Emerging derivatives markets?

Only 10% of global derivatives turnover is in contracts denominated in the currency of an emerging market economy (EME), much lower than the share of these economies in global GDP or world trade. Derivatives in EME currencies also tend to be less complex and more likely to be traded outside the home economy than those in advanced economy currencies. Differences persist even if we control for key drivers of derivatives turnover such as the size of the bond market, the openness of the capital account, the amount of foreign trade and the size of external liabilities. Instead, the small size of EME derivatives markets appears to reflect differences in per capita income. Large external asset holdings by residents of a country go hand in hand with lower turnover, perhaps because they are used as a hedge against country risk.¹

JEL classification: F31, G12, G23.

The economies and financial markets of emerging market economies (EMEs) tend (with some exceptions) to be more volatile than those of advanced economies. This is true whether one looks at output growth, exchange rates, interest rates or capital flows. Given this volatility, one would expect hedging markets in EMEs to be well developed. But this does not seem to be the case. EMEs make up about one third of the global economy when measured at market exchange rates and just under one half when measured at purchasing power parity. Their share in global trade is 36%. Still, derivatives referencing their currencies or interest rates account for only 10% of the global turnover of such contracts, despite notable growth in some cases in recent years.

But the desire to hedge need not result in liquid markets unless there is somebody willing to take the other side. For idiosyncratic risks that are easily diversifiable, a financial intermediary could sell the appropriate insurance. But creating a market for macroeconomic risks, such as exchange rate and interest rate risk, is much more difficult unless there are agents who are exposed to such risk in opposite ways.

Consider the example of an EME with a volatile exchange rate. Residents and foreign investors alike will probably be interested in protecting themselves against a depreciation in the exchange rate, but it is less obvious who would be willing to take the other side. Of course, foreign financial institutions with no particular interest in gaining exposure to the currency could offer such protection, but they would probably charge a very high price. More attractive terms could be obtained from market participants wanting to protect themselves against an appreciation - for instance, domestic holders of foreign assets or foreigners issuing domestic currency debt.

This article uses the latest vintage of the BIS Triennial Central Bank Survey to examine how derivatives trading on EME exchange rates and interest rates has evolved and why it remains lower than for advanced economies. Our focus is on the ability to trade EME risks rather than on trading derivatives locally, so we focus on derivatives on EME currencies and interest rates (including those traded offshore) and not just on contracts traded domestically in EMEs. Given the focus of the Triennial Survey, we limit ourselves to exchange rate and interest rate contracts and do not discuss other risk categories such as equities and commodities, despite their importance for both EME residents and foreign investors.

The article is divided into two parts. In the first part, we compare derivatives markets in EMEs with those in the advanced economies using data from the Triennial Survey. In the second part, we look at the factors that explain cross-country differences in the development of these markets.

Derivatives markets in emerging and advanced economies

The sharp movements in exchange rates during the past few years masked a solid increase in activity in EME derivatives. Average daily turnover in FX and interest rate derivatives denominated in EME currencies stood at $0.8 trillion in April 2016, almost the same as in April 2013, on the cusp of the bout of sharp market moves that started a month later (the so-called taper tantrum).² But this apparent stagnation paints a misleading picture of the underlying dynamics of the market, as the sharp depreciation of many EME exchange rates between the last two Triennial Surveys depressed volumes measured in US dollars. Measured in local currencies, turnover actually expanded by 25%, albeit with large differences across currencies (Graph 1).

The size and depth of derivatives markets vary considerably across currencies. The largest and most rapidly growing market is the one for contracts on the Chinese exchange rate and Chinese interest rates. Its activity grew by 50% between the 2013 and 2016 surveys (55% when measured in local currency) to reach $150 billion a day in April 2016. This makes the Chinese renminbi the eighth most traded currency in the global derivatives market in 2016, overtaking the Brazilian real ($108 billion), Korean won ($83 billion) and Mexican peso ($81 billion). The rapid growth of the renminbi market is part of a broader internationalisation of the currency (see "The rise and financialisation of the renminbi" in Moore et al (2016), and Box A). Other EME currencies with significant market activity are the Turkish lira, the Indian rupee and the South African rand with just over $50 billion turnover a day each.

