## _wp12265

## Source details

**Canonical URL:** [_wp12265](https://www.imf.org/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2012/_wp12265.pdf)

## Other formats

- [Markdown version](/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2012/_wp12265.pdf.md)
- [Structured JSON version](/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2012/_wp12265.pdf.json)

---

### I. Introduction and objectives
- Definition:
  - The monetary transmission mechanism: how policy-induced changes in the short-term interest rate affect real variables such as investment, output, and employment; implicit is that changes in short-term policy rates affect other interest rates that matter for economic activity such as long-term interest rates.
- Problem statement:
  - The link between the short-term policy rate and longer-term rates can be impaired in economies with open financial accounts because international factors also play a role in determining returns on domestic financial assets.
  - Decoupling of short- and long-term interest rates in open emerging markets is a stylized fact over the last decade.
  - During periods of large capital inflows, banks and other financial institutions can obtain cheap foreign financing in large quantities, depressing domestic long-term yields and potentially impairing policymakers’ ability to tighten monetary policy.
- Regional focus and research questions:
  - Emerging Asia; emphasis on developing capital and bond markets.
  - Key questions: What determines market interest rates in Asia? What is the relative contribution of global versus domestic factors in driving yield curves? Are short-term or long-term interest rates more important in driving economic fluctuations?

### II. Data, methodology and sample
- Sample:
  - Eight emerging Asian economies: China, India, Indonesia, Korea, Malaysia, the Philippines, Taiwan Province of China, and Thailand.
  - Hong Kong SAR and Singapore excluded (monetary nominal anchor is the exchange rate).
- Methods:
  - Generalized Dynamic Factor Model (GDFM) to estimate unobserved common component of long-term bond yields.
  - Structural Vector Autoregression (SVAR) models to study relationships between domestic long-term interest rate, domestic policy rate, foreign interest rates, and output.
  - Identification: VAR with ordering (VIX, U.S. 10-year yield, estimated common factor); SVAR identified via Choleski decomposition with policy rate assumed contemporaneously more exogenous than market rates.
  - Robustness: generalized impulses and alternative global interest rate (average of United States, the European Union, and Japan) yield similar results.

### III. Empirical findings — drivers of Asian long-term bond yields
- Correlations and common factor:
  - Correlation between changes in the estimated common factor and changes in the U.S. 10-year government bond yield: 0.65.
  - Correlation between the estimated common factor and the VIX: -0.45.
  - If the crisis period is excluded, the correlation between the estimated common factor and the VIX is about positive 0.4.
- Share of variation explained by global/common factors:
  - About 40 percent of the variation in long-term Asian bond yields on average over the last 10 years can be explained by the “common factor.”
  - In a VAR decomposition of the estimated common factor (over 10 quarters):
    - U.S. long-term interest rates explain 35 percent of the variation.
    - VIX explains 25 percent of the variation.
  - Contributions of foreign interest rates and global risk aversion increase with maturity; foreign investors are more active at the longer end of the yield curve. At the short end, domestic factors are more prominent.
- SVAR results (U.S. versus domestic drivers):
  - 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.
  - For one-year yields, the contribution from U.S. interest rates is lower and domestic variables matter more.
  - Cross-country heterogeneity:
    - Contribution of U.S. interest rates to domestic bond yields is smaller for less financially integrated countries (India, China).
    - Contribution is higher in countries with large foreign presence in domestic government bond markets (Indonesia, Malaysia).
    - Regression of share of variance in long-term yields explained by U.S. interest rates on the Chinn-Ito index yields R² = 0.7604.

