The resurgence of currency mismatches in emerging market

International Economics and Economic Policy (2021) 18:721–742 https://doi.org/10.1007/s10368-021-00500-x
ORIGINAL PAPER
The resurgence of currency mismatches: Emerging market economies are not out of the woods yet?
Hari Venkatesh1  · Gourishankar S. Hiremath1
Accepted: 29 April 2021/ Published online: 30 June 2021 © The Author(s), under exclusive licence to Springer-Verlag GmbH Germany, part of Springer Nature 2021
Abstract The emerging market economies (EMEs) are experiencing significant financial distress due to the rapid accumulation of foreign currency-denominated debt in recent years. We develop the foreign exposure indicators such as original sin and currency mismatches using a novel data set. Our computations suggest that Latin American economies suffer from the original sin problem, followed by Central European countries. We find a higher degree of currency mismatches in Argentina, Chile, Colombia, Indonesia, Poland, Mexico, and Turkey. The resurgence of currency mismatches and the Covid-19 pandemic is a stress test for monetary policy frameworks. We find that country’s size, inflation volatility, and exchange rate depreciation cause currency mismatches. We show that the currency mismatch and original sin problem are lower in countries following de-dollarization policies such as limiting debt exposure, effective monetary and fiscal policies, better institutional quality, and export openness. The EMEs need to adopt policies to control currency mismatches, which are consistent with their growth-oriented policies. We suggest the independence of monetary policy, the implementation of macroprudential policies, and the development of offshore bond markets in a local currency. These policies control currency mismatches without changing the growth orientation of the EMEs. South Africa, Hungary, and Asian economies hold lessons for EMEs in controlling currency mismatches.
Highlights  • Emerging market economies (EMEs) exposed to foreign currency risk and currency mismatches at
an alarming level • Latin America has greater original sin followed by central European economies • Countries following de-dollarization policies lower the balance sheet vulnerabilities • Forex reserves can safeguard EMEs from shocks emanating due to Covid-19 pandemic • Independence of monetary policy, macroprudential policies, and development of offshore bond
markets in a local currency control currency mismatches.
* Hari Venkatesh [email protected]
Extended author information available on the last page of the article
1 3 Vol.:(0123456789)

722

H. Venkatesh, G. S. Hiremath

Keywords  Currency mismatches · Original sin hypothesis · Emerging market economies · Foreign currency debt · Financial vulnerability · Covid-19 pandemic
JEL Codes  F30 · F31 · F34 · G01

1 Introduction
Since 2008, emerging market economies (EMEs) are experiencing significant financial distress and large output losses due to the large ebb and flow of foreign currencydenominated debt (FCD). The dollar debt issuance of EMEs has substantially grown to 134 percent from $1.57 trillion in 2008 to $3.67 trillion in 2018 (BIS 2019).1 Such growth has been owing to the low-cost source of capital but at the cost of currency risk. In addition, the share of FCD in GDP increased from 10.74 percent in 2007 Q4 to 16.5 percent in 2018 Q3, indicating an alarming sign of external sector vulnerability in EMEs (Fig. 1). The high level of FCD possibly leads to a currency mismatch problem, which, in turn, increases the likelihood of financial crises. On this count, it is clear that the Covid-19 and consequent lockdown of economies can lead to destabilizing effects on EMEs. Therefore, there is a need to measure foreign currency exposure to assess recent external vulnerability and financial distress. In this context, we evaluate the external vulnerabilities in EMEs by developing a method to measure exposure. We also investigate the factor responsible for such exposures.
The EMEs raise the capital from international markets in foreign currency as they find it difficult to borrow in their own currency. This inability to borrow in own currency is termed as ‘original sin’ (Eichengreen et al. 2005a, b). The original sin results in a ’currency mismatch’ in the assets and liabilities of the country. Currency mismatch is defined as the mismatch between currency composition of liabilities and assets where the assets are denominated in domestic currency but liabilities in foreign currency (Goldstein and Turner 2004). We measure the original sin and currency mismatches in EMEs to examine the extent of external vulnerability. Nevertheless, the foreign currency borrowings boost the production and output in EMEs (Hiremath 2016). Therefore, while the currency mismatches suggest vulnerability, the debt cannot be completely averted as leverage is essential for the economic growth.
The EMEs are posed with significant challenges, such as a slowdown in growth rate, higher sensitivity to the exchange rate risk, and capital volatility. The exchange rate pass-through is higher in EMEs than in advanced economies, and therefore the fallout of currency mismatch will be devastating in the former than the latter (BIS 2019). Thus, the growing FCD and original sin can hurt the balance sheet of EMEs in the event of exchange rate depreciation. The local currency bond spread sharply rises in the event of exchange rate depreciation and capital outflows due to Covid-19 (Hofmann et al. 2020).
On the other hand, foreign currency risk can be hedged by earning forex reserves through exports (Goldstein and Turner 2004). For instance, Chinese Taipei and
1  The debt of EMEs is primarily denominated in the US dollar ($3.67 trillion), Euro ($792 billion), and Japanese yen ($72 billion).
1 3

