Submissions – Housing and Housing Finance Joint RBA-APRA Submission to the Inquiry into Home Lending Practices and Processes

Introduction

This submission brings together the factual material available to the Reserve Bank of Australia and APRA which is relevant to the four questions asked by the Committee. The material mainly relates to questions 1, 2 and 4. The information collected by the Reserve Bank and APRA in pursuit of their policy responsibilities does not extend in any great detail to the issues raised in question 3; we do, however, outline information on the extent to which households are accessing superannuation.

The main conclusions we draw from the material presented here are as follows:

• Competition in the home loan market has increased markedly in recent years. This has reduced the cost, and increased the availability, of housing finance to households. Financial innovation has played a big role in this, with lenders introducing new products to cater for a wider range of potential borrowers and finding new ways to assess their borrowing capacity. Some of this innovation is often described as a ‘reduction in lending standards’, a term which has negative connotations. While it is true that some of the innovation has resulted in an increase in risk for both borrowers and lenders, its overwhelming effect has been to widen the range of households who can get access to finance.
• The greater availability of credit means that, for a given level of unemployment and interest rates, a higher share of loans is likely to be in arrears than in the past. While arrears rates have risen somewhat recently, they remain low by historical and international standards.
• Overall, the household sector remains in sound financial shape, supported by the ongoing strength in the economy. Notwithstanding this, there is a small proportion of borrowers whose finances are overextended.
• Any impact on the stability of the Australian financial system from changes to lending standards and the increased availability of credit is likely, on current assessments, to be small. Various stress tests of the financial system over recent years have indicated its resilience to events that are well outside historical experience.

Question 1: To what extent have credit standards declined in recent years?

Before addressing this question, it may be useful to provide some background on developments in the home loan market. The key point is that competition in the housing finance market has increased considerably over the past decade. Competition was boosted in the mid 1990s by the entry of specialist mortgage lenders. More recently, the emergence of mortgage brokers, who act as intermediaries between lenders and borrowers and make it easier for borrowers to compare the costs and features of different loans, has heightened competition between lenders. Foreign banks have also become more active in the housing finance market during the past few years.

Interest margins (relative to the cash rate) on prime variable-rate housing loans have fallen from a peak of around 450 basis points in 1993 to an average 120 basis points at present (Graph 1). During the 1990s, most of the competition-induced decreases in the cost of housing finance were in the form of reductions in the margin between indicator lending rates and the cash rate. More recently, competition has manifested itself in increases in the size and availability of discounts being offered on housing loan indicator rates, rather than changes in indicator rates themselves. The average new borrower is now paying 60 basis points below the standard variable indicator rate.

As well as reductions in interest margins, establishment fees on housing loans have fallen, and are now regularly waived by lenders, particularly when the loan is part of a ‘professional package’ (which combines discounts on the housing loan interest rate and on other banking products in return for an annual fee of around $200–$300).

Competition has also led to various product innovations, many of which were first introduced by non-mainstream lenders but which are now widely available. For example, lenders have introduced home-equity loans, redraw facilities and reverse mortgages, all of which allow households to borrow against the equity they have built up in their homes. Lenders have allowed higher loan-to-valuation (LVR) loans and introduced interest-only loans and shared-equity loans, which make it easier for households, particularly first home buyers, to purchase their home.

On the specific question of lending standards, it is the case that authorised deposit taking institutions (ADIs) and other unregulated lending institutions have made a number of changes to their lending practices over recent years. These have included changes to their debt serviceability criteria; introducing a wider range of loan products; increasing their reliance on mortgage brokers to originate loans; and making greater use of alternative valuation methodologies.

Debt serviceability criteria

Until about the mid 1990s, a common rule-of-thumb was that lenders were willing to lend up to the amount where the debt-servicing ratio – the ratio of interest and principal repayments to gross income – was 30 per cent. It is difficult, however, to know the extent to which this was a constraint upon lending; it was not universally applied, and some borrowers took out a second mortgage that attracted a higher interest rate to finance their property purchase.

