17 Responses to “Falling dollar saga still has a long way to go”

Comments

1. Jeff Frankel: Martin is probably right that we have entered a new phase of adjustment, reminiscent of the late 1980s, during which the dollar will continue to depreciate in order to limit the widening of the US current account deficit and the rise in net US indebtedness. I also agree that currencies in Asia and among oil-producers will have to bear a larger share of the adjustment vis–vis the dollar than they have so far.

   I think the end of the Bretton Woods regime in the 1970s may be even a better precedent than the 1980s. Three economists at Deutschebank (Michael Dooley, David Folkerts-Landau and Peter Garber) have deservedly received a lot of attention for their claim that the current period resembles the Bretton Woods period, with Asian central banks now playing the role that European central banks played in the 1960s, that is, accumulating dollar reserves because they did not want to see their own currencies appreciate and their export industries lose competitiveness.

   But I think we are closer to the end of the Bretton Woods system than the beginning. The Johnson-Nixon economic policy of fiscal and monetary expansion (driven by both the Vietnam War and increasing domestic spending, without a willingness to raise taxes to pay for it) led to declining trade balance and balance of payments. It thereby accelerated the end of the Bretton Woods system in 1971 and the depreciation of the dollar during the remainder of the decade. For the first time, the mark and the yen took on roles as international currencies, reflected in the foreign exchange reserve holdings of central banks, and steadily gained share during the 1970s and 1980s at the expense of the dollar.

   The same is happening in the current decade. A combination of fiscal and (until recently) monetary expansion, driven by both the Iraq War and increasing domestic spending without a willingness to raise taxes to pay for it is accelerating the deterioration in the trade balance. The gradual trend decline in the importance of the dollar in the reserve holdings of central banks has resumed, after a reversal in the 1990s. For the first time in 50 years, there is a credible rival for the dollar: the euro.

   As Martin says, the story is not yet over. Until now, the US has continued to enjoy the “exorbitant privilege” of being able to borrow unlimited amounts in its own currency. Most of the lively academic research on the topic takes such a privilege as a structurally given eternal characteristic of the universe (1). But the next stage might be the gradual loss of the dollar’s role as unrivalled international currency (2).

   (1) Frankel, “Global Imbalances and Low Interest Rates: An Equilibrium Model vs.a Disequilibrium Reality,” Comments in response to the paper by R. Caballero, E. Farhi & P. Gourinchas, BIS Annual Research Conference, Brunnen, Switzerland, June 19-20, 2006.
   (2) With Menzie Chinn, “Will the Euro Eventually Surpass the Dollar as Leading International Reserve Currency?” NBER WP 11510, Forthcoming, in G7 Current Account Imbalances: Sustainability and Adjustment, edited by Richard Clarida (University of Chicago Press: Chicago), 2006. Both available at http://ksghome.harvard.edu/~jfrankel/AEP.htm#Recent Developments in American International Economic Policy.

   Posted by: FT Forum - Jeffrey Frankel | December 6th, 2006 at 5:39 pm | Report this comment
   

2. Brad Setser: Martin Wolf highlights an important and often overlooked point: the dollar’s depreciation against most European currencies - combined with strong global growth - has led to a strong acceleration in US export growth.

   But, Unlike the OECD, I am not convinced that the deterioration in the US current account balance has ended. Given the sizeable gap between the US import and the US export base, US exports have to grow substantially faster than imports just to stabilize the trade balance - something that hasn’t yet consistently happened. The $200b q3 goods and services trade deficit suggests that the q3 US current account deficit is likely to be closer to 7% of US GDP than to 6.5% of US GDP. The latest data flow from the US seems to suggest that the US may be slowing in a way that may not lead to an immediate improvement in the trade deficit - note current difficulties in US manufacturing and oil’s recent move back above $60.

   Moreover, even if the trade deficit stabilizes, I suspect that the deterioration in the income balance will lead the current account balance to continue to widen. Over the past couple of years, the average interest rate on US lending to the rest of the world has increased more rapidly than the average interest rate on US borrowing from the rest of the world. The most likely explanation for this phenomenon is that US lending tends to be very short-term. Now that the Fed has stopped hiking rates, I would expect the average interest rate on US external borrowing (around 4% in early 2006) to converge with the average interest rate on US external lending (around 5% in early 2006), leading to a significant deterioration in the income balance.

   That underscores Dr. Wolf’s broader argument: keeping the current account deficit constant at 6.5% to 7% of GDP requires a fall in the trade deficit - and that most likely requires further falls in the dollar.

   It also makes the issues raised by Dr. Frankel all the more important. In some sense, the euro and pound already have emerged as important reserve currencies. I would estimate that the world’s central banks added over $200b of euros and pounds to their portfolios in both 2005 and 2006 - a noticeable increase over previous years. But the world’s central banks also likely added around $400b to their dollar reserves in 2005 and are on track to add something more like $500b to their dollar reserves this year (my estimate). If Bretton Woods 2 is on its last legs, no one seems to have told Asia’s central banks - they seem to have intervened quite heavily in November.

   I don’t actually think that contradicts Dr. Frankel’s argument. Right now, I suspect a lot of central banks want to do something (reduce their dollar exposure) that is quite different from what continuing Bretton Woods 2 requires them to do (increase their dollar exposure). And if the dollar continues to fall against the euro, many central banks who now intervene heavily (and their political masters) will have to decide whether or not they are willing to increase the dollar balances even more rapidly than they are now or allow their currencies to appreciate a bit more against the dollar.

   Posted by: FT Forums | December 7th, 2006 at 4:22 pm | Report this comment
   

3. Fred Bergsten: Martin Wolf correctly focuses this week on the international financial unsustainability of the US current account deficit and continuing substantial overvaluation of the exchange rate of the dollar. He also rightly concludes that “So far, [the adjustment is] not too bad.”

   There is a second unsustainability in the US external position, however, that could turn out to be much more worrisome. This is the domestic political unsustainability of the trade deficit and dollar overvaluation. History clearly demonstrates that these two variables are by far the most accurate predictors of US protectionism, with much greater impact than the unemployment rate or other domestic variables. The reason is that dollar overvaluation substantially alters the domestic politics of trade policy, adding many normally competitive sectors to the roster seeking relief from imports and undercutting the credibility (and even effort) of pro-trade exporting interests.

   It was this domestic political unsustainability in the United States that brought the previous periods of large global imbalances to the crisis and then resolution stages. Secretary of the Treasury James Baker organized the Plaza Agreement in 1985, and Secretary of the Treasury John Connally took the dollar off gold and imposed an import surcharge in 1971, not because of a collapse of foreign confidence in the dollar but because the Congress was threatening to pass massively protectionist legislation and wreck the global trading system. In both cases, it was the United States rather than foreign governments or the markets that decided that adjustment rather than continued financing of its deficits had become imperative, and it did so largely for trade policy reasons.

   A similar situation may be evolving today. The United States has already enacted protectionist measures against a half dozen Chinese industries. The Congress was already boiling with protectionist sentiments before the recent elections and the Democratic victories will increase this proclivity. All this is occurring despite the booming economy enjoyed by the United States over the past four years and the presence of full employment.

   US trade policy, cyclically adjusted, is thus extremely worrying. Envisage a scenario over the next year or so where the economy slows substantially or even falls into recession, joblessness rises sharply, the global current account deficit increases to $1,000bn, the politically toxic bilateral deficit with China climbs to $300bn, and the Democrats win the White House as well as the Congress. The prospects for US trade policy, and hence the global trading system, are not pretty to contemplate.

   Hence it remains imperative for the major parties to the global disequilibria, especially the United States via fiscal policy plus China and a number of other Asian countries via currency appreciation, to initiate policy actions that will start correcting the misalignments and imbalances. The alternative is to risk the openness of the global economy as well as its financial stability.

