Introduction
Being asked to comment on a paper by Professor John Williamson, Chief Economist for the South Asia Region of the World Bank and an academic of formidable reputation on matters related to exchange rate policy, is a daunting assignment. This is especially so for a mere central banker, and one from a small country on the fringe of the recent dramas in Asia to boot.
So let me begin by complimenting Professor Williamson on a very stimulating paper, and one which will, I am certain, provoke a great deal of discussion and debate. Many people, and I include myself among them, believe that the nature of the exchange rate regimes operated by several of the Asian crisis economies was a material factor in the extent of their difficulties. It therefore makes sense to give careful consideration to whether alternative exchange rate regimes might have avoided the crises, or at least reduced the extent of the social and economic damage caused by them.
I myself do not intend to reach final conclusions on the issues before us. There are still many matters on which the evidence is not yet complete. In New Zealand, where we have had a floating exchange rate since March 1985, there are both political leaders and industrialists who are arguing the case for a currency union, either with Australia or the United States. The optimal currency regime is on the table for debate in many countries.
`Free floating' is least vulnerable to crisis
I have only 10 minutes, so I will focus mainly on those points where I have a rather different perspective to that of Professor Williamson. But at the outset let me agree with his observation that in terms of avoiding crises arising from speculative attacks on the currency, a free floating regime seems optimal and a pegged exchange rate regime the most susceptible to crises.
In my own view, that is partly because the exchange rate can adjust more quickly and more smoothly to real shocks than is the case with other regimes. Both the New Zealand dollar and the Australian dollar fell sharply from around mid-1997 to late 1998 in response to the Asian crisis and the effect which that crisis had on the demand for the commodity exports of both countries. I would be surprised if the central bank of either country would have reduced its exchange rate band as quickly if it had been operating a highly managed float, or crawling band.
A second crucially important reason why floating rate regimes are less susceptible to crisis is that domestic banks tend to hedge the currency risk of the foreign borrowing they undertake. Professor Williamson notes that one of the major causes of difficulty in Thailand was that, prior to the sharp fall of the baht, Thailand had `built up a large stock of external debt, which was imprudently structured with a large volume of short-term dollar-denominated liabilities'. But that build-up of debt, and the US dollar denomination of most of it, owed a great deal to the expectation that the peg would be maintained.
In New Zealand, we have absolutely no prudential requirement that banks hedge their foreign borrowings and foreign borrowing by the banks and many corporates has been very large. However, in contrast to Thailand, the sharp depreciation in our currency produced absolutely no damage to bank balance sheets. The banks had hedged all of their foreign currency exposure (with most of the exchange rate risk being carried by retail investors in North America, Europe and Japan). Meanwhile, most domestic corporations were either hedged or net beneficiaries of the currency depreciation. Professor Williamson acknowledges this point, though seems to me to downplay the benefit which a floating rate regime confers in this respect.
A third reason why floating rate regimes may be less susceptible to crisis I advance more tentatively. It seems to me that exchange rate regimes where the central bank is committed to refrain from intervening in the foreign exchange market may tend to make big speculative plays more difficult. In such a situation, any speculator taking a big currency position (as has occurred in New Zealand) has to find some other private sector party to take the other side. This is more difficult than being almost guaranteed of a central bank intervention to provide the liquidity to get out of a speculative position.
In any event, Professor Williamson and I are agreed that a free floating exchange rate regime is less susceptible to crisis than a pegged rate.
The dangers of fixed exchange rate regimes
I think we also very largely agree on the pros and cons of fixed exchange rate regimes. These regimes are particularly susceptible to crisis unless backed by very strong political and institutional rules, as in the case of a currency board or complete dollarisation, which of course is the ultimate in irrevocable exchange rate fixing.
The Reserve Bank of New Zealand recently listed some of the pros and cons of currency union in New Zealand's situation. In broad terms, our list suggests that the decision on currency union is balanced between the benefits of macroeconomic flexibility afforded by a country's own currency, versus the possible microeconomic efficiency gains of currency union, such as reduced transactions costs and exchange rate uncertainty with the partner country.
Although there are some possible advantages in forming a currency union, when the currency union involves a very small country forming a `union' with a very large one, the small country inevitably accepts the monetary policy of the large one. If, for example, monetary policy has been very poor for a number of years in the small country, with persistently high inflation, this may well be a rational policy choice for the small country. But where domestic monetary policy has been reasonable, and can be expected to remain so, handing over monetary policy to a different sovereign entity inevitably entails the risk that the interest and exchange rates appropriate for the large country are quite inappropriate for the small country. This is often the case between different regions of the same country of course, but where several different countries are involved in a currency union the `automatic fiscal stabilisers' and the regional movement of people work very much less effectively than within a single country.
It is interesting to compare the different decisions on this issue made by, say, Ireland and Canada. Ireland has decided to accept a monetary policy designed in Frankfurt for the average of the European Monetary Union. This clearly means that, for the moment at least, interest rates in Ireland are significantly lower than needed to keep Irish inflation under control. Ireland's real exchange rate may soon be appreciating quite strongly, and at some point this is likely to threaten Ireland's competitiveness within the European Union. But Ireland has obviously judged that there are more pros than cons in joining the EMU.