### IV. Which interest rates matter more for monetary transmission in Asia?
- VAR results on contribution to output variation:
  - After 1 year:
    - Changes in 3-month interest rates account for about 25 percent of the average variation in output across Asian emerging economies.
    - Changes in 10-year rates account for about 5 percent of the average variation in output.
  - At a 6-month horizon:
    - Role of the short-term policy rate in driving economic activity: about 30 percent.
    - Role of long-term interest rates: about 3 percent.
  - The remainder of output variation is explained by foreign demand, inflation expectations, and exchange rate changes.
- Mechanisms explaining short-term dominance:
  - Bank loans to businesses are often priced in reference to interbank rates with short-term maturities (typically three months).
  - Many loans in the region have shorter maturities (three to five years); more than half of corporate debt is short-term in most sample countries; bulk of mortgages is at variable rates priced in reference to short-term rates.

### V. Impact of large capital inflows on monetary transmission
- Conceptual channels weakening pass-through:
  - Large inflows may lower risk premia, blunting the impact of monetary tightening on lending rates.
  - Abundant foreign capital may lead banks not to raise lending rates when domestic policy tightens.
- Empirical setup:
  - Fixed-effects panel model regressing three-month market rates on policy rates and lags; estimated for entire sample and for episodes classified as “surge in capital inflows.”
  - Surge definition: period of two or more quarters when net private capital inflows (as share of GDP) are significantly larger (1 standard deviation) than historical trend or above the 75th percentile of their distribution.
  - Nineteen surges in net private capital inflows to Asia occurred during the last 10 years; 13 of these occurred before the global financial crisis.
- Quantitative results on pass-through during surges:
  - On average across Asian economies:
    - Short-term pass-through coefficients decline by about 40 percent during large inflow episodes.
    - Long-term pass-through coefficients decline by about 30 percent during large inflow episodes.
  - Despite weakening, transmission remains material:
    - Pass-through coefficients under large capital inflows are 0.3 in the short-term and 0.6 in the long-term for the region.
  - Visual axis ticks referenced: 0.0, 0.1, 0.2, 0.3, 0.4, 0.5, 0.6, 0.7, 0.8, 0.9, 1.0.

### VI. Effectiveness of interest rate transmission and heterogeneity
- Long-term rates:
  - Long-term interest rates in Asia are determined to a large extent by global factors including foreign interest rates and global risk aversion.
  - Exposure to global factors varies considerably across countries; more financially open economies are more exposed to the global interest rate cycle.
- Short-term rates:
  - The interest rate channel of monetary transmission remains powerful in Asia because it works mainly through short-term interest rates.
  - Even though long-term rates are substantially influenced by global factors, monetary policy can still affect the economy via short-term policy rates.

### VII. Policy implications and recommendations
- Main conclusions:
  - The policy interest rate remains a powerful tool for macroeconomic stabilization in Asia.
  - Continued integration of the region’s capital markets with the world economy could make external developments the dominant factor in determining domestic rates.
  - There could be constraints in implementing monetary policy when faced with large capital inflows; tightening monetary policy could reinforce more inflows.
- Policy prescriptions:
  - Policymakers should remain ready to use other policy instruments, such as macroprudential policies, as a complement to appropriate macroeconomic policy settings.
  - Theoretical work (Unsal (2012)) analyzes the use of macroprudential measures to help monetary policy stabilize the economy under large capital inflows.

*Source: _wp12265 - References .............................................................................................................*

### References .............................................................................................................

##_wp12265 - References .............................................................................................................

### Figures and Content Inventory
- Figures listed:
  - 1. Secondary Market Yield of 10-Year Government Bond
  - 2. The Estimated Common Factor and U.S. 10-Year Bond Yield and the VIX
  - 3. Contributions of U.S. 10-Year Yield and VIX to Estimated Common Factor
  - 4. Variance Decomposition of Domestic 10-Year Yield by Sources During 2005–10
  - 5. Contribution of U.S. Long-Term Interest Rates to Variance of Domestic Yields by Maturity
  - 6. Importance of U.S. Interest Rate and Capital Account Openness
  - 7. Variance Decomposition of Industrial Production in Response to Shocks to Domestic Interest Rates
  - 8. Short-Term Corporate Debt
  - 9. Effect of Capital Flows on Monetary Transmission Mechanism

