RBA Annual Conference – 2007 Discussion

1. Saul Eslake

Chris Ryan and Chris Thompson have provided a thorough and thoughtful summary of the major developments in the Australian financial system and its interactions with the household and business sectors over the past 15 years.

As they see it, the most important trends to emerge in this period are:

• the substantial increase in household indebtedness, paralleled for much of the period by a decline in business gearing;
• the shift in the composition of household assets towards asset classes that are more exposed to price fluctuations as a result of market movements;
• the correspondingly greater exposure of the financial system to the household sector, and in particular to housing loans (which, in ANZ's case at least, was also partly a conscious strategic choice);
• the greater reliance of the banking system on wholesale funding and, within that category, overseas borrowings; and
• the rapid growth in the funds management industry.

I cannot think of any substantial omissions there.

One of the more important conclusions that they draw from these trends, and one that is particularly apposite given the developments in global financial markets in recent weeks, is that ‘market disruptions [may] have more wide-ranging and detrimental effects than in the past’ and that there may be ‘more frequent bouts of volatility than in the past’.

Another, perhaps more contentious conclusion, to which I want to return anon, is that ‘at the macro level, there is not a great deal that can be done about occasional bouts of mispricing of risk’.

A third important conclusion is that more could be done ‘at the micro level to bolster households' risk management capabilities’.

Along the way they draw a number of other conclusions which, though not highlighted in the same way as the three I have just mentioned, nonetheless seem particularly important in view of contemporary concerns:

• the households that have done the bulk of the borrowing should be best able to service it and appear to be well placed to repay it (a point that often appears to be omitted from public discussions about the trend in household indebtedness over the past decade and a half);
• banks face a lot of competition in the retail market, including from foreign-owned banks and from mortgage brokers, as well as in the business loan market, yet (in contrast to the experience during the 1980s) arrears rates have (thus far at least) remained ‘low by historical and international standards’; and
• banks' risk management techniques are better than in the past, and the Australian banking system is very sound and well-placed to weather adverse events.

The authors appropriately highlight the importance of the macroeconomic environment and financial innovation to the evolution of household balance sheets over the past 15 or so years. There are undoubtedly strong linkages between these two factors, the increase in household indebtedness and the increase in house prices since the early 1990s; and in my view ‘cause and effect’ run in both directions.

The combination of the sharp decline in interest rates during the 1990s and 15 years of strong growth in disposable incomes roughly trebled the maximum amount which a ‘typical’ home buyer could borrow without breaching the ‘rules of thumb’ that lenders typically use to determine the maximum amount which they are willing to lend. In addition, these ‘rules of thumb’ became somewhat more elastic during this period, so the borrowing capacity of the ‘typical’ home buyer more than trebled. Since the stock of housing and the number of households requiring accommodation increased by roughly the same amount over this period, virtually all of the increase in the borrowing capacity of households went into inflating the nominal value of the stock of housing.

Thus, by the early years of this decade, not only could would-be home buyers afford to borrow substantially more in relation to their income than 10 years previously – they needed to in order to realise their housing aspirations.

This is, of course, one of the main reasons why the increase in the ratio of debt to assets (or ‘gearing’) has been much more modest than the increase in the ratio of debt to income. I would argue that unless one takes a bearish view of the outlook for house prices – which I think requires a more pessimistic view of the outlook for interest rates than I think is warranted, and ignores the absence of any excess supply of housing as in the United States – then it is appropriate to take some comfort from the modest increase in the level of household gearing.

Another consequence is that, in contrast to the United States, there has not been any significant increase in home ownership rates in Australia over the past 15 years; rather, home buyers are taking longer to pay off their larger mortgages (ABS 2007a, 2007b). This is consistent with the view that, although financial innovation and enhanced competition have undoubtedly led to some relaxation of lending standards in Australia, the effect of this has in practice largely been to allow those already able to access mortgage finance to borrow bigger sums, rather than to allow significant numbers of people previously precluded from mortgage finance to gain access to it.

That, in turn, helps to explain why, as Chris and Chris point out, the increase in household debt has been concentrated among households who have the capacity to service it; and why default rates have continued to be much lower in Australia than in the United States.

Indeed, as Kent, Ossolinski and Willard (this volume) note, higher debt does not necessarily lead to greater vulnerability; and even if it does, it may still be welfare-enhancing.

