Populism and the economics of globalization

Populism and the economics of

globalization

Dani Rodrik

John F. Kennedy School of Government, Harvard

University, Cambridge, MA 02138, USA

Correspondence:

D Rodrik, John F. Kennedy School of

Government, Harvard University,

Cambridge, MA 02138, USA

e-mail: [email protected]

Abstract

Populism may seem like it has come out of nowhere, but it has been on the rise

for a while. I argue that economic history and economic theory both provide

ample grounds for anticipating that advanced stages of economic globalization

would produce a political backlash. While the backlash may have been

predictable, the speciﬁc form it took was less so. I distinguish between left-wing

and right-wing variants of populism, which differ with respect to the societal

cleavages that populist politicians highlight. The ﬁrst has been predominant in

Latin America, and the second in Europe. I argue that these different reactions

are related to the relative salience of different types of globalization shocks.

Journal of International Business Policy (2018).

https://doi.org/10.1057/s42214-018-0001-4

Keywords: populism; globalization; Latin America; Europe

INTRODUCTION

‘‘Populism’’ is a loose label that encompasses a diverse set of

movements. The term originates from the late nineteenth century,

when a coalition of farmers, workers, and miners in the US rallied

against the Gold Standard and the Northeastern banking and

ﬁnance establishment. Latin America has a long tradition of

populism going back to the 1930s, and exempliﬁed by Peronism.

Today populism spans a wide gamut of political movements,

including anti-euro and anti-immigrant parties in Europe, and

Syriza and Podemos in Greece and Spain, respectively, Trump’s

antitrade nativism in the US, the economic populism of Chavez in

Latin America, and many others in between. What all these share is

an antiestablishment orientation, a claim to speak for the people

against the elites, opposition to liberal economics and globaliza-

tion, and often (but not always) a penchant for authoritarian

governance.

I address two questions in this paper. First, what are the

economic roots of populism? Second, what are the factors that

affect the emergence of right- versus left-wing populism?

It may seem like populism has come out of nowhere. But the

populist backlash has been on the rise for a while, for at least a

decade or more (see Figure 1). More importantly, the backlash was

perfectly predictable. I will focus in this paper on the economic

Received: 2 November 2017

Revised: 6 December 2017

Accepted: 8 December 2017

Journal of International Business Policy (2018)

www.jibp.net

roots of populism, in particular the role of eco-

nomic globalization. I do not claim that globaliza-

tion was the only force at play – nor necessarily

even the most important one. Changes in technol-

ogy, rise of winner-take-all markets, erosion of

labor-market protections, and decline of norms

restricting pay differentials all have played their

part. These developments are not entirely indepen-

dent from globalization, insofar as they both

fostered globalization and were reinforced by it.

But neither can they be reduced to it. Nevertheless,

economic history and economic theory both give

us strong reasons to believe that advanced stages of

globalization are prone to populist backlash. I will

examine those reasons below.

The populist backlash may have been pre-

dictable, but the speciﬁc form it took was less so.

Populism comes in different versions. Here I will

distinguish between left-wing and right-wing vari-

ants of populism, which differ with respect to the

societal cleavages that populist politicians high-

light and render salient.

The US progressive

movement and most Latin American populism

took a left-wing form. Donald Trump and European

populism today represent, with some instructive

exceptions, the right-wing variant. A second ques-

tion I address below is what accounts for the

emergence of right-wing versus left-wing variants

of opposition to globalization.

I will suggest that these different reactions are

related to the forms in which globalization shocks

make themselves felt in society. It is easier for

populist politicians to mobilize along ethno-na-

tional/cultural cleavages when the globalization

shock becomes salient in the form of immigration

and refugees. That is largely the story of advanced

countries in Europe. On the other hand, it is easier

to mobilize along income/social class lines when

the globalization shock takes the form mainly of

trade, ﬁnance, and foreign investment. That in turn

is the case with southern Europe and Latin Amer-

ica. The US, where arguably both types of shocks

have become highly salient recently, has produced

populists of both stripes (Bernie Sanders and Don-

ald Trump).

