Body
China may not welcome the attention that has been bestowed on the question of the renminbi exchange rate in recent months, but it is a symptom of the fact that the country has emerged as a significant force in the global economy. That attention is not going to go away. China must learn to live with its new role, just as the rest of the world has to learn to live with an emergent China. It is an honor to have this opportunity of contributing to the process of thinking through what that implies.
I propose to address three topics in this lecture. First, I will try and explain what I mean by a "fundamental equilibrium exchange rate" (FEER) and describe how it is calculated. Second, I will examine what this approach suggests about the desirable value of the renminbi at the present time. I will argue that a substantial renminbi appreciation is indeed desirable, including from the standpoint of China's own needs. Third, I will explain why I believe the Chinese authorities were right to reject the suggestion that they should abolish capital controls and float the renminbi.
The Concept and Measurement of the FEER
The term "fundamental equilibrium exchange rate" was introduced to emphasize that in thinking about what a country's exchange rate should be one ought not to focus on a momentary market equilibrium. One ought instead to think about what in the IMF's original Articles was referred to as "fundamental disequilibrium", which meant an exchange rate at which it would not be possible to achieve simultaneously both of the basic objectives of macroeconomic policy. These were usually formulated as "internal balance", meaning non-inflationary full employment or a similar concept, and "external balance", meaning a balance of payments position that appeared sustainable and desirable. The idea was that an overvalued exchange rate would imply either deflating the economy or running an unsustainable payments deficit; conversely for an undervalued rate. Fundamental equilibrium holds where neither of those undesirable outcomes prevails.
Naturally the FEER refers to the real effective exchange rate, since this is what influences macroeconomic outcomes. China is currently following an exchange rate policy that results in a highly unstable exchange rate in that sense. By holding the bilateral nominal exchange rate against the dollar constant in a period when the dollar has strongly depreciated, China has itself depreciated (it has "ridden the dollar down"). Some of us were surprised, and dismayed, to hear senior Chinese officials describing their government's policy as one of holding a fixed exchange rate; it is most certainly not fixed in the sense that matters to macroeconomic policy.
One can often get a rough idea of whether an exchange rate is overvalued, in fundamental equilibrium, or undervalued by a cursory examination of a country's macroeconomic situation. If a country's economy is overheating at the same time that it has a bigger current account surplus (or a smaller current account deficit) than is needed to maintain a sustainable balance of payments position for the foreseeable future, then its currency is undervalued. If the economy is underemployed and it is losing reserves because of a current account deficit larger than can be financed by sustainable capital inflows, then its currency is overvalued.
More difficult cases occur when different indicators point in different directions. Then a macroeconometric model of some sort is essential to reach a judgment. If, for example, the economy is overheating but the country has an unsustainable payments deficit, then one needs to ask what the situation would be when the excess demand had been eliminated. To answer that question convincingly demands some sort of a model, though in that case a rather simple model might suffice. If one concludes that mere elimination of excess demand pressure would look after the balance of payments, then the currency is not overvalued; indeed, it is conceivable that once the economy reached internal balance the payments position might be in an unsustainable surplus, in which case the currency would actually be undervalued.
If one concludes that a currency is misaligned, then the next issue is to decide by how much. This again inevitably requires the use of some form of quantitative macroeconometric model, though again it may be crude (or even implicit). The sort of model needed here is rather old-fashioned, being one that explains current account outcomes by a measure of economic activity and the real exchange rate.1 That enables one to calculate the change in the real exchange rate needed to achieve a target current account when activity is at some normal, sustainable level. No one doubts that there are lots of other interesting issues to consider, such as how to achieve the exchange rate that the analysis indicates to be desirable, but that does not make the simple model illegitimate. One also needs, either as part of the same model or independently, a way of estimating the change in the nominal exchange rate needed to achieve the desired change in the real exchange rate, given that depreciation can be expected to cause some inflation that will reduce the change in the real exchange rate.
However, exchange rates are inherently the business of more than one country. A small country can expect that the rest of the world will allow it to choose its exchange rate without too much interference, but large countries have to recognize that their partners have a legitimate interest in what they choose because their choice influences those partners' effective exchange rates. It is for this reason that I have always emphasized that FEERs need to be calculated simultaneously for all the large countries, as a general equilibrium exercise. One needs to ask what set of exchange rates would yield a set of mutually satisfactory balance of payments outcomes when all the economies are at reasonably full employment. To do that one certainly needs to use a formal macroeconometric model.
