Abstract
Property price inflation has accelerated during the last few years, propelled by low mortgage rates, poor returns from alternative investments, strong employment and immigration, and tax incentives. Strong housing demand has led to a substantial rise in housing prices. Most industrial countries face significant fiscal pressures over the longer term associated with population aging and rising health care costs. Significant uncertainty surrounds projections of future fiscal costs. One comprehensive way to take account of this uncertainty for formulating policy decisions is through stochastic simulation.
I. The Determinants of Property prices in Australia1
1. Property price inflation has accelerated during the last few years, propelled by low mortgage rates, poor returns from alternative investments, strong employment and immigration, and tax incentives (Table 1)2. Low interest rates and the First Home Owners Scheme reduced the cost of housing for owners. Investor demand has also driven housing prices, particularly in the central business districts of the major cities, as increased expected returns on housing far outstripped returns on other assets. Consequently, investor housing accounted for 30 percent of the stock of housing loans by the end of 2002, compared to 18 percent a decade ago.
Australia: New and Established House Price Index during 1998:4-2002:4
(in percentage change)
Australia: New and Established House Price Index during 1998:4-2002:4
(in percentage change)
| 1Y | 2Y | 3Y | 5Y | ||||||
|---|---|---|---|---|---|---|---|---|---|
| Nominal | Real | Nominal | Real | Nominal | Real | Nominal | Real | ||
| Brisbane | 16.1 | 13.7 | 37.1 | 30.4 | 45.1 | 30.2 | 48.8 | 28.3 | |
| Melbourne | 11.6 | 9.3 | 32.0 | 25.6 | 46.6 | 31.6 | 848 | 59.3 | |
| Sydney | 17.0 | 14.6 | 40.5 | 33.6 | 47.1 | 32.0 | 75.9 | 51.7 | |
| Average of 8 Capital Cities | 14.1 | 11.7 | 33.9 | 27.4 | 43.2 | 28.5 | 65.9 | 43.1 | |
Australia: New and Established House Price Index during 1998:4-2002:4
(in percentage change)
| 1Y | 2Y | 3Y | 5Y | ||||||
|---|---|---|---|---|---|---|---|---|---|
| Nominal | Real | Nominal | Real | Nominal | Real | Nominal | Real | ||
| Brisbane | 16.1 | 13.7 | 37.1 | 30.4 | 45.1 | 30.2 | 48.8 | 28.3 | |
| Melbourne | 11.6 | 9.3 | 32.0 | 25.6 | 46.6 | 31.6 | 848 | 59.3 | |
| Sydney | 17.0 | 14.6 | 40.5 | 33.6 | 47.1 | 32.0 | 75.9 | 51.7 | |
| Average of 8 Capital Cities | 14.1 | 11.7 | 33.9 | 27.4 | 43.2 | 28.5 | 65.9 | 43.1 | |
Median Price of Houses and Units
(1980:1=100)
Citation: IMF Staff Country Reports 2003, 336; 10.5089/9781451801996.002.A001
2. The recent run-up in housing prices is largely explained by economic fundamentals; however, further changes in fundamentals would not appear to support additional sharp increases in prices. The estimation results of an econometric model show that real Australian housing prices over the last two decades are well explained by movements in real mortgage interest rates, real disposable income, real returns on equities, and population growth. In particular, the real mortgage rate decline of 4½ percentage points during the last five years alone accounted for about 30 percent of the increase in real property prices during the period. Mortgage rates are not expected to decline significantly in the period ahead, nor should expected changes in other fundamental determinants contribute to a further sharp rise in housing prices.
A. Factors Influencing Housing Demand
3. Financial innovation and increased competition in the mortgage industry have lowered the cost of borrowing and enhanced household access to credit. In the early 1990s, financial institutions were keen to expand their portfolios of relatively high-return, low-risk housing loans, especially considering the losses that they had incurred on their corporate loan portfolios. This interest in housing finance, as well as strong competition from new entrants, reduced borrowing costs and spurred innovation. The development of the mortgage brokering industry over the past couple of years has added further to the competitive environment for housing loans. The brokers have also substantially lowered entry costs for new lenders, by reducing the need for extensive branch networks and advertising.
