RDP 2006-12: Housing and Housing Finance: The View from Australia and Beyond 2. Global Trends

2.1 Disinflation, Deregulation and Financial Innovation

One of the most important common factors driving housing developments internationally has been the wave of deregulation and product innovation taking place in financial sectors in most countries. This has reduced interest margins on housing loans, lowering real interest rates paid by mortgage borrowers. Greater competition and product innovation has also encouraged lenders to make finance available to a wider range of potential borrowers than before. At the same time, declines in inflation in a number of countries over the past decade and a half have lowered nominal interest rates, thereby magnifying these effects.

A BIS working group on housing finance recently drew some of this evidence together (BIS 2006).^[1] The main theme of the findings of that group was that globalisation of financial markets, and particularly innovations in funding and risk management, had resulted in a substantial expansion in the supply of mortgage loans. Table 1 summarises some of the major developments across countries.

As an illustration of the role of increased competition in lowering mortgage interest rates, Figure 1 shows the evolving difference between indicator rates advertised by Australian lenders, and the rates that borrowers actually pay. So-called mortgage managers entered the market in the mid 1990s, funded via wholesale markets and securitisation. As shown in the figure, they were initially offering rates that were well below the standard variable interest rate advertised by the major Australian banks. The banks were forced to respond to this competition, and margins on mortgage rates relative to the cash rate narrowed considerably over this period. Since then the indicator rates advertised by the two classes of lender have shown a reasonably stable spread, but the prevalence of discounting from these rates has increased. Data on the average rate new borrowers actually pay are only available with a lag, but show that this rate is now below the average standard variable rates of all major lenders. Lower mortgage interest margins increase borrowers' capacity to pay at any given level of the policy interest rate, and at the same time make mortgage borrowing accessible to a wider range of households, for a given level of housing prices.

2.2 Mortgage Tilt and Disinflation

An important part of this increased capacity to pay reflects nominal developments, not just real ones such as the squeezing of interest margins on mortgage loans. It is well-known that lending markets involve information asymmetries, so lenders do not know exactly who the good credit risks are. One way lenders traditionally dealt with this is that they imposed lending limits based on repayment ratios. That is, they determined how much they would lend to a particular borrower by working out the ratio of the initial required repayment to the borrower's income. The ratio was chosen to ensure that the repayment was a manageable obligation for a well-intentioned borrower. One consequence of this credit-rationing practice is that, because the nominal interest rate determines the size of the repayment (not the real interest rate), it also determines the maximum loan size granted.

A substantial disinflation therefore increases borrowers' capacity to pay by reducing nominal rates even when real rates remain constant. This seems to have been an important driver of the increased average sizes of new mortgages and higher housing prices in a number of countries. Australia and New Zealand were particularly affected by this process given the extent of the reduction in inflation that occurred in these two countries in the 1980s and 1990s.

Figure 2 shows the effect of this on the maximum loan size available to a potential borrower. The top two panels present the standard characteristics of a fixed-term amortising (credit-foncier) loan: the amount outstanding declines at an increasing rate, as the interest component of the constant total repayment falls and the share that goes to pay off the principal can therefore rise.

The bottom two panels of Figure 2 show the effects of disinflation. The line in the left-hand panel traces out the loan sizes that generate the same repayment as for a loan of $100,000 at 10 per cent interest per annum, paid monthly over 20 years. In other words, if the borrower could afford a loan of $100,000 when the interest rate was 10 per cent, she could afford to service a loan of nearly $160,000 if rates were to fall to 4 per cent. This relationship is slightly non-linear but less than proportionate to the change in the interest rate. In other words, a fall in rates from 6 per cent to 5 per cent makes more difference than one from 10 per cent to 9 per cent, but halving the interest rate less than doubles the maximum loan size.

This effect also means that the higher nominal interest rates are, the more front-loaded the repayment burden; this front-loading is known as mortgage tilt. This is shown in the lower right-hand panel of Figure 2: assuming lower inflation implies lower nominal income growth, the repayment-income ratio diminishes more gradually. It is not clear if household behaviour fully adjusts to this fact after a disinflation. It may take time for them to recognise that the burden of a given repayment stays high for longer when inflation, and thus nominal income growth, is low (see Modigliani 1976 and Stevens 1997 for more discussion of this point).

The net result of this property of amortising loans is that as inflation falls, aggregate debt-income ratios rise. This occurs partly because capacity to pay has risen substantially, and partly because the ratio of remaining debt to income falls more slowly over the life of each individual loan. Earlier Bank work (RBA 2003a) shows the effect of changes in nominal rates and income growth on the equilibrium aggregate debt-income ratio for the whole household sector, given various assumptions; Ellis (2005) presents an analytical expression for the same ratio.

