دلار چه زمانی بر استرلینگ به عنوان ارز بین المللی فائق آمد؟ شواهدی از بازار اوراق قرضه

We offer new evidence on the emergence of the dollar as the leading international currency focusing on its role as currency of denomination in global bond markets. We show that the dollar overtook sterling much earlier than commonly supposed, as early as in 1929. Financial development appears to have been the main factor helping the dollar to overcome sterling's head start. The finding that a shift from a unipolar to a multipolar international monetary and financial system has happened before suggests that it can happen again. That the shift occurred earlier than commonly believed suggests that the advantages of incumbency are not insurmountable. And that financial deepening was a key determinant of the dollar's emergence points to the challenges facing currencies aspiring to international status.

The global economic and financial crisis has lent new impetus to discussions of the future of the international monetary and financial system. Policy makers in countries like China and Russia have openly questioned the viability of the current dollar-based global system. Some advocate moving to a multipolar system in which the dollar shares its international currency role with the euro, the Chinese yuan and/or the IMF's Special Drawing Rights. At the Cannes Summit of November 2011, G20 Leaders committed to taking “concrete steps” to ensure that the international monetary system reflects “the changing equilibrium and the emergence of new international currencies”.1 Others expect this change to develop more spontaneously; they see it as a natural result of the declining economic and financial dominance of the United States and the increasingly multipolar nature of the global economy, together with the advent of the euro and rapid internationalization of the yuan (e.g. Angeloni et al., 2011, Bini Smaghi, 2011a, Bini Smaghi, 2011b, Constâncio, 2011, Dorrucci and McKay, 2011, Eichengreen, 2011, Fratzscher and Mehl, 2011 and Subramanian, 2011). Sceptics object that prospect of a shift to a multipolar monetary and financial system is in fact remote; if it occurs, such a transition will take many decades to complete (Frankel, 2011 and Kenen, 2011). The view that a shift to a multipolar system is unlikely to occur rapidly is rooted in theoretical models where international currency status is characterized by network externalities giving rise to lock-in and inertia, which benefit the incumbent (see e.g. Hartmann, 1998, Krugman, 1980, Krugman, 1984, Matsuyama et al., 1993, Rey, 2001 and Zhou, 1997).2 These models rest, in turn, on a conventional historical narrative, epitomized by Triffin (1960), according to which it took between 30 and 70 years, depending on the aspects of economic and international currency status considered, from when the United States overtook Britain as the leading economic and commercial power and when the dollar overtook sterling as the dominant international currency. The US, it is observed, surpassed Britain in terms of absolute economic size already in the 1870s. It became the leading commercial power, gauged by the value of foreign trade, already in 1913. It was the leading creditor nation by the conclusion of World War I. And yet sterling remained the dominant international currency, not simply during this period, but also throughout the interwar years, according to the conventional narrative, and even for a brief period after World War II. Recent studies, referred to by Frankel (2011) as the “new view,” have challenged this conventional account. Eichengreen and Flandreau (2009), relying on new data on the currency composition of global foreign exchange reserves, show that the dollar in fact overtook sterling as the leading reserve currency already in the mid-1920s — that is to say, more than two decades prior to the date assumed by previous scholars. Eichengreen and Flandreau's “new view” also challenges broader implications of the conventional narrative. First, it suggests that inertia and the advantages of incumbency are not all they are cracked up to be. Second, it challenges the notion that there is room for only one international currency in the global system. Eichengreen and Flandreau show, to the contrary, that sterling and the dollar accounted for roughly equal shares of global foreign exchange reserves throughout the 1920s. Third, the new view challenges the presumption that dominance, once lost, is gone forever. Eichengreen and Flandreau's data indicate that sterling re-took the lead from the dollar for a brief period after 1931. This reinforces the point that the advantages of incumbency in the competition for reserve currency status may be less than commonly supposed. In a companion piece, Eichengreen and Flandreau (2012) show that what was true of reserve currencies was true also of the use of currencies for financing international trade. The dollar overtook sterling as the leading form of trade credit (as the currency of denomination for what were known as “trade acceptances” or “bankers' acceptances”) already in the mid-1920s, not only after World War II. The US achieved this from a standing start — that is to say, despite the fact that dollar-denominated trade credits had been virtually unknown as recently as 1914. Both market forces (financial market development) and policy support (with the Federal Reserve System as a market maker in the New York market for bankers' acceptances) were instrumental in helping the dollar rival and overtake sterling. That said, both New York and London, and both the dollar and sterling, remained consequential sources of trade credit in the 1920s. This again challenges the notion of international currency competition as a winner-take-all