II. Capital Flows to Asia: Comparison with Previous Experience and Monetary Policy Options

International Monetary Fund. Asia and Pacific Dept

Jump to Content Jump to Main Navigation

Author: International Monetary Fund. Asia and Pacific Dept

Language:: English

Keywords:: REO; economy; price; China; export; rate; interest rate channel; flows to Asia; transmission mechanism; monetary policy stance; net capital capital inflow; Capital inflows; Macroprudential policy instruments; Central bank policy rate; Long term interest rates; Asia and Pacific; Global

Download PDF (1.1 MB)

Abstract

The rapid resumption of capital flows into emerging Asia since mid-2009 has posed two sets of challenges to policymakers in the region. First, although many regional economies are experiencing inflation pressures, policymakers have been reluctant to increase policy rates for fear of attracting more capital inflows. This is in line with previous surges of capital flows to the region, when real policy rates across Asia have on average taken more than 8 quarters to regain their pre-“surge” levels (Figure 2.1). However, as Chapter I discusses, exchange rate appreciation and tighter fiscal policy can play a role in combating overheating pressures, but so must monetary policy.

A. Introduction

A second challenge for policymakers in several Asian economies is to manage the financial stability implications of large capital inflows. Unusually strong cyclical and policy differences between advanced economies and emerging Asia in 2009 and 2010, and a gradual shift in portfolio allocation toward emerging markets, have led to portfolio inflows that, for a few Asian economies, are large in relation to the absorptive capacity of domestic markets. Despite the slowdown since late 2010, portfolio inflows to emerging Asia are expected to continue over the next two years, especially in India and the ASEAN-5 economies. As noted in Chapter I, however, risks of greater volatility in capital flows have intensified compared with six months ago. Policymakers in the region thus need to remain mindful of the risks that a sharp reversal or sudden stop of these inflows could pose for domestic financial markets.

Against this background, this chapter focuses on the following three main questions:

How does the surge of capital flows to Asia since the global recession compare with previous episodes of large capital flows to the region?
Do large capital flows tend to weaken the monetary policy transmission mechanism in Asia and, therefore, the effectiveness of monetary tightening?
What is the role of macroprudential regulations in helping monetary policy reduce economic volatility?

The chapter has three main conclusions. First, although as a share of GDP net capital flows to Asia are below previous peaks, and there is still little evidence of a significant buildup of financial imbalances, the rapidity of the surge and its concentration in potentially volatile portfolio inflows do raise concerns in a few Asian economies. Second, whereas large capital inflows complicate monetary policy by depressing long-term yields, monetary policy in the region still has an important influence on economic activity, as the interest rate channel of monetary policy in Asia largely relies on short-term interest rates. Third, macroprudential measures have a useful role to play in reducing economic instability that could arise from surges in capital inflows. These measures, however, should not be a substitute for monetary tightening.

B. How Does the Current Episode of Capital Inflows Compare with Previous Episodes?

Net capital flows to emerging Asia have rebounded at a record pace since the global financial crisis. There have been two other major episodes of inflows to emerging Asia over the past two decades. The first episode began in the early 1990s and ended abruptly with the 1997–98 financial crisis; the second began in the early 2000s and again ended abruptly with the global financial crisis (Figure 2.2). What is remarkable about the current episode is the speed of the recovery. Within just 5 quarters, net inflows rose from a recent trough (in early 2009) to their recent peak (in mid-2010). In contrast, the length between the troughs and peaks was about 25 quarters during the pre-Asian crisis period and the period before the global financial crisis.

Overall net capital flows to emerging Asia as a share of GDP have been generally lower than in previous surges. As a share of regional GDP, net capital inflows peaked at about 4¼ percent of GDP in the last quarter of 2010, against 6¾ percent in both the second quarter of 1996 and the first quarter of 2004. Gross flows do not paint a starkly different picture, as the preglobal crisis period had higher gross inflows and outflows than the current period.¹ Unsurprisingly, the aggregate numbers hide some sizable variations. As a share of GDP, net capital inflows have reached record highs after the global crisis in the NIEs, reflecting extraordinary banking-related flows to Hong Kong SAR and portfolio debt flows to Korea. By contrast, net capital flows to the ASEAN-5 economies have remained below the peaks reached in the previous two surges, as the increase in portfolio debt flows has been more than offset by declines in FDI and banking flows. Net capital inflows have also remained below precrisis peaks in both China and India.

Several regional economies, however, have experienced record-high portfolio flows. The post-crisis wave of capital inflows has been more geared toward debt flows compared with previous ones. This trend has been most pronounced in portfolio debt flows into countries where local bond markets are relatively large, such as Indonesia, Korea, and Malaysia. In Malaysia, record high portfolio inflows after the global crisis have been offset by other large outflows. Debt flows to the banking sector have remained important also for Hong Kong SAR, reflecting the importance of the banking sector in this country. India has also experienced a record surge in portfolio inflows, but of the equity kind.

To put the recent surge of capital inflows into historical perspective, this chapter identifies episodes of large net private capital flows to Asia over the last two decades. The focus is on net capital flows after stripping out official flows, and a “surge” in capital flows is defined by following the methodology outlined in IMF (2007a). Broadly speaking, under this definition, an episode of large net private capital flows for a particular country is a period of two or more quarters during which these flows (as a share of GDP) are significantly larger (one standard deviation) than their historical trend, or above the 75th percentile of their distribution over the whole sample.

The event analysis confirms that the recent surges have generally been smaller than in the past. On the basis of the definition above, 31 surges in net private capital flows to Asia have occurred during the last 20 years. Most of these episodes (13) occurred before the global financial crisis (Table 2.1). There were fewer episodes in the run-up to the Asian crisis (10) but they were of longer duration (about 25 quarters, on average). Only 8 episodes of large capital inflows have occurred after the global financial crisis, averaging about 4 percent of GDP (compared with about 5 percent in the 1990s) and lasting only 5 quarters. Indeed, of these episodes, only the ones that began in China and the Philippines in the second quarter of 2009 were still ongoing as of December 2010.

Table 2.1.

Episodes of Large Net Private Capital Flows to Emerging Asia: Summary Statistics

Sources: IMF, Balance of Payment Statistics and staff calculations.

