The Dollar versus the Euro as

International Reserve Currencies

Faculty Research Working Paper Series

Menzie D. Chinn
University of Wisconsin, Madison and NBER

Jeffrey A. Frankel
Harvard Kennedy School and NBER

Hiro Ito
Portland State University

March 2024
RWP23-005

Visit the HKS Faculty Research Working Paper Series at: https://ken.sc/faculty-research-working-paper-series

The views expressed in the HKS Faculty Research Working Paper Series are those of the author(s) and do not
necessarily reflect those of the John F. Kennedy School of Government or of Harvard University. Faculty Research
Working Papers have not undergone formal review and approval. Such papers are included in this series to elicit
feedback and to encourage debate on important public policy challenges. Copyright belongs to the author(s). Papers
may be downloaded for personal use only.

www.hks.harvard.edu
 The Dollar versus the Euro
as International Reserve Currencies

Menzie D. Chinn*
University of Wisconsin, Madison and NBER

Jeffrey A. Frankel**
Harvard University and NBER

Hiro Ito***
Portland State University

February 14, 2024

Abstract: We begin by examining determinants of aggregate foreign exchange reserve holdings by

central banks (size of issuing country’s economy and financial markets, ability of the currency to hold
value, and inertia). But understanding the determination of reserve holdings probably requires going
beyond the aggregate numbers, instead observing individual central bank behavior, including
characteristics of the holding country (bilateral trade with the issuing country, bilateral currency peg, and
proxies for bilateral exposure to sanctions), in addition to the characteristics of the reserve currency
issuer. On a currency-by-currency basis, US dollar holdings are somewhat well explained by several
issuer characteristics; but the other currencies are less successfully explained. It may be that the results
from currency-by-currency estimation are impaired by insufficient sample size. This consideration offers
a motivation for pooling the data across the major currencies and imposing the constraints that reserve
holdings are determined in the same way for each currency. In this setting, most economic determinants
enter with significance: economic size as measured by GDP, size of financial markets as measured by
foreign exchange turnover, bilateral currency peg, and bilateral trade share. However, geopolitical
variables (bilateral alliance, bilateral sanctions) usually do not enter with significance.
Keywords: dollar, euro, international currencies, reserve currencies, network externalities, foreign
exchange turnover
JEL number: F33
Acknowledgements: We thank participants at the 25th Anniversary ECB Symposium (Erasmus
University) for helpful comments.
───────────────────────────────────────────────────────────

Correspondence:
* [email protected]
** [email protected]
*** [email protected]

0
1. Introduction

In the early years of this century, international use of the dollar began a gradual relative
decline, particularly as measured by the currency shares in which central banks hold their
international reserves. Some suggested that another currency might overtake the dollar as the
leading international currency one day. But the theory of international currencies was based on
network externalities: everyone uses the dollar because everybody else uses the dollar. This
implied a lot of inertia supporting the incumbent number one currency.

The literature concluded that any rivals could catch up to the dollar only with a
substantial lag behind the fundamental determinants. Three fundamental determinants were:
size of home economy; size, depth, liquidity and openness of its financial markets; and ability of
the currency to hold its value, requiring the self-discipline to resist abusing exorbitant privilege
by excessive fiscal and monetary expansion. An historical precedent dominated: it was said that
the fundamental determinants underlying the dollar had caught up with those underlying pound
sterling by 1918 and yet international use of the dollar had not caught up with the pound until 30
years later.

From its creation in 1999, the euro had immediately become the number two currency.
Initially, it appeared to be gaining on the dollar. It satisfied the three key criteria almost as well
as the dollar did: the eurozone economy was almost as large as the US economy, European
financial markets were almost as big as US financial markets, and the euro country governments
had committed to fiscal and monetary discipline at a time when the US was reverting to large
budget deficits. 1

1
Chinn and Frankel (2007) even suggested that the euro might rival the dollar for the number one slot by 2022, if
either of two conditions were met: the UK and Sweden were to join the euro, thus expanding the size of the
eurozone’s economy and financial markets; and the US were to continue to exploit its exorbitant privilege as
reflected in trend depreciation of the dollar. Neither of these conditions were met subsequently: The UK and
Sweden, decided not to join; and the euro began a long trend depreciation against the dollar in July 2008. Papers on
the euro’s growing role included: Portes and Rey (1998), Frieden (2000), Posen (2008), and Goldberg (2010),
among many others.

1
 The prospects for the euro as leading international currency started to dim at the time of
the Global Financial Crisis. For one thing, the direction of flight of risk-averse investors in late
2008 was away from the euro and toward the safe haven of the US, notwithstanding that the sub-
prime mortgage crisis had originated in the US. Moreover, the crisis that erupted in Greece in
early 2010 called into question the enforceability of the European fiscal compact and seemingly
even the viability of the euro itself.

China’s renminbi became the new purported challenger to the dollar. After all, it
increasingly met at least two of the criteria: size of the home economy and ability to maintain its
value. The renminbi increased in value steadily from 2005 to 2014. Three decades of rapid
growth, 1980-2010, seemed to foreshadow that the Chinese economy would surpass the US
economy in size by 2021, even when the GDPs were compared at nominal exchange rates rather
than PPP rates. 2 Meanwhile, China resolved to internationalize its currency. Some forecast that
the renminbi would eclipse the dollar as number one international currency as early as 2020. 3

A lot has changed since the forecasts of euro or renminbi ascendancy were made.

1. The “new view” of Eichengreen (2010, 2011a) and Eichengreen and Flandreau (2009,
2012) argued that the literature had over-emphasized network externalities and the
impregnability of the incumbent lead currency. It concluded, first, that the dollar had in
fact caught up with the pound very quickly after World War I and, second, that a unipolar
currency system was not the only possible global equilibrium. Rather, the world might be
moving in a multi-polar direction.

2. A second interpretation was added to the notion of a key-currency country’s abuse of

exorbitant privilege. In the past, it was thought to be a matter of monetary and fiscal
profligacy, as reflected in a tendency for the currency in question to lose value, as
measured either by its inflation rate or its trend depreciation against other currencies. But
geopolitical developments involving the use of financial sanctions by the US (especially

2
Frankel (2014).
3
Subramanian (2011a, 2011b). See also Dobson and Masson (2009); Ito (2010); Park and Song (2010);
Eichengreen (2011b); Prasad and Ye (2012); Frankel (2012); and Zhang (2022).

2
 against Iran and Russia) led some countries to fear that the freedom to use their dollar
reserves could be curtailed in time of crisis. In reaction, Russia and China, in particular,
shifted some of their international reserves out of dollars.

3. The euro lost momentum as an international currency. The UK and Sweden, after
respective deliberative processes, decided not to join the euro. Indeed, the UK even
exited from the EU.

4. Next, the renminbi lost some of its aura as a rising star. In 2014, ten years of Chinese net
capital inflow turned to net capital outflow, as manifest in a peaking of both foreign
exchange reserves and the foreign exchange value of the currency. The Chinese
authorities resolved the conflicting goals of international currency status and the
insulation afforded by capital controls by opting for the latter, so that they would be better
able to slow capital outflows. Furthermore, the growth rate of the Chinese economy
slowed after 2010. It became increasingly clear by 2024 that the decades in which the
economic growth rate had averaged 10 percent had come to an end. The debate over the
date when China’s economy would surpass the US economy -- whether it would be
sooner versus later (evaluated at market exchange rates) -- suddenly had to accommodate
the possibility that the changeover might not happen at all. These factors all undermined
the perceived inevitability of the rising international role of the renminbi.

5. Gold, which had been demoted with the end of Bretton Woods in 1971, returned as a
relevant component of the international monetary system. Some central banks, especially
in Asia, resumed active buying and selling gold, as a means of diversifying their
international reserves out of dollars. 4

6. Important new data became available. First, a few holdout countries, most importantly
China, for the first time began to allow the IMF to include their central banks’ holdings in the
global totals for the currency composition of foreign exchange reserves (Prasad, 2019).

4
Arslanalp, Eichengreen, and Simpson-Bell. (2023).

3
 Subsequently, it also became possible to derive from public sources data on the currency
composition of individual central banks’ reserve holdings, as in Ito and McCauley (2020) and
Chinn, Ito and McCauley (2022).

In light of these six developments, it is a good time to update the investigation of the
empirical determinants of foreign exchange holdings.

In the second section of this paper, we elaborate on the developments of the dollar vs. the
euro as the key reserve currencies, as well as the Japanese yen, the British pound, and the
Chinese yuan, in recent years (since Chinn and Frankel, 2007).

In section 3, we investigate the determinants of individual central bank holdings of

reserve currencies, currency-by-currency, bringing to bear information on bilateral relationships.
This information includes geopolitical factors that have come to prominence in recent policy
debates, namely sanctions. In using individual central banks as the unit of interest, we can exploit
the cross-country variation in reserve holdings that was not possible with aggregate reserves.

Nonetheless, despite exploiting the cross-central bank information, the results are not as
informative as would be desired. The results from currency-by-currency estimation might be
impaired by insufficient sample size or insufficient variation in the data. Refer to Figure 1 for
visual evidence as to how little the reserve shares vary over time, as compared to the variation
across currencies. This consideration offers a motivation for pooling the data across the major
currencies and imposing the constraints that their reserve holdings are determined in the same
way across currencies.

In Section 4, we move to investigating currency holdings, imposing the same

specification for all reserve currencies, i.e., making the relationship for the USD reserve share
and US share of world GDP be the same (possibly nonlinear) relationship as for the JPY reserve
share-Japan share of world GDP. In this set of results, a lot of the identified relationships
obtaining for reserve shares are driven by the cross-reserve currency variation, rather than the
over-time variation.

2. Aggregate Reserves and the Aggregate Relationships

4
 At the aggregate level, the dominance of the dollar along various dimensions of an
international currency has remained in place. In writing this, it is perhaps useful to recapitulate
what constitutes an international currency.

There is a list of uses to which an international currency like the dollar or euro is put. Some
familiar functions include: a currency in which central banks and sovereign wealth funds hold
international reserves; an anchor currency to which smaller countries’ currencies can peg; a
currency to use in denominating or invoicing trade and financial transactions; and a vehicle
currency for foreign exchange trading.

A standard linguistic analogy can help explain (Kindleberger, 1967). Filipinos are unlikely to
speak Portuguese and Brazilians are unlikely to speak Tagalog. If a Brazilian wants to
communicate with a Filipino, they are likely to find it more convenient to do so via some third
language like English or Spanish. Similarly, if one of them wishes to do business with the other
country, they will find it more convenient to transact via some third currency like dollars or euros
(the vehicle currency), than to try to find someone who wants to take the other side of a peso-real
trade.

A two-by-three table lays out these international uses, as either private or public instances of
the classic three functions of money: 5

[Table 1 about here]

The first criterion is the most salient, partly because the data on central banks’ reserve
holdings are readily available, and the one we focus on in this paper. Figure 1 confirms the
dollar’s share of foreign exchange reserves has been declining gradually since the turn of the
century, resuming an earlier decline during 1978-1992.

[Figure 1 about here]

5
The schematic table began with Cohen (1971) and was adopted with slight modifications by Kenen (1983) and
Frankel (1992), among others.

