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.
II. Housing, Consumption, and Output1
1. Housing activity in Australia has had a substantial macroeconomic impact in recent years. Strong housing investment in advance of the introduction of the goods and services tax (GST) contributed substantially to GDP growth in 1999 (Figure 1), but GDP growth dropped in 2000 as housing investment slumped after the GST was imposed in July 2000. Since the government established the First Home Owners Scheme (FHOS) to offset the impact of the GST, housing investment has increased substantially and made significant contributions to Australia’s GDP growth. In turn, the strength in housing activity has directly impacted consumption of housing-related durable goods and services, has contributed to a sharp rise in housing prices, and has influenced private consumption indirectly through increasing household net wealth (the traditional wealth effect). In addition, innovations in the credit market, which have allowed homeowners to more easily and less expensively borrow against their housing equity, have played an important role in stimulating consumption.
Australia: Housing-Related Activity
Citation: IMF Staff Country Reports 2003, 336; 10.5089/9781451801996.002.A002
2. The rapid increase in housing prices in recent years has raised concerns about the potential economic consequences of a sharp decline in prices. From understanding the channels through which the housing boom has affected real GDP, it is possible to gain some insight into how economic activity might be affected by a sharp decline in housing prices. The impact of such a decline would lead to a fall in housing investment and consumption of housing-related durables. The wealth effect would also operate in a symmetric fashion to reduce consumption. In contrast, although net equity withdrawals and the consumption they financed would decline when housing prices fall, households would not be expected to make substantial repayments of previous housing equity withdrawn. Accordingly, there would not be an additional negative effect on consumption. Hence, the impact on consumption of a sharp fall in housing prices could be significantly less than the rise in consumption that was associated with a similar rise in prices. However, a decline in housing prices could also affect output and consumption through its impact on the soundness of the banking system, which could potentially have a large and lasting impact on economic activity. Household financial distress could lead to rising loan defaults affecting both the ability and willingness of banks to lend, not just to the housing sector, but across the whole economy. At present, this is not expected to have a substantial negative effect because the stress tests done on the Australian banks suggest that a large decline in housing prices is not likely to lead to systemic problems.
A. Recent Developments in Housing Prices and Household Wealth
3. Strong housing demand has led to a substantial rise in housing prices in recent years. The weighted average of housing prices across the eight capital cities has risen at an average annual rate of 13 percent since 1999, with a price increase of 18 percent in the last year alone (Figure 2). In addition to the stimulus provided by the government’s FHOS, solid growth in long-term immigration and a rising number of households have increased housing demand. The rise in housing prices also has occurred during a time when returns from alternative investments, such as equities, have weakened significantly.2 As a result, strong investor demand for housing has developed. This demand primarily reflects purchases of single housing units by individuals seeking to capture capital gains in the housing market. The housing price increase also reflects a long-term adjustment by households to a new low inflation and low interest rate environment, which has allowed households to borrow two times more relative to their disposable income than they could in the 1980s, while maintaining their debt-servicing costs largely unchanged.3
Australia: Housing Price Developments
Citation: IMF Staff Country Reports 2003, 336; 10.5089/9781451801996.002.A002
4. The housing price increase also has been a major contributor to a rise in household net wealth (Figure 3). Household net wealth rose from 430 percent of disposable income in June 1992 to 490 percent in June 2002. While the initial rise in wealth during this period was driven by increases in the prices of financial assets (mainly equity prices), household net worth in more recent years has been largely propelled by higher nonfinancial wealth (mainly housing). This rise in net housing wealth has taken place despite a sharp rise in gross indebtedness. Household debt has more than doubled over the 1990s, from just over 50 percent of disposable income in 1990 to some 125 percent by the end of 2002—a debt level that is now broadly comparable to that in the United States and the United Kingdom—with most of this debt being housing-related.4
Australia: Household Wealth Indicators
Citation: IMF Staff Country Reports 2003, 336; 10.5089/9781451801996.002.A002
5. Historically, there are similarities and distinct differences in the behavior of housing price in Australia compared to other industrial countries. Since the late 1990s, Australia has experienced a rate of increase in housing prices very similar to that which has taken place in the United Kingdom (Figure 4). In both the United States and Canada, housing prices have risen at significantly slower rates over this period. Moreover, in 2002, the deviations in housing prices from their long-term trend were broadly similar in the United Kingdom and in Australia (around 3 to 4 percent; Figure 5). However, the behavior of housing prices in Australia following the end of previous booms differs sharply from that in the United Kingdom. Housing price booms in Australia have been followed by periods of relatively stable prices; in the United Kingdom, they have been followed by significant declines in prices. In this, housing price behavior in Australia has been closer to that in the United States.
