Globalization and urban polarization

• DOI: http://doi.org/10.1111/roie.12366
• Date: 01 November 2018
• Published date: 01 November 2018
• Author: Anthony J. Venables

Rev Int Econ. 2018;26:981–996. wileyonlinelibrary.com/journal/roie © 2018 John Wiley & Sons Ltd

|

981

1 | INTRODUCTION

The benefits of globalization for consumers are spread widely across income groups and regions in

a country, but the production impact is concentrated on particular sectors and firms. This sectoral

concentration maps into differential impacts on towns and cities, as well as on skill types. Adjustment

to such spatially concentrated shocks seems, from the experience of many countries, to be particu-

larly difficult. Cities in the rustbelt of the United States or former manufacturing areas in the north

DOI: 10.1111/roie.12366

SPECIAL ISSUE PAPER

Globalization and urban polarization

Anthony J. Venables1,2,3

Paper written for the conference in celebration of Jim Markusen on the occasion of his retirement, September 2017, Boulder

CO. Thanks to participants in the seminar and discussant Bob Staiger for helpful comments.

1Department of Economics, University of

Oxford, Oxford, UK

2Centre for Economic Policy Research,

London, UK

3International Growth Centre, London, UK

Correspondence

A. J. Venables, Department of Economics,

University of Oxford, Manor Road, Oxford

OX1 3UQ, UK.

Email: tony.venables@economics.ox.ac.uk

JEL Classification

F12, R11, R12

Abstract

External trade affects the internal spatial structure of an

economy, promoting growth in some cities or regions and

decline in others. Internal adjustment to these changes has

often proved to be extremely slow and painful. This paper

combines elements of urban and international economics

to draw out the implications of trade shocks for city per-

formance. Localization economies in production of inter-

nationally tradable goods mean that cities divide into two

types, those producing tradables and those specializing in

sectors producing just for the national market (nontrada-

bles). Negative trade shocks (and possibly also some posi-

tive ones) reduce the number of cities engaged in tradable

production, increasing the number producing just nontra-

dables. This has a negative effect across all nontradable

cities, which lose population and land value. Remaining

tradable cities boom, gaining population and land value.

Depending on the initial position, city size dispersion may

increase, thus raising the share of urban land rents in na-

tional income and reducing the share of labor.

982

|

VENABLES

of the United Kingdom and northern France have spent decades in relative, and sometimes absolute,

decline. The more recent experience of the China trade shock has been mapped into regional impacts,

in particular by the work of Autor et al. (for a survey see Autor, Dorn, & Hanson, 2016). In this work,

adjustment failure is viewed as a consequence of labor market rigidities, typically a combination of in-

adequate labor mobility, insufficient wage flexibility, or both, leading to an outcome with persistently

low levels of employment in affected areas.1

This paper explores a different mechanism through which trade or technology shocks create per-

sistent regional and urban inequalities. We assume a perfectly functioning labor market—with factor

mobility and wage flexibility—and focus on the difficulties faced in establishing new tradable sectors

in cities that have lost a traditional tradable goods sector. The difficulties arise as many of these sectors

are subject to agglomeration economies that create a first‐mover problem; in the absence of coordinated

behavior, it is not profitable to start a new activity in a new place. This problem is compounded by the

fact that, within a country, factor mobility limits the magnitude of spatial wage variation meaning that

costs do not fall enough to attract activity into a place that has experienced a negative shock. In short, the

usual mechanisms of comparative advantage do not operate to draw activity into an adversely affected

area.

We show circumstances under which trade shocks—positive as well as negative—lead to a polar-

ization of the urban structure as some cities boom and others decline. Declining cities maintain eco-

nomic activity, but it is in sectors producing nontradable goods and services. As an increasing number

of cities are relegated to specializing in these sectors the relative price of nontradables declines, am-

plifying the gap between cities. The counterpart of cities that have lost tradable sectors is booming

cities that have retained tradable goods or service sectors. These expand, driving up land prices in

these cities and possibly also the share of rent in income (actual and imputed to land and property

owners). These booming cities are as much a feature of trade shocks as are declining cities that have

lost competiveness in tradable goods production.2

The paper presents a simple model of this process, combining elements of urban economics and

international trade. At the heart of the model are three features. The first is the presence of agglomer-

ation economies, that is, external economies of scale, operating at the city or region level and within

a particular productive sector (localization economies). It is this that gives rise to spatial clustering of

sectors, meaning that different cities within a country will undertake different economic activities and

that trade shocks will impact different areas in different ways. The model assumes that all tradable

sectors are subject to these economies, although non‐tradable sectors are not. A further consequence of

agglomeration economies is that it is difficult to establish new centers of production for tradable goods.

A city that has lost a traditional sector offers neither low enough factor prices nor any productivity ad-

vantage in a new sector. Coordinated action by firms—or a “large developer”—can solve the problem,

but without this no firm wants to be the first in the sector to establish in a new location, uncertain as to

whether it will be followed by other firms which will create the cluster and raise productivity.

The second feature is that we distinguish between tradable and nontradable sectors, where by non-

tradables we mean goods produced for the national rather than the international economy, and hence

have their price set domestically rather than fixed at world prices.3 A central feature of the model is

that cities in a country will divide between those producing tradables and those producing nontrad-

ables. Trade shocks may cause cities to switch activity with, for example, cities that lose their tradable

sector due to import competition defaulting to non‐tradable production.

The third feature is that, within a country, the performance of a city depends largely on its absolute

advantage, not its comparative advantage. If a country’s export sector has a negative shock the ad-

justment mechanism is a real depreciation, that is, a reduction in its wage and unit costs relative to its

trading partners, and this reduction continues until other sectors become competitive. If a city within