Anwar Shaikh and Isabella Weber The U.S.-China Trade Balance and the Theory of Free Trade: Debunking the Currency Manipulation Argument May 2018 Working Paper 05/2018 Department of Economics The New School for Social Research The views expressed herein are those of the author(s) and do not necessarily reflect the views of the New School for Social Research. © 2018 by Anwar Shaikh and Isabella Weber. All rights reserved. Short sections of text may be quoted without explicit permission provided that full credit is given to the source. Anwar Shaikh* The New School for Social Research Department of Economics Isabella Weber Goldsmiths University of London Institute of Management Studies The U.S.-China Trade Balance and the Theory of Free Trade: Debunking the Currency Manipulation Argument Abstract The U.S.-China trade imbalance is commonly attributed to a Chinese policy of currency manipulation. However, empirical studies failed to reach consensus on the degree and kind of RMB misalignment. We argue that this is not a consequence of poor measurement but of theory. The conventional principle of comparative advantage suggests real exchange rates will adjust so as to balance trade. Therefore, the persistence of trade imbalances tend to be interpreted as arising from currency manipulation. In contrast, the Smithian-Harrodian theory explains trade imbalances as the outcome of free trade and sees unequal real competitiveness as the root cause of the U.S.-China trade imbalance. JEL Codes: B17, F10, F31, F32, F60 * Corresponding author: Anwar Shaikh, Professor of Economics, The New School for Social Research, 6 East 16th Street, New York, NY 10003.
 1 1 Introduction The history of globalizing capitalism shows the re-occurrence of lasting trade imbalances under different monetary regimes (Bordo, 2005). Large trade imbalances accumulated again on the eve of the 2008 global economic crisis with the U.S.-China imbalance being the most drastic case (Marchetti, Ruta, & Tech, 2012, p. 1). The overall trade deficit of the United States almost quintupled within ten years, reaching its peak of close to USD 828 billion in 2006 (Figure 1). China on the other hand, turned from a net importer into a net exporter in 1994 and its current account surplus surpassed 420 billion at its peak in 2008 (Figure 2). This reversed position of the U.S. and Chinese external imbalances is reflected in the bilateral trade balance. In the period leading up to the crisis, China increased its share in the U.S. trade deficit from around one fifth in 2002, the first year after China’s accession to the World Trade Organisation (WTO), to about one third in 2008. Ten years after the crisis global trade continues to be highly unbalanced and the U.S. Treasury (2018) notes in a recent report: “The Administration remains deeply concerned by the significant trade imbalances in the global economy” (p. 1). Reducing the U.S. trade deficit is on the top of the Trump Administration’s foreign economic policy, to such an extent that the U.S. President is threatening a trade war (e.g. Trump, 2018). U.S. governments have long been concerned with trade imbalances and administrations of different political orientations have consistently attributed the U.S. trade deficit to currency manipulation by their trade partners. The 1988 Omnibus Trade Act While requires the U.S. Treasury to conduct semi-annual evaluations of unfair exchange rate devaluations by major trading partners. While China has increasingly been the focus, other trade surplus countries including Japan, Korea, Germany, Switzerland, and most recently also India, are also on the Treasury’s ‘Monitoring List’ (U.S. Department of the Treasury, 2018, p. 3). Many leading economists perceive global imbalances as a threat to the world’s economic stability and as a root cause of the 2008 global economic crises (e.g. Obstfeld & Rogoff, 2009; Obstfeld & Rogoff, 2005; Lin, Dinh & Im, 2010). The most widespread explanation for the accumulation of large external imbalances on the eve of the crisis, specifically between the U.S. and China, which were the worlds’ largest in 2008, points to currency manipulation. According to this argument, the Chinese government reduced the value of the Renminbi (RMB) through exchange rate interventions. This lowered the costs of China’s exports to the United States and raised the costs of U.S. imports to China, thereby artificially causing the tremendous trade imbalance between the two countries. This view deems the Chinese “beggar-thy-neighbor devaluation” (Krugman, 2009a) to be the “single largest cause of the U.S. trade deficit and of unemployment” (Scott, Jorgensen, & Hall, 2013, p. 3). In order to accelerate growth and restore full employment the United States would 2 have to reduce its large trade deficit. This could be done at no cost to the U.S. budget, if the United States prevented other countries, primarily China, from manipulating their currency and allow for the Renminbi to return to a competitive level (Bergsten & Gagnon, 2012). Proponents of the currency manipulation hypothesis point to China’s massive accumulation of international