5
 Each use of the currencies internationally is highly correlated with the other uses, both
causally and statistically (Aiyar et al., 2023). Causally, Gopinath and Stein (2018, 2021), for
example, show that international use of a currency to invoice trade (unit of account) is bi-
directionally related to use of the currency financially (store of value, particularly in the case of a
currency with safe-haven properties, like the dollar and yen). Statistically, a correlation is evident
as well. The ranking is similar by different measures: the dollar remains number one, not only by
the criterion of reserve holdings, but also by the criteria of denominating or invoicing trade 6, as
well as denomination of international debt and loans, foreign exchange turnover (BIS, 2022), and
global payments. According to an overall measure of international currency use computed at the
Federal Reserve, the dollar remains three times as important as the euro, and far more important
than the yen, pound, or renminbi. 7

Consider foreign exchange turnover, that is, the amount of all trading of foreign exchange
taking place within the home country. It is shown (on a scale of 0.0 to 2.0), in Figure 2.

[Figure 2 about here]

The correlation between reserves and foreign exchange turnover (redefined to be out of
1.00) is shown below:

[Figure 3 about here]

Here is a comparison of the dollar vs. the euro for 2023 for a set of indicators.

[Figure 4 about here]

6
Engel (2006), Goldberg and Tille (2008), Gopinath (2015), and Boz et al (2020).
7
Bertaut, et al, (2021), Figure 10. The measure is a weighted average of five criteria.

6
 Only in the case of global SWIFT payments, 8 does the euro come anywhere close to the
dollar’s share as of 2023: 32.6 % of payments versus 41.7%. The SWIFT rankings again put the
renminbi in fifth place, at a mere 2.3 %. Admittedly, recent growth in the renminbi (RMB) would
show up more strongly if data on the Cross-Border Interbank Payment System, which China
launched in 2015, or other non-SWIFT alternative payments systems could be included.

Given these correlations, in this study we focus on reserve holdings and their
determinants. Nearly two decades ago, two of the coauthors (Chinn and Frankel, 2007) estimated
pre-euro relationships between reserve holdings and a set of variables that we thought were
important determinants, basing the specification on the empirical literature on international
currencies.9

1. Patterns of output and trade. The currency of which the home country has a large share in
international output and trade has a big natural advantage. For some measures of
international currency use – how often a vehicle currency is used in the invoicing and
financing of international trade -- other aspects of the pattern of trade may also be relevant.
For reserve currency holdings, at the aggregate level, we relied on GDP as our measure of
size.

2. The country's financial markets. To attain international currency status, capital and money
markets in the home country must be not only open and free of controls, but also deep and
well-developed. It is surprisingly difficult to come up with a proxy for size, depth, or
development that is available for all the financial centers. As in our 2006 paper, we have
opted to use data on foreign exchange turnover in the respective financial centers: New York,
London, Frankfurt, Tokyo, etc. This measure differs from worldwide turnover of the

8
SWIFT (April, 2023). The Society for Worldwide International Financial Telecommunications is a messaging
system that accompanies inter-bank transactions. It provides the highest-frequency measure of international currency
use. Chau, Ilzetski, and Rogoff (2022) and Perez Saiz, Zhang and Iyer (2023).
9
Among the relevant references are Aliber (1966), Alogoskoufis and Portes (1992), Bergsten (1975), Black (1989),
Eichengreen and Frankel (1996), Eichengreen and Mathieson (2000), Frankel (1992, 1995), Kenen (1983), Krugman
(1984), Kindleberger (1981), Matsuyama, Kiyotaki and Matsui (1993),McKinnon (1969, 1979), Portes and Rey
(1998), Rey (2001), Swoboda (1969), Tavlas (1993), and Tavlas and Ozeki (1992).

7
 currencies (dollar, pound, euro, etc.), a variable that would be closer to a measure of
international currency use determined simultaneously with the international currency status
that we are trying to explain, more than a determinant.

It captures, for example, the pre-eminence of London, which has continued despite the small
role of the pound. The turnover measure has the virtue of reflecting all kinds of international
financial transactions (both long-term and short-term, banking and securities, bonds and
equities).

3. Confidence in the value of the currency. A necessary qualification for a candidate currency
is that its value not fluctuate erratically. We relied on three measures – the long-term
inflation differential vs. average for OECD countries, the appreciation of the currency in the
foreign exchange market, and the volatility of the exchange rate. All three criteria for
stability are measured over the preceding five years.

4. Network externalities. An international money, like domestic money, derives its value
because others are using it. It is a classic instance of network externalities. In this sense, the
current characteristics of a currency are of less importance than the path-dependent historical
equilibrium. There is a strong inertial bias in favor of using whatever currency has been the
international currency in the past. Hence, the lagged currency share enters. Another
implication is that the relationship between the extent of currency importance and the
fundamental determinants is nonlinear. In Chinn and Frankel (2007), the preferred
specification involved a logit transformation of the reserve shares, thus fitting the interval
0.00-1.00 and also allowing for a sort of tipping point in the middle.

How far off were our estimates of the determination of reserve holdings? It is important
to recognize that many of the variables changed values with revisions. That being said, a
straightforward application of the currently reported values of the variables -- such as GDP share,
inflation, exchange rate volatility, and foreign exchange turnover by location -- to the estimates
in column 3 of Table 8.6 (estimated 1973-99) reveals that previous estimates were not
qualitatively wrong. Using this specification:

8
 𝑧𝑧𝑖𝑖𝑖𝑖 = −0.648∗ + 2.768∗ 𝑦𝑦𝑖𝑖𝑖𝑖 − 2.639∗ 𝜋𝜋𝑖𝑖𝑖𝑖 − 0.981∗ 𝜎𝜎𝑖𝑖𝑖𝑖 + 0.446𝑡𝑡𝑡𝑡𝑖𝑖𝑖𝑖 + 0.919∗ 𝑧𝑧𝑖𝑖,𝑡𝑡−1 + 𝑢𝑢𝑖𝑖𝑖𝑖

𝑅𝑅𝑅𝑅𝑅𝑅
where 𝑧𝑧𝑖𝑖𝑖𝑖 ≡ �1−𝑅𝑅𝑅𝑅𝑅𝑅
𝑖𝑖𝑖𝑖
�
𝑖𝑖𝑖𝑖

we obtain the following estimates of USD and EUR shares.

[Figure 5 about here]

[Figure 6 about here]

These specifications give a big role to observed lagged reserve shares, as the
autoregressive coefficient is 0.92.

How well does this equation work for the post-euro period? There are several problems in
answering this question. The first is that the persistence in USD and EUR shares seems to be
much higher than it was for the USD and other shares, pre-euro. Logit transformations imply that
the unit root can be rejected at the 10% level for the USD and EUR. Estimating a panel while
imposing the same value on the autoregressive coefficient results in an estimate that is not easily
distinguished from unity.

The estimated equation over the 1999-2022 period, with USD, EUR, JPY, GBP and CNY,
is:

𝑧𝑧𝑖𝑖𝑖𝑖 = −0.017 − 0.071𝑦𝑦𝑖𝑖𝑖𝑖 + 8.047∗ 𝜋𝜋𝑖𝑖𝑖𝑖 + 2.837𝜎𝜎𝑖𝑖𝑖𝑖 − 0.319∗ 𝑡𝑡𝑡𝑡𝑖𝑖𝑖𝑖 + 0.968∗ 𝑧𝑧𝑖𝑖,𝑡𝑡−1 + 𝑢𝑢𝑖𝑖𝑖𝑖

Note that this is an unbalanced panel, with only 5 observations for the RMB. Excluding China
does not change the estimates appreciably.

In these estimates, reserves have over the last two decades not behaved in a way
consistent with earlier correlations, whether or not the level of reserves is transformed by the
logit function. GDP share, inflation, exchange rate volatility and turnover have unanticipated
signs.

9
 Some of the puzzling results disappear if all the variables are first differenced. Then the
proportion of variance explained (omitting lagged dependent variable) is about 12%, and the
coefficient on income becomes significant. Higher GDP of the country issuing the reserve
currency is associated with a higher share or reserves.

The last twenty years have been marked by momentous events which the previous study
could not incorporate, with its sample ending in 2003. These include the 2007-08 global financial
crisis, the Euro area crisis, the Covid-19 pandemic, and (at the end of the sample) Russia’s
expanded invasion of Ukraine. 10

Especially relevant is confidence in the value of the currency, which was gauged in Chinn
and Frankel by inflation and exchange rate depreciation. Average rates – as well as variation -- of
inflation were generally much lower during the 1999-2023 period than in the earlier sample
(1973-98), thus perhaps explaining the loss of the negative coefficient.

However, the Euro crisis highlighted fiscal and banking unions as determinants of
credibility. Instead of attempting to augment the specification with an ad hoc measure of a
sovereign debt crisis (which would end up being essentially a dummy variable), it makes sense to
conclude that we have hit the limits of what aggregate foreign reserves data can tell us.

3. Individual Central Bank Behavior, By Reserve Currency

We now turn to explaining how individual central banks determine their holdings of
foreign exchange reserves. While there have been several papers that investigate this question,
they have focused on limited aspects. For instance, Arslanalp et al. (2022) examine a set of
determinants that are separate from those that were considered in Chinn and Frankel (2007), and
that focus more on factors that are specific to the holder of reserves (pegs, shares of trade).
Goldberg and Hannaoui (2023) bring into the analysis geopolitical and return variables, but focus
only on holdings of US dollars.

10
This is a separate issue from why the prediction of the euro overtaking the dollar did not occur. The requirements
for this outcome included rapid dollar depreciation over a long span of time, as well the UK joining EMU. In the
absence of this condition, Chinn and Frankel (2007) predicted continued dollar dominance.

10
 For this analysis, we rely on the Ito-McCauley data set, described in Ito and McCauley
(2020). They draw on data from annual reports, financial statements, and other relevant materials
of central banks across the world and collect data on the currency composition of FX reserves of
individual countries. They calculate the currency shares of international reserves excluding gold
and SDRs. For the Latin America central banks, they relied on data provided by the Latin
American Reserve Fund (FLAR). The dataset encompasses 58 countries: 13 advanced
economies; 45 emerging and developing economies, as defined by the IMF. By region, 10 Asian-
Pacific; 12 African and Middle Eastern; 6 Western European; 17 Eastern European and Central
Asian; and 12 Western Hemisphere. The holdings by the issuers of the key reserve currencies
themselves, US, Euro, Japan, UK and China, are omitted from the analysis.

Our dependent variable is foreign exchange reserve shares. Since there are few instances
of any particular central bank’s currency predicted reserves holding at 100% or below 0%, we
focus on results based on this variable, allowing easier interpretation of the coefficients. But, as
alternatives, we consider two nonlinear specifications. The first is the logit transformation of the
shares. The second is a recursive definition of shares, such that the USD share is expressed out of
total foreign exchange reserves, the EUR share is divided by total foreign exchange reserves
subtracting USD reserves; the JPY share is divided by total foreign exchange reserves minus
both USD and EUR reserves; and so on. The recursive approach, an innovation of this paper, has
the advantage that it uses data on the holdings of all currencies, while yet obeying the constraint
that their shares must add to 1.0.

We estimate the following specification, which includes the reserve currency issuer i
variables described in Section 2, augmented with country j specific variables related to the
country of central bank that is holding the reserves.

