Cross-country Transmission Effect of the U.S. Monetary Shock under Global Integration

Cross-country Transmission Effect of

the U.S. Monetary Shock

under Global Integration

Yoshiyuki Fukuda

yoshiyuki.fukuda@boj.or.jp

Yuki Kimura

yuki.kimura@boj.or.jp

Nao Sudo

***

nao.sudou@boj.or.jp

Hiroshi Ugai

****

hiroshi.u[email protected].jp

No.13-E-16

November 2013

Bank of Japan

2-1-1 Nihonbashi-Hongokucho, Chuo-ku, Tokyo 103-0021, Japan

International Department

Financial System and Bank Examination Department

***

Financial System and Bank Examination Department

****

International Department

Papers in the Bank of Japan Working Paper Series are circulated in order to stimulate discussion

and comments.

Views expressed are those of authors and do not necessarily reflect those of

the Bank.

If you have any comment or question on the working paper series, please contact each author.

When making a copy or reproduction of the content for commercial purposes, please contact the

Public Relations Department (post.prd8

@boj.or.jp) at the Bank in advance to request

permission. When making a

copy or reproduction, the source, Bank of Japan Working Paper

Series, should explicitly be credited.

Bank of Japan Working Paper Series

Cross-country Transmission E¤ect of the U.S.

Monetary Shock under Global Integration



Yoshiyuki Fukuda

, Yuki Kimura

, Nao Sudo

and Hiroshi Ugai

{

November 26, 2013

Abstract

Monetary policy shocks in the United States are considered a signi…cant cause

of economic ‡uctuations in other countries. We study empirically how the spillover

e¤ects of such shocks have changed as a result of the recent deepening of global

integration. We consider shocks to the Federal Funds rate and examine how do-

mestic production in a number of advanced, Latin American, and Asian countries

were a¤ected by these shocks during the 1990s and 2000s. We show that contrac-

tionary U.S. monetary policy shocks reduced domestic production in most of the

sampled countries during the 1990s. During the 2000s, by contrast, the adverse

e¤ects were moderated. To explore the reasons behind the weakened spillover ef -

fects, we construct a DSGE mod el and examine the theoretical implications of the

recent changes in economic structure, including global integration. In addition, we

estimate response of trade and …nancial variables as well as policy instruments to

U.S. monetary policy shocks. Our model combined with the empirical exercises

suggests that, despite being enhanced by deepened trade integration, spillover ef-

fects may be decreasing due to a decline in the relative importance of the U.S.

economy, and to regime switches in domestic monetary and exchange rate policy



The authors would like to thank Rasmus Fatum, Ichiro Fukunaga, Jan Groen, Naohisa

Hirakata, Takushi Kurozumi, Jouchi Nakaj ima, Akira Otani, Toyoichiro Shirota, Masashi Saito,

Tomohiro Sugo, Kozo Ueda, Yohei Yamamoto, and sta¤ of the Bank of Japan, for their useful

comments. The views expressed in this paper are those of the authors and do not necessarily

re‡ect the o¢ cial views of the Bank of Japan.

Deputy Director, International Department, Bank of Japan (E-mail:

yoshiyuki.fukuda@boj.or.jp).

Financial System and Bank Examination Department, Bank of Japan (E-mail:

yuki.kimura@boj.or.jp).

Director, Financial System and Bank Examination Department, Bank of Japan (E-mail:

nao.sudou@boj.or.jp).

{

Deputy Director-General, International Department, Bank of Japan (E-mail: hi-

roshi.ugai@boj.or.jp).

in non-U.S. countries. Though we empirically …nd a sign of short-run …nancial

contagion during the 2000s, its e¤ect upon the real economy was minor, possibly

re‡ecting its low persistence.

Keywords: U.S. Monetary Policy; Spillover E¤ect; Financial and Trade Linkages

1 Introduction

Global integration has increasingly deepened in terms of trade and …nancial transactions

over the current years. We display in Figure 1 the evolvements of the degree of the

two classes of integration, measured by total trade volume and banks’ cross border

claims relative to the world GDP over time. The trade integration has increased steadily

throughout the last two decades, enhancing the degree of integration in 2010 1.5 times

larger than that in the early 1990s. The …nancial integration has also increased so that the

degree of integration has doubled over the twenty years. One important consequence of

such a global integration is changes in the spillover mechanism across countries. That is,

a disturbance originated in one country today may be transmitted to the other countries

in a di¤erent manner from the early days. Along this line, a growing attention is paid

to the transmission of the U.S. monetary policy shocks in the global economy, re‡ecting

the disproportionately large size and the interconnectedness with the rest of the world.

Theoretically, there are two main channels in the transmission of the U.S. monetary

policy shock to the rest of the world. The …rst channel works through a change in the

nominal exchange rate. The text book Mundell–Flemming–Dornbusch (MFD) model

predicts that the U.S. monetary shock a¤ects expenditure decisions of households and

…rms in the U.S. and non-U.S. countries. For instance, other things being equal, a

contractionary U.S. monetary policy shock enhances a value of the U.S. dollar with

respect to the non-U.S. currency and makes the relative price of the U.S. goods higher. As

expenditure demand shifts toward non-U.S. goods, the domestic output in the non-U.S.

countries increases (expenditure-switching e¤ect). The second channel works through

a change in the real interest rate. Because of its dominant role in the international

…nancial market, the U.S. monetary policy shock a¤ects the world interest rate. When

domestic …nancial market is open to the U.S. and to the rest of the globe, a change in the

world interest rate a¤ects the intertemporal expenditure decisions of agents in non-US

countries. A higher world interest rate stemming from the contractionary U.S. monetary

policy shock therefore may dampen the domestic exp enditure in the non-U.S. countries

as well as that of the U.S. (expenditure-reducing e¤ect).

The spillover e¤ect of the U.S. monetary policy shock has been empirically explored

by a good number of studies, such as Kim (2001), Canova (2005), and Mackowiak (2007).

These existing studies agree that the spillover e¤ect is an important source of business

cycle ‡uctuations in the non-U.S. countries. The relative importance of the two channels

are, however, often mixed across countries and ranges of time horizon studied. Kim

(2001) and Bluedorn and Bowdler (2006) claim that the expenditure-reducing e¤ect gen-

erally dominates the expenditure-switching e¤ect based on the analysis for the non-U.S.

G7 countries. By contrast, Canova (2005) and Mackowiask (2007) point out that the

latter e¤ect prevails in some emerging countries. From the perspective of changes in

spillover e¤ects over time, Liu, Mumtaz, and Theophilopoulou (2011) investigate in‡u-

ence of the world interest rate shocks on the U.K. output and report that the former

e¤ect has been dominant until the 1990s and the e¤ect has become nonsigni…cant during

the 2000 and beyond. Ilzetzki and Jin (2013) study the spillover e¤ects of shocks to the

U.S. monetary policy, upon other advanced countries and observe that the former e¤ect

has been dominant before 1990 and the latter e¤ect has taken over after 1990.

In this paper, we revisit the issues of spillover e¤ects of the U.S. monetary policy

shock by shedding light on the in‡uence of the recent deepening of global integration.

To this end, we distill a time series of exogenous shocks to the Federal Funds rate in

the U.S. economy by the factor augmented VAR developed by Bernanke, Boivin, and

Mihov (2005) and extended by Boivin, Giannoni, and Mihov (2009).

We then estimate

response of macroeconomic variables in the non-U.S. countries, including advanced, Latin

American, and Asian countries, to the distilled U.S. monetary policy shocks following a

methodology proposed by Romer and Romer (2001). To gauge the impacts of deepened

global integration over the current years, we split our sample period into two sub-sample

periods, from January 1990 to December 1999 and from January 2000 to December 2007.

We estimate the responses for the two sub-sample periods separately. Our key …nding

is the changes in the spillover e¤ects. We …nd that the domestic production in most

of the non-U.S. countries fall in response to the contractionary U.S. monetary shocks

in a statistically signi…cant manner during the 1990s. During the 2000s, by contrast,

these adverse e¤ects are less pronounced. That is, only a limited portion of countries see

statistically signi…cant adverse consequence in the domestic pro duction and the estimated

magnitudes of the recessionary impacts across countries are reduced during the 2000s

compared with the 1990s.

Why under the global integration has the spillover e¤ects become weaker? Do tight-

ened trade and …nancial linkages dampen the spillover e¤ects instead of enhance them?

To address this question, we theoretically explore implication of two other developments

of global economic structure, in addition to global integration, that have advanced dur-

ing the last two decades. The …rst development is a change in the relative signi…cance

of the U.S. economy in the global economic activities. Figure 2 displays the evolvement

of the relative country size of the U.S. GDP compared with the world GDP. The rel-

Along the similar line, Dées and Saint-Guilhem (2009) investigate the spillover e¤ects of the shocks

to the U.S. GDP to the rest of the world and report that a change in the U.S. GDP delivers a weaker

but persistent impact on the non-U.S. e conomy in recent years.

Boivin, Giannoni, and Mihov (2009) utilize the methodology of Bernanke, Boivin, and Mihov (2005)

and distill the shocks to the U.S. monetary policy rule for the period covering from Janurary 1976 to

June 2006. We construct the balanced panel following Boivin, Giennoni, and Mihov (2009) for a period

that covers from Janurary 1976 to December 2007 for our estimation.

ative country size of the U.S. has been maintained around 25% throughout the 1980s

and gradually declined since the early 1990s. It has fallen at even quicker rate in the

subsequent periods, reaching 20% in the late 2000. From a di¤erent perspective, Figure

3 displays the relative importance of the U.S. economy in trade and …nancial transac-

tion. The bilateral trade volume with the U.S. relative to the total trade volume with

the rest of the world has risen from 14% in the early 1990s up to 16% in 2000 and has

continuously declined throughout the 2000s to 11%. By contrast, the relative size of

…nancial transaction with the U.S. has b een maintained around 11% up to year 2000

and has risen to 13% during the latter half of the 2000s after the sharp fall to 9% in

the early 2000s. Emergence of large economy and/or trading partner, such as Brazil and

China, together with the global …nancial integration that underscores the role of the U.S.

as the international …nancial center is the possible explanation for these changes in the

relative signi…cance of the U.S. economy in the world. The second development is the

regime switches of domestic monetary and nominal exchange rate policies in the non-U.S.

countries. As shown in Table 1, during the last two decades, some advanced countries

have adopted in‡ation targeting so as to stabilize in‡ation expectation

and some Asian

countries have adopted both in‡ation targeting policy and ‡oat nominal exchange rate

regime particularly after the Asian currency crisis in the late 1990s.

There are also

regime switches in Latin American countries. Because the U.S. monetary policy shocks

are clearly source of variations in price and exchange rate in the non-U.S. countries, the

endogenous policy responses to these variations matter to the transmission of the U.S.

monetary policy shocks to the domestic economic activities in the non-U.S. countries.

To see how these developments a¤ect the spillover e¤ects, we construct a DSGE

(Dynamic Stochastic General Equilibrium) model and derive the model’s implications

regarding these developments. Our model shows that these developments bring about two

opposing consequences to the spillover e¤ects. First, deepening of global trade integration

causes a larger decline of domestic production in the non-U.S. countries in response to the

same size of a U.S. contractionary monetary policy shock. On the contrary, a weakening

of the bilateral relationship with the U.S., a decline in the relative country size of the U.S.,

or regime switches to more ‡exible exchange rate regime or in‡ation targeting regime

moderates the adverse consequence on the domestic production. We also examine the

implication of …nancial integration by considering a case when households in the two

countries are more insured. The model implies that the adverse spillover e¤ects on the

domestic production in the non-U.S. countries becomes weaker though interest rate pass-

through b ecomes quicker with …nancial integration. According to the model analysis,

therefore, weakening of the spillover e¤ects on domestic production is the product of

changes in relative importance of the U.S., …nancial integration, and domestic policy

regimes.

See Ueda (2009) for details.

Countries listed in the table 1 are on ly for those that are studied in our empirical exercise and

therefore the coverage of the list is not comprehensive.

Having the model’s prediction in hand, we ask if the explanations are consistent with

the economic surroundings of our sampled countries and estimated impulse responses.

We construct a country-by-country data set of trade linkage, GDP, and …nancial trans-

action as well as historical episo des for each of the sampled non-U.S. countries regarding

policy regimes. We show that in most of the sampled countries, a decline of the trade

relationship with the U.S. quantitatively dominates the deepening of trade integration

with the rest of the world, a decline of the country size is slower than the U.S., and a

…nancial transaction with both the U.S. and the rest of the world increases over the last

two decades. In addition, we estimate the response of trade variables, …nancial market

variables and policy instruments to the contractionary U.S. monetary policy shock. We

observe the weakening of spillover e¤ects for the most of the trade variables and …nancial

variables. Though we …nd some evidence of increased …nancial contagion in the non-U.S.

countries due to the contractionary U.S. monetary policy shock over the two decades,

their propagation to the domestic productions are minor. This may be partly because

…nancial contagion after the shock is only short-lived. Regarding estimated resp onses

of policy instruments, we …nd that they are more contractionary during the 1990s than

2000s in Asian countries, suggesting that the policy responses may have caused weaken-

ing of spillover e¤ects in these countries.

Our paper stems from two strands of the literature. The …rst strand discusses spillover

e¤ect of monetary policy shock in one country to the other countries.

A pioneering work

by Kim (2001) utilizes a VAR methodology of Chrisitiano et al. (1996 and 1998) to ex-

plore how the expansionary monetary policy shock in the U.S. a¤ects economic activities

in the non-U.S. G7 countries. Subsequent work by Canova (2005) and Mackowiask (2007)

investigate the spillover e¤ect of the U.S. monetary policy shock on the non-G7 coun-

tries.

The second strand explores the implication of the deepening of integration in

terms of …nancial and trade linkages to the economic activities around the globe. While

existing studies agree that the integration has deepened in the current years, there is

no agreement empirically yet as to what is the fruit of integration. Kose et al. (2008),

discussing the globe is more integrated in current days than old days, decompose source

of variations in domestic variables into a global factor common to all variables and all

countries, a factor common to sp eci…c group of countries, and a country factor, and

document that the proportion of variations in domestic variables accounted for by the

country-speci…c factor has declined when current data is used for estimation. Dee and

Saint-Guilhem (2009) discuss contribution of the U.S. sho ck to output variations in the

There are several studies that discu ss the international spillover of the monetary policy shock in

the non-U.S. countries. Ma

ckowiak (2008) studies the impact of Japanese quantitative easing on East

Asian countries and cast the doubts on the view that the policy has beggar-thy-neighbor e¤ects. See

also Ko

zluk and Mehrotra (2009) for the spillover of expansionary Chinese monetary policy shock to

East Asian countries.