If it were true that countries cannot in principle achieve a set of mutually satisfactory balance of payments outcomes, because countries' have antithetical payments objectives, then the exercise I have described would be impossible. There are economists -mostly self-styled Keynesians who do not believe that output is constrained by supply but only by demand-who believe that all countries have a national interest in securing a bigger current account surplus. Most of us regard this as mercantilist rubbish. If output is usually constrained from the supply side, then a more competitive exchange rate has a disadvantage in curbing the resources available for investment as well as an advantage in simulating the desire to invest. The growth-maximizing exchange rate will be that at which these two forces balance one another at the margin.
Assuming that it is not impossible in principle, one still needs to devise a set of current account outcomes that all countries would recognize to be desirable. This is something that (in association with Molly Mahar) I attempted to do as one of the bases for the calculation of FEERs by Simon Wren-Lewis and Rebecca Driver (1998). I have repeated the exercise in a draft paper for a conference that the Institute for International Economics plans to hold in the coming months about the completion of the correction of the dollar's overvaluation of recent years (Williamson 2003). I will refer to the assumptions made about China in the following section. For the moment the point is merely that one needs a set of consistent current account outcomes as an input for a model of the world economy that will calculate a set of target exchange rates. If the inputs add up (allowing for the world statistical discrepancy), then the resulting target exchange rates (the FEERs) will also be consistent.
Application to the Renminbi
Recent evidence would seem to indicate that the Chinese economy is overheating. Bank loans are expanding at a frenetic rate and the pace of investment recalls what happened in other Asian countries in the days before the East Asian crisis.
The statistics tell us that China has been in current account surplus for a long time. Not only that, but it has been a net capital importer, as befits a country that is still underdeveloped but has the fundamentals in place to permit rapid growth if it invests a lot. Because of those two factors together, Chinese reserves have grown to a point where they are now the second highest in the world (after Japan). In other words, China has an external surplus on any criterion one can think of.
China thus falls into one of the cases that are easy to diagnose, where both internal and external considerations point to the desirability of a currency revaluation. This would relieve tensions in the domestic economy as well as diminish the external surplus and reserve accumulation.
A number of reasons have been advanced for resisting an appreciation and maintaining the dollar peg at an unchanged value.
1. "China's current account surplus has recently diminished and will fall further as a consequence of the import liberalization being undertaken to qualify for WTO membership." Yes, the surplus has fallen, but it is still a surplus rather than the deficit that is appropriate for a country in China's situation. And if the domestic economy were to be stabilized (excess demand were to be eliminated), the surplus would grow again. Yes, import liberalization is helping to expand imports and thus reduce the surplus, but the forthcoming abolition of the Multi-Fibre Arrangement will give scope for a substantial expansion of China's exports of textiles and apparel, thus tending to increase its surplus.
2. "If China were to liberalize capital outflows, then there could be a large increase in the outflow of capital and reserve accumulation might well go into reverse." That is true, but it constitutes an argument against precipitate liberalization of capital outflows, not against a revaluation. The reason that many Chinese might want to export capital is not that the domestic return on capital is low, but that they fear a banking crisis once other individuals gain the freedom to withdraw their savings and ship them abroad. The answer is to delay liberalization of capital outflows at least until the banking system has been cleaned up. That clean-up will be delayed the longer the flood of reserves keeps enabling the banking system to make excessive loans, many of which are likely to go bad when times turn difficult.
3. "China's policy of exchange rate stability has helped to stabilize the Asian economy, e.g. during the Asian currency crisis in 1997." It is certainly true that China's refusal to devalue at that difficult time helped limit the scope of the crisis. However, the fact that China was able to pursue that policy even at that time shows that its payments position was strong even then, and it has become much stronger since with the turnaround in Asia's position. A policy of riding the dollar down makes all other Asian competitors fearful of allowing their currencies to join in the needed general appreciation against the dollar, which is hardly a contribution to stability in the present context.
4. "A revaluation would threaten Chinese growth and, because of that, might not even diminish the balance of payments surplus." If growth depended solely on demand, this would make sense. But in fact supply, not demand, limitations are currently constraining Chinese growth. Investment is very high (42 percent of GDP, according to my colleague Morris Goldstein) and being financed by a dangerously rapid growth in bank lending. A renminbi revaluation would provide a little bit more scope for this investment, and would encourage the expansion of consumption. It would also slow the reserve accumulation that underlies the rapid credit expansion. This would diminish the chance of the process ending with a financial crisis, which really would damage Chinese growth.
In short, the reasons that have been advanced for rejecting an appreciation are unpersuasive.
How large a revaluation would be desirable? That depends on how large an adjustment is deemed to be appropriate. In 1996, Molly Mahar and I postulated a current account deficit of 2.8 percent of GDP as a reasonable objective for China over the following years, on the grounds that China is a developing country that can expect to invest productively an excess of domestic investment over the savings generated at home.2 We thought that China would have no difficulty in importing enough capital to finance such a deficit. In fact, China has imported capital, but on a somewhat smaller scale than this, around 1 percent of GDP on average. Perhaps it could have imported more capital if it had been trying to, of course. But in the draft of my new paper on the topic (Williamson 2003), I postulate a target of current account balance-admittedly more as a result of applying mechanically an algorithm that seemed to give reasonable results in general than because I was intellectually convinced that China ought not to be targeting a current account deficit.
