Capital flows, exchange rates and policy frameworks in emerging Asia

(1)

ASIAN CONSULTATIVE COUNCIL of the

BANK FOR INTERNATIONAL SETTLEMENTS

Capital flows, exchange rates and policy frameworks in emerging Asia

A report by a Working Group established by

the Asian Consultative Council of the Bank for International Settlements 27 November 2020

Reserve Bank of Australia People’s Bank of China

Hong Kong Monetary Authority Reserve Bank of India

Bank Indonesia Bank of Japan

Bank of Korea

Central Bank of Malaysia Reserve Bank of New Zealand Bangko Sentral ng Pilipinas Monetary Authority of Singaore Bank of Thailand

Bank for International Settlements (Working Group secretariat)

(2)

Preface

The Asian Consultative Council (ACC) of the Bank for International Settlements (BIS) was established in March 2001 to facilitate communication between the BIS shareholding central banks in the Asia-Pacific region and the BIS's Board and Management on matters of interest to the Asia-Pacific central banking community. As of September 2020, the ACC comprised the Governors of the central banks and monetary authorities of Australia, China, Hong Kong SAR, India, Indonesia, Japan, Korea, Malaysia, New Zealand, the Philippines, Singapore and Thailand.

Under the direction of the ACC at its February 2019 meeting, the BIS Representative Office for Asia and the Pacific set up a Working Group of regional central banks to examine their policy frameworks, focusing on capital flows, exchange rates, and the joint use of monetary, macroprudential, exchange rate and capital flow management policies. The Working Group is made up of members from the central banks and monetary authorities of China, Hong Kong SAR, India, Indonesia, Korea, Malaysia, the Philippines, Singapore and Thailand, as well as observers from the central banks of Australia, Japan and New Zealand.

This report is the summary of the responses to two questionnaires, organised by different elements of policy frameworks. First, a detailed questionnaire, intended to provide a stocktake of how central banks model exchange rates and capital flows and incorporate these into their policy frameworks, and how they use various policy instruments to deal with challenges related to capital flows and exchange rates. Second, a short supplementary survey, added to assess changes in response to the Covid-19 pandemic, a shock that is serving as a severe stress test of policy frameworks in many jurisdictions in the Asia-Pacific region.

(4)

Executive summary

The BIS set up a Working Group on “Capital flows, exchange rates and policy frameworks in emerging Asia” under the direction of its Asian Consultative Council (ACC) to focus on the joint use of monetary, macroprudential, exchange rate and capital flow management policies to deal with capital flows and exchange rate volatility. This report is based on responses to two surveys of ACC members.

The Working Group members view ample global liquidity as the most important driver of capital flows, followed by the higher growth prospects of the recipient countries and, for some economies, capital account liberalisation. In general, exchange rates are considered important because of what they imply for monetary and financial stability, rather than because they are a target in their own right.

The members agree that the effects of the exchange rate can be summarised into three channels:

trade competitiveness, pass-through to inflation, and financial channels. The trade channel and the inflation pass-through channel have become less important over time, while the financial channel has increased in importance. Moreover, the importance of the different channels is state-dependent: during normal times, no single channel is dominant across members, while the financial channel is dominant during volatile times. The overall effect is that currency depreciation is expansionary during normal times, but it is contractionary during volatile times for most members. In approximate order of importance, the largest sources of spillovers to domestic financial conditions are the monetary policy decisions of major economies, global investors’ risk appetite and the strength of the US dollar.

The member central banks’ modelling efforts, as they relate to capital flows and exchange rates, divide into two broad camps: large-scale, theory-based models used to model the macroeconomy and produce forecasts of main macro variables; and smaller-scale models with less theory behind them (eg vector autoregressions, composite indices and stress-testing exercises) used to assess financial stability risks. One reason behind this distinction is that theory-based models do not generally account for the possibility of the relationships between macroeconomic variables changing when there are threats to financial stability. Relatedly, theory-based models generally exclude the effects of policy tools other than interest rates.

Determining the appropriate policy response to exchange rates and capital flows generally relies on the careful monitoring of FX liquidity, including the speed of exchange rate change, and the effects of capital flows on asset prices, with a view to ensuring orderly market functioning. Many Working Group members report that they allow exchange rates to be flexible and market-determined during normal times, but all stand ready to intervene in FX markets in response to excessive FX volatility to maintain external stability. In addition, some are prepared to utilise capital flow management measures when intervention is insufficient. Meanwhile, reliance on macroprudential measures to target specific domestic financial stability objectives has generally increased over time.

A majority of Working Group member central banks come close to the Tinbergen principle of one instrument for one objective. At the same time, in practice, some tools can affect multiple objectives.

Moreover, employing a combination of tools in a complementary manner can strengthen the effectiveness of policies, and also help to mitigate some of the unwanted side effects of policies.

The Covid-19 pandemic has served as a stress test of current policy frameworks. Central banks from the region used the full range of conventional policy tools in response to the crisis, and also expanded their toolbox, to ensure sufficient liquidity, both in their own currency and in US dollars, as well as bought assets, provided lending to key sectors and relaxed regulatory requirements, all in an attempt to prevent negative feedback loops between the real and financial sectors. Cooperation with the government has been a key element of the policy response. The member central banks generally view their responses as having delivered a positive impact on external and financial stability in the near term, but such unprecedented measures are also seen as set to have a significant impact on their economies for some time to come.

(5)

Financial channels are weaker overall for the three Working Group observers than for the members.

Therefore, the implications of exchange rates and capital flows for financial stability are less of a concern.

(6)

1. Introduction

The relationship of exchange rates and capital flows with monetary policy is a critical issue for central banks in emerging Asia. Changes in advanced economy monetary policy, trade tensions and, most recently, major challenges from the economic and financial fallout from Covid-19 point to heightened uncertainty, with greater volatility of both exchange rates and capital flows going forward.

