Marketing material for professional and advisers only

Why investors are wrong about the role of the dollar July 2018

It is widely accepted that dollar strength is bad for emerging Craig Botham Emerging Market markets, something that seems to have been confirmed by a Economist very tough 2018 for emerging market (EM) assets. However, the traditional reasons why, such as investors’ attitude to risk (“risk sentiment”) or prevailing financial conditions, are increasingly being questioned. A growing body of evidence suggests that the dollar more than any other currency dominates trade invoicing and this is challenging the way economists and markets have previously thought about currencies and their impact on economic and financial variables. We argue in this paper that this new model provides a more complete picture of how and why the dollar is so important to emerging markets, and why it has a greater impact on global trade and inflation than is generally appreciated. For investors, this has implications across all EM asset classes.

Introduction A resurgent dollar in 2018 has piled particular pressure Sometimes the link is explained as a question of risk on emerging market assets. Historically, dollar strength sentiment: in a risk-off environment, capital flows to the is associated with the relative underperformance of EM safe haven of the US dollar, and out of riskier emerging equities compared to their developed market peers, as well markets, reversing direction when investors’ risk appetite as widening EM bond spreads and weaker EM currencies returns. This may be part of the story, but is rather (Figure 1). While this relationship is widely accepted, less unsatisfying because it relies on ambiguously determined often discussed is why it should be so. In some ways it “animal spirits” rather than anything that could be forecast. seems counterintuitive: a stronger dollar should mean that foreign goods become cheaper in the US, boosting EM exports and growth, and so providing a tailwind for EM corporate earnings.

Figure 1: EM assets tend to track the dollar 70 180 120 1000

115 80 160 110 800

140 105 90 100 600 120 100 95 100 90 400 110 80 85

80 200 120 60 75 130 40 70 0 98 00 02 04 06 08 10 12 14 16 18 00 02 04 06 08 10 12 14 16 18 Dollar index* (inverted) MSCI EM Index/World Index (rhs) Dollar index* JP Morgan Emerging Market Bond Index spread (rhs)

Past performance is not a guide to future performance and may not be repeated.

*This is the DXY index, a measure of the value of the US dollar relative to a basket of trade partner currencies. Source: Schroders Economics Group and Thomson Datastream, as at 12 June 2018 1 Another explanation is that, since the dollar typically This would suggest, on an empirical basis, that it is wrong strengthens as interest rates rise, it is the higher money costs to think about the world using the Mundell Fleming rather than the higher currency that is the real cause of EM mindset. In the extreme, the euro-yen pain. This makes sense: higher US rates draw capital from the should have no impact on European-Japanese trade if all rest of the world, buoying the dollar and weighing on other trade goods are priced in dollars. Building on this, Gopinath currencies. It certainly seems to fit the story this time round, suggests a new “dominant currency paradigm” (DCP) with US rate differentials versus the rest of the world picking model, in which import prices are driven by the currency up. Again though, if the US Federal Reserve (Fed) is becoming most used for invoicing imports and exports (i.e. the more hawkish it is often because the US economy is gaining “dominant” currency) and not the trade partner currency. strength, which should be a positive for global demand. As emerging markets are typically a high play on global Choosing the right model growth, it seems odd that this is negative for EM assets. To see which model is a better approximation of the real world, we can look at how the prices of tradable goods Perhaps it could be argued that risk sentiment and rate react to changes in exchange rates. If the Mundell Fleming differentials between them can explain the relationships we model is correct, we should find that import prices in most see. However, we think that we can suggest a more directly economies move in line with the bilateral (trade partner) causal relationship, where it is the dollar strength itself exchange rate. If the DCP model is correct, import prices that causes problems for other economies, particularly will move with the dollar exchange rate independently of emerging markets. the trade partner exchange rate.

