Submission to the Productivity Commission Inquiry on First Home Ownership 1. The Australian Housing Market: Prices, Ownership and Affordability

This chapter discusses the basic facts relating to housing prices, household indebtedness, first-home buyers and the ownership and composition of the housing stock. These facts provide the basis for the discussion in the following two chapters. The focus, wherever possible, is on those characteristics of the recent developments in Australia that are unusual by historical or international standards.^[1] The main points are as follows:

• The increase in housing prices since the mid 1990s has been unusually large, both by the standards of Australia's past and by comparison with experience abroad.
• Higher prices have been accompanied by a rapid increase in household indebtedness, driven by the growth of housing-related debt. The household debt-to-income ratio is now relatively high by international standards. This is not the case, however, for the ratio of debt to housing assets; although this ratio has risen, the extent of the rise has been held down by the cumulative increases in house prices.
• The current boom in house prices has been characterised by very strong demand by investors to purchase properties for rental purposes. This has seen borrowing by investors increase considerably faster than that by owner-occupiers and a significant increase in the share of households that own an investment property. The Australian experience in this respect has been quite different to that of other countries, where private investors have not played a comparable role in the market.
• In contrast to housing prices, rents have grown only modestly since the mid 1990s. As a result, gross rental yields have declined considerably, and are now well below historical and international norms.
• Australia has for many decades had a high owner-occupation rate. This has shown little net change over recent decades, although the owner-occupation rates for most age groups have fallen.
• The number of first-home buyers has declined over the past two years, largely explained by a surge in the number of first-home buyers in 2001.
• Measures of affordability of home ownership have declined since the mid 1990s, with house price increases more than offsetting the impact of lower interest rates on affordability.
• The stock of apartments in Australia has grown more quickly than that of detached houses over recent years, though detached houses still form the bulk of the Australian housing stock. In all countries, detached houses tend to be relatively expensive.

1.1 Housing Prices

The current focus on housing prices and affordability comes at a time when house prices in Australia have been rising rapidly over a prolonged period. In broad terms, the current upswing in house prices began in the mid 1990s, after a period of relative stability in the first half of the decade. Taking the March quarter of 1996 as a broadly representative starting point for the current upswing, the median house price has since increased at an average annual rate of 12 per cent, with prices of apartments increasing almost as quickly. In the latest two years, the trend has accelerated, so that the broad measures of house prices are currently showing annual rates of increase of close to 20 per cent.^[2] As a result of this growth, the median house price is currently around 2.3 times that in early 1996, and 2.5 times that in early 1990 (Graph 1).^[3]

A period of strongly rising house prices is not in itself unprecedented. The housing market has shown pronounced cycles in the past, including strong upswings in the early and late 1970s and a particularly sharp increase in the late 1980s. These have been interspersed by periods of greater stability, and occasionally falls, in prices. The current rate of increase in house prices, though well above average, is less than the peak rates of increase reached in some previous cycles, particularly that in the late 1980s, when the annual increase peaked at around 40 per cent (Graph 2). What is unusual about the current upswing, however, is that it has continued over a more prolonged period than has been typical in the past, and has occurred at a time when the general inflation rate has been low. As a result of these characteristics, the cumulative increase in house prices in inflation-adjusted terms has been larger than those occurring in previous cycles. For example, there was a cumulative increase in real house prices of the order of 50 per cent in the late 1980s, while in the current episode the average real house price has almost doubled (Table 1).

Another feature of the current boom in house prices is its broad geographic spread. Initially, the strong upward pressure on house prices was confined to Sydney and Melbourne. Between 1996 and 2000, the median price increased in both cities by around 13 per cent per year, on average, while in the other capitals, increases averaged below 5 per cent. However, since early 2002, upward pressure on house prices has become much more widespread. Over the past year, for example, prices have increased by at least 20 per cent in Adelaide, Brisbane, Canberra and Hobart (Graph 3). Similar price increases have been recorded in a number of regional centres.

