2017年-BIS国际清算银行_Market_volatility_monetary_policy_and_the_term_premium_42页_1004kb

Summary of BIS Working Paper No. 606: Market Volatility, Monetary Policy and the Term Premium

Core Content

This working paper investigates the relationship between market volatility, monetary policy, and the term premium in the context of the US bond market. The study uses empirical VAR models to analyze how changes in financial market volatility and monetary policy affect the term premium, which is a component of long-term interest rates that reflects the compensation investors demand for holding long-duration government securities.

Main Findings

1. Impact of Monetary Policy on Market Volatility and Term Premium

• Pre-crisis period (1988Q1–2007Q4):
  An unexpected loosening of monetary policy (e.g., a cut in the federal funds rate) leads to a decline in both expected stock and bond market volatilities, as well as the term premium.

  • Conventional monetary policy is found to have a statistically significant real effect on economic activity.
  • The term premium is negatively correlated with monetary policy easing.

• Post-crisis period (2008M1–2013M12):
  An unexpected loosening of monetary policy (e.g., an increase in bond purchases) also leads to a decline in both expected stock and bond market volatilities and the term premium.

  • However, unconventional monetary policy (such as quantitative easing) is found to have no statistically significant real effect on economic activity, despite a significant negative impact on the term premium.
  • This suggests that the effectiveness of quantitative easing in stimulating the real economy may be limited.

2. Role of Expected Volatility in the Term Premium

• VIX (Stock Market Volatility):

  • Expected equity market volatility (VIX) is more important than bond market volatility (MOVE) in influencing the term premium.
  • During the pre-crisis period, a VIX shock leads to a rise in bond volatility (MOVE), a contraction in economic activity, a decline in broker-dealer leverage, and an increase in the term premium.
  • This is consistent with pro-cyclical swings in market liquidity.
  • In the post-crisis period, a VIX shock is associated with a drop in the term premium, indicating a "flight to quality" or a shift from riskier assets to safer Treasury bonds.
  • The VIX is a model-free volatility index, reflecting investor fear due to its negative correlation with S&P 500 return dynamics.

• MOVE (Bond Market Volatility):

  • A shock to the MOVE leads to little or no change in the term premium.
  • It has no statistically significant effect on other variables, both pre- and post-crisis.
  • The MOVE is an average implied normal volatility for a one-month forecast horizon and is based on at-the-money options using the Black (1976) model.

3. Leverage and Market Liquidity

• Broker-dealer leverage:

  • Leverage responds to both monetary policy shocks and VIX shocks.
  • In the pre-crisis period, it increases with monetary easing and is negatively correlated with the VIX.
  • In the post-crisis period, it decreases in response to asset purchases, suggesting a reduced importance of the market-making role of bond dealers.

• Primary dealers' net position:

  • Used as a proxy for leverage in the post-crisis weekly model.
  • Their net positions (long minus short) are more informative about risk-taking than simple leverage measures.

4. Model Structure and Data

• The study uses a VAR model with different time frequencies for pre- and post-crisis periods:

  • Quarterly model for 1988Q1–2007Q4.
  • Weekly model for 2008M1–2013M12, incorporating high-frequency data on asset purchases and primary dealers' positions.

• The VAR model includes the following variables:

  • Economic activity (industrial production index and ADS index).
  • CPI inflation.
  • Federal funds target rate (pre-crisis).
  • Fed asset purchases (post-crisis).
  • VIX and MOVE as measures of expected market volatility.
  • Broker-dealer leverage and primary dealers' net position.
  • US term premium.

• The Cholesky decomposition is used to identify structural shocks, with the variables ordered in a recursive sequence to ensure identification.

5. Robustness and Implications

• The results are robust across different identification mechanisms and model specifications.
• The term premium is found to be counter-cyclical with respect to real economic conditions and negatively correlated with monetary policy.
• The VIX is more relevant than the MOVE in determining the term premium.
• The findings suggest that quantitative easing (unconventional monetary policy) may not have the same real economic impact as conventional monetary policy, despite affecting the term premium.
• The flight to quality effect is more prominent in the post-crisis period, as investors prefer safer assets like Treasury bonds during times of increased uncertainty.

Key Information

• JEL classification: E43, E44, E52
• Keywords: bond market volatility, VIX, unconventional monetary policy, quantitative easing, long-term interest rate, term premia
• Time periods analyzed:
  • Pre-crisis: 1988Q1–2007Q4
  • Post-crisis: 2008M1–2013M12
• Data sources: Bloomberg, Datastream, Chicago Board Options Exchange, Merrill Lynch, and Federal Reserve Bank of New York.
• Methodology: VAR models with recursive ordering and Cholesky decomposition for shock identification.
• Main variables:
  • Term premium (10-year US Treasury yield minus the risk-free rate).
  • VIX and MOVE as measures of equity and bond market volatility.
  • Fed funds rate, asset purchases, and leverage indicators.
• Conclusion:
  • Monetary policy affects the term premium, but its real economic impact is less pronounced in the post-crisis period.
  • Expected equity volatility (VIX) is more important than bond market volatility (MOVE) in influencing the term premium.
  • The study highlights the need to consider risk-taking and uncertainty in the analysis of monetary policy transmission.

Structure of the Paper

1. Introduction:

   • Motivates the study by emphasizing the importance of market volatility and monetary policy in shaping the term premium.
   • Notes the lack of attention to bond market volatility and term premium in existing literature.

2. Empirical Model and Data:

   • Describes the VAR model used for analysis, including variable selection and identification methods.
   • Discusses the use of the VIX and MOVE indices, as well as the ADS index and Fed asset purchases.

3. Results from VAR Models:

   • Presents impulse response functions for pre- and post-crisis periods.
   • Highlights the differential impact of monetary policy and market volatility on the term premium.

4. Robustness Checks:

   • Conducts tests using alternative identification methods, including time-varying parameter structural VAR.
   • Finds that the results are robust to different model specifications.

5. Conclusion:

   • Summarizes the main findings and implications for monetary policy and market volatility.
   • Suggests that risk and uncertainty channels are important in the transmission of monetary policy, especially in the post-crisis context.