The evolution of Japanese real GDP components in the time–frequency domain: The disappearance of the investment accelerator effect?

• Published date: 01 September 2024
• Author: Patrick M. Crowley,David Hudgins
• Date: 01 September 2024
• DOI: http://doi.org/10.1111/asej.12336

ORIGINAL ARTICLE

The evolution of Japanese real GDP components in the

time–frequency domain: The disappearance of the

investment accelerator effect?

Patrick M. Crowley | David Hudgins

Department of Decision Sciences and Economics, College of Business, Corpus Christi, Texas, USA

Correspondence

Patrick M. Crowley, Department of Decision Sciences and Economics, College of Business, OCNR 323, 6300

Ocean Drive, Texas A & M University–Corpus Christi, Corpus Christi, TX 78412, USA.

Email: patrick.crowley@tamucc.edu

Abstract

Understanding the interaction between national income GDP expenditure components is

a key part of the study of the macroeconomics of any country. The general aim of this

study is to explore this interaction between the GDP expenditure components in the

time–frequency domain for the country of Japan, obtain some stylized facts and then

compare them with those of the United States. The main result shows that the cyclical

interactions between consumption and investment are typical of an “accelerator”effect

and operate at several different frequencies, but that the several different cycles evident in

the data phase out in the 1990s, once secular stagnation started to take hold. We hypoth-

esize that Japan’s investment drivers then changed to be dependent on the external sector

by the 2000s. A secondary result is that Japanese fiscal policy appears to have been

largely ineffective in terms of creating cyclical interactions with other GDP components.

KEYWORDS

accelerator effect, business cycles, economic policy, Japan, macroeconomic synchronization, time–

frequency domain

JEL CLASSIFICATION

C49, C61, C82, E32, E66, F40

1|INTRODUCTION

In this article, we use the continuous wavelet transform (CWT) with Japanese

data to gain insights on business and growth cycles and the interactions between

the GDP components. Given Japan’s special economic circumstances in

DOI: 10.1111/asej.12336

© 2024 East Asian Economic Association and John Wiley & Sons Australia, Ltd.

Received: 30 December 2023; Accepted: 27 August 2024

Asian Econ. J. 2024;38:341–370. wileyonlinelibrary.com/journal/asej 341

experiencing many decades of secular stagnation, and the fact that this period

now appears to be coming to an end, the analysis is particularly relevant to

assess what our expectations should be in relation to a return to more “normal”

economic circumstances. Previous studies which used time–frequency domain

methods, such as Kimura and Shiotani (2009) and Funashima (2014), have con-

tributed by explaining output fluctuations and policy effects, as many of these

phenomena cannot be extracted by traditional time series analyses.

To be more specific, this study explores the time–frequency interactions

between the GDP expenditure components in the time–frequency domain to

obtain stylized facts for Japan, and then compares them with those of the

United States. In particular, we investigate the relationship between the cyclical

components of growth in quarterly real GDP component data for Japan by

applying continuous wavelet analysis on cross comparisons of the data, and

then conduct the same exercise for the relationship between the Japanese com-

ponents and their US counterparts so as to detect any international cyclical

drivers. Our motivation is to explore the relationships between GDP compo-

nents based on phenomena such as the so-called investment accelerator effect,

fiscal policy mechanisms, as well as international business and growth cycle

propagation. This time–frequency approach allows for much better identifica-

tion of cyclical relationships in the data, and uncovering these relationships is

an important part of macroeconomic research.

The results show that the cyclical interactions between consumption and

investment are the most sustained and substantial at several different frequen-

cies, but that the several different cycles evident in the data at frequencies not

associated with the conventional business cycle peter out in the late 1980s and

early 1990s, and then by the early 2000s there appears to have been a reor-

ientation of the “accelerator”effect from the domestic to the external sector.

Besides the uniqueness of the methodological application to the Japanese econ-

omy, this latter observation is our main contribution to the literature on the evo-

lution of the Japanese macroeconomy over the last several decades, and it has

some important implications.

Section 2provides a literature review, while Section 3considers our data and

the methodology used. Section 4presents analysis of the results, Section 5con-

siders observations and implications, while Section 6concludes.

2|LITERATURE REVIEW

In this brief literature review, we first explore the academic literature in relation

to Japanese GDP components and then provide some fiscal and monetary pol-

icy backdrop to the analysis. Next, we then review the literature in relation to

the time–frequency methodology utilized here.

342 CROWLEY and HUDGINS