Speech Booms, Busts, Cycles and Risk Appetite
Head of Financial Stability Department
Financial Risk Day 2016 Conference
Thanks very much to Macquarie University for the opportunity to speak today. Given this event is titled ‘Risk Day’, I would like to make some observations about the risk appetite of society, and the environment that creates those risks. I won't be discussing current developments: Malcolm Edey covered these in a speech last month (Edey 2016), and we will have plenty more to say when we release the Financial Stability Review in a few weeks' time.
Australia, along with New Zealand, is one of the few countries to mandate that adults wear bicycle helmets. When Australia adopted the ISOFIX European standard for children's car seats, an extra tether was added to the local standard, for additional safety. The guidance in New South Wales for the food pregnant women should avoid includes some items, such as ricotta cheese, that the equivalent UK guidance explicitly says are safe. And when the Murray Inquiry recommended that Australian banks should be ‘unquestionably strong’ and suggested that this meant being in the top quartile of large global banks by capital ratio, the only question was about the practicality of that metric, not the substance of the goal.
Anecdotally at least, across a wide range of domains, it seems that Australia, as a society, chooses more safety than the average industrialised country does.
I'm not going to speculate today on where that social preference has come from or what implications it might have, other than to note that individual and social choices can differ for a variety of reasons. It is what it is. It is society's prerogative to make that choice. Our job at the Reserve Bank is to do what we can to help achieve that social objective of safety. This preference is especially relevant to the Bank's financial stability function. For if Australian society desires more than the average level of safety, by implication it probably means that it has less than average tolerance for the bad economic outcomes that characterise episodes of financial instability.
Thankfully, Australia has also seen less financial instability than average. The last serious banking crisis here was in the 1890s. We came close to crisis in the early 1990s. Some of the banks were distressed and this clearly dragged on economic performance at the time. In fact, the early 1990s was the time of the last really serious recession in Australia. Though I wouldn't regard the last 25 years as a period of entirely uninterrupted expansion, the downturns that did occur were brief and shallow. Many Australians now in the workforce have never known anything other than relatively benign economic conditions. Although growth has been below average in recent times, we are a long way from being in the kind of downturn that some countries have experienced in recent years. But taking a longer view, it is reasonable to ask if a society that is less tolerant of adverse conditions than average, is sufficiently prepared for and resilient to them. Even in the good times, some people and firms will face more negative outcomes. In economic downturns, distress becomes much more widespread. Are we prepared to manage that?
It is also reasonable to wonder if a society that chooses more safety can actually achieve it. And that depends on the dynamics of booms and busts. It is these dynamics that are the main topic of my talk today. To achieve greater stability, we need to know if bust inevitably follows boom, and if so, on what timescale. We also need to know if the severity and timing of a bust are things that policy can affect.
In discussions about monetary and other aspects of macroeconomic policy, it is usually assumed that policy cannot eliminate the business cycle entirely, though it might smooth the cycle out a bit. It is commonly assumed that this is also true for financial boom–bust episodes. To the extent that financial conditions are simply part and parcel of the business cycle, that is a reasonable assumption. If policy can't do away with recessions completely, and recessions are the main cause of financial distress, neither can policy entirely eliminate that distress. But the Australian (and Canadian) experience does suggest that policy and institutional settings can make those episodes of distress less frequent and less painful.
That experience sits against some of the assumptions about and implicit models of financial boom–bust dynamics that shape the international debate on financial stability policy.
Where Does the Boom–Bust Come From?
One assumption implicit in much of the debate is the idea that financial conditions have their own boom–bust shape, repeating on a timetable that is sometimes assumed to be largely unaffected by policy. This is the embedded assumption in most discussions of the ‘financial cycle’, which is presumed to be separate from the business cycle. This cycle is also usually supposed to have a longer duration than the business cycle.
A second, closely related assumption is that the longer a country goes without a bust, the worse that bust will be. This is the ‘growing imbalances’ model of financial instability. The presumption, usually unstated, is that these imbalances grow at roughly the same rate in all expansions.