2.3 Supply of Housing is Inherently Sticky

The combination of disinflation, deregulation and financial innovation can generate a substantial boost to the supply of housing loans, thereby stimulating the demand for housing. For example, the data presented in Figure 2 showed that a fall in inflation and interest margins similar to that experienced in Australia over the 1990s could increase individual homebuyers' capacity to pay by as much as 60 per cent. While not everyone will increase their borrowings immediately, many first-home buyers and existing owners have availed themselves of their greater borrowing capacity over time.^[2] Thus demand for housing in dollar terms could increase by this order of magnitude within a few years.

It seems reasonable at this point to ask why households would increase their mortgage to the full extent of the expansion in their capacity to pay, rather than using some of the savings from lower mortgage repayments to purchase other goods and services. One reason might be that households have preferences such that they maintain their expenditure shares constant as relative prices shift; a Cobb-Douglas utility function would have this characteristic. In the context of housing, they would then maintain their mortgage repayments (flow of expenditure) constant as a fraction of income when interest rates changed, and expand their borrowings one-for-one with their increased capacity to do so. A more compelling reason might be that households remain credit-constrained even after some relaxation of these constraints, so they would still choose to borrow up to the maximum allowable.

The supply of housing is inherently slow to adjust, and would certainly not be able to adjust quickly to a surge in demand of this size. The increase in demand is for the whole housing stock, because it affects (almost) the whole household sector. The available supply of housing is the existing stock, which is fixed, plus whatever building and renovating work is done over a given period. So the only increment to supply is the flow of new dwellings and renovations of existing dwellings, which represents just a few percentage points of the size of the stock (Table 2).^[3] Even the most flexible and least regulated construction sector would struggle to lift its output from something equal to a few percentage points of the dwelling stock to accommodate a surge in demand of 50 per cent or more.

It is therefore inevitable that housing prices would rise in the face of such a surge in demand. Part of this would take the form of higher building costs as renovation work increases, and part would reflect a higher average quality of dwelling as the stock gets renovated. But much of it would feed through to the price of existing, unrenovated dwellings, and implicitly the price of land. Transactions in the market should be expected to rise, as households try to express their increased demand for housing services by moving to a more desirable location.

Comparatively little of the increased supply would take the form of additional new dwellings. Household formation rates may well increase in the face of an increase in the capacity to pay, but this would be a small effect compared with a change in the demand for housing services by existing households.^[4] This also demonstrates the importance of distinguishing between the number of dwellings and the average quality of those dwellings when trying to disentangle supply of, and demand for, housing. Adding a large number of extra houses of a given quality does little to meet the demand of existing households for a higher-quality home than they already have.

2.4 Result: Higher Housing Prices

The increase in the relative price of housing that occurs as a result of such an increase in demand is in large part an equilibrium phenomenon. The average quality of housing will rise over time and absorb some of the increase in demand, but this should be expected to take a very long time. And even once this process has worked through, median and average house prices will be higher because average quality has risen.^[5]

Both sides of the household balance sheet will expand relative to income as a result of this transition. Assuming households own most of the dwelling stock, either as owner-occupiers or landlords, the value of their holdings of housing assets will increase as prices rise. The amount of debt funding should also be expected to increase relative to income in order to fund this more expensive housing, even if the gearing on these housing assets does not rise much. This certainly seems to have been the experience in Australia, as well as in a number of other countries. Housing prices have risen much faster than consumer prices in many countries over the past decade (Figure 3) and ratios of debt to income have also increased (Figure 4; see also Debelle 2004).

However, the nature of the transitions can vary a lot between countries, for example in terms of their speed and duration. The consequences have also varied: some of these transitions (UK in the late 1980s, Netherlands in the late 1990s, US and, arguably, Spain in the early 2000s) were associated with significant increases in the owner-occupation rate. Others have tended to squeeze out first-home buyers, perhaps reducing ownership rates in younger age groups a little; the recent episodes in Australia and the UK seem to be examples of this. Likewise, some of these transitions have involved a speculative element, driven by investors' expectations of future capital gains, resulting in a boom/bust cycle, while others seem to have involved little speculation activity. These differences probably reflect the many national differences between housing markets. The next section discusses these differences in detail, along with their probable effects on the expansion in housing demand seen across many countries over the past decade or so.

Footnotes

For more details on the housing finance systems of the individual countries represented in the working group, see the supplementary material posted with BIS (2006). [1]

Some older households that own their homes outright are probably less affected by this change in incentives, but most existing owners have the option to refinance or renovate even if they do not move. [2]

These figures exclude depreciation and demolitions, so the net new supply of housing services is even smaller relative to the stock than is shown in Table 2. [3]

Increased household formation from an existing population may actually reduce average demand for housing services a little, since it would reduce the average number of persons per household. [4]

The desire for higher housing services, and resulting higher average quality of dwellings, will also be picked up in repeat-sales indices and price indices based on hedonics. This is because it is not possible to have complete information about all the characteristics of a property that affect the price, and which can be changed in a renovation. [5]