International Comparison of Housing Price Developments
Citation: IMF Staff Country Reports 2003, 336; 10.5089/9781451801996.002.A002
International Comparison of Actual and Trend Housing Prices
(In logs)
Citation: IMF Staff Country Reports 2003, 336; 10.5089/9781451801996.002.A002
B. Effects of Housing on Consumption
6. Housing influences consumption directly or indirectly through at least three different channels. First, the rise in housing investment directly affects consumption of housing-related durable goods and services. Second, housing influences consumption indirectly through the effect of rising housing prices on the net worth of households. Third, housing also could have an indirect influence through innovations in credit markets that allow households to tap equity in their homes to finance consumption relatively easily and at lower interest cost than alternative means of borrowing.
7. In Australia, housing activity has had a strong direct effect on consumption (Figure 6). Housing investment has been highly correlated with the consumption of housing-related durable goods and services (correlation coefficient of 0.88 since March 1990).5 The consumption of these goods have made a fairly significant contribution to the growth in private consumption in recent years.
Australia: Housing-Related Consumption
Citation: IMF Staff Country Reports 2003, 336; 10.5089/9781451801996.002.A002
8. Econometric estimates of a consumption function suggest that the net housing wealth effect has significantly impacted Australian household consumption. According to the traditional wealth effect, an increase in “permanent” household wealth would induce households to spend more out of current income (i.e., save less). An equation derived from a long-run consumption function consistent with permanent income and life-cycle models was estimated based on cointegration theory using quarterly data from 1981:4 to 2002:3.6 The equation specifies consumption as being determined by a set of variables including personal disposable income, net housing wealth, and households’ access to credit to capture the effects of financial deregulation and innovation.7 The results suggest that housing wealth has had a significant impact on private consumption in Australia with annual private consumption rising by some 5 cents for a dollar increase in net housing wealth (Table 1 and Figure 7).8 These results are broadly consistent with those presented in other studies. For example, Dvornak and Kohler (2003) found that a dollar increase in housing wealth in Australia raises consumption by some 3 cents. In a cross-country panel study that included Australia, IMF (2002) found that in market-based economies, consumption increased by 7 cents for every dollar increase in housing wealth.
Australia: Consumption Equations
Estimated Equation for Household Consumption 1/
Estimation period: 1981:4–2002:3
All variables except access to credit are in real per capita terms.
Australia: Consumption Equations
Estimated Equation for Household Consumption 1/
Estimation period: 1981:4–2002:3
| Coefficient | Std. error | Prob. Value | |
|---|---|---|---|
| Intercept | -5.523 | 5.053 | 0.278 |
| Disposable income | 0.802 | 0.150 | 0.000 |
| Net housing wealth | 0.012 | 0.004 | 0.011 |
| Access to credit (-1) | 0.061 | 0.036 | 0.094 |
All variables except access to credit are in real per capita terms.
Australia: Consumption Equations
Estimated Equation for Household Consumption 1/
Estimation period: 1981:4–2002:3
| Coefficient | Std. error | Prob. Value | |
|---|---|---|---|
| Intercept | -5.523 | 5.053 | 0.278 |
| Disposable income | 0.802 | 0.150 | 0.000 |
| Net housing wealth | 0.012 | 0.004 | 0.011 |
| Access to credit (-1) | 0.061 | 0.036 | 0.094 |
All variables except access to credit are in real per capita terms.