The resurgence of currency mismatches: Emerging market…

723

30

25

20

15

10

5

0 2005-1.Q 2006-3.Q 2008-1.Q 2009-3.Q 2011-1.Q 2012-3.Q 2014-1.Q 2015-3.Q 2017-1.Q 2018-3.Q

EMEs % of GDP

Asian EMEs

Other EMEs

Fig. 1  Foreign currency-denominated debt (FCD % of GDP). Source: Prepared from the data collected from the Bank of International Settlements (BIS)
China sustained larger FCD levels as they accumulated a considerable amount of foreign currency assets (215% and 166% of exports, respectively). Therefore, measuring both assets and liabilities in foreign currency reveals the severity of currency mismatch problems in aggregate balance sheets. Further, computation of currency mismatches at the aggregate level can help policymakers to make the right decisions apposite to external risk.
Less is known about the recent developments in foreign currency exposure of EMEs, especially that of the corporate sector. Studies such as Goldstein and Turner (2004), Eichengreen et al. (2007), and Chui et al. (2018) document that EMEs lengthened their foreign currency exposure. The FCD and currency mismatches have been associated with a higher incidence of financial crises in the past. Still, little attention is paid to the topic of currency mismatches in the recent past despite staggering growth in such mismatches. Eichengreen et al. (2005a, b, 2007), Goldstein and Turner (2004), Park (2011), and Chui et al. (2018) emphasize the need for a comprehensive method of measuring currency mismatches in EMEs. To the best of our knowledge, no study probes recent developments in currency mismatches and implications of such trends. Hence, this paper fills these gaps in the literature by analyzing the current evolution in foreign currency exposure and currency mismatches in EMEs.
We contribute to the literature on foreign currency exposure and external vulnerabilities by analyzing the evolution of FCD, currency mismatches, and policy implications. To do so, we put together and analyze the variety of data sets on foreign currency exposure. First, we examine the issuance of international debt securities encompassing the currency composition of external debt. We compute the original sin index for the EMEs using granular data on global debt securities. Second, we complement the
1 3

724

H. Venkatesh, G. S. Hiremath

information on original sin values with country-level data on the share of foreign currency debt in total debt outstanding. Third, we also use the balance sheet information to calculate the net asset position in foreign currency and examine currency mismatches. Using these indices, we take a fresh look at the role of original sin and foreign currency debt in influencing currency mismatches. One of the main contributions of this paper is to show the resurgence of currency mismatches in EMEs since the post-global financial crisis (GFC). Finally, we examine the factors responsible for the currency mismatches and original sin and offer policy inputs to manage external vulnerability.
The computation of currency mismatches reveals a higher degree of currency mismatches in Argentina, Chile, Colombia, Indonesia, Poland, Mexico, and Turkey, indicating the likelihood of financial crises. We show that the currency mismatch problem is lower in countries that follow de-dollarization policies such as limiting debt exposure, effective monetary and fiscal policies, better institutional quality, and export openness. The findings suggest the crucial role of foreign currency assets, macroprudential policies, and domestic bond markets in limiting currency mismatches and enhancing the resilience of the financial system in the event of a crisis.
The rest of the paper is organized as follows. The conceptual framework and analytical issues of currency mismatches are discussed in Section 2. In Section 3, we present the recent trends in currency mismatches and important insights into policy conflicts. We discuss the factors responsible for currency mismatches and original sin in Section 4. The last section concludes the paper with policy implications.

