Speech Australia's Recent Exchange Rate Experience
Talk to the Association Cambiste Internationale Congress
It is a pleasure to have this opportunity to address such an influential gathering. I am particularly pleased that Hans Tietmeyer and many other central bank colleagues are able to be here.
I take this opportunity to tell you something about exchange rate movements in Australia over recent years and the Reserve Bank's role in that.
Winston Churchill once said:
There is no sphere of human thought in which it is easier for a man to show superficial cleverness and the appearance of superior wisdom than in discussing questions of currency and exchange.
I hope what I have to say goes beyond the superficial. I hope too that I shall manage to avoid the many perils that can beset central bankers who talk publicly about exchange rates.
The Foreign Exchange Market in Australia
The foreign exchange market in Australia has developed rapidly in the period since the $A was floated in December 1983. Aggregate foreign exchange turnover now averages about US$30 billion a day, a little below the peak of US$34 billion a day in 1990. Around 40 per cent of the turnover is against the Australian dollar.
At the time of the 1989 BIS survey, the Australian foreign exchange market was the eighth largest in the world and the Australian dollar was the sixth most heavily traded currency. This relative prominence reflects several factors, including a time zone which spans the closure of New York and the opening of Tokyo, and well developed domestic markets in a resource rich economy, which provide good opportunities for international investors seeking to diversify their portfolios.
The initial post-float years saw a good deal of volatility in the foreign exchange market and we had our share of cowboys. The market has since matured and volatility has declined. For some time the Australian dollar has been steadier against the US dollar than the experience of the main European currencies and the yen. In trade-weighted terms, volatility over recent years has been similar to that of other major currencies. Unlike foreign exchange traders, policy makers and others generally welcome reductions in volatility.
We expect the Australian foreign exchange market to develop further and will be encouraging it to do so. The exchange rate plays an important role in the economy's adjustment to terms of trade and other external ‘shocks’. The deeper and more mature the market becomes, the greater confidence we can all have in the exchange rate movements it produces.
What Determines the Exchange Rate?
Economists usually think of the exchange rate as a price which keeps our balance with the rest of the world in some kind of equilibrium. Traditionally, the current account was put at the centre of thinking about the exchange rate, which was seen as being determined in markets for exports and imports.
This view had its heyday in the days of fixed exchange rates and widespread restrictions on the international flow of capital.
In these days of much greater capital mobility the relationship between the current account and the exchange rate is more complicated. In the short run, however, exchange rates are determined in financial markets, essentially by the portfolio decisions of international investors.
This does not mean that the exchange rate can be permanently decoupled from its current account-based fundamentals; at the end of the day the fundamentals will always assert themselves. But it does mean that, in a world of mobile capital, countries might be able to sustain large current account deficits for longer than they could in the past, provided international investors remain confident about the conduct of domestic economic policies. It may also mean that views about what is a sustainable current account, and hence the exchange rate, will be more variable than in the past. Certainly, the importance of expectations in determining exchange rates has been heightened; portfolio managers will always try to anticipate changes in asset prices, political developments and many other variables in addition to the ‘fundamentals’.
This framework has been described as a ‘rope and anchor’ theory of the exchange rate. The anchor is that, in the long run, the exchange rate has to move to a level consistent with a sustainable current account balance, and financial markets should be able to acquire some feel for that, however imprecisely. The rope which stops the exchange rate from drifting too far from that point is the pressure of market forces: significant departures of exchange rates from their expected long-run future values should not persist, because they would be wound in by profitable speculation.
It is a neat analogy but two points should be emphasised:
- First, the anchor – the long-run equilibrium exchange rate – itself shifts in response to changes in the terms of trade and productivity trends (which means that market participants need to continually reassess their views about ‘fundamentals’).
- Secondly, the rope – the speculative process of the market – will provide a somewhat flimsy link to the anchor. Far from having their eyes firmly fixed on the long run, many investors are notoriously shortsighted and can be swept away by the psychology of the market.
Against that background, a useful starting point in seeking to understand changes in the exchange rate is to distinguish between nominal and real exchange rates. It is intuitively obvious that if a country inflates much faster than its trading partners, its exchange rate will at some point have to depreciate to reflect this. Exchange rate movements can therefore be divided into two elements – one part which reflects inflation rate differentials, and one which reflects other, more fundamental factors.
The focus here is on the second component: what determines the real exchange rate?
For countries like Australia, changes in terms of trade are clearly important. Falls in the terms of trade, for example, reduce export income and put downward pressure on the exchange rate. Where falls in the terms of trade are perceived to be permanent and cause the real exchange rate to fall, this will set off an adjustment process which operates to discourage imports and encourage exports.