This report lays out how monetary policy frameworks in Asia have responded to volatile exchange rates and capital flows, using a rich survey of central banks in the region.¹ In addition to discussing conventional policy in the form of short-term interest rates, this report documents how foreign exchange intervention has been used to lean against undesired exchange rate developments. Some central banks have also relied on macroprudential tools and capital flow management measures at times.

The report focuses primarily on the practices of the nine central banks in emerging Asia that are members of the Working Group, while Annex A compares them against those of the advanced economies in the region that were observers of the Working Group.²

The following key findings emerge. Increasing exposure to swings in global risk appetite and increased exchange rate and capital flow volatility since the Great Financial Crisis (GFC) of 2007–09 have occasioned an evolution in Working Group member central banks’ policy frameworks. The Working Group members view ample global liquidity as the most important driver of capital flows, followed by the higher growth prospects of the recipient countries. In addition, all members view the level of the effective exchange rate as crucial for trade competitiveness, while most members regard the volatility of the bilateral exchange rate against the US dollar as central to financial stability.

Among the three channels through which exchange rates affect the real economy, all members agree that the trade channel and the inflation pass-through channel have become less important over time, while the financial channel has become more important. All members recognise the financial channel as the most important during volatile times, when exchange rate depreciations tend to have contractionary effects on the domestic economy. This contrasts with more normal periods, when most members view exchange rate depreciations as expansionary due to the trade channel. Most members see advanced economy monetary policy as the most important source of spillovers to domestic financial conditions, with global investors’ risk appetite and swings in the US dollar also playing a role. For the Working Group observers, the financial channel of the exchange rate is weaker than in the member economies, or could even work in the opposite direction, in part due to structural factors such as widespread hedging of FX exposures.

From an analytical point of view, the financial channel is increasingly captured in stress testing and other scenario analyses, and to a lesser extent in larger macro models. The financial channel then feeds into decision-making through considerations about the effects of exchange rates and capital flows for domestic monetary conditions and financial stability.

For most economies, the reported policy response entails allowing exchange rate flexibility in normal times, but remaining vigilant and ready to use FX intervention and/or capital flow management measures (CFMs) during episodes of excessive volatility. Indeed, all Working Group members report using FX intervention, at least occasionally, to maintain external stability. Meanwhile, reliance on macroprudential measures to target specific domestic financial stability objectives has generally increased. Moreover, a majority of Working Group members tend to use each instrument mainly with the aim of affecting a particular, well defined objective. At the same time, employing a combination of

1 The questionnaires are contained in Annexes C and D. Sections 2 to 6 are based on survey responses received in mid- December 2019. Any policy changes since then are excluded from the discussion in this report.

2 See Annex A for a comparison of the responses by the Working Group members with those by the Reserve Bank of Australia and the Reserve Bank of New Zealand and, in the case of responses to Covid-19, also the Bank of Japan.

(7)

tools in a complementary manner can strengthen the effectiveness of policies, as suggested by the large number of instruments that many central banks report using to maintain external stability.

The Covid-19 pandemic has served as a stress test of current policy frameworks. To manage public health risks, governments have taken drastic measures, including lockdowns and social distancing rules, which have frozen domestic economic activity to varying degrees. Disruptions in global value chains, in turn, have led to plummeting exports, investment and consumption. The overall effect is extraordinarily adverse monetary and financial conditions. Central banks from the region responded forcefully. They used the full range of conventional policy tools in response to the crisis, including policy rate cuts, reserve requirement reductions, increased liquidity injections using repos, and intervention in FX markets. They also expanded their toolbox to ensure sufficient liquidity, both in their own currency and in US dollars, bought assets, supported lending to key sectors, especially to small and medium-sized enterprises (SMEs), and relaxed regulatory requirements. Meanwhile, financial stability policies were adapted to facilitate continued access to funding. Several of these policies required high levels of cooperation with governments and other financial authorities.

This report is structured as follows. The next section provides background as to why exchange rates and capital flows matter for policy, followed, in Section 3, by central banks’ views on the different transmission channels. Section 4 reports on the analytical frameworks in place at central banks to assess exchange rates and capital flows, while Section 5 discusses the information that staff provide to decision-makers to inform their views. In Section 6 we examine how central banks respond to exchange rate volatility and capital flows in terms of the choice of policy tools, the ordering of their use and calibration of the response. Section 7 provides an early evaluation of how monetary policy frameworks fared during the fallout from the Covid-19 pandemic. Finally, Section 8 offers some concluding comments.

2. Why exchange rates and capital flows matter for policy

This section discusses the reasons central banks give for focusing on capital flows and exchange rates.

The volatility of capital inflows into emerging Asia is perceived to have increased since the GFC, posing challenges for central bankers seeking to insulate their economies from destabilising external shocks.

In addition to the increase in volatility cited by most Working Group members (Table 1), the high level of inflows is seen as important for Thailand. China notes that sharp exchange rate fluctuations and large capital flows would threaten financial stability and have negative real economic consequences. In Korea these developments have been partially mitigated through the strengthening of FX sector macroprudential policy measures, and in Malaysia through restrictions on the facilitation of speculative offshore FX trading. In the Philippines, Bangko Sentral ng Pilipinas (BSP) reports that the volatility of foreign direct investment (FDI) inflows has decreased even as their level has increased. Bank Indonesia (BI) cites the high growth of short-term flows.

Key contributors to these capital flow developments include global liquidity, differential growth prospects and structural reforms in the region, including the easing of capital account restrictions (Table 1). Most respondents cite ample global liquidity, reflecting advanced economies’ expansionary monetary policies and especially quantitative easing, as an important driver of portfolio inflows. In addition, relatively advantageous growth prospects for the region have attracted capital in search of higher returns, notably in Indonesia, Malaysia, the Philippines and Singapore. A final factor mentioned by some respondents is structural reforms, in particular capital account liberalisations, within emerging Asia. According to the central banks surveyed, such reforms have made the region a more attractive destination for some kinds of flows, and increased financial integration.