exchange reserves as evidence for undervaluation (Bergsten & Gagnon, 2012; IMF, 2010, p. 19; Bergsten, 2006; Bergsten, 2007; Bergsten, 2010; Krugman, 2009b; Bergsten & Gagnon, 2012). China is indeed accumulating international exchange reserves at a high rate (see Figure 3 in the Appendix). The total of Chinese international exchange reserves increased almost 17-fold from 2000 to 2010. In 2013, the Chinese reserves reached a level of USD 3.6 trillion, 60 percent of which is estimated to be held in dollar-denominated assets making China the biggest creditor of the U.S. (Chinn, 2013; Reuters, 2013). However, the link between China’s exchange rate intervention and both the actual currency devaluation as well as the Chinese current account surplus is not directly established. Instead, the vast literature bases itself on the standard assumption that trade would be automatically balanced in unfettered exchange, which in turn leads to various attempts to estimate the undervaluation of the Chinese RMB that must account for its persistent trade surplus. Yet such attempts have failed to achieve any consensus. Estimates of the degree of misalignment vary widely and some studies even find that the RMB is overvalued (Cheung, 2012; Cheung, Chinn, & Fujii, 2010a; 2010b; Dunaway & Leigh, 2006; Cline & Williamson, 2007). This paper argues that the great unity across the political spectrum with regard to currency manipulation as the cause of trade imbalances is rooted in a shared Ricardian outlook on international trade. We analyze the Ricardian theoretical underpinnings of the currency manipulation argument and presents the Smithian-Harrodian theory of trade as an alternative view. The currency manipulation argument is stems from David Ricardo’s radical departure from Sir James Steuart and Adam Smith’s theory of international trade. Ricardian trade theory predicts trade to be balanced in the long run. This is based on the assumption that real exchange rates will adjust so as to guarantee equal competitiveness. Following this logic, the diversion from this ‘natural’ state of affairs reflected in the persistence of trade imbalances can only be due to government intervention impeding the free movement of exchange rates. But the ‘natural’ state of global balances is a fragile theoretical construct. As pointed out by Harrod (1957), the assumption of automatic exchange rate adjustment breaks down when the quantity theory of money is rejected. If a liquidity theory of the rate of interest is assumed instead, a change in the quantity of money induced by trade imbalances under free trade and well-functioning markets leads to capital flows in the opposite direction: from surplus to deficit countries. From this viewpoint international 3 creditor and debtor positions as well as persistent long-term imbalances can be understood as an outcome of global capitalist competition among countries with unequal real costs of production rather than being ‘artificially’ induced by market distortions (Shaikh, 2016, pp. 520-522). As an implication, real exchange rates will be regulated by relative real costs (Shaikh and Antonopoulus, 2012a; Shaikh, 2016, Ch. 11 Sections V-VI). The paper is structured as follows. The next section briefly reviews the central position the currency manipulation argument has occupied in U.S. foreign economic policy over the last three decades. The third section evaluates the empirical studies which estimate the degree of exchange rate misalignment based on the extended purchasing power parity (PPP) and the macroeconomic balance approach. The fourth section analyzes the Ricardian theoretical foundation of these models and contrasts the currency manipulation argument with the Smithian-Harrodian view. The final section summarizes the opposed perspectives on the U.S.-China trade imbalance arising from the two theories of international trade. 2 Monitoring Currency Manipulation – An Integral Element of U.S. Foreign Economic Policy The accusation against China for manipulating its currency has been a central theme in Trump’s belligerent foreign trade policy until a recent, unexpected turnaround. When Donald Trump announced his candidacy in 2015 he pronounced: “They’re [the Chinese] devaluing their currency to a level that you wouldn't believe. It makes it impossible for our companies to compete, impossible. They’re killing us.” (CNN, 2015) At the Republican convention at which Trump was nominated as the candidate, he pledged he would stop China’s “devastating currency manipulation” and added “they are the greatest currency manipulators ever!” (Rauhala, 2016). President Trump’s Director of Trade and Industrial Policy, Peter Navarro (2011), proclaims in his book Death by China under the heading “Death by Currency Manipulation”: “China’s manipulation of its currency, the yuan, is the tap root of everything wrong with the U.S.