𝑧𝑧𝑖𝑖𝑖𝑖𝑖𝑖 = 𝛽𝛽0 + 𝛽𝛽1 𝑦𝑦𝑖𝑖𝑖𝑖 + 𝛽𝛽2 𝜋𝜋𝑖𝑖𝑖𝑖 + 𝛽𝛽3 𝜎𝜎𝑖𝑖𝑖𝑖 + 𝛽𝛽4 𝑡𝑡𝑡𝑡𝑖𝑖𝑖𝑖 + 𝛽𝛽5 𝑧𝑧𝑖𝑖𝑖𝑖,𝑡𝑡−1 + 𝛾𝛾1 𝑡𝑡𝑡𝑡𝑡𝑡𝑡𝑡𝑡𝑡𝑖𝑖𝑖𝑖𝑖𝑖 + 𝛾𝛾2 𝑝𝑝𝑝𝑝𝑝𝑝𝑖𝑖𝑖𝑖,𝑡𝑡 + 𝑢𝑢𝑖𝑖𝑖𝑖

Following Ito and McCauley (2020) and Chinn, Ito, and McCauley (2022), this
specification includes country j’s trade share with country i as well as a dummy variable
indicating whether country j’s currency is pegged to country i.

Finally, we want to assess geopolitical factors which have been recently highlighted,
motivated by increased fear of sanctions. One would hypothesize that, the more a country is at

11
odds with the United States or Europe in geopolitical terms, the more vulnerable it is to sanctions
from these places and thus the less dollars and euros it would choose to hold. Goldberg and
Hannaoui (2023), Mosler and Potrafke (2020) and Perez-Salz, Zhang and Iyer (2023) use
geopolitical proximity as proxied by the frequency with which the country votes in agreement
with the United States in United Nations General Assembly resolutions and we use this data in
our analysis. Eichengreen, Mehl, and Chitu (2017) and Arslanalp, Eichengreen, and Simpson-
Bell (2022), use a dummy variable reflecting whether a country has a defense pact with the
United States. The case of Russia notwithstanding, the general finding in these studies is little
significant positive effect of these measures of geopolitical proximity on dollar holdings.

The literature on vulnerability to sanctions features an exception for gold. Economists

had long believed that central bank holdings of gold were an anachronism. Although monetary
authorities in many countries still held some gold, they did not treat it as an active part of their
international reserves. More recently, central banks, especially in Asia, have returned to actively
buying (and selling) gold. Arslanalp, Eichengreen, and Simpson-Bell 2023) and Ferranti (2023)
find that, in recent years, countries that faced a higher risk of US sanctions increased the share of
gold in their international reserves more than countries facing a lower risk of US sanctions.

𝑧𝑧𝑖𝑖𝑖𝑖𝑖𝑖 = 𝛽𝛽0 + 𝛽𝛽1 𝑦𝑦𝑖𝑖𝑖𝑖 + 𝛽𝛽2 𝜋𝜋𝑖𝑖𝑖𝑖 + 𝛽𝛽3 𝜎𝜎𝑖𝑖𝑖𝑖 + 𝛽𝛽4 𝑡𝑡𝑡𝑡𝑖𝑖𝑖𝑖 + 𝛽𝛽5 𝑧𝑧𝑖𝑖𝑖𝑖,𝑡𝑡−1 + 𝛾𝛾1 𝑡𝑡𝑡𝑡𝑡𝑡𝑡𝑡𝑡𝑡𝑖𝑖𝑖𝑖𝑖𝑖 + 𝛾𝛾2 𝑝𝑝𝑝𝑝𝑝𝑝𝑖𝑖𝑖𝑖,𝑡𝑡 + 𝜃𝜃1 𝑣𝑣𝑣𝑣𝑣𝑣𝑣𝑣𝑣𝑣𝑣𝑣𝑖𝑖𝑖𝑖𝑖𝑖
+ 𝜃𝜃2 𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑖𝑖𝑖𝑖,𝑡𝑡 + 𝑢𝑢𝑖𝑖𝑖𝑖

where voting is a dummy variable which indicates whether country j voting in the UN is similar
to that of reserve currency issuer i and sanctions is a dummy variable which indicates whether
country j has had sanctions imposed by country i . The sanctions could pertain to either trade or
financial transactions. 11

Tables 2.1 through 2.5 present the estimates from these specifications, on the basis of
reserve-currency-by-reserve-currency: USD, EUR, GBP, JPY, and CNY.

11
We have used an “alliance” variable as an alternative geopolitical distance proxy. These results are not as strong as
those obtain using the voting variable.

12
 [Table 2.1-2.5 about here]

For the USD (Table 2.1), most of the variables that were included in Chinn and Frankel
(2007), as explained in Section 2, enter in with expected sign -- in contrast to the aggregate
results. The exception is the GDP share; it does not show up as statistically significant.
Exchange rate volatility reduces the attractiveness of the reserve currencies: each one percentage
point increase in nominal trade-weighted exchange rate volatility reduces holdings by an
estimated 3.3-3.6 percentage points. Inflation differentials and foreign exchange turnover
location enter in as predicted, but not statistically significantly. Finally, as in the aggregate data,
there is a great deal of inertia. The coefficient on lagged reserves is 0.89, implying that
deviations have a six-year half-life.

Consistent with the literature, the share of trade with the reserve currency issuer enters
significantly. A one percentage point increase in trade share with the US raises the USD share by
an estimated 0.07-0.08 ppts. A peg to the USD raises the USD share by an estimated 0.04 ppts.
Both of these coefficients are statistically significant and are robustly so across specifications.
The adjusted R2 is fairly high, at 0.89, with 935 observations, over 56 central banks.

One geopolitical variable is included in column (2): UN voting distance. This variable
enters in significantly, but with a positive sign. Being less aligned with the US seems to result in
higher dollar holdings. This is the same (surprising) result as in Goldberg and Hannoui (2023) 12.
A dummy variable for military alliance, which is time invariant, does not enter significantly. In
columns (3-5), sanctions imposed on the country by the US do not have significant effects,
whether they be, trade sanction, financial sanction, or any sanction. 13

The Euro is examined in Table 2.2. The basic specification results shown in Column (1)
yield coefficients with signs as anticipated, excepting foreign exchange turnover, the coefficient

12
Their explanation seems to be that the set of countries that vote at odds with the US in the UN tend to have reserve
holdings in total that are worryingly low from a precautionary viewpoint, and thus can’t afford to diversify out of the
dollar, which is the most liquid of the reserve currencies
13
Sanctions of any sort do have a statistically significant and negative effect when using the logit specification.

13
of which is not statistically significant. The proportion of variation explained is even higher, with
adjusted R2 at 0.94, for 852 observations over 52 central banks.

A one percentage point increase in GDP share results in a 0.24 ppt increase in EUR
reserve share. An increase in inflation by one ppt reduces holding share by 5.7-5.9 ppts. An
increase in trade share with the Euro area increases EUR reserve share by 0.07 ppts (the same
value as in the USD regression). In contrast to the USD result, a EUR peg has no statistically
significant effect, probably because there are so few observations of countries pegged to the
euro. The degree of persistence is about the same as for the USD, around 0.92. These coefficient
estimates are largely insensitive to inclusion of geopolitical variables.

Greater voting distance from the European Union (column 2) has a negative and
statistically significant impact on EUR holdings. This is the effect hypothesized. Meanwhile,
European Union sanctions on country j have no significant impact.

The results for the British pound (Table 2.3) show some similarities to those for the Euro.
The degree of persistence is high, but not as high as for the USD or EUR. GDP share is
statistically significant. A one percentage point increase in the UK GDP share results in a 0.7 ppt
increase in the GBP share of holdings. The trade share with the UK is significant in the baseline
specification (column 1) at 0.04. However, this result is not robust to the addition of covariates.

The inflation differential has the wrong sign, with statistical significance. Foreign
exchange turnover share has no significant impact. Finally, all geopolitical variable coefficients
are statistically insignificant.

Japanese yen results are reported in Table 2.4. The sample size drops to 465, with only 33
central banks included. The adjusted R2 is at 0.74. In column (1), only lagged JPY reserves,
exchange rate volatility and foreign exchange turnover location enter significantly. The last two
coefficients also have the wrong sign.

Geopolitical distance as measured by voting patterns is associated with decreased yen

holdings. This effect is statistically significant. So too is the impact of sanctions, with the
imposition of sanctions associated with a 0.005 ppt reduction in yen holding share. On the other
hand, the coefficient is only significant for trade sanctions.

14
 Finally, the Chinese yuan (Renminbi) results are reported in Table 2.5. The data for CNY
holdings typically pertains to only the last few years and applies to only 20 countries. The
sample size is then only 231, only a bit more than a quarter of the sample size for USD holdings.
The bottom line is that CNY holdings are not explained.

Geopolitical distance is negatively associated with CNY holdings, but not significantly
so. One seemingly anomalous result is the finding that imposition of financial sanctions results in
an increase in CNY holdings. This result is, however, not robust to alternative specifications (i.e.,
logit).

Other than the indicated exceptions, the results are largely invariant to using alternative
measures of holdings. Regressions using logit shares, or shares defined recursively, are reported
in the Appendix.

To recap, while panel regressions across central banks are able to provide some insights
not obtainable in the aggregate (total holdings), there is a limit to what relationships can be
identified when running the OLS equations for each reserve currency separately. We suspect that
the currency-by-currency results are handicapped by insufficient data or insufficient variation in
the data. Hence, we turn to examining these reserve currencies pooled simultaneously.

4. Individual Central Bank Behavior: Pooled Analysis

In this section, we adduce the results from pooled regressions where we incorporate
variation across currencies to explain central bank behavior, rather than just variation across
time. This is our preferred set-up for estimation, because variation across currencies is where the
action is.

Table 3 reports the results for the same specifications used in Table 2, except that the data
is pooled, each currency with its own intercept. Column (1) shows the estimate for the baseline
specification. All six base-case coefficients are correctly signed. Four of them are statistically
significant: GDP share, exchange rate volatility, anchor currency dummy, and trade shares. The
inflation differential has the expected negative effect on the demand for the currency, but the
estimate is not statistically significant, suggesting that exchange rate volatility may better capture
the stability of a currency.

15
 A one percentage point increase in GDP share induces an estimated 0.07 ppt increase in
currency holding shares, while a one percentage point increase in trade share increases reserve
holding share by 0.05 ppts. Finally, in line with the other results, a bilateral currency peg raises
reserve currency shares by 0.04 ppts. (This is relevant essentially only for USD and EUR.)

None of the coefficients on vulnerability to sanctions (distance in UN voting behavior)

nor actual sanctions (whether trade sanctions, financial sanctions or both) is statistically
significant.

The coefficient on the lagged dependent variable is estimated at 0.91, implying a speed of
adjustment of 0.09 per year and a half-life of about 7 years.

Next, the results in Table 4 pertain to the dependent variable transformed using the logit
transformation. Recall the two motivations for this functional form: it corresponds to a
dependent variable that is restricted to the interval (0.00-1.00) and its non-linearity allows for a
sort of tipping point. GDP share retains its significance, as does the currency peg dummy and
trade share.

Goldberg and Hannoui (2023) point out that (i) Switzerland holds a high fraction of its
reserves in euros, because it trades so much with euro countries, which is in turn largely
attributable to geographic proximity to the eurozone; and (ii) Switzerland’s weight in the
conventional results increases over time, because the country has run such large balance of
payments surpluses. This would produce an apparent shift from dollars to euros in the aggregate
numbers, even if no individual central bank shifted from dollars to euros. Further, as noted, (iii)
the composition of Russia’s holdings has shifted away from dollars. Goldberg and Hannoui find
that these two countries, Switzerland and Russia, by themselves explain the downward trend of
the dollars share in the aggregate statistics. We re-estimated our equation while excluding
Switzerland or Russia. We find little effect on our estimates.