What we think we know about currencies: Happily, Boz et al (2017)2 have done this work for us. textbook theory Specifically, they find that across a large panel data set the To begin, we should quickly recap how economists dollar exchange rate dominates the bilateral exchange and markets normally view exchange rates. The most rate in terms of transmitting inflation and determining common way of thinking about the impact of currencies trade volumes. The DCP model turns out to be a better on macroeconomics is the Mundell Fleming model, or approximation of the world today than the Mundell Fleming producer currency pricing. In this model, exporters sell model. This suggests we need to reframe our thinking. goods in their own currency. Consequently, a weaker currency leads to cheaper exports and dearer imports. Practical applications I: why do we care? This causes a reduction in imports, an increase in exports, We have been trying to find a causal link between dollar and more consumption of domestic goods. Growth and strength and emerging market underperformance. When inflation then both increase. it comes to EM equities, a key driver for earnings is export performance. Normally, we would expect weaker EM This view informs the way we think about monetary policy currencies, on a trade-weighted basis, to boost exports and, as a result, investing. Economists tend to focus on how and thereby support equities. But in the DCP model, the currencies move in trade-weighted terms as the Mundell bilateral exchange rate against the dollar dominates the Fleming model implies that these are what matters for the trade partner exchange rate when it comes to movements trade balance and inflation. In these terms, a sharp move in in trade volumes and prices. When investing, this model the dollar can be offset by moves in other major currencies suggests that we should focus on the exchange rate versus like the euro or yen, or the trade- weighted basket as a the dollar rather than versus the trade partner, or the whole. However, evidence increasingly suggests that this trade-weighted exchange rate. may not be an accurate picture of what is happening. However, remember that under this view imports and What we know about currencies: market practice exports are all priced in dollars. So while we need to focus A series of academic papers have looked at whether on the dollar exchange rate, we also need to change how exporters really do invoice goods in their own currencies. we think about it. Dollar appreciation makes imports more The chief finding, including the most recent by Gopinath expensive, as before, but it does not make exports cheaper, (2015)1, is that actually many exporters choose to price in because our trade partner is also using dollar pricing. a third country’s currency. That is, they price in a currency Dollar strength now means imports are more expensive for that neither they nor their trade partner uses. everyone (except the US), and no one receives a competitive advantage for their exports, even though their currencies First and foremost among these “third country currencies” are weaker against the dollar. This would imply that a is the dollar. This dominance is above and beyond the stronger dollar can weaken trade activity and vice versa. dominance of the US in global demand: dollar invoicing is 4.7 times the US share of global imports, according to As with all models, DCP is a simplification, and it is Gopinath. The comparable figure for the euro is just 1.2 recognised as such. Yet it does seem to be a better times. Nor is this dominance limited to emerging markets: approximation of the world than previous models, as the high dollar invoicing is also found in Japan, France, and paper’s results indicate. We can see this for ourselves in other developed markets. Strikingly, Gopinath found that, Figure 2, which compares the dollar index to the global for many countries, the share of their own currency in their trade multiplier, which predicts how trade growth will own trade is close to zero. translate into global growth. To quantify this, Boz et al (2017) found that a 1% dollar appreciation leads to a 0.6- Simply put, countries are trading with each other using 0.8% decline in global trade volumes within a year. dollars as the invoicing currency, even when neither country uses the dollar as a domestic medium of exchange.

1 Gita Gopinath, ”The international price system”, Jackson Hole Symposium, 27, 2015, Federal Reserve Bank at Kansas City. 2 2 Emine Boz, Gita Gopinath and Mikkel Plagbord-Møller, “Global Trade and the Dollar”, IMF Working Paper 17/239, November 2017. Figure 2: The dollar is a reasonable predictor of This relationship would explain EM underperformance trade volumes during periods of dollar strength, as it leads to weaker

ear on year ear on year earnings growth in the year ahead. It also suggests that 3 -20 the benefits to growth from currency weakness are much smaller than commonly assumed (and minimal in the -15 extreme cases where all goods are invoiced in dollars). 2.5 It implies that central banks should intervene more -10 aggressively against currency weakness, particularly as 2 -5 the dollar exchange rate also dominates price effects. If a weaker currency delivers more inflation (because imports 0 become more expensive) and no growth (because there 1.5 5 is no boost to exports), then a central bank targeting inflation, or even growth and inflation, has an incentive to 1 10 lean against depreciation by hiking rates. As investors, this matters because we may be misinterpreting the effects of 15 0.5 central bank activities, and in doing so placing erroneous 20 bets on central bank and market rate moves.