In practice, fine judgments are involved in determining whether a change in the terms of trade is permanent or temporary. For present purposes, however, the point to note is that countries like Australia with volatile terms of trade can expect to have significant swings in their real exchange rates.
Relative productivity changes will also influence the real exchange rate in a fundamental – if very long term – way. A country which achieves an increase in relative productivity, for example, will benefit by seeing its real exchange rate appreciate over time, thereby enabling it to buy imports more cheaply.
An important shorter term influence on the real exchange rate is the interest rate differential with other countries. Relatively high interest rates attract foreign capital, which puts upward pressure on the exchange rate. Of course, those higher interest rates will not be attractive to foreign investors if they are expected to be offset by currency depreciations. Consequently, and as a shorthand measure, the focus is usually on inflation adjusted (or real) interest differentials. All this is necessarily imprecise in practice but the basic idea is simple enough: if a country maintains relatively high real interest rates, it will tend to have a higher real exchange rate.
The Exchange Rate in Practice
So much for the main factors which influence the exchange rate in theory: how has the $A behaved in practice? And how well has the market performed in delivering the ‘right’ exchange rate – that rate which will help to keep the economy in reasonable internal and external balance?
Perhaps the most constant factor over the past two decades has been the difference between Australia's inflation rate and that of our trading partners. For most of the 1970s and 1980s, Australia's inflation exceeded the average of our trading partners. And the nominal exchange rate has followed a downward trend to compensate for this (see Graph 1). At the start of the 1970s, the $A was worth US$1.12; today it is worth about US$0.76. In terms of the trade-weighted index, the Australian dollar has fallen from 100 in 1970 to about 57 today.
As shown in Graph 1, when allowance is made for our past relatively high inflation rate, the fall in the real exchange rate is much less than in the nominal rate. On this measure, much of the potential improvement in competitiveness from falls in the nominal exchange rate in the past has been whittled away by a relatively poor performance on inflation.
Even so, the real exchange rate has been subject to some large fluctuations. Changes in the terms of trade clearly have been important here, as can be seen from Graph 2. In particular, the sharp fall in the terms of trade in 1985 and 1986 – and the equally sharp recovery over 1987 and 1988 – were broadly mirrored by the real exchange rate.
Not all this movement in the real rate, however, can be attributed to the terms of trade. (The fall in the terms of trade in the mid-1980s was about 15 per cent while the fall in the real TWI was about 30 per cent.) Another reason for the very large fall in the exchange rate in 1985 and 1986 was the decidedly bearish sentiment towards the $A at that time, in part reflecting a re-assessment of Australia's external position. General concerns as to whether the economy was on a sustainable path were crystallised in the then Treasurer's “banana republic” comment of 14 May 1986.
The role of interest rates is more difficult to disentangle, if only because there are various measures to consider (ie short versus long, nominal versus real). Changes in short term rates as a result of monetary policy actions can influence the exchange rate through their effects on both rates of return and expectations.
At the longer end, interest rates are determined in the market and the linkages can run both ways. At times the exchange rate has moved up in response to large capital inflows pursuing attractive bond yields; at other times bond yields have risen on the back of a weakening exchange rate and expectations of higher inflation (as in January this year). In short, the relationship between interest rates and the exchange rate is two-way and causality can vary from episode to episode.
Some observers have been surprised by the underlying strength of the $A since early 1990, given the substantial reductions in nominal interest rates which have occurred. The differentials in real interest rates, however, have narrowed by lesser amounts and in those terms there is less to be surprised about (see Graph 3).
Summarising all this, the exchange rate over the post-float period seems to have responded in roughly the way the theory would suggest. Inflation differentials and terms of trade changes can explain much of what has happened, but interest rate differentials also have been important. At times the market has developed a life of its own, driven by expectations and sentiments which have sometimes caused the rate to overshoot (in both directions).
Monetary Policy and the Australian Dollar
What role does the exchange rate play in setting monetary policy in Australia?
One of the main reasons for floating the $A in 1983 was to regain control over monetary policy after a period when foreign capital flows had, at times, frustrated attempts to exercise monetary control. Floating the currency has avoided the automatic importation of foreign inflation in the manner that had occurred in the first half of the 1970s (and, even more dramatically, during the Korean War).
Internationally, the thinking on links between monetary policy and exchange rates seems to be evolving in two directions. Some countries, like Australia, see virtue in retaining exchange rate flexibility, and using monetary policy to home in directly on domestic economic objectives. Other countries prefer to have a basically fixed exchange rate as an anchor for their inflation objective, and address their monetary policies to keeping their exchange rates stable. Most western European countries now have either a formal link to the Deutschemark through the Exchange Rate Mechanism (ERM), or a strong informal link which keeps their exchange rates steady against ERM currencies. Their reasons for adopting this approach include the desire to promote European economic integration, to avoid exchange rate volatility, and to import German-style monetary discipline.