^¹Market GDP-weighted average across episodes.

Table 2.1.

Episodes of Large Net Private Capital Flows to Emerging Asia: Summary Statistics

		Emerging Asia	ASEAN-5	NIEs	China	India
Number of episodes		31	12	13	3	3
	1989–1998	10	5	3	1	1
	1999–2008	13	5	6	1	1
	Current	8	2	4	1	1
Average size (in percent of GDP)¹
	1989–1998	5.0	7.0	5.3	4.8	2.4
	1999–2008	4.3	3.2	5.7	3.7	4.8
	Current	4.3	2.3	7.9	3.5	5.0
Duration (in quarters)		7	9	5	20	23
	1989–1998	25	30	10	20	23
	1999–2008	6	7	5	22	23
	Current	5	4	5	8	5

Sources: IMF, Balance of Payment Statistics and staff calculations.

^¹Market GDP-weighted average across episodes.

Table 2.1.

Episodes of Large Net Private Capital Flows to Emerging Asia: Summary Statistics

		Emerging Asia	ASEAN-5	NIEs	China	India
Number of episodes		31	12	13	3	3
	1989–1998	10	5	3	1	1
	1999–2008	13	5	6	1	1
	Current	8	2	4	1	1
Average size (in percent of GDP)¹
	1989–1998	5.0	7.0	5.3	4.8	2.4
	1999–2008	4.3	3.2	5.7	3.7	4.8
	Current	4.3	2.3	7.9	3.5	5.0
Duration (in quarters)		7	9	5	20	23
	1989–1998	25	30	10	20	23
	1999–2008	6	7	5	22	23
	Current	5	4	5	8	5

Sources: IMF, Balance of Payment Statistics and staff calculations.

^¹Market GDP-weighted average across episodes.

However, there is sizable cross-country variation. For the ASEAN-5 countries, net capital flows were substantially higher in the 1990s than at any time subsequently. This is because although portfolio debt flows have been sizable recently, FDI and banking flows have fallen substantially since the mid-1990s. In India, on the other hand, net capital flows have been on a secular rise since the capital account liberalization of 1991. The NIEs offer yet another variation, with net flows peaking after the global financial crisis. In China, the surges have gradually decreased over time, with the composition shifting from FDI to banking flows.

There are fewer signs of imbalances in Asian asset markets now than during previous capital inflow surges. Comparing the deviations from long-term averages in early 2011 asset valuations with peak levels in previous episodes of large capital inflows suggests that (Table 2.2):²

Equities and bonds. Across all economies in Asia, equity valuations (forward-looking price-earnings ratios) reached significantly higher peaks during the previous episodes of large capital inflows, particularly in the buildup to the Asian crisis. The picture is almost identical for bonds. Ten-year sovereign bond spreads were wider in early 2011 compared with the trough reached before the global financial crisis and at the eve of the crisis (2007:Q4; Figure 2.3).
Real estate markets. There were strong signs of overheating in the buildup to the Asian crisis according to house price-to-rent indicators, with the possible exception of Indonesia.³ There were fewer such signs before the global financial crisis, except for price-to-rent ratios in Indonesia, Malaysia, and Taiwan Province of China. As of 2010:Q4, price-to-rent ratios appear relatively strong only in China and Hong Kong SAR.
Credit growth. Most countries showed signs of excessive credit expansion during the capital inflows episodes of the 1990s. Although there were less signs of excessive credit growth before the global financial crisis, in late 2010 growth of credit to GDP ratios was particularly strong in a few regional economies including China, Hong Kong SAR, and Vietnam.
Corporate sector. Firms have deleveraged markedly since the Asian crisis; before then, corporate debt-to-equity ratios were in the red or orange zone for all countries except for Taiwan Province of China.

Table 2.2.

Financial Indicators Across Episodes of Large Net Capital Flows to Emerging Asia

Sources: CEIC Data Company Ltd.; Haver Analytics; Thompson Reuters I/B/E/S database; Organization for Economic Co-operation and Development; Worldscope; and IMF, International Financial Statistics and staff estimates.

^¹Colors represent the extent of the deviation from long-term averages expressed in number of standard deviations (z-scores). Green denotes less than 1.5 standard deviations above long-term averages, orange between 1.5 and 2 standard deviations, and red greater than 2 standard deviations. For methodologies, see Annex 1.9 of IMF (2010a).

^²For Indonesia, Korea, Malaysia, and Thailand, the period 1998–2000 is excluded in determining the peaks.

^³Indexes equal to 100 in 2002:Q3 for Taiwan Province of China and in 2008:Q4 for other economies.

^⁴Annual changes in credit-to-GDP ratios.

However, Asian economies are still at an early stage of the capital flow cycle, and concerns about the volatility of capital flows remain. Moreover, as suggested in Chapter I, the few signs of recent overheating pressures more likely reflect domestic imbalances than capital inflows, suggesting that imbalances often develop outside of capital inflow surges. A perennial concern of policymakers in emerging market countries is the volatility of capital flows, which, as noted in the April 2011 World Economic Outlook, has generally increased across time for all types of flows and regional groupings. Given the relatively shallow markets in some countries, this suggests that asset price bubbles can still form quickly, and that sudden stops remain a real possibility.

C. How Effective Is Monetary Policy in the Face of Large Capital Flows?

When confronting volatile and potentially destabilizing capital inflows, the first line of defense is macroeconomic policies, including both monetary and exchange rate policy, and fiscal policy. If the economy is overheating, with high or rising inflation or a developing credit or asset price boom, monetary policy should be tightened, although this can attract further inflows, to the extent that they are being driven by yield differentials. Greater exchange rate flexibility offers an important buffer against the risks posed by large capital inflows, as it can reduce the contribution to domestic demand overheating from large capital inflows; curb expectations of a large step appreciation and thus discourage further speculative inflows; and lessen the need for foreign exchange intervention and the resulting risk of excess liquidity and credit booms. Countercyclical fiscal policy also has an important role to play in weakening the impact of capital flows on the domestic cycle, reducing both appreciation and overheating pressures.

By depressing local long-term yields, however, the rapid resumption of capital flows to Asia after the global crisis has raised concerns about policymakers’ ability to tighten monetary stances. Households and firms often base their decisions to consume or invest on long-term interest rates. However, central banks have relatively limited influence on longer-term rates, which are also influenced by global factors, term premiums, and inflation expectations.

