Optimal Taxation Theory and the Taxation of
Housing
Report to the
Department of Communities and Local Government
Professor ALAN W. EVANS
Centre for Spatial and Real Estate Research
The University of Reading
1
Abstract
The report begins by setting out the basic principles of optimal tax theory, in particular
the idea that taxes should be imposed so as to distort the economy as little as possible by
minimising welfare losses. We then survey recent research in the field dealing with the
optimal taxation of housing. The most important finding is the under-taxation of owner
occupied housing will distort investment. Owner occupiers are given an incentive to try
buy more expensive homes than they would otherwise occupy, diverting investment away
from productive investment in business capital. The theory demonstrates that the taxation
of owner occupied housing should be at least as high as for any other form of investment,
if not higher, and, especially it should, so far as possible be tenure neutral as between
renters and owner occupiers.
We go on to rehearse recent history in the UK showing that up to the middle of the
century the taxation of housing did indeed try to be tenure neutral. Various changes in the
tax system have occurred since then, but, in sum, it would appear that that system is
probably tenure neutral for younger and poorer households but that older and wealthier
owner occupiers are almost certainly under-taxed. Nevertheless their favoured position
serves as an example which encourages the younger households to save to reach their tax
favoured position.
Finally we turn to the problem of land, a topic which is mentioned but not analysed in the
earlier review and survey. We note that the stringent controls on land use enforced by the
planning system differentiate the UK from other countries. These controls mean that the
price of housing and of land has increased substantially over the past forty years. As a
result each generation holds an asset which increases in value while it is being held but
where the profit is only realised when it is sold on and thus has to be paid by a
succeeding generation. This implicit tax has reduced investment in housing in the UK,
which now has the smallest new homes in Europe. We conclude by pointing out that
planning policies and fiscal policies tend to be contradictory. Planning policies promote
smaller homes. Fiscal policy, through the Council Tax, favours buying large homes. The
policies together have resulted in high land values, and at various times local and central
governments have attempted to obtain a share of the increase in land values. The author
recommends that one part of a solution would be the imposition of VAT on new housing
construction.
2
Optimal Tax Theory: the Basic Analysis
1. Development of the theory of optimal taxation began in the nineteen seventies,
though its origins go back forty years to a 1927 paper by the short lived genius, Frank
Ramsey. The basic premises of the theory are that a government wishes to raise a
given sum through taxation, that taxes distort economic choices, and the question at
issue is how the taxes should be set in order to minimize the distortion which does
occur. An indication of the nature of the loss arising from tax distortion is
demonstrated in the figure on this page which is derived from Atkinson and Stiglitz,
Lectures on Public Economics (1980, Figure 12.1). There is a downward sloping
demand curve and the supply curve is assumed to be horizontal, that is there are no
economies or diseconomies of scale. A tax is imposed raising the supply curve from
DE to AB. The increase in price means that the consumer has to pay more and
consumes less.
2. The welfare loss is represented by the triangle BCE, the consumer surplus on the
goods consumers would have liked to buy but will not because the price is now too
high. Note that from a welfare point of view the revenue derived from the tax, the
rectangle ABCD, is not a welfare loss to society because it is used by government to
provide services which are presumably desired by consumers.
3. It can be seen that the size of the triangle representing the welfare loss is dependent
on the slope of the demand curve, i.e. the elasticity of demand. If demand is inelastic
then the size of the triangle will be small, if demand is highly elastic the triangle will
be large. From this it follows that the welfare loss resulting from taxation will be least
if taxes are imposed on goods for which demand is inelastic.
Price to the
consumer
Demand curve
B
A
Supply curve with tax
Revenue
E
D
Supply curve
(perfectly elastic)
C
O
G
F
Quantity
3
4. But this only deals with the question of efficiency. But as always with welfare
economics one has also to consider the question of equity. And the answer to the
question as to the most equitable way of imposing taxes on commodities depends not
on the price elasticity of demand but on its income elasticity. If the income elasticity
of a good is close to one then as people’s incomes increase their consumption of the
good increases proportionally. Thus a tax on such a good is neither progressive nor
regressive. It follows that the tax structure is most equitable if taxes are imposed on
goods for which the demand is income elastic so that the rich who spend more of their
income on such goods will pay proportionately more of their income in taxes.