Large as they may seem, these numbers are relatively modest when set against GDP or proxies for hedging needs, as measured by international trade or investment. On all those measures, even EMEs with relatively developed and deep derivatives markets sustain lower turnover than most advanced economies. Median EME derivatives turnover is roughly 5% of annual GDP, well below the median for advanced economies, which is close to 20%. Even advanced economies that are neither financial centres nor issuers of a reserve currency, such as Australia, Canada, New Zealand, Norway and Sweden, have much higher turnover-to-GDP ratios than any EME (Graph 2, top panel). The only EME currencies coming close to the advanced economy median are the South African rand and the Hungarian forint, with ratios of 15-20%. The picture does not change much when derivatives turnover (FX contracts only, as interest rate contracts are unlikely to be used in this context) is set against foreign trade (Graph 2, middle panel). On both measures, CNY contracts rank close to the bottom of the distribution. When derivatives volumes are compared with cross-border investment positions, the differences between advanced and emerging economies are less stark (Graph 2, bottom panel).

The low ratios of FX derivatives turnover to foreign trade in EMEs could indicate that exporters and importers in these countries still find it relatively difficult to hedge the exchange rate risk associated with their transactions, although one cannot say this for sure in the absence of detailed data on their use of derivatives.

Derivatives trading in EMEs is heavily geared towards FX contracts, which account for 75-90% of turnover in most cases (Graph 3, first panel). FX contracts also outweigh interest rate contracts in advanced economies, though by a smaller margin. Currencies with relatively low FX shares are the Brazilian real (with a share of 56%), Hungarian forint (59%), Chilean peso (66%), South African rand (67%) and Mexican peso (68%).

EME derivatives markets are not only smaller than their advanced economy counterparts, they also feature less complex contracts. At least this is the message from the relatively coarse instrument breakdown of the Triennial Survey. More sophisticated contracts, such as cross-currency swaps, FX options, interest rate swaps and interest rate options, account for a lower share in total turnover in emerging than in advanced economy derivatives markets (Graph 3, second panel). But there are signs that such contracts are becoming more common in some EMEs. For example, Mexico and South Africa have comparatively liquid cross-currency swap markets. In the interest rate segment, many economies have developed active over-the-counter (OTC) markets for options.

Another feature that sets some, albeit not all, EME derivatives markets apart is the prevalence of non-deliverable contracts. Three quarters of Brazilian real FX forwards and swaps and two thirds of Korean won contracts are non-deliverable in the sense that they are settled in a single currency. The other markets with a significant share of non-deliverable contracts are the New Taiwan dollar (56%) and the Indian rupee (47%). The prevalence of non-deliverable contracts reflects the presence of restrictions on currency convertibility, which hinder the development of contracts settled in two currencies. That said, non-deliverable markets may continue to exist well after controls have been lifted. Developments in the non-deliverable FX market are discussed in McCauley and Shu (2016); for background on the Brazilian market, see Box B.

With some notable exceptions, EME FX and interest rate derivatives are mostly traded outside the home economy, usually in a major global financial centre such as New York, London, Hong Kong and Singapore (Graph 3, third panel). Very few EMEs have onshore markets that come close, in terms of activity, to the offshore market. Only in the Brazilian real, Russian rouble and Korean won does onshore OTC turnover exceed offshore turnover, although the Indian rupee, Israeli new shekel and South African rand come close. Offshore trading tends to take place in the same time zone. For instance, Latin American currencies tend to be traded in New York, central and eastern European currencies in London, and Asian currencies in Hong Kong and Singapore. Box A compares on- and offshore renminbi markets.

Finally, EME derivatives are disproportionately traded over the counter (Graph 3, fourth panel). OTC trading exceeds exchange-based trading by a factor of three, with Brazil being the only major market in which derivatives exchanges account for a large fraction of turnover (Box B). Many other EMEs have active futures exchanges, but these tend to concentrate on stock index and commodity contracts rather than FX and interest rate instruments. For example, only 2.5 million of the close to 1 billion contracts traded on Chinese derivatives exchanges in the first quarter of 2016 referenced interest rates. According to Euromoney TRADEDATA, there are no listed contracts on exchange rates in China. And even in Brazil, equity contracts account for 50% of total turnover (measured by the number of contracts); interest rate contracts follow with 28% of turnover and FX contracts with 11%.

Why is EME derivatives turnover so low?

What explains the comparatively low level of activity in EME currency derivatives? In the introduction, we argued that the development of a liquid derivatives market - of any market, in fact - requires the existence of regular two-way flows. So, what explains the existence of such flows, and thus the size of the market? In this section, we aim to address these questions by estimating the determinants of activity in derivatives markets across a wide range of economies - emerging and advanced - on proxies for the hedging needs of different investor types. Our sample comprises 34 economies (24 emerging and 10 advanced economies)³ and the data from the Triennial Survey responses in April 2013 and April 2016. Unfortunately, the incomplete currency breakdown in earlier surveys precludes us from going back further.

