RDP 2006-12: Housing and Housing Finance: The View from Australia and Beyond 4. Policy Issues

Housing prices have risen substantially in many countries over the past 15 years, both in the English-speaking world and in a number of countries in continental Europe. Related to this, household indebtedness (the ratio of debt to income) has increased substantially in most major economies. This expansion in household balance sheets has raised some concern for policy-makers. Lower inflation, reduced financial regulation and ongoing competition and innovation have allowed higher debt burdens. Even if most or all of the expansion so far has been a rational response to a new equilibrium, there is a risk that households or lenders may succumb to over-exuberance and a tendency to extrapolate past price gains into the future. On the lending side, there is evidence that financial institutions in some countries have eased lending standards at the same time as households have been able to borrow more. It is therefore of interest to disentangle how much of the increase in household indebtedness and gearing against housing represents a normal reaction by borrowers and lenders to the easing of constraints and regulation, and how much implies an increase in their risk profiles.

Even if systemic risks to the financial system have not increased, these developments have implications for policy-makers' understanding of macroeconomic behaviour. There is a growing body of literature suggesting that the easing of financial constraints makes household spending more sensitive to shocks, not less (Almeida 2000; Aoki, Proudman and Vlieghe 2002; Debelle 2004, for example). Similarly, city-level evidence from the US suggests that households respond more to income shocks when the average LTV ratio is high (Lamont and Stein 1999). It is therefore important for policy-makers to have good information on how the transmission of shocks might be changing.

4.1 Need for More and Different Data

To assess the implications of these developments in household balance sheets for macroeconomic and financial stability, policy-makers are finding that they need to expand the types of data they analyse (BIS 2006; Schwartz et al 2006). In particular, disaggregated information about household financial positions and stresses is likely to become increasingly important. This is because it is not the average household that will experience distress in the event of a shock, but rather those households at the tails of the distribution. At this stage, the conclusions drawn from analysis of disaggregated data are fairly benign. For example, stress tests using Swedish data (BIS 2006) and analysis of disaggregated data on balance sheets of Australian households (for example, RBA 2003a; Ellis, Lawson and Roberts-Thomson 2003; and Kohler and Rossiter 2005) suggest that the holders of housing debt in Australia and elsewhere are those most able to afford it.

In addition, if balance sheet developments have a stronger influence on household spending than in the past, identifying turning points in housing prices will be of greater concern for macroeconomic analysis than previously. This means that the quality of housing price indicators will become increasingly important. In Australia, the RBA has put considerable resources not only into analysing existing indicators, but also into encouraging data providers to improve the range and quality of their data. Some of the outcomes of these efforts included the introduction of a new set of mix-adjusted house price indices (Prasad and Richards 2006) and a deeper understanding of repeat-sales and hedonic price indices (Hansen 2006). Also largely in response to the RBA's advocacy, many data providers shifted to reporting house prices using the date of sale as the reference period, rather than the date the transaction was reported to land titles agencies. This improves the usefulness of the data because it relates the price to the period in which the price was agreed.

Data improvements not only help policy-makers, but might also assist private decision-making as well. The increased use of data-driven risk assessments through credit scoring implies that data quality has become increasingly important for lending and underwriting decisions. For this reason, there may be positive externalities to efforts by policy-makers to improve data.

4.2 Old Rules of Thumb for Balance Sheets Might be Misleading

As has been discussed earlier in this paper, much of the expansion in borrowing has been a response to reductions in inflation, financial market deregulation and financial product developments that have all allowed borrowers to manage larger debts. These are permanent changes. Ratios of debt or housing prices to income should therefore not be expected to revert to past historical averages, and it would be a mistake to enact policies designed to bring this about.

It would also be misleading to assume that past relationships between household balance sheet variables ought to reassert themselves. Ratios of household debt to assets have been much more stable over recent years than the ratios of debt and housing prices to income shown in Figures 3 and 4. This is because they are not affected by the mortgage tilt effects of disinflation. But they have nonetheless increased, as households have responded to the removal of financial constraints that resulted from deregulation and increased competition.

Old rules of thumb about individual mortgages have likewise been rendered obsolete. As lenders have used a wider range of information to ascertain different borrowers' credit risk more precisely, the amounts they are willing to lend are no longer linked to repayment-income ratios or loan-to-valuation ratios in a simple way. Particularly in North American markets, simple ratios have given way to credit scoring and risk-based pricing, so that loan sizes and pricing are more closely tailored to individual borrowers' circumstances. To the extent that this reduces the margin of safety for some borrowers who are now able to borrow more than the older practices would have implied, this might mean that more households are facing greater financial risks than previously. But overall, this easing of financial constraints is a reflection of their ability to repay and withstand those risks. Therefore it cannot be assumed that a shift away from the earlier lending practices based on rigid ratios implies that financial vulnerability has increased in any significant way. On the other hand, most credit-scoring models have been based on data drawn from the unusually benign experience of recent history, and have not been tested in more turbulent times.