(8)

Why exchange rates and capital flows matter for policy Table 1

Increasing exchange rate volatility (Y/N) Increasing exposure to global risk-on, risk-off dynamics (Y/N) Key factors driving

capital flows and exchange rates by rank

(1 = highest)

Which exchange rate matters more:

For competitiveness For financial stability

Ample global liquidity Higher growth prospects Structural changes in Asia1 Effective (trade- weighted) exchange rates Bilateral (USD) Level (L) or volatility (V)? Effective (trade- weighted) exchange rates Bilateral (USD) Level (L) or volatility (V)?

CN Y Y 1 2 √ L √ V

HK Y Y 1 2 √ L √ L

ID Y Y 1 2 √ L √ V

IN Y Y 1 2 3 √ L √ V

KR Y Y 1 2 √ L √ V

MY Y Y 1 2 √ L √ V

PH Y Y 1 2 3 √ L √ V

SG Y Y 1 1 3 √ L √ V

TH Y Y 1 2 √ L √ V

1 Including the easing of capital account restrictions.

The easing of capital flow restrictions has allowed increased borrowing from abroad. In India, regulations on the use of external borrowing were relaxed, encouraging inflows. Banks have also become more regionally integrated. In the Philippines, the expansion of foreign banks accounts for half of all FDI inflows into the financial intermediation sector, while external funding of banks more widely has also increased.

One consequence of increased financial integration is greater exposure to global risk-on, risk- off dynamics. While the build-up of inflows tends to be gradual, their reversal can be sudden and destabilising, as illustrated by the taper tantrum episode. The growing role of global asset management companies and benchmark tracking funds is also cited by the Central Bank of Malaysia (CBM) as a source of greater co-movement in asset prices across the region. By contrast, the Monetary Authority of Singapore (MAS) reports that institutional investors have tended to act countercyclically in Asian bond markets and thus to support financial stability. Even then, not all economies are treated equally:

Thailand, with the perception of being a relative safe haven, has seen capital inflows even during risk- off periods. More generally, most jurisdictions cite a preference for longer-term investors and direct investment over portfolio flows to try to limit destabilising dynamics.

With increased international financial exposures – especially in the form of external funding in foreign currency or foreign investment in local currency assets – both exchange rates and capital flows have become increasingly important inputs into domestic policy discussions. Exchange rates are key for policy, not because they are themselves a policy target or instrument (except, notably, in Hong Kong SAR and Singapore) but because of what they imply for monetary and financial stability, including the effective transmission of monetary policy. Capital flows are important drivers of interest rates and asset prices, including the exchange rate, which in turn influence the quantity and price of trade and, via exchange rate pass-through, the overall price level.

(9)

So, while most central banks characterise exchange rates as being primarily market-determined, they report a role for policy intervention to avoid excessive FX and capital flow volatility. Volatile capital flows pose particular risks to financial stability. Many respondents point out that large capital outflows tighten financial conditions and may affect financial intermediation and hence the effective stance of monetary policy, leading to a reliance on macroprudential or capital flow management measures. In the case of the Hong Kong Monetary Authority (HKMA), with foreign exchange interventions triggered automatically under the Linked Exchange Rate System (LERS), macroprudential measures are the primary means of discretionary intervention.

Most central banks rely on multiple ways of measuring exchange rate movements to inform different aspects of their policy discussions. Nominal and/or real effective exchange rates (NEER/REER) are used to assess competitiveness, and hence the real effects of the value of the currency, in all member jurisdictions (Table 1). In Singapore, the NEER is also the monetary policy instrument, and intervention is used to ensure that the NEER stays within a path projected to be consistent with medium- term price stability.

The bilateral exchange rate against the US dollar plays a pre-eminent role in Hong Kong SAR due to the LERS, but it is also important elsewhere due to the dollar’s dominance as an invoicing currency and use in trade finance. Moreover, the bilateral exchange rate is important for comparing asset returns, such that large changes may trigger destabilising financial dynamics. Furthermore, as an easily understood price, particular levels of the bilateral US dollar exchange rate can play psychological roles.

Most jurisdictions stress the importance of higher moments – the volatility and/or rate of change in the exchange rate – as important variables in considering intervention, primarily due to the implications for tail risks and financial stability (Table 1). The Bank of Thailand (BoT) notes that while too much exchange rate volatility is clearly costly, too little volatility will deter financial market development, including the availability and use of hedging instruments, and can encourage destabilising hot money inflows from non-resident investors.

In summary, exchange rates and capital flows play a key role for monetary policy for most Working Group members, and one that has been increasing in importance since the GFC. We next discuss the transmission channels via which they affect the economy.

3. Transmission channels

This section discusses the different transmission channels of exchange rates and capital flows to the wider economies in emerging Asia, as highlighted in questionnaire responses. Exchange rates and capital flows affect the economy through three main mechanisms: trade competitiveness, pass-through to inflation, and financial channels. The relative strength of these three channels evolves over time. A priori, the overall effects of exchange rate changes are likely to be multiple and state-dependent.

Exchange rate depreciations tend to increase competitiveness due to their effect on export prices, import prices and firm profits, and hence support net export increases and economic growth.

The price effects may be somewhat muted by firms seeking to protect their market shares, and can be weakened when exchange rate volatility is high. Larger firms may be able to offset some of these effects through hedging, whereas SMEs are likely to face greater consequences from exchange rate changes.

The size of these effects also depends on the structure of the economy. In Indonesia, exports and investment rely heavily on imported raw materials and capital goods, so that a weaker domestic currency does not necessarily lead to an increase in exports and output, and may actually reduce growth. In the Philippines, remittances are a persistent source of foreign income that affects a large share of the population and tends to increase in response to a depreciation.