One geopolitical variable is included in column (2): UN voting distance. This variable
enters in significantly:, being farther in proximity results in higher holdings. This is the same
result as in Goldberg and Hannoui (2023).

As with the simple shares regression, none of the sanction-related variables enters the
logit equation with statistical significance.

16
 Finally, in Table 5, we report the results when the recursively defined foreign exchange
shares are used as the dependent variable (and a different constant is allowed for each recursively
defined share). The results are qualitatively the same. However, the degree of persistence is
lower, with the coefficient on the lagged dependent variable at about 0.78. Moreover, the
proportion of variation explained for this dependent variable is also lower, at 0.9.

In addition, we evaluated whether the shares regression results are robust to the use of
tobit, where we take into account the fact that shares below 0% or above 100% shares do not
make much sense. The regressions using logit variables also take this aspect into account, but
tobit facilitates the comparison of coefficient estimates. We find that the magnitude of
coefficients, as well as instances of statistical significance, do not change much with the
application of tobit.

5. Conclusion

Some conclusions, in brief:

• Relative GDP is important.

• Along with the high estimated level of persistence, this helps explain why the dollar
retains dominance as reserve currency.
• Of the variables that measure the ability of a currency to hold its value compared to other
currencies, exchange rate volatility is generally of the right sign (negative) and
significant, while the inflation differential is generally of the right sign (also negative),
but insignificant.
• The hypothesized effect of financial market size is important, because it could help
explain why the dollar is still well ahead of the euro and why the renminbi still lags
behind the pound and the others. But we find little evidence of its effect, as proxied by
the amount of foreign exchange trading in the home country.
• The euro is held in large part due to bilateral trade (which follows from bilateral
geographic distance).
• Bilateral geopolitical proximity (congruence in UN voting) induces higher reserve
currency holdings in the cases of the euro and pound. But, apparently, not the dollar.

17
• Otherwise, the dummy variable for sanctions and other geopolitical variables do not show
up as significantly important…yet.

18
References

Aizenman, Joshua, Yin-Wong Cheung, and Xingwang Qian. 2020. The Currency Composition of
International Reserves, Demand for International Reserves, and Global Safe Assets. Journal of
International Money and Finance, 102, 102-120.
Alogoskoufis, George, and Richard Portes. 1992. European Monetary Union and International
Currencies in a Tripolar World. In Establishing a Central Bank: Issues in Europe and Lessons
from the US, ed. Matthew Canzoneri, Vittorio Grilli, and Paul Masson, 273–302. Cambridge:
Cambridge University Press.
Arslanalp, Serkan, Barry Eichengreen, and Chima Simpson-Bell. 2022. "The stealth erosion of
dollar dominance and the rise of nontraditional reserve currencies." Journal of International
Economics 138: 103656.
Arslanalp, Serkan, Barry Eichengreen, and Chima Simpson-Bell. 2023. Gold as International
Reserves: A Barbarous Relic No More? Working Paper WP/23/14. Washington: International
Monetary Fund.
Aysan, Ahmet Faruk, Ender Demir, Giray Gozgor, and Chi Keung Marco Lau. 2019. Effects of
the Geopolitical Risks on Bitcoin Returns and Volatility. Research in International Business and
Finance 47: 511–18.
Bapat, Navin A., and T. Clifton Morgan. 2009. "Multilateral Versus Unilateral Sanctions
Reconsidered: A Test Using New Data." International Studies Quarterly 53, no.4: 1075-1094.
Bertaut, Carol, Bastian von Beschwitz, and Stephanie Curcuru. 2021. The International Role of
the US Dollar. Fednotes, October 6. Washington: Federal Reserve System. Available at
https://www.federalreserve.gov/econres/notes/feds-notes/the-international-role-of-the-u-s-dollar-
20211006.html.
BIS (Bank for International Settlements). 2022. Triennial Central Bank Survey of Foreign
Exchange. December. Basel. Available at http://www.bis.org/statistics/rpfx22.htm..
Boz, Emine, Camila Casas, Georgios Georgiadis, Gita Gopinath, Helena Le Mezo, Arnaud Mehl,
and Tra Nguyen. 2020. Patterns in Invoicing Currency in Global Trade. Washington:
International Monetary Fund.
Chau, Vu, Ethan Ilzetzki, and Kenneth Rogoff. 2022. Zipf's Law for International Currencies.
March. Cambridge, MA: Harvard University.
Chinn, Menzie, and Jeffrey Frankel, 2007, “Will the Euro Eventually Surpass the Dollar as
Leading International Reserve Currency?”, in G7 Current Account Imbalances: Sustainability
and Adjustment, edited by Richard Clarida (University of Chicago Press). NBER WP No.
11510, 2005.
Chinn, Menzie D., Hiro Ito, and Robert N. McCauley. 2022. "Do central banks rebalance their
currency shares?." Journal of International Money and Finance 122: 102557.
Chinn, Menzie D., Hiro Ito, and Robert N. McCauley. 2022. Do Central Banks Rebalance Their
Currency Shares? Journal of International Money and Finance 122.

19
Chitu, Livia, Barry Eichengreen, and Arnaud Mehl. 2014. When Did the Dollar Overtake
Sterling as the Leading International Currency? Evidence from the Bond Markets. Working
Paper 1433. Frankfurt: European Central Bank. Available at
https://www.ecb.europa.eu/pub/pdf/scpwps/ecbwp1433.pdf.
Cohen, Benjamin J. 1971. The Future of Sterling as an International Currency. London:
Macmillan.
Dobson, Wendy, and Paul Masson. 2009. Will the Renminbi Become a World Currency? China
Economic Review 20, no. 1: 124–35.
ECB. 2023. The international role of the euro.
Dooley, Michael, J. Saul Lizondo, and Donald Mathieson. 1989. The Currency Composition of
Foreign Exchange Reserves. IMF Staff Papers 36: 385–434. Washington: International Monetary
Fund.
Eichengreen, Barry. 2010. Managing a Multiple Reserve Currency World. In The Future Global
Reserve System: An Asian Perspective A Project of the Asian Development Bank, Manila, ed.
Jeffrey Sachs, Masahiro Kawai, Jong-Wha Lee, and Wing Thye Woo: Chapter 4. {The project is
at https://aric.adb.org/grs/about.php.}
Eichengreen, Barry. 2011a. Exorbitant Privilege: The Rise and Fall of the Dollar and the Future
of the International Monetary System. Oxford: Oxford University Press.
Eichengreen, Barry. 2011b. The Renminbi as an International Currency. Journal of Policy
Modeling 33, no. 5: 723–30.
Eichengreen, Barry. 2021. Will Central Bank Digital Currencies Doom Dollar
Dominance? Project Syndicate, August 9.
Eichengreen, Barry, and Marc Flandreau. 2009. The Rise and Fall of the Dollar (or When Did the
Dollar Replace Sterling as the Leading Reserve Currency)? European Review of Economic
History 13, no. 3: 377–411.
Eichengreen, Barry, and Marc Flandreau. 2012. The Federal Reserve, the Bank of England and
the Rise of the Dollar as an International Currency, 1914–39. Open Economies Review 23, no. 1:
57–87.
Eichengreen, Barry, and Jeffrey Frankel. 1996. The SDR, Reserve Currencies, and the Future of
the International Monetary System. In The Future of the SDR in Light of Changes in the
International Financial System, ed. Michael Mussa, James Boughton, and Peter Isard.
Washington: International Monetary Fund: 337-378.
Eichengreen, Barry, and Donald Mathieson. 2001. The Currency Composition of Foreign
Exchange Reserves: Retrospect and Prospect. In The Impact of EMU on Europe and the
Developing Countries, ed. Charles Wyplosz: 269-293. Oxford: Oxford University Press.
Eichengreen, B., Mehl, A. and Chiţu, L., 2019. Mars or Mercury? The geopolitics of
international currency choice. Economic Policy, 34(98), pp.315-363.

20
Eichengreen, Barry, Arnaud Mehl, and Livia Chitu. 2017. How Global Currencies Work.
Princeton, NJ: Princeton University Press.
Eichengreen, Barry, Livia Chiţu, and Arnaud Mehl. 2016. "Stability or upheaval? The currency
composition of international reserves in the long run." IMF Economic Review 64: 354-380.
Engel, Charlers. 2006. "Equivalence results for optimal pass-through, optimal indexing to
exchange rates, and optimal choice of currency for export pricing." Journal of the European
Economic Association 4.6: 1249-1260
Farhi, Emmanuel, and Matteo Maggiori. 2018. A Model of the International Monetary
System. Quarterly Journal of Economics 133, no. 1: 295–355.
Ferranti, Matthew. 2023. Hedging Sanctions Risk: Cryptocurrency in Central Bank Reserves.
Chapter 2 of PhD thesis, Department of Economics, Harvard University, Cambridge, MA.
Frankel, Jeffrey. 2012. Internationalization of the RMB and Historical Precedents. Journal of
Economic Integration 27, no. 3: 329–65.

Frankel, Jeffrey, 2015, "China is Not Yet Number One," Frontiers of Economics in China, 10, 1:
1-6.
Frankel, Jeffrey. 1992. On the Dollar. In The New Palgrave Dictionary of Money and Finance.
London: Macmillan Press Reference Books.
Frankel, Jeffrey. 1995. Still the Lingua Franca: The Exaggerated Death of the Dollar. Foreign
Affairs 74, no. 4: 9–16.
Frieden, Jeffry. 2000. The Political Economy of the Euro as an International Currency. In The
Euro as a Stabilizer in the International Economic System, ed. Robert Mundell and Armand
Clesse. Boston: Kluwer Academic Publishers: 203-213.
Goldberg, Linda. 2010. Is the International Role of the Dollar Changing? Current Issues in
Economics and Finance 16, no. 1.
Goldberg, Linda. 2010. Is the International Role of the Dollar Changing? Current Issues in
Economics and Finance 16, no. 1.
Goldberg, Linda S. and Oliver Hannaoui, 2023. “Drivers of Dollar Share in Official Foreign
Exchange Reserves,” mimeo (October).
Goldberg, Linda S., and Cédric Tille. 2008, Vehicle Currency Use in International Trade. Journal
of International Economics 76, no. 2: 177–92.
Gopinath, Gita. 2015. The International Price System. Jackson Hole Symposium, vol. 27.
Federal Reserve Bank of Kansas City.
Gopinath, Gita, and Jeremy C. Stein. 2018. Trade Invoicing, Bank Funding, and Central Bank
Reserve Holdings. American Economic Review Papers and Proceedings 108: 542–46.
Gopinath, Gita, and Jeremy C. Stein. 2021. Banking, Trade, and the Making of a Dominant
Currency. Quarterly Journal of Economics 136, no. 2: 783–830.