Over recent years, this rule-of-thumb has been relaxed. There has been a change in the way that lenders undertake their loan assessments from simple rules based on gross income to ‘net-income surplus’ models. These models require borrowers to have a minimum amount of after-tax income after taking account of debt-servicing requirements (including some allowance for an increase in interest rates) and a basic level of living expenses. Based on these models, lenders' online housing loan calculators suggest that lenders are now willing to provide loans with debt-servicing ratios up to around 50 per cent for higher income borrowers.^[1]

While some borrowers utilise the higher allowable debt-servicing ratios, the majority of borrowers taking out new loans continue to have debt-servicing ratios significantly below the allowable limit. APRA survey data show that, in September 2006, the most common debt-servicing ratio for new borrowers was 21–25 per cent, and three-quarters of these borrowers had debt-servicing ratios of less than 30 per cent; only 9 per cent had ratios greater than 40 per cent (Graph 2). There are no historical data to allow a comparison of how these figures have changed over time, but it would not be surprising if the proportion of households with higher ratios had risen, reflecting a willingness by households to steer a greater proportion of their rising disposable income to housing. It is also important to note that, in aggregate, real household disposable income, after allowing for mortgage interest payments, is continuing to grow solidly, which indicates that households as a whole are doing well (Graph 3).

A wider range of loan products

Lenders have also introduced new loan products that better meet the needs of certain types of borrowers, such as those with irregular income streams or those who do not meet the standard lending criteria due to impaired credit histories. These products include: low-doc loans, high LVR loans, interest-only loans and non-conforming loans. As well as increasing their availability, lenders have also been reducing the interest rate premiums that they charge on these loans.

Low-doc loans, for which borrowers self-certify their income in the application process, have become more common over recent years. In 2006, 10 per cent of newly approved housing loans were low-doc, compared with less than ½ per cent in 2000 (see Table). Low-doc loans currently account for almost 7 per cent of all outstanding housing loans. These loans, designed mainly for the self-employed and those with irregular income who do not have the documentation required to obtain a conventional housing loan, have benefited borrowers who, in the past, would have been denied access to finance. However, the process may be more open to abuse by some who overstate their income to obtain a larger loan.

There has also been an increase in non-conforming loans for borrowers who do not meet the standard lending criteria of mainstream lenders. These borrowers typically have poor credit or payment histories. In 2006, non-conforming loans comprised 2 per cent of new loan approvals, compared with less than ½ per cent in 2000. Almost all non-conforming loans are provided by lenders that are not subject to prudential regulation. (The Attachment highlights the differences between non-conforming loans in Australia and sub-prime loans in the US.)

Maximum allowable LVRs on traditional prime, full-doc housing loans have risen, with low – including zero – deposit home loans now offered by a wide range of mainstream lenders. Survey data from APRA indicate that 15 per cent of loans approved in September 2006 had LVRs greater than 90 per cent, and that loans with LVRs in excess of 95 per cent have increased in recent years (Graph 4). Most lenders do not charge an interest rate premium for these high-LVR loans, though the lenders' typical requirement for borrowers to take out lenders' mortgage insurance pushes up the overall cost of these loans to borrowers.

A few specialist lenders offer loans with maximum LVRs of around 105 per cent. These loans allow borrowers with little or no deposit savings to purchase a house, and cover the associated costs such as legal fees, stamp duty and loan extension fees. These loans have higher interest rates than standard housing loans.

Interest-only loans have also become more common, particularly for investors. These loans do not require borrowers to make any repayments of principal for up to 10–15 years (after which the loan typically converts to a principal-and-interest loan), and hence have an initial repayment amount that is lower than on a principal-and-interest loan. We estimate that, in 2005, a little over 15 per cent of new owner-occupier loans were interest-only, up from 10 per cent in 2003.^[2] The share of investor loans that are interest-only is much higher, at around 60 per cent in 2005. This presumably reflects the tax deductibility of interest payments on these loans. Overall, in 2005, interest-only loans accounted for around 30 per cent of new housing loans and a slightly lower share of outstanding loans.