4.3 Limited Effect of Extra Marginal Supply

In some countries, particularly Australia and the UK, the upswing in housing prices has made it difficult for young households to achieve homeownership. At times, this has become a contentious issue, with various interested parties arguing that regulations and other frictions have brought about an affordability crisis by preventing an expansion of supply at the fringes of cities.^[15] But such an argument fails to distinguish between an increase in the number of dwellings demanded, and an increase in demand for average quality of dwellings. If income growth, demographic change or immigration is boosting the number of households, there will clearly be demand for extra dwellings. However, this is not the main cause of the expansion in housing demand and mortgage borrowing seen in recent years. As shown in Table 2, it is simply physically infeasible for new supply to expand enough to have accommodated the expansion in households' capacity to pay. Instead, there have necessarily been large increases in the cost of housing and land.

This is not to say that government regulation has not had a role in determining the level of house prices. As mentioned in the previous section, cities that are more sprawled do seem to have lower house prices. However, they are not immune from price cycles (Chinloy 1996), or from the level shift in equilibrium housing prices that occurs when inflation falls and financial sectors are deregulated. It seems indisputable that government regulation can increase housing prices (Glaeser and Gyourko 2002, 2003), but regulation designed to prevent sprawl is not the only kind that does so. Regulations setting high minimum standards for housing quality or block size, or preventing medium-density and apartment development – which effectively enforce sprawl – also raise prices (Schill 2002).^[16] As an illustration of this, Figure 5 shows that Portland, Oregon, which is widely cited as a case where planning policy has been inspired by anti-sprawl ‘Smart Growth’ ideas, does not appear to have particularly high housing prices compared to other cities of comparable size. In addition, the cumulative price growth in the recent upswing has been much less there than for some fast-growing cities where sprawl has occurred, such as Las Vegas and Phoenix. This example is not intended as a substitute for a thorough econometric investigation, which is beyond the scope of this paper.^[17] Nonetheless, the facts suggest that allowing for more spread out cities or, more generally, untrammelled supply of extra dwellings, would not have prevented a large increase in Australian housing prices over the past decade.

The debate about the role of housing supply highlights the fact that the expansion in housing prices and debt has meant that issues previously considered to be specific to housing policy now have implications for macroeconomic policy. Policy-makers may need to become more expert in topics they previously considered outside of their portfolio in order to evaluate claims and debates that have some bearing on macroeconomic and financial developments.

4.4 Rising Indebtedness

The most important lesson to draw from recent international experience is that a run-up in housing prices and debt need not be dangerous for the macroeconomy, was probably inevitable, and might even be desirable. As emphasised by the BIS Committee on the Global Financial System's Working Group report, the expansion in household borrowing has in many cases reflected better pricing of risk and credit scoring, implying that credit is being allocated more efficiently than in the past. This should improve the economy's resilience to adverse shocks. In addition, the product innovation summarised in Table 1 implies that households now have greater choice about the kind of mortgage they take out, which ought to be welfare-improving (BIS 2006).

If a macroeconomic downturn were to occur, it could be exacerbated by a correction of an extended housing boom. However, the experience of Australia and the UK seems to suggest that booms in housing price growth can subside without themselves bringing about a macroeconomic downturn. In Australia, nationwide average housing prices fell for around two years from about the end of 2003. Household consumption growth did slow during this period, but from rates that were unsustainably strong (Figure 6). The picture for the UK was similar, even though it did not benefit from the cushioning effect on incomes from the sharply rising terms of trade, as Australia experienced.

These relatively benign outcomes point to the underlying robustness of the financial systems in these economies. Even where there was evidence of speculative demand (or panic buying), and an apparent belief in some quarters that housing prices never fall, households adapted to the turn in the market reasonably well. Although there have been anecdotal reports of home buyers experiencing negative equity, it seems that much of this can be attributed to the normal idiosyncratic risk inherent in a heterogeneous product like residential housing.

In contrast, consumption slowed more sharply in the Netherlands when housing price growth slowed there.^[18] However, the slowdown in consumption was not caused directly by households reacting to housing prices; rather, both developments were driven by a more general macroeconomic slowdown brought about by other causes, namely the slowing in Germany and other trading partners. There seems to be little evidence that households that have rapidly expanded both sides of their balance sheet will autonomously decide to contract it again, cutting back on their consumption and thereby generating a general economic slowdown. Rather, it seems households only re-evaluate their balance sheets when they are forced to by a macroeconomic slowdown. While this may not be a desirable pattern of behaviour in cases where household balance sheets are in genuine need of repair, it might provide some comfort that an economy-wide contraction is not the inevitable outcome of a substantial increase in property prices.

Footnotes

See, for example, some of the submissions made to the 2003 Productivity Commission Inquiry into First Home Ownership in Australia, at <http://www.pc.gov.au/inquiry/housing/subs/sublist.html>. [15]

This is not to say that such regulations do not provide other social benefits. [16]

At least part of the variation in price growth across cities reflects differences in population growth, that is, the contribution of demand for extra dwellings. [17]

See, for example, Box B ‘The Housing Market Slowdown in the Netherlands’ in the RBA's September 2004 Financial Stability Review for more discussion of the Dutch experience. [18]