However these factors do not explain why, as the authors also note, the pace of borrowing by Australian households has been ‘unusually rapid by … international standards’. Although they do not say so, I think that this is at least partly attributable to some unusual features of the Australian tax system – in particular, the unlimited extent to which investors can offset net borrowing costs against other income for tax purposes (‘negative gearing’) which, to the best of my knowledge, has no parallel in OECD countries other than New Zealand.

Particularly following the halving of the capital gains tax rate in 1999 – which converted ‘negative gearing’ from a strategy which merely facilitated tax deferral into one which permits both deferral and permanent reduction in income tax payable – borrowing for property investment rose significantly, exceeding 45 per cent of all lending for the purchase of housing in 2003/04.

Moreover, since a larger proportion of the borrowing for investment housing than of the borrowing for owner-occupied housing has been applied to the purchase of existing rather than new housing, the investment boom exacerbated the upward pressure on dwelling prices (see, for example, RBA 2003), while doing little to alleviate the shortage of rental housing.

Another curiosity of the Australian experience which the authors note, is the relatively high proportion of household debt at variable rates, something which they attribute to the non-deductibility of interest payments by owner-occupiers leading to a preference for the capacity to make prepayments of principal.

That may be so, although it does not explain why fixed-rate mortgages have become more popular in New Zealand, where the tax treatment of interest payments is similar to Australia but where a much higher proportion of mortgages are at fixed rates, albeit for shorter periods than are common in the US or Europe. It could also result from the fact that fixed-rate mortgage products on offer in Australia are much more ‘fixed’ than those in the US, in particular effectively precluding the option of refinancing at lower rates as has been common in the US during periods of declining long-term interest rates – one example perhaps of products being ‘ill-suited to many households’, as the authors note later in their paper.

The trend decline in business sector gearing, which the authors note, is important in one other additional respect. Because the financial position of the business sector is, in aggregate, much less directly sensitive to fluctuations in interest rates than it was towards the end of the 1980s, aggregate employment should also be much less vulnerable to increases in interest rates than it was during that period. Indeed, although there are other reasons for the considerable strength in employment over the past five years, it has nonetheless occurred through a period of rising interest rates. The enhanced security of employment which has been promoted by, among other things, the stronger financial position of the business sector, has probably contributed to the greater willingness of households to take on additional debt and to their ability to continue to service it in the face of higher interest rates.

The banking system has become the conduit through which the bulk of the financing of Australia's current account deficit has been accomplished. Over the past 10 years, for example, overseas borrowings by ‘depository institutions’ have financed 80 per cent of Australia's current account deficit, compared with around 45 per cent over the preceding 8½ years; while private-sector ‘financial corporations’ account for 82 per cent of Australia's net foreign debt, compared with less than 30 per cent 20 years ago (ABS 2007b). As Chris and Chris note, nearly all of the banks' offshore borrowings are hedged (which was not the case when most of the net foreign debt was owed by governments or non-financial corporations two decades ago), so that a sharp fall in the exchange rate would not, of itself, have any significant consequences for the health of the Australian financial system. On the other hand, as we have seen in recent days, any diminution in overseas lenders' appetite for Australian bank debt can have implications for the exchange rate.

Let me turn finally to two of the authors' policy conclusions. As I mentioned earlier, I am not sure I entirely agree with the conclusion that ‘at the macro level, there is not a great deal that can be done about occasional bouts of mispricing of risk’. For some years now there has been a minority opinion in academic and official circles suggesting that central banks could and should pay more regard to asset prices in formulating monetary policy (see, for example, Borio and Lowe 2002; Bean 2003; Cecchetti 2003; Borio 2006, this volume).

It is at least arguable (with the admitted benefit of hindsight) that the current crisis in the US sub-prime mortgage market may have been less severe had US monetary policy not been eased by as much or for as long in the early years of this decade. The contrast with the Australian experience, where the Reserve Bank did not ease monetary policy nearly as much as most other central banks (in part, to be fair, because the Australian economy was much less affected by the collapse of the ‘tech bubble’) and was the first central bank to begin ‘normalising’ interest rate settings, may be instructive on that point. But I accept that central banks may have difficulty reconciling a desire to use monetary policy to correct perceived mispricing of risk with the inflation-targeting mandate that most of them have been given by elected governments.

However, there is perhaps a role for other policy instruments in at least reducing the propensity for speculative excesses which are intrinsically associated with the ‘mispricing of risk’.