Increased reliance on mortgage brokers

Lenders' reliance on brokers to originate loans has also risen. Mortgage brokers are heavily used by those lenders which have limited branch networks, such as the smaller regional banks or other ADIs. Broker-originated loans are estimated currently to account for a third of new housing loans, up from a quarter of new loans in 2003.^[3] This is a potential source of risk for lenders, as the link between borrower and lender is weaker, and brokers' incentives may be aligned more closely with the volume of loans rather than their quality. (This has been an issue in the US sub-prime market.) In addition, APRA has found that some lenders were less diligent in verifying borrower information on broker-originated loans than they were on branch-originated loans and is addressing this issue through its routine supervision process.^[4]

Greater use of alternative valuation methodologies

There has been a trend towards the use of lower cost alternative valuation methodologies in credit assessment. Traditionally, valuations were based on a full external and internal inspection of the property. But an APRA survey of approximately 100 lenders in December 2004 found that around one third of the valuations requested by lenders were based on only an external inspection, or were conducted off site using information from the contract of sale, Valuer General records, or desk-based electronic methods. There is anecdotal evidence that this share has risen further in recent years. These alternative valuation methodologies tend to be used by larger lenders for fully documented, low-LVR mortgages. The greater availability of data on housing prices in recent years may have been one factor reducing lenders' reliance on physical inspections.

Question 2: Have declining credit standards caused an increase in the number of loans in arrears and in the number of repossessions?

The level of loan arrears in Australia remains low by international standards, though it has increased over the past few years. To some extent, this increase reflects credit being made available to a broader range of borrowers. The vast bulk of these borrowers are meeting their loan obligations.

As at end March 2007, 0.38 per cent of the value of housing loans on the banks' domestic balance sheets were classified as non-performing. (Comparable figures for building societies and credit unions were slightly higher). While this has increased since the low point in 2003, it remains lower than was typical in the past (Graph 5). Of these non-performing loans, banks report that around three quarters are fully covered by collateral.

The arrears rate for housing loans underlying residential mortgage-backed securities (RMBS) has increased by more than those on banks' books. As at end April 2007, the 90-day arrears rate for all securitised loans (regardless of lender) stood at 0.47 per cent, up from 0.16 per cent in early 2004. (While this ratio is higher than in the mid 1990s, the RMBS market at that time was very small.)

The higher arrears rate for securitised loans partly reflects the larger share of low-doc loans in the pool of securitised loans: the 90-day arrears rate for securitised low-doc loans was 1.07 per cent in April 2007, compared to 0.40 per cent for securitised full-doc loans. The RMBS data also show that non-conforming loans represent a significant step up in risk, with 6½ per cent of these loans currently in arrears by more than 90 days.

The pattern is similar in terms of the numbers of loans in arrears. We estimate that in March 2007, roughly 12,000 of the 5½ million loans in Australia were in arrears by at least 90 days. This represented an arrears rate of 0.22 per cent, lower than the arrears rate by value, as loans that fall into arrears tend to have been taken out over the past few years (and so are larger). Consistent with the data on the value of loan arrears, arrears rates are higher for non-conforming and prime low-doc loans (Graph 6).

There is a noticeable geographical dispersion in housing loan performance. Data drawn from securitised full-doc loans show that arrears rates are higher in New South Wales than in other states (Graph 7).

The bulk of loans that go into arrears eventually get back on track, often with the lender's assistance. Even in those cases where the borrower is unable to remedy the arrears, it is more common for the loan to be refinanced with another lender (often a non-conforming lender) or for the property to be voluntarily sold, than for the lender to seek repossession.

