Sources of heterogeneous gains from trade: Income differences and non‐homothetic preferences

• Published date: 01 November 2018
• DOI: http://doi.org/10.1111/roie.12368
• Author: Peter H. Egger,Sergey Nigai
• Date: 01 November 2018

Rev Int Econ. 2018;26:1021–1039. wileyonlinelibrary.com/journal/roie

|

1021

© 2018 John Wiley & Sons Ltd

1

|

INTRODUCTION

Global wage inequality is tightly linked to productivity‐ and remoteness‐related income differences

within and across countries. These sources of heterogeneous effects of globalization on wages are

customary in quantitative work in trade which aims at examining the link between trade and inequal-

ity. However, the second source of heterogeneous effects of trade on income difference around the

globe—non‐homotheticity in preferences—has been largely ignored. We explore the relative impor-

tance of these two factors in explaining the link between globalization and global wage inequality.

DOI: 10.1111/roie.12368

SPECIAL ISSUE PAPER

Sources of heterogeneous gains from trade: Income

differences and non‐homothetic preferences

Peter H. Egger1

|

Sergey Nigai2

1Department of Management, Technology,

and Economics,ETH Zurich, Zurich,

Switzerland

2Department of Economics,University of

Colorado Boulder, Boulder, CO

Correspondence

Sergey Nigai, Department of Economics,

University of Colorado Boulder, 256 UCB,

Boulder, CO 80309‐0256.

Email: sergey.nigai@colorado.edu

Abstract

This paper considers the interrelationship between a hier-

archy of consumption needs of agricultural goods, manu-

factures, and services as a central feature of the demand

side of economies and the demand for skilled and un-

skilled workers, whose incomes differ. The relationship is

established as skilled and unskilled labor are used in con-

junction with capital goods in a nested constant‐elastic-

ity‐of‐substitution (CES) production function, which

features capital‐skill complementarity and a substitutive

relationship between the bundle of capital and skills on the

one hand and unskilled labor on the other hand. Incomes

differ across countries both within and between types of

workers. As trade costs change, relative and absolute in-

comes of worker types change and so does the worker‐

type‐specific pattern of consumption. A calibration and

simulation exercise documents that a consideration of the

supply‐and‐demand‐side linkage in this hierarchy‐of‐

needs model matters quantitatively for the role of trade

liberalization in the world economy.

1022

|

EGGER and NIGAI

For that, we employ a multi‐sector, multi‐country model of international trade, where workers of

different wages are linked to their skill‐level and sector affiliation. The model features capital‐skill

complementarity such that an increase in the trade (e.g., through a decline in trade costs) of capital

goods results in a relative increase in the nominal wage of high‐skilled workers. We also introduce

non‐homotheticity such that the consumption patterns of workers depend on the level of their income

which allows us to address real wage inequality taking into account wage heterogeneity as well as

consumer‐price heterogeneity within and across different countries. We use the model to find the

impact of changes in trade costs on measures of global income inequality as well as welfare across all

workers, skilled workers and unskilled workers. We also contrast the results predicted by the model

here with more customary models that employ homothetic preferences or homogeneous workers.

We find that under the counterfactual scenario of no changes in trade costs between 1995 and 2009,

global income inequality as measured by the Gini coefficient would have been lower than observed

in the data. However, the models that lack both non‐homothetic preferences and capital‐skill comple-

mentarity (or one of these two mechanisms) significantly overestimate the effect of changes in trade

costs on global wage inequality. This is because such models tend to underestimate the gains from

trade of the workers in the top 10 percent and overestimate them for those in the bottom 10 percent of

the wage distribution. While this is true for all workers, we find that the results are more pronounced

for skilled workers. For example, a model with homothetic preferences and no capital‐skill comple-

mentarity mechanism overestimates the gains from trade of skilled workers in the 10th percentile by

33 percentage points and underestimates them for the 90th percentile by 6 percentage points.

The present paper relates to the recent work on the role of non‐homothetic preferences in open‐

economy quantitative models that alludes to the heterogeneity of the gains from trade due to income‐

related demand patterns. However, most papers in that literature focus on demand‐driven differences

in international trade patterns across countries (see), whereas this paper incorporates both margins of

unequal gains from trade: between and within countries. Our work is most closely related to Caron,

Fally and Markusen (2014), Markusen, 1986) which is an exceptional example of a situation in which

income differences accruing to skill‐related income heterogeneity and non‐homothetic preferences

are considered jointly in a quantitative model. We complement their approach by introducing an alter-

native preference structure and a capital‐skill complementarity mechanism that governs the nominal

wage distribution. Our focus is also largely different from theirs, as we aim at explaining the link

between changes in trade costs across countries and sectors and global wage inequality.

Traditional quantitative work on non‐homothetic preferences permits consumption patterns to vary

with the income level of factor owners (for example in ), and typically such work specifically distin-

guishes between agricultural goods, for which there is a basic need, and manufactures which will only

be consumed beyond a certain income level. In this literature, services tend to be in the shadow of

attention.1 In contrast to that, the present paper specifies a demand structure with specific minimum

income threshold levels not only pertaining to manufactures but also to services. This provides for a

hierarchy of needs where consumers first satisfy their needs related to agricultural consumption fol-

lowed by the consumption of manufactures and then by services.2 In general, the contribution of this

paper relative to earlier work on non‐homothetic preferences is threefold. First, we combine across‐

and within‐country heterogeneity in skills and wages with a novel hierarchical three‐layer utility func-

tion. This combination allows us to calculate global inequality measures that take into account both

sources of heterogeneity which turns out to be important for understanding the evolution of global

inequality. Second, in our counterfactual experiments we isolate and examine pure effects of changes

in trade costs across different sectors on global wage inequality in general equilibrium, while keeping

all other factors constant. Finally, we contrast the results implied by the model with non‐homothetic