5. As it happens the income elasticity of demand for housing is low, about one, which
suggests that it is not, on equity grounds, a candidate for high taxation. On the other
hand the price elasticity of demand is also close to one which suggests that, on
efficiency grounds, it might be a candidate for higher taxation. The analysis so far has
been fairly elementary however, assuming a static economy. But if housing is being
considered the analysis should be more complex. It will need to be dynamic to take
account of the fact that housing is an investment good as well as a consumption good.
It might need to take into account differences in tastes, and the externalities which are
the economic basis for the land use planning system, and the fact that housing is not
just a product but is necessarily built on land, a factor of production which exists but
is not produced. In the survey of work on the optimal taxation of housing we will see
that the analysis has been made dynamic, and differences in tastes have been taken
into account but neither land nor externalities have really been considered
Optimal Taxation and Housing: a Survey
6. A problem with the design of an optimal tax system is that the taxing authorities do
not have information on individual preferences since purchases tend to be
anonymous. So the person who bought three hundred bottles of whisky a year, even
though a potential alcoholic, could not be charged a tax per bottle which was higher
than that charged to someone who bought only one bottle per year. After all, the
alcoholic could buy one bottle at a time or get his friends to buy on his behalf.
7. It can be recognized, however, that housing is a commodity where the total amount
purchased by a household is identifiable, and thus that it could be feasible to
differentially tax housing. In one of these Cremer and Gahravi (1998) explore the idea
of housing subsidies for the less well off. They demonstrate that if the housing
preferences of the wealthier and the poorer households can be distinguished, and if
the latter have a stronger preference for lower quality housing than the former, then
there may be an argument for subsidising the housing consumption of the less well
off.
8. This question of ‘social’ or ‘affordable’ housing appears to have been a side issue,
however. The more obvious problem has been seen to be the way in which the tax
4
systems in many countries appear to favour owner occupation over renting, and, in so
doing, distort the pattern of housing consumption and investment. Certainly a number
of researchers have, over the years, identified a welfare loss arising from the implicit
subsidisation of owner occupied housing. The first of these, Laidler (1969), identified
the welfare loss as being less than half of one per cent, and later studies have
generally also found the welfare loss to be low. More recently, however, it has been
argued by Skinner (1996) that these low estimates result from using static models and
that the dynamic effects are substantial. Using a dynamic model (overlapping
generations with a bequest motive), Skinner estimated the welfare loss to be about 2
per cent.
9. Skinner’s argument is that the low taxation of (owner occupied) housing makes
housing an attractive investment and so its price rises. This delivers a windfall gain to
existing owner occupiers, but, as a capital gain on housing this is largely untaxed. But
‘the loss to future homeowners erodes the government tax base, because a larger
fraction of their saving takes the form of the now more expensive non-taxable
housing. The attenuation of future tax revenue implies that this (lump sum)
intergenerational transfer has real efficiency effects’ (p. 398).
10. In another recent paper on housing taxation Gervais (2002) also uses a dynamic
model (a dynamic general equilibrium life cycle economy populated by
heterogeneous individuals). He argues that the favourable tax burden of owner
occupied housing causes two potential distortions. Firstly there is an incentive for
individuals to own rather than rent. Secondly, if they own there is an incentive to own
larger homes than they would otherwise occupy. Thus the tax system distorts people’s
savings and investment because the imputed rent is not taxed and, in the US,
mortgage interest is tax deductible.
11. Using simulations of the US economy Gervais finds that if imputed rent were taxed at
the same rate as the income from business capital then the stock of business capital
would be greater by 6 per cent and that the stock of housing capital would be 8 per
cent less. His simulations also show that, as one might perhaps expect, more
households would rent, indeed that owner occupation would be reduced by a quarter,
but that, as one might not expect, reducing incentives for owner occupation would
have very small distributional effects.
12. Another argument raised by Gervais is one that might easily be ignored. That is that
the tax system distorts people’s savings and consumption profiles. Young people
cannot buy because they cannot put down a substantial down payment so they save
whilst they are renting. Having saved the down payment and bought the house they
no longer need to save, and as has been identified by Englehardt (1996), their
consumption rapidly expands.
13. Relevant to this is an argument put by Englund (2003). Because of the tax incentives
young people try to buy as soon as possible and this results in many taking positions
where their housing equity is low, only five or ten per cent. This means that falls in
5
house prices can leave them in a position where bankruptcy and foreclosure is very
feasible. In economic terms they are incentivised by the tax system to take on a very
risky, highly leveraged, investment earlier than they otherwise should.
