Printer Friendly

Housing and the wider economy in the short and long run.

The Review is pleased to give hospitality to CLARE Group articles, but is not necessarily in agreement with the views expressed; responsibility for these rests with the author. Members of the CLARE Group are M.J. Artis, A.J.C. Britton, W.A. Brown, C.H. Feinstein, C.A.E. Goodhart, J.A. Kay, R.C.O. Matthews, D. Miles, M.H. Miller, P.M. Oppenheimer, M.V. Posner, W.B. Reddaway, J.R. Sargent, M.F-G Scott, Z.A. Silberston, J.H.B. Tew, S. Wadhwani and M. Weale.

Introduction: The importance of housing in the economy

This article analyses the links between developments in the housing market and in the wider economy. Three questions are addressed. First, what have been the forces behind the dramatic increase in the proportion of wealth held by the personal sector in the form of owner occupied housing over the century and what are the implications for future saving? Second, what does the interaction of past investment decisions in the housing stock and future demographic trends suggest about the resources which will be channelled into housing over the next fifty years? Third, what are the short run and long run effects of changes in the price of houses relative to other goods? These questions are analysed by focussing on long-term trends-in the stock of houses, in house prices, in the demand for houses, in bequests of property and in financing patterns. It is important to take a long-term view of the housing market because the process whereby housing wealth has been accumulated, and the ways in which any discrepancy between actual and desired holding of housing wealth are removed, are slow-moving. More fundamentally, it is worth thinking about the links between housing and other markets because they are likely to be quantitatively significant, reflecting the economic importance of houses which are: expensive(1) last a long time(2) yield a substantial flow of real services each year(3) and, in the UK, constitute a significant part of national net wealth(4). These facts are, of course, not unconnected.

Economists have always been interested in houses but over the last decade the links between developments in the UK housing market and in the wider economy have been the subject of enormous attention. Housing economics is no longer the preserve of micro-economists focussing upon the welfare effects of the system of housing provision or on analysis of the implications of the tax treatment of housing; macroeconomists now see developments in the housing markets as of prime importance to the wider economy. Modelling the housing market has become central to the process of macroeconomic forecasting. In some ways this is puzzling, for although total UK housing wealth is substantial the net worth of all other assets remains greater. As regards aggregate consumption and saving it might be thought that it is the total value of wealth rather than the value of one part of wealth, albeit a large part, that is important. But there are reasons for thinking that houses are special.

First, owner occupied houses are both an asset and a consumption good; they are valuable, durable assets which also directly yield real benefits in the form of housing services. One implication of this is that changes in the price of houses relative to other commodities simultaneously increase the value of the wealth of home owners and the price of housing services, both measured in terms of other goods. The impact of such price changes on the level of spending is likely to be different from the effect of a change in the value of equities which in itself leaves unchanged the price of all goods consumed.

Second, housing wealth has become relatively fungible; it can be transformed into other commodities more easily than other forms of wealth which are important to the personal sector. For example, personal sector wealth in the forms of pensions and life assurance policies now accounts for approximately 20 per cent of their net worth; net (of mortgage debt) owner-occupied housing accounts for around 40 per cent. It is still difficult for households to liquidate that part of their wealth which is in pensions or life policies whilst it became easy in the 1980s for households to take out second mortgages or top-up mortgages against the collateral of accumulated equity. Again, this makes trends in housing wealth of particular importance.

Third, the fact that prior to the 1980s it was relatively hard for households to extract housing wealth, at least without moving to the rented sector or else to a smaller house, also means that the stock of net housing wealth held may have been (and even ten years later might still be) significantly different from its equilibrium value. Had the existence of rent controls and the presence of tax breaks on owner occupied housing not made the rented option unattractive the difficulty of extracting accumulated net housing equity might not have had much effect on saving and consumption. But rent controls have made private sector renting uneconomic (see, for example, Ashton, Minford and Peel (1987) and Black and Stafford (1988)). For these reasons the financial liberalisation of the 1980s in the housing finance market may have had, and might continue to have, a significant effect on spending plans. Thus, it is both the fungibility of housing wealth relative to that of other important components in total wealth and the recent change in the degree of fungibility of housing wealth that makes housing wealth special. in the light of this it is not surprising that the combined effects of rapidly rising real house prices in the 1980s and the greater ability of households to borrow against housing wealth-phenomena which I will argue are not unconnected-have been widely seen though not convincingly proved) to have caused the sharp fall in the personal sector saving rate. (See Muelibauer and Murphy (1989), Congdon and Turnbull (1982), Miles (1992), Lee and Robinson (1990)).

These factors which make housing wealth spccial-the degree of, and changes in, fungibility and the dual aspect of houses as durable assets and consumption goods-prompt the questions which I will address in this article.

The plan of this article is as follows. Section I describes long-run trends in the UK housing market and makes some international comparisons. Section 11 analyses the effects of housing inheritances. Section Ill focuses on financial liberalisation and equity withdrawal. In Section IV demographic factors are considered. The effects of taxes, interest rates and expected house price changes on the real cost of housing are analysed in Section V. The impact of changes in real house prices on consumption, saving and bequests is analysed in Section VI.