Australia: Estimates of Household Consumption
Citation: IMF Staff Country Reports 2003, 336; 10.5089/9781451801996.002.A002
9. Housing also appears to have played a role in stimulating consumption through a credit market channel. Financial deregulation and greater competition among banks have led to the development of new products, such as home equity loans and mortgages with a redraw facility, that have reduced credit constraints on household borrowing.9 During the period from end-2001 to March 2003, the cumulative mortgage equity withdrawal by households was around $A 165 billion (Figure 8).10 Housing wealth increased by around $A 650 billion during this period, which, based on the estimated wealth effect, suggests that consumption could have been increased by as much as $A 33 billion, if households considered this increase in wealth to be permanent. Thus, the resources released from mortgage equity withdrawal were around five times more than the funds available from the estimated housing wealth effect. The positive cash flow generated from equity withdrawal represents money available to finance other spending and consumption activities. According to an ABS survey (ABS, 2001), some 20 percent of borrowers used their home refinancing during 1997–99 to finance purchases such as cars and holidays (Table 2). While more recent surveys are not available, it is likely that a more substantial part of the housing equity withdrawal was used for consumption spending in recent years, due to the large interest rate differential between home equity loans and other personal loans, which would make it cheaper for households to finance purchases, such as cars, using home equity loans (Figure 8).
Australia: Housing Equity Withdrawal and Lending Rates
Citation: IMF Staff Country Reports 2003, 336; 10.5089/9781451801996.002.A002
Australia: Reasons for Refinancing, 1997-1999
Some owners report more than one reason for refinancing. Thus, components do not add up to total.
Australia: Reasons for Refinancing, 1997-1999
| In percent | |
|---|---|
| Better interest rate | 23.3 |
| Better loan conditions | 25.1 |
| Extension of loan period | 5.1 |
| Home renovations | 9.5 |
| Other purchase (car, holiday) | 21.0 |
| Consolidation of debts | 15.2 |
| Business related reasons | 7.5 |
| Other reasons | 17.4 |
| All homeowners who refinanced 1/ | 100.0 |
Some owners report more than one reason for refinancing. Thus, components do not add up to total.
Australia: Reasons for Refinancing, 1997-1999
| In percent | |
|---|---|
| Better interest rate | 23.3 |
| Better loan conditions | 25.1 |
| Extension of loan period | 5.1 |
| Home renovations | 9.5 |
| Other purchase (car, holiday) | 21.0 |
| Consolidation of debts | 15.2 |
| Business related reasons | 7.5 |
| Other reasons | 17.4 |
| All homeowners who refinanced 1/ | 100.0 |
Some owners report more than one reason for refinancing. Thus, components do not add up to total.
10. Rising housing prices might also affect housing investment and consumption by influencing the willingness of banks to lend and households to borrow. It is particularly difficult to try to model this effect because of the difficulty in quantifying this “willingness” factor and to identify “overlending” or “overborrowing” that might result from it. While this may have relatively minor effects when housing prices are rising, more significant problems could emerge in the event of a pronounced price decline. In such circumstances, substantial financial stress could develop in the household sector with rising default rates and repercussions for the soundness of financial institutions. In turn, there could be wider ramifications for the economy, if problems in home loan portfolios were to induce a more broad-based retrenchment of bank lending. In Australia’s case, housing mortgages comprise roughly half of total bank loans of the major banks. Default rates on mortgages are low, but this could reflect the fact that, historically, Australian housing prices have not fallen significantly, and the vast majority of loans have been to owner-occupiers. In the event of a sharp fall in housing prices and with a larger number of investors in the housing market (whose behavior in this circumstance is not clearly known), default rates could rise and financial system problems emerge. To assess this possibility, the Australian Prudential Regulation Authority conducted special stress tests to see how individual banks might fare if housing prices were to decline sharply. The preliminary results of these tests suggest that, while the net income of individual banks could be significantly affected by a housing price fall, bank capital appeared to be sufficient to avert the risk of failure of any major institution and systemic difficulties.
References
Davey, M., 2001, “Mortgage Equity Withdrawal and Consumption”, Bank of England Quarterly Bulletin, Spring 2001.