2 Conceptual framework

2.1 Original sin hypothesis

In their theory of original sin, Eichengreen et al. (2005a) define sin as “the inability of a country to borrow abroad in its own currency.” The original sin occurs when the economies exhibit a lack of monetary creditability, low credit ratings, and volatile capital flows and output. Overall, the incompleteness in financial markets is the cause of original sin. Many of the EMEs are unable to borrow for long-term maturity from the domestic markets. Such inability can be termed as the domestic original sin. Eichengreen et al. (2005a) construct three indices of original sin (OSIN):

OSIN1 = 1 − Securities issued by country i in currency i
Securities issued by country i

(1)

when the county issues all the securities in domestic (foreign) currency, OSIN1 is zero (one). However, OSIN1 does not include hedging instruments. Hence, the OSIN2 index includes hedging instruments such as swaps. Finally, the OSIN3 is a comprehensive measure that encompasses OSIN2 and long-term debt indexed to prices and imposes lower bounds:

OSIN3 = max(1 − Securities in currency i , 0)
Securities issued by country i

(2)

1 3

The resurgence of currency mismatches: Emerging market…

725

The original sin index ranges between zero and one. The value closer to one implies an extreme level of original sin position, whereas the lower value suggests a secure position. Nevertheless, Goldstein and Turner (2004) criticize the original sin indices as a measure of currency mismatches on several grounds. First, original sin indicators consider the liability side of the balance sheet effect. In the real-world, both assets and liabilities are used to hedge foreign exchange positions. Second, the original sin framework ignores the essential inputs such as the differences in export openness, the size of the foreign assets, and reserve holdings to assess currency risks. Third, original sin indicators exclude foreign currency assets and receipts over time. Finally, the original sin index does not consider international bonds and bank loans. In order to overcome these limitations, we measure currency mismatches.
2.2 Currency mismatches
In seminal work, Goldstein and Turner (2004) define the currency mismatches as “how the change in the exchange rate will affect the present discounted value of
the future income and expenditure flows.” Currency mismatches refer to the mismatches between assets and liabilities of a country, sector, or firm, in which the liabilities are denominated in foreign currency and assets or revenue in domestic currency. In the event of exchange rate depreciation, the value of liabilities increases, thereby aggravating the currency mismatch problem further. As a result, currency mismatches lead to financial instability in EMEs. Moreover, the exchange rate changes affect the financial position through stock and flow channels. The sensitivity of the balance sheet to changes in the exchange rate is called a ‘stock aspect of currency mismatches.’ On the other hand, the sensitivity of income statements to changes in the exchange rate is known as the ‘flow aspect of currency mismatches.’

2.3 The measurement of currency mismatches

The literature discusses various methods to measure currency mismatches. The earliest measure of currency mismatches in the literature is the original sin hypothesis. Later, Goldstein and Turner (2004) construct an aggregate effective currency mismatch (AECM) index to overcome the drawbacks of original sin indicators. They consider external vulnerability indicators at the aggregate level using the residence principle and include both sides of the balance sheet items. The AECM is calculated as follows:

AECMi,t = NFMCAi,t × FCTDi,t if NFCAi,t > 0

(3)

i,t

AECMi,t = NFXCAi,t × FCTDi,t if NFCAi,t < 0

(4)

i,t

1 3

726

H. Venkatesh, G. S. Hiremath

where NFCA is the net foreign currency assets.2 M and X are the country’s imports and exports of goods and services, respectively; FCTD represents the foreign currency share of the total debt.3 AECM > 0 indicates the net asset position in a country’s foreign currency, whereas the AECM < 0 suggests the net liability position. When foreign currency liabilities are equal to assets of the country i and year t, AECM = 0 (no currency mismatch). The exchange rate depreciation causes a negative balance sheet as well as competitiveness effect when there is a net liability position. On the other hand, the net asset position can have a positive balance sheet and competitiveness effect.
The AECM method, although a sophisticated measure, has some limitations. First, the external vulnerabilities ideally need to be based on currency denomination rather than residence principle since each country’s financial relations are associated with the rest of the world (Levy-Yeyati 2006; Eichengreen et al. 2007; Tobal 2013). Second, the AECM underestimates the balance sheet problem in the case of net asset position in foreign currency. Third, Lane and Shambaugh (2010) argue that AECM neglects the components of capital flows such as foreign direct investment (FDI) and portfolio investment (FPI).4 Hence, this method does not capture the full currency composition of an international balance sheet.
Further, the trade-weighted exchange rate indices are insufficient to understand the financial impact of currency mismatches. Therefore, Lane and Shambaugh (2010) consider the dual role of the exchange rate and its variation in international currency exposure.5 Moreover, AECM covers the internal foreign currency exposure, i.e., one resident to another resident’s bank and bond financing in foreign currency. This index does not include offshore finance vehicles.
Unlike Goldstein and Turner’s (2004) measure of currency mismatch. Kuruc et al. (2016) and Chui et al. (2018) develop a new method to overcome the limitations of ACEM. This method combines two distinct components of currency mismatches:

MRi,t = FXCGTDDPi,t

(5)

i,t

AECMi,t = MRi,t × NGFDCPAi,t

(6)

i,t

2  NFCA > 0 implies net asset position in foreign currency, and net liability position in foreign currency represents NFCA < 0 that leads to currency mismatches. The NFCA consists of "net foreign assets of monetary authorities and deposits of money banks, and foreign currency assets of non-banks held with BIS reporting banks minus foreign currency liabilities of non-banks to BIS reporting banks international debt securities outstanding". 3 The FCTD is comprised of "liabilities of non-banks and non-banks to BIS reporting, domestic credit to the private sector, international and domestic debt securities outstanding". 4  The equity-related instruments like FDI and FPI are excluded in AECM as they may not have the characteristics of FCD and FCA. 5  Lane and Shambaugh’s (2010) innovative contribution is to construct the financial weight of exposure for relevant currency and each country.
1 3

The resurgence of currency mismatches: Emerging market…

727

Table 1  List of selected EMEs

Asia
China Chinese Taipei India Indonesia Malaysia Philippines South Korea Thailand

Central Europe
Czech Republic Hungary Poland

Latin America
Argentina Brazil Chile Colombia Mexico Peru Venezuela

Other EMEs
Russia Israel Turkey South Africa

where MRi,t is the mismatch ratio; FCTDi,t , and XGDPi,t denote foreign currency share of total debt and the ratio of exports of goods and services to GDP of country i and year t. The FCTDi,t is a much broader component than external debt denominated in foreign currency. The NFCAi,t stands for the net foreign currency asset position.6 Kamil (2012) and Montoro and Rojas-Suarez (2012) use the proxies such as a ratio of dollar debt to the sum of exports and dollar assets; and the ratio of foreign currency debt to total debt as a percentage of exports in GDP. However, these currency mismatch ratios do not capture the foreign currency exposure in its entirety.
In a nutshell, the extant literature treats currency mismatch with financial risk indicators like original sin and ratio of broad money to reserves. Moreover, the measurement of currency mismatches is challenging due to the lack of data and a comprehensive method to capture foreign currency exposure. The AECM method is based on the residency principle and covers internal foreign currency exposure. Kuruc et al. (2016) and Chui et al. (2018) extends the AECM methodology. This improved method is not yet empirically tested. Therefore, we analyze the foreign currency exposure and currency mismatches by employing these methods. Besides, we probe the factors responsible for aggregate currency mismatches.
3 The empirical analysis
3.1 Data
We employ a novel method suggested by Kuruc et al. (2016) and Chui et al. (2018) to measure the currency mismatches. Our sample includes 22 EMEs for the period 2008–2018. The sample comprises seven Latin American economies, three Central European countries, eight Asian economies, and four other EMEs (Table 1). The Bank for International Settlements (BIS) publishes statistics on cross-border
6  NFCA computed as "the sum of (i) the net foreign assets of the central banks and other depository corporations plus (ii) non-bank foreign currency cross-border assets with BIS reporting banks minus (iii) non-bank foreign currency cross-border liabilities (excluding debt securities) to BIS reporting banks, minus (iv), non-bank international debt securities outstanding in foreign currency".
1 3

728

H. Venkatesh, G. S. Hiremath

Table 2  Original sin by country groupings (average)

Region

2008–12

2013–18

Latin America

0.85

0.90

Central Europe

0.63

0.84

Asia

0.73

0.72

Other EMEs

0.57

0.66

Source: Authors’ own calculations based on BIS data
This table shows the average original sin index values computed for 20 EMEs during 2008–2018. The original sin index value ranges between zero and one. The value closer to one indicates severe original sin problem implying a higher currency risk