Canada, by contrast, despite a huge proportion of its trade being with the US, has opted to retain its floating regime, presumably reasoning that the factors which drive demand for Canada's major exports are rather different from the factors which drive demand for America's exports. In other words, they see a higher probability of asymmetric shocks occurring.
I suspect that, in a world where the major currencies are floating against each other, there are some real risks in fixing irrevocably to any one of them unless a very high proportion of the country's trade is with the country with which a currency union is sought. Canada's situation suggests that even then, the cons may still outweigh the pros.
Having said that, each country has to make its own calculation of the pros and cons, and some of these decisions may not be economic, or at least related to monetary policy. If, for example, forming a currency union with the United States and foregoing the domestic seigniorage income was the price of a free trade deal with them then it might look more tempting!
Susceptibility to volatility and serious misalignment
Professor Williamson's paper argues that susceptibility to crisis is only one of the two main factors to be taken into account in choosing a currency regime. The other is the susceptibility of the regime to volatility and serious misalignment. On that score, he rates free floating the `worst of all', with his own preferred option of a crawling band the best regime.
I think we both agree that the issue of short-term volatility is probably not the real issue with a free-floating regime. Empirical work does not suggest that this short-term volatility has a significant effect on trade, presumably because of the ease with which companies can hedge themselves against this short-term volatility. In New Zealand's case of free floating, short-term volatility has in fact not been particularly high, at least in recent years.
Looking at the volatility against the US dollar over rolling 30 day periods, the New Zealand dollar was more volatile than the yen, the mark, the pound, the Canadian dollar and the Australian dollar in the period from the float in March 1985 till August 1988. In the subsequent decade, the New Zealand dollar was less volatile against the US dollar than any of the other currencies mentioned except the Canadian dollar. This lower volatility was despite the fact that we have not intervened in the foreign exchange market at all. We could even suggest that the lower volatility is because we don't intervene.
I'm not sure I know all the reasons for this, but I suspect it may have something to do with the fact that the market understands our free floating exchange rate in terms which are a little different from those used by Professor Williamson. He suggests that a country with a true free float does not `change interest rates or monetary policy with a view to influencing the exchange rate' and describes New Zealand as an `extreme example'.1 We do not change interest rates or monetary policy in order to achieve a particular exchange rate outcome, but if the exchange rate moves, either up or down, we take that into account as a factor which may affect our inflation rate. Because financial markets understand this point, this may tend to stabilise the exchange rate that, as in all small open economies, has important direct and indirect effects on the inflation rate.
The bigger issue is the `serious misalignment' issue. Professor Williamson makes the point that free floating currencies are particularly susceptible to serious misalignment and illustrates his case by a comparison of India and New Zealand. Both countries have recently undertaken micro-economic reforms, but India has managed its currency heavily and maintained strong controls on capital movement, while New Zealand has had a free floating regime with no capital controls.
India, he points out, has had a much better growth performance than New Zealand, and indeed in the early years after the commencement of New Zealand's reforms New Zealand barely grew at all. He attributes this to the `very differing macroeconomic policy stances (of the two countries), of which their differing exchange-rate policies were an integral part'. He makes the point that free floating exchange rates tend to appreciate strongly with capital inflow (itself attracted by reforms reflecting `an ideological stance congenial to Wall Street'), squeeze export sectors and lead to large balance of payments deficits. He argues that crawling pegs, by contrast, can be used to maintain a competitive real exchange rate, to the benefit of the country's balance of payments and the country's overall growth rate. Allow me several comments.
First, there are a number of eminent economists who would argue that if the capital inflow is going to the private sector and the exchange rate is floating, current account deficits should be of no concern to policy-makers. I have considerable sympathy with this view.
But secondly, while having sympathy with that view, I am sufficiently cautious to feel uneasy about large balance of payments deficits, and high ratios of external debt to GDP. New Zealand currently has both. And yes, the New Zealand dollar did appreciate strongly, in 1985-88 and again in the mid-nineties. That recent currency appreciation in turn almost certainly has some connection to our balance of payments deficit (around 6½ per cent of GDP in the year to March 1999), though two successive years of drought and very weak world prices for our major exports are also major contributing factors.
But I certainly do not accept for a moment the proposition that the main reason why New Zealand grew relatively more slowly in the second half of the eighties was because of the exchange rate regime. I am more sympathetic to the proposition that New Zealand's monetary policy had to be undesirably tight (and the exchange rate high) because fiscal policy in the mid-eighties was unhelpfully loose.
