Abstract
This paper analyzes likely effectiveness of the fiscal policy in supporting aggregate demand in Australia. The simulation analysis illustrates that the type of fiscal measure and the underlying behavioral responses have an important impact on the magnitude of fiscal multipliers. The Global Integrated Monetary and Fiscal Model simulations also suggest that the cumulative impact after five years could be close to 10 percentage points of GDP for the announced stimulus measures that cumulate to almost 8 percentage points of GDP.
II. Australian Household Vulnerabilities4
A. Introduction
10. If the economic downturn in Australia is to be milder than in other industrial countries, resilience in private consumption spending will be essential. To this point in the downturn, private consumption growth in Australia has slowed, but not nearly to the same extent as in the United States, the United Kingdom, or Canada (Figure II.1). This chapter examines the position of Australian households relative to those in several other advanced countries to assess how well household spending can be expected to hold up going forward.
Australia: Private Consumption
(Q/Q percentage change, seasonally adjusted)
Citation: IMF Staff Country Reports 2009, 249; 10.5089/9781451802214.002.A002
11. The comparison suggests that, in addition to rising unemployment and slower income growth, the key vulnerability that Australian households face is the value of their housing wealth. House price appreciation has been the single largest contributor to the growth in Australian household wealth over the last decade. The metrics examined suggest overvaluation in house prices as of March 2009 in the range of 0–20 percent. Although a sharp correction in house prices toward the upper end of this range could prove disruptive for private consumption because of its implications for household wealth, a number of factors should contribute to gradual and orderly adjustment. First, a significant moderation in house price inflation has already occurred and prices appear to have stabilized. Second, the significant decline in mortgage interest rates, a government subsidy for first-time home buyers, and continued strong net immigration will all provide considerable support for housing demand and thus prices. This in turn should help maintain households’ wealth, confidence, and consumption expenditure. Nevertheless, a sharp fall in house prices over the next few years remains a tail risk.
B. Household Saving and Wealth
12. Although Australian households accumulated significant debt prior to the downturn, household consumption appears to be in a less precarious position than in some other countries. Similar to the situation in other advanced countries, the household savings rate in Australian declined over the first half of the last decade. However, the savings rate subsequently recovered and has returned to positive territory (Figure II.2). There has also been a significant increase in household debt relative to disposable income. While this debt accumulation occurred in many countries, the rise in debt service costs, owing to a prolonged period of tight monetary policy, was dramatic in both Australia and New Zealand.5 Australian households, however, appear to have covered a portion of their increased debt-service burden by reducing consumption as a share of disposable income. Consequently, Australian households may have entered the downturn with consumption expenditure closer to a sustainable level.
Australia: Household Savings
Citation: IMF Staff Country Reports 2009, 249; 10.5089/9781451802214.002.A002
13. Despite the run-up in debt, Australian households’ net wealth position has risen over the last decade, but the improvement has been largely housing wealth (Figure II.3). Australian and New Zealand households currently have net wealth positions that are well above their levels a decade earlier. However, in part this reflects their relatively low holdings of financial wealth, which has been hit hard in many countries by falling equities prices. Australian households, like those in New Zealand and the United Kingdom, hold a large portion of their wealth as housing. As outlined in Kohler and Smith (2005), the concentration of the population in large urban centers in Australia can partially account for the higher proportion of housing in wealth. However, this doesn’t mitigate the risk to household consumption. With housing such a large share of their wealth and house prices having risen so significantly over the last decade, Australian household wealth, and thus spending, could be vulnerable to a large decline in house prices.
Australia: Household Wealth
Citation: IMF Staff Country Reports 2009, 249; 10.5089/9781451802214.002.A002
C. Metrics of House-Price Valuation
14. As in other countries over the last ten years, real house prices have risen rapidly in Australia, and here we use two metrics to assess house-price valuation (Figure II.4). First, an econometric approach is used to derive estimates of the house-price gap (i.e., the extent to which the increase in house prices in recent years cannot be explained by economic factors). Second, the ratios of house prices to disposable income and house prices to rents are compared to various estimates of their sustainable levels (long-run moving averages).