I. UK housing in the twentieth century Tables 1 and 2 reveal some of the main trends in UK housing over the century. If able I shows that the stock of dwellings has risen dramatically, though far from smoothly, over the past ninety years. The proportionate increase in the number of dwellings over the century has been almost twice as great as the rise in the population. Although the increase in the real stock of housing per capita is very imperfectly measured by the ratio of dwellings persons (for example houses will vary through time both in average quality and size) it is clear that there has been substantial accumulation over the century; the ratio of rooms to persons, a no more precise (but different) measure of real housing per person, has also doubled. Table 2 shows that owner occupation has also risen enormously over the past eighty years. Almost 70 per cent of households are now owner occupiers. A comparison with the UK on the eve of the first world war reveals just how dramatic the change in housing tenure has been; only 10 per cent of households were owner occupiers in 1914. Much of the rise in owner occupation over the century has been concentrated in the post-war period; mass home ownership in the UK is a relatively recent phenomenon. Nor is the rise in the rate of home ownership over. Surveys undertaken by the Building Societies Association (BSA) consistently reveal a strong preference for owner occupation. Hamnett et al (1991) report that the 1989 BSA survey showed that 85 per cent of the sample interviewed expressed a desire to be home-owners; an even higher proportion of the young aimed to own homes. By 2014 it is plausible that more than three quarters of households might be owner occupiers, a century before 90 per cent of households lived in rented accommodation.

Table 2 also shows that real house prices have been on an tipward trend through the century. Comparisons of real prices across such a long period are, however, problematic because adjustments for quality changes become increasingly difficult, so chart I focuses on the past 25 years. Over that period real prices using either the Department of Environment mix adjusted price index or the average price of homes purchased with mortgages, both relative to retail prices) have doubled, even allowing for the decline in nominal house values in 1990 and 1991. The mix adjusted series makes some allowance for changes in the quality of homes by standardising for the number of habitable rooms; but other quality changes, for example double glazing and central heating, are not accounted for. At least part of the apparent rise in the real price of houses is therefore due to home improvements. As regards the real value of the housing stock-the product of the number of dwellings and the average real price of homes-it is appropriate that quality changes should be reflected and it is right that that part of the rise in real house prices that reflects tangible improvements be included.

The combination of an increasing stock of dwellings both absolutely and relative to the population), rising rates of owner occupation, mortgages which (at least in the past) have had to be fully repaid over periods rarely exceeding 25 years and rising real house prices has meant that real, owner occupied housing net wealth held directly by the personal sector has increased sharply. Chart 2 shows net wealth in residential housing (the value of the owner occupied housing stock net of outstanding mortgage debt) relative to total personal sector net worth (the total value of all tangible and financial assets net of all liabilities). That ratio has risen from around 20 per cent in the early 1960s to around 40 per cent by the end of 1990. Over a longer time horizon the increase in this ratio would be far more dramatic-tables 1 and 2 reveal that-ince 1939 real house prices have more than trebled; over that period the stock of dwellings per person has risen by 70 per cent and the rate of owner occupation-doubled. The combination of these three factors will have increased the value of directly held per capita real housing wealth in terms of consumer goods) by over tenfold. Clearly that part of this rise simply due to the switch in ownership of existing houses from landlords to owner occupiers does not represent a real increase in overall national housing wealth. But as the tables reveal the rise in total dwellings (owner occupied and rented) and in real house prices over the century has generated enormous increases in per capita housing regardless of the change in tenure. These increases have been substantially greater than the rise in real GDP-in the period since 1939 the total number of dwellings has doubled and the real price of houses more than trebled generating a sixfold increase in the value of the real stock of housing; over the same period real GDP has risen roughly threefold.

Tables 3 and 4 reproduced from Miles (1992)) compare the UK housing market with those of other major economies. Although there are enormous difficulties in comparing ownership patterns, house prices and mortgage debt across countries (see Holmans (1990) and Miles (1992)) certain conclusions are robust to even significant measurement errors. First, the owner occupation rate in the UK is amongst the highest in the developed world. Second, the stock of mortgage debt relative to household income is also one of the highest in the world and is far higher than in most other countries with comparable levels of owner occupation. Indeed what is surprising about table 3 is that there is little relation between owner occupation and the mortgage debt to income ratio; both Spain and Italy have high rates of owner occupation but little mortgage debt. Third, the net housing equity of the personal sector in the UK is relatively high; amongst the countries for which estimates can be made only in japan is the ratio of equity to GDP likely to be higher (table 4).

To summarise, the UK is a country where the personal sector holds a comparatively large part of its wealth in the form of residential property and where mortgage debt is relatively freely available.

II Housing bequests into the 21st century

One implication of the trends in home ownership and real house prices outlined in the previous section is clear-while most children born in the early part of this century would have lived in rented accomodation for much of their lives and could not have expected to receive a bequest of a house from their parents, a child born in the post-war period has a high probability that their parents will die as owners of a very valuable property. The importance of housing wealth in total bequests is already great. The study by Hamnett, Harmer and Williams reports the results of a 1988 survey into inheritance. 60 per cent of a random sample of households who had ever received an inheritance of over L1,000 had inherited housing or money derived from the sale of housing. But although a high proportion of households who had received a substantial inheritance had been bequeathed housing wealth the number of such households was quite small-only 15 per cent of the random sample of households had ever inherited significant wealth.