As I noted earlier, in the Australian context the income tax system explicitly encourages speculative activity by providing a subsidy for the borrowing costs incurred in the course of engaging in it, and by taxing the returns to it at a lower rate than the income accruing to labour, for example – despite the fact that encouraging a higher rate of participation in the labour force is ostensibly an aim of government policy.

I do agree with the authors' conclusion that there is scope for further progress at the micro level to bolster households' risk management capabilities, and that a regulatory approach based on disclosure places a premium on financial literacy. I unhesitatingly endorse the conclusion that the proliferation of lengthy and densely-worded product disclosure statements in response to the Financial Services Reform Act 2001 has done little to enhance understanding on the part of retail investors and consumers of the products with which they are dealing. Only last week, in response to an application for a trauma insurance product, I received a 96-page product disclosure statement from the insurer as well as a 34-page statement of advice from the insurance broker, neither of which materially enhanced my understanding of the characteristics of the product I was contemplating.

I am a little sceptical of the authors' suggestion that credit rating agencies could play an enhanced role in summarising the risks attached to debt securities. As was seen during the Asian crisis, and is again becoming apparent in the context of the US sub-prime mortgage crisis, credit ratings are a lagging indicator and have not provided consistently reliable warnings of default. Moreover, credit ratings are paid for by the issuer of the securities; and, the authors correctly note, households are reluctant to pay directly for financial advice.

Research undertaken by ANZ (2005) indicates that it is not only people with low levels of educational attainment or on low incomes who lack adequate knowledge of financial matters or who find themselves in financial difficulty.

ANZ has also accepted that responsible lenders need to do more to assist customers who do get into financial difficulties and to enhance financial literacy among various segments of the population, and in recent years has introduced a number of new programs with those objectives in mind (ANZ 2007).

In Australia, at least, the relaxation of credit standards and the subsequent deterioration in credit quality has largely occurred outside of traditional mortgage lenders. Indeed the introduction at ANZ in 1999 of application and behaviour scoring, and the use of ‘default’ cost-of-living expenses (that is, using expenses based on data from the Australian Bureau of Statistics rather than those advised by the customer if the latter are lower) in calculating the servicing margin for mortgages resulted in practice in a tightening of credit standards.

ANZ does not actively participate in the ‘sub-prime’ market, and does not extend ‘low-documentation’ mortgages on loan-to-valuation (LTV) ratios of more than 60 per cent without mortgage insurance (in which case an LTV ceiling of 80 per cent applies). The delinquency rate on ANZ's ‘low-doc’ loan portfolio is actually lower than that for our traditional loan portfolio (ANZ 2007, p 3).

There is a strong case for bringing mortgage brokers and other non-traditional providers under the same national regulatory system as applies to traditional intermediaries. However I am fearful that, as was the case with the collapse of the ‘tech bubble’, there will be a political and regulatory over-reaction to the US sub-prime mortgage crisis and we will end up with a Sarbanes-Oxley for mortgages, with an inevitable echo in Australia. It would indeed be unfortunate if the amplitude of the swings in financial market sentiment were to be mirrored in the regulatory framework.

References

ANZ (Australia and New Zealand Banking Group) (2005), ‘ANZ Survey of Adult Financial Literacy in Australia’, Research report, November.

ANZ (2007), ‘ANZ Submission – Inquiry into Home Loan Lending Practices and Processes’, paper submitted to the House of Representatives Standing Committee on Economics, Finance and Public Administration, July.

ABS (Australian Bureau of Statistics) (2007a), ‘Australian Social Trends’, Cat No 4102.0, p 174.

ABS (2007b), ‘Balance of Payments and International Investment Position, Australia’, March quarter 2007, Cat No 5302.0.

Bean C (2003), ‘Asset Prices, Financial Imbalances and Monetary Policy: Are Inflation Targets Enough?’, in A Richards and T Robinson (eds), Asset Prices and Monetary Policy, Proceedings of a Conference, Reserve Bank of Australia, Sydney, pp 48–76.

Borio C and P Lowe (2002), ‘Asset Prices, Financial and Monetary Stability: Exploring the Nexus’, BIS Working Papers No 157.

Borio C (2006), ‘Monetary and Prudential Policies at a Crossroads? New Challenges in the New Century’, BIS Working Papers No 216.

Cecchetti SG (2003), ‘What the FOMC Says and Does When the Stock Market Booms’, in A Richards and T Robinson (eds), Asset Prices and Monetary Policy, Proceedings of a Conference, Reserve Bank of Australia, Sydney, pp 77–96.