The suggestion of my colleagues Morris Goldstein and Nicholas Lardy that China should target a current account deficit of some 1 percent of GDP, to put it in overall balance given the 1 percent capital inflows, seems as reasonable as any. They have stated (in the Asian Wall Street Journal, 12 September 2003) that their calculations suggest a revaluation of 15 to 25 percent. The Director of my Institute, Fred Bergsten, has taken the top half of this range (20 to 25 percent) as his figure. My own guess is that even this may be somewhat conservative: I think it important that any adjustment be big enough to convince market participants that the change is complete, not a first step, since otherwise a revaluation may simply serve to encourage more speculative inflows. But none of us has a satisfactory macroeconometric model at our disposal to back up our estimates or guesses.
After a first step revaluation I agree with my colleagues that China should operate its foreign exchange market with wider margins and should change its peg from the US dollar to a basket of the three major currencies. It should be more willing to contemplate changes in the value of the peg, while keeping each individual change in the peg small so as to avoid creating strong speculative pressures.
The Dangers of Premature Liberalization
While I am critical of the Chinese resistance to a revaluation of the renminbi, I believe that the Chinese authorities were absolutely right to resist the call of US Secretary of the Treasury John Snow for a liberalization of the capital account and a float of the currency. This is because of the danger that liberalization of the capital account would cause many Chinese savers to decide to switch a part of their portfolio to some foreign country, to avoid the fate that may befall those holding money in Chinese banks if and when a financial crisis finally hits. The attempt of many savers to safeguard their position in this way is likely to bring on the very calamity they were trying to avoid having an impact on them, as is normally true in a bank run. The obvious policy response for the Chinese government is to delay capital account liberalization until after the banks have been cleaned up. This is in fact in accordance with the conventional wisdom on the sequencing of liberalization of the financial system, which calls for the liberalization of the capital account to come at the end of the process.
If capital account liberalization did in fact cause a substantial capital outflow, then it might well be that the result of the US Treasury's program would be a depreciation of the renminbi rather than an appreciation. While market fundamentalists may shrug their shoulders and say so be it, more pragmatic economists would regard such an outcome as utterly perverse. It would do little to help China overcome the financial crisis that had been thrust upon it, and it would do nothing at all to help overcome the problem that is posed to the world by the US payments deficit.
I was the World Bank's Chief Economist for South Asia during the Asian crisis. Just after the crisis had broken, there was still some pressure to implement the recommendations of the Tarapore Committee that had reported to the Indian government shortly before the crisis, and which recommended a "gradual" process of capital account liberalization lasting 3 years. I told an Indian newspaper that I thought 30 years would be more realistic. It seems to me that China should be thinking of a similar timescale to complete the liberalization process. That would then put Asian countries on the same sort of timescale that Europe followed in moving back to a liberal economy after World War Two. While the process was slow, it was also thorough, with the result that today no one worries that Europe may revert to a repressed economy. That is vastly better than a rush to liberalize followed by backtracking, as has often occurred in Latin America (with Argentina being a dramatic recent example).
Concluding Remarks
It is now widely acknowledged that China has undertaken by far the most successful transition from socialism to a market economy of any of the formerly centrally planned economies. It would be truly tragic if that success were to be jeopardized by a banking crisis, or for that matter by an international confrontation with countries that conclude they cannot live with a super-competitive China. Both threats seem altogether too likely to materialize if China retains an unchanged peg to the dollar. A substantial revaluation would be good for both China and the rest of the world.
References
Bergsten, C. Fred, and John Williamson, ed. 2003. Dollar Overvaluation and the World Economy. Washington: Institute for International Economics.
Williamson, John. 2003. Is There a Problem of Inconsistent World Payments Objectives? Draft paper prepared for a forthcoming conference of the Institute for International Economics.
Wren-Lewis, Simon, and Rebecca L. Driver. 1998. Real Exchange Rates for the Year 2000. Washington: Institute for International Economics.
Notes
1. Nowadays the tendency is to build models that take variables that are clearly policy instruments as the exogenous variables, rather than activity levels or exchange rates. One can still use that type of model to answer the question posed, by introducing shocks into the model. See Martin Baily's paper in Bergsten and Williamson (2003) for an example.
2. See Table A.11 in the Williamson-Mahar appendix to Wren-Lewis and Driver (1998).
Commentary Type