The data available on property repossessions is relatively limited. The information available suggests a pick up in the number of court applications for property repossession in the past few years. Applications for repossessions are above their level in the 1990s even though arrears on loans are lower. This suggests there may have been some change in the behaviour of some lenders in terms of seeking repossession, which may in turn be associated with lenders moving into new loan products, such as non-conforming loans. There are anecdotal reports that some lenders, particularly non-traditional newer entrants to the market, are acting faster to obtain and execute repossession judgments.

The data on court applications for repossessions cover all types of property. They include residential properties – including properties used as security for small business loans – as well as some commercial properties. In New South Wales and Victoria, the number of applications increased by around 50 per cent in 2005 compared with the previous year, and by about 10 per cent in 2006 (Graph 8). Some more recent data are available for Victoria, and these suggest that there has not been a further increase in the number of applications in 2007.

It should be noted that the figures on court applications for repossession – 5,368 in New South Wales in 2006 and 2,720 in Victoria in the year to June 2007 – overstate the number of repossessions that eventually occur, as the borrower can still voluntarily choose to sell the property or refinance the loan with another lender before the existing lender has gone through all the remaining legal steps to take possession.

There has been an increase in the number of personal administrations (mainly bankruptcies and personal debt agreements) in the past couple of years. New South Wales has accounted for the bulk of the national increase. Caution needs to be exercised in interpreting these trends given that a reasonably large proportion of personal administrations probably relate to personal credit problems rather than mortgages directly.

In summary, while there has been an increase in housing loan arrears in recent years, the figures remain low by historical and international standards. Part of the rise is due to the fact that lenders have broadened the range of households to which they are prepared to extend credit. This is a structural change taking place in the housing loan market. The softness in economic conditions in some regions – for example Western Sydney – may also have contributed to the rise in arrears, but the vast majority of households are in a strong financial position with rising incomes (even after mortgage repayments), higher levels of confidence and solid growth in spending.

Question 3: Are borrowers in financial difficulty being treated appropriately by lenders?

As noted, the limited information available to the Reserve Bank and APRA does not allow us to form a clear view on this question. APRA does collect data on the extent to which households in distress are accessing their superannuation.^[5] These data are shown in Graph 9. Superannuation members may apply to APRA for the early release of their benefits for a number of reasons, including helping to meet mortgage repayments to prevent a lender from taking possession of a member's principal place of residence. In 2006, APRA approved 13,871 applications in total for the release of $135 million, compared with 10,459 applications for the release of $77 million the previous year. The approvals in 2006 were equivalent to around 0.05 per cent of the number of superannuation accounts and about 0.01 per cent of the value of all superannuation assets.

Question 4: Are declining credit standards likely to have any long-term implications for the Australian financial system?

Housing lending has traditionally not been a source of significant risk for Australian ADIs. Nonetheless, APRA has been monitoring banks' risks carefully as greater competition among lenders over the past decade and the associated change in lending standards mean that the past may not be a particularly good guide to the future risks facing lenders. Analysis by APRA and the RBA suggests that developments to date do not pose a significant risk to the stability of the Australian financial system. This conclusion is based on a series of surveys, stress tests, and research projects which shed light on the potential for ADIs to suffer significant losses from their housing lending activities.

In 2002 and 2003, APRA conducted an extensive stress test covering most ADIs in the housing market.^[6] Despite the fact that the stresses applied in this test were well beyond historical experience in Australia – including a fall of 30 per cent in property prices – no ADI would have failed as a result of this stress, although the capital bases of a small number of ADIs would have fallen below their APRA-mandated minimum.

In 2004 and 2005, APRA extended its stress testing to lenders' mortgage insurance companies (LMIs). These culminated in APRA adopting higher and more risk-based minimum prudential capital requirements for LMIs.