14. To complete this brief survey one should note a paper by Eerola and Maattanen
(2005), so far only published by the Bank of Finland (and therefore not yet peer
reviewed), which also explores the idea that housing is under-taxed. They point out
that housing is an investment but also provides a service, and using a dynamic general
equilibrium model they conclude that ‘in general the optimal tax rate on the imputed
rent should not equal the tax rate on business capital income’. Indeed they suggest
that ‘both housing and other consumption should be taxed at relatively high tax rates,
whereas the tax on business capital income should be close to zero’ (p.27).
15. The foundation of their argument would seem to be the imputed income from housing
should, at the least, be taxed at the same rate as the income from business capital. But
then if commodities and services are being taxed then the housing services which are
obtained from the housing should also be taxed. Thus we arrive at the notion that
housing should not be exceptionally undertaxed, but should, if anything, be taxed
highly.
16. Thus the theory of optimal taxation can be called in aid, by Cremer and Gahravi, as a
justification for the provision of subsidised housing for the less well off but also, by
other authors, as a justification for a policy of high tax rates on housing, at least as
high as for other commodities. Note however that land is ignored in these analyses.
As Eerola and Maattinen put it in the last sentence in their paper, ‘one possibly
important extension would be to take land into account by assuming that housing
services are provided by the combination of land and structures’ (p. 28).
17. The exclusion of land from these models may be important, and we shall return to the
problem in a later section. Meanwhile we note the discussion by Englund (2003) of
the possibility that differences in tax rates may be capitalised into land values. As he
notes, ‘in an extreme case there [may] be no supply adjustment in the long run and
the subsidies will all be capitalised in land prices’ (p 946). Indeed, even in the USA,
across US cities Capozza et al (1999) find that the variations in property values are
consistent with virtually full capitalisation of variations in property tax rates. If this is
so in the US context where the supply of housing is elastic, capitalisation is even
more likely in the UK where the supply is very inelastic.
18. Variations in land values can also have intergenerational effects to which we shall
also refer later. Thus Feldstein (1977), using a life cycle model, shows that a
reduction in the rate of tax on land can benefit existing land owners, whose land
increases in value, at the expense of later generations who lose out from higher land
prices.
6
The Optimal Taxation of Housing: Two Conclusions
19. From the above survey of the theory and the evidence we can at least draw two
conclusions regarding the optimal taxation of housing. Firstly, although an argument
for subsidising housing for lower income housing can be derived, the argument is
artificial, based on the possibility that lower income households have a higher
preferences for lower quality housing than do wealthier households. This argument is
at best suggestive, but not conclusive. For the rest of this report we shall set it aside.
20. Secondly, there is a very strong argument that the taxation of owner occupied housing
should be at least at the same level as that of any other business investment, if not
higher. The reasoning is that otherwise there is an incentive for people to save and
invest in their homes, which yield a socio-psychological return (the imputed rent, in
economic terms), rather than investing their savings in business assets which yield a
return in terms of income. If the imputed rent is untaxed whilst the return from
investments is taxed the pattern of investment will clearly be distorted. Overall people
would be better off if owner occupied housing was taxed since other taxes could then
be reduced and people would have an incentive to invest in industry and commerce.
Thus the under-taxation of owner occupied housing distorts the pattern of saving and
investment in an economy, as well as the pattern of consumption. The taxation of
housing should be, at the least, tenure neutral in that neither owner occupation nor
renting should be favoured by the tax system.
Tenure Neutral Taxation in the United Kingdom: Some History
21. If we go back some seventy years or more then the UK tax system did, at that time,
set out to be tenure neutral. Prior to World War II the vast majority of the UK
population rented their homes. Moreover a large number of the houses which were
rented out were owned by the wealthy minority – renting out homes was a way of
ensuring a pension. (In Switzerland today the same situation exists – 80% of the
population rents their home and much of this housing is owned by the middle class,
particularly self employed professionals like lawyers or doctors who rent out homes
to ensure an income after retirement).
22. In the United Kingdom, at that time, owner occupiers were taxed on the imputed rent
of the property they occupied. The logic was as follows. Suppose someone rents their
own home, but invests their capital in another property which they rent out. The rental
income from the second property would be subject to tax. Suppose now that this
person now gives the tenants notice to quit and moves out of his or her rented home
and into the house which had been let out. He or she will no longer have to pay any
rent, but also no longer receives rent and so no longer pays tax on it. He or she is now
clearly better off as an owner occupier because no tax has to be paid on what is, in
effect, the rent paid, as occupier, to oneself, as owner.