Indeed, global interest rates have been a key driver of long-term bond yields in emerging Asia (Figure 2.4). To assess the relative importance of domestic versus foreign factors in determining long-term interest rates in Asia, two methodologies are used in this chapter: a generalized dynamic factor model and a structural vector autoregression (SVAR) model (Appendix 2.1 for further details).⁴ The main results of the analysis are as follows:

A large proportion of the change in long-term yields in Asia over the last decade can be explained by global factors. The estimated common factor model shows that about 40 percent of the variation in Asian bond yields on average over the last 10 years can be explained by a “common factor.” U.S. long-term interest rates and global risk aversion (measured by the VIX) explain a large share of the variation in this common factor (35 percent and 25 percent, respectively; Figure 2.5).
U.S. interest rates are a more important determinant of changes in Asia’s long-term yields than short-term domestic interest rates. The response of domestic long-term yields to shocks to U.S. yields and domestic policy rates is assessed within an SVAR model that also includes inflation expectations, exchange rate changes, global risk aversion, and GDP growth.⁵ The results suggest that, on average across Asia, about half of the variation in long-term yields can be attributed to shocks to U.S. long-term interest rates (Figure 2.6). For shorter-term yields (1 year), the contribution from U.S. interest rates is lower, and domestic variables matter more (Figure 2.7).

The contribution of U.S. interest rates to domestic bond yields varies noticeably across Asia. In particular, the contribution is smaller in countries that are less financially integrated and have relatively less open capital accounts, such as India and China. The contribution is higher in countries with a large foreign presence in domestic government bond markets (such as Indonesia and Malaysia). Indeed, the increasing foreign participation in Asian bond markets appears to explain to a larger extent the correlation between U.S. interest rates and Asian long-term yields (Box 2.1). Plotting the contributions to Asian bond yields from the U.S. interest rate (from the SVAR model) together with an index of capital account openness (as measured by the Chinn-Ito index; see Chinn and Ito, 2008) shows that countries that are more financially integrated tend to be more exposed to the U.S. interest rate cycle (Figure 2.8).

Notwithstanding the important role of long-term rates in monetary transmission in other parts of the world, the interest rate channel in Asia works mostly through short-term interest rates. A vector autoregression model shows that after 1 year, changes in 3-month interest rates account for about 25 percent of the average variation in output across Asian emerging economies, compared with about 5 percent explained by changes in 10-year rates (Figure 2.9).⁶ The relatively greater importance of short-term rates in Asia may be explained by bank loans to businesses in Asia being often priced in reference to interbank rates with short-term maturities, typically 3 months.⁷ Furthermore, for most countries in the sample, more than half of corporate debt is short term, and the bulk of mortgages is at variable rates and also priced in reference to short-term rates (Figure 2.10).⁸

Box 2.1.

Do Nonresident Bond Holdings Affect Long-Term Interest Rates in Emerging Markets?

The surge of capital inflows to emerging market economies that began in mid-2009 has been characterized by a spike in nonresident investment in domestic bond markets. Nonresident investment in bond holdings has reached new peaks in Indonesia and Poland, and more than doubled in Korea, Malaysia, and Thailand (figure). However, even after the current surge in capital flows, nonresident investors still represent a minority share of emerging markets’ bond markets.

In several emerging market economies, the sharp increase in nonresident bond holdings has been accompanied by a decline of long-term yields, raising questions on the strength of the monetary policy transmission mechanism in the presence of large capital inflows, as discussed in the text.

To what extent was the decline in long-term yields in emerging markets related to the spike in nonresident investment in bonds market? There are two reasons why the two phenomena can be related. First, nonresident “real money” investors extended the maturity of their fixed income investments in emerging markets, relative to before the global crisis, attracted by the higher potential for gains at the longer end of the yield curve. Second, since domestic holders of long-term bonds typically hold them to maturity, even a small reallocation of foreign investment toward this segment of the market may be enough to bring down yields significantly.

Econometric analysis by IMF staff (Appendix 2.1) suggests that nonresident investment has contributed significantly to the observed decline in emerging market long-term yields. Each percentage point increase in foreign participation reduces long-term bond yields by about 5 basis points on average across emerging markets. This result is robust to changes in specification and estimation methods, and nearly identical to that in Peiris (2010). The results are also not significantly different for Asian economies compared with non-Asian emerging markets; nor is there a significant difference between economies that have a high or low share of nonresident investment. Finally, the results show that a 25 basis point increase in policy rates could offset the impact on long-term yields from a 2 percentage point increase in nonresident bond holdings. In other words, a modest tightening of the monetary policy stance could maintain long-term rates unchanged in the presence of moderate capital inflows.

Note: The main author of this box is Ceyda Oner.

Although still effective, the interest rate channel of transmission may be somewhat weaker in periods of large and volatile capital inflows. First, large inflows may lower the risk premium, blunting the impact of monetary tightening on lending rates. Second, if foreign capital is abundant, banks may choose not to raise lending rates when domestic monetary policy is tightened. To assess whether the pass-through from policy rates to market interest rates in Asia is different when an economy is facing large capital inflows, a fixed-effects panel model is estimated that regresses 3-month rates on policy rates and lags of both.⁹ Other factors that determine the pass-through from policy rates to market rates, such as the degree of competition within the banking sector, and financial market development and openness, are accounted for by including country-specific fixed effects in the model. The analysis finds that on average across Asian economies, the short-term pass-through coefficients decline from about 0.5 to 0.3 in periods of large capital inflows, whereas the long-term pass-through coefficients decline from 0.9 to 0.6 (Figure 2.11).

D. What Role Can Macroprudential Measures Play?

When capital inflows are large, conventional monetary policy still has a role to play in counteracting overheating pressures but it may not be sufficient to guard against the risks of financial instability. Indeed, the global financial crisis has shown that macroeconomic stability is not sufficient to ensure financial stability. Financial imbalances built up in advanced economies notwithstanding stable growth and low inflation. Prudential regulation and supervision, with its focus on individual firms, provided no guarantee that systemwide risks could be contained. In this context, there have been increased calls for the development of macroprudential measures globally, with an explicit focus on systemwide financial risks (Bank for International Settlements (BIS), 2010; Ostry and others, 2011).