(10)

Exchange rate depreciations also imply increased prices of imports which, via exchange rate pass- through, fuel higher rates of domestic inflation. For India, a 10% depreciation of the trade-weighted rupee is estimated to translate into a 1.5 percentage point increase in headline inflation in 2014, and the exchange rate pass-through had increased over time up to 2014 due to greater openness and exchange rate volatility. However, exchange rate pass-through has since declined. In Thailand, pass- through is most clearly observed for oil prices, with an asymmetric effect: lower oil prices bring down inflation, whereas the effects of oil price rises tend to be offset by domestic oil subsidies. For the HKMA, any disequilibrium in the foreign exchange market is corrected primarily through the adjustment of wages and prices, given the LERS.

The central banks surveyed generally perceive that, over time and with economic development, the macroeconomic importance of both exchange rate pass-through and export competitiveness has tended to decline across the region (Table 2).

Transmission channels _{Table 2}

Becoming more (M) or less (L)

important through time?

Ranking of importance during normal

times (1 = highest)

Ranking of importance during

volatile times (1 = highest)

Depreciation is expansionary

(E) or contractionary

(C) during...

Sources of spillovers to domestic financial conditions;

ranking of importance (1 = highest)

Trade competitiveness Pass-through to inflation Financial channels Trade competitiveness Pass-through to inflation Financial channels Trade competitiveness Pass-through to inflation Financial channels …normal times …volatile times Monetary policy decisions of major central banks Global investors' risk appetite USD appreciation or depreciation Changes in AE regulatory frameworks, inflation rates and bond yields

CN L L M 1 3 2 2 3 1 E E 1 3 2 4

HK L L M 1 3 2 2 3 1 E E 1 1 1 4

ID L L M 1 3 1 2 3 1 C C 2 1 2 2

IN L L M 2 1 3 3 2 1 E C 1 1 1 4

KR L L M 1 2 3 2 3 1 E C 1 3 2 4

MY L L M 1 3 2 2 3 1 E C 1 1 3 4

PH L L M 2 1 3 3 2 1 E C 1 1 3 4

SG L L M 2 1 3 2 2 1 E C 1 4 2 3

TH L L M 1 2 3 2 3 1 E E 1 3 2 4

Capital flows work through multiple channels. First, capital outflows exert downward pressure on the price of domestic currencies, increasing competitiveness and boosting inflation pressures.

Second, capital flows have important financial effects. These financial channels in general work in the opposite direction to the competitiveness channel, and may contribute to resource misallocation.

Outflows are associated with revaluations of foreign currency-denominated assets and liabilities. In economies with net short foreign currency positions, the negative wealth effect tends to lead to credit contractions. Even if there is no foreign currency debt, exchange rate depreciation may be contractionary if sovereign yields rise when the local currency depreciates. Capital inflows, in contrast, contribute to asset price overvaluation and excessive risk-taking. The effects of capital flows on asset

(11)

prices tend to be self-propagating, and can influence the real economy through wealth effects and collateral channels. Market expectations about the direction and volatility of capital flows and exchange rates have an important role to play in this context. In Singapore, easier domestic financing conditions associated with capital inflows can encourage overinvestment, especially in real estate. In the Philippines, a period of strong capital inflows following the GFC saw a divergence in market interest rates from the policy rate due to excessive liquidity growth, necessitating refinements to the monetary policy framework.

Central banks see the financial channels of the exchange rate as having important financial stability implications, especially at times of large depreciations driven by sudden changes in the risk appetite of global investors. Large appreciations, when driven by capital inflows, also have negative effects – artificially boosting asset prices and causing resource misallocations. Most of the time, when exchange rate movements are modest, competitiveness and inflation channels receive the most focus.

But larger changes in capital flows and exchange rates lead to financial stability tail risks and negative effects on growth. In Thailand, these effects work through prices in bond and equity markets. The Bank of Korea (BoK) reports that the effects of the financial channel on real economic variables have declined since the GFC due in large part to a persistent current account surplus and macroprudential policy measures that have reduced foreign loans to the banking sector. In Singapore and Thailand, financial channels appear to have been more muted recently.

Sensitivity to all three channels varies by both the size of the economy and the extent of the exposures that the economy faces. In particular, the size of financial channels reflect balance sheet compositions, the level of foreign participation in domestic asset markets, exposure to foreign currencies, and the responsiveness of capital flows. For China, the People’s Bank of China (PBC) reports that the size of the country’s economy and levels of foreign participation in domestic capital markets impact all three channels. The BoT finds that the financial channel was the most important one at the time of the regional crisis (in 1997–98), but its prominence subsequently declined as the dependence on external funding fell. More generally, a heavy reliance on external funding, whether in terms of foreign currency borrowing or foreign investment in local currency bond markets, is seen to complicate the transmission of exchange rate and capital flow shocks.

Taking all the channels together, most central banks consider exchange rate depreciations to be expansionary during normal times, and the exchange rate to work as a shock absorber during such periods (Table 2). However, the dynamics switch with large exchange rate fluctuations and capital outflows, when non-linear dynamics, working through amplification mechanisms, strengthen financial channels. Further, several central banks mention that depreciations tend to be contractionary when the economy is slowing.

Different factors, however, are highlighted for different economies. For India, as a net commodity importer, the exchange rate generally acts as a shock absorber, although may act as a shock amplifier when conditions are volatile. For the Philippines, buffers built into the economy against external shocks have reportedly allowed the central bank to avoid reacting aggressively to exchange rate developments without a deterioration in macroeconomic performance.

Another way to interpret the survey responses is in terms of the implications for financial stability.