-China trade relationship.” (p. 67) Yet while the rhetoric of Trump and his advisors represents a new level of aggression, the claim of currency manipulation has been a central concern of U.S. governments long before Trump. The currency manipulation argument is deeply engrained in U.S. foreign economic policy. Since the 1988 Omnibus Trade Act the U.S. Treasury issues semi-annual reports to “consider whether countries manipulate the rate of exchange between their currency and the U.S. dollar for purposes of preventing effective balance of payments adjustments or gaining unfair competitive advantage in international trade” (Section 3004). Initially the reports targeted the Asian Newly Industrialized Economies (NIE) (i.e. Korea, Taiwan, Hong Kong and Singapore), with whom the U.S. was 4 running increasing trade deficits. In the first report of 1989 the Treasury “concluded that Taiwan and Korea engaged in such ‘manipulation,’ within the meaning of the legislation.” (U.S. Department of the Treasury, 1989, p. 12). China was added to the monitoring list in 1991, just after the demise of the Soviet Union, and has stayed on the list to the present day. Since 1994, when China unified its dual-track exchange rate, the Treasury has not found the legal criteria for currency manipulation fulfilled (Gao, 2005; U.S. Department of Treasury, 2018). Nevertheless, the Treasury has consistently pointed to the possibility that China might use currency manipulation and the accusation of unfair RMB devaluation has been a recurring claim in the U.S. policy discourse. A general pattern emerges from the U.S. Treasury reports over the last three decades: The U.S. government exerts continues pressure by threatening to accuse the East Asian surplus economies of currency manipulation and demands that they enter into bilateral negotiations on a wide array of market liberalization policies, such as exchange rate and capital account liberalization, elimination of import restrictions, and market determined domestic prices. In the aftermath of the 2008 crisis the tensions between the U.S. and China over the question of currency manipulation reached a new height under the presidency of Obama. Obama’s Treasury secretary Geithner stated in 2009: “President Obama – backed by the conclusion of a broad range of economists – believes that China is manipulating its currency.” (Wearden, 2009) Obama followed the logic of currency manipulation when he warned in 2010: “One of the challenges that we’ve got to address internationally is currency rates and how they match up to make sure that our goods are not artificially inflated in price and their goods are artificially deflated in price. That puts us at a huge competitive disadvantage.” (Weaver, 2010, emphasis added). The Obama administration enacted the Trade Facilitation and Trade Enforcement Act of 2015 in addition to the 1988 Act to strengthen the U.S. ability to take action against currency manipulation. The Group of 20 also adheres to the currency manipulation thesis in the context of its commitment to “a lasting reduction in global imbalances” and proposes “to move more rapidly toward more market- determined exchange rate systems and exchange rate flexibility to reflect underlying fundamentals, and avoid persistent exchange rate misalignments” (G-20, 2013, emphasis added). In more recent years it has been acknowledged that China has taken steps to appreciate the RMB but the adherence to the currency manipulation argument continues. The Treasury (2017) finds that as a consequence of China’s slow and gradual strengthening of the RMB the “distortion in the global trading system resulting from China’s currency policy over this period [i.e. before the RMB appreciation] imposed significant and long-lasting hardship on American workers and companies” (ibid.). In April 2018, Trump made a U-turn and announced that he would not label China a currency manipulator (Baker, Lee, and Bender, 2018). Around the same time, the most recent 5 Treasury (2018) report once more did not legally establish that China is manipulating its currency. Nevertheless, the accusations that previous currency intervention caused the piling up of global imbalances continues and China is again warned “to refrain from engaging in competitive devaluation and to not target China’s exchange rate for competitive purposes” (U.S. Department of the Treasury, 2018, p. 4). The Treasury has not found any country to fulfil its own criteria for currency manipulation since 1994 while observing persistent global trade imbalances, but it continues to attribute these imbalances to misaligned currencies. 3 Measuring the RMB Misalignment Despite great attention attributed to the RMB misalignment by policy makers, the degree of adjustment needed to reduce the Chinese current account surplus to a certain target level remains highly contested. The logic of the currency misalignment argument assumes that if there were no market distortions, the real exchange rate would converge to a certain equilibrium level which would balance the current accounts in line with selected macroeconomic fundamentals. However, several literature reviews have demonstrated that there is no consensus on how to determine the equilibrium exchange rate that brings about this external balance nor on the level of adjustment needed (Cheung, 2012; Cheung, Chinn, & Fujii, 2010a; 2010b; Dunaway & Leigh, 