### I. INTRODUCTION — context and objectives
- Definition:
  - The monetary transmission mechanism refers to how policy-induced changes in the short-term interest rate affect real variables such as investment, output, and employment, implicit in this definition is that changes in short-term policy rates have a significant impact on other interest rates that matter for economic activity such as long-term interest rates.
- Problem statement:
  - The link between the short-term policy rate and longer-term rates can be impaired in economies with open financial accounts because international factors also play a role in determining returns on domestic financial assets.
  - Decoupling of short- and long-term interest rates in open emerging markets is a well-known stylized fact over the last decade (Pradhan and others, 2011).
  - During periods of large capital inflows, the link between short-term policy rates and long-term rates could be particularly weak because banks and other financial institutions can obtain cheap financing from abroad in large quantities, depressing domestic long-term yields and potentially impairing policymakers’ ability to tighten monetary policy.
- Regional focus:
  - Emerging Asia, with emphasis on developing capital markets and bond markets that could make long-term yields more sensitive to global developments.
- Research questions:
  - What determines market interest rates in Asia? What is the relative contribution of global versus domestic factors in driving yield curves in the region?
  - Are short-term or long-term interest rates more important in driving economic fluctuations in the region?

### Data, methodology and sample
- Methods:
  - Generalized Dynamic Factor Model (GDFM) to estimate the unobserved common component of long-term bond yields in Asia and associate it with foreign interest rate (proxied by the U.S. long-term rates) and global risk aversion (measured by the VIX).
  - Structural Vector Autoregression (SVAR) models to study structural relationships between domestic long-term interest rate, domestic policy rate, and foreign interest rates, and between interest rates and economic activity.
- Sample:
  - Eight emerging Asian economies: China, India, Indonesia, Korea, Malaysia, the Philippines, Taiwan Province of China, and Thailand.
  - Hong Kong SAR and Singapore excluded because their nominal anchor for monetary policy is the exchange rate rather than the interest rate.

### Key empirical findings
- Role of global factors in long-term yields:
  - Global factors are important drivers of long-term interest rates in Asia.
  - About 40 percent of the changes in the long-term rates can be explained by external factors.
  - U.S. interest rates are a more important determinant of changes in long-term yields than short-term domestic interest rates in Asia.
- Monetary transmission and relative importance of maturities:
  - Monetary policy in the region is still effective; the interest rate channel of transmission largely relies on short-term interest rates.
  - Changes in short-term policy rates account for a much more significant variation in output than longer-term rates in the region.
- Effect of capital inflows:
  - Large capital inflows weaken the link between the short-term policy rate and lending rates by about a third, but the transmission mechanism still remains powerful.

### Literature connections and prior findings cited
- U.S. long-term yield explanations:
  - Kim and Wright (2005) and Cochrane and Piazzesi (2006) attribute decline in U.S. long-term yields to falls in risk premium.
  - Backus and Wright (2007) suggest a more stable macroeconomic and financial environment, including more predictable monetary policy, contributed to the decline.
  - Bernanke (2005, 2007) argued that global forces can help explain long-term interest rates.
  - Craine and Martin (2009) and Warnock and Warnock (2009) point to increased foreign demand lowering U.S. yields in 2004–05.
  - Bernanke and others (2004) found evidence that Japanese intervention in currency markets pushed the ten year bond yield down.
- Cross-country and emerging market evidence:
  - Diebold, Li, and Yue (2008) and Byrne, Fazio, and Fiess (2010) find global factors explain a large fraction of the yield curve, especially at the long end.
  - Moreno (2008), Ciarlone and others (2009), and Pradhan and others (2011) provide evidence on the key role of global factors in determining long-term interest rates in a number of emerging markets.
  - Moreno (2008) finds evidence for an increasing role played by global factors over the last decade.
- Regional reference:
  - See Felman and others, 2011 for related material.