Whereas small shocks are of little importance, large exchange rate and capital flow shocks have a material effect on financial stability, especially in the presence of currency mismatches, requiring different policy responses. Memories of 1997–98 remain informative for how policymakers in the region think about these issues. In many jurisdictions, financial stability concerns are the basis for attempts to reduce exchange rate volatility. In some, macroprudential tools play an important role. In Singapore, the monetary policy framework automatically implies a stronger response to large, disorderly shocks: a policy band around the target exchange rate path provides room for the exchange rate to fluctuate in response to modest shocks, but triggers intervention in the short term if the exchange rate swings sharply towards the boundaries of the band.

(12)

Regional central banks see the financial spillovers from advanced economy monetary policy, and the bilateral exchange rate against the US dollar in particular, as key drivers of domestic financial conditions (Table 2). The most important mechanism for this is through the effect of advanced economy monetary policy on global risk appetite. BSP reports, for example, that US monetary policy and the global risk appetite (proxied by the VIX) affect foreign portfolio inflows, equity returns and domestic credit growth. US quantitative easing policies saw large, sharp increases in gross foreign portfolio inflows. For the CBM, capital flows at risk (defined as downside risks to capital outflows under a severe adverse shock) indicate that global factors such as investor risk aversion and broad dollar strength are significant predictors of large capital outflows. The HKMA was concerned that the taper tantrum would result in serious financial market disruption, in particular in the FX swap market, if it led to a dollar shortage in the global financial markets. Liquidity risks associated with the flow of international US dollar credit can be high due to the non-linear dynamics they can generate, and can pose significant ongoing challenges for policymakers.

Changes in regulatory frameworks in the advanced economies also play a role, as mentioned by India, Indonesia, Korea, the Philippines and Singapore: regulation affects risk sentiment and the degree of global liquidity, and hence influences financial conditions globally.

4. Modelling and analytics

Using the survey responses on the transmission channels discussed in Section 3 as a backdrop, this section outlines the modelling and analytics mentioned by central banks as being useful for assessing exchange rate dynamics and capital flows. Models are used to produce forecasts, conduct policy simulations, compare different scenarios, provide structure to policy discussions and measure policy effectiveness. Most respondents rely on multiple types of models, including some that are largely empirical in nature (eg based on vector autoregression (VAR) models or error correction mechanisms (ECMs)), and others that have New Keynesian theoretical foundations. Models intended to address longer-horizon questions tend to be more structural in nature, while shorter-term forecasting models are primarily empirical.

The types and use of models are linked to the monetary policy frameworks whose decisions they inform. For economies where the exchange rate plays a central role, exchange rate objectives are generally derived based on models. For example, MAS utilises a suite of models that includes a computable general equilibrium model, a reduced-form dynamic stochastic general equilibrium (DSGE) model and a variant of the global VAR model. These models incorporate key macroeconomic variables, including the exchange rate as the monetary policy instrument.

Forecasting inflation occupies a core role in modelling efforts, especially for economies with inflation targets such as Korea and the Philippines. Models provide a means to translate exchange rate and capital flow behaviour into implications for inflation outcomes at horizons appropriate for both setting policy and assessing the performance of the inflation targeting framework. The BoK relies on three different models to analyse the effects of exchange rates and capital flows on domestic variables.

BSP uses a multi-equation model based on ECMs for comprehensive assessment of the inflation outlook, and less structural approaches for short-term nowcasting. At the CBM, partial equilibrium models based on the Phillips curve are applied to forecast inflation over a one- to two-year horizon, while autoregressive integrated moving average (ARIMA) models are used for shorter horizons.

Most central banks also rely on a suite of different models to improve their understanding of exchange rate behaviour and gauge the effects of shocks, and a mixture of calibration and estimation (Table 3). The more theoretically grounded models generally have New Keynesian foundations and, in the cases of India and the Philippines, are based on the Forecasting and Policy Analysis System developed in consultation with the International Monetary Fund. They feature a forward-and-backward-

(13)

looking Phillips curve and monetary policy described using an augmented Taylor rule. These models allow the assessment of the dynamic path of key macroeconomic variables in a theoretically consistent manner. BSP also uses a DSGE model incorporating financial frictions in the form of a credit constraint on banks’ balance sheets as a complementary tool to assess both monetary policy and macroprudential regulation. The BoT relies on both a semi-structural DSGE model and an ECM-based model. Meanwhile, the CBM applies a structural VAR model to capture the transmission channels through which capital flows and exchange rate movements affect key domestic real and financial variables.

Empirical models are used in many jurisdictions to assess financial stability risks (Table 3).

Most of the theoretically based models discussed above are linear in macroeconomic variables, and hence not well suited to assess episodes of financial instability. Thus, other approaches are used instead.

For example, the HKMA uses a VAR-based Financial Conditions Index (FCI) to track overall financial conditions and draw macro-financial implications. Similarly, MAS uses an FCI along with a Financial Vulnerabilities Index and Growth-at-Risk for financial stability purposes. The BoK compiles a Financial Stability Index based on 20 variables to comprehensively assess the stability of financial markets and financial institutions, and also uses a foreign currency liquidity stress test model to analyse the impact of abrupt capital outflows in times of crisis through scenario analyses. BSP uses the Philippine Composite Index of Financial Stress, composed of 13 indicators, as an early warning indicator.

Modelling and analytics _{Table 3}

Primary means for forecasting

inflation and output Primary means for assessing

risks to financial stability Regarding primary structural model:

Large-scale structural

models (eg DSGE)

Small-scale empirical

models (eg ECM, VAR)

Large-scale structural

models (eg DSGE)

Small-scale empirical

models (eg ECM, VAR)

Assumes uncovered interest parity

(UIP)? (Y/N)

Captures FX intervention, CFM and

macroprudential tools? (Y/N)

CN √ √ N N

HK √ √ Y N

ID √ √ Y Y

IN √ √ Y N

KR √ √ Y N

MY √ √ Y N

PH √ √ Y N

SG √ √ Y N

TH √ √ Y N

Another use of models at some central banks is to estimate the equilibrium exchange rate, either to assess the degree of exchange rate misalignment or for more general monitoring purposes. This continues even as trade competitiveness channels are perceived to have weakened in recent years.