21
Grubel, Herbert G. 1963. World Monetary Reform: Plans and Issues. Stanford, CA: Stanford
University Press.
Iancu, Alina, Lucine Lusinyan, Yiqun Wu, Andrea Gamba, Sakai Ando, Gareth Anderson, Neil
Meads, Ethan Boswell, and Shushanik Hakobyan. 2022, Reserve Currencies in an Evolving
International Monetary System. Open Economies Review, 33(5), 879-915.
Ito, Takatoshi. 2010, China as Number One: How about the Renminbi? Asian Economic Policy
Review 5, no. 2: 249–76.
Ito, Hiro, and Robert N. McCauley. 2020. Currency Composition of Foreign Exchange
Reserves. Journal of International Money and Finance 102: 102104.
Kenen, Peter. 1983. The Role of the Dollar as an International Currency. Occasional Paper No.
13. New York: Group of Thirty.
Kenen, Peter. 1987. Changing Views about the Role of the SDR and Implications for Its
Attributes. In The International Monetary System and Its Reform, Part II, ed. Sidney Dell, 373–
85. Amsterdam: North Holland.
Kindleberger, Charles. 1967. The Politics of International Money and World Language. Essays
in International Finance, No. 61. Princeton, NJ: Princeton University Press.
Kindleberger, Charles. 1981. International Money. London: George Allen & Unwin.
Kindleberger, Charles. 1995. Is the Dollar Going the Way of Sterling, the Guilder, the Ducat, and
the Bezant? The International Economy, no. 3: 609–11. Reprinted in Kindleberger. 1999. Essays
in History: Financial, Economic, Personal. Ann Arbor: University of Michigan Press.
Krugman, Paul. 1984. The International Role of the Dollar: Theory and Prospect. In Exchange
Rate Theory and Practice, ed. John Bilson and Richard Marston, 261–78. Chicago: University of
Chicago Press.
Lindert, Peter. 1969. Key Currencies and Gold: 1900-1913. Princeton Studies in International
Finance 24: 16–22.
Matsuyama, Kiminori, Nobuhiro Kiyotaki, and Akihiko Matsui. 1993. Toward a Theory of
International Currency. Review of Economic Studies 60 (April): 283–07.
Mosler, Martin, and Niklas Potrafke. 2020. International Political Alignment During the Trump
Presidency: Voting at the UN General Assembly. International Interactions 46, no. 3: 481–97.
Obstfeld, Maurice, and Haonan Zhou. 2022. The Global Dollar Cycle. Brookings Papers on
Economic Activity (Fall). Washington: Brookings Institution.
Park, Yung Chul, and Chi-Young Song. 2010. RMB Internationalization: Prospects and
Implications for Economic Integration in East Asia. Asian Economic Papers 10, no. 3.
Portes, Richard, and Hélène Rey. 1998. The Emergence of the Euro as an International Currency.
Economic Policy 13, no. 26: 306–43.
Park, Yung Chul, and Chi-Young Song. 2010. RMB Internationalization: Prospects and
Implications for Economic Integration in East Asia. Asian Economic Papers 10, no. 3.

22
Perez Saiz, Hector, Longmei Zhang and Roshan Iyer, 2023, “Currency Usage for Cross Border
Payments,” IMF Working Paper No. 2023/072.
Posen, Adam S. 2008. Why the Euro Will Not Rival the Dollar. International Finance 11, no. 1:
75–100.
Prasad, Eswar. 2019. Has the Dollar Lost Ground as the Dominant International Currency?
Washington: Brookings Institution. Available at https://www.brookings.edu/wp-
content/uploads/2019/09/DollarInGlobalFinance.final_.9.20.pdf.
Prasad, Eswar, and Le (Sandy) Ye. 2012. The Renminbi’s Role in the Global Monetary System.
Washington: Brookings Institution.
Rey, Hélène. 2001. International Trade and Currency Exchange. Review of Economic Studies 68,
no. 2: 443–64.
Subramanian, Arvind. 2011a. Eclipse: Living in the Shadow of China’s Economic Dominance
Washington: Peterson Institute for International Economics.
Subramanian, Arvind. 2011b. Renminbi Rules: The Conditional Imminence of the Reserve
Currency Transition. Working Paper Series 11-14. Washington: Peterson Institute for
International Economics.
SWIFT. RMB Tracker: Monthly Reporting and Statistics on Renminbi (RMB) Progress towards
Becoming an International Currency. Available at http://www.swift.com/our-
solutions/compliance-and-shared-services/business-intelligence/renminbi/rmb-tracker/rmb-
tracker-document-centre (accessed April 2023).
Takagi, Shinji. 2011/ Internationalizing the Yen, 1984–2003: Unfinished Agenda or Mission
Impossible? In Asia and China in the Global Economy, ed. Y.W. Cheung and G. Ma, 219-244.
Singapore: World Scientific Publishing.
Tavlas, George. 1993. The Deutsche Mark as an International Currency. In International
Finance: Contemporary Issues, ed. Dilip Das, 566–79. London: Routledge.
Tavlas, George, and Yuzuru Ozeki. 1992. The Internationalization of Currencies: An Appraisal
of the Japanese Yen. IMF Occasional Paper 90. Washington: International Monetary Fund.
Voeten, Erik, Anton Strezhnev, Michael Bailey. 2009. "United Nations General Assembly Voting
Data."
Zhang, Longmei. 2023. Capital Account Liberalization and China’s Financial Integration. M-
RCBG Associate Working Paper Series. Cambridge, MA: Harvard University.

23
 Data Appendix

Variable Description
Ratio of GDP to total Ratio of GDP of five major currency issuers in USD (converted
World GDP at official exchange rates) to GDP of world aggregate. Sources:
IMF, International Financial Statistics. Euro area, world GDP
data from IMF, World Economic Outlook.
Inflation differentials Calculated as log difference of monthly CPI of each major
currency issuer, averaged with moving 60-month windows,
subtracted by the log difference of monthly CPI of industrialized
countries. Source: IMF, International Financial Statistics.
Rate of appreciation Calculated as the 60-month moving average of the log first
difference of the nominal effective exchange rate (NEER) of
each major currency issuer. Source: Bank for International
Settlements (BIS).
Exchange rate Calculated as the standard deviation of the log first difference of
volatility the NEER of each major currency issuer over moving 60-month
windows. Source: BIS.
FX turnover ratio Is daily turnover (in billions of dollars) divided by the total of
turnover of the 5 major currency issuers. The data are available
for 1998, 2001, and every three years. Observations in-between
survey years log-linearly interpolated. Source: BIS.
Stock market liquidity Stock market capitalization (SMKC) as a share of GDP of each
of the five major currency issuers; stock market total value
(SMTV) as a share of GDP; stock market turnover (SMTO) ratio
of domestic shares (%).
Military alliance It is the dummy variable that takes a value of one for a country
signs treaty with relevant country issuer country. Source:
Voeten, Streszhnev, Bailey (2009).
Sanction The dummy taking a value 1 if there are any kind of financial
sanctions between sender i and target j, 0 otherwise. The data
source is the Global Sanction Data Base (GSDB).
financial sanction The dummy taking a value 1 if there are financial sanctions
between sender i and target j, 0 otherwise. GSDB
trade sanction The dummy taking a value 1 if there are trade sanctions between
sender i and target j, 0 otherwise. GSDB
Dependent Variable
Share of gold in total It is the ratio of gold to total foreign exchange reserves, i.e.,
international reserves 𝑔𝑔𝑔𝑔𝑔𝑔𝑔𝑔 ℎ𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜
.
𝑇𝑇𝑇𝑇𝑇𝑇𝑇𝑇𝑇𝑇 𝑟𝑟𝑟𝑟𝑟𝑟𝑟𝑟𝑟𝑟𝑟𝑟𝑟𝑟𝑟𝑟

Share of USD in FX It is the share of USD in FX reserves (RVUSD; 0 < RVUSD < 1).

24
 reserves The data source is Ito and McCauley (2020) and Chinn, Ito, and
McCauley (2021).
Share of EUR in FX It is the share of EUR in FX reserves that excludes RVUSD,
reserves calculated as:
𝑅𝑅𝑅𝑅𝐸𝐸𝐸𝐸𝐸𝐸
.
�1−𝑅𝑅𝑅𝑅𝐸𝐸𝐸𝐸𝐸𝐸 �

Share of GBP in FX It is the share of GBP in FX reserves that excludes RVUSD and
reserves RVEUR, calculated as:
𝑅𝑅𝑅𝑅𝐺𝐺𝐺𝐺𝐺𝐺
.
�1−𝑅𝑅𝑅𝑅𝑈𝑈𝑈𝑈𝑈𝑈 −𝑅𝑅𝑅𝑅𝐸𝐸𝐸𝐸𝐸𝐸 �

Share of JPY in FX It is the share of JY in FX reserves that excludes RVUSD, RVEUR,

reserves and RVGBP, calculated as:
𝑅𝑅𝑅𝑅𝐽𝐽𝐽𝐽𝐽𝐽
.
�1−𝑅𝑅𝑅𝑅𝑈𝑈𝑈𝑈𝑈𝑈 −𝑅𝑅𝑅𝑅𝐸𝐸𝐸𝐸𝐸𝐸 −𝑅𝑅𝑅𝑅𝐺𝐺𝐺𝐺𝐺𝐺 �

The estimations are done with the OLS method. The major currency issuers: the US, the euro
member countries, the UK, Japan, and China are not included in the estimations.

25
 Table 1. Roles of an International Currency
Function of money Governments Private actors
Store of value International reserves held Currency substitution
by central banks (private dollarization)
Medium of exchange Vehicle currency for Vehicle currency for private
foreign exchange foreign exchange trading
intervention
Unit of account Anchor for pegging smaller Denominating trade and
currencies financial transactions

26
 Table 2-1: USD Share in FX reserves (simple ratios)
Baseline Baseline Baseline Baseline Baseline
(1) (2) (3) (4) (5)
Share (t – 1) 0.890 0.879 0.879 0.878 0.878
(0.022)*** (0.022)*** (0.022)*** (0.022)*** (0.022)***
GDP ratio -0.097 -0.098 -0.104 -0.092 -0.111
(0.111) (0.115) (0.132) (0.104) (0.115)
ER volatility -3.253 -3.334 -3.378 -3.248 -3.546
(1.259)** (1.313)** (1.368)** (1.260)** (1.329)**
Inflation diff. -0.635 -0.523 -0.567 -0.530 -0.545
(1.295) (1.336) (1.343) (1.318) (1.338)
Share of trade w US 0.072 0.078 0.077 0.077 0.079
(0.019)*** (0.020)*** (0.020)*** (0.019)*** (0.020)***
USD as Anchor 0.042 0.036 0.036 0.036 0.035
(0.010)*** (0.009)*** (0.009)*** (0.009)*** (0.009)***
FX turnover, location 0.284 0.224 0.216 0.224 0.214
(0.390) (0.397) (0.411) (0.397) (0.393)
Political distance US 0.010 0.010 0.010 0.011
(0.005)* (0.005)* (0.005)* (0.005)**
US_sanctions 0.001
(0.007)
US_trade 0.004
(0.019)
US_financial -0.006
(0.009)
N 935 896 896 896 896
Adj. R2 0.89 0.89 0.89 0.89 0.89
# of countries 56 54 54 54 54
Years covered 1999 - 2022 1999 - 2022 1999 - 2022 1999 - 2022 1999 - 2022
* p<0.1; ** p<0.05; *** p<0.01
Note: The major currency issuers, the US, the euro member countries, the UK, Japan, and China, are not included in
the estimations. Political distance reflects how distant a country is from a major currency issuer based on voting
behavior at the UN. A smaller value means closer geopolitical distance.