23. It is to remedy this apparent inequity that a tax was imposed on the imputed income
from the investment in the family home. Such a tax, collected as part of Schedule A
7
(the schedule in the Income Tax Acts which deals with the income from land or
property), existed in the United Kingdom until the early sixties when it was abolished
24. The reasons for its abolition were practical but also still relevant. Under Schedule A
the tax was imposed on owner occupied properties in the same way as on rented
properties. The owner of a property would be entitled to deduct from the rent received
the costs of maintaining the property and keeping it in good repair, and so these costs
could be deducted from the imputed rent. Similarly mortgage interest was an
allowable deduction for tax purposes when a property was rented out and so this too
was an allowable deduction from the imputed rent. The last revaluation of properties
in the UK for the purposes of Schedule A took place in 1935/36. Though prices did
not change very much during the years of the Depression in the thirties, price
inflation was high during and after World War II in the forties and fifties. But during
this period no revaluations took place because of the War and subsequent economic
problems.
25. As a result, by the late fifties, because prices had risen so much, most owner
occupiers could set against the imputed income maintenance costs and mortgage
interest which were so high relative to the Schedule A valuation of twenty years
before that little or no tax had to be paid. Thus accountants, owner occupiers and tax
inspectors, spent a great deal of time and effort calculating that there was little or no
tax to pay. The Conservative government of the day, faced with the fact that, firstly,
as things stood the tax was not worth the trouble of collecting, and that, secondly, if
there had been a revaluation there would have been political uproar, and, finally, that
it wished to encourage owner occupation, quietly abolished the tax on imputed rent.
26. One feature of the tax remained. Although expenditure on maintenance was no longer
tax deductible, mortgage interest remained deductible, even though this was illogical
if the imputed income was not to be taxed. But then, as we have already said, the
government was seeking to encourage owner occupation, and keeping mortgage
interest tax deductible helped. But as owner occupation increased this tax relief
became increasingly expensive. The amount of the loan for which tax could be
deducted was limited from 1976 onwards, and, in a number of steps, mortgage
interest relief was finally abolished twenty five years later. The logic of a tax on
imputed rental income remains correct, however, and it was only abolished in
Denmark in 1998.
27. One further point should be made. Schedule A coexisted with the domestic rates,
which was itself also a tax based on an estimate of the rental which could be obtained
from a property, and which was charged on both rented and owner occupied property.
It is evident that, from an optimal tax theory point of view, the two taxes, Schedule A
and the domestic rates, together resulted in a tax on residential property which was
greater than the tax on other commodities and, as mooted by Eerola and Maattinen,
certainly greater than that on other business income. The reasons for this seem to be
political and practical rather than economic. The rates came into existence as a tax
which could be levied on local residents, which was simple to administer and was
8
roughly related to their ability to pay, and which did not require the sort of
sophisticated accounting records which other types of tax might need. Nevertheless
the result was that, up to the middle of the twentieth century, housing was highly
taxed relative both to other goods and to other forms of income. Only in the middle of
the century, with the decreasing impact and eventual abolition of Schedule A, and
with mortgage interest remaining tax deductible, owner occupied housing in the UK
became relatively low taxed.
28. It was during this period from the fifties through to the eighties that owner occupation
was most favourably taxed, though it also has to be said that during the same period
local authority housing was heavily subsidised, and rents in the private rented sector
were tightly controlled. Since letting houses was highly unprofitable during this
period the supply of privately rented housing fell substantially.
29. From the late eighties onward the position changed again. Firstly, during the nineties
mortgage interest relief for basic rate tax payers was gradually phased out. But,
secondly, from 1989/90 onwards the domestic rate was abolished, and then replaced
by the Council Tax. Thus the taxation of owner occupied housing was substantially
reduced, particularly for the better off and older households whilst it was not reduced
so much for the poorer and for the young who were saddled with mortgages.
Explicit Taxes and Housing in the UK: the Current Position
30. At the time that Schedule A was abolished, fifty years ago, the taxation system was
different. We have outlined some of the changes which have occurred since then in
the previous section, the most important of which is that mortgage interest is no
longer tax deductible. But other taxes have come into existence. Then there was no
Value Added Tax or Capital Gains Tax or Council Tax. In this section we look at
several direct taxes as they impact on property and consider their impact on rented
and owner occupied housing. We also look at recent changes to Income Tax which
are intended to encourage saving.