   Posted by: C. Fred Bergsten | December 7th, 2006 at 4:51 pm | Report this comment
   

4. Andrew Smithers: I would like to extend the discussion set off by Martin’s article on the dollar.

   If we assume, as seems eminently reasonable, that the US external trade deficit needs to fall from nearly 7% to say 3% of GDP, there needs to be a rise in US net domestic savings of around 4% of US GDP, and a fall in the rest of the world net domestic savings by around 1.3% of RoW GDP.

   These are not, however, the only adjustments needed. The US also needs to have capacity for the production of traded goods and services, in excess of that required to meet domestic demand. This can either be achieved through a US recession, or a switch of investment from producing non-traded goods and services to traded ones. Martin has previously referred in his article to the work of Obstfeld and Rogoff (1). They have pointed out that this switch in investment is likely to require a large fall in the dollar, partly because this will enable the needed rise in investment returns on traded goods and services to rise relative to those of non-traded ones.

   A straight switch in investment is, however, unlikely to be sufficient. This is because the capital/output ratio of traded goods and service production seems very much higher than that of non-traded ones. It has been a marked feature of the world economy over the past 15 years or so that, among G5 countries, the UK and US have invested least and grown fastest. They have also been the ones with large and growing trade deficits.

   The two features are likely to be connected. Rapid growth has been possible with low investment in the US and UK because it has been concentrated in low capital/output production. (The problem is perhaps more acute in the UK, where housing investment, which has a particularly high capital/output ratio, needs to rise, whereas a fall in housing investment in the US should help the adjustment problem.)

   If this is correct, an additional adjustment is needed in the world economy. If recessions are to be avoided, not only must US (and UK) net domestic savings rise to finance lower trade deficits, they must rise even more to finance the higher levels of domestic investment which are needed to create capacity in traded rather than non-traded goods and services.

   Equally, the rest of the world will have to adjust by reducing investment as it switches to more non-traded investment and, to avoid recession, this switch must be accompanied by an even greater fall in domestic net savings than is needed simply to match the adjustment in external trade. If recession is to be avoided, all these adjustments must take place in a smooth and balanced fashion.

   The problem posed by different capital/output ratios for traded and non-trade output is not, of course, the only other difficulty in adjusting to lower trade imbalances, but I thought that one was enough for now!

   (1) E.g. NBER working paper 10869 “The Unsustainable US current account position revisited” by Maurice Obstfeld and Kenneth Rogoff.

   Posted by: Andrew Smithers | December 8th, 2006 at 3:20 pm | Report this comment
   

5. Martin Wolf: The posts by Jeffrey Frankel and Brad Setser raise two important questions.

   The first is whether the depreciation of the dollar so far would, with the right macroeconomic environment in the US, generate a stable trade deficit as a share of GDP while the economy is growing in line with potential. In other words, can we envisage real output and real domestic demand growing at the same rate, at the current real exchange rate. The evidence suggests to me that this is conceivable, but not yet certain. As I noted in the column, and Brad has underlined, real export growth has picked up. But, given that trade tends to grow faster than gross domestic product and imports of goods and non-factor services were 56 per cent bigger than exports last year, exports have to grow considerably faster than imports if this is to happen. Let me give a simple example: assume potential GDP grows at 3 per cent and imports grow at 6 per cent a year (both in real terms). Then, if one ignores any possible shifts in relative prices, for the trade deficit to remain a constant share of GDP, exports must grow initially at 7.6 per cent in real terms, falling to 6.9 per cent fifteen years from now.

   The second question is how long we can expect Bretton Woods II to last. I would argue that the present episode has elements of the collapse of Bretton Woods I, as Jeff argues, but it also has elements of the 1980s, because of the big initial real appreciation of the currency and the scale of the subsequent current account deficit. Either way, one of the big questions is how long the massive foreign currency intervention of the Bretton Woods II period will endure. Jeff thinks we are closer to the end than the beginning. The question then obviously is when did it begin. Was it in 1994, when the Renminbi was devalued and pegged? Or was it in 1998, after the Asian financial crisis? Or was it in 2002, when the dollar started to fall and reserve accumulations exploded?

   I have really no clear idea how long it will last, because I do not know how much more money the Chinese and others are prepared to pour down this particular rat hole. But what is quite clear is that as long as they wish to preserve their competitiveness against the US by targeting the dollar, they cannot also achieve the massive diversification of reserves of which Brad Setser speaks. At some point, however, they will indeed stop offering their open-ended line of credit to the dollar and the currency will then plunge. For reasons I have argued on many occasions, the sooner that happens the easier the subsequent period of adjustment will be. At the end of it, the global role of the US dollar will surely have diminished substantially, as both Jeff and Brad suggest.

   I will try to comment on Andrew’s interesting post soon.

   Posted by: FT Forum - Martin Wolf | December 8th, 2006 at 7:25 pm | Report this comment
   

6. Michael Dooley: Connally’s famous comment can be updated: “The sky is not falling in Texas, the Euro and the Pound have a problem but the dollar is doing just fine.” I can see why a 5% appreciation of the Euro and the Pound over two weeks is a cause for concern especially if this is interpreted as the beginning of the long awaited slide in the dollar. But the dollar’s average value fell much less, about 2 percent against all currencies, and only 1.3 percent against what the New York Fed calls other important trading partners and what we call BW II emerging markets.

   The Dooley, Folkerts-Landau, Garber BW II framework suggests a straightforward interpretation of recent developments in exchange markets. The Euro’s recent strength reflects the market’s view that US rates will fall early in 2007 while rates in Europe continue to rise. We don’t share this forecast but there is no doubt that investors around mid-November placed heavy bets that the US economy would stumble and the Fed would ease in the first quarter of 2007. The standard open economy model would predict an appreciation of the Euro against the dollar given this change in expectations about interest differentials.

   Martin is of course right that the dollar has depreciated by 22% since early 2002. Moreover, we agree that the 50% appreciation of the Euro against the dollar during 2002-2004 was related to the very large projected change in the net foreign asset position of the US relative to Europe that would be generated by the Bretton Woods II system. But that major adjustment of the Euro/dollar rate is over. We expect the Euro and the Pound to follow a normal cyclical pattern against the dollar within the narrow ranges observed for the past three years.

   The dollar will depreciate in real terms against the “other” managed currencies but the adjustment will be glacial. In the BW II system some emerging markets have at last found a development strategy that works. They will find it in their interests to continue to subsidize export led growth and finance the US current account deficit. Call this mercantilism if it makes you feel better but that system also lasted a long time. As Brad Sester points out in his blog, governments bought $70 billion in October, and we agree with his evaluation of this fact: Wow!

   (We also agree with Jeff Frankel that the important question is where we are in this adjustment process and we respect but disagree with his conclusion. A development strategy that works should “last” until incomes are roughly equalized across countries. We clearly have a very long way to go in this process.)

   Posted by: FT Economist Forum | December 10th, 2006 at 7:00 pm | Report this comment
   

7. Martin Wolf: Andrew has raised a very interesting question. But I would also think the relative capital intensity of US tradeables and non-tradeables is not so obvious. How does the (physical) capital-output ratio of restaurants or houses (which produce non-tradeable services) compare with that of international finance (which produces skill-intensive tradeable services) or factories (which produce tradeable goods)? I would have thought it would depend on the nature of the comparative advantage of the country in question. US exports are likely to be intensive in human capital, but not necessarily intensive in physical capital. If so, the (physical) capital output ratio of US tradeable goods and services may not be much higher than that of non-tradeables.

   Posted by: FT Forum - Martin Wolf | December 10th, 2006 at 7:05 pm | Report this comment
   

8. Martin Wolf: The Bretton Woods II framework proposed by Dooley, Folkerts-Landau and Garber is deservedly celebrated. It is helpful, therefore, to have Mike Dooley spell out its implications for the present conjuncture. I would like to make the following comments.

   My first is that the proposition that exchange rate movements between the US and the eurozone can be explained by expected differentials in short-term interest rates is not entirely persuasive. As I noted in my column, a high proportion of the financing of the US deficit takes the form of bonds. Now if one expects the US economy to slow and interest rates to fall, one would expect the price of dollar-denominated bonds to rise. This then is a reason to buy dollar-denominated assets, not sell them.