Macroprudential measures are designed to increase the resilience of the financial system to credit or asset valuation boom-bust cycles. They are defined as regulatory policies that aim to reduce systemic risks, ensure stability of the financial system as a whole against domestic or external shocks, and ensure that it continues to function effectively (BIS, 2010). During boom periods, perceived risk declines, asset prices increase, and lending and leverage become mutually reinforcing. Conversely, during a bust phase, a vicious spiral can arise between deleveraging, asset sales, and the real economy. Given Asia’s past experience with these cycles, macroprudential measures could be particularly useful in reducing the procyclicality of financial systems and, therefore, the amplitude of business cycles.¹⁰ Indeed, several economies, in Asia and elsewhere, strengthened macroprudential regulations during 2010 in an effort to minimize risks associated with large capital inflows (see Box 2.2 for details of a recent IMF survey).¹¹

Macroprudential measures differ from traditional monetary policy in some key respects. Changes in both policy rates and macroprudential measures are likely to affect aggregate demand and supply as well as financial conditions. However, the two instruments are not perfect substitutes and can usefully complement each other, especially in the presence of large capital inflows that tend to increase vulnerabilities of the financial system:

First, changes in policy rates are “blunt” instruments, as they impact all lending activities regardless of whether they represent a risk to the economy (Ostry and others, 2010). The increase in interest rates required to induce specific sectors to deleverage might be so large as to amplify aggregate economic volatility. By contrast, macroprudential measures are aimed specifically at markets in which the risk of financial instability is deemed to be excessive (BIS, 2008; Ingves, 2011).
Second, in economies with open financial accounts, increases in the interest rate may have only a limited impact on credit expansion if firms can borrow at a lower rate abroad. Moreover, although monetary transmission works well through the asset price channel in “normal” times, in “abnormal” times sizable rapid changes in risk premiums could offset or diminish the impact of policy rate changes on credit growth and asset prices (Kohn, 2008; Bank of England, 2009).
Third, interest rate movements aimed at ensuring financial stability could be inconsistent with those required to achieve macroeconomic stability, and this inconsistency could risk destabilizing inflation expectations (Borio and Lowe, 2002; Mishkin, 2007). For example, under an inflation-targeting framework, if the inflation outlook is within the target, a response to asset market fluctuations to maintain financial stability may damage the credibility of the policy framework.

Box 2.2.

Macroprudential Policy–An International Perspective

A recent survey conducted by the IMF shows that “macroprudential policy” is becoming an important element of economic policy making.¹ Macroprudential policy is intended to limit the buildup of systemic risk in the financial sector that may arise from either domestic imbalances or external shocks. The survey finds that Asia has had extensive experiences with the implementation of macroprudential policy.

A key responsibility of the macroprudential policy is to identify and monitor various risks that may have a systemic impact. According to the survey, a large array of indicators is used in that regard. In Asia and Europe, many country authorities are concerned with credit risk, and monitor quite closely the ratio of nonperforming loans to total loans (figure). Reflecting the structure of balance sheets, in Latin America, respondents tend to be more worried about currency risk and frequently monitor the net open currency position to capital. Emerging markets generally tend to track credit risk more than advanced countries, where the focus is on leverage.

Authorities have used a large toolkit to address the various risks given the wide policy perimeter they assign to macroprudential policy. A total of 30 different instruments is cited, some of which extend beyond traditional prudential tools. Increasing government-owned land sales to boost land supply, for example, is cited by Singapore as an instrument to prevent house price bubbles. While some of the instruments have been in use since the early 1990s, more countries have started to deploy and/or made adjustment to their instruments since the global crisis, a clear indication that such instruments are gaining importance as the macroprudential policy framework evolves. The most widely used instruments include caps on the loan-to-value ratio and limits on net open currency positions, but there are regional variations (figure).

For many emerging market economies, a frequently cited macroprudential policy objective is to address the impact of large capital inflows on the financial system. Asian countries often implement measures that are aimed at credit growth and the associated asset price inflation. Of the six countries that have total or sector-specific credit limits, three are in Asia (China, Malaysia, and Singapore), compared with one each in Latin America and Europe (figure). In contrast, Latin American countries tend to apply measures that target nonresidents, that is, capital controls.² These include unremunerated reserve requirements for nonresidents, taxation of capital flows, and minimum holding periods for capital inflows. Of the eight countries that have implemented capital controls, seven are in Latin America, compared with one in Asia (Thailand). European countries tend to focus on the currency risk aspect, and often impose caps on foreign currency lending.

Note: The main authors of this box are Cheng Hoon Lim, Xiaoyong Wu, and Akira Otani (of the IMF Monetary and Capital Markets Department).¹The survey was conducted in December 2010. It was sent to 63 countries and the European Central Bank, including all countries in the G–20. The response rate was close to 80 percent, of which 24 percent (12 countries) were from Asia (see IMF, 2011a).²The IMF, the Financial Stability Board, and the Bank for International Settlements do not consider capital controls to be part of the macroprudential toolkit.

The tradeoffs and complementarities between monetary policy and macroprudential measures are analyzed in an open-economy, New Keynesian macroeconomic model (see Appendix 2.2 for details). In the model, firms can finance their investment through retained earnings or borrowing from domestic or foreign sources.¹² Macroprudential policy entails higher costs for financial intermediaries that are likely to be passed on to borrowers. Hence, in the model, macroprudential measures are defined as an additional “regulation premium” to the cost of borrowing, rises with credit growth.¹³ This is meant to capture the notion that such measures make it harder for firms to borrow during boom times, and therefore make a subsequent bust less dramatic. Monetary policy is assumed to follow a Taylor rule, with the central bank reacting to changes in inflation and output gaps.

The analysis allows an assessment of alternative monetary and macroprudential responses to capital inflow surges. The initial shock is modeled as a decline in investors’ perception of risk, and it plays out through the familiar financial accelerator mechanism. As financing costs decline, firms borrow and invest more. Stronger final demand and higher asset prices boost firms’ balance sheets and reduce the risk premium further. As capital inflows surge, the currency appreciates, which helps limit overheating and inflation pressures. Eventually, higher leverage triggers an increase in risk premium, and financial conditions normalize. But both monetary and macroprudential policies have a nontrivial role in mitigating the impact of the shock.