31. Capital Gains Tax We start with one of the simplest. Capital Gains Tax is not levied
on the sale of the primary residence of an owner occupier (although it is levied on
second homes). But the sale of properties which are owned by landlords for
investment purposes would be liable to CGT. It therefore appears to be non-tenure
neutral. But if a property is held for investment purposes and the proceeds of the sale
are reinvested in similar investments then any capital gain can be rolled over, and
payment delayed. Further it is not levied if the property has to be sold on the owner’s
death. And the rate of tax for most owners has recently been substantially reduced.
The implication is that the actual rate of taxation of the increase in value of an
investment property is not very high and the extent of non tenure neutrality is small,
difficult to calculate and uncertain. It does nevertheless exist, however. In the United
States, where the principal residence is subject to CGT, it has little effect because the
capital gain can be rolled over, and, of course, invariably is.
9
32. Stamp Duty Another tax which is relatively simple to analyse is the stamp duty.
Stamp duty is levied on any sale of a residential property. It follows that it is charged
when the residents of an owner occupied property move house but not when the
tenant of a rented house moves. However the existence of such a tax almost certainly
encourages the more mobile tenants, notably the young, to rent and not buy, since
they do not intend to stay long in one place. But this is a pattern of behaviour which
we would expect to occur anyway, both because the young are more mobile anyway,
and because of other transaction costs associated with owner occupation such as
agents’ and lawyers’ fees. Thus, given the relatively low level of stamp duty it does
not seem likely that its existence distorts a pattern of home ownership which would
occur anyway.
33. Value Added Tax In the UK VAT is not charged on the construction of residential
properties, but it is charged on their extension, repair, refurbishment, and
improvement. It is charged on the construction of industrial and commercial
properties, and on all other commodities except food and children’s clothing. Since it
is charged whether the property is to be owner occupied or tenanted it would appear
to be tenure neutral. On the other hand optimal tax theory would suggest that it should
be charged on all forms of building construction.
34. Council Tax Since this is levied on both owner occupiers and on tenants it is overtly
tenure neutral. On the other hand it is a non-linear (regressive) tax on housing
consumption and as such it is very much not in line with the principle of optimal
taxation theory. The tax is highest as a rate per £ or per sq. ft. on small properties and,
as a marginal tax rate, it becomes lower as the amount of housing consumed
increases, falling to zero when the value of the property is about a million pounds or
more. It is possible that a theorist could come up with a set of assumptions where this
would be an optimal tax but it is difficult to conceive that such assumptions would be
remotely plausible.
35. Income Taxes and ISAs. There is one further point which needs to be made. PEPs
were introduced in the early nineties and were intended to encourage saving. In the
late eighties ISAs were introduced. Over the past fifteen years people have been able
to invest substantial amounts through ISAs and PEPs so that neither the dividends
received nor the capital gains made are taxable. During most of this period a couple
could invest £14,000 per annum through the stock market. Thus, investment which is
not in housing is now somewhat more favoured, relative to owner occupation than it
used to be.
Tenure Neutrality and Explicit Taxes: a Tentative Conclusion
36. At this point we can attempt to summarise the position with respect to tenure
neutrality in the United Kingdom. Clearly there is an under-taxation of owner
occupied housing in that there is now no system in place for taxing the imputed rent
of owner occupied housing. Unlike the United States, however, mortgage interest is
10
not now tax deductible. In the paper by Gervais (2002) which we referred to earlier he
notes that, with regard to his simulations of the US situation, ‘since mortgage interest
deductibility increases the incentive to become a homeowner, the distortion
associated with [the undertaxation of housing is] greatly reduced by the removal of
mortgage interest deductibility’(p.1464). If Gervais’ simulations are accepted as
accurate it follows that the degree of distortion in the UK has been much reduced by
the elimination of mortgage interest tax deductibility over the last quarter of the
twentieth century. The analysis of the explicit taxes set out above does not appear to
challenge this view. The distortions caused by Stamp Duty, Capital Gains Tax, and
Value Added Tax would seem to be minor.
37. The major distortion would appear to result from the way in which the UK tax system
favours the wealthier and older as against the younger and poorer. In this it both
distorts and is inequitable. It largely exists because of the structure of the Council
Tax. The tax is high relative to the value of cheaper properties and low relative to the
value of more expensive properties. Thus the taxation of housing to younger
households will, relative to its value, be much greater than it will be to older and
richer households. They will have paid off their mortgage and so will be paying no
interest, will be paying a relatively low Council Tax and have no incentive to move.