Why does Australia not adopt a similar approach?
The short answer is that, given Australia's economic structure, the costs would be too great. As a country susceptible to highly volatile terms of trade, we need greater exchange rate flexibility. With a largely fixed exchange rate, a sharp fall in the terms of trade would be highly contractionary. The automatic exchange rate depreciation that tends to occur with a floating rate buffers those contractionary effects to some extent.
As Graph 4 shows, Australia is very vulnerable to changes in terms of trade, along with New Zealand and Japan. The composition of each country's imports differs substantially from the composition of its exports, opening the way for wide swings in relative prices to occur. (The mix of Australia's imports and exports is, of course, very different from that of Japan; a major change in oil prices, for example, would have opposite implications for the two countries.) The terms of trade of the ERM countries, on the other hand, tend to move in the same direction.
So the broad strategy for Australia is clear enough. Whereas the ERM countries use the exchange rate as an anchor for their inflation objectives, and have to absorb the (relatively small) swings in their terms of trade, I believe Australia is better served by allowing the exchange rate to respond to the changes in the terms of trade and focussing monetary policy directly on containing inflation and supporting economic adjustment.
With the exchange rate free to move, monetary policy can be set primarily having regard to where inflation and domestic activity are heading. Australia has not, since the float, subordinated monetary policy to hitting any exchange rate target. To have done so would have negated one of the main reasons for floating the currency. There have been episodes when the exchange rate has been a major consideration in setting monetary policy, most notably in 1985 and 1986. Even then, however, the main objective was to restrain the fall in the exchange rate so as to limit the inflationary consequences of excessive depreciation, not to target any particular exchange rate.
Although not an objective in determining monetary policy in Australia, the exchange rate is affected by monetary policy settings. I noted earlier that real interest differentials can influence the exchange rate. Rising real interest rates throughout 1988 and 1989 no doubt influenced the exchange rate, although they were aimed squarely at hosing down an overheated domestic economy.
This side effect of firm monetary policies on the exchange rate was criticised in the late 1980s on the grounds that the brunt of the adjustment was being borne by the traded goods sector. The exchange rate response to tight monetary policies, however, can be an important channel through which monetary policy operates: a higher exchange rate encourages an increased flow of imports which adds to the domestic supply of goods and restrains inflation. Even if it was possible to cut off this channel, it would not be desirable to do so.
An extension of this criticism is the view that monetary policy should be adjusted to engineer a currency depreciation. This view is usually based on a belief that the current account deficit is unsustainably high or that we are accumulating an unsustainable foreign debt burden. Proponents of this view see a lower exchange rate as the panacea which would deliver faster growth and remove the external constraint. The American journalist H.L. Mencken might well have had this sort of argument in mind when he said that, for every complex problem, there is a solution that is neat, plausible and wrong.
Putting to one side the issue of the Bank's ability to deliver any particular exchange rate target, this approach begs the question of whether Australia's real exchange rate is overvalued. There can be no definitive answer to this question but the framework outlined earlier is relevant. The exchange rate might be judged ‘about right’ if it is consistent with reasonable prospects for achieving medium term internal and external balance – that is, with an acceptable rate of economic growth and a sustainable current account position.
While precise answers are not possible, several points can be made to pessimists who assert that our present path is not sustainable:
First, exports of goods and services have been remarkably strong over recent years, including manufacturing exports (see Graph 5). There is a cyclical component here but much of the growth is explained by new and concerted efforts to pursue export markets, and by the good fortune to be close to the fastest growing region of the world.
- Secondly, the interest burden on Australia's overseas borrowings has fallen substantially as interest rates have come down. This has been reinforced by the recent switch away from debt to equity in capital flows.
- Thirdly, Australian exporters can expect to reap substantial productivity benefits in the years ahead from recent and on-going structural changes in the labour market, the transport sector and elsewhere.
- Fourthly, the relative stability of the $A, notwithstanding the large reductions in official interest rates since early 1990, suggests that international investors retain their confidence that Australia is on a sustainable path.
While these trends continue, it is difficult to substantiate the view that the current exchange rate is inconsistent with a sustainable adjustment path for the Australian economy. Certainly to manipulate the exchange rate through artificially low interest rates in search of improved competitiveness would be a fool's errand.