Real House Prices
(1997-2008, cumlative growth rate)
Citation: IMF Staff Country Reports 2009, 249; 10.5089/9781451802214.002.A002
Source: OECD database.15. The econometric approach models house price growth as a function of the following variables: growth in per capita disposable income, changes in working age population (or alternatively lagged net immigration), changes in equity prices and credit, short-term and long-term interest rates (or alternatively mortgage rates). All versions of the model include an affordability ratio (the lagged ratio of house prices to disposable income) as a medium-term anchor. Some versions of the model considered also include lagged real house price growth to better capture the observed persistence.
16. As is the case with all reduced-form econometric models, the results from this model need to be interpreted cautiously, but can still be informative. First, exogeneity of variables assumed to be independent is not guaranteed. Second, such models rarely perfectly capture the underlying theoretical relationships that they are used to proxy. Finally, but not unrelated, results can be highly dependant on sample periods considered and specification. That being said, this type of modeling approach has been used elsewhere to provide estimates of house price overvaluation, for example see Terrones (2004), and should be interpreted a just another piece of evidence contributing to an assessment of potential disequilibrium in house prices.
17. The resulting econometric estimates suggest that house-price overvaluation ranges between 5 and 15 percent as of March 2009 depending on different specifications and starting dates for the comparison. The estimated house-price gaps from five models (Appendix Table II.1) are presented in Figure II.5 under three different start dates for generating the models’ predicted levels for house prices.6 All models and start dates imply some degree of overvaluation, the simple average of which is 11 percent. These results should be interpreted cautiously because models selected on the basis of their ability to fit the data, may not actually be effectively capturing the underlying fundamentals determining sustainable prices. For example, the low current interest rates help explain current house prices, but looking ahead Australian interest rates will undoubtedly return to a more neutral level.
House Price Gaps as of March 2009
(In percent) 1/
Citation: IMF Staff Country Reports 2009, 249; 10.5089/9781451802214.002.A002
1/ The specifications are reflected in the estimates reported in Appendix Table II.1.Source: Fund staff estimates.18. The simple metrics of the price-to-rent ratio and the price-to-income ratio also point to some house price overvaluation in Australia (Figure II.6). Over the last couple of decades many industrial countries witnessed rapid house price appreciation. While prolonged high house price appreciation does not in itself constitute overvaluation, the ratios of house prices to income and to rent provide some simple measures of potential misalignment. However, care must be taken when evaluating the current level of these ratios relative to their long-run trends. Structural changes such as permanently lower nominal interest rates, rising incomes, and the increasing scarcity of land close to main urban centers can contribute to sustainable increase in these ratios. In Australia, the scarcity of land close to main urban centers is an important factor. As argued in Ellis and Andrews (2001), the fact that such a high proportion of Australia’s population live in two major centers tends to drive up average house prices and thus the ratio of prices to incomes. One way to allow for this is to evaluate the current levels relative to a range of long-run moving averages that proxy for the sustainable level. The gaps relative to 10-, 15-, and 20-year moving averages presented in Table II.1 indicate that in all but one case, the price-to-income and price-to-rent ratios appear to be above sustainable levels.