2006; Cline & Williamson, 2007). Since currency misalignment is defined as the deviation of the real exchange rate from its equilibrium level, there is consequently also no common measure of the currency misalignment (Cheung, 2012). While there was a proliferation of studies aiming to estimate the RMB misalignment in the decade following China’s accession to the WTO in 2001 and the analysis in this section focuses on this period, “the search for a consensus on whether the Renminbi is undervalued continues” (Almås et al., 2017, p. 19). This lack of consensus results from the current state of exchange rate economics. There is no generally accepted exchange rate model (Cheung, Chinn, & Fujii, 2010b). Most widely applied are various incarnations of the Purchasing Power Parity (PPP) and the macroeconomic balance approaches (Ahmed, 2009; Cheung, Chinn, & Fujii, 2010a; Cheung, Chinn, & Fujii, 2010b; Dunaway & Leigh, 2006; Cline & Williamson, 2007). Both leave considerable room for judgment with regard to the model specifications (Cheung, 2012; Dunaway & Leigh, 2006). Consequently, studies using these approaches to estimate the misalignment of the RMB yield widely varying results. The following two sections introduce the PPP and the macroeconomic balance approach on a theoretical level and provide an overview of the estimation results by studies conducted since China’s accession to the WTO in 2001. 6 3.1 The Purchasing Power Parity (PPP) Approach The PPP approach derives the equilibrium exchange rate directly from a comparison of price levels and estimates the degree of misalignment as the deviation of the actual real exchange rate from this equilibrium level, which also indicates the adjustment needed to overcome the current account imbalance. The PPP hypothesis is based on the law of one price (LoP), which states that in the absence of trading barriers and transaction costs competition equalizes the prices of tradable goods across economies. If is the nominal exchange rate, the foreign price index and the domestic price ∗ index, then with no misalignment, the same bundle of goods would have the same price across 𝑒𝑒 𝑃𝑃 𝑃𝑃 countries denominated in a common currency1: 1) ∗ 𝑒𝑒 𝑃𝑃 = 𝑃𝑃 Since price level data is generally in terms of index numbers, this implies that the real exchange rate (Q) converges to a stationary value around some constant level2: 2) = constant ∗ 𝑒𝑒 𝑃𝑃 𝑄𝑄 = 𝑃𝑃 The nominal exchange rate is expected to move to adjust the price levels in the long run (Cheung, Chinn, & Fujii, 2010a). If currencies are ‘misaligned’ by PPP standards, the real exchange rate as defined above will be non-stationary. According to standard trade theory, this might be either due to market interventions that prevent the equalization of prices or to currency interventions that hinder the nominal exchange rate from adjusting. The crucial challenge for an empirical evaluation of PPP is to find price indexes that represent an identical basket of goods in the countries of comparison (Shaikh & Antonopoulos, 2012; Shaikh 2016, pp. 528-535)3 Another fundamental obstacle to the PPP approach in the context of the debate over the RMB misalignment with respect to the dollar is that by definition the United States can never be blamed for over- or undervaluation, as the dollar is the numeraire with an exchange rate always equal to the market rate (Cline & Williamson, 2007). 1 Note that the composition of goods as well as the ratio of tradables to nontradables must be the same in both price indexes and . ∗ 2 See Shaikh (2016, pp. 517-527) for a detailed discussion of this point. 𝑃𝑃 3 Th𝑃𝑃e simplest and methodologically most questionable PPP approach is the Big Mac index, compiled by The Economist. It tries to get around the problem by using a basket containing just one good, the Big Mac. 7 Since the PPP hypothesis was found to be very weak empirically, in particular, real exchange rates do not converge to any stationary level (Rogoff, 1996; MacDonald & Ricci, 2001)4, a number of extensions were introduced in hope of a better determination of the trade balancing equilibrium exchange rate. These are usually classified as extended (Dunaway & Leigh, 2006) or enhanced (Cline & Williamson, 2007) PPP approaches. This paper will refer to them as extended PPP approaches. Most common is the introduction of the Balassa-Samuelson effect to accommodate the deviation of the real exchange rate from purchasing power parity (Bosworth, 2004; Dunaway & Li, 2005). Note that this effect is based on the assumption that wages grow with productivity and that the economy is at full employment.5 The Balassa-Samuelson effect is intended to rationalize the empirical regularity, that in contrast with the LoP, higher income countries which are assumed to have higher productivity levels show higher price levels (Penn Effect). In an attempt of a more precise calibration of the equilibrium exchange rate so-called macroeconomic fundamentals are introduced in PPP based approaches often labeled as behavioral effective exchange rates (BEER) (Cline & Williamson, 2007).6 These include for example net foreign assets (NFA), terms of