### Structure of the paper (as provided)
- Methodologies discussion (next section)
- Empirical analysis on role of global and domestic factors in driving long-term interest rates in Asia (Section 3)
- Investigation of relative importance of short-term versus long-term interest rates in driving economic activity (Section 4)

*Source: _wp12265 - References .............................................................................................................*

### Section 5 analyzes the impact of large capital inflows on the transmission of changes in the

### _wp12265 - Section 5 analyzes the impact of large capital inflows on the transmission of changes in the

### Methodology: GDFM and SVAR
- Generalized Dynamic Factor Model (GDFM)
  - Assumes unobservable latent variables: common factors (influence more than one data point) and specific factors (influence only one).
  - The data set is a balanced panel of yields on government bonds of different maturities.
  - Each observation is decomposed into a common component χ and an idiosyncratic component ξ.
  - The common factor f is estimated via dynamic principal component analysis; the estimated unobserved “common” factor represents the impact of global factors on bond yields.
  - To identify drivers of the estimated common factor, a VAR is constructed with global risk aversion (VIX), U.S. 10-year government bond yields, and the estimated common factor (in that ordering).

- Structural Vector Autoregression (SVAR)
  - SVAR estimated for domestic long-term interest rate, domestic policy rate, and foreign interest rates; Y includes expected changes in exchange rates, changes in capital flows, GDP growth, and inflation expectations; Z includes exogenous variables such as the VIX and foreign demand.
  - Data from January 2000 to November 2010 are first differenced; lags chosen by standard information criterion.
  - Identification via Choleski decomposition; ordering assumes output reacts more slowly than financial variables and that policy rate is contemporaneously more exogenous than market rates.
  - Robustness: generalized impulses give similar results; using a global interest rate (average of United States, the European Union, and Japan) does not change findings.
  - A second set of VARs examines output responses to shocks in 3-month, 1-year, and 10-year yields; interest rates ordered by maturity so shorter maturities affect longer-term yields contemporaneously.

### Empirical findings on drivers of Asian long-term bond yields
- Estimated common factor and correlations
  - Correlation coefficient between changes in the estimated common factor and changes in the interest rate on 10-year U.S. government bonds: 0.65.
  - Correlation coefficient between the estimated common factor and the VIX: -0.45.
  - If the crisis period is excluded, the correlation between the estimated common factor and the VIX is about positive 0.4.

- Share of variation explained by common/global factors
  - About 40 percent of the variation in long-term Asian bond yields on average over the last 10 years can be explained by the “common factor.”
  - In a VAR decomposition of the estimated common factor (over 10 quarters):
    - U.S. long-term interest rates explain 35 percent of the variation.
    - VIX explains 25 percent of the variation.
  - Contributions of foreign interest rates and global risk aversion increase with maturity; foreign investors have been more active at the longer end of the yield curve.
  - At the short-end of the yield curve, domestic factors play a more prominent role.

- SVAR results: U.S. interest rates versus domestic policy rates
  - 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.
  - For shorter-term yields (one year), the contribution from U.S. interest rates is lower and domestic variables matter more.
  - Cross-country heterogeneity:
    - Contribution of U.S. interest rates to domestic bond yields is smaller for less financially integrated countries such as India and China.
    - Contribution is higher in countries with large foreign presence in domestic government bond markets, such as Indonesia and Malaysia.
    - Plotting the share of variance in long-term bond yields explained by U.S. interest rates against the Chinn-Ito index shows R² = 0.7604, indicating more financially integrated countries tend to be more exposed to the U.S. interest rate cycle.

### Which interest rates matter more for monetary transmission in Asia?
- Relative importance of short-term versus long-term rates for output
  - VAR results: after 1 year, changes in 3-month interest rates account for about 25 percent of the average variation in output across Asian emerging economies.
  - After 1 year, changes in 10-year rates account for about 5 percent of the average variation in output.
  - At a 6-month horizon:
    - Role of the short-term policy rate in driving economic activity: about 30 percent.
    - Role of long-term interest rates: about 3 percent.
  - The remainder of the variation in output is explained by foreign demand, inflation expectations, and exchange rate changes.
- Explanation for greater short-term role
  - Bank loans to businesses in Asia are often priced in reference to interbank rates with short-term maturities (typically three months).
  - Many loans in the region have shorter maturities (three to five years); more than half of corporate debt is short-term in most sample countries; bulk of mortgages is at variable rates priced in reference to short-term rates.