You might have noticed that so far in this talk, I have spoken of booms and busts, but not bought into the idea of a separate financial cycle. That was deliberate. I'm open to further evidence on this, but from what I've seen, a literal cycle does not fit the data. There clearly are dynamics in the financial system that mean that booms often overshoot and lead to a partial correction or bust. But it is almost always only a partial correction in those financial variables, such as asset prices or the level of credit. Much of the effect of the original boom remains intact. And there is no known mechanism for the fact of the bust to necessarily spark a fresh boom, which is what would be needed for a true cycle.
So, yes, we see boom–bust dynamics in the financial system, but a repeating cycle, not so much. Where, then, do these dynamics come from? I believe that there are two plausible sources, both of which could be relevant in any given episode.
The first is over-exuberance in the face of a genuine structural shift. This is the mechanism that Kindleberger identified: first, the ‘displacement’, the real change, then the ‘mania’ or overreaction to that change (Kindleberger and Aliber 2000). This is why the busts almost never fully reverse the original boom. There was almost always something real at the beginning that doesn't go away. Sometimes when the mania turns to panic, you see an overreaction to the down side. But even if you do, it's unlikely to be as big as the fundamental shift that started it all.
The second source of financial boom–bust dynamics will be well known to anyone who has studied complex dynamical systems in other fields: simply put, it is lags. In particular, it is the lags inherent in stock–flow dynamics. As I have emphasised on other occasions (Ellis 2014), many of the relationships important to financial stability depend on the balance of demand for and supply of a stock quantity that evolves only slowly, like the stocks of different kinds of property. The flow is the thing we can influence, but often it is just a small fraction of the stock. Because of this, the flow variable will typically ‘boom’ when some shift in the fundamentals means the stock variable needs to be higher. And when the stock transition is over because it has reached that equilibrium, the flow will ‘bust’. In the simplest stylised example one can construct, the stock and flow will look something like this (Graph 1).
Structural Shifts: This Time, It's the Same Kind of Different
It doesn't take much to come up with historical examples of those real shifts that spark booms but end badly. New regions open up, spurring a land speculation boom. New technologies are invented that boost investors' enthusiasm for the firms that make or use them. Newly emerging countries open themselves to trade and investment, inducing capital flows in amounts too great for their institutions to manage well. We have seen these kinds of stories over and over again.
Yet perhaps the most important structural shift for sparking a subsequent bust is financial liberalisation. The historical evidence is that many of the banking crises of the past half-century occurred within five years of a significant liberalisation of the financial sector (Kaminsky and Reinhart 1999). That's more or less true for the near crisis in Australia in the early 1990s as well.
A liberalised financial sector is worth having. It is better for economic performance in the longer run. A financial sector free of artificial constraints can more effectively serve its intended purposes: ensuring that savers have a range of suitable assets to invest in; that credit goes to those who can make best use of it; and that people facing risks can insure against them. The highly regulated system of the 1950s, 1960s and 1970s had its own drawbacks. But there can be problems in the transition to a more liberalised system. These come from human over-enthusiasm and the learning curve involved to do credit risk management properly. In the old, highly regulated world, credit was so rationed that only the lowest-risk borrowers received any. Lenders didn't have to assess the risk of the less obvious cases, so they didn't know how to do so. Freed from those artificial quantity constraints, it is easy to start off making choices that are less than ideal, especially when everyone around you is optimistic about the new regime.
Some constraints must remain. There will always need to be some limits on how much people can borrow, because there's a limit to how much they can repay, and uncertainty about whether they will repay. There will always need to be some limits on the desire of creditors to lend, because, in the end, it is other people's money at stake. In the financial system, almost everybody is acting as an agent for somebody else, sometimes a long chain of other people. These agency problems are a key reason why a completely deregulated, untrammelled system is not ideal. Regulation and supervision of the system will remain essential. Even so, a fairly liberalised system, such as the one Australia and many other countries have, does seem to deliver more benefits than the old, highly regulated regime did. It's just that getting from the old system to the current one almost always seems to involve transition problems. Some things that are good in the long run can be dangerous on the way through.
