Financial Stability Review – April 2017
Interest-only Mortgage Lending
Interest-only (IO) loans account for a sizeable and growing share of total housing credit in Australia, now representing around 23 per cent of owner-occupier lending and 64 per cent of investor lending (Graph B1). IO lending has the potential to increase households' vulnerability in part due to the higher average level of indebtedness over the life of an IO loan compared with a regular principal-and-interest (P&I) loan.
Measures to address some risks associated with IO lending practices were among those taken in late 2014 by the Australian Prudential Regulation Authority (APRA) and the Australian Securities and Investments Commission (ASIC), in conjunction with the Council of Financial Regulators, to reinforce sound housing lending practices. While the share of IO loans in total lending approvals subsequently declined, IO loans have since started to rise again, especially for investors, which has again attracted regulator attention. In March 2017, APRA announced new measures requiring authorised deposit-taking institutions (ADIs) to limit new IO lending to 30 per cent of total new residential mortgage lending and, within that, to tightly manage new IO loans extended at high loan-to-valuation ratios (LVRs). This box outlines in more detail recent trends in IO lending and the nature of the potential risks that can arise from this type of lending.
Characteristics of and Demand for Interest-only Loans
For P&I loans, the balance of the loan must be paid down over the entire term of the loan. In contrast, for IO loans repayments of principal are not required during the IO period, which is typically the first five to ten years of the loan. Instead, scheduled principal repayments start at the end of the IO period, with the balance of the loan then paid off over the residual loan term. As a consequence, for a typical 30-year P&I loan of $400,000 with an interest rate of 4 per cent, borrowers would be ahead on principal repayments by around $38,000, or about 10 per cent of the initial balance, after five years compared with an IO loan (Graph B2). Further, because the scheduled balance on an IO loan is always higher than on a comparable P&I loan, IO loans incur a greater interest cost over the term of the loan (around 9 per cent extra in the previous example). An IO loan can potentially also be refinanced at the end of the IO period into another IO loan, prolonging the period before scheduled principal repayments start.
Both investors and owner-occupier borrowers make extensive use of IO lending for a range of reasons. Housing investors make the most use of IO loans. Since interest payments on investment loans are tax deductible, the incentive to pay down a loan's principal is reduced. IO loans also enable investors to maintain a higher level of leverage and so magnify potential gains or losses if housing prices rise or fall. As noted earlier, the share of owner-occupier lending that is IO has also risen noticeably over time. This has been due to increasing numbers of owner-occupiers using IO loans and the increasing average sizes of IO loans (relative to P&I loans).
Another reason borrowers may prefer IO loans to P&I loans is because they can offer greater repayment flexibility. Borrowers with lumpy income or those wanting to build buffers or save for planned expenditures, such as renovations, can use IO loans with an offset or redraw facility to minimise the effective interest costs over the period of the loan while still ensuring funds are readily available for other uses. In particular, offset accounts and redraw facilities allow borrowers to effectively amortise loan balances during the IO period and so reduce (or eliminate) the extra interest cost associated with the higher principal balance on IO loans compared with P&I loans. However, borrowers need to be disciplined in their repayment behaviour to receive these benefits; otherwise they may incur greater interest costs, and remain more indebted for longer. IO loans are also routinely used for bridging finance and construction loans to minimise repayments for the short duration of these loans.
For some time regulators have highlighted the potential risks associated with IO compared with P&I loans. Because IO loans allow borrowers to remain more indebted for longer, there may be greater credit risks associated with such loans. When loan balances stay high, there is an increased risk of borrowers falling into negative equity should housing prices decline.
Another risk is that borrowers may find it difficult to service higher required payments at the end of the IO period, which increases the chance of default. For example, repayments on a $400,000 loan with a 4 per cent interest rate and a five-year IO period would typically increase by around 60 per cent at the end of the IO period. While some borrowers may have planned to refinance into another IO loan at the end of the IO period, this may be difficult if circumstances have changed.
Borrowers who anticipate future price rises can use IO loans to maintain a higher level of leverage for a given servicing payment, thereby magnifying their returns from rising housing prices but also magnifying any losses. More generally, at an aggregate level this behaviour could induce a more pronounced cycle in housing prices than would otherwise occur, amplifying the size of any subsequent downswing in housing prices.
In recognition of the higher risks associated with IO loans, some lenders have introduced premiums on advertised interest rates for IO loans. For example, the four major banks have announced, on average, an 18 basis point premium for IO owner-occupier loans and a 15 basis point premium for IO investor loans (in addition to premiums for investor loans relative to owner-occupier loans).
Lenders' practices in assessing the ability of borrowers to repay their loans are important to manage the systemic risks posed by IO lending. These practices determine the maximum loan size that a borrower could sustainably repay out of their income. In particular, APRA serviceability guidance for ADIs sets out prudent practices for IO loans, with the capacity to repay assessed at the higher repayment amount required when the IO period ends (known as the residual-term method). Under this residual-term method, borrowers seeking IO loans receive a lower maximum loan size than would be available for an equivalent P&I loan.
Nonetheless, prior to the 2014 measures, some lenders assessed serviceability based on lower hypothetical P&I repayments calculated from the entire term of IO loans (including the IO period; known as the full-term method). This approach risks borrowers being unable to meet their repayment obligations when the IO period ends and higher repayments commence. This is potentially in breach of the National Consumer Credit Protection Act 2009, which requires that lenders make loans that consumers will be able to repay without undue hardship. ASIC has found that around 40 per cent of loans reviewed in 2014 used the full-term approach; lenders have since undertaken to change their practices in order to meet their responsible lending obligations and APRA's guidance. There have also been some recent reports of borrowers applying for P&I loans to maximise their borrowing capacity, and then switching soon after approval to an IO loan. APRA has recently issued guidance to address this behaviour.
The Council of Financial Regulators agencies are APRA, ASIC, the Reserve Bank of Australia (RBA) and The Treasury. For further details of the measures announced in 2014, see RBA (2015) ‘Box B: Responses to Risks in the Housing and Mortgage Markets’, Financial Stability Review, March, pp 45–47. 
APRA (2017), ‘APRA Announces Further Measures to Reinforce Sound Residential Mortgage Lending Practices’, Media Release No 17.11, 31 March. 
APRA started the regular collection of data on IO loan approvals in 2008, though other data indicate that the IO share of housing credit had been rising for several years prior. In 2003, it was estimated that IO lending accounted for almost 50 per cent of new investor housing loans and a little over 10 per cent of new owner-occupier loans. See RBA (2006), ‘Box B: Interest-only Housing Loans’, Financial Stability Review, September, pp 42–43. 
For example, for a 30-year loan with a five-year IO period, a residual-term serviceability assessment would use repayments based on a 25-year payback period, whereas the full-term method would use the (lower) repayments from a 30-year payback period. An IO loan assessed using the full-term method would result in the same maximum loan size available for an equivalent P&I loan. 
APRA (2017), ‘APG 223 Residential Mortgage Lending’, Prudential Practice Guide, February.