The number of properties which are bequeathed to succeeding generations, rather than to a surviving spouse, and are released onto the market is estimated to be currently of the order of 160,000 a year (Hamnett et al, p 73). But the dramatic rise in owner occupation over the post-war period and the increases in life expectancy suggest that the surge in bequests is yet to come and that property inheritance will increase sharply over the next fifty years. The rise in the real price of housing in the post-war period further boosts the total value of homes which will be bequeathed and implies that the proportion of the total value of estates accounted for by property is likely to be higher in the future. Whilst it is the total value of bequests, rather than just that part which is in the form of housing wealth, which really matters for the spending decisions of inheritors, the rises in home ownership, in the stock of dwellings and in the real value of dwellings means that property inheritance is likely to be the key factor behind trends in bequests. Again, the Hamnett study provides fascinating evidence. Using population projections and standard rates of life expectancy they estimate that the average annual number of bequests of housing will increase consistently over the next forty years; the number will be more than twice as high in the period 2026-31 than in the period 1986-91. The projection is consistent with earlier estimates by Morgan Grenfell (1987) who forecast that by the end of this century the number of estates including residential property wealth would be in excess of 200,000 a year. Although the precise rate at which the level of annual bequests of residential property will rise is hard to forecast it is clear that in the longer term that rate must increase substantially. This is a direct implication of the increase in the rate of owner occupation.

The long-run implications for the value of inheritances and for the rate of non-housing consumption of a greater stock of housing wealth being directly owned by the personal sector are potentially great. Consider a stylized example which brings out some of the longer-run effects of the changing pattern of home ownership through the twentieth century. In 1918 our typical boy and girl are born into households in rented accommodation. In 1948 these typical children are thirty, meet, get married, buy a new house and have children. (For simplicity let's assume they do all this in the same year-making 1948 rather hectic).

By 1978 the mortgage has been paid off for some years and their own children are getting married and buying a house. Unlike their parents, however, these newly-weds can expect between them to receive (as a bequest) a house unencumbered with a mortgage just before the end of the century (assuming an eighty year life span for their parents). The lifetime budget constraint of this second generation of home owners is different from that of their parents in the crucial respect that the total value of repayments of capital on their own mortgage is offset by the receipt (around fifteen to twenty years after the first repayments are made) of an asset whose value is likely to be of the same order of magnitude. (House price inflation, the level of mortgage interest rates and the relative sizes of houses bought by the successive generations clearly influence the exact relation between the total value of mortgage repayments and the value of the inheritance). This first generation to inherit an (unmortgaged) house could spend the proceeds from its sale over the remainder of their life and still hand on a bequest to the next generation equal to the value of a home, since by the early part of the 21st century their own mortgage will be fully repaid.

The thing which makes the difference here is that only one generation of this dynasty-the one born just as the First World War ends-has to pay off a mortgage without receiving an unmortgaged house as a bequest. Lifetime consumption of non-housing goods by that generation must, other things specifically net acquisition of non-housing wealth) equal, be lower by roughly the full value of a home. If that home was newly built in 1948 then the lower life-time consumption of non-housing goods by that generation is the counterpart to the extra resources used to build the house. But all future generations have to do to keep the housing stock constant is to cover the depreciation.

One way to see the macroeconomic implications of this fable is to assume that the first generation of new households that were formed after the Second World War moved into newly built homes, but that once that generation was housed the stock of dwellings was in line with the population of households. Net investment in residential housing could be lower once this post-war generation was housed and could stay lower so long as the population of households was constant. The steady state savings ratio-part of which is the counterpart to residential investment-would be lower or else investment in non-housing assets higher.

The simple example ignores numerous factors which are of great quantitative importance. I will consider these shortly. But fables are helpful and what this one shows is that the increase in owner occupation over the twentieth century-a period of rising population and falling household size which, as table 1 showed, has implied a sharp rise in the number of dwellings-has involved substantial household resources being devoted to accumulating housing equity. If the number of households were to grow by much smaller amounts in the next fifty years then the national saving required to preserve a balance between the number of households and dwellings could fall (an issue to which I return below); in itself this is a good thing and one for which we should thank our parents and grandparents.

What the family fable brings out is that the saving done by the original generation of home-buyers generates a tangible asset, the income from which which we can think of as the real net, of depreciation, return on houses) can accrue to all future generations. But the benefits of that saving would fail to stretch into the 21st century if the recipients of bequests spend the proceeds and fall to bequeath an unmortgaged house to their heirs. In the past that would have proved hard to do unless a household was prepared to move to the rented sector late in life and use the proceeds of the house sale to finance retirement consumption. But during the 1980s it became much easier for home-owners to extract housing equity whilst remaining within the owner occupied sector. Second mortgages, top-up mortgages, interest-only mortgages (which allow home owners to borrow against housing equity without necessarily accumulating net equity as the loan matures) and equity-release schemes have been widely marketed. (See Miles (1990 and 1992) and Drake (1991)). If substantial amounts of housing equity are being extracted we should expect to see a sharp fall in the savings rate for the generation with housing wealth at the time credit restrictions are eased. Measuring the extent of equity extraction, and assessing the uses to which equity extracted from the housing market is put, is therefore important.