The good news for Australia is that you can usually only deregulate once. But if you find a way to deregulate a second time, for example by reversing a subsequent tightening or finding a new part of the system to liberalise, you might be courting trouble. This is one lesson of the US experience in the lead-up to the crisis. The already quite liberalised US financial system was further deregulated. In particular, the Glass-Steagall Act was repealed in 1999, and the Securities and Exchange Commission (SEC) relaxed leverage restrictions on investment banks in 2004. You could also argue that some of the constraints imposed on the government-sponsored mortgage insurers around that time opened up opportunities for less-regulated private-label securitisation. Those changes might well have been the right thing to do in the long run, but the transition cost was large.
Stock–Flow Dynamics: The Bathtub and the Tap
Now suppose a displacement has happened: a discovery, say, or a deregulation. What then? Simply put, something big and slow-moving has to shift. Yet (as I noted earlier) we can only affect the flows into and out of that stock. We can only change the water level in the bath tub using the tap or the plug. So it is with the stock of credit, or ships, or buildings. We can borrow and repay; we can build and demolish; but we cannot simply wish for the stocks of these things to reach a new, higher equilibrium immediately.
That means there is a transition path. Along that path, the flow into the stock is much higher than average – a construction or lending boom. To induce that boom in the quantity flow, prices need to adjust. Does this mean that at the end of the transition, prices could subside? Of course they could. And yet both the upswing and the downswing will appear to be driven by fundamentals. Does that mean it must all end in tears? Not necessarily. It depends on how much of the increase is transition to the new equilibrium and how much is the over-exuberant extrapolation of recent growth. I will discuss later how we can minimise the latter on the way to achieving the former.
Another point I'd like to make about stock–flow dynamics is that the stock and flow of the same process will have different dynamics when measured in the usual ways one measures these things. We can illustrate this with a simple imaginary example.
Let's pretend that the flow of credit really does follow a very simple cycle that repeats every eight years, which is the long end of the range of what people often call the business cycle. And let's pretend that this credit is paid down over time using the normal credit foncier schedule that would apply to a home mortgage. So the stock is the sum of today's flow, plus last period's flow adjusted for one period of principal repayments, plus the period before that's flow adjusted for two periods of principal repayments and so on.
If the term of those loans is three years, then the stock and flow together would look like this graph (Graph 2). If the flow cycle ranges between +1 and –1, then the stock ranges between about +5 and –5. More importantly, the eight-year cycle in the flow translates into a slightly longer cycle in the stock, closer to nine years.
If the loan term is longer – say 25 years as would be typical for mortgage finance – then the stock cycles are even larger and even longer (Graph 3).
The data are a lot more complicated than this in reality, and so are the techniques used to analyse them. But this simple approach is just intended to remind everyone that differently shaped cycles do not necessarily imply different underlying causes. They could just be the result of different propagation mechanisms. It should also be obvious that bigger cycles might imply destabilising procyclicality, but they do not have to.
How Can We Choose More Safety?
None of what I've just said is to deny that high leverage – especially high and rapidly rising leverage – can be dangerous. If indeed we are a society that chooses more safety, how can we get it? How can we reduce the chance that a transition to a new stock equilibrium ends in tears? Do we have to forgo booms, or even expansions?
It should be no surprise that my answer to this last question is ‘No’. Otherwise we should all just go home. It should also be no surprise that I do believe that there are things we can do to achieve more safety, if that is what we all choose. Some of these things are at the system level: the policies, regulations and structures we create for Australia at large. Some of them are individual to the person, firm or financial institution. For if we want more safety, we need to act like we do.
