wealth tax in addition to a consumption tax. 69 Although the view that wealth indicates taxable capacity distinct from consumption has some intuitive appeal, the notion of taxable capacity is not at all straightforward to define. The authors of the Meade Report acknowledged this. To illustrate the difficulty, they considered an example involving Mr Smith and Mr Brown. 70 In the example, both men start with equal wealth and income, but Mr Smith chooses to save his income whereas Mr Brown spends it. Mr Smith now has more wealth than Mr Brown, but the puzzle is which of them has more taxable capacity. The authors of the Meade Report put it in the following way. 'Both have the same power of command over resources. Can the fact that Mr Smith chooses to use this power one way and Mr Brown in another way affect their present capacities to pay tax?' There are several parts to this puzzle. Before addressing the one that is of direct relevance for present purposes, some ground-clearing is required. First, the example only has force where Mr Brown spends in such a way that dissipates his wealth (e.g. on holidays or meals out). If instead he spends on durable assets -even ones that also provide him with consumption benefits, such as a car, a yacht or artwork -then his wealth will not have decreased as a result of the purchase, except to the extent that the asset depreciates over time and there is an opportunity cost of the money used (which could otherwise have been invested). 71 At certain high levels of wealth, it seems likely that a large proportion of spending 65 For example, in the US context, Saez and Zucman (2019) recommend a wealth tax that would only start at a threshold of $50 million, with the highest rates applying to individuals with fortunes exceeding $1 billion. 66 Advani et al., 2020. 67 For now I use the terms 'welfare' and 'ability to pay' interchangeably; below, I explain the different ways of thinking about tax policy that each of these reflect. 68 Meade et al., 1978, Chapter 3. 69 Meade et al., 1978, Chapters 10 and 16. 70 Meade et al., 1978, p. 34. 71 Although this opportunity cost does not detract from Mr Brown's nominal wealth, it would over time cause his wealth to fall behind Mr Smith's, on the assumption that the latter received the normal rate of return on his savings whereas the durable assets purchased by Mr Brown did not. takes forms in which wealth is preserved or increased rather than dissipated; if so, the example may not take us very far in the real world of the super-rich. Second, although this example primes us with the intuition that Mr Smith's choice to save rather than spend should not leave him liable to additional tax, it is reasonable to ask how widespread such examples are in real life. If it turns out that there is not much variation in the savings decisions of individuals at similar initial levels of wealth and income, then there may not be many cases in which the horizontal inequity posited by the example actually bites. However, it is not sufficient to point to a lack of variation in savings rates (by initial wealth and income) under our current system as decisive, if future savings behaviours would be substantially affected by reforms that increase the effective tax rate on savings. Still, the available evidence indicates that taxes only have a very weak effect on savings rates in practice. 72 Leaving these two issues aside, the answer to this puzzle depends mainly on whether one looks at the information as a snapshot or over the life cycle. Viewed as a snapshot, it appears that if ability to pay depends on income, then Mr Brown and Mr Smith should pay the same tax; if it depends on wealth, then Mr Smith should pay more tax than Mr Brown; and if it depends on consumption, then Mr Brown should pay more tax than Mr Smith. However, the authors of the Meade Report pointed out that when considered from a life-cycle perspective, these positions may even out, on the assumption that Mr Smith eventually consumes his wealth. The residual case for thinking that wealth indicates taxable capacity is if the wealth conferred some additional benefit on Mr Smith besides his eventual consumption of it. The Meade Report concluded that it did, on the basis that wealth also provided the holder with 'security, independence, influence and power'. 73 The Mirrlees Review rejected the Meade Report's conclusions about the distinctive benefits of holding wealth. To understand the root of this divergence, it is important to appreciate the differences in methodology between these two landmark reviews. Contributing to the Mirrlees Review, Banks and Diamond (2010, p. 610) considered but explicitly rejected 'the Meade Report view […] that taxes should relate monotonically to some measure of taxable capacity'. They argued that: '[i]n addition to finding taxable capacity not well-enough measurable and not sufficiently uniformly evaluated to be usable for this purpose, we also do not see an underlying normative basis for reaching the conclusion that taxes should be related to taxable capacity without full consideration of the equilibrium consequences [for social welfare] of following such an approach.' Instead, they concluded that 'the starting place for thinking about taxation should be the impact of taxes on the utilities of people in the economy', reflecting the economic approach of optimal tax theory.