At an aggregate level, prices in the apartment market have tended to show similar trends to those in the market for detached houses (Graph 4). However, it is likely that the broad aggregates for apartment prices overstate the average rate of increase since, over time, they are affected by a significant shift in apartment turnover towards inner-city areas where prices are relatively high. In addition, it should be noted that, at any point in time, the broad aggregate indices can mask considerable divergences in conditions across different parts of the market. Over the past couple of years, for example, there has been increasing evidence of oversupply in parts of the apartment market, particularly in the inner areas of Melbourne and possibly Sydney, where prices in the June quarter were around 4 per cent below their peaks (Graph 5). However, there is no indication that this easing of price pressures has become more widespread, and the broad city-wide indices are continuing to show rapid rates of price increase both for apartments and for detached houses.

Increases in house prices can be notionally separated into two components representing the land (or location) value and the cost of constructing a house. As is usually the case, the increases in established house prices observed over recent years have been mainly attributable to the first of these components. Since 1996, the cost of building a standard project home, for example, has increased at an average annual rate of 4.2 per cent. While higher than the general inflation rate, this rate of increase is around 8 percentage points lower than the average annual increase in the median price of established houses.

Unlike earlier episodes, the current upswing in house prices has not been associated with a broader boom encompassing the commercial property sector. In the early 1970s and late 1980s, when house prices rose quickly, so did prices of commercial property. Indeed, in the late 1980s, the rate of increase in commercial property prices far outstripped that in housing prices. In contrast, since the mid 1990s, prices of commercial office buildings have increased at an average annual rate of just over 4 per cent, considerably below the rate of increase in housing prices. The price indices for commercial office buildings remain below their late-1980s peaks in all capital cities, with the recent gains not yet having made up for the large fall in prices during the first half of the 1990s (Graph 6).

Across countries, movements in house prices have shown marked divergences over recent years (Graph 7). At the low end of the range are a number of countries where house prices have been flat or declining. In Japan, for example, prices at the national level have declined by around 30 per cent since their peak in 1991, and considerably more in the large cities. Prices have also fallen significantly in Hong Kong and Singapore in recent years. In Germany and Switzerland, prices are broadly unchanged from 1995 levels. At the other end of the range, a number of countries have experienced increases in house prices broadly similar to those in Australia. For example, in the United Kingdom and the Netherlands residential property prices have increased at an average annual rate of more than 10 per cent since 1995, although the rate of increase has recently moderated considerably in the Netherlands.

The rate of growth of house prices in Australia over recent years has significantly exceeded growth of household incomes, so that the ratio of house prices to incomes has shown a strong upward trend since the mid 1990s (Graph 8). This ratio has, in the past few years, risen to levels not previously seen in Australia. On the latest available REIA data, for the June quarter 2003, the median house price was equivalent to around 6 times the average annual household income, compared with a ratio of just over 4 times income at the peak of the late 1980s boom. Using the measure of prices based on transactions financed by the Commonwealth Bank, the increase is even larger. Some international comparisons of housing asset values relative to incomes are presented in Table 2.^[4] While it is difficult to ensure that these data are compiled on a fully comparable basis, the available evidence suggests that the ratio of house prices to incomes in Australia is now relatively high by international standards, whereas a decade ago it had been similar to that observed in a number of other countries.^[5]

1.2 Household Borrowing

The rapid increase in housing prices in Australia has been accompanied by strong growth in borrowing by the household sector. Since the beginning of 1996, household debt has increased at an average annual rate of 14½ per cent, and has accelerated more recently to a pace of around 20 per cent over the past year. Most of this increase has been in loans for the purchase of housing. This component of household debt now accounts for 84 per cent of the total debt of the household sector, up by around 15 percentage points since 1990. As discussed below, housing-related debt can in turn be separated into borrowing for owner-occupation and borrowing for investor housing. While borrowing for owner-occupation is still the larger of these components, the investor component has been growing much more quickly over recent years (Graph 9).