Australia: House Prices
Citation: IMF Staff Country Reports 2009, 249; 10.5089/9781451802214.002.A002
Australia: Percent Deviation from Estimated Sustainable Levels as of March 2009
Australia: Percent Deviation from Estimated Sustainable Levels as of March 2009
| Relative to 10-Year Moving Average | Relative to 15-Year Moving Average | Relative to 20-Year Moving Average | |
|---|---|---|---|
| Price-to-Income Ratio | −6.1 | 3.3 | 7.4 |
| Price-to-Rent Ratio | 6.7 | 20.9 | 30.4 |
Australia: Percent Deviation from Estimated Sustainable Levels as of March 2009
| Relative to 10-Year Moving Average | Relative to 15-Year Moving Average | Relative to 20-Year Moving Average | |
|---|---|---|---|
| Price-to-Income Ratio | −6.1 | 3.3 | 7.4 |
| Price-to-Rent Ratio | 6.7 | 20.9 | 30.4 |
D. Assessment of Potential Overvaluation and Risks
19. A reasonable assessment of the metrics considered would suggest overvaluation in the range of 0 to 20 percent as of March 2009. Although the price-to-rent ratio was roughly 30 percent above its 20-year moving average, this is likely too a long a period to use for the estimate of the sustainable level given the significant structural change that has occurred over the period. Further, using only a 10-year moving average may be too short, particularly given the duration of Australia’s most recent economic expansion. Based on the 15-year moving averages, the gaps at the end of the first quarter of 2009 would suggest a range of overvaluation of roughly 0–20 percent. This range also encompasses the range of estimates of overvaluation coming from the econometric analysis. While the range of overvaluation may appear to suggest significant adjustment in house prices could be required, it is useful to remember that sustainable levels can be achieved if incomes and rents grow faster than house prices, or other fundamentals, such as migration, change. Thus house prices alone need not do all the adjustment and the adjustment that is required in prices, can occur gradually.
20. Looking ahead, several factors suggest that any required house price adjustment is likely to be orderly. First, after declining q/q by 2 percent in 2009 Q1, preliminary data for the second quarter of 2009 indicate that house prices have started to rise again. Second, variable mortgage interest rates have fallen by almost 400 basis points since the RBA started to ease monetary policy in September 2008 and are expected to remain low for an extended period of time. Third, the government’s subsidy for first-time home buyers, which lasts through 2009, will also provide support for housing demand. Finally, the gap over the last several years between population growth and housing starts should result in strong demand for housing over the next few years (Figure II.7).
Australia: Housing Demand and Supply
Citation: IMF Staff Country Reports 2009, 249; 10.5089/9781451802214.002.A002
Appendix II.1. Australia: House Price Models
The growth rate of real house prices was modeled as a function of the following variables:
Past growth rates of real house prices. If the growth rate of house prices is persistent, then the current growth rate must be serially correlated with the past growth rate. Higher values of this correlation coefficient imply higher persistence;
Past housing affordability ratio. If the growth rate of house prices shows long-run reversion to fundamentals, this implies that prices would tend to fall when they are out of line relative to income levels. Hence, the coefficient of the housing affordability ratio—the ratio of real house prices to (per capita) real income—must be negative;
Fundamentals. The growth rate of house prices is positively affected by household real disposable income growth—as this increases households’ purchasing power and borrowing capacity. Other fundamentals influencing house prices include the growth rate of real credit, a proxy for mortgage debt, as this indicates that households are less credit rationed; the past growth rate of real stock prices—which captures households’ efforts to rotate their portfolio in favor of housing; working age population (or alternatively migration) growth, as this proxies for the growth rate of households; and
Interest rates. The growth rate of house prices is negatively affected by interest rates (including mortgage rates) because lower rates increase households’ capacity to borrow.
This model was estimated using quarterly data over the period Q1 1987-Q1 2009.
The econometric results confirm that real house prices in Australia show high persistence, long-run reversion to fundamentals, and dependence on economic fundamentals, especially working-age population growth (Appendix Table II.1).
Australia: Determinations of House Prices in Australia: Empirical Results, 1987Q1-2009Q1
Lagged ratio of house prices to disposable income.