III Measuring equity withdrawal

Equity withdrawal is usually defined as the excess of net (of repayments) new lending for house purchase over all forms of investment in residential property (see, for example, Bank of England (1985)). Table 5 below, and chart 3, show the nominal and real value of equity withdrawal on this definition, using a comprehensive measure of investment in the housing stock. The chart shows clearly that equity withdrawal has risen dramatically in the 1980s; withdrawal was over 6 per cent of total personal disposable income in 1988. Since 1985 around one half of total new lending each year has been used for purposes other than financing housing investment, repairs or home improvements. The scale of equity withdrawal over that period has been greater than the size of the current account deficit and is, in a purely accounting sense, sufficient to account for the decline in the personal sector saving ratio from the level of the late 1970s and early 1980s. The personal sector saving rate fell from around 12 per cent of income in the early 1980s to under 4-5 per cent in 1988. Savings would have needed to have been just over E20 billion higher in 1988 to have generated a savings rate equal to the average from the early part of that decade. The table reveals that equity extraction in 1988 was approximately equal to this 'savings gap .

Prior to 1980 equity withdrawal was small. The massive expansion in withdrawal has coincided with the period in which borrowing for house purchase has, on a variety of measures-loan to value ratios, the length of mortgage queues, availability of interest-only and second mortgages-become easier (see Bank of England, 1989). It is tempting to interpret the rise in borrowing as a response to the easing of credit restrictions which has allowed the personal sector to finance consumption, and that this was the driving force behind the decline in the savings ratio and its counterpart in the massive current account deficits of the past five years. But the figures in table 5 are consistent with many different stories. In the first place, it is not clear what the personal sector has done with funds which have been raised on the back of housing collateral. One route for equity withdrawal is the sale of properties by the middle-aged which they have been left by their parents. If the buyers of this property finance the purchase with mortgage debt, and assuming the deceased left an un-mortgaged home, measured equity withdrawal would rise by the value of the mortgage. Hamnett and his co-authors report that out of a sample of 295 households interviewed in 1988 who had inherited housing wealth almost 70 per cent sold the property more or less immediately, which is unsurprising since the overwhelming majority of beneficiaries were themselves home owners. They found that the most common initial use of sale proceeds was for financial investment. Around 50 per cent of their sample used most of the funds released to acquire financial assets; only 24 per cent used most of the funds to finance consumption. If it is through the sale of inherited property that most equity is released and if most of the proceeds are saved the implications of equity withdrawal of L20 billion a year for consumption may not be very significant.

But alternative means of equity withdrawal exist-for example, second, or top-up, mortgages; mortgages for 'home improvement'; and home equity plans. The propensity to consume out of equity withdrawn in these ways may be different from that on funds released through the sale of bequeathed property. The Hamnett evidence suggests a low short run propensity to consume out of bequeathed housing wealth. But evidence reported in Maciennan (1990), who studied the uses to which a sample of households who had taken out further advances on their existing mortgages put the funds, suggests a high propensity to consume. His survey suggests that only a relatively small proportion (between 25 per cent and 50 per cent) of funds borrowed for home repairs or improvements may have been used for that purpose; non-housing consumption was the main use of funds. Maclennan's findings suggest a high propensity to consume out of funds raised and is consistent with evidence from the US. Poterba and Manchester (1989) analyse the balance sheets of a large sample of US households who had taken out second mortgages in the 1980s; they estimate that the propensity to consume out of funds raised is around 80 per cent. A crude attempt was made by Miles (1992) to estimate the overall propensity to consume out of equity withdrawn-by whatever means-from the UK housing market. His strategy was, essentially, to assess the correlation between equity withdrawn from housing and that part of aggregate UK consumption which could not be explained by a forecasting equation designed by Pesaran and Evans in the early 1980s) which was constructed using data from a period when extracting equity had been difficult. The correlation implied a high propensity to consume out of equity withdrawn and that the contribution of credit liberalisation in housing markets to the consumption boom of the 1980s was significant. The implied propensity to consume withdrawn equity of 78 per cent was almost exactly equal to the estimates made by Poterba and Manchester for the us.

These pieces of evidence suggest that a substantial proportion of the equity that has already been released from the housing market has financed consumption. But we really only have scraps of circumstancial evidence about what has happened over the past few years. The real issue about the phenomenon of equity withdrawal and the uses to which that equity might be put is this: will a significant part of the wealth accumulated in housing over the past hundred years be used to finance a one off rise in consumption for the current generation of home owners, who are now better able to borrow against the value of their homes and who are also the prime recipients of bequests of houses, or will the net value of housing wealth held by the personal sector be largely preserved for future generations? The question is hard to answer because there has been no other similar episode in UK history-the recent relative to life expectancy) rise of mass home ownership makes the phenomenon of widespread property inheritance a new one. What is clear is that the level of equity withdrawn thus far is relatively small compared with the value of the owner occupied housing stock. Even if we assume that as much as 80 percent of measured equity withdrawal so far has been spent on non-durable consumer goods the implied reduction in net assets held by the personal sector as a proportion of the value of the owner occupied housing stock is unlikely to exceed 10 per cent. So if the decline in equity withdrawal apparent in 1990 and 1991 marks the end of significant transformation of housing wealth into consumption it is clear that the net wealth in housing will not have been seriously diminished. Whether the recent declines in equity withdrawal will be reversed should real rates of interest fall is the big question.

