Structural change: Pace, patterns and determinants

• Author: Pedro M. G. Martins
• Published date: 01 February 2019
• DOI: http://doi.org/10.1111/rode.12555
• Date: 01 February 2019

REGULAR ARTICLE

Structural change: Pace, patterns and

determinants

Pedro M. G. Martins

The World Bank, Dili, Timor-Leste

Correspondence

Pedro M. G. Martins, The World Bank.

Avenida dos Direitos Humanos, Dili,

Timor-Leste.

Emails: pmartins@worldbank.org or

pedromgmartins@gmail.com

Abstract

This paper provides a comprehensive assessment of struc-

tural change in the world economy. The analysis relies on

a newly constructed dataset comprising 169 countries and

covering the period from 1991 to 2013. Shapley decom-

positions are employed to evaluate the pace and pattern

of structural change across regions and sub‐regions.

Country‐level estimates are then used to conduct an origi-

nal empirical exercise on the determinants of structural

change. The results suggest that labor reallocations (struc-

tural change) have played a critical role in enhancing eco-

nomic performance since the early 2000s, even if they

remain comparatively less important than within‐sector

productivity improvements. The widespread reallocation

of labor from agriculture to the services sectors has been

the key driver of structural change. Finally, we find

robust evidence that the pace of structural change is sig-

nificantly shaped by human and physical capital. The pol-

icy implication is that investments in education and

economic infrastructure are crucial to accelerating

structural change.

KEYWORDS

labor productivity, structural change

Disclaimer: The views expressed in this paper are those of the author and do not necessarily reflect the views of The World

Bank or the United Nations (where the author worked at the time of the writing).

DOI: 10.1111/rode.12555

Rev Dev Econ. 2019;23:1–32. wileyonlinelibrary.com/journal/rode © 2018 John Wiley & Sons Ltd

|

1

1

|

INTRODUCTION

The economic growth literature has largely relied on theoretical models underpinned by an aggre-

gate production function (such as Solow's neoclassical growth model), thus emphasizing the role

of economy‐wide factor accumulation and productivity. These one‐sector models have provided a

theoretical foundation for countless empirical studies investigating the determinants of econom ic

growth through econometric methods and growth accounting frameworks. In particular, the seminal

work of Barro (1991) on cross‐country growth regressions opened up a vast and prolific field of

empirical research.

1

However, several studies have shown that the empirical results tend to be sen-

sitive to model specification, sample data, and estimation method; see, for example, Levine and

Renelt (1992) and Pritchett and Summers (2014). This lack of robustness might be partly due to

one‐sector models not accounting for the large sector heterogeneity that is characteristic of devel-

oping economies. In fact, Eberhart and Teal (2013) demonstrate that the aggregation of heteroge-

neous sectors in cross‐country growth regressions can have a considerable impact on inference.

These critiques have contributed to a renewed interest in dual‐economy models and the role of

structural change in the growth process (McMillan & Heady, 2014). While one‐sector growth mod-

els were originally conceived with developed economies in mind, it can be argued that structural

(dual‐sector) models provide a better representation of developing economies. Temple (2005), for

instance, asserts that dual‐economy models should take center‐stage in the analysis of economic

growth in developing countries.

2

These models assume the coexistence of a relatively ‘advanced’

sector and a relatively ‘backward’sector in the economy –for example modern versus traditional,

industry versus agriculture, capitalist versus subsistence, or formal versus informal (Fields, 2007).

Moreover, they acknowledge that productivity gaps across sectors can be an important source of

economic growth (Lewis, 1954). These gaps can be seen as allocative inefficiencies and thus

opportunities to catalyze growth. The reallocation of labor across sectors assumes particular impor-

tance. Changes in the structure of employment not only are important for boosting economic

growth, but also can ensure that the benefits of growth are equitably distributed across society –

since workers in the lagging sector are unlikely to experience significant increases in living

standards.

3

The early literature on structural change dates back to the 1950s and 1960s. Kuznets (1957),

Chenery (1960), and Chenery and Taylor (1968) uncover important stylized facts on the relation-

ship between a country's economic structure and its income level. This literature posits that struc-

tural change is a key characteristic and driver of economic and social development. In fact, the

historical experience of developed and emerging economies confirms that sustained economic

development requires structural change. The reallocation of factors of production across sectors

with different productivity levels can induce economic gains or losses, depending on the direction.