Under the optimal tax approach adopted in the Mirrlees Review, the goal of tax policy is to maximise social welfare, defined as the weighted sum of individuals' welfare. In turn, it was assumed that an individual's welfare depends exclusively on the amount they consume (increases welfare) and how much they work (decreases welfare). 74 However, one might reasonably question this assumption. What if, as Meade posited, wealth provides additional welfare-enhancing benefits beyond the consumption that it funds (even over a lifetime) -benefits such as security, independence, influence and power, for example? If so, these benefits are not accounted for by the definition of welfare that drives the policy recommendations of Mirrlees -including the review's rejection of the case for an annual wealth tax. 75 It will be apparent from this chain of argument that we are essentially back to the debate about taxable capacity with which the Meade Report was concerned, except now under the guise of defining welfare. For this reason, the authors of the Mirrlees Review did not succeed in sidestepping the difficulty of defining taxable capacity; they just reframed the same problem; for further details, see 72 Advani and Tarrant, 2021. 73 Meade et al. 1978, p. 351. 74 Banks and Diamond, 2010. 75 Mirrlees et al. 2011, Chapter 15. Kay (2010. 76 If, as the Meade Report concluded, wealth is relevant to taxable capacity (distinct from consumption), this is surely because it is relevant to welfare (distinct from consumption). The issue must be resolved either way. Whereas the Meade Report argued for its conclusion that wealth was relevant to taxable capacity, the Mirrlees Review merely stipulated that wealth was not relevant to welfare, in the way that it defined individual welfare in the social welfare function.
In this respect, the policy question of whether an annual wealth tax can be justified in addition to comprehensive consumption tax treatment seems to turn on whether wealth confers benefits on individuals beyond the consumption that it funds: and this is so regardless of whether the puzzle is framed in terms of taxable capacity or welfare. The difficulty is that there is currently not much empirical evidence on this question, although there is some suggestion in the empirical literature on estate taxes that 'wealth as welfare' could explain (part of) the observed desire to retain control of wealth whilst alive. 77 Given its potential importance in shaping the future direction of optimal tax theory -on which the tax policy prescriptions of economists depend -this would be a useful area for further research.
Notwithstanding this evidential uncertainty, if a wealth tax were justified on the basis that it affected taxable capacity or welfare distinct from consumption, this again has various implications for the design of the tax. First, it ought to be recurring rather than one-off, on the assumption that the added benefits of wealth (over consumption) would persist for those continuing to hold wealth. Second, the tax should be comprehensive unless there is reason to believe that particular types of asset are more likely to confer the relevant benefits (such as security, independence, influence and power) than others. Again (as with correcting the past), it seems plausible that some particular types of asset could be singled out, at least from a static perspective: for example, owning a newspaper business may yield more influence and power than merely owning houses. However, from a dynamic perspective, it may be unwise to differentiate tax treatment on this basis.
Finally, in relation to the CIT approach to reforming existing taxes on wealth, one might reasonably ask about taxable capacity from another perspective: does wealth indicate ability to pay distinct from income? If we take taxable income from a snapshot (annual) perspective, it seems quite obvious that wealth provides additional information about ability to pay -although even this claim is somewhat in tension with the liquidity concerns that are often raised when a taxpayer has high wealth but low income. 78 Over a lifetime, the difference between (comprehensive) income and final wealth is by definition consumption (excluding bequests).