Although it is very hard to know what proportion of inherited housing wealth will be transformed into consumer goods we can at least think about what difference the answer would make. A decision by current households to use a significant part of housing equity to finance consumption clearly reduces the saving ratio. If domestic investment is financed, at the margin, predominantly from domestic savings the implications of this equity withdrawl is that the non-housing capital stock is reduced relative to what it would have been. In effect the release of excess accumulated wealth in housing, itself a product of the interaction of the benefits to owner occupation and the past difficulty in financing consumption with mortgage debt, reduces the stock of other wealth. But in a small open economy which, like the UK, has no capital restrictions it is more plausible that profitable investment can be financed relatively easily with overseas' savings. In this case a collective decision by UK households to liquidate part of their tangible housing wealth need have no effect on domestic investment and can succeed if the overseas sector is prepared to swap consumer goods for claims on houses. This process does not, of course, involve the transfer of houses across the Channel nor the direct purchase of UK homes by Germans, the French or the Japanese. The process whereby net housing wealth is exchanged for consumption is via equity withdrawal whereby mortgage claims upon houses increase-the funding for extra lending coming from overseas' accumulation of claims upon UK banks and building societies which in turn is the capital account counterpart of the current account flows generated by the extra UK consumption.

The implication of this process is clear-if a substantial proportion of the current net housing equity held by UK households comes to be held by those who would like to transfer that wealth into current consumption then given that financial liberalisation has provided means for the release of excess equity in ways which have not previously been feasible, we might expect that housing equity will be reduced in a process which generates low savings and high current account deficits for a, possibly prolonged, transitional period.

Ultimately, the choices made by current home-owners on the uses to which they put net housing wealth depend upon their degree of benevolence to subsequent generations. the first generation of home-owners to receive substantial inheritances of housing equity from their parents aim to bequeath an equivalent amount of real housing to their heirs, the dramatic and sustained, but nonetheless transitional, decline in savings and increase in current account deficits described above will not materialise. But as noted above steady state savings might be reduced relative to the levels over the period when real housing wealth was accumulated. The scale of such effects would depend upon the future levels of gross investment in residential capital required to generate an equilibrium housing stock. A key determinant of that investment figure are population trends.

IV Population and household formation in the next fifty years

The growth in the number of households over the next half century will be a major determinant of net new housing investment. Tables 6, 7 and 8 bring together some key data relevant to the question. Table 6 shows the rise in the stock of dwellings and the balance between dwellings and potential households in the 40 year period from the Second World War. Table 7 shows gross investment in residential dwellings relative to GDP through the century. Table 8 shows forecasts reported by john Ermisch (1991) and Adrian Coles 1991), based on Office of Population Census and Surveys projections and Department of the Environment forecasts, of changes in population and in the number of households over the next thirty years.

Several things emerge from these tables. First, that major investment in dwellings over the twentieth century in the UK has reflected two things: a significant trend increase in the number of households and the transition from a position, in 1945, of a major gap between the potential number of households and the stock of dwellings to one-by 1981-when that gap had been, in aggregate, removed and when the number of dwellings exceeded the number of households. The projections reported by Ermisch and Coles also suggest that the net increase in new households over the next 40 years will fall steeply and be substantially lower than the average annual increases since the second war. Ermisch reports that predictable changes in the age distribution will in themselves lead to a slowing in the rate of annual net household formation from 160,000 to 40,000 over the next 15 years.

The implication of the tables is that whereas throughout most of the twentieth century there has been substantial net investment in housing in order to keep the number of dwellings from lagging too far behind the number of households tables 6 and 7), in the 21st century there is likely to be a substantially reduced requirement for new dwellings (table 8). Provided that domestic ownership of the housing wealth that currently exists is not substantially reduced by a process of equity withdrawal of the sort which generates an increase of net overseas holdings of liabilities of UK financial institutions, the implication of a reduced level of net residential investment in new dwellings is that resources available for other purposes by UK residents are likely to be higher. Of course the demand to improve the size and quality of existing dwellings will still be a factor, so even if the number of new dwellings required to match the increase in households were to fall sharply total net investment in the housing stock would fall by proportionately less. But we might still expect some decline in net residential investment. Consequently, investment in other assets can be higher or national saving can be lower.

V. House prices and the cost of housing services

We have looked back at the process whereby housing wealth has been accumulated and forward at the implications of that accumulation process for consumption and saving. One factor which is crucial to an understanding of both the past and future are trends in housing costs-both house prices and the user cost of housing. Two long-run issues seem particularly important to me. First, to what extent do trends in house prices; the ways in which housing is taxed; and in interest rates help explain the tenure choices made over the century? Second, what is the impact of changes in house prices upon the consumption choices faced by individuals?

The best way to evaluate how the tax system, financial markets and housing markets have combined to generate incentives to owner occupation is to evaluate the user cost of owner occupied housing. Chart 4 shows a measure of the ex-post user cost of housing, that is the cost of housing allowing for what actually happened to house prices rather than what might have been expected to happen when tenure decisions were made. The series measures the real resources which would have been used by an individual as a result of owning a typical home in each year.