27
 Table 2.2: EUR Share in FX reserves (simple ratios)
Baseline Baseline Baseline Baseline Baseline
(1) (2) (3) (4) (5)
Share (t – 1) 0.916 0.910 0.909 0.910 0.908
(0.023)*** (0.026)*** (0.025)*** (0.026)*** (0.026)***
GDP ratio 0.243 0.235 0.240 0.238 0.241
(0.107)** (0.107)** (0.110)** (0.111)** (0.109)**
ER volatility -2.024 -1.698 -1.681 -1.673 -1.677
(1.581) (1.626) (1.641) (1.661) (1.641)
Inflation diff. -5.924 -5.751 -5.786 -5.743 -5.804
(2.103)*** (2.140)*** (2.129)*** (2.144)*** (2.130)***
Share of trade w EURO area 0.074 0.072 0.069 0.071 0.070
(0.026)*** (0.025)*** (0.026)*** (0.025)*** (0.024)***
EUR as Anchor 0.016 0.010 0.011 0.010 0.011
(0.011) (0.011) (0.011) (0.012) (0.011)
FX turnover, location -0.113 -0.128 -0.128 -0.128 -0.127
(0.171) (0.173) (0.172) (0.174) (0.171)
Political distance euro -0.009 -0.009 -0.009 -0.009
(0.005)* (0.005)* (0.005)* (0.005)*
Euro_sanctions 0.003
(0.007)
Euro_trade 0.003
(0.009)
Euro_financial 0.006
(0.009)
N 852 836 836 836 836
Adj. R2 0.94 0.93 0.93 0.93 0.93
# of countries 52 51 51 51 51
Years covered 2001 - 2022 2001 - 2022 2001 - 2022 2001 - 2022 2001 - 2022
* p<0.1; ** p<0.05; *** p<0.01
Note: The major currency issuers, the US, the euro member countries, the UK, Japan, and China, are not included in
the estimations. Political distance reflects how distant a country is from a major currency issuer based on voting
behavior at the UN. A smaller value means closer geopolitical distance.

28
 Table 2.3: GBP Share in FX reserves (simple ratios)
Baseline Baseline Baseline Baseline Baseline
(1) (2) (3) (4) (5)
Share (t – 1) 0.850 0.852 0.851 0.852 0.852
(0.023)*** (0.024)*** (0.024)*** (0.024)*** (0.024)***
GDP ratio 0.751 0.732 0.735 0.730 0.735
(0.395)* (0.403)* (0.405)* (0.403)* (0.403)*
ER volatility 0.084 0.050 0.023 0.047 0.031
(0.467) (0.479) (0.475) (0.483) (0.474)
Inflation diff. 1.260 1.322 1.331 1.320 1.327
(0.513)** (0.519)** (0.522)** (0.519)** (0.520)**
Share of trade w UK 0.036 0.024 0.023 0.024 0.023
(0.015)** (0.016) (0.016) (0.016) (0.016)
FX turnover, location -0.097 -0.107 -0.104 -0.107 -0.104
(0.115) (0.117) (0.118) (0.118) (0.119)
Political distance UK -0.002 -0.002 -0.002 -0.002
(0.002) (0.002) (0.002) (0.002)
UK_sanctions -0.003
(0.002)
UK_trade -0.000
(0.003)
UK_financial -0.002
(0.003)
N 657 641 641 641 641
Adj. R2 0.80 0.80 0.80 0.80 0.80
# of countries 44 43 43 43 43
Years covered 1999 - 2022 1999 - 2022 1999 - 2022 1999 - 2022 1999 - 2022
* p<0.1; ** p<0.05; *** p<0.01
Note: The major currency issuers, the US, the euro member countries, the UK, Japan, and China, are not included in
the estimations. Political distance reflects how distant a country is from a major currency issuer based on voting
behavior at the UN. A smaller value means closer geopolitical distance.

29
 Table 2.4: JPY Share in FX reserves (simple ratios)
Baseline Baseline Baseline Baseline Baseline
(1) (2) (3) (4) (5)
Share (t – 1) 0.835 0.824 0.823 0.822 0.823
(0.033)*** (0.034)*** (0.035)*** (0.035)*** (0.035)***
GDP ratio 0.237 0.218 0.218 0.221 0.218
(0.191) (0.209) (0.210) (0.210) (0.210)
ER volatility 0.337 0.371 0.359 0.353 0.367
(0.182)* (0.189)* (0.192)* (0.194)* (0.189)*
Inflation diff. 0.213 0.223 0.221 0.221 0.225
(0.264) (0.281) (0.281) (0.281) (0.282)
Share of trade w Japan 0.002 0.010 0.010 0.011 0.010
(0.020) (0.020) (0.020) (0.020) (0.020)
FX turnover, location -0.523 -0.507 -0.504 -0.509 -0.508
(0.289)* (0.309) (0.309) (0.309) (0.309)
Political distance Japan -0.004 -0.004 -0.004 -0.004
(0.002)** (0.002)* (0.002)** (0.002)*
Japan_sanctions -0.004
(0.002)
Japan_trade -0.005
(0.003)*
Japan_financial -0.005
(0.003)
N 465 456 456 456 456
Adj. R2 0.74 0.74 0.74 0.74 0.74
# of countries 34 32 32 32 32
Years covered 1999 - 2022 1999 - 2022 1999 - 2022 1999 - 2022 1999 - 2022
* p<0.1; ** p<0.05; *** p<0.01
Note: The major currency issuers, the US, the euro member countries, the UK, Japan, and China, are not included in
the estimations. Political distance reflects how distant a country is from a major currency issuer based on voting
behavior at the UN. A smaller value means closer geopolitical distance.

30
 Table 2.5: RMB Share in FX reserves (simple ratios)
Baseline Baseline Baseline Baseline Baseline
(1) (2) (3) (4) (5)
Share (t – 1) 0.861 0.860 0.859 0.860 0.857
(0.043)*** (0.044)*** (0.043)*** (0.044)*** (0.043)***
GDP ratio 0.284 0.285 0.259 0.287 0.262
(0.247) (0.247) (0.255) (0.248) (0.254)
ER volatility -2.030 -1.984 -1.841 -1.987 -1.785
(1.866) (1.789) (1.840) (1.793) (1.862)
Inflation diff. -0.066 -0.087 -0.087 -0.090 -0.108
(0.190) (0.182) (0.180) (0.184) (0.181)
Sh of trade w/ China 0.007 0.012 0.018 0.011 0.015
(0.018) (0.022) (0.022) (0.023) (0.021)
FX turnover, location -1.875 -1.874 -1.697 -1.881 -1.653
(2.265) (2.260) (2.323) (2.266) (2.338)
Political distance China -0.002 -0.002 -0.002 -0.002
(0.004) (0.004) (0.004) (0.004)
China_sanctions 0.013
(0.009)
China_trade -0.003
(0.006)
China_financial 0.020
(0.007)**
N 231 231 231 231 231
Adj. R2 0.68 0.67 0.67 0.67 0.67
# of countries 20 20 20 20 20
Years covered 2001 - 2022 2001 - 2022 2001 - 2022 2001 - 2022 2001 - 2022
* p<0.1; ** p<0.05; *** p<0.01
Note: The major currency issuers, the US, the euro member countries, the UK, Japan, and China, are not included in
the estimations. Political distance reflects how distant a country is from a major currency issuer based on voting
behavior at the UN. A smaller value means closer geopolitical distance.

31
 Table 3: Pooled Regression: Major Currency Share in FX reserves (simple ratios)
Baseline Baseline Baseline Baseline Baseline
(1) (2) (3) (4) (5)
Share(t-1) 0.902 0.900 0.900 0.900 0.900
(0.018)*** (0.019)*** (0.019)*** (0.019)*** (0.019)***
USD 0.023 0.019 0.019 0.018 0.019
(0.012)* (0.013) (0.013) (0.013) (0.013)
EUR -0.006 -0.009 -0.009 -0.011 -0.010
(0.007) (0.008) (0.007) (0.008) (0.008)
JPY 0.007 0.005 0.005 0.004 0.005
(0.006) (0.006) (0.006) (0.006) (0.006)
GBP -0.011 -0.013 -0.014 -0.012 -0.013
(0.022) (0.023) (0.023) (0.023) (0.023)
RMB -0.000 -0.002 -0.002 -0.004 -0.003
(0.006) (0.006) (0.006) (0.007) (0.007)
GDP share in world 0.063 0.067 0.067 0.072 0.068
(0.035)* (0.036)* (0.035)* (0.040)* (0.037)*
NEER volatility -0.774 -0.742 -0.742 -0.689 -0.731
(0.335)** (0.342)** (0.343)** (0.348)* (0.348)**
Inflation diff. -0.064 -0.049 -0.051 -0.048 -0.049
(0.221) (0.224) (0.220) (0.225) (0.224)
Share of trade w Ctry i 0.053 0.057 0.057 0.056 0.057
(0.014)*** (0.015)*** (0.015)*** (0.015)*** (0.015)***
Anchor Currency 0.036 0.035 0.035 0.035 0.035
(0.007)*** (0.007)*** (0.007)*** (0.007)*** (0.007)***
FX turnover, loc 0.056 0.054 0.054 0.047 0.053
(0.066) (0.068) (0.067) (0.066) (0.068)
Political distance from Ctry i 0.002 0.002 0.001 0.002
(0.002) (0.002) (0.002) (0.002)
Sanctions by Ctry i 0.000
(0.003)
Trade sanctions by Ctry i 0.006
(0.010)
Financial sanctions by Ctry i 0.001
(0.004)
N 3,140 3,060 3,060 3,060 3,060
Adj. R2 0.97 0.97 0.97 0.97 0.97
# of countries 56 54 54 54 54
Years covered 1999 - 2022 1999 - 2022 1999 - 2022 1999 - 2022 1999 - 2022
* p<0.1; ** p<0.05; *** p<0.01
Note: The major currency issuers, the US, the euro member countries, the UK, Japan, and China, are not included in
the estimations. Political distance reflects how distant a country is from a major currency issuer based on voting
behavior at the UN. A smaller value means closer geopolitical distance.

32
 Table 4: Pooled Regression: Major Currency Share in FX reserves (logit ratios)
Baseline Baseline Baseline Baseline Baseline
(1) (2) (3) (4) (5)
Share(t-1) 0.909 0.909 0.909 0.909 0.909
(0.016)*** (0.016)*** (0.016)*** (0.016)*** (0.016)***
USD -0.354 -0.283 -0.284 -0.292 -0.282
(0.260) (0.288) (0.295) (0.292) (0.287)
EUR -0.650 -0.612 -0.612 -0.622 -0.610
(0.201)*** (0.211)*** (0.215)*** (0.217)*** (0.211)***
JPY -0.562 -0.536 -0.536 -0.545 -0.535
(0.167)*** (0.173)*** (0.175)*** (0.179)*** (0.173)***
GBP -0.183 -0.145 -0.145 -0.136 -0.150
(0.419) (0.447) (0.450) (0.448) (0.457)
RMB -0.321 -0.300 -0.300 -0.311 -0.298
(0.204) (0.211) (0.215) (0.216) (0.212)
GDP share in world 1.595 1.582 1.583 1.625 1.572
(0.778)** (0.804)* (0.824)* (0.828)* (0.809)*
NEER volatility -4.661 -4.204 -4.203 -3.829 -4.312
(6.375) (6.547) (6.542) (6.718) (6.570)
Inflation diff. -2.450 -1.989 -1.984 -1.947 -1.984
(4.922) (4.947) (4.908) (4.961) (4.945)
Share of trade w Ctry i 0.601 0.544 0.545 0.540 0.547
(0.126)*** (0.146)*** (0.145)*** (0.145)*** (0.145)***
Anchor Currency 0.248 0.260 0.260 0.262 0.260
(0.048)*** (0.051)*** (0.050)*** (0.051)*** (0.051)***
FX turnover, loc -0.507 -0.492 -0.492 -0.539 -0.473
(1.235) (1.268) (1.269) (1.277) (1.304)
Political distance from Ctry i -0.032 -0.032 -0.033 -0.032
(0.025) (0.026) (0.025) (0.026)
Sanctions by Ctry i -0.001
(0.041)
Trade sanctions by Ctry i 0.039
(0.047)
Financial sanctions by Ctry i -0.012
(0.055)
N 2,729 2,654 2,654 2,654 2,654
Adj. R2 0.93 0.93 0.93 0.93 0.93
# of countries 56 54 54 54 54
Years covered 1999 - 2022 1999 - 2022 1999 - 2022 1999 - 2022 1999 - 2022
* p<0.1; ** p<0.05; *** p<0.01
Note: The major currency issuers, the US, the euro member countries, the UK, Japan, and China, are not included in
the estimations. Political distance reflects how distant a country is from a major currency issuer based on voting
behavior at the UN. A smaller value means closer geopolitical distance.