Central banks generally rely on several different measures, which can give very different answers. Some, such as those used by Hong Kong SAR, Malaysia and Thailand, are based on estimating the value of the REER that may be justified based on fundamentals, be they the terms of trade and productivity differentials, a sustainable current account or stabilising net foreign asset positions. In Indonesia and Singapore, consistency with inflation objectives is also a consideration. BSP relies on a number of empirical models, including deviations from a long-run trend, ARIMA and VAR models.

Within the models, exchange rates are assumed to affect the economy through two main channels: the value and volume of trade and, via exchange rate pass-through from the prices of commodities and/or other imports, the overall price level. The pass-through channel has been seen to weaken in several economies over time as inflation performance has improved. But key relationships involving financial channels are not part of workhorse macro models, and are typically missing

(14)

from those models used at central banks that have theoretical foundations. For instance, uncovered interest parity (UIP) is an assumption made by most members (Table 3). Further, interactions among the different policy tools – especially capital flow management tools and foreign exchange intervention – are generally not explicitly modelled. Where there is room for judgment in the calibration of model inputs, they can be implicitly taken into account. In some economies, such as India and the Philippines, work is ongoing to incorporate these effects more explicitly. Such absences from models are hard to avoid, given the complexity of the underlying relationships, leaving an important role for judgment in assessing the output of models.

5. Monitoring indicators and information provided to decision-makers

This section discusses the survey respondents’ description of information that is made available to policymakers – partly arising from the modelling efforts discussed in Section 4 – in order for them to decide how to respond to exchange rates and capital flows. We discuss the various indicators that central banks monitor and the rationale for paying attention to them, focusing first on measures of FX liquidity and then on other financial market and macroeconomic indicators.

In response to the questionnaire, central banks report that they monitor various FX liquidity and market development indicators on a regular basis. While the list of indicators is long, it includes both price and quantity indicators: those in spot and derivatives markets, as well as demand for FX by both residents and non-residents. It also features factors that affect supply and demand in the FX market, such as import and export data, and FX flows stemming from current, capital and financial transactions.

The liquidity indicators that are being monitored include FX volumes, bid-ask spreads on currencies, measures of FX volatility, and data on net open positions related to FX.

The central banks report that they monitor FX liquidity largely with a view to promoting orderly market functioning (Table 4). This objective features prominently in responses by most central banks, although details vary across the economies. In the case of Malaysia, the central bank’s financial stability mandate includes maintaining orderly market conditions. The BoK notes that foreign currency liquidity conditions affect the currency and rollover risks of non-financial corporations and financial institutions with foreign currency-denominated liabilities. In China, low levels of FX liquidity can be seen as hindering the normal operation of the foreign exchange market and price discovery. And, in the case of the Philippines, BSP assesses FX liquidity in terms of its implications for price stability and financial stability, including the orderly functioning of markets.

At the same time, some central banks monitor FX liquidity also because of its exchange rate implications. Indeed, the RBI argues that, while important for orderly market functioning, FX liquidity will have a bearing on the exchange rate. BI reports that developments in FX liquidity may generate exchange rate dynamics of overshooting and undershooting. In Thailand, US dollar funding liquidity affects the exchange rate and market functioning through the Thai baht implied swap rate. Additionally, US dollar liquidity is one of the key determinants of the BoT’s FX swap operations. By contrast, in Hong Kong SAR, the rule-based exchange rate regime is seen to provide the basis for currency stability, rendering FX liquidity less directly relevant for either the exchange rate or orderly market functioning.

Not surprisingly, decision-makers are provided with a broad spectrum of macroeconomic and financial market indicators to inform their views. These indicators comprise a number of risk factors and vulnerabilities, as well as forecasts, typically of key macroeconomic variables. The CBM mentions the use of scenario analyses on both exchange rates and capital flows. For capital flows, this includes the adequacy of international reserves and the estimated impact on various macroeconomic and financial indicators under stressed conditions. The indicators provided to policymakers can vary with the state of the economy: in the case of China, the PBC dynamically adjusts indicators according to economic and financial developments.

(15)

Information for decision-makers _{Table 4}

Do you monitor FX liquidity because

of its implications for… Macro forecasts provided with horizon of one to two years (Y/N)

…the exchange

rate? …orderly

market functioning?

Inflation Output Exchange rate Capital flows

CN N Y Y Y N N

HK N N Y Y N N

ID Y Y Y Y Y Y

IN Y Y Y Y N N

KR Y Y Y Y N N

MY N Y Y Y Y Y

PH N Y Y Y Y Y

SG N Y Y Y N N

TH Y Y Y Y Y N

At some central banks, in addition to other macro variables, decision-makers are provided with forecasts of exchange rates and capital flows (Table 4). In the case of Indonesia, this includes a projected path for both the exchange rate and the financial account of the balance of payments; for Malaysia and the Philippines, the outlook for both capital flows and exchange rates. At the BoT, model- based simulations are occasionally used to assess the effect of the Thai baht exchange rate on the economic outlook. A number of central banks report that the typical forecasting horizon for macroeconomic variables is two years.

In addition to forecasts, exchange rates and capital flows also feature in stress testing, vulnerability analyses and other assessments that are reported to decision-makers. The effects on financial stability are prominent. In particular, in Korea, Malaysia and the Philippines, policymakers receive information about the implications of exchange rates and capital flows for domestic financial stability. In the case of the CBM, this information includes implications for the institutional resilience of the banking system and the need for prudential guidelines or supervisory intervention. At the BoT, assessments about the sensitivity of the economy to the exchange rate are provided to decision-makers. Moreover, in India, Korea, Malaysia, the Philippines and Thailand, among others, the decision-makers in monetary policy meetings receive information on exchange rate developments.