Along the similar line, Ncube et al. (2012) discuss the impact of the U.S. monetary policy shock on

the South African economy.

non-U.S. countries and show that the contribution has been declining over the years.

The rest of the paper is organized as follows. The section 2 describes our estimation

procedure that is employed to extract the U.S. monetary policy shocks and gauge the

macroeconomic response to the shocks. It then documents the estimated response of

domestic production in the non-U.S. countries to the shocks. The section 3 explores

the linkages between changes in economic environment and spillover e¤ects of the U.S.

monetary policy shock using a simple open economy model. Section 4 discusses the

candidate explanations for the weakened spillover e¤ects, and the section 5 concludes.

2 Estimating Spillover E¤ects

2.1 Estimation Methodology

Our estimation methodology contains two steps. We …rst distill shocks to the U.S.

monetary policy rule using the U.S. macroeconomic time series. We then estimate the

response of a set of macroeconomic variables including production-related variables, trade

and …nancial variables, and policy instruments, in the non-U.S. countries to the distilled

shock series.

Obtaining the U.S. monetary policy shock series

We identify a historical time series of the U.S. monetary policy shock by making use of

the factor-augmented VAR approach proposed by Bernenke, Boivin, and Eliaz (2005) and

Boivin, Giannoni, and Mihov (2009). Closely following their estimation methodology,

we construct a balanced panel that consists of 120 monthly macro economic time series

running from February 1976 up to December 2007, extract the time paths of 5 latent

factors from the balanced panel, and obtain a historical realization of innovations to

Federal Funds rate up to December 2007 that are orthogonal to these latent factors and

the own lag of the Federal Funds rate.

We denote the identi…ed shock series by

hereafter:

The time path of the Federal Funds rate together with the identi…ed shocks

See also Eickmeier et al. (2011) that estimate the consequence of favorable shock to the U.S.

…nancial conditions index to the non-U.S. advanced countries using time-varying FAVAR and …nd that

the degree of transmission has increased since the 1980s to the 2000s.

Bernenke, Boivin, and Eliaz (2005) display the estimation results for the case when the number of

latent factors is three and …ve. Though we choose an estimation methodology that extracts …ve factors

from the balanced panel, our estimation results are little altered under an alternative setting where

three latent factors instead of …ve latent factors are extracted.

Because our analysis concentrates on the spillover e¤ect of shocks to the conventional U.S. monetary

policy rule, we drop a sample period beyond December 2007 where the unconventional monetary policy

has been implemented. To check the robustness of our empirical analysis, however, we also conduc t the

analysis based on the shock series to the Federal Funds rates that are distilled from the sample period

that covers the time series up to December 2012. The estimated results are little altered by the use of

this alternative monetary policy shocks.

to the U.S. monetary policy rule

are displayed in Figure 4.

Estimating cross-country responses to the U.S. monetary policy sho ck

Our sample country includes four classes of countries; G7 countries other than the

U.S. plus Australia, other advanced European countries, Asian countries, and Latin

American countries. To see the impact of trade and …nancial integration, we estimate

economic dynamics of twelve macroeconomic series in these countries for the two sample

periods, the 1990s and the 2000s. The …rst sample period covers from January 1990 to

December 1999 and the second sample period covers for the January 2000 to December

2007. We drop a country from our analysis of a speci…c macroeconomic variable for a

speci…c sample period when the corresponding monthly time series is not available. Our

sample macroeconomic variable includes three production-related variables, industrial

production index, employment, and unemployment rate, three trade-related variables,

real export index, real import index, and real trade balance index, six …nancial variables,

stock price, nominal exchange rate, long term nominal interest rate, volatility stock price,

exchange rate, and long term nominal interest rate, and three policy instruments, policy

rate, narrow money, and in‡ation rate. All of the macroeconomic series other than stock

price series and interest rate series are seasonally adjusted.

Using the measure of the U.S. monetary policy shock estimated by the factor-augmented

VAR approach, we estimate the impulse response functions of macroeconomic variables in

the non-U.S. countries following closely the estimation methodology proposed by Romer

and Romer (2004). For each of the macroeconomic variables for each country and for

each sample period, we regress the macroeconomic variable x in country j on its own

lags and lagged values of the U.S. monetary policy shock,

. Our regression equation is

then given by

x

j;t

= c

`=1

^

x

j;t`

`=1



t`

+ "

; (1)

where x

j;t

denotes the …rst di¤erence of the macroeconomic variable x

j;t

; c

denotes the

constant, ^

denotes the estimated coe¢ cients attached to l-th lagged macroeconomic

variable x

j;t

in country j,



is the estimated coe¢ cient attached to l-th lag of the

estimated U.S. monetary policy shock, and "

is the innovation that is speci…c to the

macroeconomic variable x in country j:

^p and ^q stand for the number of lags attached

to the macroeconomic variable and the U.S. monetary policy shock, respectively. In all of

the estimations, twelve lags for each macroeconomic variable x and thirty six lags for the

U.S. monetary policy shock are included.

When estimated parameters are explosive,

All of the response of macroeconomic variables are estimated in the log di¤erence form, except f or

interest rates and the unemployment rate for which log form is employed.

The di¤erence between Romer and Romer (2004) and our speci…cation is that numbers of lags and

the use of seasonally adjusted variables. Romer and Romer (2004) make use of settings where ^p = 24

however, shorter lags are included.

2.2 Estimation Results

To gauge the changes in domestic resp onse to the U.S. monetary policy shock in the non-

U.S. countries in details, we construct two statistics. The …rst statistics is a proportion of

countries that experience a statistically signi…cant increase or decrease at the signi…cant

level of 5%, or countries that see no statistically signi…cant deviation from zero after the

shock among countries that b elong to the same country group. This statistics conveys

information regarding signs of the responses that are statistically meaningful.

The

second statistics computes distribution of impulse response of macroeconomic variables

across countries. While the sizes and signs of domestic responses exhibit a substantial

cross-country heterogeneity, this statistic conveys overall distribution of the estimated

responses. To obtain the statistics, for each sampled country, we collect a value of

estimated impulse response at four years after the sho ck and construct the distribution

using the responses for the sample period covering the 1990s and the 2000s.

13 14

Figure 5 displays the response of macroeconomic variables that capture the size of

domestic production, industrial production index (IIP), employment, and unemployment

rate, in the non-U.S. countries to the contractionary U.S. monetary policy shock. To

highlight the heterogeneity of responses across di¤erent country group and subsample, we

categorize sampled countries into …ve country groups, G6 + Australia, Other advanced

European countries, Latin American countries, Asian countries, and all countries, and

document the results for the sample covering the 1990s and the 2000s separately.

For

each variable, the upper panel displays a proportion of countries within each country

group that experience a statistically signi…cant positive response, depicted in a black

bar, signi…cant negative response, depicted in a gray bar, and no signi…cant di¤erence

from zero, depicted in a white bar. The lower panel displays the distribution of estimated

impulse response in months around four years after the impact period across all of the

and ^q = 36 for output and ^p = 24 and ^q = 48 for price level and use non-seasonally adjusted variables

for estimation.

To compute the con…dence interval of the impulse response fun ction of each of domestic macroeco-

nomic variables to the shock, we assume that innovations "

in the equation (1) is normally distributed

and there is no measurement errors in es timating

t`

in the …rst s tep.

We choose four years horizon for the purpose of comparison with existing studies including Kim

(2001) and Canova (2005).

For variables which we use its …rst di¤erence in estimating the impulse response, we convert the

impulse response into level by taking the cumulative sum of the impulse response over a period from the

initial period up to a k period after the shock. For variables which we do use level variable in estimating

the impulse response, such as unemployment rate, we do not take the cumulative sum but instead use

the impulse response function itself in calculating the numbers.

See Table 9 for a list of countries that are studied in this paper. Note that because the length of

time series of macroeconomic variables available for the analysis di¤ers across countries, the number of

listed countries is not equal across macroeconomic variable.

sampled countries.

In each panel, the X-axis denotes the estimated size of impulse

responses and Y-axis denotes marginal density of the corresponding size of the estimated

impulse responses.

During the 1990s, for all of the three production-related variables and in each of the

…ve country groups, the most of countries witness signi…cant recessionary impact by the

contractionary U.S. monetary policy shock. By contrast, during the 2000s, such adverse

e¤ects are observed only in a limited number of countries. The mitigation of the recession-

ary e¤ects is pronounced the most in advanced countries and less so in Latin American

and Asian countries. The similar observation is obtained from the cross-country distrib-

ution of estimated impulse responses. During the 1990s, estimated impulse responses in

majority countries are concentrated around points below zero for I IP and employment

and above zero for the case of unemployment rate, indicating that general in‡uence of the

contractionary U.S. monetary policy shock on domestic production activities in the non-

U.S. countries is contractionary. During the 2000s, by contrast, the peaks of marginal

densities are ‡attened for all of the production-related variables and they are shifted to

the right for IIP and employment and to the left for unemployment rate. These …ndings

suggest that the contractionary spillover e¤ect is weakened.

Two remarks are noteworthy regarding the estimated results. First, the results are

consistent with early studies including Kim (2001) and Bluedorn and Bowdler (2011).

They both document the quantitative dominance of expenditure-reducing e¤ect on expenditure-

switching e¤ect in the transmission of the contractionary U.S. monetary policy shock to

the non-U.S. countries. Our contribution regarding this point is to have shown that

the same statement holds true in a broader set of countries that contains Latin Ameri-

can and Asian countries. Second, our result indicates that the e¤ects of contractionary

U.S. monetary policy shocks to the domestic production in the non-U.S. countries have

changed over time. The adverse spillover e¤ects have become less contractionary during

the 2000s than 1990s. This …nding is in line with observations made in Liu, Mumtaz, and

Theophilopoulou (2011) for the U.K.

In the subsections below, we investigate reasons

To obtain cross-country distribution of responses, we employ averaged impulse response fun ctions

over 40th month to 50th month after the shock for all of the macroeconomic variable except for the long

and short term interest rate. Because the impulse responses of these variables are short-lived, we make

use of averages over periods shortly after the impact period, from 1st month to 24th month.

Note that our estimation results only imply that contractionary U.S. monetary policy shocks identi-

…ed in the way describ ed above may deliver adverse e¤ect on production-related variables in the non-U.S.

countries and do not indicate that a rise in the Federal Fun ds rate itself hampers domestic productions

in these countries.

Our estimation result is, in some aspects, analogous to that obtained by Ilzetzki and Ji (2013).

They document that in recent years there has b ee n switches in the relative quantitative importance of

expenditure-switching e¤ect and expenditure-reducing e¤ect in the transmission of the U.S. monetary

policy shock to abroad. That is, the contractionary U.S. monetary policy shock delivers contractionary

e¤ects in the non-U.S. countries during the early years while delivers favorable e¤ects during the current

years. Admittedly, some of our sampled countries exhibit the similar pattern of the responses as theirs.

It is important to note, however, that in our analysis the most of our sampled countries see adverse

behind the changes in adverse e¤ects of the contractionary U.S. monetary policy shock

on the production activities in the non-U.S. economy.

3 Explanations for Weakened Spillover E¤ects and

Their Theoretical Implications

Why has the spillover e¤ects changed over time? To see the reasons, in this section, we

theoretically analyze the implications of three recent developments of economic struc-

ture to the spillover e¤ects using a plain vanilla Dynamic Stochastic General Equilibrium

(DSGE) Model. The developments that we focus include a change in the relative sig-

ni…cance of the U.S. economy in the global economic activities, the regime switches of

domestic monetary and nominal exchange rate policies in the non-U.S. countries, as well

as deepening of trade and …nancial integration. Clearly, these developments may be

interrelated from each other. For instance, deepening of global integration and less re-

liance on the trade with the U.S. originate from the same economic reason. We however

treat these developments as exogenous in the model and investigate how these exogenous

developments are related to the spillover e¤ects.

Our model is built upon the text book open economy model studied by Obstfeld and

Rogo¤(1996) or Corsetti and Pesenti (2001a,b). Our model consists of three countries, A,

B, and C. We analyze how responses of domestic production in country B to a monetary

policy shock in country A vary with degree of global integration, relative importance

of the trade relationship with country A, and domestic policy regimes in country B.

Admittedly, the transmission mechanism between country A and B depends on economic

structure surrounding country C as well. Because of the three-country framework, our

model is able to capture the third-country e¤ects. The model answers to a question such

as how an increase in country size of country C in‡uences the monetary transmission

from country A to country B.

3.1 Household

Because all of the three countries are symmetric, we describe settings for households,

…rms, and government sector only in country A unless otherwise noted.

A represen-

tative household in country A is an in…nitely-lived representative agent with preference

over the …nal consumption goods, C

) and work e¤ort, h

), as described in the

expected utility function, (2) :

consequence during both of the two sample periods and experience the mitigation of the adverse e¤ect

in the current years maintaining the qualitative aspect of the shock’s e¤ect unaltered.

Only di¤erence between country A and other two countries is that the internationally traded bond

is denominated by the currency in country A: This di¤erence, however, does not a¤e ct the spillover

e¤ects considered in the current paper.

= E

j=0



t+j

)  bC

t+j1

)]

1

1  

 

t+j

))

1+

1 + 

; (2)

where s

stands for a state in period t;  2 (0; 1) is the discount factor,  > 0 is the inverse

of intertemporal elasticity of substitution, b 2 [0; 1) is the persistency of consumption

habit,  > 0 is the inverse of the Frisch labor-supply elasticity, and  > 0 is the utility

weight assigned to labor disutility.

The budget constraint for the household is

) C

)

) + B



)



) h

) + R

t1

) B

t1

)



t1

) B



t1

) (1  

t1

)) :

(3)

Here P

) is the nominal price of …nal consumption goods C

) ; W

) is the

nominal wage rate paid for the work e¤ort h

) : Each household is the monopolistic

supplier of a di¤erentiated work e¤ort to domestic intermediate goods …rms discussed

below. Wage contracts are subject to nominal rigidity and a household chooses her

wage so as to maximize her utility under the constraint.

Household holds two assets,

domestically traded bond B

) and internationally traded bond B



) : The former

bond is in zero net supply at each country level and the latter bond is in zero net

supply worldwide. For simplicity, we assume that the internationally traded bond is

denominated in the currency of country A: The corresponding short-term nominal rates

paid for the two bond holdings are denoted as R

t1

) and R



t1

) ; respectively.