33
 Table 5: Pooled Regression: Major Currency Share in FX reserves (recursively defined
ratios)
Baseline Baseline Baseline Baseline Baseline
(1) (2) (3) (4) (5)
Share(t-1) 0.781 0.777 0.777 0.775 0.777
(0.062)*** (0.064)*** (0.064)*** (0.064)*** (0.064)***
USD 0.004 -0.017 -0.016 -0.020 -0.016
(0.034) (0.039) (0.039) (0.039) (0.039)
EUR 0.000 -0.012 -0.012 -0.016 -0.011
(0.024) (0.026) (0.027) (0.027) (0.026)
JPY 0.032 0.025 0.026 0.022 0.026
(0.023) (0.024) (0.025) (0.025) (0.024)
GBP 0.105 0.095 0.094 0.100 0.093
(0.076) (0.077) (0.077) (0.078) (0.078)
RMB -0.010 -0.019 -0.018 -0.022 -0.018
(0.023) (0.026) (0.026) (0.027) (0.026)
GDP share in world 0.521 0.548 0.544 0.567 0.544
(0.173)*** (0.180)*** (0.178)*** (0.189)*** (0.182)***
NEER vol -1.893 -1.919 -1.917 -1.755 -1.961
(0.940)** (0.968)* (0.968)* (0.999)* (0.975)**
Inflation diff. -0.812 -0.834 -0.858 -0.841 -0.831
(0.711) (0.727) (0.732) (0.729) (0.728)
Share of trade w Ctry i 0.131 0.148 0.147 0.146 0.149
(0.039)*** (0.047)*** (0.046)*** (0.046)*** (0.047)***
Anchor Currency 0.058 0.054 0.055 0.055 0.054
(0.020)*** (0.018)*** (0.019)*** (0.019)*** (0.018)***
FX turnover, loc -0.069 -0.076 -0.075 -0.098 -0.070
(0.201) (0.208) (0.208) (0.211) (0.211)
Political distance from Ctry i 0.007 0.007 0.007 0.008
(0.007) (0.007) (0.007) (0.007)
Sanctions by Ctry i 0.004
(0.009)
Trade sanctions by Ctry i 0.018
(0.019)
Financial sanctions by Ctry i -0.004
(0.009)
N 3,081 3,005 3,005 3,005 3,005
Adj. R2 0.92 0.91 0.91 0.91 0.91
# of countries 56 54 54 54 54
Years covered 1999 - 2022 1999 - 2022 1999 - 2022 1999 - 2022 1999 - 2022
* p<0.1; ** p<0.05; *** p<0.01
Note: The major currency issuers, the US, the euro member countries, the UK, Japan, and China, are not included in
the estimations. Political distance reflects how distant a country is from a major currency issuer based on voting
behavior at the UN. A smaller value means closer geopolitical distance.

34
Figure 1: Shares of major currencies in the world’s aggregate foreign exchange reserves

Figure 2: FX turnover as a measure of size of financial markets in the home country for each
currency

35
Figure 3: Reserve holdings correlate strongly with size of home financial market (turnover)

Figure 4: International currency metrics for US dollar and euro. Source: ECB (2023).

36
Figure 5: US dollar share predicted by logit specification

Figure 6: Euro share predicted by logit specification

37
Appendix: Estimation Results with Logit Shares or Recursively Defined Shares

A1-1: USD Share in FX reserves (Shares in Logit Transformation)

Baseline Baseline Baseline Baseline Baseline
(1) (2) (3) (4) (5)
Share (t – 1) 0.927 0.924 0.921 0.925 0.917
(0.026)*** (0.027)*** (0.026)*** (0.027)*** (0.023)***
GDP ratio -0.262 -0.226 0.330 -0.337 -0.561
(1.345) (1.432) (1.448) (1.443) (1.470)
ER volatility -14.636 -14.721 -11.026 -16.472 -20.549
(10.700) (11.134) (11.814) (11.408) (10.520)*
Inflation diff. -8.394 -8.123 -4.296 -8.040 -9.050
(6.492) (6.704) (7.758) (6.756) (6.551)
Share of trade w US 0.347 0.362 0.401 0.379 0.407
(0.154)** (0.161)** (0.163)** (0.161)** (0.165)**
USD as Anchor 0.238 0.208 0.196 0.207 0.196
(0.056)*** (0.058)*** (0.059)*** (0.059)*** (0.063)***
FX turnover, location 2.618 2.334 3.019 2.348 2.038
(3.022) (3.110) (3.212) (3.107) (3.046)
Political distance us 0.043 0.058 0.046 0.064
(0.035) (0.036) (0.036) (0.037)*
US_sanctions -0.083
(0.052)
US_trade sanctions -0.077
(0.075)
US_financial -0.157
Sanctions (0.080)*
N 903 864 864 864 864
Adj. R2 0.90 0.90 0.90 0.90 0.90
# of countries 56 54 54 54 54
Years covered 1999 - 2022 1999 - 2022 1999 - 2022 1999 - 2022 1999 - 2022
* p<0.1; ** p<0.05; *** p<0.01
Note: The major currency issuers, the US, the euro member countries, the UK, Japan, and China, are not included in
the estimations. Political distance reflects how distant a country is from a major currency issuer based on voting
behavior at the UN. A smaller value means closer geopolitical distance.

38
 A1-2: USD Share in FX reserves ( Recirsively Defined Shares)
Baseline Baseline Baseline Baseline Baseline
(1) (2) (3) (4) (5)
Share (t – 1) 0.890 0.879 0.879 0.878 0.878
(0.022)*** (0.022)*** (0.022)*** (0.022)*** (0.022)***
GDP ratio -0.097 -0.098 -0.104 -0.092 -0.111
(0.111) (0.115) (0.132) (0.104) (0.115)
ER volatility -3.253 -3.334 -3.378 -3.248 -3.546
(1.259)** (1.313)** (1.368)** (1.260)** (1.329)**
Inflation diff. -0.635 -0.523 -0.567 -0.530 -0.545
(1.295) (1.336) (1.343) (1.318) (1.338)
Share of trade w US 0.072 0.078 0.077 0.077 0.079
(0.019)*** (0.020)*** (0.020)*** (0.019)*** (0.020)***
USD as Anchor 0.042 0.036 0.036 0.036 0.035
(0.010)*** (0.009)*** (0.009)*** (0.009)*** (0.009)***
FX turnover, location 0.284 0.224 0.216 0.224 0.214
(0.390) (0.397) (0.411) (0.397) (0.393)
Political distance us 0.010 0.010 0.010 0.011
(0.005)* (0.005)* (0.005)* (0.005)**
US_sanctions 0.001
(0.007)
US_trade sanctions 0.004
(0.019)
US_financial -0.006
Sanctions (0.009)
N 935 896 896 896 896
Adj. R2 0.89 0.89 0.89 0.89 0.89
# of countries 56 54 54 54 54
Years covered 1999 - 2022 1999 - 2022 1999 - 2022 1999 - 2022 1999 - 2022
* p<0.1; ** p<0.05; *** p<0.01
Note: The major currency issuers, the US, the euro member countries, the UK, Japan, and China, are not included in
the estimations. Political distance reflects how distant a country is from a major currency issuer based on voting
behavior at the UN. A smaller value means closer geopolitical distance.

39
 A2-1: EUR Share in FX reserves (Shares in Logit Transformation)
Baseline Baseline Baseline Baseline Baseline
(1) (2) (3) (4) (5)
Share (t – 1) 0.909 0.901 0.894 0.900 0.895
(0.024)*** (0.026)*** (0.024)*** (0.026)*** (0.024)***
GDP ratio 4.833 4.842 5.058 4.987 4.957
(1.771)*** (1.788)*** (1.781)*** (1.868)** (1.768)***
ER volatility -7.297 -3.640 -3.560 -2.529 -3.771
(21.606) (22.664) (22.809) (23.084) (22.784)
Inflation diff. -39.730 -37.444 -39.485 -37.071 -38.860
(26.763) (27.259) (27.180) (27.209) (27.176)
Share of trade w EURO area 0.933 0.868 0.728 0.842 0.823
(0.269)*** (0.256)*** (0.266)*** (0.250)*** (0.247)***
EUR as Anchor 0.081 -0.008 0.052 0.008 0.011
(0.083) (0.094) (0.107) (0.095) (0.094)
FX turnover, location -5.696 -6.068 -6.056 -6.127 -6.032
(3.614) (3.704) (3.680) (3.730) (3.682)
Political distance euro -0.133 -0.146 -0.133 -0.143
(0.057)** (0.061)** (0.058)** (0.060)**
Euro_sanctions 0.142
(0.098)
Euro_trade sanctions 0.104
(0.102)
Euro_financial 0.135
Sanctions (0.122)
N 798 782 782 782 782
Adj. R2 0.86 0.86 0.86 0.86 0.86
# of countries 52 51 51 51 51
Years covered 2001 - 2022 2001 - 2022 2001 - 2022 2001 - 2022 2001 - 2022
* p<0.1; ** p<0.05; *** p<0.01
Note: The major currency issuers, the US, the euro member countries, the UK, Japan, and China, are not included in
the estimations. Political distance reflects how distant a country is from a major currency issuer based on voting
behavior at the UN. A smaller value means closer geopolitical distance.

40
 A2-2: EUR Share in FX reserves (Recursively Defined Shares)
Baseline Baseline Baseline Baseline Baseline
(1) (2) (3) (4) (5)
Share (t – 1) 0.605 0.591 0.587 0.587 0.591
(0.190)*** (0.193)*** (0.195)*** (0.196)*** (0.193)***
GDP ratio 1.401 1.410 1.461 1.476 1.410
(0.762)* (0.757)* (0.779)* (0.819)* (0.765)*
ER volatility -4.363 -3.852 -3.748 -3.429 -3.851
(4.369) (4.507) (4.566) (4.742) (4.514)
Inflation diff. -16.279 -15.903 -16.220 -15.795 -15.907
(7.480)** (7.410)** (7.700)** (7.381)** (7.505)**
Share of trade w EURO area 0.335 0.321 0.294 0.311 0.321
(0.146)** (0.143)** (0.126)** (0.135)** (0.141)**
EUR as Anchor 0.004 -0.035 -0.025 -0.028 -0.035
(0.037) (0.040) (0.040) (0.041) (0.040)
FX turnover, location 0.232 0.222 0.220 0.203 0.222
(0.840) (0.848) (0.839) (0.861) (0.845)
Political distance euro -0.039 -0.041 -0.039 -0.039
(0.024) (0.025) (0.024) (0.025)
Euro_sanctions 0.028
(0.031)
Euro_trade sanctions 0.044
(0.059)
Euro_financial 0.000
Sanctions (0.027)
N 839 823 823 823 823
Adj. R2 0.54 0.54 0.54 0.54 0.54
# of countries 52 51 51 51 51
Years covered 2001 - 2022 2001 - 2022 2001 - 2022 2001 - 2022 2001 - 2022
* p<0.1; ** p<0.05; *** p<0.01
Note: The major currency issuers, the US, the euro member countries, the UK, Japan, and China, are not included in
the estimations. Political distance reflects how distant a country is from a major currency issuer based on voting
behavior at the UN. A smaller value means closer geopolitical distance.