game. Some critics have questioned the new view. Ghosh et al. (2011) suggest that the interwar gold standard was special in that gold, not foreign exchange, was the dominant reserve asset, accounting for some two-thirds of international reserves. The fact that gold played a large monetary role then but plays only a small one today may limit the inferences about prospective changes in international currency status that can be drawn from this earlier experience, in other words. Forbes (2012) suggests that, compared to the past, international financial transactions may play a larger role in driving the decision of which unit or units to use internationally. Merchandise transactions, and the importance of a currency and market as a source of trade credit, play a correspondingly smaller one. Thus, inferences about the future are less convincing insofar as they are drawn from the past behavior of trade credits and not from the use of currencies in international financial transactions. In this paper we address these objections and complete the story. We provide new evidence from the interwar years on the use of the leading international currencies, sterling and the dollar, in international financial transactions. This sheds light on a third dimension of international currency status, namely the use of currencies as vehicles for international financing. We focus on the international bond market, bonds being the principal instrument for foreign lending and borrowing in this earlier era prior to the advent of syndicated bank lending.3 Looking at yet an additional aspect of international currency competition is useful for establishing the generality (or otherwise) of the so-called “new view” of international currency competition. In addition, because international bonds were typically denominated in national currencies and not gold, the earlier objection that evidence from reserves data is not insightful for today no longer applies.4 Lastly, we try to go deeper than in previous studies in understanding the factors that helped the dollar surpass sterling. We provide a systematic empirical analysis of the determinants of currency choice in international bond markets during the interwar years. We employ data on the currency denomination of foreign public debt for 33 countries in the period 1914–1946. We focus on bonds issued in foreign markets (“international bonds”) because they were only rarely denominated in the issuer's own currency. Instead, these were denominated in international currencies so as to appeal to international investors. It is thus the denomination of these foreign bonds that shed light on international currency use.5 Our analysis supports the new view in that the dollar had a share almost equal to that of sterling as a currency of denomination for international bonds already in the interwar years. When excluding the Commonwealth countries, which were heavily inclined towards sterling issuance due to their dominion status, the dollar overtook sterling as early as 1929. Our results further call into question two other tenets of the conventional narrative. While sterling lost its pre-eminence in 1929 (again, abstracting from the Commonwealth countries), it subsequently ran neck and neck with the dollar as the dominant currency of denomination for international bonds at least for a brief period. This is at odds with the conventional view that dominance, once lost, is gone forever. Second, much of the 1920s and the 1930s saw the use of both sterling and the dollar as currencies of denomination in international debt markets. This was a bipolar rather than a unipolar currency system. This finding is at odds with the presumption that there is room for only one dominant international currency in the market. Finally, our empirical results suggest that inertia effects in international currency use, while strong, are not insurmountable. In addition to incumbency, financial development was an important macroeconomic determinant of the ability of the dollar to ultimately overcome sterling's initial advantage. Its impact dwarfed that of country size or that of monetary policy and the exchange rate regime. We also find evidence of the importance of more microeconomic factors, such as market liquidity, which strengthens our emphasis on financial development further, but not of hedging or funding cost considerations, whose importance then seem to be confined to the more recent period. The interwar years being the only period since the onset of the industrial revolution when one incumbent unit was dethroned by a competitor as the world's currency, these findings are relevant to discussions of the future of the international monetary system. They suggest that a shift from a dollar-based system to a multipolar system is not impossible. While it will still take time, the shift could occur sooner than commonly believed. Our results point to financial deepening and market liquidity as key determinants of how and when additional units strengthen their international currency status. Contemporary data on the currency of denomination of international bonds are consistent with the existence of this kind of shift. The share of the euro in the stock of international debt securities increased to some 30% in the 2000s from approximately 20% in the 1990s.6 Again, this is inconsistent with the presumption that there is room for only one international currency in the financial domain. The rest of the paper is structured as follows. Section 2 presents our dataset and Section 3 some stylized facts. Section 4 describes our empirical methodology, while Section 5 presents the baseline empirical results. Following a discussion of robustness in 6 and 7 concludes and draws implications for policy and future research.