4. Nominal and real interest rates have declined significantly since 1990. The standard variable mortgage rate has averaged 6.8 percent over the past five years, compared to 10.8 percent in the period 1990–97. From a peak of 17 percent in early 1990, the mortgage rate has fallen to 6.6 percent in March 2003. This decline largely reflects the decline in inflation over the period. The mortgage rate has declined in real terms from an average of 8.4 percent in 1990–97 to an average of around 4 percent during the last five years.3 The substantial decline in interest rates has significantly increased households’ borrowing capacity.
Nominal and Real Mortgage Rates
(In percent)
Citation: IMF Staff Country Reports 2003, 336; 10.5089/9781451801996.002.A001
Annual Growth in Real Credit Growth
(In percent)
Citation: IMF Staff Country Reports 2003, 336; 10.5089/9781451801996.002.A001
5. Household mortgage credit has grown at a rapid pace during the last five years. Real credit grew at an average of 12½ percent per year during the last five years compared to 11 percent in 1990–97. However, the pace of growth has accelerated to reach more than 17 percent by the first quarter of 2003. Considering the relatively slow adjustment in the supply of housing, the rapid growth in credit (reflecting housing demand), has translated into higher housing prices. Some of the increase in housing prices reflects quality improvements, but a large part of the increase in housing prices probably reflects pure inflation.
6. Strong immigration and falling household size have boosted housing demand. Permanent and long term net immigration flows have averaged close to 125,000 persons a year during the last five years, more than 40 percent higher than the average over the 1990–97 period. Strong immigration flows, in conjunction with falling household size, boosted demand for housing.4 This has been particularly true for capital cities such as Sydney, Melbourne, and Brisbane where housing prices have increased the most.
Population Growth and Net Arrivals
(In thousands)
Citation: IMF Staff Country Reports 2003, 336; 10.5089/9781451801996.002.A001
Annual Growth in Real Disposable Income and Unemployment Rate
(In percent)
Citation: IMF Staff Country Reports 2003, 336; 10.5089/9781451801996.002.A001
Annual Real Returns on Equities and Housing Units
(In percent)
Citation: IMF Staff Country Reports 2003, 336; 10.5089/9781451801996.002.A001
7. Strong growth in disposable income and a declining unemployment rate have also contributed to sustaining the boom in the housing market. Real disposable income grew by 3.3 percent a year during the last five years compared to 2.1 percent in 1990–97.5 The unemployment rate has declined from almost 11 percent in the early 1990s to about 6 percent by the end of 2002.
8. Poor returns on alternative investments have also contributed to the rapid rise in housing prices, especially in the investor segment of the housing market. Average yearly returns on equities was less than one percent during the last five years compared to 7½ percent for housing investments.6 In addition, housing investments benefited from the halving of the capital gains tax rate in 2000.
B. An Econometric Model for the Relative Price of Housing
9. Inherent features of the real estate market—in particular, supply rigidities and imperfect information—combined with the procyclical nature of bank lending contribute to making the housing market vulnerable to periods in which actual prices may deviate from their fundamental value.7 Since the price of housing depends on the future value of fundamentals, investors may either underestimate or overestimate the fundamental price in an environment with imperfect information. In particular, housing purchasers may become overly optimistic about expected capital gains, driving the price above its replacement cost. In efficient financial markets, these deviations would rapidly be eliminated by a relatively quick adjustment of supply and demand conditions. However, in the housing market, optimistic housing purchasers will remain in the market as long as prices are rising and financing is available, owing to the slow supply response, and the lack of natures and options markets for housing prevent a quick adjustment to equilibrium. Finally, as prices move farther and farther away from their fundamental value, more and more investors would eventually move to the sell side, dampening price inflation. As this process gathers momentum, prices could drop abruptly.
10. For owner-occupied housing, the demand for housing can be specified as a function of real disposable income, the cost of borrowing, and demographic changes. In the case of investors, the demand for housing can be specified as depending on the cost of borrowing, real disposable income, expected capital gains in the housing market, and expected yields on alternative investments. The effect of the latter variable is uncertain given the presence of substitution and income effects.8 Another potential explanatory variable is housing credit growth, but it is not possible to distinguish between an increase in the demand for credit driven by a decline in mortgage rates and an increase reflecting greater availability of credit for housing.