0 25 02 03 04 05 06 07 A dominant dollar increases the incentive to manage Global trade multiplier Dollar index* 12m lag (rhs) the currency It is hardly a secret that many EM central banks intervene ear on year ear on year to calm in their exchange rates, even if this is 2 -20 not always made explicit. According to the International Monetary Fund’s Annual Report on Exchange Arrangements 1.8 -15 and Exchange Restrictions 20163, there are currently 39

1.6 -10 countries following an explicit dollar peg, with a further 40 following a “managed float”, in which the currency is 1.4 -5 free to move with the market but is subject to central bank 1.2 intervention. Perhaps one way to see this in emerging 0 markets is to look at the behaviour of central bank foreign 1.0 5 exchange reserves in the face of dollar moves (Figure 0.8 3). Dollar weakness tends to be associated with reserve 10 0.6 growth, and dollar strength with reserve shrinkage; this 15 is consistent with central banks buying dollars to prevent 0.4 domestic currency appreciation when the dollar weakens, 0.2 20 and selling them in periods of dollar strength to fight 0 25 depreciation. 11 12 13 14 15 16 17 18 In buying dollars from the market, the central bank injects Global trade multiplier Dollar index* 12m lag (rhs) domestic currency, boosting system liquidity. The opposite *This is the DXY index, a measure of the value of the US dollar relative to a basket is true when dollars are sold. In this way dollar weakness of trade partner currencies. Crisis period removed for reasons of scaling. Source: and dollar strength lead to domestic monetary policy Schroders Economics Group and Thomson Datastream, 12 June 2018. easing and tightening respectively in ways which may not be reflected by the headline policy rate on its own.

Figure 3: Dollar strength tracks reserve accumulation quite closely

Month-on-month yearly Month-on-month yearly moving average -2.0% 3.5%

-1.5% 3.0%

-1.0% 2.5% 2.0% -0.5% 1.5% 0.0% 1.0% 0.5% 0.5% 1.0% 0.0% 1.5% -0.5%

2.0% -1.0% 2.5% -1.5% 00 01 02 03 04 05 06 07 08 09 10 11 12 13 14 15 16 17 18 Dollar index* EM foreign exchange reserves

*This is the DXY index, a measure of the value of the US dollar relative to a basket of trade partner currencies. Source: Schroders Economics Group and Thomson Datastream, 12 June 2018

3 3 International Monetary Fund, Washington, October 2016. This effect is known and understood. However, if the DCP of a single dominant currency in both trade and finance. model is the more accurate one, then it should mean That is, the dominance of the dollar in trade can lead to its that dollar moves will drive domestic liquidity conditions dominance in the global financial system, edging out the in emerging markets by exerting pressure on the central currencies of other major economies, and even leading to bank to intervene. As discussed above, if local currency some use of dollar credit within non-dollar economies. weakness yields no growth benefits, and higher inflation costs than commonly assumed, central banks will want To see how this happens it is worth considering why firms to fight it. By the same token, local currency strength may wish to invoice in a foreign currency. We will use should have little negative effect on growth, while reducing dollars for ease of reference, but the same could apply to inflation. This suggests an asymmetric response from the euro or RMB. those central banks whose aim is purely to target inflation: Faced with a large trade partner pricing in dollars, dollar weakness should prompt less intervention than importers will wish to hold any cash reserves as dollars dollar strength, because the costs of dollar strength are to provide the best hedge against exchange rate moves. higher than those of dollar weakness. If we expect periods This creates demand for dollar deposits, which domestic of dollar strength, we should also expect EM liquidity to banks will seek to provide. To do this, however, banks must tighten, and market rates to rise. It is not immediately clear seek collateral for these dollar liabilities, and so will offer that dollar weakness should see the opposite effect. loans in dollars to other customers. As these customers However, the chart in Figure 3 does not seem to suggest will not have access to dollar revenues, they must be this is what happens. Central banks seem to buy dollars on compensated for the exchange rate risk via a lower interest the currency’s weakness as much as they sell them when it rate. (This argument has a certain appeal, as it provides an strengthens. The simplest explanation consistent with the explanation for the “exorbitant privilege” enjoyed by the DCP model is that EM central banks are not pure inflation US, whereby the US is able to borrow more cheaply than targeters. When their own currency strengthens against other countries.) The effect is to reinforce the dominance of the dollar, reducing inflation but having no effect on the dollar in trade invoicing: any exporter taking advantage growth, they are happy to inject liquidity to provide a boost of the lower interest rates offered on dollar financing has a to growth, because the currency has bought them space on large incentive to price their goods in dollars to reduce the the inflation side. currency risk.