The growth in household debt in Australia has been unusually fast by international standards. Among other advanced economies, only the Netherlands and Spain have recorded comparably rapid credit growth in the period since 1995. In other advanced countries, credit growth rates of the order of 5 to 8 per cent per year have been typical over this period (Table 3).

The rapid growth of household debt in Australia has resulted in a strong upward trend in the ratio of debt to income. In the 1980s and early 1990s, this ratio was low by international standards but, after a decade of significantly stronger-than-average debt accumulation, it is now at the top end of the range seen in most other countries (Graph 10). In the few countries where the ratio is significantly higher, particularly the Netherlands, this can be largely explained by tax arrangements, in the form of tax-deductibility of home mortgage interest payments for owner-occupiers, that discourage the repayment of that type of housing debt. As a result of such arrangements, households in these countries tend to use funds that would otherwise have been used to repay a housing loan to build up their financial assets. In contrast, in Australia, the tax system creates an incentive for households to repay owner-occupier housing loans quickly (this same incentive does not apply to investor loans).

The upward trend in the debt-to-income ratio has meant that the debt-servicing ratio – the ratio of interest payments to disposable income – has also trended upward over recent years (Graph 11). Mortgage interest costs now represent 6½ per cent of aggregate household disposable income, a level that exceeds its peak of around 5¾ per cent in the 1990s and is still increasing as mortgage debt continues to rise more quickly than incomes.^[6] The total interest costs of the household sector (i.e. including interest on other forms of household borrowing) now stand close to 8 per cent of household income. This ratio is still about a percentage point below the peak reached at the end of the 1980s boom, a point when interest rates were at much higher levels than they are at present and when a greater proportion of household debt was in higher-cost forms of personal loans rather than mortgages. Nonetheless, on current trends, the debt-servicing ratio will continue to rise. For example, assuming a continuation of the current rates of growth in household debt and incomes, the overall debt-servicing ratio would exceed its previous peak sometime in 2004.

These measures of the debt-servicing ratio represent averages across all households, many of whom do not have a mortgage. Hence, the debt-servicing ratios for the subset of households that have a mortgage are considerably higher. Around 30 per cent of Australian households have a housing loan, a figure which has been broadly unchanged for the past two decades. For these households, the total servicing payment (interest plus required payment of principal) averages 20 per cent of disposable income. Again, there is considerable variation within this group. Among households with a mortgage, it is the relatively young households, those with lower than average incomes and, more generally, those with recently acquired mortgages, that tend to have relatively high ratios of interest payments to incomes. Given the relatively constant proportion of households with a mortgage, the rise in the aggregate debt-to-income ratio is accounted for by an increase in the average debt level of those households with a mortgage. Much of this additional debt appears to have been taken on by mid-life households with relatively high incomes.^[7]

A consequence of rising house prices and the associated growth of debt has been that the increase in borrowing secured against housing has exceeded net new spending on housing assets. This implies that households, in aggregate, have been drawing on their accumulated equity in the housing stock to release funds for other purposes. This housing equity withdrawal has amounted to an average of around 4½ per cent of household disposable income since mid 2001; prior to this, the usual pattern was for the household sector to inject equity into the housing stock (Graph 12).^[8] The shift towards equity withdrawal over the past few years has been facilitated by the combination of rising house prices, which have increased the collateral available to the household sector, and the development of new lending products, such as home-equity loans and loans with redraw facilities, which have enabled households to borrow more easily against the equity in their homes (see Chapter 2).

The increase in housing-related debt has also meant an increase in the average housing gearing ratio – the ratio of housing debt to the value of housing assets. This ratio has approximately doubled since the 1980s, to be currently around 20 per cent (Graph 13), though it remains relatively low by international standards. Much of the increase in gearing in Australia occurred in the early to mid 1990s. In more recent years, although debt has been growing rapidly, the gearing ratio has been relatively constant, due to the similarly rapid increase in house prices. Obviously many individual households – those that have only recently acquired mortgages – would have much higher gearing than the average.