Central banks also use model-based simulations to compare the effects of alternative policy decisions. At some central banks, including those of Indonesia, India, Korea, the Philippines and Singapore, these simulations are done regularly, to coincide with policy meetings. Sometimes they are undertaken without a predetermined schedule, should risk assessments or other factors suggest a need.

At the CBM, simulations are used as a complement to the overall risk assessments. In the case of the BoT, they are employed to assess the impact of the exchange rate on the economic outlook. At the same time, the BoT notes that the results from such simulations are used with caution, as the models require a number of assumptions and cannot address uncontrollable market factors.

In regimes where the exchange rate is targeted, information related to intervention activity is important. In Hong Kong SAR, a currency board subcommittee, which is responsible for ensuring that the currency board’s operations are in accordance with established policy, reviews reports on intervention operations, and also risk and vulnerability reports. The subcommittee may also recommend improvements to the currency board system and ensure a high degree of transparency in its operation.

Similarly, at MAS, decision-makers receive information on FX intervention operations and market developments in regular reports on the implementation of monetary policy.

(16)

We now turn to how this information is used by policymakers, in terms of formulating policy responses.

6. Responding to exchange rate volatility and capital flows

This section examines how central banks describe their responses to exchange rate volatility and capital flows in terms of the choice of policy tools, the ordering of their use and the calibration of the response.

Working Group members have used a mixture of policy tools in pursuit of their policy objectives, including interest rates (Annex Graph B1), intervention in FX markets (as proxied by changes in FX reserves and net forward positions; Graph B2), CFMs (Table B1) and macroprudential measures (Table B2).

6.1 When and how central banks intervene in FX markets

Many central banks in the region allow their exchange rates to be flexible and market-determined during normal times. For India, Korea, Malaysia, the Philippines and Thailand, the exchange rate generally works as a shock absorber, so that the initial response in the face of exchange rate pressures is to let the exchange rate move. Another consideration is that central banks may be willing to accommodate persistent exchange rate movements due to productivity advances – for example, while seeking to offset more temporary fluctuations. Of course, Hong Kong SAR and Singapore differ from other economies in the region, as the exchange rate plays a role as a target.

That said, even central banks with flexible exchange rates stand ready to intervene in response to excessive FX volatility. Notably, all Working Group members report using FX intervention, at least occasionally, to maintain external stability (Table 5). Indeed, the BoT notes a willingness to trigger FX intervention to ensure that markets function well. The RBI intervenes in response to excessive volatility, while the CBM’s interventions are driven by an assessment that excessive and volatile FX movements create risks. In addition to tempering sharp fluctuations by FX intervention, BSP mentions that, when warranted, the central bank provides liquidity and ensures that legitimate demands for FX are satisfied.

While many central banks monitor measures of equilibrium exchange rates (Section 4), the level of the real exchange rate is not itself a direct objective of policy. However, policy responses may be considered warranted if inflation or financial stability objectives are threatened.

Besides intervention in spot markets, other markets also play a role. For instance, the RBI does operations in over-the-counter and exchange-traded currency derivatives markets, while the BoT also mentions verbal intervention.

When intervention is insufficient, some authorities are also willing to use CFMs in response to large FX movements. Three Working Group members report that they have CFMs in their policy frameworks (Table 5). The BoT mentions that if FX volatility results from speculative flows from non- residents, appropriate CFMs are considered. BI may use CFMs in combination with interest rate policy when there is persistent exchange rate volatility. Similarly, the RBI’s past interventions were often combined with CFMs to help ensure that reductions in FX volatility are durable. Regarding other central banks, the CBM has not used CFMs for external stability post-GFC, but they remain an option in the policy toolkit.

(17)

Instruments and policy objectives in the policy framework (pre-Covid-19) _{Table 5}

Objective and instruments CN HK ID IN KR MY PH SG TH

External stability (including exchange rate stability and capital flow issues)

CFM √ √ √

FX intervention √ √ √ √ √ √ √ √ √

Intervention in bond and money markets √

Macroprudential measures √ √ √ √

Policy interest rates √ √ √

Capital account liberalisation, excluding (cyclical)

CFMs √ √ √ √

Liquidity provision √ √ √

Domestic financial stability CFM

FX intervention

Intervention in bond and money markets

Macroprudential measures √ √ √ √ √ √ √ √ √

Policy interest rates √ √

Capital account liberalisation, excluding (cyclical) CFMs

Liquidity provision √

Macroeconomic stability (including price stability) CFM

FX intervention √

Intervention in bond and money markets

Macroprudential measures √ √

Policy interest rates √ √ √ √ √ √ √

Capital account liberalisation, excluding (cyclical) CFMs

Liquidity provision

6.2 Determining the response to capital flows

Central banks do not treat all capital flows equally: some require stronger responses than others.

Two relevant dimensions relate to the types of flow and the types of investor. Five out of nine Working Group members report that policy responses depend on either or both aspects (Table 6). For example, Korea applies macroprudential limits on financial institutions’ short-term FX positions, including on currency-related derivatives. This is done to curb risks from the foreign currency borrowing of financial institutions, as well as to encourage them to lengthen the maturity structure. The BoT, meanwhile, is likely to use CFMs when capital flow volatility arises from the behaviour of non-resident investors, such as from speculative short-term flows from non-residents. In China, the financial sector has been opened up in order to promote more stable two-way flows, with recent measures covering institutions such as wealth management companies and pension funds.