At this point, having paid off their mortgage, they will also be able to invest savings
in ISAs so that their investment will yield a tax free return.
38. A peculiarity of this situation is that it appears to be intended to encourage people to
buy and remain in bigger houses. In this it runs counter to planning policies which set
out to encourage higher density housing and tends to base this partly on a prediction
that small properties are necessary because of all these older households who should
move into smaller properties now that the family has grown up and left. These older
households, on the other hand, are given no incentive to move by the tax system.
Thus, with respect to the UK housing market, so far as explicit taxes are concerned,
the position would appear to be complex. It would seem to be close to tenure
neutrality for smaller houses and younger households but very divergent from it with
regard to larger houses and older households.
39. The above analysis deals only with explicit taxes on housing. But in our earlier
review of recent work on optimal taxation and housing we noted that at various points
authors indicated either that they were not considering land (Eerola and Maatanen,
2005), or that capitalisation through the land market significantly increased the
distortionate effects of tax policy (Skinner, 1996) or that the intergenerational effects
of changes in land values may be significant (Feldstein, 1977). In the US economy
the supply of land has been relatively elastic, whilst in the UK the supply of land for
housing as become increasingly inelastic over the last half century, and, although we
here start to deal with implicit rather than explicit taxes it is necessary, for the sake of
both completeness and accuracy to consider the British land market and UK planning
policies.
11
Land, Land Use Planning and Intergenerational Transfers
40. In an analysis of the cost of housing in the United Kingdom at the present time it is
probable that the most important tax is an implicit tax, one which is there but is not
levied by HMRC but by one generation on the next. But first it is necessary to explain
the nature of the problem and how such an implicit tax can occur.
41. The British do not favour building on green field sites; they favour, as far as is
possible, the preservation of the countryside. If asked, an economist might have set
out to protect such land by imposing a tax on its use. And if such a tax had been
imposed back in the fifties, then, over the years, the level of the tax would have
gradually increased as the demand for housing increased, because of increasing
incomes, more households and a larger population, in order to choke off the increased
demand for housing and housing land and to ensure that no more land was used than
had been planned for. Such a tax might have been directly on housing or on land. If
on either, and accurately set, the price of land would not have risen but would have
remained close to agricultural value. But the cost of housing would have risen
substantially.
42. The choice actually made was to constrain the availability of land for development.
This is understandable – the results are more certain than with a tax, and constraint is
politically acceptable where heavy taxation would not be. Nevertheless the policy of
constraint has some of the same economic consequences as a policy of taxation,
particularly as successive governments have attempted, over the years, to appropriate
parts of the increase in land values, through the Development Land Tax in the
seventies and eighties and then through Section 106 agreements requiring the
construction of affordable housing and other forms of contribution, a Planning Gain
Supplement, or a Community Infrastructure Levy.
43. As Kate Barker (2003) demonstrates in the first volume of her Review of Housing
Supply the price of housing in England has risen in real terms by about 3.3 % per
annum between 1971 and 2001, considerably faster than the average rate of increase
for Western Europe of 1.8 %, also in real terms. The increase in the cost of housing
has had an impact on consumption exactly equivalent to a tax on housing. Thus new
housing in England is the smallest in western Europe, and the average size of all
existing housing is smaller and older than the European average.
44. To the household buying a house it matters little, in economic terms, whether the
price of the house is high because of a land tax or because of land supply restrictions.
But, given that the house is both a capital asset and provides a service the position is
quite complex. Would be house owners in the UK recognise that housing is expensive
but they also assume that the value of the house will increase over time. Past
experience suggests this so that even now, in the summer of 2009, it would not be
overly optimistic to assume that property values will come back to current levels in
five or six years time. As we noted in the previous paragraph experience over the past
thirty years is of a real increase of 3.3 % per annum. The high price that has to be
12
paid suggests reducing consumption; the anticipated increase in price suggests buying
as much house as one can, and borrowing as much as one can to do so.
45. Looking at this situation with regard to the question of tenant neutrality it would
appear that the policy of constraint certainly favours owner occupiers as against
tenants. The latter pay a rent which is related to the current capital value of the
property; any increase in the value of the property will go to the landlord while the
tenants will have to pay higher rents. Owner occupiers, on the other hand, certainly
pay a high price for their property, but then they gain from the increase in its price
over the succeeding years, particularly since this is not subject to Capital Gains Tax.