   My second comment is to recall that half the current account surpluses in the world economy are being generated by oil exporters. So the analysis by Dooley et al applies only to a part – albeit a very important part – of the global balance of payments picture. The oil exporters are wealth-preservers. They may well wish to diversify out of dollar assets on a long-term basis.

   My third reaction is to the implications of the BW II framework for the policies of the export-oriented economies. The view that Mike Dooley puts forward is that Asian developing countries will persist with mercantilist economic policies until incomes are equalised. That would be for at least three and perhaps for four or five decades.

   I find that impossible to believe. By then China’s economy would probably be bigger than that of the US and its current account surplus, in today’s prices, might be over $1,000bn a year. The system would surely blow up before that. China – let alone the whole of Asia – is too big to grow by accumulating increasingly worthless paper claims on the US. That is, in any case, not a growth strategy, but a gift strategy.

   Yet even if one doesn’t believe BW II will last for more than three decades, it now seems conceivable that it might last for another decade or so. But would this be a sensible strategy? I would still argue that the answer to this question is a resounding no.

   First, a deliberate policy of undervaluing the real exchange rate is subsidising export capacity (overcapacity, in fact) that would not be efficient in the long term. This is not a way to create a world-beating manufacturing sector.

   Second, this strategy is not in fact proving a good way to increase employment. An emphasis on rural services would do much more to generate employment.

   Third, as Professor Yu Yongding has argued in a previous forum, foreign currency intervention creates large and growing distortions in the financial system, because nominal and real interest rates are too low. Monetary sterilisation is manageable now, but what would happen over another decade?

   Fourth, at the current real exchange rate, it is overwhelmingly plausible that the current account surplus consistent with rough macroeconomic equilibrium will explode upwards. It is easy to imagine its reaching $300bn-$400bn in just a few years.

   Fifth, a sizeable part of the wealth of the Chinese people is being given away, instead of being used profitably at home.

   Sixth, because China is so large and so open, the protectionist backlash could be enormous: just wait for the next serious slowdown in the US. The backlash might even break up our open world economy.

   I think Professor Dooley and his colleagues have captured part (though certainly not all) of the motivation of the Chinese authorities. I recognise, too, that their prediction about China’s behaviour is more persuasive now than it seemed to me when it was first proposed. I applaud them for that. But I cannot believe this policy will last for decades and certainly do not think it should do so.

   The US blew up Bretton Woods I in 1971 and it will blow up Bretton Woods II. I am not sure when. But it will be well before 2035.

   Posted by: FT Forum - Martin Wolf | December 11th, 2006 at 8:04 pm | Report this comment
   

9. Martin Wolf: : I somehow missed Fred Bergsten’s cogent remarks in my comments on Mike Dooley. But Fred has spelled out in detail the sixth point in my post immediately above. I can only say “amen” to his comments. The proponents of BW II are too complacent about the political economy of mercantilist policies by a country as big and (from the US perspective) potentially threatening as China.

   Posted by: FT Forum - Martin Wolf | December 12th, 2006 at 11:09 am | Report this comment
   

10. Andrew Smithers: I suggested in my previous comment that the high growth and low investment ratios of the US and UK in recent years were the corollary of their rising trade deficits and that a reduction in those deficits required either a recession or a rise in the overall investment levels to create tradeable output capacity.

    Martin’s response was to question the relative capital/output ratios of traded and non-traded goods and services.

    I set out some relevant data on this, for which I give many thanks to James Mitchell for his help.

    In Q3 2006, US goods’ imports ($1.938bn) were 30 per cent greater than all its exports ($1,420bn.) and nearly five times US service exports ($427.7bn).

    Average COR in the US Traded and Non-Traded Goods Sectors, 1998-2003.1
    Type Sector COR p-value for Test of Mean Equality.2
    Traded Manufacturing 1.560 -
    Non-Traded Construction 0.396 0.000
    Non-Traded Retail trade 0.977 0.000
    Non-Traded Finance, insurance,
    real estate, rental
    and leasing 1.095 0.000
    Non-Traded Professional and
    business services 0.553 0.000
    Non-Traded Educational services,
    health care and
    social assistance 1.342 0.001

    It takes, on average, 70 per cent more capital to produce output from manufacturing than it does from sectors which are primarily or exclusively non-traded.

    From the above data it seems reasonable to assume that the US trade deficit cannot be significantly reduced within a reasonable time frame without an improvement in the balance of trade on goods and that, to avoid recession, this will require a rise in US domestic investment.

    The case does not, however, rest on this argument alone. There is also strong supporting evidence in the form of the relative growth rates and investment ratios of the US and UK compared with other G5 countries.

    From 1994 - 2004, US and UK investment was 25 per cent below the average of France, Germany and Japan, while their growth rate was almost exactly double.

    This can be seen either as a temporary, albeit decade sustained, phenomenon or as an example of the enormous superiority in business of “Anglo-Saxons”. If, as we argue, growth with low capital output ratios can be sustained if trade deficits rise, the phenomenon will be temporary; though, as we have seen, it can be sustained for a decade or more.

    As the capital income shares appear to be fairly similar, the alternative also requires that the return on capital in the US and UK should be permanently way above that in other G5 countries, without a compensating change in investment levels.

    Footnotes: 1. The COR is defined as the ratio of the current-cost net capital stock of private non-residential fixed assets to national income. The ratios are derived from the NIPA Table 6.1 for national income by industry, and the “detailed fixed asset tables at replacement cost”.

    2. This test shows that the differences in the capital/output ratios are statistically significant. The p-values indicate the probability of accepting the hypothesis that the average COR in the selected industry is equal to the average COR in the manufacturing sector. The test is a so-called ANOVA test based on the idea that, if the sub-groups have the same mean, then the variability between the sample means (between groups) should be the same as the variability within any sub-group (within group).

    Posted by: FT Economist Forum | December 12th, 2006 at 12:12 pm | Report this comment
    

11. Ronald McKinnon: I feel goaded to write to register my disagreement.

    First, the fall of the dollar in 2006 (about 12.5 per cent against the euro) is not significantly different from a random statistical fluctuation. The figure in Martin’s article showing the dollar decline of 35 per cent was deceptive because he picked as the base January 2002, when the dollar was a local maximum from the high-tech bubble. If before January 1, 1999 you splice the old German Mark on to the euro as in the figure, you see that the euro value of the dollar today is about where it was in the mid-1990s.

    Second, a major nominal depreciation of the dollar can be sustained only if there is inflation in the United States or deflation abroad. In the long run, a real depreciation cannot be sustained at all. And there is no short
    run over which devaluation itself will reduce America’s trade deficit.

    Unfortunately, the FT is full of negative references to the dollar: strange front-page large-type headlines on downward (random) blips in the dollar, and then lesser inner page references such as “bruised dollar recovers” to
    random upward blips in the dollar. Your editorial page (apart from Martin)often refers to America’s trade deficit not being sustainable and with a big correction in exchange rates being inevitable.

    Just recently(December 11), there was a bit more dollar bashing on the front page where “Oil Producers Shun the Dollar” appeared in large type. But if you look closely, the BIS had just reported that OPEC countries had cut their dollar holdings from 67 per cent to 65 per cent of their total reserves between the first and second quarters of this year. However, because the euro has risen against the dollar this year, this modest change might well be explained by just a book keeping valuation adjustment rather than any deliberate policy change in purchasing official reserves.

    Posted by: FT Forums | December 13th, 2006 at 9:52 am | Report this comment
    

12. Martin Wolf: The last post above, by one of the world’s two most distinguished (and justly admired) international monetarists, Ronald McKinnon, raises important questions. (The other highly distinguished economist in professor McKinnon’s camp is the Nobel-laureate Robert Mundell of Columbia University.)

