The Long‐run Relationship Between Trade and Population Health: Evidence from Five Decades

• Author: Dierk Herzer
• Date: 01 February 2017
• Published date: 01 February 2017
• DOI: http://doi.org/10.1111/twec.12419

The Long-run Relationship Between Trade

and Population Health: Evidence from

Five Decades

Dierk Herzer

Department of Economics, Helmut Schmidt University, Hamburg, Germany

1. INTRODUCTION

THE world has witnessed a dramatic increase in trade over the last 50 years. This increase

has sparked a debate, mainly among epidemiologists and political scientists, about the

impact of international trade on health. In 2009, The Lancet, a leading public health journal,

published a series on trade and health, calling for increased attention to the potential adverse

consequences of trade for health. In an introductory comment to The Lancet Series on Trade

and Health, MacDonald and Horton (2009, p. 273) state: ‘[T]he fact that trade directly and

indirectly affects the health of the global population with an unrivalled reach and depth

undoubtedly makes it a key health issue that the global health community can no longer

ignore’.

However, although the World Health Assembly passed a resolution in 2006 requesting the

Director General of the World Health Organization to work with other international organisa-

tions to generate and share evidence on the relationship between trade and health (World

Health Assembly, 2006), only two published studies have empirically evaluated the effect of

total trade on population health in a large sample of countries. Owen and Wu (2007), using

panel data based on five-year intervals for the period 1960–95, find that growth in trade

improves population health, as measured by life expectancy and infant mortality, but the effect

diminishes as income rises to higher levels and is not significant for developed countries. Simi-

larly, Stevens et al. (2013) conclude, based on five-year data for the period 1970–2005, that

trade has a positive effect on population health, but their estimates imply that the effect

decreases as the level of income increases and becomes negative at high levels of income.

Given the small number of time-series observations (every five years), the results of these

studies are primarily driven by the cross-sectional rather than time-series variation in the data.

However, both trade and health are dynamic phenomena that vary considerably across coun-

tries and over time. It is therefore an open question whether these results hold whe n more

attention is paid to the time-series dimension of the data, and this is one question this study

attempts to answer.

The objective of this paper is to examine the long-run relationship between trade and

health by applying panel time-series methods to annual data from 74 countries over the period

1960–2010. Specifically, this paper makes the following contributions.

The first and main contribution is to estimate the long-run effect of trade on health using

panel cointegration estimation methods. These methods are known to be robust (under cointe-

gration) to a variety of estimation problems that often plague empirical work, including omit-

ted variables and endogeneity (as discussed in more detail in Section 3). To preview the main

I am grateful to an anonymous referee for many helpful suggestions and constructive comments.

©2016 John Wiley & Sons Ltd

462

The World Economy (2017)

doi: 10.1111/twec.12419

The World Economy

result, I find that trade openness has a statistically and quantitatively significant positive

impact on population health. This result is robust to a variety of sensitivity tests, including

the use of different cointegration techniques, potential outliers, different samples and sample

periods, and alternative definitions of the dependent and independent variables.

Another question this study addresses is whether countries with certain characteristics ben-

efit more, on average, from trade than others. To answer this question, and thus gain insight

into possible factors that determine the effect of trade on life expectancy, on average, across

countries, I present panel estimates of the long-run trade-life expectancy coefficients for

selected groups of countries. This is the second contribution of this paper. My results suggest

that the effect of trade on life expectancy tends to be greater in countries with lower develop-

ment levels, higher taxes on income, profits, and capital gains (as a percentage of GDP), and

less restrictive business and labour market regulations.

The paper is organised as follows. In Section 2, I discuss hypotheses concerning the effects

of trade on health. Section 3 sets out the basic empirical model and describes the data. The

econometric implementation and the estimation results are pres ented in Section 4, and

Section 5 concludes.

2. HYPOTHESES ABOUT THE EFFECTS OF TRADE ON HEALTH

Epidemiologists, political scientists and economists have proposed several mechanisms by

which increased trade could result in a change in health status over time (see also Dollar,

2001; Blouin et al., 2009). In the following, I group these mechanisms under seven headings:

income (Section 2a), inequality (Secti on 2b), access (Section 2c), insecur ity (Section 2d),

unemployment (Section 2e), pollution (Section 2f) and aid (Section 2g).

a. Income

If an increase in income leads to an increase in expenditures, both public and private,

on goods that improve health (such as food, clean water and sanitation, education and med-

ical care), the effect of trade on population health through changes in the level of income

will be positive unless trade has no effect on income or the effect of trade on income is

negative.

According to the textbook model of international trade, with perfect factor mobility, perfect

competition and no market failures (such as external effects or increasing returns), trade has a

positive effect on real income. It is well known, however, that if these standard assumptions

do not hold, the theoretical effect of trade on real income is not necessarily positive, which is

also supported by empirical evidence. Borrmann et al. (2006) and Freund and Bolaky (2008)

find that trade does not increase income in economies with excessive business and labour reg-

ulations. According to the results of Herzer (2013), trade raises income in most countries, but

in some countries, particularly in the developing world, increased trade is associated with

lower levels of income. Herzer (2013, p. 209) also finds that many developing economies are

highly regulated and speculates that this ‘may, at least partly, explain why [...] some devel-

oping countries lose from trade’.

Another qualification is the observation that increases in income are strongly associated

with increases in life expectancy among poorer countries, but as income rises, the relationship

flattens out and is weaker or even absent among the richest countries (Deaton, 2003, 2004). If

health gains from an extra unit of income diminish as income increases, as the concave

©2016 John Wiley & Sons Ltd

TRADE AND HEALTH 463