The simulations suggest that macroprudential measures could be a useful complement to, but not a substitute for, monetary policy in stabilizing the economy. Figure 2.12 shows the response to an unanticipated 1 percent reduction of perceived risk, which results in an increase in capital flows of about 0.1 percent of output.¹⁴ Three different policy responses are compared, with the parameters of the policy rules and their stabilization properties presented in Tables 2.3 and 2.4, respectively.

Table 2.3.

Parameters of the Policy Rules

Source: IMF staff estimates.

Table 2.3.

Parameters of the Policy Rules

	Monetary policy rule			Macroprudential rule
	Lagged interest rate	Inflation	Output gap	Nominal credit
Taylor rule	0.5	1.5	0.5	0.0
Taylor rule and macroprudential rule	0.5	1.5	0.5	0.5
Optimized Taylor rule and macroprudential rule	0.0	2.4	0.8	1.3
Macroprudential rule without monetary response	0.0	0.0	0.0	2.5

Source: IMF staff estimates.

Table 2.3.

Parameters of the Policy Rules

	Monetary policy rule			Macroprudential rule
	Lagged interest rate	Inflation	Output gap	Nominal credit
Taylor rule	0.5	1.5	0.5	0.0
Taylor rule and macroprudential rule	0.5	1.5	0.5	0.5
Optimized Taylor rule and macroprudential rule	0.0	2.4	0.8	1.3
Macroprudential rule without monetary response	0.0	0.0	0.0	2.5

Source: IMF staff estimates.

Table 2.4.

Performance of Policies in Reaction to a Financial Shock

Source: IMF staff estimates.

^¹Welfare loss is calculated as the sum of the volatility of inflation and output gaps, multiplied by 100 so that the results indicate welfare losses as a percent of steady-state consumption.

Table 2.4.

Performance of Policies in Reaction to a Financial Shock

	Consumer prices (standard deviations)	Output gap (standard deviations)	Welfare loss¹ (in percent)
Taylor rule	0.03	0.15	2.46
Taylor rule and macroprudential rule	0.05	0.12	1.25
Optimized Taylor rule and macroprudential rule	0.00	0.02	0.02
Macroprudential rule without monetary response	0.19	0.53	31.50

Source: IMF staff estimates.

^¹Welfare loss is calculated as the sum of the volatility of inflation and output gaps, multiplied by 100 so that the results indicate welfare losses as a percent of steady-state consumption.

Table 2.4.

Performance of Policies in Reaction to a Financial Shock

	Consumer prices (standard deviations)	Output gap (standard deviations)	Welfare loss¹ (in percent)
Taylor rule	0.03	0.15	2.46
Taylor rule and macroprudential rule	0.05	0.12	1.25
Optimized Taylor rule and macroprudential rule	0.00	0.02	0.02
Macroprudential rule without monetary response	0.19	0.53	31.50

Source: IMF staff estimates.

^¹Welfare loss is calculated as the sum of the volatility of inflation and output gaps, multiplied by 100 so that the results indicate welfare losses as a percent of steady-state consumption.

In the first—baseline—scenario (Taylor rule only), policy rates are increased in response to higher output and inflation gaps. The higher policy rates partially offset the impact of the lower risk premium on lending rates, and stabilize output as investment and consumption become more costly. The stabilization of demand helps to reduce inflation, whereas the welfare loss is estimated at about 2½ percent of steady state consumption.
In the second scenario (Taylor rule and macroprudential measures), policymakers also adopt macroprudential measures that directly counteract the easing of the lending standards and thus the financial accelerator effect. Indeed, both domestic debt and foreign debt increase less than in the first scenario, and the increase in asset prices is also lower. The responses of output and inflation are therefore more muted, and the welfare loss after the shock decreases by more than half, compared with the simple Taylor rule.
In the third scenario (optimized Taylor rule and macroprudential regulation), the parameters of the Taylor rule and macroprudential responses are optimized so as to minimize the variation in inflation and output gap after the shock. Hence, the policy response in this case is most successful in stabilizing the economy and reducing welfare loss. This optimal response involves a tighter monetary policy stance, as the inflation term has a higher weight in the optimized Taylor rule (2.4) than in the previous two cases (1.5). But there is also room for macroprudential measures: indeed, the weight on nominal credit growth in the macroprudential rule is higher (1.3) than under the second scenario (0.5).

To illustrate that macroprudential measures alone are not a sufficient response and are not a substitute for monetary policy, we model a policy regime with macroprudential regulation while maintaining policy rates unchanged. Under this scenario, the regulation premium is calibrated to replicate the path of the lending rate under the baseline (Taylor rule) scenario, to reflect policymakers’ objective of achieving the same increase in the lending rate through macroprudential measures only. This policy would constraint firms’ borrowing and investment, but not consumption, as it would leave interest rates constant. Demand and inflation would thus be higher than in the other policy regimes, and the welfare loss would be excessively large. The size of the required macroprudential measure is likely to be too far reaching, significantly constraining the financial sector and damaging productive investment and potential growth.

As noted in Chapter I, it is too early to assess the effectiveness of the macroprudential measures adopted in many emerging Asian economies over the past few quarters. Using the IMF’s Annual Report on Exchange Arrangement and Exchange Restrictions (AREAER), however, it is possible to identify similar measures that were adopted in the past (since the mid-1990s), and to assess where they have been associated with changes in capital flows and key financial variables. To do this, measures in the AREAER database have been classified across three categories: (i) foreign exchange-related measures (aimed at reducing banks’ foreign currency exposure, including, for example, higher reserve requirement on foreign currency deposits); (ii) housing market-related measures (including lower loan-to-value ratios); (iii) other measures taken to address financial stability concerns and that did not discriminate between domestic and foreign residents. The result suggests that foreign exchange-related and other measures have been generally associated with some moderation in net capital inflows, although only the latter with statistical significance (Table 2.5). The adoption of housing-related measures has been followed by lower residential price-to-rent ratios.

Table 2.5.