1.3 Investors in the Housing Market

Perhaps the most important distinguishing feature of the current housing price boom has been the very strong demand by household investors for the purchase of residential properties to rent. The extent of this demand is unprecedented, both in terms of previous experience in Australia and experience overseas. The prominent role of investors in the Australian housing market can be seen both in the high and rising share of housing finance going to investors, and in the relatively high proportion of households in Australia owning rental properties.

Since 1996, the value of investor housing loans outstanding has grown at an average annual rate of 22 per cent, and the pace has accelerated in more recent years to be currently around 33 per cent. Investor loans now account for around one-third of banks' outstanding housing loans, up from around 15 per cent at the beginning of the 1990s (Graph 14). Investors account for an even larger share of new loans approved. Since mid 2002, around 40 cents of every dollar of new housing loans approved by financial institutions has been for investment properties. The high proportion of housing finance accounted for by investors in the Australian housing market stands out as quite different from the experience of other countries. In most countries, the percentage of housing loans accounted for by investors is estimated to be only in single figures. In the United Kingdom, for example, gross lending for the ‘buy-to-let’ market rose to a peak of around 6½ per cent of gross mortgage lending over the first half of 2003. Comparable data for other countries are difficult to obtain, partly reflecting the fact that, in these countries, individual investors play only a small role in the overall rental market, and hence their activities are not systematically recorded by statistical agencies as they are in Australia. Notwithstanding the lack of official figures, financial and housing authorities in a range of countries report that the role of household investors in their housing markets is relatively minor.

The strong investor demand for residential properties has seen the number of Australian households owning an investment property increase significantly since the early 1990s. Surveys suggest that the share of households with an investment property has risen from around 8 per cent in the early 1990s to around 12 per cent in 2001 (Graph 15).^[9] Ownership rates appear to have risen for almost all income groups, although the increases have been largest for households in the upper part of the income distribution (Graph 16). Similarly, according to the Australian Taxation Office (ATO), there has been a significant rise in the proportion of taxpayers reporting rental income. In the 2000/01 financial year, around 13 per cent of taxpayers, or 1.3 million individuals, reported earning rental income, compared with a figure of around 9 per cent in the early 1990s.

The proportion of Australian households owning an investment property is considerably higher than that in most other countries. In the United States, for example, taxation data indicate that around 6½ per cent of individuals earn rental income, and both tax and survey data indicate that this proportion has been falling over time.^[10] In Canada, ownership rates appear to be broadly similar to those in the United States. In the United Kingdom, less than 2 per cent of households own a rental property, although the rate of ownership has been increasing recently.^[11] In these and other countries, a larger proportion of the rental stock is typically owned by institutions, government agencies and charities. In part, the low rate of individual ownership in other countries reflects the history of rent controls and other rigidities in the rental market.

1.4 Rents and Vacancy Rates

In contrast to the large gains in housing prices, rents have increased only modestly over recent years, and the rate of increase has recently been declining. Since the March quarter of 1996, the Consumer Price Index measure of rents has increased at an average annual rate of 2.8 per cent, only slightly faster than the overall rate of inflation (Graph 17). Over the past year, this measure has increased by 1.9 per cent. An alternative measure of rents constructed from city-based data published by the REIA shows slightly faster average growth, but broadly in line with the growth in average household disposable income over this period.

As with housing prices, growth in rents was strongest in Sydney and Melbourne over much of the second half of the 1990s, with vacancy rates having been at relatively low levels in the middle part of the decade. More recently, however, rents have been subdued in both cities, and under downward pressure in some localities, reflecting higher vacancy rates as the increased supply of apartments comes on stream. In contrast, rents have been increasing more quickly in Canberra, Brisbane and Adelaide, with low vacancy rates in these cities still indicating a relatively tight rental market.