New Zealand reduced its inflation rate from well above the OECD average (where it had been for nearly a decade) to marginally below the OECD average over a period of just a few years, and that obviously involved a loss of output. New Zealand removed subsidies and import protection from a very much wider part of its economy than India has done, and again the short-term effects of that were adverse for both output and employment. Yes, New Zealand's aggregate economic performance was very disappointing in the second half of the eighties, but over the 1991 to 1999 period, based on data in the OECD's December 1998 Economic Outlook, New Zealand's growth will significantly exceed the OECD average, following two decades when our growth rate fell well short of the OECD average. That we have not grown as rapidly as India has done, or indeed as many other developing countries have done, should not be the slightest cause for surprise, given our respective starting points.2
It may also be relevant to note that, prior to the New Zealand dollar being floated in March 1985, New Zealand experimented with almost every currency regime known to man - including both a fixed peg regime and a crawling peg. Our growth performance does not appear to have been sustainably superior at that time. Indeed, rather the reverse, although of course there are many other factors relevant to this.
But granted that the strong appreciation of the currency was a source of considerable pressure on the tradeable sector, and was a contributing factor to the balance of payments deficit, was that exchange rate appreciation in some sense extreme or abnormal? Apparently not. From trough to peak, the real trade-weighted measure of the New Zealand dollar appreciated by 29 per cent between early 1993 and early 1997. On the same basis, sterling appreciated from trough to peak by roughly the same amount at one point in the nineties, and the yen appreciated by very much more than this in the nineties. The US dollar and the mark appreciated by rather less, but not by much less.
Using a rather different methodology and data set, JP Morgan has estimated that New Zealand actually experienced a real appreciation of 42 per cent over one period in the nineties. However, several Latin American countries had an even stronger real exchange rate appreciation (including Columbia 49 per cent, Ecuador 65 per cent and Argentina 109 per cent) while Hong Kong, like Argentina with a currency board peg to the US dollar, experienced a 45 per cent real appreciation. So strong real appreciations appear to be the feature of a variety of currency regimes, both floating and fixed. And this has considerable relevance of course to those who think that strong appreciation of the real exchange rate can be avoided by pegging to a major currency, or by dollarising.
Even granted that policy-makers wish to avoid strong exchange rate pressure, how do they do it? Jaw-boning financial markets has a strictly limited effect, as I found in the mid-1990s when the New Zealand dollar was at its strongest. The evidence that sterilised intervention is capable of having any enduring effect on the real exchange rate is at least still subject to fierce debate. Chile's attempts to slow the inflow of certain kinds of capital may have had a beneficial effect (though that too is hotly contested, both inside Chile and outside). However, Chile also, according to JP Morgan data, experienced a real appreciation of 45 per cent at one point in the nineties, and now has a current account deficit similar to New Zealand's as a proportion of GDP. Moreover, as financial markets gradually become more sophisticated in all the countries of the region, it is even less clear that controls of the Chilean type would have the desired effect for long.
The harsh reality is that monetary policy alone can not deliver both internal balance and external balance. Indeed, monetary policy can't even have an enduring effect on the real exchange rate, so can't sustainably influence the current account deficit.
To be sure, changing the mix of monetary and fiscal policies is a more promising route, and Professor Williamson mentions that. I have already conceded that fiscal policy was rather too loose in New Zealand in the eighties, and this contributed to upward pressure on the exchange rate given that monetary policy was dedicated to radically reducing our inflation rate. On a previous occasion, I have also acknowledged that, with the wonderful gift of hindsight, it now seems likely that the easing of fiscal policy which occurred in New Zealand in 1996 and 1997 contributed to prolonging the period of tight monetary policy, and the misalignment of the exchange rate. But there is a real issue about both the feasibility - political and technical - and the desirability of using fiscal policy in a counter-cyclical fashion.
Tentative conclusion
Perhaps at the end of the day there is no perfect exchange rate regime. I am reminded of Churchill's famous comment that democracy is the worst form of government in the world, with the exception of all the alternatives. Or as George Soros recently observed `Currency systems are like marriage, whichever one you find yourself in, you think another one might be better'.3
Faced with a choice between a heavily managed float (or crawling band) on the one hand and a clean float on the other, on balance I still opt for the clean float. That is partly because the clean float seems very much less susceptible to crisis, and that surely must be of major concern to policy-makers, but also because I am not yet persuaded that floating currencies are inherently more susceptible to serious misalignment. Certainly, fixing one's currency to a larger, but still floating, currency does not seem to be a way of avoiding serious misalignment, as the very strong real appreciations of, say, Hong Kong and Argentina in recent years illustrate.
1 Incidentally, Professor Williamson's suggestion that we boast about having dismantled our dealing room is incorrect, since we maintain a capacity to intervene in foreign exchange markets in the event of disorderly market conditions, and have both (modest) foreign exchange reserves and a small number of dealers.
2 There are, incidentally, a number of minor factual errors in Professor Williamson's account of New Zealand's experience. Thus, unemployment was certainly not as low as 2 per cent when reforms began, and the argument that income distribution became `noticeably more unequal' is quite hotly debated in New Zealand. He suggests that `only in 1992 did inflation fall to the range of less than 2% which had been mandated'. In reality, inflation measured as required by my agreement with the Minister of Finance fell below 2 per cent in 1991, and this was two years ahead of the required deadline, not a belated achievement as Professor Williamson seems to imply.
3 Quoted in Time magazine, 15 February 1999.