11. Table 2 provides the Augmented Dickey-Fuller unit root test for all the variables included in the two housing demand equations that are estimated. Generally, variables fail to reject the unit-root hypothesis at conventional significance levels. Table 3 reports the estimation results for the housing demand equations.9 For the owner-occupied segment of the housing market, the log of the relative price of houses (P_h)—where the latter is the median price for new and established houses divided by the consumer price index (CPI)—is expressed as a function of the expected real mortgage rate (r), the log of real disposable income (Yd), population growth (POP), and the lag of the dependent variable in order to capture persistence in the relative price of houses.10 Similarly, for the investor segment of the housing market, the log of the relative price of units and flats (P_u)—where the latter is the median price for new and established units and flats divided by the CPI—is expressed as a function of the expected real return on units and flats, expected real return on equities, the expected real mortgage rate, real disposable income, and the lag of the dependent variable in order to capture persistence in the relative price of units and flats. As is customary in the literature, the expected future value of a variable is computed as its moving average.
Australia: The Dickey Fuller-GLS Unit Root Test
Australia: The Dickey Fuller-GLS Unit Root Test
| Variables | DF-GLS Statistic |
|---|---|
| P_h | -2.22a |
| P_u | -1.00a |
| H_r | -1.93a |
| M_r | -1.75a |
| Yd | -2.35a |
| E_r | -2.58a |
| POP | -1.73a |
| N | 82 |
Australia: The Dickey Fuller-GLS Unit Root Test
| Variables | DF-GLS Statistic |
|---|---|
| P_h | -2.22a |
| P_u | -1.00a |
| H_r | -1.93a |
| M_r | -1.75a |
| Yd | -2.35a |
| E_r | -2.58a |
| POP | -1.73a |
| N | 82 |
Australia: Determinants of Median Price for New and Established Houses and Units Deflated by CPI
Australia: Determinants of Median Price for New and Established Houses and Units Deflated by CPI
| Independent Variables | Dependent Variable | |
|---|---|---|
| P_h | P_u | |
| Constant | -2.502 (-4.69)a |
-3.364 (-8.15)a |
| Lagged Dependent Variable | 0.666 (8.72)a |
0.324 (4.29)a |
| H_r | 0.218 (2.95)a |
|
| -Mr | -0.376 (-2.33)b |
-0.678 (-4.09)a |
| D95*M_r | -0.569 (-3.32)a |
-1.157 (-6.85)a |
| Yd | 0.554 (4.75)a |
0.981 (9.5 l)a |
| E_r | -0.166 (-4.57)a |
|
| POP | 9.128 (2.82)a |
|
| Long-Elasticity for M_r (after 1995) | 2.83 | 2.71 |
| Adjusted R2 | 0.977 | 0.985 |
| Shin’s Cointegration Statistic | 0.025 | 0.019 |
| N | 77 | 77 |
Australia: Determinants of Median Price for New and Established Houses and Units Deflated by CPI
| Independent Variables | Dependent Variable | |
|---|---|---|
| P_h | P_u | |
| Constant | -2.502 (-4.69)a |
-3.364 (-8.15)a |
| Lagged Dependent Variable | 0.666 (8.72)a |
0.324 (4.29)a |
| H_r | 0.218 (2.95)a |
|
| -Mr | -0.376 (-2.33)b |
-0.678 (-4.09)a |
| D95*M_r | -0.569 (-3.32)a |
-1.157 (-6.85)a |
| Yd | 0.554 (4.75)a |
0.981 (9.5 l)a |
| E_r | -0.166 (-4.57)a |
|
| POP | 9.128 (2.82)a |
|
| Long-Elasticity for M_r (after 1995) | 2.83 | 2.71 |
| Adjusted R2 | 0.977 | 0.985 |
| Shin’s Cointegration Statistic | 0.025 | 0.019 |
| N | 77 | 77 |
12. In the equation for owner-occupied housing, all variables have the expected sign and are highly statistically significant, and the estimated equation fits the data quite well (Figure 1). The relationship between the real mortgage interest rate and the relative price of housing is nonlinear. This relationship strengthened significantly after 1995 with the short-term semi-elasticity between these two variables increasing from 0.4 to 1.1 after 1995.11 The corresponding long-run semi-elasticity after 1995 is 2.8. This implies that the decline in real mortgage interest rates of 4½ percentage points during the last five years accounted for about 12¾ percentage points increase or 30 percent of the total increase in the relative price of housing during the period. As expected, demand for housing is strongly procyclical. An increase in real disposable income of 1 percent would increase the relative price of housing by 0.6 percent on impact and by 1⅔ percent in the long run. Population growth also has a powerful effect on the relative price of housing. A one percentage point increase in population growth is associated with an increase in the relative price of housing of more than 9 percent on impact and by 27 percent in the long run.