The combination of a rapid increase in housing prices and low increases in rents has meant that average rental yields have fallen to very low levels (Graph 18). In the mid 1980s, it was not uncommon for gross rental yields to exceed 8 per cent, with rents being under strong upward pressure due to very low vacancy rates. Rental yields began to fall during the housing boom of the late 1980s and have continued their downward trend since. Currently, gross rental yields are typically around 3–3½ per cent for houses and a little higher for apartments. After taking into account costs such as council rates, strata levies, management fees, repairs and maintenance, net rental yields are at least a percentage point lower.

Rental yields in Australia are very low by international standards. In the United Kingdom the average gross rental yield is currently estimated to be around 7½ per cent, down from around 9½ per cent at the start of 2002.^[12] Rental yields in the United States tend to be around the same level as in the United Kingdom, while in Canada, it is not uncommon for rental yields to be as high as 12 per cent.

1.5 Owner-occupation Rates

Owner-occupation rates in Australia have traditionally been high by international standards, remaining relatively stable over the past few decades at around 70 per cent. Even so, ownership rates have fallen for most age groups, with the falls being most noticeable for younger households. In particular, ownership rates for households in the 25–29 and 30–34 year age groups have declined by a little less than 10 percentage points since the early 1980s. In contrast, ownership rates for households in the 45–65 year age group have changed little in net terms over this period. The fact that the overall ownership rate has remained relatively steady, while the rates for most age groups have declined, is explained by an increase in the share of older households as the population ages, with these households having higher ownership rates (Graph 19).

The fall in the ownership rates for younger age groups partly reflects an increase in the number of single-person households – as the average age of marriage has increased and the number of university students has risen – and a rise in the number of single-parent households. The ownership rate for couples with children has remained broadly unchanged since the early 1980s, at around 80 per cent.

1.6 First-home Buyers

According to data on loan approvals, there have been around 110,000 first-home buyers purchasing a house with debt, on average, each year since the mid 1990s. The total number of first-home buyers (including those who did not borrow) peaked in 2001 at 187,000 following the introduction of the First Home Owner Grant (FHOG) (Graph 20). This peak partly reflects delays in purchasing by some households until after the FHOG was available, as well as the bringing forward of purchases made possible by the grant. The introduction of the Commonwealth Additional Grant (CAG), which operated between March 2001 and June 2002 would also have contributed to the bringing forward of some first homeowner purchases into the qualifying period. At a quarterly rate, the number of first-home buyers peaked at around 55,000 in the December quarter of 2001. More recently the unwinding of these effects has contributed to a decline in the number of first-home buyers to quarterly levels of around 30,000, or a little below the medium-term average. This decline has occurred at a time when financing for repeat buyers has been growing very rapidly and, as a result, the share of first-home buyers in total finance approvals for owner-occupied housing (excluding refinancing) has fallen sharply (Graph 21).

There are no comprehensive time-series data on the prices paid by first-home buyers, although data are available on the average size of the loans they take out to fund the purchase. Assuming that the ratio of the loan size to the price paid has remained unchanged, these data suggest that the average price paid by first-home buyers has increased by somewhat less than the median price of all houses over recent years (Graph 22). Since the March quarter 1996, for example, the average loan size of first-home buyers has risen by 100 per cent while, as discussed above, the median house price has risen by 130 per cent. Thus, first-home buyers are buying homes that are further below the median price of all homes than was formerly the case.

The data on the average size of loans taken out by first-home buyers are broadly consistent with information on prices paid for properties purchased under the FHOG scheme. Data on FHOGs paid also confirm that the average price paid by first-home buyers is lower than the median metropolitan house price. This partly reflects the inclusion of apartments in the FHOG data, as well as the fact that purchases by first-home buyers tend to be in relatively less expensive areas.

1.7 Affordability

There is no single universally applicable measure of affordability.