Selected Asia: Impact of Macroprudential Measures¹

Sources: IMF, Annual Report on Exchange Arrangements and Exchange Restrictions (AREAER), International Financial Statistics, and staff estimates.

^¹The impact of each measure is assessed within six quarters following its introduction. A standard one-sided t-test is used to assess whether financial indicators are significantly different after the introduction of the measures relative to the same number of quarters before. The table reports the impact during the first quarter the policy is effective, or–for measures not statistically significant–most effective (the number in brackets). An asterisk mark denotes statistical significance at 10 percent level.

Table 2.5.

Selected Asia: Impact of Macroprudential Measures¹

		FX-related prudential measures	Other prudential measures
Net private capital inflows (in percent of GDP)		- 1.3 (4)	- 1.9 * (5)
	Portfolio investment	- 0.6 (1)	- 1.2 * (4)
	Bank loans/ other investment	- 1.8 (4)	- 1.7 * (5)
Index of residential price/rent ratios (in percent)		- 11.5* (4)

Sources: IMF, Annual Report on Exchange Arrangements and Exchange Restrictions (AREAER), International Financial Statistics, and staff estimates.

Table 2.5.

Selected Asia: Impact of Macroprudential Measures¹

		FX-related prudential measures	Other prudential measures
Net private capital inflows (in percent of GDP)		- 1.3 (4)	- 1.9 * (5)
	Portfolio investment	- 0.6 (1)	- 1.2 * (4)
	Bank loans/ other investment	- 1.8 (4)	- 1.7 * (5)
Index of residential price/rent ratios (in percent)		- 11.5* (4)

Sources: IMF, Annual Report on Exchange Arrangements and Exchange Restrictions (AREAER), International Financial Statistics, and staff estimates.

E. Conclusions

Net capital flows to emerging Asia have surged after the global crisis but not reached previous peaks, and there are only isolated signs of pressures thus far, but policymakers should remain focused on potential risks to the real economy and financial stability from capital inflow surges. There is significant variation across countries in both the magnitudes and types of inflows experienced so far. Signs of risks from asset valuations and corporate indicators remain largely muted, and external buffers are large. Nevertheless, there are isolated pockets of concern, such as credit dynamics in some countries and certain segments of property markets around the region. Despite the fact that the current capital flow cycle is only a few quarters old, the strong pace of the surge until late 2010 and the continuing volatile nature of capital flows warrant special attention.

Even in periods of large capital inflows, monetary policy in Asia remains effective at macroeconomic stabilization. In addition, other complementary tools can be useful at times in helping to achieve the twin objectives of macroeconomic and financial stability. In particular:

Although long-term interest rates in Asia are predominantly determined by global factors, the interest rate channel of the monetary transmission mechanism remains powerful, as it works mainly through short-term interest rates.
Macroprudential measures have a useful role to play in addressing the risks of macrofinancial instability from capital inflow surges. However, these measures are not a substitute for tighter monetary policy.

Appendix 2.1. Econometric Methods: Global Dynamic Factor Model, Structural VAR, and Panel Regression

The Global Dynamic Factor Model (GDFM) provides an estimation of the unobserved common factor among a given set of sample elements, following Forni and others (2005). A vector of time series for each country is represented as the sum of two mutually orthogonal components: a common component and an idiosyncratic component. The common component here would correspond to the variation in yields that is not directly linked to the specific macroeconomic characteristics of the country but to developments in the global economic and financial system. Next, the determinants of this common component are analyzed, including the VIX, foreign interest rates, and the slope of the U.S. yield curve as a proxy for foreign growth prospects.

A structural VAR (SVAR) is first used to estimate the relationship between the domestic long-term interest rate, domestic policy rate, and foreign interest rates.¹⁵ The SVAR includes a number of variables that influence the dynamics of interest rates, including the VIX (assumed to be exogenous), expected changes in exchange rates, changes in capital flows, growth differentials, and inflation expectations.¹⁶ Data are from January 2000 to November 2010 and are first differenced to ensure stationarity. The lags are chosen using the standard information criterion.

The SVAR is identified using the Choleski decomposition of the variance-covariance matrix of the residuals. The ordering of the model follows other papers in the literature, including Christiano, Eichenbaum, and Evans (1996) and Kim (1999). The identifying assumption is that output reacts more slowly than financial variables and is contemporaneously more exogenous. In the ordering, output is followed by inflation expectations, interest rates, and exchange rates. The policy rate is assumed to be contemporaneously more exogenous than market rates, because the latter are assumed to react to the policy rate.

The robustness of the results is tested in various ways. First, using generalized impulses also yields the result that long-term yields in Asia are driven to a greater extent by foreign interest rates. Second, the results are based on the U.S. 10-year yield, which correlates well with the common factor. However, a robustness check using a global interest rate (average of the United States, the European Union, and Japan) does not change the nature of the findings.

A second set of VARs is used to assess the relationship between output and interest rates across the yield curve in Asia, similar to Christiano, Eichenbaum, and Evans (1996) and Lange (2005). In particular, the responses of output to shocks in interest rates of different maturities are examined. The variables in the model include 3-month (proxy for short-term rate), 1-year, and 10-year government bond yields, industrial production (measure of output), inflation expectations, exchange rate changes, and foreign demand. The variables in the VAR are ordered so that output is assumed to be contemporaneously the most exogenous (as above), whereas interest rates are considered more endogenous as they react more rapidly to changes in nominal variables. To identify the VAR, we assume that interest rates are ordered in terms of maturity (Lange, 2005), so that the yields of shorter maturity affect the longer-term yields contemporaneously. The relative importance of shocks to different interest rates for aggregate demand is shown by the forecast error variance decompositions.

To analyze the impact of nonresident investment on long-term yields, a fixed-effects panel model is estimated for eight emerging markets over 2000:Q1–2010:Q4. This analysis follows the specification similar to Warnock and Warnock (2009), which explored the impact of nonresident purchases of U.S. Treasury bonds on their yields.¹⁷ The specification estimated is as follows:

\begin{matrix} i_{i, t}^{L T} = c + α_{i} + β_{1} i_{i, t}^{S T} + β_{2} π_{i, t}^{e} + + β_{3} x_{i, t}^{e} + β_{4} ρ_{t} + β_{5} f_{i, t} + β_{6} y_{i, t}^{e} + β_{7} B_{i, t} + ε_{i, t} & (1) \end{matrix}

where for country i, $i_{i, t}^{L T}$ is the long-term government bond yield, $i_{t}^{S T}$ is the short-term interest rate, $π_{t}^{e}$ is the 12-month ahead inflation expectations, $x_{i, t}^{e}$ is the expected change in the exchange rate vis-à-vis the U.S. dollar over the next 12 months, f_{i, t} is the fiscal balance, ρ_t is a measure of the risk premium on the long-term bond, $y_{t}^{e}$ is the one-year ahead expected output gap, and B_t is the foreign holdings of bonds in percent of total outstanding.