In the time remaining in this talk, I cannot cover all the things that can be done or are being done to promote financial safety and stability. But I want to cover a few important ones, at least, which I will organise around three pieces of advice.
Go in with your eyes open
The first piece of advice is ‘go in with your eyes open’. Very often, the bad outcomes happen when people didn't appreciate all the risks they were taking.
Sometimes this is because they didn't fully understand the environment or the drivers of those risks. Or maybe they were focused on the wrong risks, much as if a person who is afraid of flying were to text while driving, which is much more dangerous. Probably the best response here is education. At the system level, that is the rationale for much of the regulation around disclosure and consumer financial protection. It is also why central banks devote so much of their financial stability communication simply to explaining how things work. At the individual level, one can also educate oneself. Thinking critically about risks and opportunities isn't always easy. But we are safer if we try. At the very least, we are safer when we remember that if something sounds too good to be true, it probably is.
At other times, people do not fully appreciate the risks because someone else made the decision for them. As I said earlier, the financial system is built on chains of agents, operating for their clients, bosses and other stakeholders. Their interests do not always align, and cleverly written contracts cannot always solve that problem.
That is why prudential supervision is so important. Stability isn't maintained by rules alone. Weak governance of risk-taking can break an institution, even one that looks good just from looking at the numbers. This is not just a matter for the public agencies. Each firm's board has a responsibility to ensure that the agents all the way down their firm's hierarchy are delivering the risk profile that the stakeholders want. But a supervisor with a view of the whole system can shine a light on that governance. It can check that policies are being followed as intended, and disabuse any boards that only think they are more conservative than their peers.
Probably nowhere is this more important than in the banking system. There are huge social benefits from having a safe, liquid asset you can use to make payments, where you don't have to worry about the riskiness of the provider. Modern societies allow certain private firms to offer such an asset – a deposit – but it's long been known that no private entity can offer immediate liquidity at par, in bad times as well as good, without public sector support. Without central bank liquidity support and some sort of government insurance, even sound banking institutions can face a panic and a bank run (Diamond and Dybvig 1983). So both depositors and the public sector have a stake in the behaviour of the banking system. The regulatory authorities are the agents that represent their interests.
Don't get ahead of yourself
The second piece of advice, closely related to the first, is ‘don't get ahead of yourself’.
Remember Kindleberger's displacement idea: history is full of shifts to new (stock) levels, but the transition doesn't last forever. At some point the (flow) boom ends. Plan on that basis. We may never know exactly how long it will last or where the end point might be. But we should never assume that boom conditions will last forever.
The Bank's analysis has long embedded this understanding. We knew a decade ago that the global mining industry would eventually expand supply capacity and so commodity prices would turn down. If anything, we underestimated how long the boom would last (Heath 2015). We knew that much of the housing price boom in the late 1990s and early 2000s was a transition to a low-inflation world; we said so at the time (RBA 2003).
Yet the longer these transitions go on, the more likely it is that some people begin to believe that the boom is permanent, or at least very long-lasting. Policy institutions can take a longer view and warn of the dangers of extrapolation. Even if you don't know where the end point is, it's possible to describe how things might change when it comes. For example, the end of the transition to the low-inflation world meant slower trend growth in housing prices, more periods where they fell a little, higher savings, slower growth in credit and bank balance sheets, and more time needed to accumulate a down payment for the purchase of a home.
With that understanding of the end point, both policymakers and private entities can take pre-emptive action. These transition dynamics, and their end points, can be built into forecasts and stress tests. They can frame business planning. And they can inform how we coordinate and regulate a range of economic activities. This is not just about finance. Stock–flow dynamics are central to property market developments, as we have long known. The actors in that market know that, too. But just as there's a Greater Fool Theory of investment that helps perpetuate booms in prices of financial assets, it sometimes seems that there is a Slower Builder Theory of property development, where everyone knows that not all the projects underway will make money, but yours will if you can just complete it before the other guys complete theirs. This coordination failure can end in painful busts in building activity, because the boom went beyond underlying demand.