INTERACTIONS WITH EXISTING TAXES
If a government decided to introduce a new wealth tax, it would need to fit within a system that already imposes various taxes on wealth. Here I focus on two specific questions about how a wealth tax should interact with existing taxes. First, I address concerns about double taxation. When -if at all -should an individual who has paid the wealth tax be able to obtain a tax credit (or relief) against payment of other taxes on wealth? 79 Second, what are the prospects for using a wealth tax as an alternative minimum tax, specifically targeting individuals who pay relatively little as a proportion of their wealth under the current system? 76 As Kay (2010) put it when commenting on Banks and Diamond (2010): 'I cannot imagine that it would be easier to secure agreement on the definition of utilities [welfare] than on the definition of taxable capacities [ability to pay]: indeed, it is likely that the two definitions would be very similar. I believe it is difficult to argue that it is possible to define utilities but not to define taxable capacities.' 77 Kopczuk, 200777 Kopczuk, , 2013. 78 Loutzenhiser and Mann, 2021. 79 Besides avoiding double taxation, tax reliefs are often advocated to achieve specific policy objectives. For example, there is a question of whether Gift Aid Relief for charitable donations (which currently applies for income tax) should be extended to a wealth tax. However, these types of relief (known as 'tax expenditures') raise different considerations: see NAO (2020) for a recent overview.

Double taxation
In principle, whether any measures to avoid double taxation are required must depend on whether the wealth tax serves the same purpose as another tax; to the extent that it does, there may be a case for allowing relief if the purpose is already fully served by the other tax. However, if the taxes serve different purposes, then there should be no objection to cumulating them (i.e. not allowing any relief), even if both taxes are assessed in respect of the same asset. To take an easy example, no one thinks that income tax should be deductible against tobacco duty, even where an individual has bought cigarettes out of their taxed income. The reason is that these two taxes clearly serve different purposes. This remains the case even where the taxes are due in respect of the thing (here, the cigarettes). Cigarette manufacturers are liable for both tobacco duty and VAT, again because they serve different purposes: VAT is a broad-based tax on consumption whereas tobacco duty is specifically aimed at reducing smoking.
Nevertheless, in the context of taxes on wealth in particular, the issue of double taxation is often raised. For example, IHT is sometimes said to involve double taxation on the assumption that the assets in the estate will have been purchased out of taxed income or gains. 80 More often, it is claimed that charging both IHT and CGT on death would amount to double taxation, and further that that this is a reason for retaining CGT forgiveness on death. 81 Whether this charge of double taxation is valid depends on whether IHT and CGT serve the same purpose or not. It does not depend on whether both taxes have been charged in respect of the same asset.
To determine whether a wealth tax should be deductible against other taxes, we therefore need to know its purpose, which takes us back to Section 4. If a wealth tax is justified to correct the past, or because wealth affects taxable capacity or welfare distinct from income or consumption, then the tax is serving a different purpose from any of the existing taxes on wealth and should be levied on top, with no reliefs. By contrast, if a wealth tax is justified only in place of reforming existing taxes on wealth, and so intended to serve the same purpose by another means, then to the extent that existing taxes serve the same purpose already, there may be some case for allowing reliefs.
However, even where reliefs may be justified in principle, it is doubtful whether they would be a good idea in practice, due to the additional legislative complexity that they would generate. There are almost no examples of wealth tax payments generating a relief or tax credit against other domestic taxes in any of the countries that operate a wealth tax. Prior to the abolition of its wealth tax, France allowed a deduction from the IHT base if wealth tax was due but not yet paid at the date of death; however, this represents a very limited timing provision only. Within the UK, there is a limited deduction for CGT in certain circumstances where the donor died within seven years of making a lifetime gift, which is therefore included in their IHT estate. There is also some deduction against IHT for offshore income gains that are crystallised on death. A handful of other obscure reliefs operate between existing taxes on wealth but experience seems to be that they can operate in arbitrary ways.