The general formula for the user cost of housing is:

[uc.sub.t], = [(PH.sub.t/P.sub.t)][B.sub.t.r.sub.m. {[Alpha.sub.t](1-[] + (1-[Beta.sub.t]) + [] + [] + d + [m.sub.t] - [P.sub.t] ]

 [uc.sub.t] is the user cost of housing at time t
 [PH.sub.t] is the nominal house price index at time t
 [p.sub.t] is the index of prices of other goods at t
 Beta is the proportion of the house financed through
mortgage debt
 [r.sub.m] is the nominal interest rate on mortgage debt
 alpha is the proportion of mortgage interest repayments
which are deductible against income tax
 [t.sub.y] is the rate of income tax
 [r.sub.i] is the post-tax return on alternative investments, or
the opportunity cost of funds invested in home
 [t.sub.p] is the rate of property tax
 d is the rate of physical depreciation of housing
 m is maintenance and repairs as a proportion of the
value of the home
 P is the rate of increase in nominal house prices.

The measure reflects the real purchase price of a home, the cost of financing the purchase (whether by mortgage or else by forfeiting the chance of holding financial assets which yield returns), taxes paid as a direct result of home ownership (for example, rates) repairs and maintenance, and depreciation. (For exact details of how the various measures are derived see the appendix). The real house price (PH.sub.t/P.sub.t] in index number form and is equal to unity in 1985; the user cost measure therefore represents an index of how many consumer goods needed to be given up in each year to enjoy the services of an owner occupied home. If the user cost index is negative the real capital gain on housing exceeded all the costs of home ownership so that owner occupation, rather than requiring a commitment of real resources, yielded a real profit in that year.

Since the cost of housing varies with the way in which the house purchase is financed, as well as with the tax status of the home owner, there is no unique measure of the user cost of housing. But if we consider the opportunity cost of wealth tied up in homes (r.sub.i) to be the net of tax return on building society deposits then variations in the proportion of the value of the home financed by mortgage have a small effect on the user cost. This is because for much of the period since the early 1960s a high proportion of mortgage interest payments have been tax deductible and the spread between the mortgage interest rate and the rate paid on deposits is relatively small. So whether we focus on houses financed primarily on mortgage (first time buyers) or houses financed largely by equity which we term the 'average' case and is shown in chart 4) the same picture emerges-a negative cost of housing during the major house price booms of the early 1970s and late 1980s with a slightly negative average cost of capital for the whole period. The picture seems to confirm the conventional view that housing has been a good investment-even for those that gained no benefit from the real services provided by home ownership. If we allow for the benefits of home ownership by taking the ratio of rents to the market value of private rented accommodation (a number we might conservatively estimate in the range 4-7 per cent 5')) the conclusion that housing has been a great investment over the past thirty years seems clear.

But two adjustments must be made to the measures in chart 5. First, it is expected house price inflation rather than actual capital gains which are relevant for tenure decisions. Second, because the capital gain (or loss) element on housing is unknown the investment which is made is risky and a risk premium should be added to the cost of funds. If we make the crude assumption that the return on housing is no less risky than the return on UK equities we need to add a premium of around 8 per cent to the cost of funds-this being the average, actual excess return on UK equities over a safe rate of interest for the period 1919-88 (see Spackman (1991) and Office of Water Services (1991)). One way to think about this risk adjustment is to assume that the alternative to sinking wealth in housing is to invest in unit trusts rather than putting the money on deposit with a building society.

Making either of these adjustments has a significant impact on the user cost of housing. Chart 5 shows the effect of replacing actual inflation with the average house price increase over the previous five years-a very crude measure of expected house price inflation. The chart also shows the further impact of adding a risk premium of 8 per cent to the user cost. Both adjustments are made to the measure for the 'average borrower'. The net impact of these changes is to make the average user cost positive, with a mean of around 5 per cent for the period 1963-90, and to reduce the volatility of the time series. Taking 4-7 per cent as an estimate for the range of values of the annual housing services provided by a home as a proportion of its market value we are left with the conclusion that the ex-ante return to housing has not been unduly attractive in the period since the early 1960s. These figures in no way suggest that the tax benefits to home ownership-the existence of mortgage tax relief and the zero tax rate on the imputed rental income-are insignificant. Rather they imply that those benefits have been capitalised in house prices; given the tax implications of home ownership house prices are driven up so that the tangible benefits of home ownership are priced so as to clear the market. One implication of this process is that because house prices come to reflect the tax benefits to owner occupation it is not easy to earn an economic rate of return by renting out houses as rental income, unlike the imputed rent to owner occupiers in the period since the abolition of Schedule A taxation in 1963, is taxed. A second implication of the capitalisation argument is that tax breaks on owner occupation may do little to help potential owner occupiers since they are, in large part, reflected in buying prices. The real gainers from favourable tax treatment are those who bought property when the expected present value of tax concessions were lower than they subsequently became. Since many of those who bought property when the value of the tax concessions was low are now dead we might more accurately identify the beneficiaries as their heirs. Yet even here it is not clear that gains are real since those who inherit houses whose real values have increased sharply may feel obliged to increase their bequest to the next generation to offset the change in the real cost of homes. This Ricardian point leads to the last of the issues raised at the beginning of this paper-who gains from increases in real house prices?

