Trade and Wage Inequality: Local versus Global Comparative Advantages

• DOI: http://doi.org/10.1111/j.1467-9701.2010.01293.x
• Date: 01 December 2010
• Author: Paolo Naticchioni,Alessia Matano
• Published date: 01 December 2010

Trade and Wage Inequality:

Local versus Global

Comparative Advantages

Alessia Matano

1

and Paolo Naticchioni

2

1

University of Rome ‘La Sapienza’ and

2

University of Cassino, University of Rome ‘La

Sapienza’, and CELEG (Luiss)

1. INTRODUCTION

IN recent decades, the effects of international trade openness on wage

inequality have been at the core of both the theoretical trade literature and

the policy debate. The reference framework, the Heckscher–Ohlin–Samuelson

(HOS hereafter) theorem, states that increasing trade between developed and

developing countries causes wage inequality to increase in developed countries

(DCs) and to decrease in developing countries. However, the empirical evi-

dence provides ambiguous findings with respect to HOS predictions. While

some authors have then explored alternative theoretical frameworks (Haskel

and Slaughter, 2001; Acemoglu, 2003; Feenstra and Hanson, 2003; Epifani and

Gancia, 2008), we claim that taking into account the geographical dimension of

trade can be crucial to reconcile the HOS framework with the empirical evi-

dence. In accordance with the theoretical contribution of Davis (1996), analysts

and policymakers should evaluate carefully the different impacts of ‘local’

trade (trade within the cone of diversification a country belongs to) and ‘global’

trade (trade with the other cone of diversification) on wage inequality. Similar

intuitions are also derived in Wood (1997). The literature is lacking in

This paper was produced within the scope of the research project IEIILM, funded by the Italian

Ministry of University and Research (MIUR, PRIN project). We thank the research partnership

between ISFOL – Area Mercato del Lavoro (Rome) – and Dipartimento di Economia – University

of Rome ‘La Sapienza’ – for access to the databases. We are grateful to Paolo Epifani for his

precious advice and support. We also thank Antony Atkinson, Massimiliano Bratti, Riccardo Mas-

sari, Paolo Piacentini, Emiliano Rustichelli and an anonymous referee for helpful comments and

suggestions.

The World Economy (2010)

doi: 10.1111/j.1467-9701.2010.01293.x

2010 Blackwell Publishing Ltd., 9600 Garsington Road,

Oxford, OX4 2DQ, UK and 350 Main Street, Malden, MA 02148, USA. 1757

The World Economy

empirical studies that might bear out these conclusions, as also stressed by

Davis (1996).

1

This paper provides a direct test for the predictions of Davis (1996) and

Wood (1997) as it investigates the impact of ‘local’ versus ‘global’ trade on

wage inequality by using data on exports and imports classified by groups of

origin ⁄destination countries: DCs and less developed countries (LDCs

hereafter).

We focus on the case of a DC, Italy, which represents an interesting case

study because of its specialisation patterns. On the one hand, Italy is special-

ised in the production of unskill-intensive goods and relies on imports of capi-

tal-intensive goods within its own cone of diversification (the DC cone).On

the other hand, Italy exports medium skill-intensive goods and imports unskill-

intensive goods from the LDC cone of diversification.

We use a matched employer–employee database provided by the Italian

Social Security Institute (INPS), merged with regional–sectoral data on exports

and imports classified according to groups of origin⁄destination countries, col-

lected by ISTAT (Coeweb: Italian International Trade Statistics).

The analysis is carried out at both individual and aggregate levels. In the

former, we regress the real-gross weekly wages on the export and import

variables and on a set of individual, firm and industrial characteristics. In

order to analyse how trade separately impacts on the wage structure of blue

collars and white collars, we carry out separate estimates for these two occu-

pational categories. The individual-level analysis also provides robustness to

the aggregate level of analysis, since it allows us to control for both individ-

ual (observed and unobserved) and firm heterogeneity. Moreover, since fixed

effect estimates are also affected by movers’ decisions, we derive our pre-

ferred estimates exploiting the exogeneity of the group of displaced workers,

thereby taking into account the aspects related to the endogeneity of mobility

decisions.

In the aggregate-level analysis, the individual information is collapsed to

derive a regional–sectoral database. Thus we compute two indices of inequal-

ity (the Gini index and the average wage ratio of white collars to blue col-

lars) at the regional–sectoral level and we aggregate the individual variables

at the same level. We then regress the indices of inequality on the trade and

control variables to analyse the impact of international trade on wage

inequality.

Our results show that the local trade – with the DCs – produces a

strong impact on wages and inequality. In particular, exports towards the DCs

1

Actually, some papers have already exploited the ideas of Davis (1996) to provide an interpreta-

tion of wage inequality patterns: Gouveia and Tavares (1995) for Portugal; Hanson and Harrison

(1999) for Mexico. However, they did not directly test them.

2010 Blackwell Publishing Ltd.

1758 A. MATANO AND P. NATICCHIONI