    Professor McKinnon found himself in such disagreement with my column above that, instead of posting his full comment here, he published his rejoinder in the Wall Street Journal of 13th December (sad, but true). Some of the flavour is, however given in the above post.

    Attentive readers will wonder how distinguished international economists (I am not, by the way, including myself in this category) can disagree so strongly on whether a devaluation of the dollar is either likely or necessary. So let me try to explain the underlying difference.

    In doing so, let me first indicate what everybody in this debate agrees on: a trade deficit reflects a difference between a country’s income and its expenditure. Then if a country spends more than its income (which is the same thing, in the Keynesian national-income-accounts framework economists use, as investing at home more than it saves), it will need to attract finance from foreigners and so sell claims upon itself (or real assets) to them. The counterpart of this financial flow - a capital account surplus - is a trade deficit. Thus, obtaining finance from foreigners finances the real flow of goods and services (the trade deficit) that a country spending more than its income needs, by definition.

    The disagreements then lie elsewhere. Economists work with simplifying models. It is the only way that the complexity of the real world can be made analytically tractable. The big differences among economists are over what their simplifying model of the economy should look like.

    The model Ronald McKinnon uses seems to have four main characteristics: each country produces just one good, which is fully tradeable; the demand for money is stable; nominal and real wages and prices are flexible in both directions; and savings and investment (spending and income) are determined by the structure of the economy, not by policy decisions. I think he also assumes that foreign demand for a country’s assets is perfectly elastic at a given expected real return. It is not clear to me what he believes happens to the terms of trade (the relative price of exports and imports) when a country’s demand for imports falls and supply of exports rises. But if he believes these would be unchanged, then exports and importables would also be the same commodity.

    I am not saying that professor McKinnon believes these things true in a strict sense, but he believes they are a reasonable approximation to the truth, for analytical purposes.

    If the world were more or less like this, the economy would always be close to full employment; changes in the price level and the exchange rate, under floating rates, would be determined by the growth of the money supply relative to the growth of the money supply elsewhere; with fixed exchange rates, change in the money supply would create a temporary trade deficit in the country issuing the currency, but ultimately a rise in the world price level; and when growth in the money supply was in line with underlying growth in nominal demand, the trade balance would reflect the differences between savings and investment at full employment. There is no real exchange rate in this model of the world. All that would be needed to alter a trade deficit (or surplus), while keeping the economy at full employment, would be more (or less) spending.

    The question is whether this is a reasonable view of the world. I would suggest it is not in at least two important respects. (Actually, I think there are some more, but I will ignore them here.)

    First, the assumption that it is easy to transform the capacity to produce goods and services from one sector into another, at a constant opportunity cost, (technically, that elasticities of substitution are perfect) is, in my view, heroic. (on this point see the forum’s discussion of my column of September 26, 2006 at http://blogs.ft.com/wolfforum/2006/09/a_slowing_us_co.html#comments.)

    In many standard models of the economy, a distinction is drawn, at the least, between non-tradeable goods and services, exportable goods and services and importable goods and services (the latter two sometimes being known, together, as tradeables). The important point is that it may take a large relative price change to shift domestic resources from non-tradeables into the production of exportables or importables. It may also need such a price change to shift demand from tradeables to non-tradeables. This is known as a change in the “real exchange rate”. The reduction in the trade deficit may also lead to a change in the terms of trade between exports and imports (relative prices) as the former increase and the latter shrink.

    Second, the price changes needed to bring that shift around may not happen easily under a fixed nominal exchange rate, particularly if they require a sizeable fall in nominal wages. If, for example, the US were to need a sizeable real depreciation of the exchange rate, to correct the external deficit, while also maintaining full employment, and it had a fixed nominal exchange rate against all other countries, then nominal unit labour costs would need to fall in the US, other things being equal, relative to those elsewhere. If the rate of growth of productivity and nominal wages were fixed, this would not happen.

    If the nominal exchange rate depreciated, however, and nominal wages were sticky, the relative price of non-tradeables would fall. (More precisely, the domestic currency price of exportables and importables would tend to rise, while that of non-tradeables would stay much the same.) That would lead to shifts towards consuming non-tradeables and producing tradeables.

    Combined with the necessary reduction in excess spending (without which these shifts would merely create excess demand and so inflation), this would sustain domestic output, while the trade deficit fell. The alternative of just reducing spending would, under these assumptions, lead to a prolonged recession, since there would be little or no increased incentive to increase output of tradeables or consumption of non-tradeables.

    It would be possible also to explore the extent to which the real exchange rate may, through policy responses, determine the balance between spending and income rather than the other way round. (I believe this inverse relationship is important in today’s world economy.) Even if that were not the case, most economists (Jeffrey Frankel of Harvard in his comments above, for example) assume that government can determine the balance between spending and income, by changing levels of government spending or revenue, if not by changes in incentives to save or spend by the private sector. But this discussion has already gone far enough to indicate why the assumptions made by those who believe there is no point to changes in nominal exchange rates may be open to question.

    Posted by: FT Forum - Martin Wolf | December 15th, 2006 at 6:58 pm | Report this comment
    

13. Martin Wolf: Finally, I thank Andrew for his additional information on capital-output ratios. I accept the point. How far this means that the growth rates of the US and UK must fall when the tradeable goods and services sector must expand relative to non-tradeables depends on how big a role total factor productivity (i.e. the part of productivity growth not determined by measured investment) plays in growth. It is not the case, that TFP necessarily accrues as a return to capital. So the US and UK may have higher TFP growth without higher returns on capital. But the determinants of TFP are obscure, since it is essentially a residual. At present, however, TFP does not seem to be accruing to labour income, at least in the US, so maybe it is accruing to capital, as Andrew argues.

    Posted by: FT Forum - Martin Wolf | December 15th, 2006 at 7:06 pm | Report this comment
    

14. Ronald McKinnon: Martin Wolf is very kind to both praise me unduly, and then provoke me to spell out more systematically my theoretical model linking exchange rates tointernational adjustment. My December 13 Wall Street Journal op-ed, ‘The Worth of the Dollar’, took a historical approach to showing the worldwide macroeconomic disorder caused by large fluctuations in the dollar’s exchange rate against other major currencies. Here, in criticizing Wolf’s suggested model of what is in McKinnon’s head, I try to summarize more carefully the theoretical model that is embedded in my several different research papers on the subject.

    Like it or not, the dollar is at the center of the world’s monetary system, while simultaneously the United States runs large current account and trade deficits. Indeed, it couldn’t have run such deficits for more than two decades if the dollar was not the definitive international money. Because much of the world is on a dollar standard, only the U.S. can borrow abroad indefinitely in terms of its own currency to cover its relatively low level of saving.This is possible as long as the U.S. Federal Reserve Bank keeps the purchasing power of the dollar fairly stable so that other countries with trade surpluses are loathe to appreciate against that currency in which most of world trade is invoiced. Thus, there is no immediate crisis and no need for precipitate action by governments - particularly on the exchange rate front - to ‘correct’ the U.S. current account deficit. Nevertheless this continualU.S. borrowing is unsatisfactory ( I do not say unsustainable) not only because the world’s richest economy is grabbing the lion’s share of international finance potentially available for economic development,but also because of tensions it sets up within the American economy itself.

    What is the transfer problem? In order to transfer resources in real (nonfinancial) terms from the rest of the world, the U.S.runs very large trade deficits in manufactures from surplus - saving industrial economies such as China, Japan,a host of smaller ones in East Asia, Germany,as well as several smaller European countries. This ‘real’ transfer of manufactures to cover the shortfall in American saving speeds the contraction in employment in U.S. manufacturing beyond the ‘natural’ rate of decline experienced by other mature industrial economies. The upshot is a protectionist backlash in the United States, particularly by members of Congresswith manufacturing constituencies. These people incorrectly blame ‘unfair’ foreign trading practices - undervalued currencies, substandard labor practices, or ‘dumping’ of subsidized exports in American markets - instead of America’s own deficient saving covered by foreign borrowing.Rather than any imminent collapse in America’s credit line with the rest of the world, it is this protectionist backlash that is the serious threat to the world trade.