Similarly to Laxton and Pesenti (2003) and Hirakata, Iwasaki, and Kawai (2013),

there are frictions associated with the transaction of the internationally traded bond;

such that





= 



exp



^B











 1





exp



^B











+ 1



;

where  and ^ are parameters that govern the cost of accessing the international bond

market and 

) stands for size of the cost facing the household.

3.2 Firms

There are two types of …rms, distributors and intermediate goods producers in each of

the countries. Most of our settings are borrowed from Bodenstein, Guerrieri, and Gust

(2010), Laxton and Pesenti (2003) and Hirakata, Iwasaki, and Kawai (2013).

See for instance Fueki, Fukunaga, Ichiue, and Shirota (2010) for the details of the nominal wage

settings.

Laxton and Pesenti (2003) assume that parameter value of  and ^ are 0.05 and 0.01, respectively. As

we study the recent economy after the 1990s, considering the current technological progress in …nancial

transaction, we set the …rst parameter at lower value than theirs.

Distributors

The distributors purchase di¤erentiated intermediate goods produced by the inter-

mediate goods producers located in each country, combine the di¤erentiated goods to

produce intermediate composites, convert the intermediate composites to the …nal con-

sumption goods, and sell the …nal consumption goods to the domestic households. They

are perfectly competitive both in their input and output market.

The production technology that yields the …nal consumption goods C

) is given

by the following equation:







)

=(1+)

))

=(1+)

+ (n

)

=(1+)

))

=(1+)

+ (n

)

=(1+)

))

=(1+)

1+

; (4)

where C

) ; C

) ; and C

) are intermediate composite that is converted

from the di¤erentiated intermediate goods produced in country A, B; and C; respectively.

; n

; and n

are positive weight assigned to each intermediate composite and  > 0

is the parameter that governs substitutability across three intermediate composite.

22 23

Each intermediate composite is produced from di¤erentiated intermediate goods with

di¤erent origin. The corresponding production technology is given by











i; s



1+



1+

;











i; s



1+



1+

; and











i; s



1+



1+

;

where y

(i; s

) ; y

(i; s

) ; and y

(i; s

) is di¤erentiated intermediate goods that is

produced by the di¤erentiated intermediate goods producer i in country A; B; and C: The

parameter  > 0 governs the elasticity of substitution between di¤erentiated intermediate

goods. Taking price of these di¤erentiated intermediate goods p

(i; s

) ; p

(i; s

) ; and

(i; s

) as given, distributors determine how much the intermediate composite and the

…nal consumption goods to produce.

We assume that n

= 1 and that the similar constraint holds for other two countries.

In the current model, we denote intermediate composite that is converted from di¤erentiated inter-

mediate goods produced in country k and used for prod uc ing …nal consumption goods in country l in

period t by C

) : The weight attached for the intermediate composite is correspondingly depicted as

Intermediate goods producers

Firms in country A produce di¤erentiated products indexed by i 2 [0; 1]. As explained

above, each di¤erentiated goods is converted by the local distributors in the destination

country to the intermediate composite C

) ; C

) ; and C

). Note that these

intermediate composite are produced from the di¤erentiated goods via the following

technology:











i; s



1+



1+

;











i; s



1+



1+

; and











i; s



1+



1+

;

where y

(i; s

) ; y

(i; s

) ; and y

(i; s

) is portion of di¤erentiated goods produced

by the di¤erentiated intermediate goods producer i that is delivered to country A; B;

and C : The demand function for the di¤erentiated intermediate goods is derived from

the optimization behavior of the distributor in each destination country, represented by

(i; s

) 



(

i;s

)



(1+)



) ;

(i; s

) 



(

i;s

)



(1+)



) ; and

(i; s

) 



(

i;s

)



(1+)



) :

(5)

where p

(i; s

) ; p

(i; s

) and p

(i; s

) are the nominal price of di¤erentiated inter-

mediate goods produced by …rm i; and P

), P

), and P

) are the nominal

prices of the intermediate composite C

) ; C

) ; and C

). All prices are

denominated in the currency of destination country.

Production input of di¤erentiated intermediate goods is labor input h

(i) provided

by household in country A and the production function of …rm i is given by:



i; s



+ y



i; s



+ y



i; s



 h

(i) : (6)

Each di¤erentiated intermediate …rm i is a monopolistic competitor in the domestic

and overseas products market, setting price of its product in the destination country in

reference to the demand equation given by (5) subject to the nominal rigidity associated

with price adjustment. That is, …rm i faces a quadratic cost à la Rotemberg (1982)

in adjusting its product price, and solves the following pro…t maximization problem in

choosing its price p

(i; s

) :

max

p(i;s

t+j

)

j=0





j1;j

(

i;s

t+j

)

t+j

)

(i; s

t+j

) 

(

t+j

)

t+j

)

(i; s

t+j

)







(

i;s

t+j

)

(i;s

t+j1

)

 1



(i; s

t+j

)

Here  is the parameter that governs size of the real cost associated with the prod-

uct price adjustment and 

j1;j

stands for the Lagrangian multiplier associated with a

representative household’s budget constraint. The prices of di¤erentiated intermediate

goods shipped to overseas p

(i; s

) and p

(i; s

) are also chosen subject to the similar

nominal rigidity constraint (local currency pricing).

3.3 Government and Resource Constraint

Monetary Policy

The central bank in country A sets the short term nominal interest rate according to

a simple Taylor rule given by

ln R





(1  

) ln R + 

ln R

t1

)

+ (1  

)







ln 

) + 

(

)



+ 

)

; (7)

where R is the short term nominal interest rate at the steady state, 

2 [0; 1] is the au-

toregressive coe¢ cient of the rate, 



> 1; 

> 0 are the policy weight attached to the in-

‡ation rate 

) of …nal consumption goods; which is de…ned as P

) =P

t1

) ; and

output gap measured by the deviation of the …nal consumption goods from its steady

state value denoted as ln(C

) =C

)) ; and 

) is an i.i.d. shock to the monetary

policy rule in country A:

Evolution of bond of international capital market

The net holding of internationally traded bond in country A evolves according to the

following law of motion.









= 





t1







t1



1  



t1



+

) C

) ="

)  

) C

)

+

) C

) ="

)  

) C

)

(8)

where "

) and "

) are nominal exchange rate that is expressed in a value of

currency A in terms of currency B and C, respectively. Here 

; 

; and 

are scalers

that determine the relative size of the three countries. The …rst term in the right hand

side of the equation denotes the net return from holding internationally traded bond



) and the sum of the last four terms denotes the net trade balance between the

country A and the rest of the world.

Asset market clearing condition

Asset market clearing condition is given by







+ B







+ B







= B





= B





= B





= 0:

Resource constraint

Resource constraints of goods produced in each of the three countries are given by





i; s



di = 





+ 





+ 





;





i; s



di = 





+ 





+ 





; and





i; s



di = 





+ 





+ 





3.4 Domestic Response to a Foreign Monetary Policy Shock

Using the model described above, we study how domestic response in the non-U.S. coun-

tries to a monetary policy shock in the U.S. is altered with changes in economic structure.

To this end, we compute equilibrium response of macroeconomic variables in country B

to a positive innovation 

) to the monetary policy rule in country A speci…ed in

equation (7) :

The computed equilibrium time paths are linearly approximated around the non-

stochastic steady state. The non-stochastic steady state is de…ned as a competitive

equilibrium with B



) and B

) are zero for k = A; B; and C, such that for,

all t, ffC

) ; y

) ; P

) =P

l=A;B; and C

; C

) ; y

(i; s

) ; W

) =P

) ;

) ; R

k=A;B; and C

, "

) P

) =P

) ; "

) P

) =P

) are constant.

In the simulation exercise, we compare two distinct economies. In the …rst economy,

which we call baseline economy, all parameters are calibrated to standard values. When

the appropriate data is available, the parameter values are chosen such that the economic

structures surrounding country A are consistent with the U.S. during the 1990s. In the

second economy, a subset of parameter values or model settings are altered so as to

characterize changes in economic structure that is focused. See Table 2 for parameter

values used for baseline model.

The role of trade linkages with country A in monetary transmission

The steady state cond ition that B



) for k = A; B; and C is zero implies that the trade is balanced

for all of the three countries.

We …rst discuss impacts of weakened trade linkage with the U.S. on the spillover

e¤ect. A trade is a leading explanation for international business cycle comovement.

Baxter and Kouparitsas (2005) empirically investigate business cycles of more than 100

countries and document that bilateral trade relationship is the only variable that accounts

for the comovement in a statistically robust manner.

That is, other things being equal,

tighter bilateral trade linkages leads to higher business cycle comovements across the two

countries. The degree of bilateral trade linkage in the model is governed by parameters in

aggregation equation (4) of …nal consumption goods such as n

and n

: For instance, a

smaller value for n

implies that di¤erentiated intermediate goods produced in country

B is less needed in constructing the …nal consumption goods in country A: We analyze

equilibrium responses under the two alternative economies together with the baseline

model. In the …rst alternative economy, values for n

and n

are reduced to 70% of

the baseline parameter values so that the bilateral trade linkage between country A and

B is weaker. In the second alternative economy, a value for n

; together with a value

of n

and n

; is reduced to 70% of the baseline parameter value so that the bilateral

trade linkage between country A and C is weaker while that between country A and

B is maintained the same as the …rst alternative economy. 70% is taken from Figure

3. Simulation exercise using the …rst economy illustrates the direct impact of weaker

trade linkage with country A on economic activities in country B. Simulation exercise

using the second economy illustrates the indirect impact of weaker trade linkage between

country A and C on economic activities on country B:

Figure 6 displays response of domestic production 

(i; s

) di, short-term real

interest rate R

) =

t+1

), trade balance, which is measured by the sum of the

last four terms of the right hand side of the equation (8) in country B divided by

) ; nominal exchange rate (a value of one unit of currency in country A measured

by domestic currency of country B) "

) ; terms of trade (price of exported goods to

country A relative to that of imported goods from country A) "

) P

) =P

) ;

and in‡ation rate 

) to a positive shock to the monetary p olicy rule in country A

under the baseline economy as well as the two alternative economies. All of the series

are expressed in terms of log-deviation from the non-stochastic steady state.

The two e¤ects highlighted in the introduction are seen in the …gure. A contractionary

monetary policy shock in country A raises the domestic real interest rate and dampens

production and consumption in country A: In the open economy environment, these

domestic developments in country A are translated to the rest of the world through

the trade and …nancial linkages. In particular, the rise in the real interest rate of the

internationally traded bond is re‡ected in the domestic real interest rate in country B,

hampering production and consumption in country B: Although the nominal exchange

rate of currency in country B depreciates against currency in country A, making export

They also study other variables including total trade in each country, sectoral structure, similarity

in export and import baskets, factor endowments, and gravity variables, and …nd the other variables do

not account for the comovement in a statistically robu st manner.

goods shipped from country B cheaper compared with goods shipped from country A, the

expenditure-reducing e¤ect dominates the expenditure-switch e¤ect, lowering production

in country B: The roles of bilateral trade linkage are illustrated in the equilibrium time

paths of variables under the alternative economies. With a weaker bilateral trade linkage

between country B and A; less portion of domestic production and consumption in

country B are a¤ected by shocks to country A: Consequently, the adverse e¤ect of the

contractional monetary policy shock in country A is softened. As depicted in dotted

line, the trade structure between country A and C also matters to the spillover e¤ect

from country A to country B: When the bilateral trade linkage between these countries

is weak, a downturn in country C following the shock in country A and subsequent

transmission e¤ect from country C to country B are both diminished, mitigating the

adverse consequence of the spillover e¤ect.

The role of global trade integration in monetary transmission

We next discuss the implication of deepened trade relationship with the rest of the

world to the spillover e¤ect. To see this, we conduct simulation under the two alterna-

tive economies. The …rst economy illustrates the case when country B becomes more

integrated with the rest of the world, which consists of country B as well as country A;

such that the sum of n

and n

becomes 1.5 times larger than the baseline case.

1.5 is taken from the progress of trade integration from 1990 to 2000 shown in Figure 1.

The second economy illustrates the case when country C becomes more integrated such

that the sum of n

and n

becomes 1.5 times larger than the baseline case. In Figure

7a, we depict the equilibrium response of macroeconomic variables in country B under

the two alternative economies together with baseline economy. With a larger weight of

imp orted goo ds from other countries in constructing …nal consumption goods, domestic

economies are a¤ected more by shocks abroad. In both of the two cases, the quantitative

impacts of the contractionary monetary policy shock in country A are more pronounced

than the baseline model.

Existing studies, such as Kose, Otrok, and Prasad (2008) stress that the global trade

integration is accompanied with specialization of industrial production structure. To

evaluate the relationship between changes in production structure and spillover e¤ects,

we construct two alternative economies with di¤erent parameter values for  and study

the equilibrium response of variables under the economies. The size of parameter  in

equation (4) captures the degree of di¤erentiation across the three intermediate composite

with distinct origin. In Figure 7b, we depict response of variables under the economy with

higher elasticity (less di¤erentiated) and lower elasticity (more di¤erentiated) compared

with baseline model. Under the premise that specialization of industrial production

enhances di¤erentiation of goods, the simulation indicates that it worsens the adverse

spillover e¤ects at the impact.

Because of the assumption that n

+ n

= 1; this implies that n

falls accordingly.

The role of relative country size in monetary transmission

A relative country size in the world economy also matters to the size of spillover e¤ect

of shock originated in country A to the rest of the world: The reason why variations in the

domestic real interest rate in the U.S. are transmitted to those in the non-U.S. countries

stems from the fact that the size of the U.S. GDP is disproportionately large and the

world interest rate is in‡uenced by the households’intertemporal decisions in the U.S.

To see the impact, we consider two alternative scenarios. In the …rst scenario, parameter

value of 

shown in equation (8) is adjusted so as to reduce the country size of country

A in the world economy at the steady state from 1/4 to 1/5. The numbers are taken from

Figure 2. In this case, relative country size of C and B in the world economy increase.

In the second scenario, parameter value of 

is adjusted so as to double the relative

country size of C in the world economy at the steady state compared with baseline:

In Figure 8, we compute equilibrium responses of macroeconomic variables in country

B under the two alternative economies as well as baseline economy. With a smaller

relative size of the country size of A; the transmission of the contractionary monetary

policy shock in country A to the real interest rate of internationally traded bond are less

pronounced. The developments in domestic real interest rate and domestic productions

in country B become mitigated accordingly. It is important to note that the increase in

relative country size of C weakens the adverse e¤ects on domestic production in country

B although initial impacts become greater than baseline model.

The role of …nancial transaction in monetary transmission

Households construct their portfolio from two assets, domestically traded bond and

internationally traded bond. To see the role played by the …nancial integration in the

spillover e¤ects, we study two alternative cases, a case when the cost of accessing to the

international bond market  becomes cheaper by 1/10 only for country B and a case

when such a change occurs to all of the countries. Figure 9 displays simulation results for

the two alternative scenarios. As a result of the …nancial integration, the consumption

decline is more pronounced. By contrast, the decline in work e¤ort, which is equivalent

to production, is less pronounced.

In order to illustrate the model mechanism, in the lower four panels, we display

the response of holding of internationally traded bond by households in country B



) =P

) ; interest rate of internationally traded bond R



), relative price of

goods produced in country A and country B P

) =P

) and the real exchange

rate: The last three variables are expressed in terms from deviation under each of the

alternative cases from the baseline. With a smaller …nancial friction, households around

the globe conduct inter-market arbitrage at cheaper cost than otherwise. As a result,

the increase in interest rate in country A is easily transmitted to that in the domestic

bond market in country B as well as in the international bond market. Because of the

higher increase in the domestic interest rate, the domestic consumption in country B

is dampened greater. Because of the higher return to the internationally traded bond

makes, households in country B save more in the form of the internationally traded

bond B



) compared with the baseline. The households therefore consume less and

work more for future consumption. As the real exchange rate of currency in country B

depreciates against that in country A; the foreign demand increases from the view points

of households in country B:

27 28

The role of policy regime in monetary transmission

Lastly, we discuss implications of policy regime switch from a simple Taylor rule that

is speci…ed by equation (7) to alternative policies, a …xed exchange rate regime and a

monetary policy rule that assigns a higher policy weight to in‡ation rate such as in‡ation

rate targeting (hereafter IT). Conventional wisdom almost agrees, at least theoretically,

that ‡oating exchange rate regime helps insulate domestic output ‡uctuations from for-

eign monetary policy shock by allowing ‡exible nominal exchange rate adjustment to

the shock and the …xed exchange rate regime does not. Domestic economy under a …xed

exchange rate regime should therefore witness greater variations after the U.S. monetary

policy shock compared with the case of otherwise. In modelling a …xed exchange rate

regime, we follow Hirakata, Iwasaki, and Kawai (2012) and modify a monetary policy

rule in country B to a policy speci…ed in equation below:

"





= 0: (9)

Policy regimes in other two countries are maintained the same as the baseline case.

Changes in policy weight in the monetary policy rule also a¤ects the spillover e¤ects.

Flood and Rose (2010), discussing the role of IT in the context of business cycle syn-

chronization across countries, point out that a central bank following IT attains more

stable in‡ation rate variations at a cost of a larger domestic output ‡uctuations. In our

simulation, we maintain the functional form of the monetary policy rule and numeri-

cally increase the policy weight assigned to in‡ation 



from two to …ve. As shown in

Figure 10, when the …xed exchange rate regime sp eci…ed in (9) is in place, the nominal

exchange rate is maintained in response to the occurrence of monetary policy shock in

country A: The domestic policy rate in country B is adjusted so as to fully o¤set nominal

exchange rate movements stemming from the shock, leaving other variables responsive

to the shock. Consequently, the real interest rate rises substantially compared with the

case of otherwise, suppressing domestic production and consumption. As predicted by

As shown in Figure 9, the cheaper cost of …nancial transaction delivers lower consumption and higher

labor input of households in c ountry B: Consequently, cross-cou ntry di¤erence in terms of consumption

and labor input between countries is mitigated compared with the case of baseline model. In this sense,

the cheaper cost enhances risk-sharing among households around the globe:

Admittedly, there is limitation in the way that the …nancial integration is analyzed in the current

paper. First, deepening of …nancial integration takes more than one forms in the actual economy. The

current model speci…cation provides only one example among them. Second, as discussed below, we do

not consider e¤ects associated with the second moment of …nancial variables. The linkage between the

…nancial integration and volatilities of the asset returns, for instance, is beyond the scope of our paper.

Flood and Rose (2010), our model generates less de‡ationary consequence under the IT

since in‡ation is more stabilized. In contrast to their results, however, in our model,

a higher weight attached in‡ation rate moderates the fall in domestic production as

well as de‡ation. Since the de‡ationary response is relatively stabilized under the IT

regime compared with the other regimes, terms of trade appreciates greater than other-

wise, giving a rise to a larger expenditure-switch e¤ect that partly o¤sets the production

decline.

4 Discussion regarding Explanations

Our simulation exercise in the last section provides two opposing predictions as to how

spillover e¤ects change with recent developments in economic structure. On the one hand,

weaker trade linkage with the U.S. economy, decline in the relative country size of the

U.S. in the globe, deepening of …nancial integration, and adoption of a monetary policy

rule with a higher weight for in‡ation rate and/or more ‡exible exchange rate regime

would reduce adverse consequence of the shock on the domestic productions in the non-

U.S. economy.

On the other hand, deepening of global trade linkage would enhance the

domestic ‡uctuations in response to the same shock. These model predictions suggest

that estimation results obtained in section 2 are brought about by that the …rst e¤ect

dominates the second e¤ect. In this section, by making use of the country-by-country

data and estimated impulse function of trade and …nancial variables, as well as policy

instruments in the non-U.S. countries, we discuss each of these candidate explanations

in details.

4.1 Developments of Economic Surroundings of the Sampled

Countries

Bearing the model’s predictions in mind, we analyze if the four classes of recent devel-

opments of economic structure, a weaker trade linkage with the U.S., a decline in the

relative country size of the U.S., deepening of …nancial integration, or regime switches

in monetary policy or exchange rate policy, are observed on country-by-country basis

for our sampled countries. In fact, the …rst three of the developments hold on country-

by-country basis for most of our sampled countries, suggesting that they have played

certain roles in mitigating the spillover e¤ects. Admittedly, however, as shown in the

above section, because there is third-country e¤ects in the monetary transmission across

countries, changes in the spillover e¤ects to a non-U.S. country depend also on the eco-

nomic surroundings of the other non-U.S. countries. The analysis here therefore provides

As shown above, although quantitatively limited, our simulation exercises suggests that the reduc-

tion in the transaction cost of internationally traded bond implies wide r decline in consumption but

moderate decline in production in country B:

a side evidence regarding the determinants of the spillover e¤ects.

Trade integration with the U.S. and the rest of the world

As we see in Figure 1 and 3, there are two changes in trade structure that have

opposing e¤ects on the spillover e¤ects. Table 3 shows the size of the two secular trends

regarding trade activity on country by country basis. That is, shrinkage of bilateral trade

with the U.S. measured by the trade shares with the United States over the total trade,

and widening of trade with the rest of the world measured by the trade activity relative

to nominal GDP. These two trends are seen in almost all countries.

To see which of

the two trends are dominant, in Table 4, we document a trade activity with the U.S.

divided by the domestic GDP in the non-U.S. countries separately for the 1990s and the

2000s. While the di¤erence of this measure over the two decades is not large in most

of the countries, in particular in G6 + Australia and Asian countries, the …rst trend

generally dominates the second trend. This suggests that from the trade perspective the

imp ortance of the U.S. becomes smaller for these countries. In the table, we also display

the changes of the estimated response of production-related macroeconomic variables

over the two decades on country-by-country basis. The numb ers are shadowed with gray

when the changes of the estimated responses over the sample periods qualitatively agrees

with the theoretical prediction regarding the trade structure. Estimated di¤erences are

consistent with the theoretical prediction at least for G6 + Australia.

Country size

Table 5 documents the relative size of own GDP to the world GDP for the sampled

non-U.S. countries and the U.S. The …gures documented in the …rst two columns of the

table display that about a half of the countries, including the U.S., see the decline in its

relative country size from the 1990s to the 2000s. To see the relative changes between

the U.S. and non-U.S. countries, we compute the di¤erence of world share of own GDP

over the two decades subtracted by the di¤erence of the world share of the U.S. GDP.

The third column of the table displays the relative size of decline in the country size

for the non-U.S. countries with respect to the decline of country size of the U.S. The

numb ers are all positive, except for Germany and Japan. The table indicates that all of

the sampled non-U.S. countries see increase in its country size with respect to the U.S.,

although some countries experience a decline in country size with respect to the world

GDP.

Similarly to the exercise above, we display the changes of the estimated response

of production-related variables over the last two decades. The numbers are shadowed

with gray when the changes in the estimated responses from the 1990s to the 2000s

qualitatively agree with the theoretical prediction about the relative country size. Again,

the relationship between the estimated di¤erence of impulse responses and changes in

The number is shadowed when it indicates there are shrinkage of bilateral trade with the U.S.,

deepening of trade with the rest of the world or both.

country size is consistent with the view that such changes may have contributed to the

mitigation of the adverse e¤ects of the contractionary U.S. monetary policy shock on the

non-U.S. economies.

Financial linkages

To see the role of …nancial linkages with the U.S. and the rest of the world, we

document in Table 6 the FDI shares of the U.S. and FDIs relative to nominal GDP.

It is seen that the linkages with the U.S. displays a decline and those with the rest of

the world displays an increase in all of the countries. The net e¤ect of the two opposing

trends are shown in Table 7. That is, for most of the countries, the …nancial linkages

with the U.S. has become larger in the current years for country by country basis. While

the decline in transaction cost in our theoretical model setting does not exactly match

with the increase in the linkages of …nancial transaction caught by the FDI share, for

almost all countries, the di¤erence of estimated impulse responses of production-related

variables across the two periods agrees with the theoretical prediction.

Regime switches

Lastly, to see the role of policy regime, we select countries that have experienced

institutional changes in the domestic monetary and nominal exchange rate policy regime

during either the latter 1990s, the early 2000s, or both from those listed in Table 1.

We then ask if the changes in the estimated impulse response to the U.S. monetary

policy shock regarding domestic production variables in the listed non-U.S. countries

di¤er from those in the non-U.S. countries that experience no regime switches over the

two decades. Table 8 documents the di¤erence between the 1990s and the 2000s for the

regime-switching countries in terms of domestic response of production variables to the

contractionary U.S. monetary policy shock. To summarize, the estimated results are

consistent with the theoretical prediction regarding institutional changes in some coun-

tries but not for all of the countries. On the one hand, in advanced countries, countries

that have employed the IT see less adverse consequence in the wake of the contractionary

U.S. monetary policy shock during the 2000s compared with the average of all countries.

In Asian and Latin American countries, on the other hand, such relationship is less pro-

nounced because the responses of the three domestic production variables disagree from

each other. In these two country groups, the in‡uence of exchange rate regime on the

spillover e¤ect is not prominent, either. In fact, this accords well with the existing stud-

ies that see a limited role of the nominal exchange rate regime in explaining the spillover

e¤ects. For instance, Canova (2005) analyzes Latin American countries with various

exchange rate regime including dollarized, partially dollarized, crawling banks, and ‡oat

to conclude that di¤erentials of domestic response in these countries due to di¤erences

in exchange rate regime is minor. Ma

ckowiak (2007) also sees minor di¤erence between

Clearly, series for FDI captures only a limited amount of …nancial transaction across borders. In

the current paper, however, we focus on the FDI because of the data limitation.

Hong Kong, regime with tight …xed exchange rate, and other Asian countries in the way

that domestic variables respond to the U.S. monetary policy shock.

4.2 Resp onses of Trade, Financial, and Policy Instruments to

the U.S. monetary Policy Shock

In addition to the country-by-country analysis, we estimate responses of macroeconomic

variables other than production-related variables in the non-U.S. countries to a contrac-

tionary shock to the U.S. monetary policy rule. The responses of these variables help

outline the sources behind the changes in the spillover e¤ects.

Response of trade variables

Figure 11 displays domestic responses of trade variables, including real export, real

imp ort, and trade balance in the non-U.S. countries to the U.S. contractionary mone-

tary policy shock.

The responses of trade variables are generally in line with those of

production-related variables. In particular, observation that most of the countries ex-

perience a signi…cant decline in real export veri…es that the expenditure-reducing e¤ect

dominates the expenditure-switching e¤ect. There is a cross-country heterogeneity in

the way that the trade variables respond to the shock. G6 + Australia experience the

similar pattern of changes to the domestic variables over the two decades. That is, a

proportion of negatively-responding countries for the trade balance decreased from the

1990s to the 2000s, indicating the trade activity helps moderate the adverse consequence

of domestic production in these countries. Other advanced countries also exhibit that

similar decline in the numb er of negatively-responding countries from the 1990s to the

2000s. In Latin American and Asian countries, moderations of adverse e¤ect are not seen

in trade balance responses. A larger portion of countries respond negatively in the cur-

rent years in Latin American countries, and the composition of countries is maintained

the same in Asian countries.

Response of …nancial variables

Figure 12 displays the estimated impulse response of nominal stock returns, nominal

exchange rate, and the long term interest rate to the shock. Here the nominal exchange

rates are expressed in terms of a dollar evaluated by the non-U.S. currency. The posi-

tive response indicates that a value of the non-U.S. currency depreciates against dollar

One candidate explanation for the results that exchange rate regime does not seem to a¤ect the

spillover e¤ect is that the class of regime is not the only determinants of the spillover e¤ect. For instance,

ckowiak (2007) discusses that Hong Kong stands out not only in terms of …xed exchange rate regime

but also in terms of …nancial openness. Canova (2005) points out the possibility that the e¤ective

di¤erence between the distinct institutional arrangements for exchange rate regime may be small.

Similarly to the treatment in the model s ection, in constructing trade balance series used for esti-

mation, we …rst take the di¤erence between the export and import and divide it by the average of GDP

of the sample period.

after the contractionary U.S. monetary policy shock. Estimated responses of …nancial

variables are in line with those of production-related variables, except for the case of

long term interest rate. Regarding stock price, while a bulk of countries see the negative

response during the 1990s, the portion of negatively-responding countries has declined

during the 2000s. Regarding nominal exchange rate, while a bulk of countries see the

positive response during the 1990s, the p ortion of positively-responding countries has

declined during the 2000s. Smaller depreciation of currencies in the non-U.S. countries

against dollar may be attributed to mitigated domestic consequence of the U.S. monetary

policy shock.

The long term interest rate exhibits qualitatively opposite responses from

what are obtained for the other variables. That is, a proportion of positively-responding

countries has risen from the 1990 to the 2000s. One interpretation for this observation

is that the contractionary U.S. monetary policy shock causes a larger …nancial conta-

gion during the 2000s than the 1990s. Combined with the estimated results regarding

production-related variables, however, the real e¤ect of such contagion on the domestic

variables are minor, if not any.

A potential explanation is that the adverse spillover

e¤ects on the long term interest rates are short-lived. To see this, we depict the dis-

tribution of impulse response function of the long term interest rates that is averaged

over three to four years after the shock. Clearly, the distribution of the impulse response

function shifts to zero, indicating that increased interest rate reverts back to zero almost

within these years.

Our model analysis above concentrates on a log-linearized approximation of economic

dynamics after the contractionary U.S. monetary policy shock. By contrast, in practice,

the shock may trigger …nancial contagion associated with the second moment of the

asset returns. Calvo and Mendoza (2000) provide a theoretical example that the global

…nancial integration brings about rational contagion in the …nancial market by solving

the optimization problem of p ortfolio manager. Along this line, suppose that volatilities

of …nancial assets in the non-U.S. country endogenously react to the U.S. monetary policy

shock, then changes in the relationship between the volatilities of assets and the U.S.

monetary policy shock may result in the weaker spillover e¤ects of the U.S. monetary

policy shock to the non-U.S. countries.

To see this, we construct a monthly series of volatility of stock return, nominal ex-

change rate, and long term bond for each of the sampled non-U.S. countries using gener-

alized autoregressive conditional heteroskedasticity (GARCH) model. For this purpose,

Diebold and Yilmaz (2009) construct s pillover indices of stock market return and its volatility that

capture the contribution of shocks originating abroad to variations of these two variables and show that

the former index increases monotonically from 1995 to 2007 while the latter series rises in response to

each economic event, such as East Asian crisis in 1997 or Terrorist Attack in 2001 and do not display

the positive trend.

From slightly di¤erent perspective, Fujiwara and Takahashi (2011) point ou t the presence of “macro-

…nance dissonance.” Based on the analysis on the linkages between Asian c ountries and developed

countries, they …nd that the spillover of the …nical variables is strong and that of the industrial production

is weak.

we construct a simple GARCH (1, 1) model for a variable X

; which in this analysis

includes a stock return, nominal exchange rate, and long term interest rate, that is spec-

i…ed by the following law of motion and estimate the unobservable time-varying second

moments based on the model:

= X

t1

+ "

; (10)

= 

; for  > 0; and



= c + "

t1

+ 

t1

: (11)

Here, "

is an innovation to a variable X

; z

is i.i.d. innovation with zero mean and

unit variance that hits the variance of a variable X

,  is an autoregressive parameter of

a variable X

; and c; ; and  are the non-negative parameters that govern the law of

motion of time-varying volatility of a variable X

denoted by 

: Note that (11) indicates

that the volatility of the …nancial variable X

is given by a one-period ahead forecast

variance based on news about volatility from the previous period measured as the lag of

the squared residual from the mean equation (10) and forecasted variance in the previous

period.

Based on the estimated series of volatility 

; we compute the responses of these

volatilities to the U.S. monetary policy shock. Figure 13 provides the evidence that the

U.S. monetary policy shock in‡uences the size of asset volatilities in a statistically sig-

ni…cant manner.

36 37

Regarding the volatilities of stock return and the long term interest

rate, the shock a¤ects them positively throughout the two sample periods. For both of

the two assets, however, a proportion of positively-responding countries has declined.

This reduced volatilities of …nancial asset may have contributed to the weakened adverse

spillover e¤ect. The distributions of the estimated responses of volatilities for the two

assets have ‡attened during the 2000s compared with the 1990s, indicating cross-country

responses have become diverse in the current years. The shock’s impact on the volatil-

ities of nominal exchange rate disagree across countries. In Asian countries, the shock

enhances the volatility in more than half of the sampled countries throughout the two

decades. In other countries, except for the G6 + Australia during the 1990s, it gives no

or negative impact on the volatilities.

Response of policy instruments

Figure 14 displays the response of policy instruments together with price level, mea-

sured by CPI, to the contractionary U.S. monetary policy shock. Estimated results are

In computing the volatilities of long term interest rate, we employ …ve years government bond yield.

In addition, because of the unavailability of the su¢ cient time series, we restrict our analysis for the

1990s on the advanced countries.

Along the similar line, Shirota (2013) analyzes the determinants of cross-border credit ‡ows and

show that the importance of global factor is increasing over the years.

mixed depending on instruments and country group.

Among advanced countries, a

larger portion of countries witness positive response of short term interest rate and neg-

ative response of narrow money to the shock during the 2000s period than the 1990s.

That is, overall policy responses in the two country groups have currently become more

contractionary. On the contrary, among Asian countries, a larger portion of countries

witnesses a positive response of short term interest rate and a negative response of narrow

money during the 1990s compared with the 2000s, suggesting that other things being

equal, the endogenous policy responses more adversely a¤ect the domestic economic

activities in the country group during the 1990s.

Admittedly, e¤ective impact of the endogenous response of policy instrument on the

domestic economy dep ends on how macro economic variables, in particular, price develops

in the wake of the shock. The bottom two panels in Figure 14 show that the contrac-

tionary U.S. monetary policy shock generally yields de‡ationary impacts to the economy

in the non-U.S. countries. The negative e¤ects are more pronounced during the 2000s

compared with the 1990s.

In Asian countries, more than half of the countries witness

negative response of price level during the 1990s while a limited portion of the countries

witness the negative response during the 2000s. These estimated results combined with

those for policy instrument responses suggest that the developments of real interest rate

and real money balance may have severely a¤ected the domestic productions in these

Asian countries during the 1990s and such e¤ects have been diminished during the 2000s.

That is, the changes in endogenous policy responses in the Asian countries to the con-

tractionary U.S. monetary policy shock may have contributed to the recent moderation

of adverse impacts of the U.S. monetary policy on these countries.

5 Conclusion

In this paper, we empirically study the spillover e¤ects of the U.S. monetary policy shock

to the rest of the world. We speci…cally focus on how recent developments about global

Along the similar line to our study, Grilli and Roubini (1995) empirically investigate the monetary

policy reactions of non-U.S. G7 countries and document that the policy rates in these countries evolve

in the same direction as the U.S. monetary policy rate in response to the shock to the U.S. monetary

policy rule, suggesting that the domestic policies in the non-U.S. countries endogenously react to the

U.S. monetary policy shock. Kim (2001), on the contrary, claims that the endogenous policy responses

of these countries are not statistically signi…cant except for the case of Canada.

Figure 14 does not document the estimation results for Latin American countries during the 1990s

because of the inavaiability of the data series in the region.

There is no agreement as to the qualitative impacts of the U.S. monetary policy shock on domes-

tic in‡ation in the non-US countries. Mackowiak (2007), analyzing emerging countries, discusses that

the contractionary U.S. monetary policy shock causes depreciation of domestic currencies and leads to

in‡ation in the non-U.S. countries. By contrast, Chen et al. (2011), while focusing on the unconven-

tional monetary policy, document that the expansionary shock to the policy causes in‡ation for most of

advanced and Asian countries.

integration a¤ect the transmission mechanism of the U.S. monetary policy. Making use

of factor augmented VAR, we distill innovations to the Federal Funds rate up to 2007

and estimate the responses of a set of macroeconomic variables in the non-U.S. countries,

including Advanced, Latin American, and Asian countries, to the contractionary U.S.

monetary policy shock. In order to gauge the impacts of the structural changes of the

economy on the spillover e¤ects, we estimate macroeconomic responses of the non-U.S.

country separately for the sample covering the 1990s and the 2000s. Our key …nding

is the weakening of adverse spillover e¤ects. The contractionary U.S. monetary policy

shock delivers adverse e¤ects on the domestic production in the most of the non-U.S.

economies during the 1990s. Such e¤ects have been, however, largely diminished during

the 2000s.

To see the causes behind the weakening of adverse spillover e¤ects, we examine sev-

eral candidate explanations including trade and …nancial global integration. Those are

changes in bilateral trade and …nancial relationship with the U.S., the relative country

size of the U.S., and regime switch of domestic policies. We …rst construct a text book

Dynamic Stochastic General Equilibrium (DSGE) model and derive theoretical impli-

cations about the linkages between these structural changes of economic environments

and the spillover e¤ects. Our model predicts that these structural changes bring about

two opposing e¤ects to the spillover e¤ects. First, other thing being equal, a deepening

of trade integration makes the non-U.S. countries more responsive to external shocks,

leading to a larger decline of domestic production in these countries in response to a con-

tractionary U.S. monetary policy shock. On the contrary, a weakening of the bilateral

trade relationship with the U.S., a decline in the relative size of the U.S. economy, risk-

sharing e¤ects associated with deepening of …nancial integration, and switches of policy

regime to more ‡exible exchange rate regime and/or in‡ation targeting rule soften such

adverse consequence to domestic production. Our theoretical exercise therefore suggests

that enhanced spillover e¤ects due to global trade integration are overturned by other

factors.

In addition to the theoretical analysis, we ask if these candidate explanations are

consistent with the economic surroundings of our sampled countries and estimated im-

pulse responses of non-production-related variables. We study country-by-country data

and show that in most of our sampled countries, a decline of the trade relationship with

the U.S. quantitatively dominates the deepening of global trade integration, a decline of

the country size is slower than the U.S., and a …nancial transaction with both the U.S.

and the rest of the world increases over the last two decades. We …nd that estimated

response of long-term interest rate is positive and larger during the 2000s, but they die

out quickly. From the estimation of policy instrument, we …nd that the policy responses

Note that our estimation result only suggests that the changes in the domestic variables in the

non-U.S. countries are mitigated over the decades in response to the same size of the contractionary

U.S. monetary policy sho ck. Admittedly, when the original contractionary shock becomes larger, then

the quantitative impacts abroad will b ecome larger as well.

to a U.S. monetary policy shock are more contractionary during the 1990s than 2000s

in Asian countries, suggesting that the policy responses may have caused weakening of

spillover e¤ects in these countries.

Our study provides some policy implications to the consequence of the monetary

easings undertaken by several advanced countries in response to the current global …-

nancial crisis. The monetary easings contain unconventional quantitative measures as

well as adjustments of the policy rates to a low level. Regarding the …rst ingredient of

the monetary easings, our theoretical and empirical studies are built upon the monetary

policy shocks that are identi…ed as innovations to the nominal short-term interest rate.

They are silent about the various types of credit spreads facing …nancial institutions,

…rms, and households, that were reduced by the unconventional measures. By contrast,

our analysis addresses the second ingredient of the monetary easings. The estimated

weakenings of adverse spillover e¤ect is attained even if the sample period is extended

to 2012.

While our pap er stresses the role played by a decline in the relative signi…cance of

the U.S. in the domestic economic activities in the non-U.S. countries in accounting

for the weakened spillover e¤ects of the U.S. monetary policy shock, our explanation

is not mutually exclusive with alternative candidate explanations. For instance, the

developments in specialization of production structure or increases in the compositional

share of services sector in the economy in the non-U.S. countries may be promising

alternative determinant of the spillover e¤ects. These are left for the future researches.

Admittedly, forward guidance of future time path of the short term interest rate was also an impor-

tant element of policy packages launche d during the …nancial crisis. The e¤ect of the forward guidance

is also beyond our scope. See Fujiwara, Sudo, and Teranishi (2010) and Fujiwara et al. (2013) for

how commitment about future domestic short term interest rate a¤ects economic activities in the other

countries when both countries fall in the liquidity trap.

References

[1] Baxter, M., Kouparitsas, M. A., (2005) “Determinants of Business Cycle Comove-

ment: a Robust Analysis,”Journal of Monetary Economics, Vol. 52.

[2] Bayoumi, T., Vitek, F. (2013) “Macroeconomic Model Spillovers and their Discon-

tents,”IMF, Working Paper NO. 13/4.

[3] Bernanke, Ben, Jean Boivin, and Piotr Eliasz, (2005) “Measuring Monetary Policy:

A Factor Augmented Vector Autoregressive (FAVAR) Approach,”Quarterly Journal

of Economics Vol. 120.

[4] Bluedorn, J. C., C. Bowdler (2011) “The Open Economy Consequences of U.S.

Monetary Policy,”Journal of International Money and Finance, Vol. 30.

[5] Boivin, J., M. Giannoni, and I. Mihov (2009) “Sticky Prices and Monetary Policy:

Evidence from Disaggregated U.S. Data,”American Economic Review, Vol. 99.

[6] Calvo, A. G., and E. G. Mendoza (2000) “Rational contagion and the globalization

of securities marke markets,”Journal of International Economics, Vol. 51.

[7] Canova, F. (2005 ) “The Transmission of Us Shocks to Latin America,”Journal of

Applied Econometrics, Vol. 20.

[8] Casares, M. (2006) “Time-to-build, Monetary Shocks, and Aggregate Fluctuations,”

Journal of Monetary Economics, Vol. 53.

[9] Chen, Q., Filardo A., HE, D., Zhu, F. (2012) “International Spillovers of Central

Bank Balance Sheet Policies,”A chapter in Are Central bank balance sheets in Asia

too large?, BIS Papers Vol. 66.

[10] Christiano, L., Eichenbaum, M., Evans, C. (1996) “The E¤ects of Monetary Policy

Shocks: Evidence from the Flow of Funds,”Review of Economics and Statistics Vol.

78.

[11] Christiano, L., Eichenbaum, M., Evans, C. (1998) “Monetary Policy Shocks: What

Have We Learned and to What End?”NBER Working Paper No. 6400.

[12] Cwik, T., Muller, J. G., Wolters, H. M. (2011) “Does Trade Integration Alter Mon-

etary Policy Transmission?,”Journal of Economic Dynamics and Control, Vol. 35.

[13] Corsetti, G., Pesenti, P. (2001a) “International Dimensions of Optimal Monetary

Policy,”National Bureau of Economic Research Working Paper 8230.

[14] Corsetti, G., Pesenti, P. (2001b) “Welfare and Macroeconomic Interdependence,”

Quarterly Journal of Economics, Vol. 116.

[15] Dees, S., Saint-Guilhem A. (2009) “The Role of the United States in the Global

Economy and its Evolution over time,”ECB Working Paper Series.

[16] Ehrmann, M., Fratzscher, M. (2006) “Global Financial Transmission of Monetary

Policy Shocks,”ECB, Working Pap er Series No. 616.

[17] Fujiwara, I., Nakajima, T., Sudo, N., Teranishi, Y. (2013) “Global Liquidity Trap,”

forthcoming, Journal of Monetary Economics.

[18] Fujiwara, I., Sudo, N., Teranishi, Y. (2010) “The Zero Lower Bound and Monetary

Policy in a Global Economy: A Simple Analytical Investigation,” International

Journal of Central Banking, Vol. 6.

[19] Hirakata, N., Iwasaki, Y., Kawai, M (2013) “Emerging Economy Supply Shocks

and Japan’s Price De‡ation: International Transmissions in a Three-country DSGE

Model,”mimeo, Asian Development Bank Institute.

[20] Ilzetzki, E., Jin, K. (2013) “The Puzzling Change in the International Transmission

of U.S. Macroeconomic Policy Shocks,”London School of Economics, mimeo.

[21] Kim, S. (2001) “International Transmission of U.S. Monetary Policy Shocks: Evi-

dence from VAR’s,”Journal of Monetary Economics, Vol. 48.

[22] Laxton, D., P. Pesenti (2003) “Monetary Rules for Small, Open, Emerging

Economies,”Journal of Monetary Economics, Vol. 50.

[23] Liu, P., H. Mumtaz, A. Theophilopoulou (2009) “International Transmission of

Shocks: a time-varying factor-augmented VAR approach to the open economy,”

Working Paper No. 425, Bank of England.

[24] Ma

ckowiak, B. (2007) “External Shocks, U.S. Monetary Policy and Macroeconomic

Fluctuations in Emerging Markets,”Journal of Monetary Economics, Vol. 54.

[25] Ma

ckowiak, B. (2006) “What does the Bank of Japan do to East Asia?,” Journal

of International Economic, Vol. 70.

[26] Koluk, T. Mehrotra A. (2009) “The Impact of Chinese Monetary Policy Shocks on

East and South-East Asia,” The Economics of Transition, The European Bank of

Reconstruction and Development, Vol. 17.

[27] Ncube, M., Ndou, E., Gumata, N. (2012) “How are the US Financial Shocks Trans-

mitted into South Africa?,”African Development Bank, Working Paper Series No

157.

[28] Obstfeld, M., Rogo¤, K. (1996) Foundations of International Macroeconomics. MIT

Press, Cambridge, MA.

[29] Romer, Christina D. and David H. Romer (2004) “New Measure of Monetary Shocks:

Derivation and Implications,” American Economic Review, Vol. 94, No. 4. 1055–

1084.

[30] Rotemberg, J. (1982) “Sticky prices in the United States,”Journal of Political Econ-

omy, Vol. 90.

[31] Shirota, Toyoichiro (2013) “What is the Major Determinant of Credit Flows through

Cross-Border Banking?,”Bank of Japan, Working Paper Series 13-E-5.

[32] Sousa, J., Zaghini, A. (2004) “Monetary Policy Shocks in the Euro Area and Global

Liquidity Spillover,”ECB, Working Paper Series No. 309.

[33] Spantig, K. (2012) “International Monetary Policy Spillovers in an Asymmetric

World Monetary System: The United States and China,”Global Financial Markets,

Working Papers No. 33.

[34] Teja, R. (2012) “2012 Spillover Report –Background Papers,”IMF, Policy papers.

[35] Ueda, K. (2009) “Changes in In‡ation Targeting – Experience in New Zealand,

Canada, United Kingdom, and Sweden,”Kinyu Kenkyu, in Japanese.

[36] Willems, T. (2011) “Using Dollarized Countries to Analyze the E¤ects of US Mon-

etary Policy Shocks,” Society for Economic Dynamics in its series 2011 Meeting

Papers, No. 200.

Table 1: Regime Switches of Monetary and Nominal Exchange Rate Policy

▼Inflation Targeting Countries

Country Start of IT framework

G7 countries + Australia Canada February 1991

United Kingdom

October 1992

Australia April 1993

Other developed economies Switzerland January 2000

Sweden January 1993

Norway March 2001

Asian emerging economies Indonesia July 2005

Korea April 1998

Philippines Januray 2002

Thailand May 2000

Latin American economies Brazil1 July 1999

Mexico September 1990

Chile January 1994

Colombia January 1999

▼

Countries managing foreign exchange

Country Timing of policy change

Asian emerging economies Hong Kong May 2005

1USD=7.8HKD

→

1USD=7.75

～

7.85HKD

India March 1993

fixed dual exchange rate system

→

managed float system

Indonesia August 1997

managed float system

→

float syst em

Korea December 1997

managed float system

→

float syst em

Malaysia September 1998

managed float system

→

fixed exchange rate system

July 2005

fixed exchange rate system

→

basket system

Thailand July 1997

basket system

→

managed float system

Latin American economies Brazil January 1999

managed float system

→

float syst em

Chile September 1999

fixed exchange rate system

→

float system

Colombia September 1999

fixed exchange rate system

→

float system

Mexico January 1999

managed float system

→

float syst em

Table 2: Baseline Parameters

Parameter Value Description

 :99 Quarterly subjective discount rate

 :2 Elasticity of substitution across di¤erentiated products

 5 Utility weight on disutility of labor

 :5 Inverse of Frisch elasticity of lab or input

 2 Elasticity of substitution across composite goods

 200 Cost associated with price adjustment



50 Cost associated with wage adjustment



2 Elasticity of substitution across di¤erentiated labor

b :8 Persistency in habit parameter

 1:5 Inverse of intertemporal elasticity of substitution

for Z = A; B; and C 5=6

Share parameter for goods produced in country Z

in consumption c omposite in country Z

for

Z; Z = A; B; and C

, and

Z 6= Z:

1=12

Share parameter for goods produced in country

in consumption c omposite in country Z



1=4 Scale parameter for country A



1=20 Scale parameter for country B



1  

 

Scale parameter for country C

R 1= Policy rate at steady state



:9 Autoregressive parameter in monetary policy rule





1:1 Policy weight on in‡ation rate



:1 Policy weight on output gap

 :01 Financial friction associated with bond trade

^ :1 Financial friction associated with bond trade

Table 3: Trade Shares with the U.S. and Trade Relative to GDP

1990's 2000's 1990's 2000's

G7 and Australia

Australia 11.65% 7.33 % 3 6.84% 41.20%

Canada 76.78% 72.09 % 55.76 % 56.14%

France 11.50% 9.15 % 45.97 % 44.42%

Germany 9.59% 7.33% 47.74% 66.33%

Italy

7.25% 5.35% 37.16% 44.35%

Japan 26.29% 16.45% 15.83% 24.71%

United Kingdom 13.07% 11.27% 41.18% 40.54%

Other developed economies

Belgium 5.85% 4.54%

109.25% 123.27%

Denmark 4.87% 4.98% 52.75% 60.51%

Finland 7.90% 5.51%

56.36% 62.09%

Netherlands 6.79% 6.21% 95.11% 112.22%

Norway 6.79% 5.80% 50.62% 51.91%

Portugal N/A 2.73% 53.37% 55.89%

Spain 4.99% 3.93% 45.92% 42.38%

Sweden

7.51% 6.16% 51.58% 65.19%

Switzerland 7.60% 7.72% 49.69% 63.74%

Latin emerging economies

Brazil 21.26% 15.57% 13.31% 20.11%

Chile 18.44% 15.60% N/A 60.95%

Columbia N/A 32.30% N/A 32.82%

Mexico 65.51% 68.01% 12.47% 13.45%

Asian emerging economies

China 15.51% 13.52% 34.65% 51.45%

Hong Kong 10.54% 6.20% 137.66% 172.12%

India 14.17% 8.78% 17.53% 33.99%

Indonesia 12.51% 8.29% 48.29% 46.81%

Korea 21.08% 11.87% 50.79% 76.29%

Singapore 20.03% 11.33% 215.27% 229.05%

Taiwan

22.30% 13.08% 75.97% 107.92%

Thailand 16.14% 10.08% 74.46% 120.11%

Notes:

As for trade activity relative to nominal GDP of 1990's, the figure for France and Spain refer to 2000, Germany, Italy, Belgium, Finland,

Netherlands, Portugal refer to 1991, Japan refers to the average from 1994 to 1999, United Kingdom refers to the average of 1996 and

1999, Mexico and Indonesia refer to the average from 1993 to 1999, China refers to the average from 1992 to 1999, India and Thailand

refer to the average from 1991 to 1999. As for those of 2000's, the figure for Chile refers to the average from 2003 to 2012, Columbia

refers to the average from 2006 to 2012, respectively.

Trade shares with the United States

Trade activity relative to Nominal GDP

Trade shares with the U.S. indicates the ratio of the sum of export to and import from the United States to the sum of export to and import

from all of the world in each country.

Trade activity relative to nominal GDP indicates the ratio of the sum of import and export to nominal GDP for each country.

As for trade shares with the United States of 1990's, the figure for France and Germany refer to 2000, Italy, Belgium, Finland, Netherlands,

Spain refer to 1991, the United Kingdom and Singapore refer to the average of 1996 and 1999, Sweden and Taiwan refer to the average

from 1998 to 1999, Mexico and China refer to the average from 1993 to 1999, India and Thailand refer to the average from 1991 to 1999.

As for those of 2000's, the figure for Portugal refers to the average from 2004 to 2012, Columbia refers to the average from 2006 to 2012,

respectively.

Table 4: Trade Linkage and Domestic Response to the U.S. Monetary Policy Shock

Difference

Response

of IIP

Response of

Employment

Response

Unemploy

ment Rate

1990s 2000s

2000s -

1990s

2000s -

1990s

2000s -

1990s

2000s -

1990s

G7 and Australia

Australia 4.3% 3.0% - NA 0.22 -0.90

Canada 42.8% 40.5% - -0.49 0.05 -1.17

France 5.3% 4.1% - 0.26 NA -1.61

Germany 4.6% 4.9% + 0.50 0.62 0.72

Italy 2.7% 2.4% - 0.27 NA -1.31

Japan 4.2% 4.1% - -0.02 0.16 -1.17

United Kingdom 5.4% 4.6% - 0.09 0.01 -0.06

Other developed economies

Belgium

6.4% 5.6% -

0.64 NA -1.31

Denmark 2.6% 3.0% + -0.10 NA -0.60

Finland 4.4% 3.4% - 1.02 0.08 -0.19

Netherlands 6.5% 7.0% + 0.22 NA -0.60

Norway 3.4% 3.0% - NA 0.57 0.40

Portugal NA 1.5% NA NA NA 0.16

Spain 2.3% 1.7% - NA NA -0.51

Sweden

3.9% 4.0% +

NA 0.30 -0.47

Switzerland 3.8% 4.9% + NA NA -2.01

Latin emerging economies

Brazil 2.8% 3.1% + -0.25 -0.10 0.20

Chile NA 9.5% NA -0.06 0.17 -1.92

Columbia NA 10.6% NA 0.07 NA NA

Mexico

8.2% 9.1% + -1.28 NA -0.11

Asian emerging economies

China 5.4% 7.0% + NA NA NA

Hong Kong 14.5% 10.7% - NA -0.36 -0.40

India 2.5% 3.0% + NA NA NA

Indonesia 6.0% 3.9% - 0.05 NA NA

Korea

10.7% 9.1% - 0.17 -0.18 -1.27

Singapore 43.1% 25.9% - 0.25 NA NA

Taiwan

16.9% 14.1% - NA 0.00 -0.36

Thailand 12.0% 12.1% + NA 0.22 NA

Notes:

Trade with the United States relative to

GDP

Trade shares with the U.S. indicates the ratio of the sum of export to and import from the United States to the sum of export to and import from all of the

world in each country. Difference indicates the difference of trade with the U.S. relative to GDP over the two decades.

Response of IIP, employment, and unemployment rate are the difference over the two decades of the estimated impulse response at the quarter four years

after the contractionary U.S. monetary policy shock. The numbers are shadowed when response becomes less adverse during the 2000s for countries that

sees a decline in trade relationship with the U.S. over the two decades.

Table 5: Relative Importance of Own GDP to the U.S. GDP

Difference

Response

of IIP

Response of

Employment

Response of

Unemployment

Rate

1990's 2000's

2000s -

1990s

2000s -

1990s

2000s -

1990s

2000s - 1990s

23% 22% -2%

G7 and Australia

Australia 1.33 % 1.71% + NA 0.22 -0.90

Canada 2.24 % 2.48% + -0.49 0.05 -1.17

France 5.12 % 4.35% + 0.26 NA -1.61

Germany 7.67% 5.70% - 0.50 0.62 0.72

Italy 4.29% 3.55% + 0.27 NA -1.31

Japan

15.50% 9.54% - -0.02 0.16 -1.17

United Kingdom 4. 38 % 4.32% + 0.09 0. 01 - 0.06

Other developed economies

Belgium 0.88% 0.78% + 0.64 NA -1.31

Denmark 0.58% 0.52% + -0.10 NA -0.60

Finland 0.44% 0.41% + 1.02 0.08 -0.19

Netherlands 1.33% 1.31%

0.22 NA -0.60

Norway 0.51% 0.66% + NA 0.57 0.40

Portugal

0.39% 0.38% +

NA NA 0.16

Spain 2.09% 2.28% + NA NA -0.51

Sweden 0.91% 0.77% + NA 0.30 -0.47

Switzerland 1.00% 0.86% + NA NA -2.01

Latin emerging economies

Brazil 2.25% 2.54% + -0.25 -0.10 0.20

Chile

0.23% 0.30% + -0.06 0.17 -1.92

Columbia 0.33% 0.38% + 0.07 NA NA

Mexico 1.56% 1.80% + -1.28 NA -0.11

Asian emerging economies

China 2.59% 7.19% + NA NA NA

Hong Kong 0.49% 0.40% + NA -0.36 -0.40

India

1.29% 2.06% + NA NA NA

Indonesia 0.58% 0.85% + 0.05 NA NA

Korea 1.53% 1.67% + 0.17 -0.18 -1.27

Singapore 0.26% 0.32% + 0.25 NA NA

Taiwan 0.91% 0.74% + NA 0.00 -0.36

Thailand 0.47% 0.44% + NA 0.22 NA

Source: IMF

Share of GDP relative to world GDP

The difference denotes the changes in relative significance of own GDP to the U.S. GDP. That is, the difference of world share of own GDP over the two

decades subtracted by the difference of that for the U.S.

Response of IIP, employment, and unemployment rate are the difference over the two decades of the estimated impulse response at the periods four years

after the contractionary U.S. monetary policy shock. The numbers are shadowed when estimated response becomes less adverse during the 2000s for

countries that see an increase in own GDP relative to the U.S. GDP over the two decades.

Table 6: FDI Shares of the U.S. and FDIs Relative to GDP

1990's 2000's 1990's 2000's

G7 and Australia

Australia 20.42% 19.70% 42.78% 65.92%

Canada 53.82% 46.44% 41.39% 70.37%

France 19.11% 14.14% 26.17% 76.97%

Germany 26.47% 15.77% 17.23% 61.26%

Italy

10.66% 6.85% 15.48% 32.64%

Japan 52.71% 47.14% 6.53% 13.92%

United Kingdom

39.43% 32.03% 50.22% 105.67%

Other developed economies

Belgium N/A N/A N/A N/A

Denmark 8.64% 5.09% 49.92% 102.73%

Finland

14.36% 5.96% 20.76% 78.31%

Netherlands 45.42% 42.52% 74.18% 184.55%

Norway

19.81% 11.75% 25.38% 57.65%

Portugal 4.15% 0.90% 24.94% 65.57%

Spain 9.58% 8.11% 23.45% 72.28%

Sweden 13.97% 13.52% 39.81% 123.21%

Switzerland 36.14% 32.16% 63.61% 202.66%

Latin emerging economies

Brazil

N/A 35.16% N/A 33.24%

Chile N/A N/A 56.81% 84.34%

Columbia N/A N/A N/A N/A

Mexico 46.61% 30.24% 10.93% 32.34%

Asian emerging economies

China N/A 3.74% N/A 26.31%

Hong Kong

N/A N/A N/A N/A

India 11.57% 11.06% 3.71% 11.56%

Indonesia N/A 12.24% N/A 14.05%

Korea N/A 17.64% N/A 20.58%

Singapore N/A N/A N/A N/A

Taiwan N/A N/A N/A N/A

Thailand N/A N/A N/A N/A

Notes:

FDI shares of the United States

FDIs relative to Nominal GDP

FDI shares of the U.S. indicates the ratio of the sum of inward FDI from the U.S. and outward FDI to the U.S. to the total of inward and outward FDI in

each country. FDIs relative to nominal GDP indicate the ratio of the inward FDIs to nominal GDP for each country.

As for FDIs relative to nominal GDP of 1990's, the figure for Denmark refers to the average of 1998 and 1999, Portugal refers to the average from 1995 to

1999, Chile and India refer to the average from 1997 to 1999, respectively.

As for FDIs relative to nominal GDP of 2000's, the figure for France, Italy, Denmark, Norway, and Indonesia refers to the average from 2000 to 2011,

Brazil and Korea refer to the average from 2001 to 2012, China refers to the average from 2004 to 2011, respectively.

As for FDI shares of the U.S. of 1990's, the figure for Denmark refers to the average of 1998 and 1999, Portugal refers to the average from 1995 to 1999,

Chile and India refer to the average from 1997 to 1999, respectively.

As for FDI shares of the U.S. of 2000's, the figure for France, Italy, Denmark and Norway refers to the average from 2000 to 2011, Brazil refers to the

average from 2001 to 2011, Korea refers to the average from 2001 to 2012, Australia refers to the average from 2005 to 2012, Portugal refers to the

average from 2007 to 2009, China refers to the average from 2004 to 2011, Indonesia refers to the average from 2007 to 2011, respectively.

Table 7: Financial Linkage and Domestic Response to the U.S. Monetary Policy Shock

Difference

Response of IIP

Response of

Employment

Response

Unemploy

ment Rate

1990's 2000's

2000s -

1990s

2000s - 1990s 2000s - 1990s

2000s -

1990s

G7 and Australia

Australia 13.92% 22.31% + NA 0 .22 -0.90

Canada 27.11% 44.07% + -0.49 0.05 -1.17

France 4.22% 9.64% + 0.26 NA -1.61

Germany 1.79% 5.96% + 0.50 0.62 0.72

Italy

1.11% 1.60% + 0.27 NA -1.31

Japan 2.49% 4.88% + -0.02 0.16 -1.17

United Kingdom 18.72% 2 9. 6 7% + 0.09 0 . 01 - 0. 06

Other developed economies

Belgium N/A N/A 0.64 NA -1.31

Denmark 4.31% 5.23% + -0.10 NA -0.60

Finland

2.98% 4.67% +

1.02 0.08 -0.19

Netherlands 33.69% 78.48% + 0.22 NA -0.60

Norway

5.03% 6.77% +

NA 0.57 0.40

Portugal 1.03% 0.59% - NA NA 0.16

Spain 2.25% 5.86% + NA NA -0.51

Sweden 5.56% 16.66% + NA 0.30 -0.47

Switzerland 22.99% 65.17% + NA NA -2.01

Latin emerging economies

Brazil

N/A 11.69% -0.25 -0.10 0.20

Chile N/A N/A -0.06 0.17 -1.92

Columbia N/A N/A 0.07 NA NA

Mexico 5.10% 9.78% + -1.28 NA -0.11

Asian emerging economies

China N/A 3.74% NA NA NA

Hong Kong

N/A N/A NA -0.36 -0.40

India 0.43% 1.28% + NA NA NA

Indonesia N/A 1.72% 0.05 NA NA

Korea N/A 3.63% 0.17 -0.18 -1.27

Singapore N/A N/A 0.25 NA NA

Taiwan

N/A N/A NA 0.00 -0.36

Thailand N/A N/A NA 0.22 NA

Notes:

FDI of the United States relative to

Nominal GDP

Response of IIP, employment, and unemployment rate are the difference over the two decades of the estimated impulse

response at the quarter four years after the contractionary U.S. monetary policy shock. The numbers are shadowed when

estimated responses become less adverse during the 2000s for countries that see an increase in financial transaction with the

U.S.

Table 8: Response of Production-Related Variables and Regime Switches

Changes in Estimated Response to U.S. Monetary Policy Shock

IIP Employment

Unemployme

Average of Countries that adopt IT (Advanced) NA 0.571 -0.801

Average of Countries that adopt IT (Asia) 0.109 -0.181 -1.269

Average of Countries that adopt IT (Latin)

-0.088 -0.104 0.201

Average of Countries that adopt IT (All)

0.011 0.095 -0.668

Average of Countries that adopt Floating (Asia)

0.109 -0.364 -1.269

Average of Countries that adopt Floating (Latin)

-0.379 0.035 -0.610

Average of Countries that adopt Floating (All)

-0.216 -0.037 -0.699

Difference from 1990 to 2000

Note: Numbers are shadowed if the numbers are greater (smaller) than the averaged number

of all sampled countries for IIP and employment (for unemployment rate).

Table 9: List of Countries used for Estimation

IIP BEL BRA CAN CHI COL DEN FIN FRA ARG BEL BRA CAN CHI COL DEN FIN

GBR GER INA IRL ITA JPN KOR MEX FRA GBR GER INA IND IRL ITA JPN

NED SIN KOR MEX NED POR SIN SPN

Employment AUS BRA CAN CHI FIN GBR GER HKG AUS BRA CAN CHI FIN GBR GER HKG

JPN KOR NOR SWE TPE JPN KOR NOR SWE TPE

Umployment Rate AUS BEL BRA CAN CHI DEN FIN FRA AUS BEL BRA CAN CHI DEN FIN FRA

GBR GER HKG IRL ITA JPN KOR MEX GBR GER HKG IRL ITA JPN KOR MEX

NED NOR POR SPN SUI SWE TPE NED NOR POR SPN SUI SWE TPE

Real Export ARG AUS BEL BRA CAN CHN FIN FRA AR G AUS BEL BR A CAN CHN FIN FRA

GBR GER INA IND ITA JPN KOR MEX GBR GER INA IND ITA JPN KOR MEX

NOR SPN SUI SWE TPE NOR SPN SUI SWE

Trade Balance AUS BEL BRA CAN FIN FRA GBR GER AUS BEL BRA CAN FIN FRA GBR GER

INA IND ITA JPN KOR MEX NOR SPN INA IND ITA JPN KOR MEX NOR SPN

SWE SWE

Stock / ARG AUS CAN CHI EUR GBR GER HKG ARG AUS BRA CAN CHI CHN EUR GBR

Volatility of Stock Return INA IND JPN KOR PER PHI SPN SUI GER HKG INA IND JPN KOR MAS MEX

THA TPE NZL PER PHI SIN SPN SUI THA TPE

Exchange Rate / AUS BRA CAN CHI CHN COL FIN GBR AUS BRA CAN CHI CHN COL FIN GBR

Volatility of Exchange Rate HKG INA JPN KOR MAS MEX NOR PHI HKG INA JPN KOR MAS MEX NOR PHI

SIN SWE THA TPE SIN SWE THA TPE

Long Term Interest Rate AUS BEL BRA CAN COL DEN FIN FRA AUS BEL BRA CAN COL DEN FIN FRA

GBR GER HKG IRL ITA JPN KOR MEX GBR GER HKG IRL ITA JPN KOR MEX

NED NOR POR SIN SPN SUI SWE NED NOR POR SIN SPN SUI SWE TPE

Volatirity of AUS CAN DEN FRA GBR GER SPN SWE AUS CAN DEN FRA GBR GER HKG ITA

Long Term Interest Rate KOR MEX NOR PHI SIN SPN SUI SWE

THA

Policy Rate AUS CAN DEN GBR JPN NOR PHI SIN BRA CAN CHI CHN COL DEN EUR GBR

SUI SWE TPE HKG IND ISL JPN KOR NOR PER PHI

SIN SUI SWE THA TPE

Narrow Money AUS BEL BRA CAN CHN FIN GBR HKG AUS CAN DEN GBR INA JPN MAS NOR

INA IND IRL JPN KOR MAS NED SIN SIN SUI SWE

SUI TPE

Inflation ARG BEL BRA CAN CHI COL DEN FRA ARG BEL BRA CAN CHI COL DEN FRA

GBR GER HKG INA IND ITA JPN KOR GBR GER HKG INA IND ITA JPN KOR

MAS MEX NED POR SIN SPN SUI SWE MAS MEX NED POR SIN SPN SUI SWE

TPE

Notes:

Abbreviations used in the chart follow the country codes recognized by the International Olympic Committee. Each abbreviation represents a country or an area

as follows:

ARG

for Argentina,

AUS

for Australia,

BEL

for Belgium,

BRA

for Brazil,

CAN

for Canada,

CHI

for Chili,

C HN

for China,

COL

for Colombia,

DEN for Denmark, EUR for Euro area, FIN for Finland, FRA for France, GBR for the United Kingdom, GER for Germany, HKG for Hong Kong, INA for

Indonesia, IND for India, IRL for Ireland, ISL for Iceland, ITA for Italy, JPN for Japan, KOR for Korea, MAS for Malaysia, MEX for Mexico, NED for

Netherland, NZL for New Zealand, N OR for Norway, PER for Peru, PHI for Philippine, P OR for Portugal, S IN for Singapore, SPN for Spain, SUI for

Switzerland, SWE for Sweden, THA for Thailand and TPE for Taiwan.

1990's

2000's

Figure 1: Trade and Financial Integration

1990 1995 2000 2005 2010 2015

100

110

120

130

140

150

160

170

Trade Integration

1990 1995 2000 2005 2010 2015

Financial Integration

Note: Degree of trade integration is de…ned as the world import volume relative to world

GDP and degree of …nancial integration is de…ned as the global banks’ cross-border claims

relative to world GDP. The data are taken from IM F, BIS, and HAVER.

Figure 2: Share of the U.S. GDP over the World GDP

1980 1985 1990 1995 2000 2005 2010 2015

Share of U.S. GDP

Note: The data are taken from IMF.

Figure 3: Relative Importance of Trade and Financial Transaction with the U.S.

1990 1995 2000 2005 2010 2015

Trade with U.S.

1990 1995 2000 2005 2010 2015

Financial Transaction with U.S.

Note: Relative importance of trade with the United States in the world trade is de…ned

as the ratio of the sum of import to and export from the United States relative to the sum of

world import and world export. Relative importance of …nancial transactions with the United

States in cross-border claims is de…ned as the amount of holdings of U.S. banks’cross-border

claims relative to overall cross-border claims around the globe. The data are taken from IMF,

BEA, and BIS.

Figure 4: Time p ath of Fed eral Funds Rate and Extracted U.S. Monetary Policy Shocks

1990 1992 1994 1996 1998 2000 2002 2004 2006 2008

Federal Funds Rate

1990 1992 1994 1996 1998 2000 2002 2004 2006 2008

-0.2

-0.15

-0.1

-0.05

0.05

0.1

0.15

0.2

Extracted Monetary Poilcy Shocks

Figure 5: Response of IIP, Employment, Unemployment Rate

-1.5 -1 -0.5 0 0.5 1 1.5

0.5

1.5

2.5

3.5

4.5

Distribution of Response Across Countries

1990s

2000s

G90 00 E90 00 L90 00 As90 00 Al90 00

0.2

0.4

0.6

0.8

Proportion of Countries by Response

Positive

Negative

Non-Significant

-1 -0.8 -0.6 -0.4 -0.2 0 0.2 0.4 0.6 0.8 1

Distribution of Response Across Countries

1990s

2000s

G90 00 E90 00 L90 00 As90 00 Al90 00

0.2

0.4

0.6

0.8

Proportion of Countries by Response

Positive

Negative

Non-Significant

-3 -2 -1 0 1 2 3

0.2

0.4

0.6

0.8

1.2

1.4

Distribution of Response Across Countries

1990s

2000s

G90 00 E90 00 L90 00 As90 00 Al90 00

0.2

0.4

0.6

0.8

Proportion of Countries by Response

Positive

Negative

Non-Significant

Note: Estimation results for IIP, employment, and unemp loyment rate are reported. For

each variable, a bar graph depicts proportion of countries whose corresponding macroeco-

nomic variable exhibit statistically increase (black) or decrease (grey) at 5%, or non-signi…cance

(white) after the contractionary U.S. monetary policy shock by di¤erent country groups and

by estimation period. Lower graph depicts distribution of impulse response functions of macro-

economic variable four years after the sh ock across all sample countries for the sample period

of the 1990s (solid black line) and the 2000s (red line with circle).

Figure 6: Domestic Response When Trade with the U.S. Shrinks

0 5 10 15 20

-6

-5

-4

-3

-2

-1

x 10

-4

Production of B

Baseline

Less Trade with US I

Less Trade with US II

0 5 10 15 20

-1

-0.5

0.5

1.5

2.5

x 10

-4

Real Interest Rate of B

0 5 10 15 20

-5

-4

-3

-2

-1

x 10

-4

Consumption of B

0 5 10 15 20

0.01

0.02

0.03

0.04

0.05

0.06

Nominal Exchange Rate (AB)

0 5 10 15 20

-0.08

-0.07

-0.06

-0.05

-0.04

-0.03

-0.02

-0.01

0.01

Real Export of B

0 5 10 15 20

-0.12

-0.1

-0.08

-0.06

-0.04

-0.02

0.02

Real Import of B

0 5 10 15 20

-1

x 10

-4

Trade Balance of B

0 5 10 15 20

-5

-4

-3

-2

-1

x 10

-4

Inflation of B

Note: Figure displays the dome stic response of production, short-term real interest rate,

real consumption, nominal exchange rate, real export, real import, trade balance, and in‡ation

rate in country B to a positive shock to the monetary policy rule in country A under the

baseline economy, the economy where trade relationship between country A and B is smaller,

and the economy where trade relationship between country A and C is smaller compared with

the case of baseline economy.

Figure 7a: Domestic Response When Trade Integration Deepens

0 5 10 15 20

-8

-7

-6

-5

-4

-3

-2

-1

x 10

-4

Production of B

Baseline

Integrated (T)-I

Integrated (T)-II

0 5 10 15 20

-1

x 10

-4

Real Interest Rate of B

0 5 10 15 20

-8

-6

-4

-2

x 10

-4

Consumption of B

0 5 10 15 20

-0.01

0.01

0.02

0.03

0.04

0.05

0.06

0.07

Nominal Exchange Rate

0 5 10 15 20

-0.07

-0.06

-0.05

-0.04

-0.03

-0.02

-0.01

0.01

Real Export of B

0 5 10 15 20

-0.1

-0.08

-0.06

-0.04

-0.02

0.02

Real Import of B

0 5 10 15 20

-1

x 10

-4

Trade Balance

0 5 10 15 20

-6

-4

-2

x 10

-4

Inflation of B

Note: Figure displays the domestic response of production, short-term real interest rate, real

consumption, nominal exchange rate, real export, real import, trade balance, and in‡ation rate

in country B to a positive shock to the monetary policy rule in country A un de r the economy

where all parameters are calibrated to benchmark value, the economy where share of traded

goods relative to GDP in country B is larger than baseline model and the ec onomy where the

share of traded goods relative to GDP in country C is larger than baseline model.

Figure 7b: Domestic Response When Trade Integration Deepens

0 5 10 15 20

-8

-6

-4

-2

x 10

-4

Production of B

Baseline

Differentiated

Less Differentiated

0 5 10 15 20

-1

-0.5

0.5

1.5

2.5

x 10

-4

Real Interest Rate of B

0 5 10 15 20

-6

-5

-4

-3

-2

-1

x 10

-4

Consumption of B

0 5 10 15 20

0.01

0.02

0.03

0.04

0.05

0.06

0.07

Nominal Exchange Rate (AB)

0 5 10 15 20

-0.07

-0.06

-0.05

-0.04

-0.03

-0.02

-0.01

0.01

Real Export of B

0 5 10 15 20

-0.1

-0.08

-0.06

-0.04

-0.02

0.02

Real Import of B

0 5 10 15 20

-1

x 10

-4

Trade Balance of B

0 5 10 15 20

-5

x 10

-4

Inflation of B

Note: Figure displays the domestic response of production, short-term real interest rate, real

consumption, nominal exchange rate, real export, real import, trade balance, and in‡ation rate

in c ountry B to a positive shock to the monetary policy rule in country A un de r the baseline

economy, the economy where goods produced in each country is highly di¤erentiated, and the

economy where goods produced in each country is less di¤erentiated compared with the case

of baseline model.

Figure 8: Domestic Response When Relative Size of the U.S. Shrinks

0 5 10 15 20

-9

-8

-7

-6

-5

-4

-3

-2

-1

x 10

-4

Production of B

Baseline

Decrease of A

Increase of C

0 5 10 15 20

-0.5

0.5

1.5

2.5

x 10

-4

Real Interest Rate of B

0 5 10 15 20

-6

-5

-4

-3

-2

-1

x 10

-4

Consumption of B

0 5 10 15 20

-0.01

0.01

0.02

0.03

0.04

0.05

0.06

Nominal Exchange Rate

0 5 10 15 20

-0.07

-0.06

-0.05

-0.04

-0.03

-0.02

-0.01

0.01

Real Export of B

0 5 10 15 20

-0.1

-0.08

-0.06

-0.04

-0.02

0.02

Real Import of B

0 5 10 15 20

-1

x 10

-4

Trade Balance

0 5 10 15 20

-5

-4

-3

-2

-1

x 10

-4

Inflation of B

Note: Figure displays the dome stic response of production, short-term real interest rate,

real consumption, nominal exchange rate, real export, real import, trade balance, and in‡ation

rate in country B to a positive shock to the monetary policy rule in country A under the

baseline economy, the economy where country size of A becomes smaller, and the economy

where country size of C becomes larger compared with the case of baseline model.

Figure 9: Domestic Response When Cost of Accessing International Market is Small

0 5 10 15 20

-9

-8

-7

-6

-5

-4

-3

-2

-1

x 10

-4

Production of B

Baseline

Integrated (F)-I

Integrated (F)-II

0 5 10 15 20

-0.5

0.5

1.5

2.5

3.5

x 10

-4

Real Interest Rate of B

0 5 10 15 20

-6

-5

-4

-3

-2

-1

x 10

-4

Consumption of B

0 5 10 15 20

-0.01

0.01

0.02

0.03

0.04

0.05

0.06

Nominal Exchange Rate

0 5 10 15 20

0.004

0.006

0.008

0.01

0.012

0.014

0.016

Bond Holding of Country B

Baseline

Integrated (F)-I

Integrated (F)-II

0 5 10 15 20

0.5

1.5

2.5

x 10

-5

Interest Rate Deviation

(F)-I - Baseline

(F)-II - Baseline

0 5 10 15 20

0.2

0.4

0.6

0.8

1.2

1.4

1.6

x 10

-4

Relative Price -P AA over P AB- Deviation

(F)-I - Baseline

(F)-II - Baseline

0 5 10 15 20

0.6

0.8

1.2

1.4

1.6

1.8

2.2

2.4

x 10

-4

Real Exchange Rate Deviation

(F)-I - Baseline

(F)-II - Baseline

Note: Figure displays the domestic response of production, short-term real interest rate, real

consumption, nominal exchange rate, internationally traded bond holding in cou ntry B and

those in country A;interest rate of internationally traded bond, relative price of goods produced

in A and B in country A; and real exchange rate to a positive shock to the monetary policy

rule in country A under the baseline economy, the economy where country B faces less cost

in accessing the global …nancial market, and the economy where all countries face less cost in

accessing the global …nancial market. The bottom three panels display the deviation from the

baseline model.

Figure 10: Domestic Response When Alternative Policy is Adopted

0 5 10 15 20

-0.025

-0.02

-0.015

-0.01

-0.005

0.005

Prouction of B

Baseline

Fixed Exchange

0 5 10 15 20

-0.005

0.005

0.01

0.015

0.02

0.025

Real Interest Rate of B

0 5 10 15 20

-0.025

-0.02

-0.015

-0.01

-0.005

0.005

Consumption of B

0 5 10 15 20

-0.01

0.01

0.02

0.03

0.04

0.05

0.06

Nominal Exchange Rate

0 5 10 15 20

-0.07

-0.06

-0.05

-0.04

-0.03

-0.02

-0.01

0.01

Real Export of B

0 5 10 15 20

-0.1

-0.08

-0.06

-0.04

-0.02

0.02

Real Import of B

0 5 10 15 20

-8

-6

-4

-2

x 10

-4

Trade Balance of B

0 5 10 15 20

-3

-2.5

-2

-1.5

-1

-0.5

0.5

x 10

-3

Inflation of B

Note: Figure displays the domestic response of production, short-term real interest rate, real

consumption, nominal exchange rate, real export, real import, trade balance, and in‡ation rate

in c ountry B to a positive shock to the monetary policy rule in country A un de r the baseline

economy, the economy where country B adopts a monetary policy that maintains the nominal

exchange rate instead of Taylor rule, and the economy where country B adopts monetary p olicy

that assigns a higher weight on in‡ation rate.

Figure 11: Response of Export, Import, and Trade Balance

-2 -1.5 -1 -0.5 0 0.5 1 1.5 2

0.5

1.5

Distribution of Response Across Countries

G90 00 E90 00 L90 00 As90 00 Al90 00

0.2

0.4

0.6

0.8

Proportion of Countries by Response

1990s

2000s

Positive

Negative

Non-Significant

-4 -3 -2 -1 0 1 2 3 4

0.2

0.4

0.6

0.8

Distribution of Response Across Countries

1990s

2000s

G90 00 E90 00 L90 00 As90 00 Al90 00

0.2

0.4

0.6

0.8

Proportion of Countries by Response

Positive

Negative

Non-Significant

-0.02 -0.015 -0.01 -0.005 0 0.005 0.01 0.015 0.02

100

150

200

250

300

350

Distribution of Response Across Countries

1990s

2000s

G90 00 E90 00 L90 00 As90 00 Al90 00

0.2

0.4

0.6

0.8

Proportion of Countries by Response

Positive

Negative

Non-Significant

Note: Estimation results for real export, real import, and trade balance are reported.

For each variable, a bar graph depicts proportion of countries whose corresponding macroeco-

nomic variable exhibit statistically increase (black) or decrease (gray) at 5%, or non-signi…cance

(white) after the contractionary U.S. monetary policy shock by di¤erent country groups and

by estimation period. Lower graph depicts distribution of impulse response functions of macro-

economic variable four years after the sh ock across all sample countries for the sample period

of the 1990s (solid black line) and the 2000s (red line with circle).

Figure 12: Response of Stock Price and Exchange Rate, and Long Term Interest Rate

-10 -8 -6 -4 -2 0 2 4 6 8 10

0.1

0.2

0.3

0.4

0.5

Distribution of Response Across Countries

1990s

2000s

G90 00 E90 00 L90 00 As90 00 Al90 00

0.2

0.4

0.6

0.8

Proportion of Countries by Response

Positive

Negative

Non-Significant

-3 -2 -1 0 1 2 3

0.5

1.5

Distribution of Response Across Countries

1990s

2000s

G90 00 E90 00 L90 00 As90 00 Al90 00

0.2

0.4

0.6

0.8

Proportion of Countries by Response

Positive

Negative

Non-Significant

-1 -0.8 -0.6 -0.4 -0.2 0 0.2 0.4 0.6 0.8 1

Distribution of Response Across Countries

1990s (first 2 years)

2000s (first 2 years)

2000s II (3-4 years)

G90 00 E90 00 L90 00 As90 00 Al90 00

0.2

0.4

0.6

0.8

Proportion of Countries by Response

Positive

Negative

Non-Significant

Note: Estimation results for stock return, exchange rate, and long-run interest rate are

reported. For each variable, a bar graph depicts proportion of countries whose correspond-

ing macroeconomic variable exhibit statistically increase (black) or decrease (gray) at 5%, or

non-signi…cance (white) after the contractionary U.S. monetary policy shock by di¤erent coun-

try groups and by estimation period. Lower graph depicts distribution of impulse response

functions of macroeconomic variable after the shock across all sample countries for the sample

period of the 1990s (solid black line) and the 2000s (red line with circle).

Figure 13: Response of Volatility of Stock Price, Exchange Rate, and Long Term Interest Rate

-20 -15 -10 -5 0 5 10 15 20

0.05

0.1

0.15

0.2

0.25

0.3

0.35

Distribution of Response Across Countries

1990s

2000s

G90 00 E90 00 L90 00 As90 00 Al90 00

0.2

0.4

0.6

0.8

Proportion of Countries by Response

Positive

Negative

Non-Significant

-20 -15 -10 -5 0 5 10 15 20

0.02

0.04

0.06

0.08

0.1

0.12

0.14

0.16

Distribution of Response Across Countries

1990s

2000s

G90 00 E90 00 L90 00 As90 00 Al90 00

0.2

0.4

0.6

0.8

Proportion of Countries by Response

Positive

Negative

Non-Significant

-5 -4 -3 -2 -1 0 1 2 3 4 5

Distribution of Response Across Countries

1990s Advanced

2000s Advanced

2000s Latin Asia

G90 00 E90 00 L90 00 As90 00 Al90 00

0.2

0.4

0.6

0.8

Proportion of Countries by Response

Positive

Negative

Non-Significant

Note: Estimation results for volatility of stock return, exchange rate, and long term interest

rate are reported. For each variable, a bar graph depicts proportion of countries whose cor-

responding macroeconomic variable exhibit statistically increase (black) or decrease (gray) at

5%, or non-signi…canc e (white) after the contractionary U.S. m onetary policy shock by di¤erent

country groups and by estimation period. Lower graph depicts distribution of impulse response

functions of macroeconomic variable four years after the shock across all sample countries for

the sample period of the 1990s and the 2000s.

Figure 14: Response of Policy Rate, Narrow Money, and In‡ation Rate

-15 -10 -5 0 5 10 15

0.05

0.1

0.15

0.2

Distribution of Response Across Countries

1990s

2000s

G90 00 E90 00 L90 00 As90 00 Al90 00

0.2

0.4

0.6

0.8

Proportion of Countries by Response

Positive

Negative

Non-Significant

-2 -1.5 -1 -0.5 0 0.5 1 1.5 2

0.5

1.5

2.5

3.5

4.5

Distribution of Response Across Countries

1990s

2000s

G90 00 E90 00 L90 00 As90 00 Al90 00

0.2

0.4

0.6

0.8

Proportion of Countries by Response

Positive

Negative

Non-Significant

-0.01 -0.008 -0.006 -0.004 -0.002 0 0.002 0.004 0.006 0.008 0.01

100

150

200

250

300

350

400

Distribution of Response Across Countries

G90 00 E90 00 L90 00 As90 00 Al90 00

0.2

0.4

0.6

0.8

Proportion of Countries by Response

1990s

2000s

Positive

Negative

Non-Significant

Note: Estimation results for short-term policy rate, monetary aggregate, and in‡ation are

reported. For each variable, a bar graph depicts proportion of countries whose corresponding

macroeconomic variable e xhibit statistically increase (black) or decrease (gray) at 5%, or non-

signi…cance (white) after the contractionary U.S. monetary policy shock by di¤e rent country

groups and by estimation period. Lower graph depicts distribution of impulse response func-

tions of macroeconomic variable four years after the shock across all sample countries for the

sample period of the 1990s (solid black line) and the 2000s (red line with circle).