In 2005 and 2006, the International Monetary Fund, APRA, RBA and Treasury stress-tested the five largest Australian banks (which account for about 65 per cent of housing loans) as part of the Financial Sector Assessment Program. The exercise comprised a macroeconomic stress test involving a large fall in house prices (down 30 per cent), a marked rise in the unemployment rate (up 4 percentage points), and banks having to pay more for their funding. The stress tests confirmed the resilience of the Australian banking system; though bank profits dipped sharply, banks remained profitable and well capitalised.^[7]

During 2005 and 2006, APRA also conducted two surveys which examined the extent to which ADIs have become more accepting of risk in their home lending. The first survey, which reviewed ADI lending policies, suggested ADIs would lend more money at higher debt servicing ratios than had typically been the case in the past. The second survey, which was based on a collection of ADI home loan approvals in September 2006, suggested an increase in the risk of mortgage portfolios, though this increase must be seen against the exceptionally low loss levels of past decades.

Data from a lenders' mortgage insurance provider show that annual losses on insured prime housing loans averaged 4 basis points between 1980 and 2000, with the highest loss in any one year being 10 basis points.^[8] These remain easily covered by the average interest margin (relative to the cash rate) on prime housing loans (120 basis points in 2006). This interest margin needs to cover all of the costs of providing a housing loan, including the cost of originating and servicing the loan, not just the expected losses from default. Losses on non-conforming loans are significantly higher than for prime loans, at around 30 basis points in 2006, but so too is the average interest margin, at 290 basis points.

Attachment: US Sub-prime Loans and Australian Non-conforming Loans

The United States sub-prime market is markedly different from the Australian non-conforming loan market (Graph A1).

• Non-conforming loans, the closest equivalent to sub-prime loans in Australia, accounted for only about 2 per cent of new loans in 2006. This was well below the 20 per cent share of sub-prime loans in total new loans approved in the United States.
• Australian non-conforming housing loans have lower LVRs than sub-prime loans in the United States. The average LVR on newly-approved Australian non-conforming loans is around 75 per cent, much lower than the average LVR of about 95 per cent on United States sub-prime loans.
• A feature of many US sub-prime loans is their use of low introductory or ‘teaser’ interest rates for a period before the rate reverts to a much higher standard rate. Also, high-risk repayment options such as negative amortisation periods have been common. These features can expose borrowers to payment shock – a large increase in their mortgage repayments – when the initial introductory interest rate period or negative amortisation period expires. Non-conforming loans in Australia do not have these features.
• The arrears rate on non-conforming loans in Australia, at 6½ per cent, is well below the corresponding rate on sub-prime loans in the United States (Graph A2).

Footnotes

For further details on these calculations, see Box D: Estimates of Borrowing Capacity from Banks' Online Housing Loan Calculators, Reserve Bank of Australia Financial Stability Review, March 2005. [1]

Box B: Interest-only Housing Loans, Reserve Bank of Australia Financial Stability Review, September 2006. [2]

Box D: The Characteristics of Loans Originated by Mortgage Brokers, Reserve Bank of Australia Financial Stability Review, September 2004. [3]

Laker, J Credit Standards in Housing Lending – some further insights, Australian Prudential Regulation Authority, 20 June 2007. [4]

Other than applying to APRA, people can also apply directly to their superannuation fund for early release of their benefits on the grounds of ‘severe financial hardship’. No data are currently collected on these applications. [5]

See Australian Prudential Regulation Authority (2003), ‘Stress Testing Housing Loan Portfolios’, APRA Insight, 3rd Quarter. [6]

See Reserve Bank of Australia (2006), Financial Stability Review, September and International Monetary Fund (2006), Australia: Financial System Stability Assessment, including Reports on the Observance of Standards and Codes on the following topics: Banking Supervision, Insurance Regulation, Securities Regulation, and Payment Systems, October. [7]

The Performance of Australian Residential Mortgage-backed Securities, Reserve Bank of Australia Financial Stability Review, March 2006. [8]