Australia: Median Price for New and Established Houses Deflated by the CPI (in log)
Citation: IMF Staff Country Reports 2003, 336; 10.5089/9781451801996.002.A001
13. In the equation for the relative price of housing in the investor segment of the market, all variables have the expected sign and are highly statistically significant, and the estimated equation fits the data quite well (Figure 2). Higher expected real returns in the housing market tend to increase demand (for a given supply of housing), which leads to an increase in the relative price of housing. An increase of 10 percentage points in real returns on housing investments will increase the relative price of housing by about 3¼ percent in the long run. The long-run effect of mortgage interest rates on housing investments is very similar to that on the owner-occupied housing. The long-run semi-elasticity for investment housing is 2.7 versus 2.8 for owner-occupied housing. An increase in expected real returns on equities (holding real returns on housing constant) exerts a negative effect on the relative price of housing which implies that the substitution effect largely dominates the income effect. A decline in real returns on equities would lead investors to move out of equities (or at least redirect new cash out of equities) and into housing, bidding up the relative price of housing. A decline of 10 percentage points in real returns on equities would induce an increase in the relative price of housing of 1.7 percent on impact and 2.5 percent after prices have adjusted fully. Interestingly, the adjustment of relative prices of housing to a shock is faster in the investor than in the owner-occupied segments of the housing markets.
Australia: Median Price for New and Established Units Deflated by the CPI (in log)
Citation: IMF Staff Country Reports 2003, 336; 10.5089/9781451801996.002.A001
14. These estimated equations can be used to assess the degree of divergence between the actual relative price of housing and the relative price of housing that is implied by fundamental factors. The relative price of housing for both the owner-occupied and investor segments of the market are broadly in line with their estimated values, deviating by only 2½ percent in the first quarter of 2003.
References
Collyns, Charles, and Abdelhak Senhadji, 2003, “Lending Booms, Real Estate Bubbles, and the Asian Crisis”, in Asset Price Bubbles: The Implications for Monetary, Regulatory, and International Policies, William Hunter, George Kaufman, and Michael Pomerleano, eds., Cambridge, Massachusetts: The MIT Press.
Dvornak, Nikola and Marion Kohler, 2003, “Housing Wealth, Stock Market Wealth and Consumption: A Panel Analysis for Australia”, Reserve Bank of Australia Research Discussion Paper No. 2003–07, Reserve Bank of Australia.
Ellis, Luci, Jeremy Lawson, and Laura Roberts-Thomson, 2003, “Housing Leverage in Australia”, Reserve Bank of Australia Research Discussion Paper No. 2003–09, Reserve Bank of Australia.
Ellis, Luci, Jeremy Lawson, and Laura Roberts-Thomson, and Dan Andrews, 2001, “City Sizes, Housing Costs, and Wealth”, Reserve Bank of Australia Research Discussion Paper No. 2001–08, Reserve Bank of Australia.
International Monetary Fund, 2003, “When Bubbles Burst”, World Economic Outlook, April.
Kiyotaki, Nobuhiro, and John Moore, 1997, “Credit Cycles”, Journal of Political Economy, Vol. 105, No. 2, 211–47.
La Cava, Gianni, and John Simon, 2003, “A Tale of Two Surveys: Household Debt and Financial Constraints in Australia”, Research Discussion Paper No. 2003–8, Reserve Bank of Australia.
Learner, Edward, 2002, “Bubble Trouble? Your Home Has a P/E Ratio Too”, UCLA Anderson Forecast, June.
Macfarlane, Ian, 2003, “Do Australian Households Borrow Too Much?”, Talk to the Sydney Institute, Sydney, April 3.
Muellbauer, John and Anthony Murphy, 1997, “Booms and Busts in the UK Housing Market”, Economic Journal, Vol. 107, No. 445, 1701–27.
Mühleisen, Martin and Martin Kaufman, 2003, “Are U.S. House Prices Overvalued?” IMF Country Report No. 03/245, (Washington: International Monetary Fund).