I argue that it is important to distinguish between

the demand and supply sides of the rise in

populism. The economic anxiety and distributional

struggles exacerbated by globalization generate a

base for populism, but do not necessarily determine

its political orientation. The relative salience of

available cleavages and the narratives provided by

populist leaders is what provides direction and

content to the grievances. Overlooking this distinc-

tion can obscure the respective roles of economic

and cultural factors in driving populist politics.

The outline of the paper is as follows. I ﬁrst

discuss what economic theory and empirics have to

say about the distributive consequences of trade

liberalization (Trade and Redistribution section). In

light of that discussion, I turn to the economics and

politics of compensation (Compensation and

Safety Nets section). In the next section, I examine

questions of fairness and distributive justice, which

economists have generally stayed away from, but

are crucial to understand the populist backlash

(Trade, Redistribution, and Fairness section). In The

Perils of Financial Globalization section, I review

the consequences of ﬁnancial globalization. The

Political Economy of the Backlash section brings

the supply side of populism into the picture,

discussing conditions under which right-wing and

left-wing variants of populism are more likely to

thrive. Finally, I provide some concluding com-

ments in Concluding Remarks section.

TRADE AND REDISTRIBUTION

Why does globalization, in its many forms, cause

political conﬂict? How does the intensity of the

conﬂict vary over time, depending on the stage of

globalization, state of the business cycle, and other

factors? In view of the contentious history of the

ﬁrst era of globalization, what enabled the later

ﬂowering after the Second World War? And how

similar (or dissimilar) is the current populist

backlash?

Figure 1 The global rise of populism. Notes: see Appendix for

sources and methods.

Populism and the economics of globalization Dani Rodrik

Journal of International Business Policy

To anyone familiar with the basic economics of

trade and ﬁnancial integration, the politically

contentious nature of globalization should not be

a surprise. The workhorse models with which

international economists work tend to have strong

redistributive implications. In fact, the real puzzle is

that the world economy could achieve such high

levels of openness in recent decades and maintain

it for so long – a question I will pick up later.

Start with trade theory. Models of trade and

distribution are essentially exercises in tracing out

the effects of price changes on the material well-

being of identiﬁable economic groups. One of the

most remarkable theorems in economics is the

Stolper–Samuelson theorem (1941), which gener-

ates very sharp distributional implications from

opening up to trade. Speciﬁcally, in a model with

two goods and two factors of production, with full

intersectoral mobility of the factors, owners of one

of the two factors are made necessarily worse off

with the opening to trade. The factor which is used

intensively in the importable good must experience

a decline in its real earnings. Note that the theorem

establishes absolute losses, and not relative losses.

Note also that the result holds regardless of con-

sumption preferences: there are losses to one group

even if they spend most or even all of their budget

on the importable good (which gets cheaper in

relative terms), although the magnitude of the

losses is reduced. Applied with some amount of

hand-waving to the US economy, the result pre-

dicts that low-skilled workers are unambig uously

worse off as a result of trade liberalization.

The original Stolper–Samuelson theorem was

derived under very speciﬁc conditions, and it is

sometimes thought that more complicated, or

more realistic, models soften the edges of this

striking conclusion. This is true to some extent. But

there is one Stolper–Samuelson-like result that is

extremely general, and which can be stated as

follows. Under competitive conditions, as long as

the importable good(s) continue to be produced at

home – that is, ruling out complete specialization –

there is always at least one factor of production that

is rendered worse off by the liberalization of trade.

In other words, trade generically produces losers.

The proof of this result is simple enough to be

stated quickly here. Let’s express the unit cost of

production for the importable sector that is being

liberalized as c = u(w

, w

, …,w

), with w

denoting

the return to the ith factor of production used in

that sector. Since payments made to the factors

must exhaust the cost of production, changes in

unit costs are a weighted average of changes in

payments to each of the factors, where the weights

(in perfect competition) are the cost shares of each

factor.