VI Changes in real house prices and national welfare At the aggregate level the impact of changes in real house prices is unclear. At the micro level, however, the answer is fairly obvious-those who own homes and are prepared to trade down to less valuable dwellings, either in the rented sector or within the owner occupied sector, are better off when the relative price of property increases. If you move from one house to another whose value is 75 per cent of the original you are able to buy goods worth 25 per cent of the value of a house; the more valuable are houses the greater the quantity of such goods can be bought from trading down. But in the immediate aftermath of a real house price rise there is only one way for society as a whole to trade down-sell off part of the housing stock to foreigners; all house moves between domestic residents simply cancel out with the gain of those trading down after real house prices rise exactly offsetting the loss of those trading up. Notice that domestic residents could not gain from real house prices by selling their homes to overseas property developers and renting them back since rental costs will generally be proportional to house values-the only way for UK residents as a whole to gain is to simply trade down en masse. The scope for this is clearly somewhat limited and the demand from overseas for UK residential property might be small. Clearly if the real value of houses were to become astronomical and overseas buyers could be found then only a rather limited degree of trading down would be needed to realise large welfare gains. But such a scenario seems rather far-fetched and is reminiscent of the story that the land in the Imperial Palace in Tokyo, when valued at the same price per square metre as adjacent land in the business district, was at least a few years ago) more valuable than the whole of California. The relevance of that story is not that the calculation was inaccurate but that no American who might be lucky enough to own even a smallish fraction of California would ever trade it to use, rather than rent out, a small piece of land on the other side of the Pacific.

(In the longer term it is feasible for a society to collectively trade down without selling off part of its housing stock to the overseas sector by reducing gross investment. But the maximum rate at which trading down of this kind is possible is determined by the depreciation rate of housing and is therefore low.)

The Ricardian argument, that changes in real house prices have distributional effects but that aggregate effects are limited, has obvious force and is sufficient to dispel the belief that the enormous rise in real house prices this century (table 2) has created real benefits. The argument also casts doubt on the belief that house price rises fuel consumption. Assuming that the overall economy-wide scope to substitute out of housing is insignificant, aggregate wealth effects from real house price rises are approximately zero. Those who believe that house price rises fuel consumption via wealth effects are really relying on the propensity to consume of those who are able to trade down for example, those who have been owner-occupiers for many years) exceeding the propensity to consume of those who aim to trade up (for example, potential first time buyers). One plausible story why this might be true is that the prime losers (those waiting to become first time buyers) are already credit restricted because they cannot easily borrow against future real employment income and are forced to consumer less than they would otherwise choose; consequently they do not reduce consumption when the future real cost of buying a home rises. The gainers, however, (established home-owners) can borrow against higher housing equity and increase consumption. This may be the mechanism that John Muellbauer has in mind in much of his careful work on house prices and consumption. But the implication of this story is that consumption responses to real house price inflation will eventually be reversed as the losers grow older and find that the real cost of housing is higher.

There remains the question as to why real house prices have risen so dramatically over the century. It seems unlikely that productivity in house building should grow sufficiently less fast than productivity in the rest of the economy to explain the increase; Holmans (1990) shows that rising construction costs cannot account for much of the rise in the real cost of new houses. A natural answer is that the real value of the one non-renewable resource used in the construction of a new house-land-has risen rapidly. Real land values have indeed risen through the century, and at a pace considerably faster than house prices (Holmans (1990)). Over the period 1963-88 the average annual increase in the real price of new houses was 3-5 per cent while the average increase in the real price of building land was 7-5 per cent. These figures are consistent with the view that limits on the supply of building land due to planning restrictions have contributed to the rise in house prices; in no way does this imply that such restrictions are bad, merely that they have a significant impact on house prices.

Whilst deregulation of planning restrictions might cause substantial reductions in the price of houses deregulation of financial markets in the 1980s can plausibly be seen as one of the factors behind the sharp rise in prices over the decade. The availability of funds which can be borrowed using housing as collateral effects the demand for houses. Greater ability to borrow against housing wealth, and more specifically to extract equity, enhances the real value of housing because owner occupied dwellings become a fungible source of wealth. It would not be surprising if a real, and probably permanent, improvement in one of the characteristics of an asset (that is, its fungibility) should increase its value. Thus, part of the rise in the real value of houses in the 1980s may be due to the process of financial liberalisation; may prove to be permanent; and can be explained in a way quite consistent with rational behaviour, rather than speculative frenzy. If this is so the downward adjustment to the real price of houses which follows the dramatic price increases of the 1980s will be smaller than those that followed earlier booms.


Houses are important; a typical house is worth around four years of average disposable income and as a result of substantial investment in housing throughout the century there are now a lot of them about. This is a major source of national wealth which yields a stream of benefits in the form of real housing services. The real resources that need to be devoted to preserving the stock of housing in line with the population of households may be lower in the next century than it has been over the last hundred years. This would represent a real gain to national welfare. A significant proportion of the benefits of the slow accumulation of housing wealth through the century may be reaped by the current generation of home owners if they choose to pass on real bequests lower than those which they receive. The welfare implications of such a decision are unclear-the argument that is undesirable implicitly assumes that the discount rate applied by the current generation to the welfare of future generations is too high; which is just a fancy way of saying that we don't care enough about our grandchildren.


The current average value of a house is in excess of three times the average earnings of those buying houses with a mortgage (Housing Finance, table 5). CSO estimated annual depreciation on dwellings, at current replacement cost, is slightly less than 1 per cent of the market value of dwellings.