It could also be the case that central banks are not aware This is not to say that all financing in the economy becomes of the invoicing issue, in which case their policy response dollar denominated, but that dollar financing becomes may simply be mistaken. This could mean they intervene more prominent than would be expected, given the share less against dollar strength than they should, necessitating of dollar revenues or dollar expenditures in the economy. rate hikes to keep inflation under control, while intervening It also means that foreign currency reserve holdings too much against dollar weakness, and so ultimately being should be skewed towards dollars, rather than euros, yen forced to suspend an easing cycle. or renminbi, including at the central bank, given its need to provide support in times of stress. The data would Interaction with banking and finance generally seem to support this: work done by the European Central Bank (ECB) in 20174 found that 64% of worldwide Looking more broadly, the next consideration is whether official foreign exchange reserves were held in dollars, the dollar’s dominance is limited to trade, or if there are against 20% in euros and 4% in yen. Meanwhile, data from other economic channels that may be affected. As part of the Bank for International Settlements (BIS) suggest that the DCP model, Gopinath argues that a currency’s role as around 49% of all cross-border bank claims, and 51% of a unit of account for invoicing is complementary to its role cross border bank liabilities, are in dollars, with similar as a safe store of value, which can lead to the emergence numbers for the non-bank private sector.

Figure 4: Dollar dominates many EM credit systems

Foreign currency debt as share of total, private sector 70

60

50

40

30

20

10

0 Mexico Turkey Hungary Indonesia Chile Malaysia Colombia South Poland Brazil Russia India South Thailand China Africa Korea Dollar* Other* Total

Source: Schroders Economics Group and Thomson Datastream, 12 June 2018

4 4 ECB staff, The International role of the euro, Technical Report, European Central Bank, July 2017. For the EM private sector, foreign currency debt seems to with faster credit growth. The literature lends some weight constitute 15-40% of total debt, though China at 4% and to this argument: Brauning & Ivashina (2017)5 found a Mexico at 60% are outliers. Most of this foreign currency typical US easing cycle sees a 32% loan volume increase debt seems likely to be in dollars, as separate BIS data in emerging markets, with a similarly large reverse effect place the dollar share of total EM debt at around 80% when policy tightens. (around 57% for developing Europe). Figure 4 provides a breakdown of debt in some of the more important The linkages we have explored can also offer insight into emerging markets. a possible connection between the dollar and EM debt spreads, as measured by the JP Morgan Emerging Market Practical implications II: the dollar and EM credit Bond Index. If a stronger dollar is generally associated with tougher credit conditions, it makes sense that investors So what are the implications for emerging markets? We would demand a higher risk premium for holding EM have touched already on how dollar strength can lead to credits over the safer US alternative. credit tightening in emerging markets as central banks sell reserves to reduce depreciation pressures, or even Figure 5: Dollar strength weighs on EM credit growth hike rates to defend their currency. We can now add to this picture the fact that dollar credit plays a large and ear on year ear on year occasionally dominant role in EM private sector financing. 35 -20 Regardless of moves, a stronger dollar should -15 therefore make debt service more expensive in local 30 currency terms, with dollar coupon payments rising in -10 value. Corporates will become less willing to borrow and 25 banks less willing to lend. -5 20 On top of this, we have the more familiar macroeconomic 0 linkages occasioned by foreign currency debt: interest on 15 5 that debt will be driven by foreign central banks, and to the extent that dollar strength is driven by higher US rates, 10 10 it will be associated with slower EM credit growth. As with 15 monetary policy, this might be expected to work with a lag, 5 something that we can observe in the data. The chart in 20 Figure 5 shows total EM credit data from the BIS, adjusted 0 25 for moves in the dollar (otherwise currency moves would 04 05 06 07 08 09 10 11 12 13 14 15 16 17 18 confound the moves in credit growth), plotted against the Change in EM credit outstanding, dollar adjusted Dollar, 9m lag lagged trade-weighted dollar. There appears to be some support for the idea that a stronger dollar should weigh Source: Schroders Economics Group and Thomson Datastream, 12 June 2018. on EM credit, while a weaker dollar seems to be associated