### Impact of large capital inflows on monetary transmission
- Conceptual channels through which large, volatile capital inflows weaken pass-through
  - Large inflows may lower risk premia, blunting impact of monetary tightening on lending rates.
  - Abundant foreign capital may lead banks not to raise lending rates when domestic policy tightens.

- Empirical assessment: fixed-effects panel model of pass-through
  - Model regresses three-month market rates on policy rates and lags of both; estimated for entire sample and for episodes classified as “surge in capital inflows.”
  - Definition of a surge in net private capital inflows: period of two or more quarters when flows (as share of GDP) are significantly larger (1 standard deviation) than their historical trend or above the 75th percentile of their distribution over the whole sample.
  - Nineteen surges in net private capital inflows to Asia have occurred during the last 10 years; most of these episodes (13) occurred before the global financial crisis.

- Quantitative results on pass-through under large inflows
  - Large capital inflows weaken the link between changes in policy rates and bank lending rates.
  - On average across Asian economies:
    - Short-term pass-through coefficients decline by about 40 percent during large inflow episodes.
    - Long-term pass-through coefficients decline by about 30 percent during large inflow episodes.
  - Despite weakening, transmission from short-term interest rate to lending rates remains strong:
    - The pass-through coefficients even under large capital inflows are 0.3 in the short-term, and

*Content based on _wp12265 - Section 5 analyzes the impact of large capital inflows on the transmission of changes in the*

### 0.6 in the long-term for the region.

### 0.6 in the long-term for the region.

### Effectiveness of interest rate transmission in Asia
- Long-term interest rates in Asia are determined by global factors to a large extent, including foreign interest rates and global risk aversion.
- Exposure to global factors varies considerably across countries.
- More financially open economies are more exposed to the global interest rate cycle.

### Role of short-term versus long-term rates
- The interest rate channel of the monetary transmission mechanism remains powerful in Asia because it works mainly through short-term interest rates.
- Even though long-term rates are substantially influenced by global factors, monetary policy can still affect the economy via short-term policy rates.

### Impact of capital inflows on pass-through and stabilization
- Pass-through from policy rates to lending rates is lower during episodes of large capital inflows.
- Despite lower pass-through in such episodes, monetary policy in Asia remains effective at macroeconomic stabilization.
- The document presents a comparison of short-term pass-through and long-term pass-through with axis values shown from 0.0 to 1.0 and labels indicating "Entire sample" and "Capital inflow episodes" in a figure titled "Effect of Capital Flows on Monetary Transmission Mechanism (Pass-through from policy rates to lending rates)".

### Policy implications and recommendations
- The policy interest rate remains a powerful tool for macroeconomic stabilization in Asia.
- Continued integration of the region’s capital markets with the world economy could make external developments the dominant factor in determining domestic rates.
- There could be constraints in implementing monetary policy when faced with large capital inflows; for example, tightening monetary policy could reinforce more inflows.
- Policymakers should remain ready to use other policy instruments, such as macroprudential policies, as a complement to appropriate macroeconomic policy settings.
- See Unsal (2012) for a theoretical analysis on the use of macroprudential measures in helping monetary policy stabilize the economy under large capital inflows.

### Supporting numerical and episodic details
- The global crisis averaged about 4 percent of GDP (compared with about 5 percent in the 1990s) and lasted only five quarters.
- Figure axis ticks referenced in the source: 0.0, 0.1, 0.2, 0.3, 0.4, 0.5, 0.6, 0.7, 0.8, 0.9, 1.0.

*Source: IMF working paper excerpt (VI. CONCLUSION and REFERENCES).*

---


_Source: https://www.imf.org/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2012/_wp12265.pdf_