For the CBM, the policy response to exchange rate- and capital flow-related risks depends on the source of stress, as each episode of exchange rate depreciation or volatility is different. During 2014–15, when the economy saw large non-resident capital flows and there was a terms-of-trade shock,

(18)

the authorities allowed the exchange rate to adjust flexibly as a shock absorber. By contrast, during the capital flow reversal of late 2016 there were spillovers from more opaque markets (eg for non- deliverable forwards) as well as imbalances in the domestic FX market, leading the authorities to take additional measures to maintain orderly FX market conditions. These considerations are consistent with the more general point that some of the variation in policy responses across countries reflect differences in the macroeconomic and financial environments that these countries face.

Aspects of policy frameworks _{Table 6}

CN HK ID IN KR MY PH SG TH

One tool is (mainly/strictly) assigned to each objective

(one-to-one mapping) √ √ √ √ √ √

There is a preferred order in using the tools √ √ √ √ √

Policies to manage capital flows or exchange rate volatility

have unwanted side effects √ √ √ √

Policy response depends on types of capital flow/types of

investor √ √ √ √ √

6.3 Responses to exchange rate volatility and capital flows within the evolving policy framework

Applying a combination of policy tools to address macroeconomic and financial stability risks, while following an inflation target, is characteristic of monetary policy in the post-GFC period for many regional economies. One common feature, as noted by BI, is policy responses to risks associated with capital flows and exchange rate dynamics. Another is the introduction of additional instruments, including FX-related macroprudential measures by the BoK. The ordering of the different policy tools is also highly relevant – five Working Group members state that they have a preferred order of using the instruments in their toolbox (Table 6).

Yet, post-GFC, the policy mix shifted also in economies that were not explicitly targeting inflation.

As noted by the CBM, the changing nature of risks necessitated a broader range of policy options to address them. For instance, the PBC and MAS introduced macroprudential measures to contain risks related to short-term capital flows.

In these post-GFC frameworks, emerging Asian central banks tend to use each instrument mainly with the aim of affecting a particular, well defined, objective. This feature – mentioned in six survey responses – is consistent with the Tinbergen principle (Table 6). Specifically, while there is some variation across institutions, monetary policy is generally aimed at maintaining macroeconomic stability;

macroprudential policy at dealing with identified threats to domestic financial stability; and FX intervention, at times together with CFMs and other tools if necessary, at targeting external stability.

However, strict compliance with the Tinbergen principle is impossible because, as the BoT mentions, in practice some tools affect multiple objectives. Approaching the zero lower bound can also complicate trade-offs. MAS discusses that while macroprudential policy is primarily meant to address systemic financial risks, it can straddle multiple objectives, as asset prices matter for both price stability and financial stability. The RBI makes use of macroprudential tools for both external and domestic financial stability purposes. Using a combination of tools in a complementary manner can also strengthen the effectiveness of policies, as suggested by the large number of instruments the central banks report employing to maintain external stability (Table 5). Finally, BI notes that its instruments are not strictly assigned to singular objectives.

(19)

6.4 Unwanted side effects and policy constraints

Central banks also acknowledge unwanted side effects of the different tools used to manage capital flows and exchange rate volatility (Table 6).

Many side effects relate to FX intervention. BSP discusses the potential financial losses that stem from valuation changes and sterilisation costs. The BoT mentions that FX intervention reduces the private sector’s incentives to manage exchange rate fluctuations with proper hedging tools, and could also lead to trade disputes and accusations of currency manipulation. In addition, excessive FX intervention could lead to price distortions and inhibit market efficiency. At the same time, not intervening could also have unwanted consequences: the BoT argues that excessive exchange rate volatility may deter economic agents’ adjustments and have adverse implications for the economy.

Liberalised capital flows can also lead to unwanted dynamics. While the objective of liberalisation is to encourage greater two-way flows, domestic firms may be reluctant to repatriate assets from abroad during periods when the domestic currency is depreciating and capital is flowing out from the economy.

Further, central banks acknowledge various constraints in the use of the different tools. Limits to FX intervention can be binding, due to either the size of the central bank’s balance sheet (capital or FX reserves) or the cost of sterilisation.

If the necessary tools are outside the central bank, policy responses could also be constrained.

Yet central banks document various ways in which this concern, relevant especially for financial stability policy, is being alleviated. In Indonesia, the central bank maintains close coordination with financial regulators, the ministry of finance and other ministries in order to ensure aligned and coordinated policies. Similarly, MAS reports a relatively high degree of coordination across the different agencies to prevent policy conflicts or arbitrage. And, the RBI notes that formal and informal coordination mechanisms exist to deal with overlaps with other regulatory agencies.

6.5 Role for international cooperation

Moving beyond domestic frameworks, member central banks consider international cooperation and information-sharing useful when responding to capital flows and exchange rates. India notes that cooperation helps to ensure that central banks obtain timely signals and it facilitates faster policy responses, while BSP mentions that varied experiences from different economies help assess policy responses to different scenarios. That said, MAS argues that domestic price and financial stability objectives should in general not be compromised by international cooperation or coordination considerations.

The region’s central banks are already active internationally along a number of dimensions. Policy discussions with other central banks, including those coordinated by regional and international organisations, play a key role. Another important component of cooperation is the financial arrangements and safety nets among regional central banks, such as the Chiang Mai Initiative Multilateralisation and repurchase agreements between EMEAP members. Moreover, arrangements for local currency settlement for trade and investment within ASEAN have been put in place.

As it is, central banks see the potential for further international cooperation. In its survey response prior to the Covid-19 outbreak, the RBI mentioned the possibility of swap lines between advanced and emerging market economies (see also Section 7). MAS also argues that policy cooperation or coordination could be in the area of global safety nets comprising multilateral swap lines and/or repurchase agreements. More generally, the RBI calls for greater recognition of the spillover effects of advanced economy monetary policies to emerging market economies.

Capital flows, exchange rates and policy frameworks in emerging Asia