Admittedly they can usually only realise this increase in value by selling up and
downsizing, but it is evident that many British regard this with equanimity,
particularly since it absolves them from much of the responsibility for saving for their
retirement.
46. On the other hand it is evident that investors in lettable property factor into their
decisions not only the rent that can currently be obtained but also the increase in
value that can be expected in the future. To cite an actual example, in July 2007 a
property in Ealing was valued at £400,000, and the monthly rent that could be
obtained for it was £1,400, i.e. 16,800 p.a. a gross rate of return of 4.2 %. But out of
this income agents had to be paid, the property had to be maintained, depreciation of
the building had to be allowed for, and, if letting on a long term basis, vacancies had
to be allowed for. It is evident that, at that time, a greater return could be obtained
with less risk by putting the money in a Building Society. Since this example seems
to be representative of the decisions being made by investors ‘buying to let’ at that
time, it is evident that these would be landlords were anticipating increases in the
value of their properties, and this has the clear implication that rents were lower than
they would have been if landlords had not taken this view.
47. However, on balance, one has to conclude from the discussion that there is an implicit
tax on housing in the UK, a tax which raises the cost of housing substantially. Its
distortionate effect is that it awards a capital gain, untaxed, to one generation, but
then the capital gain has to be paid for by the next generation so that much of the
latter’s saving is directed towards paying for housing. These distortionate
intergenerational effects have been noted by the US authors cited above but they are
less important in the US economy than they are in the UK economy. Finally, of
course, the implicit tax favours owner occupation but, as we have argued above, the
effect is not as great as one might expect.
The Taxation of Housing in the UK: Conclusions
48. The literature on the optimal taxation of housing argues that housing should be taxed
as highly as the income from business investments, possibly more highly. The
argument is that otherwise investment is diverted into housing and away from
‘productive’ investment in industry and commerce. Other literature, primarily relating
13
to the US economy, argues that housing is undertaxed. The main reason for this
conclusion is that mortgage interest in the US is tax deductible. The second reason is
that the property tax is regarded as a local tax yielding local benefits – if the benefits
equal the cost of the tax then they cancel each other out.
49. What is the position in the UK? As we noted earlier mortgage interest is no longer tax
deductible in the UK. On the other hand the Council Tax is considerably lower than
the US property tax, but is relatively higher for those occupying cheaper housing. The
position would seem to be that for those starting out, the young, they do pay the full
cost of their investment in property because they will have had to borrow some ninety
per cent of the cost and their mortgage interest is not tax deductible. Indeed what they
are paying may be greater than the cost of renting. This position is reinforced by the
fact that they are likely to be buying a cheaper property where the Council Tax is a
higher percentage of the cost of the investment than it is for more expensive
properties. And there is the implicit tax on land which these households have to pay
to enter the market.
50. The position of older households is different. They will have paid off most if not all
of their mortgage, and are likely to be occupying more expensive homes so that the
Council Tax is a relatively small percentage of the capital cost. For these households
housing investment is undertaxed. And when they sell they expect to receive the
benefit of the implicit tax on housing from the next generation seeking to move into
larger properties. But, on the other hand, because families do not tend to move much
once children have reached school age the social cost of moving on and into larger
housing becomes too great. Overinvestment is not likely.
51. I have to conclude that the combination of taxation and planning policy does not
undertax housing. After all, let us look at the evidence. As was mentioned earlier,
new housing in England is smaller than in the rest of western Europe, and existing
housing is generally smaller and older than in Europe. And the differences are not
small. The latest figures for England date from 1996 when the average size of a new
property was 76 sq. m. At about the same time the average size of a new property in
Germany was 109 sq. m. and in France was 113 sq.m. (Erostat, 2002)
52. Relative to other countries these figures do not suggest over investment. Nor do the
figures on housing construction which have been relatively low for the past twenty
years, the reason why Kate Barker was asked to review policy on housing supply.
Housing completions peaked at over 400,000 a year in the mid sixties, and fell to less
than half that figure in the nineties, over a period during which both house prices and
land prices doubled in real terms. (Barker, 2003)
53. In only one way might there be an indication of overinvestment. The older families
cited earlier may not wish to move, and may find housing in the UK too expensive to
upgrade into. But housing abroad is cheaper. It is possible to use the increase in the
value of one’s primary residence to invest in a second home abroad or, less likely, in
the UK. Statistics on this do not seem to be available but casual empiricism suggests
14
that large numbers of middle class Britons have homes in France, Italy, Spain,
Greece, and Florida. The economic impact of this flow of investments abroad is
difficult to calculate.
