Speech Opening Remarks to Plenary Panel at the Australasian Housing Researchers Conference
Assistant Governor (Economic)
Australasian Housing Researchers Conference
I'd like to thank RMIT and the organisers of this conference for inviting me to participate in this session. I'm very glad to be here and to share the stage with Dr Rogers.
Looking at the program, I see many interesting papers. A huge range of housing-related topics are represented, many well outside the Reserve Bank's remit. But housing, as a sector, is crucial to a number of our core policy functions.
In financial stability, my old role, housing is certainly crucial. Housing is the biggest asset most households will own and the bulk of household sector wealth. Lending against housing is also a large part of financial sector assets. Housing outcomes are therefore central both to the welfare of households and the stability of the financial system.
The housing sector is also crucial on the monetary policy side, my new role. As a sector, it is especially sensitive to interest rates. It is a key part of the transmission mechanism of monetary policy. Housing prices respond to interest rates, and in turn affect household consumption through wealth and collateral effects. And the housing construction industry is one of the most interest-sensitive.
The theme of this conference is ‘Housing and Inequality’. I'd like to talk about two aspects of that. First, there's the inequality – or rather, diversity – between countries. Second, there's the inequality represented by the distribution of housing outcomes across residents of a particular country.
Institutions Matter: The Cross-country Dimension
On the first, we can summarise by saying that institutions matter. Country-specific features are very important to housing outcomes, including the outcomes most relevant to monetary and financial stability policymakers.
- A range of institutions determine the idiosyncratic risk that households face, and thus the credit risk they pose to lenders. These include labour market institutions, the funding of health care, and the structure and generosity of the social safety net.
- Geographic and other factors determine the urban structure, which in turn affects the relative price of housing. We know from economic geography that as city populations rise – and the amenities that go together with that – housing prices rise even relative to the higher incomes that often go along with big-city living. So urban concentration affects macro-level variables, like the ratio of housing prices to household incomes, as well as individual-level outcomes, such as people's lifetime housing experiences.
- The structure of the mortgage market matters, for example whether fixed or variable-rate mortgages predominate. These features are in turn affected by a range of tax and other institutional factors. An example of a feature quite specific to Australia is the popularity of linked offset accounts. These change the interpretation of debt data that do not adjust for balances in these accounts. They also change assessments of the resilience of the household sector to various kinds of shocks.
Taken together, these factors suggest that we can't assume a single value of any particular macro-level ratio exists that is ‘right’ for all countries or for all time. It depends. We can't assume that a particular level of debt is always and everywhere sustainable. It depends on who has the debt and what kind of credit risk they pose. We can't assume that a particular level of prices is ‘correct’ or ‘sustainable’ in all circumstances.
What we can do is get some sense of the relativities between countries that you might expect, given those institutional and other differences. For example, we can reasonably expect that countries where much of the population lives in smaller, cheaper cities will have lower national aggregate ratios of housing prices to incomes than other countries. That might partly explain why the price-to-income ratio for the United States is relatively low (Graph 1). By contrast, Australia is somewhere around the middle of the pack of mid-sized countries on this metric.
Similar comparisons of household debt-to-income ratios across countries also put Australia in the middle of the pack. The United States is lower, consistent with the lower average price-to-income ratios there. Several small European countries, such as Sweden and the Netherlands, have much higher ratios than Australia does. I would not expect or want Australia's ratio to be as high as it is in those countries. Their institutions are different. In particular, a wider range of debt is tax-deductible in those countries, so households have less incentive to pay debt down.
Distributions Matter: The Micro-level Dimension
The concept of inequality perhaps of most interest to this conference, though, is the differences across households. The way people access housing is not the same. These distributional issues can be considered over several dimensions.
- We can consider the distribution of outcomes in the cross-section of households: how outcomes differ for different types of households at a point in time.
- We can consider outcomes for different cohorts: whether the experience of people in a particular age group changes over time.
- We can consider issues of intergenerational distribution: whether outcomes for people in one generation depend on the experiences of their parents.