The most commonly cited measure is the ratio of average household income to the income required to meet debt repayments on a typical house. As the cost of debt finance, or the size of loans, falls, the denominator in this ratio also falls, and housing is said to be more affordable.

There are a number of measures of this type. A prominent one is that jointly produced by the Commonwealth Bank of Australia (CBA) and the Housing Industry Association (HIA). It is constructed to measure the affordability of the median-priced established dwelling purchased by first-home buyers.^[13] Similar measures can be constructed using other house prices series.

These measures show that affordability has declined over recent years as housing prices have increased (Graph 23). Comparing the latest period with the early 1990s, the level of housing interest rates is now significantly lower, but the impact of this on affordability has been more than offset by the cumulative increase in house prices relative to incomes. The current level of affordability is still above the low point reached at the end of the 1980s boom, a period when interest rates were exceptionally high and the ratio of house prices to incomes was also at a peak. The affordability measure produced by CBA/HIA has fallen by more than other measures over recent years, reflecting the faster rate of increase in the CBA measure of house prices.

The fact that the average price of houses purchased by first-home buyers has increased less quickly that the average price of all houses suggests that an affordability index for first-home buyers would not have declined as much as the indices in Graph 23. There are two interpretations of this result. The first is that, as prices have risen, first-home buyers have had to purchase properties that are less expensive, relative to the median, than was the case in the past. This could be occurring through the purchase of apartments, rather than houses, or the purchase of houses in less desirable areas than was previously the case. The second is that, first-home buyers are still purchasing the type of properties that they always have, but that the prices of these properties have not increased as quickly as overall prices. Unfortunately, the data available to the Bank does not allow us to distinguish between these two explanations.

An alternative measure of affordability is based on the size of the deposit required to purchase a given home; in particular the ratio of the average required deposit to household income. Assuming a fixed loan-to-valuation ratio, this measure is directly determined by the ratio of house prices to income. As discussed above, this ratio has increased considerably over recent years and, as a result, households now need to save a larger amount relative to their income than was previously the case before they can purchase a home. For example, in 1990, a 10 per cent deposit on the median-priced house was equivalent to around 25 per cent of average annual household income (Graph 24). Today, the figure is around 45 per cent.^[14] The increase is even more significant if stamp duty is taken into account, reflecting the fact that the average rate of stamp duty has risen as nominal prices have increased. In 1990, stamp duty (calculated at NSW rates) on the median-priced house in Australia was equivalent to 6 per cent of annual income. Today, the figure is around 14 per cent. For many people who may be able to meet the ongoing repayment burden, these upfront costs have become a significant constraint on their ability to purchase a home.

For first-home buyers the up-front deposit constraint has been eased by the First Home Owner Grant. Taking account of this grant reduces the savings required to fund a deposit on the median-priced house from around 45 per cent of average household income to around 35 per cent. Given that many first-home buyers purchase houses with prices lower than the median, the FHOG has made a significant contribution to easing the deposit constraint.

1.8 The Housing Stock

The Australian housing stock has two particularly distinguishing characteristics relative to that in other countries. The first is the high proportion of detached houses and the second is the high proportion of the housing stock located in major cities. Overall, the quality of the housing stock appears to be broadly comparable with that in a number of major industrialised countries. Although the average size of dwellings is larger than in some European countries, this partly reflects the larger average size of households in Australia.

Around three-quarters of the housing stock in Australia is accounted for by detached houses. The comparable figure for the United States is 60 per cent and for the United Kingdom it is 25 per cent (Table 4).^[15] While data on the average amount of land used by each dwelling are not available, the above figures suggest that, on average, dwellings in Australia take up more land than is typically the case elsewhere. The large geographical size of Australian cities, relative to their populations, provides further support for this conclusion.