The first, and most obvious, hedging motive derives from international trade, which we proxy by summing the exports and imports of the home economy.

The second motive concerns the hedging of financial exposures. We proxy those exposures by the size of the bond market in domestic currency. Investors in this market may want to hedge the interest rate risk associated with these bonds, so we would expect a larger market to go hand in hand with higher turnover in the interest rate segment. Foreign investors in particular may also want to hedge the currency risk associated with their bond holdings, so we would expect bond market size to be related to turnover in the FX segment too. Unfortunately, we do not have data on the currency composition of domestically issued bonds, so we limit our attention to international issues. Since investors, even those who hold securities until maturity, tend to follow overlay strategies when hedging interest rate and currency risks, we look at amounts outstanding rather than new issuance.

The third motive is related to the level of integration with global markets as measured by the size of external assets and liabilities. We would expect high external liabilities to be associated with a high demand for (FX) hedging instruments on the part of domestic borrowers and thus larger turnover. The impact of large external assets on turnover is more ambiguous. On the one hand, the holders of these assets may want to hedge the associated exchange rate risk in derivatives markets,⁴ thus providing a natural counterpart to external borrowers. On the other hand, external assets may provide a natural hedge against currency fluctuations, thus dampening derivatives turnover. Since these effects centre on FX exposures, ideally we would like to separate each of these variables further, into domestic currency- and foreign currency-denominated assets and liabilities. Unfortunately, data availability precludes us from doing so for a large enough sample of countries.

The fourth financial variable is the gross issuance of domestic currency bonds by non-residents. Several international financial institutions, such as the World Bank and the European Investment Bank, as well as other public and private entities issue bonds in EME currencies and swap the proceeds into their base currency. In contrast to the previous variables, this one does not refer to a stock but to a flow. This is because associated derivatives trades tend to be one-off transactions, often with tenors equal to the maturity of the underlying bond. We proxy these hedging needs with the gross issuance of local currency bonds by non-residents in the first half of the year.

In addition, we include measures for capital account controls and restrictions on cross-border derivatives transactions. Both are taken from Fernández et al (2015) and correspond to de jure restrictions, irrespective of how tightly they are enforced. Finally, we include GDP per capita of the home economy to control for income level.

Simple bivariate regressions suggest that per capita GDP, capital account openness and bond market size are better predictors of market development than trade (Graph 4). This is in contrast with earlier work by Mihaljek and Packer (2010), who found a close association between (FX) derivatives turnover and trade. Non-resident bond issuance and large external assets and liabilities are also correlated with higher turnover, although their explanatory power is weaker.

These results are largely confirmed by multivariate regressions (Table 1). International trade remains insignificant regardless of the specification. By contrast, many of the financial variables turn out to be highly significant.

The total stock of bonds outstanding in a particular currency is an important driver of both interest rate and total derivatives turnover but is insignificant for FX turnover. Raising the stock of bonds outstanding to GDP by 1 percentage point is on average associated with a 0.3 percentage point higher ratio of derivatives turnover to GDP.

Large external liabilities are associated with significantly higher turnover in both FX and total derivatives, although the elasticity is smaller than for bond market size. High non-resident issuance in domestic currency is also associated with higher derivatives turnover, especially in the interest rate segment. Finally, economies with large external asset holdings tend to have lower derivatives turnover.

It is important to note that our results reflect correlations, not necessarily causality. While we measure all variables (except non-resident issuance) during or at the end of the previous year, this does not entirely eliminate endogeneity since the variables are mostly quite persistent. For instance, the negative coefficient on external assets could indicate that holders of external assets do not need to hedge their exposure to foreign currency; or, quite the contrary, that the limited ability to hedge through derivatives makes residents hold more external foreign currency assets. But in either case, there is little support for the notion that residents with external asset holdings provide a natural counterparty to those wanting to insure against currency fluctuations. The experience of Chile, where pension funds convert their external asset holdings into domestic currency and, in the process, serve as natural counterparts to residents wanting to hedge their foreign currency liabilities, appears to be the exception rather than the rule.

Controlling for such financial factors, the differences between advanced and emerging economy derivatives markets appear to reflect mainly differences in income. The level of per capita GDP is highly significant, both statistically and economically, whereas a dummy variable that takes the value of one if a country is an EME and zero otherwise turns out to be insignificant. Variables measuring capital account controls and restrictions on cross-border derivatives transactions also turn out to be insignificant. It is worthy of note that GDP per capita is significant for determining only turnover in the foreign exchange market, not that of interest rate contracts.⁵ This is quite puzzling because only relatively rich economies have sizeable interest rate derivatives markets.