It turns out that these sorts of distributional issues matter for policy. Certainly they matter for financial stability policy. It is rarely the median borrower who defaults. Risk comes from the tails of the distribution. But even something as ‘macro’ as the transmission of monetary policy turns out to depend strongly on distributional considerations (Hughson et al 2016). So the Reserve Bank spends a lot of time and resources examining household survey data and other micro-data to inform our macro understanding; some of that work has explicitly touched on inequality as the subject of interest (Dollman et al 2015), while other work has noted the effect of macro-level developments on inequality (La Cava 2016). The rest of my talk today will cover a few interesting findings from this type of analysis. I'm not setting out to be comprehensive here, because that would not be feasible in the time available. But hopefully it will give a flavour of what's possible and the kinds of issues we've looked at.
- Firstly, and unsurprisingly, housing ownership and housing wealth are not equally distributed across the cross-section of households.
- Secondly, fewer young people are home owners now than previous cohorts at the same age. This is partly about demography. If it were purely a story of people being priced out, other outcomes would be different.
- Finally, concerns about access to appropriate housing aren't only about ownership. We should also think about how housing is experienced, including security of tenure.
One issue of inequality that generates a lot of public discussion is whether people can purchase their own home. Housing affordability is clearly a concern more generally, because adequate shelter is so important to human welfare. But clearly there is great social interest in how that housing is delivered, who owns it, and who pays for it. Of the many inquiries into housing affordability over the past decade or so, at least two – the Productivity Commission's 2003 Inquiry and the House of Representatives Standing Committee on Economics Inquiry in 2015 – were specifically on home ownership.
At a national level, Australia actually has a home ownership rate that is neither unusually high nor unusually low. Many of the countries that have higher rates seem to manage it by having very high proportions of young adults living with their parents, even into their thirties, rather than renting. But the outcomes do vary across households. We can see this in several independently collected data sets, but for today I'm going to rely mainly on the HILDA survey.
Unsurprisingly, higher-income households are more likely to own their own home, whether with a mortgage or outright (Graph 2). The highest-income group are not only more likely to live in their own home; the wealth represented by their own home, net of debt, is also higher than for lower-income and middle-income households.
More important than income, though, is life stage. Not many teenagers or young adults are heads of households; those that are, are not that likely to own their own home. Over time, people do start purchasing their home. Home ownership rates rise rapidly with age, reaching 80 per cent for households headed by someone aged 55 or more (Graph 3). Housing wealth has the same pattern, largely because the older you are, the longer you have had to pay your mortgage down.
Because higher-income households are more likely to be owner-occupiers, they are also more likely to have mortgage debt (Graph 4). This is why we say that most of the mortgage debt in Australia has been borrowed by those most able to service it. That said, the lower-income households that do have debt, tend to have quite a lot of debt relative to their incomes. Sometimes this is because they have temporarily low incomes, but overall it does speak to the need to be aware of pockets of potential stress within a more benign overall picture.
Participation in the housing market need not be about owning your own home. Many people rent; someone else has to own those dwellings as well. In Australia, most private rental properties are owned by other households. In other words, there are not many companies operating as residential landlords. A good indication of who owns rental properties comes from looking at people who have housing debt other than debt on their own primary residence. In this graph, we have split up the share of people who have any housing debt into those who have only debt on their own home, and those who have some other kind of housing debt; most, but not all, of this second group also have owner-occupier debt. As you can see, this other property debt is even more likely to be held by high-income earners than owner-occupier debt is.
The pattern of rising ownership rates for older age groups does not give everyone comfort that most people will be able to afford a home eventually. People worry that it is becoming harder to achieve home ownership at a given life stage than it used to be. This would be an issue of inequality across cohorts. It is true that home ownership rates have fallen for most age groups over time (Graph 5). The overall ownership rate has stayed fairly steady because the population is ageing. It is also true that ownership rates shifted down between 2006 and 2011; we'll know what happened in 2016 when the Census data come out in a few months.
The 25–34 age group is typically seen as the core group of first home buyers. We can see that ownership rates in this group have fallen by a bit more than 10 percentage points since the 1970s. But we can also see that most of the decline had already happened by the early 1990s, before the big increase in housing prices relative to incomes. What changed during that earlier period was that people started partnering and settling down later in life. The post-war Baby Boom had been characterised by an anomalously young average age at first marriage. Through the 1970s and 1980s, the normal historical pattern started to reassert itself. Ownership rates for young people declined as a result, because many people wait to settle down before they buy a home.