Conclusion We have seen how a stronger dollar reduces global import demand by simultaneously raising import prices for all countries, except the US. At the same time, it raises the cost of credit, particularly in economies where dollar-denominated debt is prevalent. A stronger dollar is therefore a clear headwind to EM growth, generating inflationary pressure regardless of what trade-weighed exchange rates may be doing. It means that central banks have a clear incentive to fight depreciation of their currencies against the dollar, given that it offers little benefit to exports, while generating higher inflation. This suggests a few lessons for investors. One is that, with EM at least, it may be better to focus on the exchange rate versus the dollar rather than the trade-weighted exchange rate when considering whether a currency is fairly valued, whether an adjustment of the currency is sufficient to correct external imbalances, and in judging the effect on inflation. Reliance on purchasing power parity, real effective exchange rates, and other “fair value” measures will not lead to robust investment conclusions. For example, if we see a large REER move but very little against the dollar, we should not expect much reaction from the trade balance or inflation, and instead fade any market based on hopes of lower inflation, or better trade, depending on the direction of the move. We may also need to adjust our thinking on how EM central banks should react, assigning more of a role to currency stability than inflation and growth. Dollar strength may lead to more policy tightening than normally expected. This has particular implications for bond and currency investors, but should not be ignored by equity investors in EM either, given the role that monetary conditions play in equity markets.