Phillips, Peter and Bruce Hansen, 1990, “Statistical Inference in Instrumental Variables Regression with 1(1) Processes”, Review of Economic Studies, Vol. 57, pp. 99–125.
Reserve Bank of Australia, 2002, “Innovation in the Provision of Finance for Investor Housing”, Reserve Bank of Australia Bulletin, December.
Reserve Bank of Australia, 2003, Statement on Monetary Policy, February.
Reserve Bank of Australia, “Housing Equity Withdrawal”, Reserve Bank of Australia Bulletin, February.
Reserve Bank of Australia, “Household Debt: What the Data Show”, Reserve Bank of Australia Bulletin, March.
Reserve Bank of Australia, Statement on Monetary Policy, May.
Vladkova Hollar, Ivanna, 2003, “An Analysis of House Prices in the United Kingdom”, IMF Country Report No. 03/47, (Washington: International Monetary Fund).
ANNEX Data Sources and Definitions
The data are quarterly for the period 1980:1 to 2003:1. The definition and sources for each variable are as follows:
Median price for new and established houses, weighted average for the eight capital cities. Source: Real Estate Institute of Australia.
Price index for new and established houses, weighted average for the eight capital cities. Source: Australian Bureau of Statistics.
Median price for new and established units and flats, weighted average for the eight capital cities. Source: Real Estate Institute of Australia.
The consumer price index (CPI). Official CPI series of the Reserve Bank of Australia, except for the computation of the real disposable income where the nonpublished CPI series adjusted for the introduction of GST in 2000 was used. Source: Reserve Bank of Australia.
Real mortgage interest rate. Nominal mortgage interest rate minus CPI inflation. Source: Reserve Bank of Australia.
Real returns on property prices. Year-on-year percentage change in the median price of new and established units and flats deflated by the CPI. Sources: Real Estate Institute of Australia and Reserve Bank of Australia.
Real returns on equities. Year-on-year growth rate of the ASX200 index deflated by the CPI. Source: Reserve Bank of Australia.
Real disposable income. Nominal deposable income deflated by the CPI. Source: The Australian Bureau of Statistics.
Population growth. Year-on-on year growth rate of total population. Source: The Australian Bureau of Statistics
Prepared by Abdelhak Senhadji (Ext. 38380).
Since the housing price indices do not fully adjust for quality, Table 1 may overstate the extent of housing price inflation.
The sharp decline in real interest rates during 2000–01 reflects the introduction of the goods and services tax and the consequent large transitory increase in CPI inflation.
According to the 2001 Census of Population and Housing, the average household size decreased from 2.8 persons in 1991 to 2.6 in 2001. The number of households increased from 15.4 million in 1991 to 17.2 million in 2001.
The CPI series used to deflate nominal disposable income has been adjusted for the introduction of the GST in 2000.
The returns for both housing and equities only include capital gains. Including dividends for equities and rents for housing is likely to show even higher returns for housing relative to equities.
Kiyotaki and Moore (1997) show how the effect of shocks to the housing market can be amplified by collateralized lending. In particular, they find that the dynamic interaction between credit limits and asset prices are a powerful transmission mechanism by which the effect of shocks persist, amplify, and spread out. See also Herring and Wachter (1999).
A decline in the relative yield of alternative investments would decrease investors’ wealth, which should reduce demand for housing. The substitution effects implies investors would move out from alternative investments and into housing as the relative yield of alternative investments decreases. The substitution effect is likely to dominate.
To take into account nonstationarity in the data, the cointegration framework of Phillips and Hansen (1991) was used. In addition, the Phillips-Hansen Fully Modified estimation method has the attractive feature of correcting for bias that may arise from potential endogeneity of the explanatory variables and/or serial correlation of the error term.
The precise definition of each variable is given in the annex.
The strengthening of the relationship between the relative price of housing and the real mortgage interest rate starting in the mid-1990s may reflect a more stable interest rate environment. Indeed, both the level and variability of interest rates declined significantly during the first half of the 1990s. Reduced variability in interest rates is particularly important when mortgage interest rates are adjustable, which is the case in Australia where variable interest rate loans account for about 85 percent of housing loans outstanding. A given decline in mortgage interest rates would likely generate a higher demand for housing (and thus lead to a larger increase in the relative price of housing) in an environment with low variability, and thus less uncertainty about the future path, of mortgage interest rates.
