Industrial Upgrading in a Multiple‐cone Heckscher–Ohlin Model: The Flying Geese Patterns of Industrial Development

• Date: 01 February 2014
• Author: Kozo Kiyota
• Published date: 01 February 2014
• DOI: http://doi.org/10.1111/rode.12077

Industrial Upgrading in a Multiple-cone Heckscher–

Ohlin Model: The Flying Geese Patterns of

Industrial Development

Kozo Kiyota*

Abstract

This paper examines empirically how the multiple-cone version of the Heckscher–Ohlin (HO) model fits

the “flying geese” patterns of industrial development: a series of industries appear, prosper, then decline

and finally disappear one after another. Using Japanese manufacturing data from 1975 to 2006, the analysis

shows that the multiple-cone model fits well with the flying geese patterns of Japanese industrial develop-

ment. The result suggests that part of the industrial upgrading can be explained by the multiple-cone HO

model. This also implies that an underlying mechanism of macroeconomic growth is industrial upgrading,

part of which can be attributed to capital accumulation.

1. Introduction

Predating Raymond Vernon’s work (Vernon, 1966) on the product cycle theory by

several decades, Kaname Akamatsu identified two empirical regularities of industrial

development (Akamatsu, 1935, 1937).1One is a time-series pattern of a given indus-

try, shifting from imports to import-substitution for the domestic market, and exports

(Figure 1, Panel A).2The other is a time-series production pattern across industries,

in which a series of industries appear, prosper, then decline and finally disappear one

after another (Figure 1, Panel B).3Because these patterns reflect wild geese flying in

ranks, they are called flying geese patterns of industrial development and are referred

to extensively in explaining the production patterns of the East and Southeast Asian

countries.4

Several studies have tried to develop a theoretical foundation for flying geese pat-

terns. For example, Kojima (1958) explained shifts in production patterns, based on a

three-good, two-factor Heckscher–Ohlin (HO) model.5Ishikawa (1992) developed a

theoretical model that describes the interdependencies between economic growth and

industrial upgrading, incorporating learning-by-doing as the source of endogenous

growth.6In contrast, empirical studies have focused on how the flying geese patterns

are different across industries or how they spread across countries over time.7Not

much attention has been paid to the link with trade theory.

* Kiyota: Keio Economic Observatory, Keio University, 2-15-45, Minato-ku, Mita, Tokyo 108-8345 Japan.

Tel: +81-3-5427-1480; Fax: +81-3-5427-1640; E-mail: kiyota@sanken.keio.ac.jp. The author has benefited

greatly from conversations on this topic with Toshihiro Okubo. Thanks are also owed to Kazuhiko Yokota,

Eng Yoke Kee, seminar participants at Chukyo University, Keio University, Tohoku University, Tokyo

Institute of Technology, Okayama University, Okinawa University and conference participants at the 6th

Annual Meeting of the Asia Pacific Trade Seminars, 2010 Japanese Economic Association Spring Meeting,

2010 Taipei International Conference on Growth, Trade and Dynamics, and the 12th International Con-

vention of the East Asian Economic Association for their helpful comments. Financial support received

from the Japan Society for the Promotion of Science Grant-in-Aid (for Young Scientists and A-22243023)

and Japan Center for Economic Research is gratefully acknowledged. The usual disclaimers apply.

Review of Development Economics, 18(1), 177–193, 2014

DOI:10.1111/rode.12077

© 2014 John Wiley & Sons Ltd

This paper attempts to fill the gap between the theoretical and empirical studies on

the flying geese patterns of industrial development. The focus of the paper is on one

aspect of the flying geese patterns: a series of industries appear, prosper, then decline

and finally disappear one after another (Figure 1, Panel B). To explain the flying

geese patterns of industrial upgrading, the analysis in this paper builds upon the

multiple-cone version of the HO model.8Using Japanese manufacturing data from

1975 to 2006, our empirical analysis shows that the multiple-cone HO model is con-

sistent with the flying geese patterns of Japanese industrial development.

The analysis in this paper also contributes to the literature on economic growth,

by clarifying an underlying mechanism of macroeconomic growth. While a number

of empirical studies, such as World Bank (1993), have found that capital accumula-

tion is an important determinant of economic growth, only a few empirical attempts

have been made to explain how economies shift their production from labor-

to capital-intensive industries with economic growth.9In the international trade

literature, however, studies such as Leamer (1987) and Schott (2003) examined

the relationship between sectoral per capita output and the capital–labor ratio

(i.e. paths of development) in a multiple-cone setting.10 However, these studies

focused on the cross-country patterns of production and not on the time-series pat-

terns.11 This paper proposes a simple framework to integrate these two strands of

research.

Panel A

Volume of

(1) imports Consumer goods Capital goods

(2) production

(3) exports

(2)

(2)

(3) (3)

(1)

(1)

Underdeveloped stage Advanced stage Time

Panel B

Volume of

production Industry 4

Industry 3

Industry 2

Industry 1

Time

Figure 1. Flying Geese Patterns of Industrial Development

Panel A. Akamatsu’s original flying geese patterns.

Panel B. Flying geese patterns of production across different industries implied by Panel A.

Source: Akamatsu (1961, p. 206) for Panel A. The author’s representation for Panel B.

178 Kozo Kiyota

© 2014 John Wiley & Sons Ltd