trade (ToT), government consumption, trade openness and price controls.7 Panel regression or single country time series analysis is used to establish an equilibrium relationship between these variables and the real exchange rate (Dunaway, Leigh, Li 2006). The equilibrium exchange rate is derived from the past average of the economy under examination or from the average over a number of countries (Dunaway, Li 2005). This method is based on the assumption that the exchange rates are in equilibrium in the long run, whereas trade may be unbalanced trade but only in so far as this is reflective of the macroeconomic fundamentals. The selection of variables to extend the basic PPP approach, as well as the selection of the proxies used for these variables are guided by data availability and depend on the judgment of the researchers. In addition, the choice of the country sample and the examination period will affect and can be used to manipulate the estimation results (Cheung, Chinn, & Fujii, 2010a; Dunaway & Leigh, 2006). 4 The real exchange rate is not stationary over the short run (Isard, 1995, pp. 63-65). It does revert to a “target level” over runs of 10-20 years, but this is not the PPP level (Engel, 2000, p. 21). Even if there was a reversion to a non-stationary mean, the “speed of convergence is extremely slow” (Rogoff, 1996, p. 647). For a more detailed discussion see Shaikh and Antonopoulus 2012a. 5 The full-employment condition is not fulfilled in the Chinese case, as discussed below. 6 The IMF calls a similar approach “reduced form equilibrium real exchange rate” in its “Methodology for CGER Exchange Rate Assessments” (2006). Ahmed (2009) subsumes what Cline and Williamson (2007) call BEER under the label extended PPP approaches. 7 With respect to NFA it is assumed that debtor countries need more depreciated real exchange rates as compared to creditor countries. Higher ToT are expected to appreciate the real exchange rate due to an income effect. Government consumption would increase the demand and ultimately the price of nontradables and appreciate the real exchange rate. Greater trade openness would decrease domestic prices and appreciate the real exchange rate. Price controls might increase or decrease domestic prices and appreciate or depreciate the real exchange rate (IMF 2006). 8 3.2 The Macroeconomic Balance Approach Whereas the PPP approach attempts to directly estimate the equilibrium exchange rate, the macroeconomic balance approach starts from the current account. The attributed misalignment is then evaluated in terms of the gap between the projected current account balance and a targeted norm.8 The question to be answered is: How much the real exchange rate needs to adjust in order to close the current account gap under conditions of internal balance? The analysis is usually conducted in three steps (Dunaway & Leigh, 2006; Dunaway & Li, 2005): First, the so-called underlying current account balance is estimated. Second, the equilibrium current account or “current account norm” is approximated which defines the targeted norm. Third, a trade model is estimated to calculate the exchange rate movement needed to close the gap between the actual and equilibrium current account balance. The underlying current account calculated in the first step does not correspond with the observable current account. Instead, it represents a ‘thought experiment’ and is defined as the current account in percent of GDP that is expected to emerge over the medium term at prevailing exchange rates assuming that the trading economies will return to noninflationary full capacity employment (i.e. internal balance), the effects of past exchange rate changes will cease and cyclical effects will be adjusted (Borowski & Couharde, 2003; IMF, 2006). In this sense the underlying current account is the state which would evolve if all trading partners were to resume internal equilibrium but at a given real exchange rate. In the second step, an equilibrium relationship is established between certain so-called macroeconomic fundamentals and the current account in order to estimate the equilibrium current account which is assumed to be sustainable in the long run (Ahmed, 2009). This equilibrium can be different form balanced trade but only in so far as this is in line with the macroeconomic fundamentals. Note that cost competitiveness is not part of the macroeconomic fundamentals. One set of fundamentals is derived by panel data techniques building on the assumption of stable structural relationships between saving, investment, and external current account balances (Dunaway & Leigh, 2006; Dunaway & Leigh, 2006; Chinn & Prasad, 2003). The starting point is the truism that in open economies national saving (S) does not have to be equal to domestic investment (I) (Obstfeld & Rogoff, 1996; Obstfeld, 2004). This can be expressed in the following accounting identity: 8 The macroeconomic balance approach was first developed by Williamson (1983) under the label Fundamental Equilibrium Exchange Rate (FEER). It is one of the three methods suggested for exchange rate assessment by the IMF (2006). 9