The Optimal Taxation of Housing: A Recommendation
54. We began this report with a discussion of the basic optimal taxation theory. To
recapitulate, the basic theory suggests that a good with a low income elasticity like
housing should be relatively lightly taxed for reasons of equity. This appears to be the
reason why VAT does not have to be paid on new housing construction, food and
children’s clothing in the UK. The problem is, as we noted above, that the UK
planning system, by constraining the availability of land, causes the price of housing
to be high, higher than in most other countries. In practice, therefore, the UK system
is inequitable, despite the (explicit) tax system.
55. The other point that was made in our discussion of basic theory was that a tax which
does not affect the total amount supplied and consumed very much also does not
distort welfare very much if at all. (In the earlier discussion the point was made with
respect to the price elasticity of goods.) But as we have noted the supply of land and
the supply of housing are highly regulated and very little subject to market forces. It
follows that taxes could be levied which would cause little welfare loss. In our
discussion of the implicit taxation of housing through land controls we noted that the
benefits went to the owners of land but that at various times attempts had been made
by central and local governments to appropriate a share of these land value increases,
through various taxes and Section 106 agreements. The recommendation would be
that VAT should be charged on new housing. The impact would be on the owners of
land, so there would be little or no welfare loss. The argument that it should not be
charged for equity reasons no longer holds, given that house prices are high anyway,
and that for years VAT has been charged on the refurbishment and extension of
existing homes. I believe that this would be better solution to an intractable problem
than the proposed Community Infrastructure Levy.
56. In this report I have surveyed the evidence regarding optimal taxation and housing,
and tried to relate the findings of the researchers to the UK economy. All of the
papers surveyed are by American or Scandinavian authors and most of the work
relates to the US economy, however. The UK economy, as I have pointed out, differs
from most other economies in the world in the strength of its land use planning
controls. I have tried to indicate the way I believe the planning controls impact on the
housing market and what might be regarded as an optimal taxation policy for that
market. It is, however, unfortunate that the analyses of optimal taxation policies
toward housing have done little to incorporate an analysis of the interaction between
the housing market and the land market, after all housing cannot really be treated as
just another manufactured commodity when at least forty per cent of the price of a
house is the cost of the (non-manufactured) land on which the house sits. I hate to
15
finish with the usual call for further research, but in this case, I am afraid that I have
to.
References
Atkinson, A.B., and J.E.Stiglitz (1980) Lectures on Public Economics (London: McGraw
Hill)
Barker, Kate (2003) Review of Housing Supply: Interim Report – Analysis (HMSO:
Norwich).
Capozza, D., R. Green and P. Hendershott (1999) Tax Reform and House Prices: Large
or Small Effect? in NTA Proceedings of the 91st Annual Conference, 19-24.
Cremer, Helmuth, and Firouz Gahvari (1998) On Optimal Taxation of Housing, Journal
of Urban Economics, 43, 315-335.
Eerola, Essi, and Niku Maatanen (2005) The Optimal Tax Treatment of Housing Capital
in the Neoclassical Growth Model, Bank of Finland Research Papers, 10.
Englund, Peter (2003) Taxing Residential Housing Capital, Urban Studies, 5-6, 937-952.
Engelhardt, G.V. (1996) Consumption, Down Payments, and Liquidity Constraints,
Journal of Money, Credit, and Banking, 28, 255-271.
Eurostat (2002) European Housing Statistics, European Commission.
Feldstein, M.S. (1977) The Surprising Incidence of a Tax on Pure Rent: a New Answer to
an Old Question, Journal of Political Economy, 85, 349-360.
Gervais, Martin (2002) Housing Taxation and Capital Accumulation, Journal of
Monetary Economics, 49, 1461-189.
Laidler, David (1969) Income Tax Incentives for Owner Occupied Housing, in A.
Harberger and M.Bailey (Eds.) The Taxation of Income from Capital, 50-76.
(Washington, DC: The Brookings Institution)
Skinner, J. (1996) The Dynamic Efficiency Cost of Not Taxing Housing, Journal of
Public Economics, 59, 397-417.
16