banking flows and debt securities of few countries. Therefore, our sample is confined to 22 EMEs. The analysis of foreign currency exposure and currency mismatches assumes importance as many EMEs are under financial stress. The COVID19 pandemic is threatening the financial stability of these economies. Moreover, currency depreciation of EMEs against the US dollar increases external debt and rollover crisis. The present study covers the recent period, which helps policymakers devise a policy framework to deal with financial stress and the crisis.
3.2 Measure of original sin (OSIN)
We compute the original sin index (Eq. 2) using a dataset on international debt securities and loans, but the index does not include cross-border bank loans. Original sin is primarily due to incompleteness in financial markets, credit ratings, and volatile capital flows. The exchange rate depreciation against the dollar tends to increase the value of the original sin and make it harder to repay the debt denominated in foreign currency. The average OSIN index indicates higher levels of original sin in the Latin American economies than that of other regions. The value of the sin increased from 0.85 to 0.90 between 2008–2012 and 2013–2018 (Table 2). Except for Asia, the average original sin rose in all the regions due to financial stress during 2008–2018. The taper tantrum crisis, China’s slowdown, trade wars, and lira/peso crisis were responsible for such stress of assets. Therefore, we can observe a clear difference in OSIN values between 2008–2012 and 2013–2018.
Further, the average original sin value at the country level presented in Fig. 2 shows a greater extent of original sin in Argentina, Chile, and Venezuela (OSIN value is close to one). Such a level of sin implies a high degree of foreign currency risk. Similarly, Hungary, Chinese Taipei, Malaysia, the Philippines, and Israel suffer from the original sin problem. Although the Czech Republic had no original sin risk during 2008–2009, the level of sin increased to 0.73 during later years. This rise in the Czech Republic’s vulnerability resulted from an increase in Euro-denominated debt to 22.2 percent of total GDP in 2018 (IIF 2019). Overall, the problem of original sin increased in all EMEs after the taper tantrum crisis during the year 2013.
In the Asiatic region, China and India have a lower value of original sin compared to other EMEs. India reduced the OSIN from 0.98 to 0.54 between 2008
1 3

The resurgence of currency mismatches: Emerging market…

729

Fig. 2  Original sin by country wise. Notes: The original sin index values are based on the sample of 20 EMEs for the period 2008 to 2018. The details of countries are also provided in Table 1. The original sin index value ranges between zero and one. The index value closer to one indicates a severe original sin problem. Such sin implies a higher currency risk. Source: Authors’ own calculations based on the Bank for International Settlements (BIS) database on international debt securities. The annual average is computed from the quarterly data
and 2018 (Fig. 2). The issuance of rupee-denominated bonds – masala bonds in the international market is one of the policy measures that reduced the debt. The OSIN in China was reduced to zero, but it rose to 0.47 after 2015, leading to the saucer-shape curve. China started the renminbi-denominated bond market, popularly known as dim sum bonds. These bond issuances substantially increased after 2010 and reached their highest level in 2014. In 2015 and later periods, the dim sum bond issuances dropped drastically due to currency devaluation, financial instability, and trade tensions (Kohli et al. 2017).
Interestingly, South Africa has the lowest OSIN values among all sample countries, and the country cautiously dealt with the foreign currency risk by issuing debt in rand-denominated bonds. South Africa borrowed only 5.5 percent of gross government debt in foreign currency. The sophisticated financial markets attracted foreign investors to local bond markets, which led to 37 percent of investment in government bonds. These markets acted as a bulwark for South Africa. When the rand depreciates against the dollar, the government need not pay foreign debtors.
Moreover, South Africa has managed the sovereign debt levels effectively than its counterparts. The country met the financial needs sustainably, using deep and liquid financial markets. With these favorable conditions, a country can easily follow the floating exchange rate regime that increases export competitiveness during the depreciation period.
1 3

730

H. Venkatesh, G. S. Hiremath

Fig. 3  Foreign currency shares of total debt outstanding (FCTD). Data Sources: Bank for International Settlements (BIS) locational banking statistics and debt securities statistics; International Monetary Fund (IMF)
3.3 Foreign currency share in total debt outstanding (FCTD)
FCTD is a comprehensive measure that includes the country’s total liability denominated in foreign currency. Moreover, this indicator is broader than external debt denominated in foreign currency. We compute the first indicator of currency mismatch – FCTD by assuming the denomination of domestic bonds and bank loans exclusively in local currency. We observe that the value of FCTD is lower than the international debt securities and cross-border bank loans in EMEs (Fig. 3). The FCTD increased after the GFCs and peaked in 2015. The FCTD is higher in the Latin American region: Argentina, Mexico, and Colombia raised FCTD to 43.7, 25.1, and 20.5 percent in 2017, respectively. On the other hand, Central Europe has moderately reduced FCTD between 2008 and 2017.
Similarly, the share of foreign currency in total debt in Asian economies such as China, Chinese Taipei, India, Malaysia, Russia, South Korea, and Thailand is lower than that of their peers. For example, India and Russia slashed their share of foreign currency debt from 10.5 to 5.7 percent and 31 to 12.4 percent, respectively. Further, Venezuela had successfully reduced FCTD to 0.3 percent in 2017 from the highest 45.5 percent in 2010.
Among the Latin American countries, Brazil has the lowest share of foreign currency in its total debt and remained flat over the period (7.2 and 6.2 percent in 2008 and 2017). To reduce the FCD, Brazil has started to issue real denominated bonds in the domestic market. Brazil has issued $7.08 billion worth of local currency bonds
1 3