Dvomak, N. and M. Kohler, 2003, “Housing Wealth, Stock Market Wealth and Consumption: A Panel Analysis for Australia”, Reserve Bank of Australia Research Discussion Paper, July 2003.
Gizycki, M., 2001, “The Effects of Macroeconomic Conditions on Banks’ Risk and Profitability”, Research Discussion Paper 2001-06, Reserve Bank of Australia, September 2001.
Gizycki, M., and P. Lowe, 2000, “The Australian Financial System in the 1990s”, Paper presented at Reserve Bank of Australia Conference The Australian Economy in the 1990s, July 2000.
International Monetary Fund, 2002, “Is Wealth Increasingly Driving Consumption?”, World Economic Outlook, April 2002, pp. 74–85.
Reserve Bank of Australia, 2002, “Recent Developments in Housing: Prices, Finance and Investor Attitudes”, Reserve Bank of Australia Bulletin, July 2002.
Reserve Bank of Australia, 2003a, “Household Debt: What the Data Show”, Reserve Bank of Australia Bulletin, March 2003.
Reserve Bank of Australia, 2003b, “Housing Equity Withdrawal”, Reserve Bank of Australia Bulletin, February 2003.
Tan, A. and Voss, G., 2003, “Consumption and Wealth in Australia”, The Economic Record, Vol. 79, No. 244, pp. 39–56.
ANNEX Data Sources and Definitions
The definition and data sources for the variables in the estimations are provided below.
All variables except access to credit are in real per capita terms.
Private consumption is the seasonally adjusted household consumption in current prices. Data from 1981:2 to 2002:4. Source: Australian Bureau of Statistics.
Disposable income is the seasonally adjusted household disposable income in current prices. Data from 1981:2 to 2002:4. Source: Australian Bureau of Statistics.
Net housing wealth is households’ housing assets minus lending for housing to persons. Data from 1981:2 to 2002:3. Source: Reserve Bank of Australia.
Access to credit is the ratio of total household credit to total banking credit. Data from 1981:2 to 2002:3. Source: Reserve Bank of Australia. The variable is lagged one period in the regression because households are expected to react with some delay to developments in banking sector lending policies.
Price deflator is the seasonally adjusted household consumption deflator. Data from 1981:2 to 2002:4. Source: Australian Bureau of Statistics.
Housing equity withdrawal is the net cash flow generated by the household sector from transactions in housing assets and mortgage debt. If the household sector in aggregate increases its mortgage debt by more than its net spending on housing assets, housing equity withdrawal is said to have taken place. Source: “Housing Equity Withdrawal”, Reserve Bank of Australia Bulletin, February 2003.”
Prepared by Uma Ramakrishnan (Ext. 35413).
Equity prices in Australia have risen at an annual average of 2¼ percent for the three years ending 2002, compared to the 13 percent increase in housing prices.
Reserve Bank of Australia Statement on Monetary Policy (August 2003).
For a more detailed discussion of Australian household debt data, see Reserve Bank of Australia (2003a).
Housing-related durable goods and services are derived from data on household expenditure as the sum of frirnishings and household equipment and other dwelling services. Quarterly data for other dwelling services were computed by applying the ratio in annual data of other dwelling services to the sum of rent and other dwelling services.
The equations were estimated using the Phi Hips-Hans en fully modified OLS (FMOLS) procedure. This method estimates the long-run parameters by correcting for serial correlation in the residuals without having to explicitly specify the dynamics of the model. It is a valid procedure when there exists a single cointegration equation and when the explanatory variables are not themselves cointegrated. For details, see Phillips and Hansen (1990).
Explanations for the derivation of the variables used and sources for the data are provided in the Annex.
Net financial wealth was dropped from the estimation due to insufficient data spanning the whole sample period. Further, Australian net worth has been increasingly held in housing, with the ratio rising from 53 percent of total wealth in 1993 to the current level of more than 70 percent, making it the dominant determinant of the wealth effect.
The redraw facility entitles the customer to automatically redraw any payments made to the loan account in excess of the repayment commitment under the loan.
For a detailed discussion on mortgage equity withdrawal in Australia, see Reserve Bank of Australia (2003b).