It is, of course, possible to let caution go too far, so that you are seeing disasters around every corner. But if one cares about the safety of the system, that is not the worst mistake one can make. To be honest, I am not too worried if there are people who shout ‘bubble’ every time asset prices rise. If it gives a few other people pause and makes them open their eyes to the real risks, it is not such a bad thing. I would rather that, than if everyone were egging the boom on.
Have something to catch you if you fall
The third piece of advice recognises that even with all our best efforts, things can still go wrong. So ‘have something to catch you if you fall’.
Probably the simplest way to do this is to avoid being too leveraged. Very roughly speaking, the alternative to more leverage is more capital. Capital absorbs losses. Capital reduces contagion. And capital ensures that the people who take the financial hit on the way down are the ones who profited from the risk-taking on the way up. It isn't the only thing that matters for stability – we've long emphasised the role of supervision of financial intermediaries – but capital regulation is very important. Society is better off if the financial institutions making promises to pay out to customers have enough capital for their risk profile.
Leverage doesn't only matter for financial entities, though. Non-financial companies and households also need to consider their leverage. So often the firms that crash and burn in the downturn are the ones that geared up most aggressively in the boom. A bit less leverage in the riskier parts of the corporate sector before the crisis might have been helpful. With the benefit of hindsight, their creditors would probably agree.
Households can't issue equity in themselves, but we can still think of their leverage in terms of their ability to service debt out of assets and future income. Anyway, there are things besides capital that can catch you when you fall. And there is one kind of ‘getting ahead of yourself’ that promotes stability – getting ahead on your loan repayments. In Australia, owner-occupiers cannot deduct mortgage interest for tax purposes, so they have an added incentive to pay back debt ahead of schedule. And because most mortgages are at variable interest rates, the lenders have less disincentive to allow this. The Australian offset account is the most tax-effective form of precautionary saving I have ever seen. Not everyone with a mortgage is ahead of schedule, but many are, and those offset balances are growing rapidly (RBA 2015). That makes those borrowers more resilient to shocks and the whole system safer.
You need not always be your own rescuer, of course. Part of the role of the financial system is to allow people to insure against the risks they face. That protection might take the form of an explicit insurance contract, or perhaps a derivatives contract.
But there are some risks – systemic risks – that no private entity can diversify away. There is a role for public policy here. Welfare systems and the automatic stabilisers built into fiscal policy are good examples of what can be done, as is monetary policy. We can also add the array of crisis management and resolution powers distributed across the various regulatory agencies.
The notion of policy doing something to cushion the bust has become unfashionable in some quarters. For sure, the idea that you only needed to clean up the bust without doing anything to lean into the boom has lost credibility since the crisis. But let's not forget that cleaning up afterwards is still a useful thing to do. There is also a bit of a moralistic view in some quarters that even if policy can do something to cushion the bust, it shouldn't – that this is somehow ‘bailing out’ an undeserving borrower. But most of the people hurt in the crisis had nothing to do with the decisions that contributed to it: they lost their jobs, their homes, their businesses in the recession that ensued. Why shouldn't policy cushion that blow? There is no virtue in collateral damage, and no room for spite in public policy.
I hope I've provided some food for thought today. I do think it's fair to say that Australia chooses more financial safety than some other countries, and I think that has implications not only for how policy should be designed and conducted, but for how we conduct ourselves. We will never eliminate risk, nor would we want to. But by making the right choices on a range of fronts and at different levels, from individual to national, we can get the balance between risk and safety that we truly want.
Thank you for your time.
Thanks to Tui Nolan and Fiona Price, whose current research informed parts of this talk. [*]
See <https://en.wikipedia.org/wiki/Bicycle_helmet_laws_by_country>. Several other countries have considered such a measure, and according to other sources, Dubai enacted a similar law in recent years. 