    However, modeling both the international monetary role of the dollar and the transfer problem together presents problems. To capture this duality, I will follow the time - honored but treacherous tradition in international economics of separating out monetary issues from ‘real’ ones. Let us model first the transfer problem and real relative price changes associated with reducing absorption (spending) relative to income in the United States while increasing it abroad. Then we can analyze the worldwide monetary repercussions from major changes in the dollar’s nominal exchange rate on U.S. and foreign monetary policies.

    The Transfer Problem

    To reduce the U.S. current account deficit from, say,7 to 3.5 percent of GDP, adjustment must start with a fall in total U.S. absorption relative to income of at least 3.5 percent - and complementary inverse changes abroad. In an earlier contribution to the Wolf forum analyzing such changes in absorption, Willem Buiter assumed, at least initially, no change in the dollar’s real exchange rate - sometimes defined as the terms of trade, i.e., the price of U.S. exports relative to its imports. Instead, for the transfer problem, Buiter in his excellent analysis correctly focuses on the more relevant price of nontradables relative to tradables (both importables and exportables),whichwould decline in response toa fall in absorption within the American economyand would increase in response to the rise in absorption in major foreign trading partners. And Buiter’s positionis essentially the same as mine and that of any economist who has thought about the matter. (Thus I reject Wolf’s attribution that I assume that each country produces just one good that is fully tradable.)

    Wolf also suggests that I assume complete wage price and wage flexibility. This is true, but only in the long run. (I am totally against abrupt or discrete changes in exchange rates or in absorption.)Let us assume that the 3.5 percent absorption decline is not abrupt, but nevertheless is fairly definite as part of corrective government program, i.e,. highly visible fiscal improvement and no more housing or other bubbles on the part of the Fed. If well signaled and spread out over some years, the fall in absorption itself would gradually bid down the price of U.S. nontradables relative to tradables - - which remain buoyed by robust external demand. This natural fall in the relative price of nontradables and slower wage growth in that sector gradually releases capital and labor for greater U.S. production of both importables and exportables. Of course nobody (least of all economists!) would know exactly how much the relative prices of non tradables and wages would eventually fall in the U.S. or increase abroad. But the American economy is flexible, with workers and firms continually adjusting to various shocks, and a 3.5 percent fall in absorption over some years isn’t all that large. In the modern world, where the distinction between tradables and non tradables is eroding, I would expect the necessary relative price changes would be quite modest in the long run.

    However absorption adjustment must be two-sided, if only becauseits an accounting identity: the gradual fall in U.S. absorption relative to income must be matched by a gradual rise in foreign absorption relative to income. Otherwise,unilateral absorption adjustment by either side to right the trade imbalancewill always be frustrated. Putting pressure on China and Japan to increase consumption is all well and good but only if matched by a reduction in consumption in the United States.

    (Note that starting ‘adjustment’ with some abrupt change in the exchange rate, e.g., a large dollar devaluationthat reduces the price of U.S. exports relative to its imports would be a bad mistake. The terms of trade is the wrong relative price, and abruptly changing it would actually make reduction in the U.S. current account deficit through absorption adjustment reduction more difficult. It would also create monetary chaos. These issues are discussed in turn below.)

    Suppose both sides begin the necessary adjustment - reducing absorption in the U.S. and raising it abroad. In long-run equilibrium, we know that relative prices of nontradables will fall in the U.S. and increase abroad. But should one presume that the U.S. terms of trade, the price of American exports relative to its imports, need fall as the U.S. trade deficit declines? Essentially, and perhaps surprisingly to most economists, the answer is ‘no’. In the long run, any change in America’s real exchange rate (its terms of trade)associated with righting the international saving imbalance would likely be small with an unpredictable sign.

    From the 1950s into the 1970s, there was a spirited ‘real’ (nonmonetary) literature on the transfer problem. Given a fall in spending in the home country (the transferor) and a rising in spending in the foreign country (the transferee), what would (need to) happen to the terms of trade? The orthodox presumption then as now was that there would be a secondary burden on the home country as its terms of trade deteriorated: the price of exportables would fall relative to importables in order for its trade balance to improve. However, several eminent authors - Paul Samuelson, Harry Johnson, John Chipman, and Ronald Jones - with some heavy mathematical artillery, in the context of a long run real model where resources remained fully employed,successfully questioned the validity of this orthodox presumption.

    In particular, Ronald Jones in his article ‘Presumption and the Transfer Problem’ (Journal of International Economics, 1975) built a model with a nontradeable and tradable sector (importables and exportables) in each country. He showed the relative price of nontradables declines in the transferor, and rose in the transferee, but what happened to the terms of trade was quite ambiguous. Only by making extremely strong assumptions about specialization in production or consumption could Jones get either the orthodox or the anti - orthodox presumption of the change in the terms of trade to hold. However, for any large economy such as the United States with well diversified production and consumption, the effect of a transfer on its terms of trade is ambiguous - and presumably a second - order effect compared to the definite changes in the prices of nontradables relative to tradables at home and abroad.

    One can get an intuitive sense of Jones’s result by noting that as absorption falls in the transferor, and the relative price of its nontradables begin to decline, then its exports will increase and imports decline as resources move into its tradeables sector. If, myopically, one stoppedat this point with adjustment only in the transferor, then it seems as if the orthodox presumption holds: the price of its exports would be bid down relative to the price of imports.

    However, absorption adjustment is (must be) two - sided. As the transferee’s absorption of both tradeables and nontradables increases, its nontradeable prices are bid up relatively so that resourcesare drawn out of its tradeables sector. The transferee’s exports will tend to fall and imports from the transferor rise. This foreign pressure by itself would tend to raise the prices of the transferor’s exports relative to its imports.he. So putting the two offsetting sides together, there isno presumption as which way the transferors terms of trade will move.

    What are the lessons from this ‘real’ long run model of the transfer problem?

    1. Balanced international adjustment in both transferor and transferee is important for preventing a secondary burden on the transferor of having the terms of trade turn against it as its trade balance improves.
    2. Precipitate action to foment a discrete major ‘real’ depreciation of the dollar - which would initially turn the terms of trade against the transferor i.e., the United States, at the start of the adjustment process - is unwarranted. This would be painful but also quite unnecessary. In the long run, when the U.S. trade deficit was substantially reduced through mutual absorption adjustment, little or no change in the initial real exchange rate need characterize the final equilibrium.

    So Martin Wolf’s query ‘the price changes needed to bring that shift [a reduction in the U.S. trade deficit] around may not happen easily under a fixed nominal exchange rate, particularly if they require a sizeable fall in nominal wages’ is not correct. When tradables prices are sticky in the short run, a fixed nominal exchange rate would have the great advantage of keeping the terms of trade fairly constant, rather than fluctuating unpredictably, as mutual absorption adjustment proceeded. There would be some downward pressure on U.S. wagesconcentrated initially in the American nontradables sector; but the 3.5 percent fall in absorption spread out over time could be readily accomplished by slowing wage growth rather than actual cuts in nominal wages.

    Clearly our long - run ‘real’ side model of the transfer problem generates important insights. But the old literature on the transfer problem, however valuable, was remiss in leaving out monetary issues. Let us now turn to the monetary side of the adjustment process involving nominal exchange rates, nominal price levels, nominal wages, and nominal money supplies both in the short and longer runs.

    Exchange Rate Clubs and the Monetary Approach

    Specialists in exchange rate economics fall into two distinct clubs: A and B. Members of Club A, by far the larger group, have been brought up since they were undergraduates on the elasticities model of the balance of trade. Besides being algebraically tractable, the microeconomics of this model seem intuitively plausible. With nominal export prices ’sticky’ in each country’s currency in the short run, the relative price effectsof adepreciation in the nominal exchange rate seem to go in the right direction for reducing a trade deficit. The depreciating country’s exports become cheaper in world markets and it sells more, and its imports become more expensive in the domestic currency so it buys less, so the trade balance allegedly improves. Members of Club A focus on this link between the real, i.e., inflation - adjusted, exchange rate and the real trade balance.

