[论文解读] Testing the effectiveness of unconventional monetary policy in Japan and the United States

Daisuke Ikeda, Shangshang Li+2 more|arXiv (Cornell University)|Dec 30, 2020

Monetary Policy and Economic Impact参考文献 30被引用 33

一句话总结

The paper develops a DSGE model and uses censored/kinked VARs to test the ELB irrelevance hypothesis, finding strong delayed effects of unconventional monetary policy (UMP) and rejecting ELB irrelevance for both Japan and the U.S.

ABSTRACT

Unconventional monetary policy (UMP) may make the effective lower bound (ELB) on the short-term interest rate irrelevant. We develop a theoretical model that underpins our empirical test of this `irrelevance hypothesis' based on the simple idea that under the hypothesis, the short rate can be excluded in any empirical model that accounts for alternative measures of monetary policy. We test the hypothesis for Japan and the United States using a structural vector autoregressive model with the ELB. We firmly reject the hypothesis but find that UMP has had strong delayed effects.

研究动机与目标

• Motivate whether unconventional monetary policy can circumvent the effective lower bound (ELB) constraints.
• Develop a DSGE framework that embeds QE and forward guidance and derives testable VAR representations.
• Design empirical tests using censored and kinked VARs to assess the irrelevance of the ELB.
• Assess whether the ELB alters the short- and long-rate responses to monetary shocks across Japan and the U.S.

提出的方法

• Construct a New Keynesian DSGE model with QE and FG operating under the ELB; define the shadow rate i_t* and the ELB constraint i_t = max{i_t*, i_lower}.
• Derive the effective interest rate as (1−λ*)i_t + λ*i_t*, with λ* capturing QE effectiveness and forward guidance.
• Show that under ELB irrelevance (ξ* = 1) the model yields VAR(1) representations for key variables, with the shadow rate censored at the ELB.
• Use censored and kinked structural VARs (Mavroeidis 2021) to test the irrelevance hypothesis by excluding the short rate from VARs that include the long rate or the shadow rate.
• Impose sign restrictions inspired by the DSGE to sharpen impulse-response identification and quantify attenuation vs. non-attenuation of long-rate responses.

实验结果

研究问题

• RQ1 Does the ELB render conventional policy ineffective, or can UMP fully circumvent it (the irrelevance hypothesis) in practice?
• RQ2 What are the dynamic effects of UMP on inflation and output in ELB vs non-ELB regimes for Japan and the United States?
• RQ3 How does the long-term rate respond to monetary policy shocks when the ELB is binding, and is there attenuation?
• RQ4 Do shadow-rate based measurements and long-rate measurements provide consistent identification of monetary policy effects under ELB?
• RQ5 How do the estimated impulse responses differ across countries and over time under UMP?

主要发现

• The irrelevance hypothesis is strongly rejected for both Japan and the United States.
• Monetary policy shocks have strong delayed effects under UMP in both countries.
• Impact effects on inflation and output within one quarter decline by more than 15% in the U.S. and more than 50% in Japan when the ELB is binding, relative to conventional policy.
• Cumulative effects over 1–2 years can be stronger under ELB regimes than non-ELB regimes, except for U.S. output gap response which remains weaker.
• UMP shows delayed but stronger effects on inflation (and output in Japan) compared to conventional policy, and responses differ across time and country.
• The results reject the ELB irrelevance, i.e., the ELB has constrained policy effectiveness in the observed data.