By saying this, I'm not suggesting that people should not worry about whether households can achieve their desired housing tenure. I am suggesting that the situation is more complex than would be suggested by a single-minded focus on a single metric of affordability, such as median housing prices.
Much of the commentary around the difficulty of achieving home ownership centres on the task of accumulating the deposit. It seems common to assume that everyone needs a 20 per cent deposit when they first buy. This isn't actually true: although lenders require some deposit coming from genuine savings, it doesn't have to be as high as 20 per cent. So it's surprising that as housing prices have risen, the distribution of loan-to-valuation ratios – the converse of the deposit – hasn't shifted up over time (Graph 6). If anything it has declined in the past few years for which we have data. It's not entirely clear why this is. And because a high loan-to-valuation ratio does imply higher risk both for the borrower and the lender, it might not be such a bad thing. But it does suggest that, again, the situation is more complex than a simple summary statistic can capture.
One reason why first home buyers haven't needed to increase loan-to-valuation ratios might be that more of them are getting help from friends and family to accumulate the deposit. Careful analysis of the HILDA Survey dataset shows that the share of first home buyers receiving that help has been increasing over the decades, but it actually remains low (Graph 7).
It is common to focus on the different outcomes for owners versus renters in purely financial terms. Another aspect of inequality I would like to talk about today is security of tenure. A range of different data sources confirm that although young people move more often than older people, the big difference is between renters and owners (Graph 8). A similar pattern occurs in countries such as the United States (Bachmann and Cooper 2014).
Looking just at the group of households who can be tracked through the whole life of the HILDA survey, and who didn't switch between owning and renting at any stage, you can also see that renters were also more likely to have moved many times in that 13-year period (Graph 9).
Some earlier literature has suggested that transaction costs induce home owners to move less often than they might otherwise do, which makes it harder for them to take advantage of changing labour market opportunities (Oswald 1999). That effect doesn't seem clear in Australian data, so I would not draw that conclusion here (Flatau et al 2002). Yet even without the transaction costs of selling a home – including tax, legal fees and agent commissions – we know that moving house can be disruptive and costly. So I question whether all those moves by renters were desired by those households. Many renters are happy with their current home, but are required to move because the lease expired or the landlord sold the property. If we are concerned about inequality of housing outcomes, perhaps we should focus less on the type of tenure, and more on security of tenure.
That's just a brief round-up of some of the issues we have been looking at recently. Housing issues will always be with us, and will always be important to the functions of the Reserve Bank. That's why we value conferences like this: the diverse perspectives offered help us see the bigger picture. I'm certainly looking forward to the rest of the discussion in this session.
Thank you for your time.
Bachmann R and D Cooper (2014), ‘The Ins and Arounds in the U.S. Housing Market’, Federal Reserve Bank of Boston Working Papers 14-3.
Dollman R, G Kaplan, G La Cava and T Stone (2015), ‘Household Economic Inequality in Australia’, RBA Research Discussion Paper No 15.
Flatau P, M Forbes, G Wood, PH Hendershott and L O'Dwyer (2002), ‘Home Ownership and Unemployment: Does the Oswald Thesis Hold for Australian Regions?’, Murdoch University School of Management and Government Working Paper 189.
Hughson H, G La Cava, P Ryan and P Smith (2016), ‘The Household Cash Flow Channel of Monetary Policy’, RBA Bulletin, September, pp 21–30.
Oswald AJ (1999), ‘The Housing Market and Europe’s Unemployment: A Non-Technical Paper', University of Warwick, mimeo, unpublished manuscript.
RBA (Reserve Bank of Australia) (2015), ‘Submission to the Inquiry into Home Ownership’, House of Representatives Standing Committee on Economics Inquiry into Home Ownership, June.
Yates J (2011), ‘Explaining Australia's Trends in Home Ownership’, Housing Finance International, 26(4), pp 6–13.
Thanks to Stephanie Parsons, Hannah Leal and Tahlee Stone for assistance with the graphs. [*]
Graphs showing this comparison are regularly published in the Financial Stability Review. 
See Graph 3 in RBA (2015). 
Who ‘heads’ a household is a matter of judgement. In these graphs, the head of each surveyed household is determined by applying certain criteria, in order, until a unique person is selected. These criteria are: in a registered marriage or de facto relationship (and still living together); a lone parent; the person with the highest income; the eldest person.