See <https://www.childcarseats.com.au/faqs>: ‘AS/NZS 1754 is one of the most stringent child car seat standards in the world. Unlike the European standard, the Australian/New Zealand Standard requires all child car seats to be tested in side and frontal impact tests, and some with inverted tests for roll-over protection. In addition, all rearward and forward facing child car seats meeting the Australian/New Zealand Standard (AS/NZS 1754) must have a top tether strap.’ 
See <http://www.foodauthority.nsw.gov.au/foodsafetyandyou/life-events-and-food/pregnancy/foods-to-eat-or-avoid-when-pregnant> and <http://www.nhs.uk/conditions/pregnancy-and-baby/pages/foods-to-avoid-pregnant.aspx>. Other examples where these sites give opposite advice include packaged ham, soft-serve ice cream and cold pre-cooked prawns. 
See Fisher and Kent (1999). 
There is an enormous literature on complex dynamical systems, but an introductory text that I particularly liked was Meadows (2008). 
For the graphs that follow, I assumed a 5 per cent annual interest rate and six payments per year to make the lines look smooth. Neither assumption affects the essential point. In reality, there will be a trend and a non-zero mean in the flows data, but it is standard to remove these before analysing the data to find cycles. 
The Bank made these points in a range of publications over the past fifteen years or so, including Stevens (1997), Ellis (2005), RBA (2011), Ellis (2013), RBA (2014) and RBA (2015). 
This emphasis on promises explains why insurance firms are prudentially supervised in almost all jurisdictions, not only deposit-taking institutions such as banks. 
And a good thing, too: if they could do so, it would imply that a strange form of self-enslavement were possible. 
Diamond DW and PH Dybvig (1983), ‘Bank Runs, Deposit Insurance and Liquidity’, Journal of Political Economy, 91(3), pp 401–419.
Edey M (2016), ‘The Evolving Risk Environment’, Australian Shareholders Association (ASA) Investor Forum, Sydney, 18 February.
Ellis L (2005), ‘Disinflation and the Dynamics of Mortgage Debt’, BIS Papers, Volume 22, pp 5–20.
Ellis L (2013), ‘Housing and Mortgage Markets: The Long Run, the Short Run and the Uncertainty in Between’, Address to the Citibank Property Conference, Sydney, 23 April.
Ellis L (2014), ‘Why Financial Stability Matters, and What We Can Do About It’, Address to the University of Adelaide, Adelaide, 4 June.
Fisher C and C Kent (1999), ‘Two Depressions, One Banking Collapse’, RBA Research Discussion Paper No 1999-06.
Heath A (2015), ‘The Terms of Trade: Outlook and Implications’, Resources and Energy Workshop hosted by the Department of Industry, Innovation and Science, Canberra, 20 November.
Kaminsky GL and CM Reinhart (1999), ‘The Twin Crises: The Causes of Banking and Balance-of-Payments Crises’, American Economic Review, 89(3), pp 473–500.
Kindleberger CP and RZ Aliber (2000), Manias, Panics and Crashes: A History of Financial Crises, 5th Edition, Palgrave Macmillan, Basingstoke.
Meadows DM (2008), Thinking in Systems: A Primer, Chelsea Green Publishing Company, White River Junction, VT.
RBA (2003), ‘Household Debt: What the Data Show’, Bulletin, March, pp 1–11.
RBA (2011), Financial Stability Review, March.
RBA (2014), ‘Key Financial Developments Since the Wallis Inquiry’, Submission to the Financial System Inquiry, March, pp 14–42.
RBA (2015), ‘Submission to the Inquiry into Home Ownership’, House of Representatives Standing Committee on Economics Inquiry into Home Ownership, June.
RBA (2015), ‘Box E: Offset Account Balances and Housing Credit’, Statement on Monetary Policy, August, pp 56–57.
Stevens G (1997), ‘Some Observations on Low Inflation and Household Finances’, Talk to the Real Estate Institute of Australia Conference on ‘New Horizons for Property’ , Canberra, 8 October.