4

Typically, growth‐enhancing structural change is narrowly defined as a process whereby labor

moves from low‐productivity to higher‐productivity sectors (McMillan, Rodrik, & Verduzco‐Gallo,

2014). This reallocation of labor raises workers’productivity, which contributes to accelerating

aggregate productivity and output growth. These ‘between‐sector’effects are in contrast to ‘within‐

sector’effects, which relate to labor productivity improvements within a specific sector, often

achieved through enhanced skills, complementary capital, improved technology, better management

practices, and resource reallocations.

5

Broader definitions of structural change go beyond changes

in economic structure (such as production and employment) as they also encompass changes in

other aspects of society (Kuznets, 1966). For instance, structural change may entail a spatial reor-

ganization of the population (through rural–urban migration) and demographic change (arising

from lower fertility rates). This paper uses a decomposition strategy that enables an empirical

2

|

MARTINS

assessment that is compatible with a broader view of structural change –by assessing the contribu-

tion of demographic and employment changes to economic performance, in addition to the relative

importance of between‐sector and within‐sector productivity effects.

The identification of key sectoral drivers can shed light on the patterns of structural change.

Historically, successful countries transformed from agrarian societies into industrial societies, an d

only subsequently into services‐based economies. Whether this ought to be the path for today's

developing countries is the subject of contentious debate. In developing countries, labor productiv-

ity in agriculture is considerably lower than in the non‐agricultural sector (Gollin, Lagakos, &

Waugh, 2014). This suggests that a reallocation of labor from agriculture to industry and/or ser-

vices would considerably boost aggregate productivity and economic growth. Meanwhile, agricul-

tural productivity is likely to rise, as labor‐saving technologies are adopted and (surplus) labor

moves out of the sector. Manufacturing is often seen as the critical sector for engendering struc-

tural change, due to increasing returns to scale, high tradability, and strong backward/forward link-

ages to agriculture and services. While the sector has certainly played an important role in the rise

of today's developed countries, growing levels of automation might be reducing its p otential to

absorb large numbers of workers.

6

Services, especially those associated with knowledge and inno-

vation, may also be able to produce structural change and thus sustain economic growth, as the

recent experience of India seems to suggest.

7

Whether services can be a substitute for manufactur-

ing, or merely a leading/lagging complement, is a key issue for policy‐making (Roncolato &

Kucera, 2014; Kucera & Roncolato, 2016). If they are a substitute, then countries may be able to

‘leapfrog’manufacturing in the traditional development path. However, services can be a leading

complement, if they increase demand for manufactured goods (e.g. an expanding IT sector requir-

ing computer hardware and other physical infrastructure), or a lagging complement, if they depend

on demand from the manufacturing sector (e.g. finance and insurance sectors relying on the perfor-

mance of manufacturing firms). Historical experience is seldom unequivocal, since even in China

manufacturing was not the sole driver of economic performance, and neither has India neglec ted

its manufacturing sector.

The recent emphasis on structural change has led to a rapidly expanding body of theoretical

and empirical work. Datasets have been compiled to document regional patterns, with varying

degrees of sectoral disaggregation and country coverage. However, the majority of studies have

small country samples and there have been very few attempts to empirically assess the determi-

nants of structural change in developing countries. This paper contributes to this emerging litera-

ture by constructing a comprehensive dataset and providing deeper insights into the recent

dynamics of structural change. The sample includes 169 countries, which enhances the representa-

tiveness of regional estimates and enables a sub‐regional perspective to evaluate the level of

heterogeneity within regions. More importantly, the paper scrutinizes the determinants of structural

change in a novel way to offer insights on how to enhance it.

2

|

METHODOLOGY AND DATA

2.1

|

Shapley decompositions

Most studies measuring the pace and pattern of structural change focus on the decomposition of

aggregate labor productivity growth. This paper adopts a broader analytical framework with a view

to providing further insights. In addition to assessing the contribution of within‐sector and

between‐sector productivity effects to economic performance, the impact of employment rates and

demographic change is also evaluated. Higher employment rates can boost economic activity,

MARTINS

|

3