Hills' (1991) estimate of the annual flow of real housing services of 3.5 per cent of the value of the stock implies a value of around L40 billion for the benefits of home ownership in 1990-over 10 per cent of total personal disposable income. Owner occupied houses currently account for over 40 per cent of the total net worth of the personal sector. 38 per cent of the total net capital stock of the UK in 1990 was accounted for by dwellings (National Income Accounts, 1991). The ratio of the average registered annual rent of furnished tenancies (excluding housing association rents) to the average value of houses has been just under 3-5 per cent over the past three years (Housing Construction, table 1. 14). This is almost certainly a substantial under estimate of the real rental services of owner occupied houses since furnished tenancies tend to be smaller than average houses and registered rents are not free market rents. An alternative means of estimating the value of the housing services consumed by owner occupiers is to use the value of imputed rent to owner occupiers reported in the national accounts. This figure is based upon assessment of the free market rent which owner occupied houses could command. Over the period since 1979 the ratio of the CSO estimate of imputed rent to the value of the owner occupied housing stock has averaged around 2-5 per cent. However, that figure relies on estimates of rateable values made in the 1970s and is not likely to be a good guide to the real services provided by owner occupied housing. Either of these estimates-3.5 per cent or 2.5 per cent-could conceivably be only one half or one third of the true rental yield on owner occupied housing. In the main text we take a, conservative, range for that yield to be 4 per cent-7 per cent.


Ashton, P., Minford, P. and Peel, D. 1987),'The housing morass', Hobart Paper, no. 25, Institute of Economic Affairs, London. Bank of England, (1985), 'The housing finance market: recent growth in perspective',Bank of Engl and Quarterly Bulletin, March

pp. 80-91. Bank of England (1 98 9), The housing market', Bank of England Quarterly Bulletin, February, pp. 66-77. Bank of England (1991), Housing finance-an international perspective', Bank of England Quarterly Bulletin, February, pp.

56-66. Black, J. and Stafford, D. (1988), Housing Policy and finance, Routledge, London. Boleat, M. (1987), National Housing Finance Systems-a Comparative Study, Croom Helm, London. Coles, A. 1991),'Household formation projections', in Housing Finance, no. 12, November. Congdon, T. and Turnbull, P. (1982), The Coming Boom in Housing Credit, Messels and Co, June. Drake, L. (1991), The Building Society Industry in Transition, Macmillan, Basingstoke. Ermisch, J. 1991),'An ageing population, household formation and housing', Housing Studies, vol. 6, no. 4, pp. 230-240. Hamnett, C. Harmer, M. and Williams, P. (1991), Safe as Houses, Paul Chapman Publishing, London. Hills, J. (1991), Unravelling Housing Finance, Oxford University Press, Oxford. Holmans, A. (1990), House prices: changes through time at the national and sub-national level', Government Economic Service

Working Paper, no. 110. Holmans, A. (1987), Housing Policy in Britain, Croom Helm, London. Inland Revenue (1990), Inland revenue statistics, HMSO, London. International Organisations for Housing Finance Institution, (1987), International Housing Finance Factbook, London. Lee, J. and Robinson, W. (1990), The fall in the savings ratio: the role of housing', Fiscal Studies, 1 1, pp.36-52, Maclellan, D. (1990), Paper presented at the first conference of the LBS Housing Research Group, London, September. Miles, D. (1990), Financial liberalisation, the current account and housing markets', Birkbeck College discussion paper in

financial economics. Miles, D. (1992), Housing markets, consumption and financial liberalisation in the major economies', forthcoming, European

Economic Review. Morgan Grenfell (1987),'Housing inheritance and wealth', Economic Review, no. 45, November. Muellbauer, J. and Murphy, A. (1989),'Why has the UK savings ratio collapsed?', Credit Suisse First Boston research paper, july. OECD (1988), Urban Housing Finance, Paris. Office of Water Services (1991), Cost of capital: a consultative report', july. Pesaran, M. and Evans, R. (1984), 'Inflation, capital gains and UK personal savings, 1953-81', Economic journal, vol. 94,

pp.237-57. Poterba, J. and Manchester, J. (1989), Second mortgages and household savings, NBER discussion paper no. 2853. Spackman, M. 1991), Discount rates and rates of return in the public sector: economic issues', UK Treasury working paper

no. 5 8. Takagi, S. (1988), Trends of Japanese savings an d wealth and recent developments in national accounts in Japan', in The Global

Role of the Japanese Economy with Affluent Savings and Accumulated Wealth, Fifth International EPA Symposium,

Economic Planning Agency, Tokyo.
COPYRIGHT 1992 National Institute of Economic and Social Research
No portion of this article can be reproduced without the express written permission from the copyright holder.
Copyright 1992 Gale, Cengage Learning. All rights reserved.

Article Details
Printer friendly Cite/link Email Feedback
Author:Miles, David
Publication:National Institute Economic Review
Date:Feb 1, 1992
Previous Article:Productivity levels in British and German manufacturing industry.
Next Article:Productivity and skills in vehicle component manufacturers in Britain, Germany, the USA and Japan.

Terms of use | Copyright © 2017 Farlex, Inc. | Feedback | For webmasters