There would be a much stronger case for allowing foreign tax credit relief where individuals who are liable to the UK wealth tax find themselves also liable to another country's wealth tax in respect of the same asset. 82 There is already a network of double tax treaties that exist bilaterally between the countries that currently have a wealth tax, to deal with this type of issue. If a wealth tax was introduced in UK, then it would be sensible to negotiate equivalent double tax treaties with these countries. The existing network of treaties indicates that this ought to be possible in principle, although it would require considerable diplomatic and administrative effort. 80 Lee, 2007. 81 Both the Mirrlees Review (2011, Chapter 15) and the Office of Tax Simplification (2019) have appropriately debunked this idea. 82 This would be most likely to occur in relation to non-residents holding UK assets that were liable to the wealth tax on the basis of the situs of the asset (such as UK land), where they were also liable to a wealth tax on their worldwide assets in their country of residency.

Alternative minimum tax
For some, the primary function of a wealth tax would be to raise revenue from those who have high wealth but who pay relatively little tax as a proportion of that wealth (or as a proportion of their comprehensive income, including accrued but not yet realised gains). One reason for this state of affairs could be if the individual had structured their remuneration in such a way as to reduce their effective tax rate on income and gains to a very low level, for example by making use of capital gains and dividends in place of taking a salary taxed as earnings. Another reason, however, may be because the individual has relatively little taxable income or gains as a proportion of their wealth, either because they own mostly low-yielding assets or they are accumulating gains (for example, by retaining profits in a company) that have not yet been realised. One way of addressing these concerns could be to use a wealth tax as an alternative minimum tax (AMT), effectively serving as a backstop for income tax and CGT. The US already operates an AMT, although it applies only to taxable income, effectively setting a floor on the tax that can be obtained by combining different income tax reliefs. In the UK, Advani and Summers (2020b) have gone further, suggesting an AMT that combines both taxable income and gains, to prevent individuals from excessively reducing their effective tax rate by repackaging income as gains. An AMT based on wealth would be a yet further extension of this idea, targeted at those with high wealth but relatively little taxable income and gains.
In the simplest terms, all individuals within the territorial scope of the wealth-based AMT would be required to pay a minimum total amount in tax, calculated as a percentage of their wealth. If the individual already paid more than this amount in other taxes, then there would be no additional tax to pay. In effect, this is equivalent to giving relief against the wealth tax for a wide range of other taxes. Clearly, this would significantly reduce the revenue that the wealth tax would raise. However, in this form, the purpose of the wealth tax would be primarily as a backstop to protect the rest of the tax base, rather than as an independent tax. Those who already pay substantial sums in tax under the current system -for example, individuals without substantial wealth who receive most of their income from employment -would be entirely unaffected by a wealth-based AMT because their existing tax payments would easily exceed the minimum threshold.
A wealth-based AMT is potentially appealing in that it would recognise some role for wealth within the notion of ability to pay, without affecting the majority of individuals for whom income or consumption is likely to be a more appropriate measure. 83 It could also serve as an anti-avoidance (or anti-tax planning) device. However, such a tax may be politically challenging because its effect would be to flip conventional concerns about the liquidity of 'asset-rich cash-poor' taxpayers on their head. 84 A wealth-based AMT would effectively target individuals who reported small (taxable) flows relative to their stock of wealth. This approach is the opposite of many countries with a wealth tax (including Spain and France), which operated a cap on wealth tax liability for taxpayers with low income. 85 The public perception that individuals with low income have low taxable capacity, irrespective of their wealth, may be an insurmountable obstacle to public acceptability of a wealth-based AMT.