41
 A3-1: GBP Share in FX reserves (Shares in Logit Transformation)
Baseline Baseline Baseline Baseline Baseline
(1) (2) (3) (4) (5)
Share (t – 1) 0.896 0.893 0.894 0.893 0.894
(0.038)*** (0.037)*** (0.037)*** (0.038)*** (0.037)***
GDP ratio 20.591 20.323 20.280 20.592 20.127
(11.435)* (11.484)* (11.468)* (11.771)* (11.435)*
ER volatility 1.925 1.022 1.488 1.342 1.319
(9.741) (9.718) (10.069) (10.263) (10.002)
Inflation diff. 16.483 19.926 19.619 20.102 19.630
(21.813) (22.646) (22.728) (22.733) (22.716)
Share of trade w UK 1.028 0.397 0.437 0.410 0.415
(0.470)** (0.394) (0.446) (0.404) (0.424)
FX turnover, location 0.161 -0.312 -0.383 -0.318 -0.380
(2.116) (2.169) (2.183) (2.173) (2.191)
Political distance uk -0.123 -0.121 -0.121 -0.124
(0.043)*** (0.044)*** (0.044)*** (0.043)***
UK_sanctions 0.050
(0.117)
UK_trade sanctions 0.028
(0.096)
UK_financial 0.044
Sanctions (0.144)
N 567 551 551 551 551
Adj. R2 0.84 0.84 0.84 0.84 0.84
# of countries 43 42 42 42 42
Years covered 1999 - 2022 1999 - 2022 1999 - 2022 1999 - 2022 1999 - 2022
* p<0.1; ** p<0.05; *** p<0.01
Note: The major currency issuers, the US, the euro member countries, the UK, Japan, and China, are not included in
the estimations. Political distance reflects how distant a country is from a major currency issuer based on voting
behavior at the UN. A smaller value means closer geopolitical distance.

42
 A3-2: GBP Share in FX reserves (Recursively Defined Shares)
Baseline Baseline Baseline Baseline Baseline
(1) (2) (3) (4) (5)
Share (t – 1) 0.844 0.844 0.843 0.844 0.844
(0.023)*** (0.024)*** (0.023)*** (0.024)*** (0.024)***
GDP ratio 4.320 4.210 4.239 4.272 4.207
(1.770)** (1.815)** (1.847)** (1.918)** (1.821)**
ER volatility 1.135 1.095 1.306 1.167 1.144
(2.519) (2.626) (2.668) (2.776) (2.608)
Inflation diff. 3.900 4.050 4.065 4.097 4.048
(2.083)* (2.103)* (2.119)* (2.112)* (2.106)*
Share of trade w UK 0.147 0.118 0.128 0.120 0.120
(0.081)* (0.111) (0.111) (0.112) (0.112)
FX turnover, location -0.298 -0.353 -0.375 -0.357 -0.361
(0.321) (0.332) (0.334) (0.330) (0.334)
Political distance uk -0.006 -0.006 -0.006 -0.006
(0.011) (0.011) (0.012) (0.011)
UK_sanctions 0.015
(0.017)
UK_trade sanctions 0.006
(0.021)
UK_financial 0.005
Sanctions (0.018)
N 642 626 626 626 626
Adj. R2 0.77 0.77 0.77 0.77 0.77
# of countries 44 43 43 43 43
Years covered 1999 - 2022 1999 - 2022 1999 - 2022 1999 - 2022 1999 - 2022
* p<0.1; ** p<0.05; *** p<0.01
Note: The major currency issuers, the US, the euro member countries, the UK, Japan, and China, are not included in
the estimations. Political distance reflects how distant a country is from a major currency issuer based on voting
behavior at the UN. A smaller value means closer geopolitical distance.

43
 A4-1: JPY Share in FX reserves (Shares in Logit Transformation)
Baseline Baseline Baseline Baseline Baseline
(1) (2) (3) (4) (5)
Share (t – 1) 0.885 0.852 0.851 0.852 0.851
(0.038)*** (0.041)*** (0.042)*** (0.041)*** (0.041)***
GDP ratio 3.817 1.198 0.916 1.119 1.065
(6.122) (7.319) (7.286) (7.313) (7.326)
ER volatility 6.229 9.318 8.627 8.320 9.376
(12.389) (13.177) (13.646) (13.876) (13.158)
Inflation diff. 3.959 3.162 2.905 2.855 3.153
(8.399) (8.960) (8.884) (8.880) (8.969)
Share of trade w Japan 2.781 3.079 3.125 3.134 3.080
(1.089)** (1.131)** (1.138)** (1.144)** (1.131)**
FX turnover, location -6.316 -3.082 -2.330 -2.624 -2.844
(9.765) (10.120) (10.451) (10.289) (10.274)
Political distance Japan -0.325 -0.324 -0.327 -0.323
(0.107)*** (0.107)*** (0.107)*** (0.107)***
Japan_sanctions -0.246
(0.198)
Japan_trade sanctions -0.313
(0.240)
Japan_financial -0.150
Sanctions (0.148)
N 369 365 365 365 365
Adj. R2 0.78 0.79 0.79 0.79 0.78
# of countries 31 30 30 30 30
Years covered 1999 - 2022 1999 - 2022 1999 - 2022 1999 - 2022 1999 - 2022
* p<0.1; ** p<0.05; *** p<0.01
Note: The major currency issuers, the US, the euro member countries, the UK, Japan, and China, are not included in
the estimations. Political distance reflects how distant a country is from a major currency issuer based on voting
behavior at the UN. A smaller value means closer geopolitical distance.

44
 A4-2: JPY Share in FX reserves (Recursively Defined Shares)
Baseline Baseline Baseline Baseline Baseline
(1) (2) (3) (4) (5)
Share (t – 1) 0.741 0.740 0.739 0.739 0.739
(0.057)*** (0.057)*** (0.057)*** (0.056)*** (0.057)***
GDP ratio 0.927 0.962 0.970 0.941 0.961
(1.269) (1.266) (1.274) (1.267) (1.267)
ER volatility -1.555 -1.712 -1.620 -1.539 -1.764
(1.435) (1.483) (1.523) (1.540) (1.489)
Inflation diff. -0.019 -0.046 -0.038 -0.038 -0.038
(1.324) (1.335) (1.336) (1.336) (1.338)
Share of trade w Japan 0.474 0.465 0.463 0.460 0.464
(0.190)** (0.189)** (0.189)** (0.188)** (0.191)**
FX turnover, location -0.504 -0.551 -0.576 -0.529 -0.563
(2.095) (2.113) (2.148) (2.135) (2.114)
Political distance japan 0.001 0.001 0.002 0.002
(0.014) (0.014) (0.014) (0.014)
Japan_sanctions 0.026
(0.054)
Japan_trade 0.047
Sanctions (0.069)
Japan_financial -0.045
Sanctions (0.014)***
N 442 437 437 437 437
Adj. R2 0.65 0.65 0.65 0.65 0.65
# of countries 32 31 31 31 31
Years covered 1999 - 2022 1999 - 2022 1999 - 2022 1999 - 2022 1999 - 2022
* p<0.1; ** p<0.05; *** p<0.01
Note: The major currency issuers, the US, the euro member countries, the UK, Japan, and China, are not included in
the estimations. Political distance reflects how distant a country is from a major currency issuer based on voting
behavior at the UN. A smaller value means closer geopolitical distance.

45
 A5-1: RMB Share in FX reserves (Shares in Logit Transformation)
Baseline Baseline Baseline Baseline Baseline

Share (t – 1) 0.907 0.909 0.908 0.916 0.905

(0.021)*** (0.021)*** (0.019)*** (0.020)*** (0.021)***
GDP ratio 43.190 43.132 38.953 39.365 41.936
(27.397) (27.757) (28.893) (29.782) (27.976)
ER volatility 599.661 601.948 607.767 605.428 604.504
(322.712)* (324.441)* (325.631)* (324.934)* (326.607)*
Inflation diff. 92.760 91.021 86.007 89.057 88.298
(63.714) (65.545) (66.944) (66.894) (65.913)
Sh of trade w/ China -0.238 0.041 0.853 0.855 0.232
(1.415) (2.152) (2.512) (2.599) (2.254)
FX turnover, location 6.781 5.685 25.484 15.633 15.336
(66.276) (66.849) (71.178) (65.276) (71.955)
Political distance china -0.064 -0.131 -0.154 -0.068
(0.238) (0.267) (0.282) (0.246)
China_sanctions 0.730
(0.446)
China_trade sanctions 1.072
(0.761)
China_financial 0.541
Sanctions (0.384)
N 92 92 92 92 92
Adj. R2 0.84 0.84 0.84 0.84 0.84
# of countries 18 18 18 18 18
Years covered 2010 - 2022 2010 - 2022 2010 - 2022 2010 - 2022 2010 - 2022
* p<0.1; ** p<0.05; *** p<0.01
Note: The major currency issuers, the US, the euro member countries, the UK, Japan, and China, are not included in
the estimations. Political distance reflects how distant a country is from a major currency issuer based on voting
behavior at the UN. A smaller value means closer geopolitical distance.

46
 A5-2: RMB Share in FX reserves (Recursively Defined Shares)
Baseline Baseline Baseline Baseline Baseline
(1) (2) (3) (4) (5)
Share (t – 1) 0.823 0.821 0.808 0.820 0.797
(0.085)*** (0.082)*** (0.080)*** (0.083)*** (0.085)***
GDP ratio 1.103 1.102 0.864 1.115 0.900
(0.953) (0.957) (0.966) (0.965) (0.957)
ER volatility -8.177 -7.941 -6.720 -7.968 -6.404
(6.951) (6.833) (7.549) (6.852) (7.682)
Inflation diff. 0.607 0.502 0.481 0.489 0.295
(0.591) (0.551) (0.512) (0.540) (0.546)
Sh of trade w/ China -0.047 -0.020 0.040 -0.025 0.013
(0.146) (0.169) (0.163) (0.180) (0.162)
FX turnover, location -6.369 -6.313 -4.526 -6.358 -4.139
(9.231) (9.212) (9.485) (9.234) (9.634)
Political distance china -0.008 -0.013 -0.008 -0.010
(0.019) (0.020) (0.020) (0.020)
China_sanctions 0.117
(0.064)*
China_trade sanctions -0.014
(0.041)
China_financial 0.182
Sanctions (0.063)***
N 223 223 223 223 223
Adj. R2 0.66 0.66 0.66 0.66 0.66
# of countries 20 20 20 20 20
Years covered 2001 - 2022 2001 - 2022 2001 - 2022 2001 - 2022 2001 - 2022
* p<0.1; ** p<0.05; *** p<0.01
Note: The major currency issuers, the US, the euro member countries, the UK, Japan, and China, are not included in
the estimations. Political distance reflects how distant a country is from a major currency issuer based on voting
behavior at the UN. A smaller value means closer geopolitical distance.

47