Australia: Determinations of House Prices in Australia: Empirical Results, 1987Q1-2009Q1
| Explanatory Variables | Dependent Variable Real House Price (growth) | ||||||
|---|---|---|---|---|---|---|---|
| Model 1 | Model 2 | Model 3 | Model 4 | Model 5 | |||
| Lagged dependent variable | |||||||
| Lagged real house price (growth) | … | … | 0.57*** | 0.55*** | 0.60*** | ||
| Reversion | |||||||
| Lagged housing affordability ratio 1/ | −0.05*** | −0.057*** | −0.03** | −0.041*** | −0.04*** | ||
| Other variables | |||||||
| Lagged real disposable income | 0.180 | 0.153* | 0.22* | 0.23* | 0.201* | ||
| (Per capita, growth) | |||||||
| Short-term interest rate (percent) | −0.003 | … | −0.022 | … | … | ||
| Long-term interest rate (percent) | 0.000 | … | 0.000 | … | … | ||
| Real credit (growth) | 0.123 | 0.015*** | 0.012 | −0.01 | −0.022 | ||
| Lagged real stock prices (growth) | −0.005 | −0.008 | 0.000 | −0.005 | −0.001 | ||
| Working-age population (growth) | 3.02** | … | 2.68** | 2.48** | … | ||
| Mortgage interest rate (percent) | … | −0.003*** | … | −0.002*** | −0.001*** | ||
| Lagged net long-term migration to population (t-3) | … | 0.01* | … | … | 0.014*** | ||
| Number of observations | 89 | 89 | 89 | 89 | 89 | ||
| R-squared | 0.23 | 0.24 | 0.51 | 0.51 | 0.52 | ||
| Breusch-Godfrey LM test | 0.000 | 0.000 | 0.09 | 0.09 | 0.050 | ||
Lagged ratio of house prices to disposable income.
Australia: Determinations of House Prices in Australia: Empirical Results, 1987Q1-2009Q1
| Explanatory Variables | Dependent Variable Real House Price (growth) | ||||||
|---|---|---|---|---|---|---|---|
| Model 1 | Model 2 | Model 3 | Model 4 | Model 5 | |||
| Lagged dependent variable | |||||||
| Lagged real house price (growth) | … | … | 0.57*** | 0.55*** | 0.60*** | ||
| Reversion | |||||||
| Lagged housing affordability ratio 1/ | −0.05*** | −0.057*** | −0.03** | −0.041*** | −0.04*** | ||
| Other variables | |||||||
| Lagged real disposable income | 0.180 | 0.153* | 0.22* | 0.23* | 0.201* | ||
| (Per capita, growth) | |||||||
| Short-term interest rate (percent) | −0.003 | … | −0.022 | … | … | ||
| Long-term interest rate (percent) | 0.000 | … | 0.000 | … | … | ||
| Real credit (growth) | 0.123 | 0.015*** | 0.012 | −0.01 | −0.022 | ||
| Lagged real stock prices (growth) | −0.005 | −0.008 | 0.000 | −0.005 | −0.001 | ||
| Working-age population (growth) | 3.02** | … | 2.68** | 2.48** | … | ||
| Mortgage interest rate (percent) | … | −0.003*** | … | −0.002*** | −0.001*** | ||
| Lagged net long-term migration to population (t-3) | … | 0.01* | … | … | 0.014*** | ||
| Number of observations | 89 | 89 | 89 | 89 | 89 | ||
| R-squared | 0.23 | 0.24 | 0.51 | 0.51 | 0.52 | ||
| Breusch-Godfrey LM test | 0.000 | 0.000 | 0.09 | 0.09 | 0.050 | ||
Lagged ratio of house prices to disposable income.
References
Ellis, L., and D., Andrews, and 2001, “City Sizes, Housing Costs, and Wealth,” Reserve Bank of Australia Research Discussion Paper, 2001–08.
Kohler, M., and K., Smith 2005, “Housing and the Household Wealth Portfolio: The Role of Location,” Reserve Bank of Australia Research Discussion Paper 2005–10.
Terrones, M., 2004, “What Explains the Recent Run-Up in House Prices?,” Box 2.1 in Chapter 2, World Economic Outlook, September 2004 (Washington: International Monetary Fund).
Prepared by Ben Hunt (ext. 36361), Khoi Viet Nguyen (ext. 37417), and Patrizia Tumbarello (ext. 34395).
With the decline in variable mortgage interest rates in Australia, debt-service costs moved below 12 percent of disposable income in the first quarter of 2009.
Because the equation is estimated in first differences, to generate estimates of the house-price gap, the fitted values from the regressions are used to generate an estimated level for house prices to compare against the level of house prices at the end of the sample period. Consequently, the start date for generating the estimated level for house prices can influence the estimate of the house-price gap.