A wealth-based AMT would also face several administrative challenges. First, there would inevitably be calls to expand the range of tax payments that would count in assessing whether the minimum payment (as a percentage of wealth) had been met. Aside from income tax and CGT, the scope could potentially be extended to include (for example) the corporation tax or employer NICs paid by firms controlled by the individual, or SDLT, or IHT on lifetime gifts into trust. However, any such extensions would rapidly generate huge complexity in determining how to allocate payments between individuals, their timing, and so on. Second, a wealth-based AMT would require assessment 83 This approach fits with the observation of Kay (2010) that 'tax liabilities are based, not on-probably unobservable-taxable capacity, but on variables which we believe to be correlated with taxable capacity'. 84 For evidence on this issue in UK public attitudes, see Rowlingson, Sood and Tu (2021). 85 See further Loutzenhiser and Mann (2021). of a very large number of taxpayers as it would be difficult to determine ex ante who might be liable to the charge, despite the fact that in most cases no tax would be due. Given the cost of valuations for both taxpayers and the revenue, 86 such an approach is unlikely to be administratively viable.

CONCLUSION
The UK's existing taxes on wealth are a mess. There is more than one possible direction for reform: I have considered both the Mirrlees approach, which reflects the idea of consumption as the ideal tax base, and an alternative approach based on the idea of (comprehensive) income as the tax base. However, it is important to emphasise that despite these different directions, there are nevertheless a long list of reforms to existing taxes on wealth that would be merited under either approach, or indeed any coherent approach. This reflects the fact that many features of the UK's current approach to taxing wealth cannot be defended on any rational basis. In this sense, the area is ripe for reform. Against this backdrop, why introduce a wealth tax on top? I have suggested three possible reasons. First, most proposals to reform existing taxes on wealth are wholly prospective: they would correct the system going forward but would do nothing to correct the mistakes of the past. A wealth tax could potentially play this role; however, it would be second-best compared with policies that addressed these mistakes directly. Second, a wealth tax might serve as an ongoing corrective to processes of excessive and/or unjustified wealth accumulation. Third, there remains a debate about whether wealth provides distinctive benefits beyond the consumption that it funds. The Mirrlees Review sidestepped this debate but it requires renewed attention.
Could a wealth tax provide an alternative to reforming existing taxes on wealth? In relation to income tax and CGT, the short answer is no. If a wealth tax is implemented, long overdue reforms to income tax and CGT should still be carried out as well. The position in relation to IHT is more nuanced. A wealth tax cannot perform the same function as IHT in advancing equality of opportunity because it applies to all wealth regardless of its source. However, if reforming IHT proves politically infeasible, a wealth tax could provide a progressive alternative. Council tax will require reform regardless of a wealth tax, although a wealth tax could contribute to a somewhat more progressive system for the taxation of main homes.
In practice, the relevant policy question is not just whether a wealth tax would be preferable to reforming existing taxes on wealth, but whether it would be preferable to the raft of other ways in which a government may be tempted to increase tax revenues. Based on past experience, the most likely may be a rise in the headline rates of one of the 'Big 3' taxes: the basic rate of income tax, NICs or VAT. Another possibility, which was tried after the financial crisis (from 2010 to 2013), could be to increase the top rates of income tax. I have not attempted to compare the distributional impacts and efficiency costs of these options in this paper, but that is an important agenda for further analysis.
The call to reform existing taxes on wealth instead of introducing a wealth tax resonates with the most force when it is accompanied by a specific list of proposals and a firm commitment to enact them. Without these, the objection rings hollow. As the contributors to the Mirrlees Review rightly put it, 'an argument that a better policy is available should only be used as an argument against a particular policy proposal if the available alternative is actively pursued'. 87 Detractors from a wealth tax would do well to heed this injunction.

A C K N O W L E D G E M E N T S
With thanks to Helen Hughson for assistance with Figure 1, and to Arun Advani, David Burgherr, Emma Chamberlain, Nick O'Donovan, Carl Emmerson and Alan McGuiness, and two anonymous reviewers, for very helpful comments. Any errors remain my own. 86 See further Troup et al. (2020). 87 Banks and Diamond, 2010.