Manufacturing Skill‐biased Wage Inequality, Natural Resources and Institutions

• Published date: 01 November 2017
• DOI: http://doi.org/10.1111/rode.12287
• Date: 01 November 2017

Manufacturing Skill-biased Wage Inequality, Natural

Resources and Institutions

Nuno Torres,



Oscar Afonso, and Isabel Soares*

Abstract

We use an extensive dataset on occupational wages to measure the manufacturing skill premium and

assess, for the first time, the influence of natural resources and institutional quality—in addition to

traditional drivers—for advanced and less-advanced countries and the full sample. The new findings,

regarding 21 countries between 1988 and 2008 in the main panel estimations, suggest the premium of

advanced countries rises with tertiary enrollment, net foreign direct investment (FDI) and institutional

quality, and falls with centralized wage negotiations and geographically diffuse natural resource activities,

mainly re-exportation related. In less-advanced countries, the premium rises with net FDI, scale effects,

centralized wage negotiations and geographically concentrated natural resource activities (absorbing

scarce skilled workers), and falls with trade, diffuse natural resource exploration (using mainly unskilled

workers) and high-technology exports, as emerging national low-end technology industrial exporters may

lower skill pay compared with foreign industrial exporters. In the full sample, the premium rises with

scale effects, trade, institutional quality and concentrated natural resources, and falls with the relative

skilled-labor supply, centralized wage negotiations and diffuse natural resources. The results account for

a wider diversity of situations compared with the previous studies.

1. Introduction

There is wide evidence that intra-country skill wage premia have increased in many

developed countries since the 1980s despite an increase in the skilled-labor

proportion (e.g. Nickell and Bell, 1996; Machin and Van Reenan, 1998; Acemoglu,

2003a,b; Autor et al., 2008). The same trends occurred in several (newly

industrialized) developing countries, such as Hong Kong, India, Thailand and

Uruguay (e.g. Zhu and Trefler, 2005). The rise in wage inequality also ocurred in

Latin America and East Asia (e.g. Avalos and Savvides, 2006) and in Russia (e.g.

Brainerd, 1998). The two major explanations—provided by the skill-biased

technological change (SBTC) literature and by the international trade literature—

contradict at least one of these observed trends (e.g. Wood, 1998; Afonso, 2012;

Burstein and Vogel, 2016) as shown in section 2.

Despite the SBTC wider acceptance in the literature, the theoretical debate

dominates the empirical research. Empirical studies usually analyze the impact of

just one explanation and ignore cross-country analysis. In order to close the gap

between the theoretical debate and the empirical research, this study uses an

extensive dataset on occupational wages (across 171 countries) to widen the

empirical research on skill–wage inequality in a panel analysis that considers intra-

and inter-country variation for 63 countries from 1983 to 2008,

1

although the main

results only cover 21 countries since 1988. The dataset on occupational wages is the

*Afonso (Corresponding author): University of Porto, Faculty of Economics and CEFAGE-UBI, Rua

Dr Roberto Frias, 4200-464, Porto, Portugal. Tel.: +351-91-999-1111; Fax: +351-22-550-5050; E-mail:

oafonso@fep.up.pt. Torres: University Lusiada – Norte and CEFUP, Porto, Portugal. Soares: University

of Porto, Faculty of Economics, Porto, Portugal.

Review of Development Economics, 21(4), e1–e29, 2017

DOI:10.1111/rode.12287

©2016 John Wiley & Sons Ltd

standardized International Labour Organization (ILO) October Inquiry 1983–2008

(2012 version), an updated and improved version of Freeman and Oostendrop

(2000, 2001).

As with most empirical studies on the subject, we focus on the skill premium of

the industrial sector, where the traditional explanations apply. However, our panel

study tests for the first time, the influence of natural resources and institutional

quality in addition to the traditional drivers for technologically advanced and less-

advanced country groups and for the full sample. In particular,

(i) We assess the effect of natural resource abundance and institutions on the

industrial skill premium taking into account institutional explanations of the

resource curse result—this is a novel approach that adds to the wage inequality

literature;

(ii) We re-ev aluate th e tradit ional dr ivers of t he indus trial sk ill premi um,

specifically SBTC, FDI and trade, while controlling for scale effects, the

relative skilled-labor supply, labor market institutions, governing

institutions and natural resources—this means that, unlike most studies on

the subject, we test the several traditional determinants at the same time

and improve the estimation by using panel data and an enlarged set of

control variables;

(iii) In addition to full-sample effects, we estimate the several skill premium drivers

for both advanced and less-advanced countries, which is also new to the best

of our knowledge.

The main results, found with an estimated panel of 21 countries between 1988

and 2008, suggest the manufacturing skill premium of technologically advanced

countries increases with tertiary enrollment, net FDI and the quality of governing

institutions. This correlates with overall institutional quality and decreases with the

centralization of wage negotiations and the use of unskilled workers by

geographically diffuse natural resource re-exportation activities. In less-advanced

countries, the skill premium augments with net FDI, scale effects, the centralization

of wage negotiations and scarcity of skilled workers absorbed by geographically

concentrated natural resource activities. It decreases with trade, the use of unskilled

workers in diffuse natural resource exploration (assessed with export shares) and

also high-technology exports, as the emergence of national low-end technology

industrial exporters may lower skill pay compared with foreign industrial exporters.

The full-sample effects show that the skill premium rises with scale effects, trade (a

result not found in the country groups), institutional quality and geographically

concentrated resource export shares. It decreases with the relative skilled-labor

supply (this estimate implies an elasticity of substitution of 10, which is large but in

line with a few studies), centralized wage negotiations and geographically diffuse

natural resources.

We believe that one of the main advantages of our paper with respect to the

previous studies lies in the ability to take into account a larger variety of situations

—namely for developing countries—by introducing several control variables not

considered before, even if this is done at the expense of a reduction in the number

of countries and observations. However, in a complex reality we prefer to have a

more robust set of explanatory variables to explain a higher diversity of situations

taking the risk of drawing conclusions from a smaller resulting set of countries and

e2 Nuno Torres, Oscar Afonso, and Isabel Soares

©2016 John Wiley & Sons Ltd