    What about absorption adjustment? True, Club A’s more sophisticated members also worry about saving - investment imbalances across countries; but these don’t easily fit into the elasticities model, which dominates their thinking. They still see a devaluation of the real exchange rate to be useful for easing the adjustment process if, say, a country with large trade and fiscal deficits reforms itself by phasing out the fiscal deficit so that its real international balance of trade can improve. Indeed, they may see the devaluation to be a useful export stimulus to the economy to offset the deflationary impact of the contraction in absorption. (But this exchange rate offset is unnecessary if one remembers from the transfer problem that absorption adjustment is two - sided: expansion in the transferee offsetting contraction in the transferor so as to preserve a rough macroeconomic balance in the two countries.)Martin Wolf is a member of Club A in very high standing.

    Club B is much smaller, and is mainly made up of monetary economists (excluding monetary cranks, of course!). Characteristically, members of Club B emphasize the linkages between national monetary policies and nominal exchange rates in financially open economies. Causality goes in both directions. A floating nominal exchange rate today is determined by what forward - looking investors think the national monetary policy will be relative to that in other countries into the future. Conversely, official action taken today to peg an exchange rate into the indefinite future - or negotiate some major change such as a devaluation or appreciation - requires that relative national monetary policies must (eventually) be changed to support it. Otherwise, the officially assigned path for the nominal exchange rate cannot be sustained. Countries that agree (perhaps by some Plaza - type negotiation) to having their currencies appreciate are also agreeing to follow a deflationary future monetary policy relative to greater inflation in the depreciating country.

    I am sorry to report that membership in Club B is quite exclusionary. It won’t admit economists who believe that governments can manipulate real exchange rates on a sustainable basis - let alone those who believe that exchange rate changes can compensate for saving - investment imbalances across countries. As narrow - minded monetary economists, members of Club B believe that central banks should aim only to stabilize the national price level. They are most fearful of having to alter the national monetary policy to support either an exchange rate appreciation or depreciation that the Ministry of Finance negotiates (mistakenly in their view) to ‘correct’a trade imbalance with some foreign country.

    Being monetary economists, Club B’s members can easily understand international currency asymmetry: why it is convenient, and even necessary, for one currency such as the dollar to dominate international finance as a vehicle and invoice currency. They understand the logic of an intervention system where other countries use the dollar as their key intervention currency, and the United States (as center country) does not intervene in order to minimize potential conflicts.

    On occasion, a central banker in good standing in Club B may opt to peg (stabilize) his country’s nominal exchange rate to the dollar in order to better anchor his country’s price level - as many countries in Asia now do. Unfortunately, members of Club A may see this as a mercantilist plot for stimulating exports by keeping its real exchange rate undervalued. But, with a fixed nominal exchange rate, Club B members know any such ‘undervaluation’ would be washed away by inflation after which the domestic price level would stabilize as long as the dollar itself was stable.

    What is truly dismaying for members of Club B is when so many members of Club A are opting for a massive dollar devaluation on the incorrect assumption that it would reduce the U.S. trade deficit. Club B members know that such a great monetary shock from nominal depreciation would eventually result in inflation in the United States or (relative) deflation in countries on its monetary periphery. Nobody would know how this division between inflation and deflation would play out. (In Japan in the late 1980s through the 1990s, it was deflation from the appreciating and then overvalued yen that led to Japan’s zero - interest liquidity trap and ‘lost decade’.) However, the end result would be to wipe out any sustained change in the dollar’s real exchange rate, although its nominal exchange rate remains depreciated.

    True, following a nominal devaluation, there are lags before prices begin to rise at home or fall abroad. But even in this short run of price ’stickiness’, the balance of trade of the depreciating country is unlikely to improve.

    First, for imports already contracted for and invoiced in a foreign currency like the euro, the U.S. importer would have to pay more (depreciated) dollars for the European goods he had agreed to buy. Economists call this the ‘J’ curve effect.

    Second, with a temporarily real depreciation of the dollar, international investors would see a window of opportunity for a year or two to undertake physical investments at lower cost in the United States. Conversely, they would see countries with appreciated currencies to be more expensive. As a consequence, an investment - led spending boom in the United States would increase imports and a slump abroad would reduce imports of American goods. The upshot is that the net effect on the U.S. trade balance in this intermediate term of two years or so would be ambiguous - even though the foreign currency prices of American exports in world markets had been (temporarily) reduced.

    However,the really big incidental negative from a deep nominal devaluation of the dollar is the monetary upheaval associated with debasing the key currency of the international monetary system. Such an event did occur in August 1971 when President Nixon imposed import tariffs in order to force all the other industrial countries to appreciate against the dollar. Because this ‘Nixon Shock’ was so well telegraphed in advance, the huge flight from dollar assets into foreign monies led to a loss of monetary control both in the United States as well as in Europe and Japan. Foreign governments had to intervene to resist having their currencies appreciate by more than what was agreed on with President Nixon, and their massive buildup of dollar exchange reserves led them to issue too much base money. In the United States, the flight from dollar assets greatly reduced the demand for U.S. base money, and the Fed, not realizing this, continued to increase the supply of base money. The result was the disastrous worldwide inflation of the 1970s into the 1980s. However, the inflation was greater in the United States, which had depreciated, than in Germany and Japan, who had been forced into appreciating. But growth in real income and productivity declined everywhere. (Ironically, the net U.S. trade balance did not improve!)

    Members of Club B lie awake at night wondering if such a calamitous event might happen again.

    Posted by: Ronald McKinnon | December 29th, 2006 at 12:45 pm | Report this comment
    

15. Martin Wolf: I appreciate Ronald McKinnon’s lengthy and valuable response to my attempt to elucidate his model. I think readers should examine what he says themselves. But it might be helpful if I spelt out a few points of disagreement.

    Let me start with the transfer problem. As will be clear from reading Ron McKinnon’s lengthy exposition, there are two relative prices that matter in these analyses: the terms of trade; and the relative price of domestically produced tradeables and non-tradeables. Ron McKinnon refers to the first of these as the “real exchange rate”. I would not do so. I always refer to the latter as the real exchange rate and the former as the terms of trade.

    I agree with Ron McKinnon that a reduction in the US current account deficit requires changes in absorption of resources (that is, in expenditure) there and an increase in absorption by its trading partners. Indeed, I said so. I agree with him, too, that on this assumption the terms of trade (price of exports relative to imports) may not have to change very much. I agree, too, that such a rebalancing of absorption on both sides is desirable and have said so in previous columns.

    So that leaves us with the internal price change – the relative price of tradeables and non-tradeables. I suggested that Ron McKinnon thinks these are the same thing and so that effectively the economy produces just one good. It is evident that he thinks they are not, but that they do have high long-run elasticities of substitution in production. It is easy, he believes, to shift resources between them without sizeable changes in relative prices. Together with the assumption of long-run wage and price flexibility, this produces the conclusion that any internal relative price changes needed to combine full employment with reduction in US absorption of 3.5 per cent of GDP are quite small and easy to achieve.

    Yet some standard academic work on the real exchange rate changes needed to accommodate the required adjustment suggests these are large (see Maurice Obstfeld and Kenneth Rogoff, “The Unsustainable US Current Account Position Revisited”, November 2004). To understand why the price changes might be large, it is important to appreciate the nature of this adjustment. A sharp reduction of 3.5 per cent in expenditure would fall predominantly on non-tradeable goods and services, generating recession rather more than an improved current account. The relative price changes are needed to shift spending towards the now unused non-tradeables, thereby ending the recession, and production towards tradeables, thereby closing the external deficit. In practice, the needed expansion in tradeable output is probably about 10 per cent.

    In any case, the first issue of disagreement concerns the size of the real exchange rate changes needed to combine the external adjustment with something close to full employment. Moreover, the real exchange rate is defined here as the relative price of tradeables and non-tradeables within a country.