5 Falk Brauning and Victoria Ivashina, “US monetary policy and emerging market 5 credit cycles”, Federal Reserve Bank of Boston Working Papers 17-9, August 2017. Important Information Note to Readers in Australia: Issued by Schroder Investment Management Australia Limited, Level 20, Angel Place, 123 Pitt Street, The views and opinions contained herein are those of Sydney NSW 2000 Australia. ABN 22 000 443 274, AFSL 226473. the authors as at the date of publication and are subject to change due to market and other conditions. Such Note to Readers in Canada: Schroder Investment Management views and opinions may not necessarily represent those North America Inc., 7 Bryant Park, New York, NY 10018-3706. expressed or reflected in other Schroders communications, NRD Number 12130. Registered as a Portfolio Manager with the strategies or funds. Ontario Securities Commission, Alberta Securities Commission, the British Columbia Securities Commission, the Manitoba Securities This document is intended to be for information Commission, the Nova Scotia Securities Commission, purposes only. The material is not intended as an offer the Saskatchewan Securities Commission and the (Quebec) Autorite or solicitation for the purchase or sale of any financial des Marches Financiers. instrument or security or to adopt any investment strategy. Note to Readers in Hong Kong: Schroder Investment Management The information provided is not intended to constitute (Hong Kong) Limited, Level 33, Two Pacific Place 88 Queensway, investment advice, an investment recommendation or Hong Kong. Central Entity Number (CE No.) ACJ591. Regulated by the investment research and does not take into account Securities and Futures Commission. specific circumstances of any recipient. The material is not intended to provide, and should not be relied on for, Note to Readers in Indonesia: PT Schroder Investment Management accounting, legal or tax advice. Indonesia, Indonesia Exchange Building Tower 1, 30th Floor, Jalan Jend. Sudirman Kav 52-53 Jakarta 12190 Indonesia. Registered Information herein is believed to be reliable but Schroders / Company Number by Bapepam Chairman’s Decree No: KEP-04/ does not represent or warrant its completeness or PM/MI/1997 dated April 25, 1997 on the investment management activities and Regulated by Otoritas Jasa Keuangan (“OJK”), formerly accuracy. No responsibility or liability is accepted by the and Financial Institution Supervisory Agency Schroders, its officers, employees or agents for errors of (“Bapepam dan LK”). fact or opinion or for any loss arising from use of all or any part of the information in this document. No reliance Note to Readers in Japan: Schroder Investment Management (Japan) should be placed on the views and information in the Limited, 21st Floor, Marunouchi Trust Tower Main, 1-8-3 Marunouchi, document when taking individual investment and/or Chiyoda-Ku, Tokyo 100- 0005, Japan. Registered as a Financial strategic decisions. Schroders has no obligation to notify Instruments Business Operator regulated by the Financial Services any recipient should any information contained herein Agency of Japan. Kanto Local Finance Bureau (FIBO) No. 90. changes or subsequently becomes inaccurate. Unless Note to Readers in People’s Republic of China: Schroder Investment otherwise authorised by Schroders, any reproduction of all Management (Shanghai) Co., Ltd., RM1101 11/F Shanghai IFC or part of the information in this document is prohibited. Phase (HSBC Building) 8 Century Avenue, Pudong, Shanghai, China, AMAC registration NO. P1066560. Regulated by Asset Management Any data contained in this document has been obtained Association of China. from sources we consider to be reliable. Schroders has not independently verified or validated such data and it should Note to Readers in Singapore: Schroder Investment Management be independently verified before further publication or use. (Singapore) Ltd, 138 Market Street #23-01, CapitaGreen, Singapore Schroders does not represent or warrant the accuracy or 048946. Company Registration No. 199201080H. Regulated by the completeness of any such data. Monetary Authority of Singapore. Schroders Korea Limited, All investing involves risk including the possible loss Note to Readers in South Korea: 26th Floor, 136, Sejong-daero, (Taepyeongno 1-ga, Seoul Finance of principal. Center), Jung-gu, Seoul 100-768, South Korea. Registered and regulated by Financial Supervisory Service of Korea. Third party data are owned or licensed by the data provider and may not be reproduced or extracted and used for any Note to Readers in Switzerland: Schroder Investment Management other purpose without the data provider’s consent. Third (Switzerland) AG, Central 2, CH-8001 Zürich, Postfach 1820, CH-8021 party data are provided without any warranties of any kind. Zürich, Switzerland. Enterprise identification number (UID) CHE- The data provider and issuer of the document shall have 101.447.114, reference number CH02039235704. Authorised and no liability in connection with the third party data. www. regulated by the Swiss Supervisory Authority (FINMA). schroders.com contains additional disclaimers which apply Note to Readers in Taiwan: Schroder Investment Management to the third party data. (Taiwan) Limited, 9F, 108, Sec.5, Hsin-Yi Road, Hsin-YI District, Taipei 11047 Taiwan, R.O.C. Registered as a Securities Investment Trust Past performance is not a guide to future performance Enterprise regulated by the Securities and Futures Bureau, Financial and may not be repeated. The value of investments and Supervisory Commission, R.O.C. the income from them may go down as well as up and investors may not get back the amounts originally invested. Note to Readers in the United Arab Emirates: Schroder Investment Exchange rate changes may cause the value of any Management Limited, 1st Floor, Gate Village Six, Dubai International overseas investments to rise or fall. This document may Financial Centre, PO Box 506612 Dubai, United Arab Emirates. contain “forward-looking” information, such as forecasts Registered Number 1893220 England. Authorised and regulated by or projections. Please note that any such information is the Financial Conduct Authority. not a guarantee of any future performance and there is no CS00166 assurance that any forecast or projection will be realised.

European Union/European Economic Area: Issued by Schroder Investment Management Limited, 31 Gresham Street, London, EC2V 7QA. Registered Number 1893220 England. Authorised and regulated by the Financial Conduct Authority.