RBA (Reserve Bank of Australia) (2003), ‘Submission to the Productivity Commission Inquiry on First Home Ownership’, RBA Occasional Paper No 16.

2. General Discussion

The papers presented in this session provoked many comments about the changing nature of risk in the global financial system. The discussion began with one participant noting that there had been a general transfer of financial risk to the household sector over the past decade and that it was unclear whether this transfer had been ideal. The participant argued that, in principle, risk should be transferred to those most willing and able to bear it, which meant that households were the correct repository for long-term risks but not short-term risks. Some concern was also expressed about whether households were even fully aware of the risks that they had taken on in recent years, with one participant agreeing that policy-makers had a greater role to play in educating households about the risks to which they are now exposed. There was a brief debate about whether changes in the riskiness of individuals' income had altered their appetite for risk, with one participant pointing out that idiosyncratic income risk was actually greater than it used to be in the United States and hence could not be a reason for greater debt accumulation.

The discussion then shifted to the broader questions of whether risk in the global financial system had increased and the extent to which there had been a general mispricing of risk. One participant noted that reduced output volatility may have encouraged the accumulation of risk and wondered whether the success of policy-makers in moderating the business cycle had helped to stimulate asset-price bubbles. Another thought was that policy-makers had successfully diagnosed the problems in the sub-prime market and other credit markets but that this diagnosis had done little to alter market behaviour. Some participants thought that rapid innovation and growth in financial markets had itself increased the incentives for market participants to favour short-term positions over long-term positions and that there was little reward for investors taking contrarian positions. There was also an acknowledgement that economists often have idealised views about how markets work, can forget that markets are sometimes ruled by waves of confidence and fear, and tend to underestimate the importance of intermediaries in the financial system.

The rest of the discussion focused on the appropriate role for policy-makers in ameliorating risk in the financial system. One participant argued that policy-makers have long known that risk is procyclical, yet policy has rarely acted to lean against this risk. The same participant went on to suggest that because financial market participants tend to be rewarded for short-term capital gains, they become advocates for a monetary policy that does not tighten during booms but does ease aggressively when the market falls; a lobbying effort that had been rewarded by the Greenspan Fed. However, this view was disputed by other participants. One argued that the problems in the sub-prime market originated in loans issued in 2005–06 – after monetary policy had been tightened – and that the apparent asymmetry of monetary policy reflected the asymmetry of financial markets. Similarly, Chris Ryan thought that central banks had implemented the monetary policy they thought was optimal, not necessarily what the markets wanted, and that sound policy was one of the reasons for good macroeconomic outcomes in recent years.

There followed a debate about the benefits of macro-prudential policy in counteracting cyclical financial risk. One line of argument broadly supported Claudio Borio's suggestion that regulators use either automatic stabilisers or discretionary mechanisms to limit the ‘speed’ of the financial system. However, some pointed to the practical problems with using prudential policy in this way. For example, it may be hard to design automatic stabilisers that deal with all relevant contingencies, but discretionary policy may also be problematic if there are political pressures to alter standards at inappropriate times. A number of participants also wondered about the overall effectiveness of speed limits given the role of the unregulated sector in recent developments. In this respect, one participant stressed the need for better margin requirements and a strengthening of counterparty risk management by regulated entities. Still on practical matters, it was unclear who would take responsibility for policies that would be politically unpopular, with one participant suggesting that financial institutions would find it hard to swallow such policies on macroeconomic grounds. Others questioned counter-cyclical prudential policies on more theoretical grounds. For example, speed limits could lull financial market participants into a false sense of security and encourage them to find new ways to accumulate risk. An alternative strategy would be to convince both regulated and unregulated financial market participants that bad decisions would ultimately have adverse consequences. The difficulty with this is the problem of time inconsistency, whereby policy-makers are willing to warn of the dangers of excessive risk-taking during expansions but cannot avoid feeling pressures to respond to difficulties during downturns. In a similar vein, another participant thought that the economic costs of prudential regulations aimed at dampening financial excesses may actually exceed the cost of dealing with the occasional financial crisis.

In response, Claudio Borio argued that despite these legitimate concerns it is important that policy-makers are more aware of the issue and devote more attention to developing optimal macro-prudential policies. In his view, such policies should in principle be no more difficult to implement than monetary policy, though there was a need to avoid the potential for abuse of discretion and to find ways to overcome time-inconsistency problems and the associated risk of forbearance during bad times.