    Now let us turn to the monetary side of the story. Again, let me repeat my view. I certainly don’t think that adjustment can occur without changes in absorption (expenditure). Nor, for the reasons already indicated, do I think the elasticities analysis (first advanced by the Cambridge economist Joan Robinson) is especially relevant. Changes in nominal exchange rates may be helpful, however, to the extent that they accelerate the needed change in the internal relative prices of tradeables and non-tradeables. This is a second area of disagreement.

    I think it is perfectly reasonable for economists to disagree on these two points. Maybe the needed internal relative price change is modest, as Ron McKinnon believes, and maybe a nominal exchange rate change isn’t needed to facilitate it. Yet let us think a little more about the monetary analysis put forward by him, because it helps to clarify the challenge now confronting the world more precisely.

    There are many floating currencies in the world. Let us assume that each is in the charge of a central banker that targets domestic inflation. There is no question of a devaluation or a revaluation in such cases. The exchange rate just does what it does. It is not clear, however, that the sole determinant of exchange rates in these cases is monetary policy. Also important is the desire to purchase assets denominated in those currencies. Monetary policy may tighten in a country, but if the desire to hold its assets falls, the exchange rate will still fall. In the column that started this discussion I was talking about the already visible consequences for the dollar of just such a shift.

    Yet there are also, as Ron McKinnon notes, currencies that are linked to an anchor currency. China is today’s most important example. Is this, as Ron McKinnon puts it, “a mercantilist plot for stimulating exports”. In itself, the answer is: no. As it is being operated, however, the answer is: yes. Fixed exchange rate regimes have rules that we have known since David Hume. When a country is accumulating reserves, its money supply should expand. That should stimulate spending and so raise prices of non-tradeable goods and services. The combination should continue until the reserve accumulation ceases (i.e. until net private capital flows balance the trade position) and a new equilibrium is reached. Correspondingly, when reserves fall, monetary policy should be tightened, spending fall and the prices of non-tradeable goods and services decline.

    If such a process had been allowed to work in China, there would be no cause for complaint. This would be a perfectly good alternative to allowing a nominal exchange rate adjustment. Indeed, given China’s undeveloped monetary system, it might well be a superior one, as Ron McKinnon argues.It is also perfectly reasonable for a country to try to smooth this process by attempting some sterilisation of the reserve inflow. It is unreasonable, however, to thwart it altogether, as China has been trying hard (and hitherto successfully) to do. China’s intervention is very large indeed, at close to 10 per cent of GDP. As a result of this success, China has prevented an appreciation of the real exchange rate (defined as I have defined it, in terms of the relative prices of tradeables and non-tradeables). This has generated an increasingly undervalued real exchange rate.

    The point here is that a fast-growing poor country, such as China, should have an appreciating real exchange rate. This is the so-called “Balassa-Samuelson” argument discussed in an earlier forum. Such an appreciation has largely been prevented. China’s underlying situation is not like that of Japan in the 1980s, but is more like that of Japan in the 1960s, when the latter’s inflation rate was consistently higher than in the US (because of the relative rise in the price of its services) and so its real exchange rate was appreciating.

    Evidently, I am claiming that China can manipulate its real exchange rate and is, in fact, doing so. Ron McKinnon insists that Club B won’t “admit economists who believe that governments can manipulate real exchange rates on a sustainable basis”. I would argue that his position is incorrect on his own principles. Club B should admit that countries can manipulate the real exchange rate. They can do so, not by manipulating the nominal exchange rate, but by manipulating savings. I would also argue (and indeed have been doing so in these forums) that China is doing just this.

    Think of a very simple example. Assume that a country’s nominal interest rate is the world interest rate (as it ought to be if it has a fixed exchange rate). The level of investment is then determined by this rate of interest and expected inflation. Assume, too, that the government is able to influence the country’s rate of saving (that is to say, so-called Ricardo equivalence does not hold). In China’s case, it can obviously do so, since government savings and savings of state owned enterprises (which are enormous) are under its control and there is no reason to suppose that households fully offset those savings.

    Given the investment rate and the government decision on the savings rate, there is an equilibrium real exchange rate and consequent current account balance that delivers full employment (more precisely, zero output gap). Moroever, for every level of saving, there is a somewhat different such equilibrium real exchange rate. As the savings rate rises and the needed current account balance needs to move further into surplus, the equilibrium real exchange rate needs to be more depreciated – that is, the price of tradeables has to be higher relative to that of non-tradeables - to achieve internal balance in the economy (roughly full employment).

    In the above example, I have ignored what determines the expected rate of inflation. Under a fixed exchange rate, inflation in tradeable goods prices is the world rate. Inflation in non-tradeable goods depends on domestic demand and so on monetary growth. The government can determine the rate of overall inflation if it can control the growth of the money supply. China can do so, by sterilising the inflow and imposing credit controls, if necessary, on the banks it owns. These have proved perfectly effective mechanisms. Even though the capital flow from China is overwhelmingly in the form of reserve accumulations that would normally be expected to undermine the policy of fixing the real exchange rate, by generating overheating through monetary expansion and expectations of high inflation, China has prevented this from happening. It has, in short, fixed the real exchange rate.

    To make my view clearer, I do not think China’s savings are driving the real exchange rate and the current account but rather the other way round. I think the same is true in the US, too. Indeed, I find it strange that Ron McKinnon does not consider this possibility. After all, unless one is a strict Ricardian one knows that government policy can affect savings rates.

    China has a more or less fixed nominal exchange rate. It also has a target of very low inflation. The two together determine the real exchange rate. At this real exchange rate, there exists a current account surplus and corresponding savings rate that delivers satisfactory levels of activity. Now suppose that the current account surplus suddenly shot up one year, because of greater than expected export capacity. China’s authorities would be concerned about overheating. So what would they do? They would tell local authorities and state-owned enterprises to invest less. They would tell banks to advance less credit. They would sterilise the reserve accumulation. These policies would generate the surplus of savings over investment needed to accommodate the current account surplus without either excess or inadequate demand. The real exchange rate tail wags the savings dog.

    Exactly the same thing applies to the US, but in reverse. Assume the Fed has an inflation target and the rest of the world delivered a certain set of nominal US exchange rates. (Remember that the US does not have an exchange rate policy.) Then the US too has a given real exchange rate that is consistent with a certain current account deficit at full employment. Policy then needs to ensure that spending in the US is large enough to employ the capacity of the non-tradeable goods and services sector fully, while also generating the needed demand for tradeable goods and services. In the US today that overall level of demand is 107 per cent of GDP. So the Fed would run a monetary policy that, given the US fiscal deficit, generated the right level of excess demand (i.e. savings relative to investment) in the private sector. If it failed to do so, the economy would fall into recession, inflation would fall and it would have failed to achieve its domestic goal. Again, the real exchange rate tail wags the savings dog.

    A big difference here then is between a view that the savings rate in an economy cannot be influenced by government policy and a view that it can be. This is why in my own policy recommendations, the starting point has not been a change of exchange rate policy in China but a change in spending policy, coupled with a willingness to let domestic prices move in response. If China wants a fixed nominal exchange rate, that is fine. But if it tries, as it is now doing, to fix the real exchange rate, too, that is not so fine. Indeed, for reasons discussed in previous forums , this creates significant distortions in the economy.

    In fact, I do agree with Ron McKinnon that a big dollar adjustment could be quite destabilising, though I see no reason why it should be as destabilising as in the 1970s, given the commitment to price stability in all the core countries. I agree, too, that monetary policy should not be targeted at a given (depreciated or appreciated) nominal exchange rate, but would argue that this is precisely the mistake China is now making. I agree, finally, that it may well make sense for China to keep a fixed nominal exchange rate, provided it accepts the full logic of such a system, including the need to allow its monetary consequences to work through the economy.

    I appreciate the effort Ron McKinnon has put into educating me and readers of this forum. I hope the discussion has at least clarified differences better than before.