Conclusions

Derivatives markets for EME currencies and interest rates tend to be much smaller than their advanced economy counterparts, in relation to both GDP and trade. EME derivatives markets are also limited to a narrower set of instruments, and a disproportionate part of trading takes place over the counter. While simple activity indicators such as the ones used in this article are only loosely related to market liquidity or trading costs, especially in the short term, there is reason to believe that residents of and investors in EMEs find it more difficult and more costly to hedge their exposures than their peers in advanced economies.

Our regression results suggest that the weaker turnover in EME reflects several factors, including a lower level of financial development (as measured by the size of the bond market in domestic currency), less integration in the global economy (as measured by the size of international liabilities) and lower per capita income. They also suggest that EME residents use their external asset holdings as a substitute for derivatives to hedge currency risk.

References

Avalos, F and R Moreno (2013): "Hedging in derivatives markets: the experience of Chile", BIS Quarterly Review, March, pp 53-63.

Bank for International Settlements (2016): "Glossary", BIS Quarterly Review, December, pp 27-33.

Chinn, M and H Ito (2008): "A new measure of financial openness", Journal of Comparative Policy Analysis: Research and Practice, vol 10, issue 3, pp 309-22.

Dodd, R and S Griffith-Jones (2007): "Brazil's derivatives markets: hedging, central bank intervention and regulation", Informes y Estudios Especiales, no 20, United Nations Publications, December.

Ehlers, T and E Eren (2016): "The changing shape of interest rate derivatives markets", BIS Quarterly Review, December, pp 53-65.

Fernández, A, M Klein, A Rebucci, M Schindler and M Uribe (2015): "Capital control measures: a new dataset", NBER Working Papers, no 20970, February.

Garcia, M, M Medeiros and F Santos (2014): "Price discovery in Brazilian FX markets", Textos para discussão, no 622, Department of Economics PUC-Rio (Brazil), May.

Garcia, M and T Volpon, (2014): "DNDFS: a more efficient way to intervene in FX markets?", Working Paper no 501, Center for International Development - Stanford University, May.

Kohlscheen, E and S Andrade (2014): "Official FX interventions through derivatives", Journal of International Money and Finance, 47, pp 202-16.

McCauley, R and M Scatigna (2011): "Foreign exchange trading in emerging currencies: more financial, more offshore", BIS Quarterly Review, March, pp 67-75.

--- (2013): "Currencies getting more financial", BIS Quarterly Review, December, p 25.

McCauley, R and C Shu (2016): "Non deliverable forwards: impact of currency internationalisation and derivatives reform", BIS Quarterly Review, December, pp 81-93.

Mihaljek, D and F Packer (2010): "Derivatives in emerging markets", BIS Quarterly Review, December, pp 43-58.

Moore, M, A Schrimpf and V Sushko (2016): "Downsized FX markets: causes and implications", BIS Quarterly Review, December, pp 35-51.

¹ The views expressed here are those of the authors and do not necessarily reflect those of the BIS. We are grateful to Claudio Borio, Ben Cohen, Ramon Moreno, Hyun Song Shin and Phil Wooldridge for helpful comments and suggestions. Diego Urbina provided excellent research support; we are grateful for his patience and dedication. The authors of Box A would like to acknowledge the help of Steven Kong, Jimmy Shek, Tsvetana Spasova and Agne Subelyte. See the Glossary (BIS (2016)) for definitions of relevant technical terms.

² Includes all contracts in the Argentine peso (ARS), Bulgarian lev (BGN), Bahraini dinar (BHD), Brazilian real (BRL), Chilean peso (CLP), Chinese renminbi (CNY), Colombian peso (COP), Czech koruna (CZK), Hungarian forint (HUF), Indonesian rupiah (IDR), Israeli new shekel (ILS), Indian rupee (INR), Korean won (KRW), Mexican peso (MXN), Malaysian ringgit (MYR), Peruvian sol (PEN), Philippine peso (PHP), Polish zloty (PLN), Romanian leu (RON), Russian rouble (RUB), Saudi riyal (SAR), Thai baht (THB), Turkish lira (TRY), New Taiwan dollar (TWD) and South African rand (ZAR). Data are subject to change.

³ The number of economies is constrained by the availability of data on international investment positions.

⁴ This might be because of having liabilities in domestic currency. Such a mechanism appears to be in place in Chile, where domestic pension funds hedge their foreign currency assets back into domestic currency (Avalos and Moreno (2013)).

⁵ McCauley and Scatigna (2011, 2013) find that FX turnover (including both spot and derivatives) increases with GDP per capita.