Appendix 2.2. DSGE Model

Model simulations are based on a fully articulated structural model. Specifically, an open economy New Keynesian dynamic stochastic general equilibrium (DSGE) model with nominal rigidities is augmented with a financial accelerator mechanism after Bernanke, Gertler, and Gilchrist (1999). While the nominal rigidities (which include sticky prices) motivate a role for active monetary policy, features such as imperfect exchange rate pass-through and foreign currency-denominated borrowing are included to better capture the economic challenges confronted by emerging market economies. The model builds on the work of Elekdag and Tchakarov (2007), Gertler, Gilchrist, and Natalucci (2007), Kannan, Rabanal, and Scott (2009), and particularly Ozkan and Unsal (2010).

There are four sectors in the model economy. Households receive utility from consumption, provide labor to production firms, and participate in domestic and international financial markets. The households also own the firms in the economy, and therefore receive profits from these firms. Final goods producers produce a differentiated final consumption good using both capital and labor as inputs. These firms engage in local currency pricing and face price adjustment costs, resulting in sticky local currency final goods prices. Similarly, importers also face price adjustment costs and have some market power. Finally, intermediate producers combine investment with rented capital to produce unfinished capital goods that are then sold to corporations.

The corporate sector plays a key role in the model—its decisions determine the production of capital. To finance their capital investments, corporations partially use internal funds. However, they also require external financing, which is more costly than internal funds. The spread between the cost of external and internal financing is defined as “the risk premium,” and links the terms of credit with condition of the corporate balance sheet. Corporations are able to borrow from both local and foreign sources, and are indifferent between the two in the absence of cost differences.

In the model, macroprudential regulations entail higher costs for financial intermediaries, which are then reflected in lending rates. The spread between the lending rate and the policy rate is affected by the risk premium, discussed above, and “the regulation premium,” which is a function of nominal credit growth. Specifically, three factors affect the lending rate, $R_{t}^{L}$ , and take the following form:

\begin{matrix} R_{t}^{L} = R t^{P} ψ (\frac{D_{t}^{D}}{N_{t}}) R P_{t} . & (2) \end{matrix}

The first factor is the monetary policy rate, $R_{t}^{P}$ , which evolves according to a Taylor-type rule:

$\begin{matrix} R_{t}^{P} = {(R_{t - 1}^{P})}^{α} {[(1 + R_{s s}) {(π_{t} / π_{s s})}^{γ}^{π} {(Y_{t} / Y_{s s})}^{γ}^{Y}]}^{1 - α}, & (3) \end{matrix}$

where R_ss is the steady-state level of the policy rate, π_t and π_SS are current and steady-state levels of inflation, Y_t and Y_ss are the current and steady-state levels of output (in logs), respectively.
The second factor is the risk premium, ψ(.), which is an increasing function of leverage, ( $D_{t}^{D}$ /N_t), satisfying the conditions ψʹ > and ψʺ > 0, where $D_{t}^{D}$ denotes the real level of debt or credit (the sum of both local and foreign borrowing), and N_t is the net worth of the borrower.¹⁸
The third factor affecting lending rates is the regulation premium, RP_t. Following Kannan, Rabanal, and Scott (2009) RP_t is a function of nominal credit growth such that:

$\begin{matrix} R P_{t} = ρ [\frac{D_{t}^{D} P_{t}}{D_{t - 1}^{D} P_{t - 1}} - 1]^{19} & (4) \end{matrix}$

Note: The main authors of this chapter are Ravi Balakrishnan, Sonali Jain-Chandra, Sylwia Nowak, Sanjaya Panth, D. Filiz Unsal, and Yiqun Wu. Souvik Gupta provided research assistance.

Gross inflows and outflows generally appear to be on a secular trend upward before the global financial crisis, likely reflecting continued financial globalization during the Great Moderation.

For methodologies, see Annex 1.9 of IMF (2010a).

Pre-Asian crisis data are not available for China, India, Malaysia, and Taiwan Province of China.

The sample consists of eight Asian economies, namely, China, India, Indonesia, Korea, Malaysia, the Philippines, Taiwan Province of China, and Thailand. Hong Kong SAR and Singapore are excluded because their nominal anchor for monetary policy is not the interest rate, but the exchange rate.

The results presented here use the U.S. 10-year yield, as it correlates well with the common factor.

The remainder of the variation in output is explained by foreign demand, inflation expectations, and exchange rate changes.

See Garcia-Herrero and Remolona (2008).

See Mohanty and Turner (2008).

This is in line with the model specification presented in IMF (2010b), although estimated as a panel rather than country by country, owing to lack of sufficient capital inflow episodes for certain countries. Robust standard errors are used to gauge statistical significance.

See Craig, Davis, and Pascual (2006) for evidence on the procyclicality of Asian financial markets.

An environment of low interest rates may also foster greater risk appetite among financial intermediaries and investors and thus contribute to the buildup of imbalances. See Borio and Zhu (2008), Altunbas, Gambacorta, and Marqués-Ibàñez (2010), and Jimenez and others (2010).

The analysis adds an open economy dimension to several studies that have incorporated macroprudential instruments into general equilibrium models, such as Angeloni and Faia (2009), Kannan, Rabanal, and Scott (2009), N’Diaye (2009), and Angeloni, Faia, and Lo Duca (2010).

The chapter does not analyze any particular form of macroprudential measures, but rather focuses on a generic case where macroprudential measures lead to additional costs to financial intermediaries, which are then reflected in higher interest rates for borrowers.

The parameters of the model are calibrated to capture several important features of emerging Asia, including trade openness, leverage of firms, and the risk premium.