    Posted by: FT Forum - Martin Wolf | January 2nd, 2007 at 5:50 pm | Report this comment
    

16. Ronald McKinnon: Dear Martin. Your comment of December 29 cracking “McKinnon in a Nutshell” made very good general points on which we are close to agreeing - including when using a fixed exchange rate to anchor the national price level may be a legitimate and preferred monetary strategy.

    On the current American dilemma, we agree that if absorption adjustment is two sided - a gradual fall in spending in the US and coupled with a similar gradual increase abroad ­there is no presumption as to which way the terms of trade have to move. But the relative price of non-tradables would gradually fall in the US and gradually rise abroad. You think that some change in the nominal exchange rate may be useful in facilitating these latter changes, whereas I think it would be unnecessary and potentially confusing. In the short run when export prices are sticky in the domestic currency, the main initial impact of a discrete dollar devaluation would be to worsen America’s terms of trade, export prices fall relative to import prices and there is a secondary burden on America in the adjustment process. Whereas in long-run equilibrium, it could well turn out that no change in the terms of trade had been necessary - as per the old literature on the transfer problem.

    However, we may disagree most on the monetary dilemma facing China. You argue that it is unreasonable for China to use a fixed exchange rate and at the same time to thwart the internal monetary expansion associated with its huge reserve accumulation. You argue that if internal monetary expansion was greater, i.e., sterilisation was less, that China’s balance of payments surplus would disappear as its internal price level rose behind the fixed exchange rate­and the price of nontradables increased more relative to tradables. This old Hume-like argument would indeed be valid if the final equilibrium was a zero net trade balance.

    Instead, suppose China is a natural surplus-saving country ­as are most others in East Asia. Whether this results from the huge pull from the saving-deficient United States or reflects the intrinsic high-saving characteristics of Chinese households and firms is difficult to sort out­it is probably some of both. If Chinese financial institutions were fully developed and there were no capital controls, one could imagine a final equilibrium where China’s current account ( saving surplus) amounted to, say, 5 per cent of GDP but almost all of it was privately (non-state) financed. Chinese insurance companies, banks, pension funds, and so on, would all be acquiring claims on foreigners that created a capital outflow that more or less balanced the current account surplus. There would be no unusual accumulation of official exchange reserves­or unusual pressure to expand the domestic money supply. This idyllic equilibrium would be similar to late 19th century Britain where its even larger current account surplus could be privately financed because of Britain’s highly developed capital markets and because the pound sterling was an internationally accepted money that dominated world finance.

    What is wrong with this idyll for describing what China can aspire to today? First and most obviously, China has yet to develop large non state financial intermediaries (although it is moving ahead on this) and its private bond market, which was the centerpiece of 19th century British finance, is just beginning. Second, it has residual two-way capital controls. These partly reflect China’s under developed capital market, although now it is trying to loosen up on capital outflows. Third, and least obviously, the renminbi is not (yet) an acceptable money for denominating international lending. Even without the capital controls, foreigners are reluctant to borrow in renminbi (a minor currency) when the dollar and the euro dominate international finance. This reluctance of foreign borrowers is much greater when there is a threat of renminbi appreciation by an unknown amount. The upshot is that, insofar as the private sector accumulates claims on foreigners, they are dollar denominated­leading to a cumulating currency mismatch within China as the trade surplus continues and these dollar claims pile up.

    The intractability of this mismatch problem can be illustrated by the plight of Japan as an international creditor­also a surplus saving economy. Technologically, Japan is a pre-eminent industrial economy with a highly developed domestic capital markets and no capital controls. But the yen is surprisingly little used in international finance. Consequently, as Japan’s trade surplus continues, Japanese banks, insurance companies, and so on, pile up dollar assets­although their domestic liabilities are mainly in yen. Eventually, just some rumor of possible yen appreciation, can cause these financial institutions to dishoard their dollar assets so that the Bank of Japan has to step in massively­as it did in late 2003 and early 2004­to buy them: Japan’s official foreign exchange reserves about doubled at that time and Japanese private dollar holding fell by the same large amount–about $340 billion. Without this intervention, the yen would have shot up in the foreign exchanges and aborted Japan’s still fragile recovery. The Japanese case is also discussed somewhat more in the attached paper.

    China private (non-state) sector is even less willing and able to intermediate China’s saving surplus internationally. So at the present time in China, the People’s Bank of China is­more or less by default­the principal (almost sole) intermediary for China’s current account surplus. Then foreigners see the pile up of official dollar reserves and complain that China’s exchange rate is undervalued, and that the PBC is rigging the foreign exchange market. Potential private holders of dollar assets within China become even more nervous and less willing to intermediate. But if the PBC lets go of the exchange rate, massive appreciation would follow­with no well-defined upper bound­because there would be no willing private Chinese holders of dollars as the trade surplus continues. The result of massive renmunbi appreciation would be a deflationary shock to the economy in the earlier Japanese mode of the late 1980s through the 1990s. This problem of “conflicted virtue” is discussed more on pages 7-9 in my forthcoming paper for International Finance “Why China should Keep its Dollar Peg: A Historical Perspective from Japan” Jan 2007 (available from me at . The virtual absence of a Balassa-Samuelson price effect in China is discussed on pages 11-12. And conflicted virtue is also discussed in more depth in my book Exchange Rates under the East Asian Dollar Standard: Living with Conflicted Virtue MIT Press 2005.

    Finally, you asserted that the Chinese government controls the national saving rate and can easily manipulate it. In China’s now highly marketized and decentralized economy, the Chinese government has no more control over its national saving rate than the American government has over its. Both governments have been continually surprised. Nobody projected the incredible fall in American personal saving towards zero over the last decade. Similarly, the current surge in corporate profits in China, including in the old “loss-making” state owned enterprises, was entirely unexpected. The attendant fall in personal disposable income relative to GNP over the past two years is a big surprise.

    Anyway, I agree that the Chinese authorities should move to counteract the recent most unexpected saving surge in the Chinese economy. The sudden increase net exports is embarrassing and disconcerting to them. (My next paper is to explain what happened and what to do about it. My tentative hypothesis is that wage growth­while still high­has now fallen below the very high productivity growth because of the threat of renminbi appreciation) I agree with you that if the Chinese government tries to cut back on the current frenetic investment surge (as many foreigners seem to think that they should), the trade surplus would balloon further in the near term­a most uncomfortable macro economic trap!

    But apart from this recent strange episode, one would still expect China to be a surplus-saving economy in some calmer state of more balanced growth­with conflicted virtue being a continual problem in intermediating that surplus. However, a satisfactory solution does not include appreciating the renminbi. The threat of renminbi appreciation has already done damage in pressuring Chinese interest rates to be two or three percentage points less than American (see attached paper). They are not yet at zero in the Japanese mode, but could get there. Because domestic financial markets are still underdeveloped and not fully liberalized, the dollar exchange rate still remains the most convenient anchor for China’s monetary policy­and it would be most unwise to let it slip.

    Martin thanks for being so constructive in putting pressure on me to make my ideas somewhat clearer. You have been most helpful
    All the best

    Ronald McKinnon

    Posted by: FT Economist Forum | January 30th, 2007 at 12:00 pm | Report this comment
    

17. Martin Wolf: This debate could run and run. I thank Ron McKinnon for his lengthy and detailed replies. While I still do not entirely agree with him, I intend to postpone the next round of the debate to another column.

    Posted by: FT Forum - Martin Wolf | February 7th, 2007 at 10:41 am | Report this comment
    

The FT Economists' Forum is a discussion among some of the world's top economists. As a general rule we accept comments from invited members only, but submissions from others will also be considered.

If you are a non-member submitting a comment, please include your relevant academic or financial background.

Post a comment

Name (required)

Mail (will not be published) (required)

As a final step before posting the comment, please type the two words you see in the image beloweight numbers in the audio clip; this test is to prevent automated robots from posting comments.

Type the two words: Type the eight numbers: Unfortunately you provided the wrong reCaptcha answer. Please try again: