On the Sources of the Feldstein–Horioka Puzzle across Time and Frequencies
| Author | Jun‐Hyung Ko,Yoshito Funashima |
| Published date | 01 August 2019 |
| DOI | http://doi.org/10.1111/obes.12293 |
| Date | 01 August 2019 |
889
©2019 The Department of Economics, University of Oxford and JohnWiley & Sons Ltd.
OXFORD BULLETIN OF ECONOMICSAND STATISTICS, 81, 4 (2019) 0305–9049
doi: 10.1111/obes.12293
On the Sources of the Feldstein–Horioka Puzzle across
Time and Frequencies*
Jun-Hyung Ko† and Yoshito Funashima‡
†Aoyama Gakuin University, 4-4-25 Shibuya Shibuya-ku, Tokyo 150-8366, Japan (e-mail:
kojunhyung@aoyamagakuin.jp)
‡Tohoku Gakuin University, 1-3-1 Tsuchitoi, Aoba-ku, Sendai, Miyagi 980-8511, Japan
(e-mail: funashima@mail.tohoku-gakuin.ac.jp)
Abstract
This study quantitatively assesses existing explanations for the Feldstein–Horioka puzzle
using time–frequency domain analyses for nine countries for the period 1885–2010. The
main findings are summarized as follows. First, large economies (e.g. the United States,
Italy) showhigher cor relations betweensaving and investment than middle-sized and small
countries do. Second, countries can be grouped into two time-changing patterns of corre-
lations: inverted U-shaped and increasing patterns. Third, the fiscal balance seems most
related to a positive saving-investment correlation in many countries. Fourth, a global
common factor plays an important role in explaining the Feldstein–Horioka puzzle.
I. Introduction
Feldstein and Horioka (1980) claim there is an absence of perfect capital mobility, based on
their regression of investment on saving for 21 OECD countries for the period 1960–1974.
Feldstein and Horioka show that the slope parameter on saving is significantly different
from zero, but not from one, which implies that domestic investment depends mainly on
domestic saving.1The authors interpret their estimated large valueof the saving–investment
coefficient in terms of the long-run immobility of international capital and the imperfect
integration of international capital markets. This so-called Feldstein–Horioka puzzle has
stimulated numerous debates on international capital mobility.2
Most of the existing literature on the Feldstein–Horioka puzzle focuses on the cross-
section or time-series relationship, relying on time-domain analyses.The seminal paper by
JEL Classification numbers: C32, F21, F32.
*The authors are very grateful to the editor in charge (Jonathan Temple) and two anonymous referees for numer-
ous helpful comments on earlier versions of this paper. The authors are also grateful to Hiroshi Gunji, KoichiYano
and participants in the 11th International Conference of Western EconomicAssociation International, 2015 Autumn
Meeting of the Japanese Economic Association, and the 6th Meiji University Economic Conference. Ko acknowl-
edges financial support from Grant-in-Aid for Young Scientists (B) (Grant Number 15K17095) and Grant-in-Aid for
Scientific Research (C) (Grant Number 18K01696) from JSPS.
1In this study, we use ‘saving’to mean the saving rate, and ‘investment’ to mean the investmentrate.
2For an extensivesur vey,see Apergis and Tsoumas (2009), among others.
890 Bulletin
Feldsteinand Horioka (1980) is an example of a standard cross-section analysis. Other stud-
ies have investigatedhow the correlation between national saving and domestic investment
has changed overtime. For example, difference-in-differences methods, constant parameter
models with a time trend, time-varying parameter models, and Markov-switching models
have been used to capture changes in the saving–investment relationship.3Another strand
of literature investigates whether the Feldstein–Horioka puzzle is a short-run or a long-run
phenomenon.4These studies investigate the cointegration relationship or use a long-run
identification in vector auto-regressive (VAR) models to capture the long-run relationship.
On the other hand, Levy (2000) uses the frequency-domain framework to simultaneously
examine long-run, cyclical, and short-run saving–investment relationships.
The nexus between saving and investment can be both time-dependent and cycle-
dependent. However, existing studies suffer from a common drawback in that their models
only partially examine saving–investment links. If they focus on the timing, for example,
they have to forgo the frequency of the Feldstein–Horioka puzzle, and vice versa. Hence,
to overcome these limitations, an alternative method is needed to investigate all perspec-
tives of the Feldstein–Horioka puzzle, including short-run, long-run, and time changes. In
this respect, wavelet analysis is a promising tool because it allows us to take into account
comovements in both time and frequency within a unified framework.
In this study, we provide new evidence on the relationship between saving and in-
vestment by examining data on nine countries for the period from the late 19th century
to the present.5Following Aguiar-Conraria, Azevedo and Soares (2008) and Rua (2010),
we use a wavelet coherency analysis to investigate the Feldstein–Horioka puzzle, taking
into account both the time and the frequency domains. In other words, we simultaneously
investigate how saving and investment are related at different frequencies, and how the
pattern of comovements between the two has changed over time.
Our three main findings based on our wavelet coherency analysis are as follows. First,
high correlations between saving and investment are more observable in large economies,
such as Italy and the United States, but less so in smaller countries, such as Argentina and
Sweden.
Second, there are two groups of countries in terms of the time-change aspect of the
Feldstein–Horioka puzzle. Larger countries, such as Italy, Spain, the United Kingdom and
the United States, exhibit inverted U-shaped patterns of saving–investment correlations.
In other words, for these countries, the correlation between saving and investment was
low in the early part of the sample period until the 1920s and 1930s. Then, it increased
substantially in the middle period, before decreasing to a low value in the 1960s and
1970s. Our evidence supports the view of Obstfeld and Taylor (2003), who find that capital
mobility exhibits a U-shaped pattern. On the other hand, for countries such as Argentina,
Australia, Canada, Finland, and Sweden, we find that the positive link between saving and
investment becomes stronger in the latter period.
3For example, see Hussein (1998), Georgopoulosand Hejazi (2005), Evans, Kim and Oh (2008), Fouquau, Hurlin
and Rabaud (2008), and Katsimi and Zoega (2016), among others.
4See, for example, Sarno and Taylor (1998), Corbin (2004), and Hoffmann (2004), among others.
5The nine countries are the United States, the United Kingdom, Italy, Canada, Spain,Argentina, Australia, Sweden
and Finland, ranked in descending order by size, measured using aggregate GDP.
©2019 The Department of Economics, University of Oxford and JohnWiley & Sons Ltd
To continue reading
Request your trial