Housing in Australia not only appears to take up more land, but is more heavily concentrated in major cities than is the case elsewhere. Fifty-five per cent of the urban population of Australia live in either Sydney or Melbourne, the two largest cities. This level of concentration is well above that in most countries. A further 30 per cent live in the other capital cities and unlike most other countries, Australia has no middle-sized cities, defined according to the UN as having between 500,000 and 1 million inhabitants. Since detached houses, and houses in large cities, tend to have relatively high prices in all countries, the preponderance of these characteristics in the Australian housing stock would provide one reason why the level of housing prices in Australia would tend to be higher than in other countries. However, this would not explain the continuing high rate of increase in Australian housing prices over recent years.

While Australians live mainly in detached houses, over recent years the stock of housing in multi-unit developments has grown at more than twice the rate of the stock of detached houses (Graph 25). This difference in growth rates is particularly noticeable in Sydney where, since 1996, almost two-thirds of new housing construction has been in medium or high-density developments. If this were to continue, as is widely expected, then the composition of the housing stock will change considerably over time. For example, within 20 years, the share of multi-unit dwellings in Sydney would increase from its current share of around 37 per cent, to above 45 per cent. Similar trends are evident in other cities, although the shares of multi-unit dwellings are considerably lower.

Footnotes

In preparing this submission, discussions were held with central banks and housing authorities in the United States, Canada, the United Kingdom and the Netherlands. The range of sources consulted is listed in the Appendix. [1]

Measures of housing prices in Australia are published by the Australian Bureau of Statistics (ABS), the Real Estate Institute of Australia (REIA) and jointly by the Commonwealth Bank of Australia (CBA) and the Housing Industry Association (HIA). All three measures show broadly similar movements through time, although the CBA/HIA measure of established house prices has tended to increase more quickly than the other measures. In this submission, we use the REIA measure as, unlike the ABS measure, it is available in value terms and not just as an index. The REIA also publishes separate measures for houses and apartments, while the ABS publishes only a house price index. Unless otherwise indicated, house price data presented in the statistical material in this submission are for established detached houses. [2]

All price indices are constructed from sales prices. The median selling price is less than the average selling price reflecting a relatively small number of very highly priced properties. [3]

The table shows the aggregate value of housing assets owned by the household sector in each country, expressed as a ratio to aggregate household disposable income (including income of unincorporated enterprises). Strictly speaking, this is not a direct measure of the ratio of house prices to incomes: while differences between these ratios across countries will primarily reflect relative housing prices, they will also reflect differences in the proportion of the housing stock owned by households, as opposed to public and institutional ownership. [4]

For further discussion of this evidence, see Ellis L and D Andrews (2001), ‘City Sizes, Housing Costs, and Wealth’, Reserve Bank of Australia Research Discussion Paper 2001-08. [5]

Repayment of principal is estimated to amount to a further 2½ per cent of income. [6]

For more details, see Ellis L, J Lawson and L Roberts-Thomson (2003), ‘Housing Leverage in Australia’, Reserve Bank of Australia Research Discussion Paper 2003-09. [7]

See Reserve Bank of Australia Bulletin, ‘Housing Equity Withdrawal’, February 2003, pp 50–54. [8]

The data for the early 1990s are from the 1993/94 Household Expenditure Survey, while the more recent data are from the 2001 HILDA Survey. [9]

The 2000 Survey of Income and Program Participation gives an alternative estimate of 5 per cent for households rather than taxpayers. [10]

This estimate is from the British Household Panel Survey 2000. [11]

This estimate is from the Paragon Mortgages Buy-to-Let Index. [12]

Until the December quarter 1987, this measure uses an estimate of the median dwelling price based on loans taken out with the CBA by first-home buyers. From 1988, with separate information on loans to first-home buyers no longer available, movements in the series are based on movements in the median price paid by all home buyers who financed with CBA. [13]

Household income used in this calculation is based on the national accounts measure but excludes imputed income on owner-occupied dwellings. [14]

For further details see Ellis L and D Andrews (2001), ‘City Sizes, Housing Costs, and Wealth’, Reserve Bank of Australia Research Discussion Paper 2001-08. [15]