This paper has provided new evidence on the emergence of the US dollar as the leading international currency, focusing on its role as a financing currency in global debt markets. This evidence challenges the three central tenets of the conventional wisdom on international currencies. First, network externalities, first-mover advantages and inertia matter, but they cannot indefinitely delay the transfer of leadership in the international monetary sphere relative to that in the economic, commercial and financial spheres; they do not dominate to the extent previously thought. Our evidence shows that, abstracting from the Commonwealth countries, the dollar overtook sterling already in 1929, at least 15 years prior to the date cited in other accounts. Even including the Commonwealth countries, which were wedded to sterling for political and institutional reasons, the dollar was already within hailing distance of sterling as a currency of denomination for international bonds by the latter 1920s. Second, our evidence challenges the presumption that once international monetary leadership is lost, it is gone forever. Although sterling lost its leadership in the 1920s it recovered after 1933 and again ran neck and neck with the dollar at the end of the decade. Third, our findings challenge the presumption that there is room for only one dominant international currency due to strong network externalities and economies of scope. International debt markets in the 1920s and the 1930s were characterized by a bipolar currency system, not a unipolar one. This is true even if one takes into account the Commonwealth countries, which were heavily oriented towards sterling for very institutional and political reasons. Our results point to the development of US financial markets as the main factor that helped the US dollar overcome sterling's incumbency advantage. We find that financial deepening was the most important contributor to the increase in the share of the US dollar in global foreign public debt between 1918 and 1932. In the case of the UK, economic stagnation (declining relative economic size) was the most important factor accounting for sterling's declining share over the period. These findings have implications for the future of the international monetary system. They suggest that a shift from a unipolar dollar-based system to a multipolar system is not impossible; that it could occur sooner than often believed; and that financial deepening and market liquidity will be key determinants of the ability of currencies other than the dollar to strengthen their international currency status. They point to addressing financial market fragmentation and deepening financial integration in the euro area as important to the evolution of the euro's international profile in the years ahead; and to the opening up of the capital account, along with further exchange rate reform and the building up of liquid domestic financial markets, as of key importance to that of the Chinese yuan. The international status of a currency will rest on solid foundations, however, only if financial deepening in the issuing country is sustainable, and not if financial innovation and liberalization simply causes a boom that eventually goes bust. The impact of finance on international currency shares in global debt markets worked both ways during the interwar period. In particular, the collapse of the US banking system and subsequent financial retrenchment were the most important factor contributing to the decline in the share of the US dollar in global foreign public debt between 1932 and 1939. The yen's experience is another cautionary tale. Attempts by the Japanese authorities to develop the international role of their currency suffered from the bursting of Japan's equity and real estate bubbles in the late 1980s and the banking and economic crisis of the 1990s. This underscores that the compass guiding the pace and scope of financial sector reform should always point to the direction of medium-term sustainability. In turn, this highlights the important role that macro-prudential policies and tools will play in shaping the international status of currencies in the new millennium.