Monetary policy is bringing inflation under control and this is its main contribution to improving competitiveness. The underlying inflation rate now is less than 3 per cent and, always subject to the proviso that appropriate policies are followed, we expect inflation to remain at low levels as the recovery gathers momentum.
Competitiveness, moreover, is not only just about the ‘real’ exchange rate. It depends on many other fundamental factors, mostly ‘micro-economic’ in nature. Changes in these areas, which require actions by businesses and unions as well as governments, will do more to improve Australia's competitiveness than any monetary policy actions to force a lower nominal exchange rate.
Reserve Bank Intervention
In normal circumstances, the market can be relied upon to deliver broadly the ‘right’ exchange rate. But the market is not infallible. As we all know, it does have a tendency on occasions to over-react to news, and to be driven by expectations and sentiments which are not always well based. On occasions, the Reserve Bank will reach a different view from the market.
This brings me to make a few observations on Reserve Bank intervention.
Usually when the Bank intervenes, its aim is to restore order in market conditions. This might involve adding some two-way business when the market is tending to be very one-sided, or dampening volatile movements in the rate. The aim is not to resist changes, but to try to ensure that adjustments are well based and reasonably orderly. Sharp changes in the exchange rate – for whatever reason – are not good for the economy.
The market can become very unsettled at times, and particularly when major policy announcements are imminent and participants are trying to guess future policy changes. This was the case last January when, following changes in political leadership, the market took a highly pessimistic view of the Government's future attitude to fiscal policy and the exchange rate. Many players in the market were drawing inferences and making pronouncements which could not be substantiated on the basis of available information.
In this situation, the Bank judged that the market was misreading the signs and should not be allowed to dictate the exchange rate outcome. At the time the exchange rate already had fallen by nearly 10 per cent in trade-weighted terms from its September 1991 peak; a further large fall would have threatened the progress being made on inflation and ruled out any additional monetary policy easings to support a still fragile recovery. The Bank backed its judgment with some solid intervention.
In the event, sentiment in the market improved quickly once the Government's policies were clarified. The Bank then ceased its intervention.
Only very occasionally has the Bank intervened to try to move the exchange rate towards what it judged to be a more sustainable level. These occasions have been rare because we can rarely be confident that the market has lost touch with the fundamentals. But three such occasions were in February 1989, October 1990 and May 1991 when we judged that the exchange rate was overly strong, having regard to some of the ‘fundamentals’. On the two latter occasions the intervention was co-ordinated with easings of monetary policy and supported by official comments that made clear our intentions. These efforts produced some useful, if limited, corrections.
We have ‘tested’ the strength of market pressures on many occasions and we will undoubtedly do so again. There may also be times when – as in January this year – we think the market has got it wrong, and some intervention will again be considered appropriate.
Impact of the Bank's Operations
We operate in the market both to meet the needs of our clients – notably the Commonwealth Government – and on our own account. The foreign currency needs of clients provide a natural base on which the Bank can structure its market operations and, in practice, operations on behalf of clients and those on our own account tend to merge.
As already indicated, we operate on both sides of the market, although the emphasis shifts from time to time. In the early post-float period, there were few market operations and the Bank met the Government's foreign currency needs from its foreign exchange reserves. This changed rapidly in 1985 and 1986 when, with the $A under severe downward pressure, substantial sales of foreign currency were made.
From 1987 to 1991, the Bank was overwhelmingly a net buyer of foreign currency (seller of Australian dollars). This buying was only partly offset by the heavy sales of foreign currency in January 1992. To the extent that the Bank's operations have been an influence on the exchange rate over this period, they have been to resist upward pressures on the $A. This is contrary to a frequently heard complaint that the Bank has been acting to hold up the currency.
After its heavy selling of foreign currency last January, some people asserted that the Bank was ‘spending’ – and even ‘losing’ – taxpayers' money. Such views are, in every sense, wrong. The Bank's balance sheet comprises Australian dollar assets and foreign currency assets. When the Bank intervenes, it is simply replacing assets denominated in one currency for assets denominated in another; its balance sheet is unchanged.
Rather than costing the taxpayers money, the Bank's foreign exchange operations have produced significant profits. We estimate that the Bank's intervention has generated realised trading profits over the post-float period as a whole of around $A500 million. This excludes valuation profits attributable to the Australian dollar's depreciation over the period.
These profits are incidental to the intervention which the Bank conducts in pursuit of its general policy objectives. I mention them only because they are one test of whether intervention has helped to stabilise the exchange rate.
Our experience suggests it has. A central bank which is successful in reducing exchange rate volatility would, on average, be buying when the rate was low and selling when the rate was high. Such a strategy would, over time, also be profitable. At the very least, it suggests that the Bank has not sought to defend the indefensible.