Previous work on a large set of Asian and non-Asian emerging market countries using VARs has found that foreign interest rates have a larger impact on domestic long-term rates than does the domestic policy rate (see Moreno, 2008).

Expected exchange rate changes are based on consensus forecasts of exchange rate movements.

The sample consists of Brazil, Indonesia, Korea, Malaysia, Mexico, Poland, Thailand, and Turkey. See Pradhan and others (2011) for details.

In the case of foreign-denominated debt, the leverage would be a function of the nominal exchange rate as well. Following Aysun and Honig (2010), we allow both foreign and domestic borrowing. Liability dollarization, in this case, is endogenous, in contrast to the existing literature.

See Borio and Drehmann (2009), Borgy, Clerc, and Renne (2009), and Gerdesmeier, Roffia, and Reimers (2009) for a specific emphasis on the potential of nominal credit growth as a regulation tool.

Foresee

Table 2.5.

Selected Asia: Impact of Macroprudential Measures¹

		FX-related prudential measures	Other prudential measures
Net private capital inflows (in percent of GDP)		- 1.3 (4)	- 1.9 * (5)
	Portfolio investment	- 0.6 (1)	- 1.2 * (4)
	Bank loans/ other investment	- 1.8 (4)	- 1.7 * (5)
Index of residential price/rent ratios (in percent)		- 11.5* (4)

Sources: IMF, Annual Report on Exchange Arrangements and Exchange Restrictions (AREAER), International Financial Statistics, and staff estimates.

	Equity forward-looking price/earnings ratios^1,2			Residential price/rent ratios^1,2,3

	Peak during the 1990s surge	Peak during the 2000s surge	Peak during the current episode	Peak during the 1990s surge	Peak during the 2000s surge	Peak during the current episode

China	24.0	32.8	14.2	…	100.0	118.8
Hong Kong SAR	15.8	21.1	17.2	170.3	126.6	144.5
India	23.6	21.7	17.5	…	103.1	91.9
Indonesia	22.2	15.6	14.4	106.6	108.9	100.9
Korea	20.7	12.3	12.2	121.8	101.0	99.5
Malaysia	27.2	17.1	15.5	…	106.0	101.8
Philippines	20.0	19.6	16.5	379.0	191.9	99.1
Singapore	27.2	22.6	14.3	171.2	130.4	117.3
Taiwan Province of China	33.2	23.9	29.4	…	120.8	112.7
Thailand	43.0	13.3	11.5	183.4	127.4	101.8

	Growth of credit-to-GDP ratios^1,4			Debt/equity ratios¹

	Peak during the 1990s surge	Peak during the 2000s surge	Peak during the current episode	Peak during the 1990s surge	Peak during the 2000s surge	Peak during the current episode

China	10.6	10.5	24.3	66.7	61.6	43.7
Hong Kong SAR	25.9	12.2	19.1	38.1	30.8	19.7
India	1.4	5.4	4.1	155.2	85.2	72.9
Indonesia	16.8	3.7	2.0	190.4	106.3	41.7
Korea	24.9	20.4	16.2	264.1	81.8	67.3
Malaysia	24.9	9.4	21.3	59.3	45.7	33.6
Philippines	12.6	2.0	4.0	67.0	39.8	16.1
Singapore	11.8	21.0	10.9	44.5	36.8	28.2
Taiwan Province of China	16.9	8.3	1.7	46.2	56.3	28.4
Thailand	19.0	5.7	6.1	166.0	68.2	34.3
Vietnam	…	22.0	22.5	…	53.8	47.4

	Equity forward-looking price/earnings ratios^1,2			Residential price/rent ratios^1,2,3

	Peak during the 1990s surge	Peak during the 2000s surge	Peak during the current episode	Peak during the 1990s surge	Peak during the 2000s surge	Peak during the current episode

China	24.0	32.8	14.2	…	100.0	118.8
Hong Kong SAR	15.8	21.1	17.2	170.3	126.6	144.5
India	23.6	21.7	17.5	…	103.1	91.9
Indonesia	22.2	15.6	14.4	106.6	108.9	100.9
Korea	20.7	12.3	12.2	121.8	101.0	99.5
Malaysia	27.2	17.1	15.5	…	106.0	101.8
Philippines	20.0	19.6	16.5	379.0	191.9	99.1
Singapore	27.2	22.6	14.3	171.2	130.4	117.3
Taiwan Province of China	33.2	23.9	29.4	…	120.8	112.7
Thailand	43.0	13.3	11.5	183.4	127.4	101.8

	Growth of credit-to-GDP ratios^1,4			Debt/equity ratios¹

	Peak during the 1990s surge	Peak during the 2000s surge	Peak during the current episode	Peak during the 1990s surge	Peak during the 2000s surge	Peak during the current episode

China	10.6	10.5	24.3	66.7	61.6	43.7
Hong Kong SAR	25.9	12.2	19.1	38.1	30.8	19.7
India	1.4	5.4	4.1	155.2	85.2	72.9
Indonesia	16.8	3.7	2.0	190.4	106.3	41.7
Korea	24.9	20.4	16.2	264.1	81.8	67.3
Malaysia	24.9	9.4	21.3	59.3	45.7	33.6
Philippines	12.6	2.0	4.0	67.0	39.8	16.1
Singapore	11.8	21.0	10.9	44.5	36.8	28.2
Taiwan Province of China	16.9	8.3	1.7	46.2	56.3	28.4
Thailand	19.0	5.7	6.1	166.0	68.2	34.3
Vietnam	…	22.0	22.5	…	53.8	47.4

Topics

Countries

Series

About

Resources

Abstract

A. Introduction

B. How Does the Current Episode of Capital Inflows Compare with Previous Episodes?

C. How Effective Is Monetary Policy in the Face of Large Capital Flows?

D. What Role Can Macroprudential Measures Play?

E. Conclusions

Appendix 2.1. Econometric Methods: Global Dynamic Factor Model, Structural VAR, and Panel Regression

Appendix 2.2. DSGE Model

Within Same Series

Other IMF Content

Other Publishers Content

Asian Development Bank

Inter-American Development Bank

Nordic Council of Ministers

The World Bank

Share Link

Table of Contents

Sections

Figures

Metrics

Foresee