In other words,

c ¼

; where a ‘‘hat’’

denotes proportional changes, h

is the cost share of

factor i, and

¼ 1:

Now consider what happens with trade liberal-

ization. The effect of trade liberalization is to raise

the domestic price of exportables relative to

importables. Let the importable described above

be the numeraire, with price ﬁxed at unity. We are

interested in what happens to the returns of factors

used in the importable. Since this good is the

numeraire, we have the equilibrium condition

c = u(w

, w

, …,w

) = 1, stating equality between

price and unit cost (the zero-proﬁt condition). As

long as the good continues to be produced, this

condition holds both before and after the liberal-

ization. Therefore,

¼ 0: Since the weighted

sum of factor price changes add up to zero, there

must be at least one factor of production, call it the

kth factor, such that

 0: (The inequality will be

strict when goods differ in their factor intensities.)

Meanwhile, exportable prices have increased

(

p [ 0), thanks to the liberalization. Hence,

 0\

p; and the return to the kth factor declines

in terms of both the importable and exportables,

producing an unambiguous fall in real returns,

regardless of the budget shares of the two goods.

This is known as the magniﬁcation effect in trade

theory and follows from the fact that factor price

changes must bracket price changes in neoclassical

equilibrium. Hence, its generality. The result that

openness to trade creates losers is not a special case;

it is the implication of a very large variety of

models, including those where labor is not partic-

ularly mobile across industries and regions.

The particular conﬁguration of gains and losses

depend on the details of the model. In the 2 9 2

case of the Stolper–Samuelson theorem,

\0\p\

; where w

and w

denote the returns

to low- and high-skill workers, respectively.

Another benchmark model in trade is the speciﬁc-

factors model, where each good has a factor that is

used only in that sector. In that model, the factors

that lose are those that are speciﬁc to the

importable sector(s).

More recent work in trade theory has emphasized

heterogeneity among ﬁrms and workers. These

models have additional margins for redistribution,

between ﬁrms and workers that otherwise look

quite similar. Grossman, Helpman, & Kircher

(2017), for example, enrich the Stolper–Samuelson

Populism and the economics of globalization Dani Rodrik

Journal of International Business Policy

framework by considering heterogeneity within

broad worker categories. ‘‘Managers’’ and ‘‘workers’’

must combine in teams, and their productivity

depends on the quality of the match. Trade liber-

alization induces re-matching and generates distri-

butional effects within occupations and industries,

in addition to the standard effects across broad

factors of production and industries.

Hence in all these models, redistribution is the

ﬂip side of the gains from trade. No pain, no gain.

This is standard economic fare – familiar to all trade

economists, even if not voiced too loudly in the

public.

Economic theory has an additional implication,

particularly germane to our discussion, which is

less well recognized. In relative terms, the redis-

tributive effects of liberalization get larger and tend

to swamp the net gains as the trade barriers in

question become smaller. In other words, the ratio

of redistribution to net gains rises as trade liberal-

ization tackles progressively lower barriers.

The logic is simple. Consider the denominator of

this ratio ﬁrst. It is a standard result in public

ﬁnance that the efﬁciency cost of a tax increases

with the square of the tax rate. Since an import

tariff is a tax on imports, the same convexity

applies to tariffs as well. Small tariffs have very

small distorting effects; large tariffs have very large

negative effects. Correspondingly, the efﬁciency

gains of trade liberalization become progressively

smaller as the barriers get lower.

The redistributive effects, on the other hand, are

roughly linear with respect to price changes and are

invariant, at the margin, to the magnitude of the

barriers. To a ﬁrst order of approximation, a

10 percent reduction in import prices has twice

the effect on factor prices of a 5 percent reduction.

Putting these two facts together, we have the result

just stated, namely that the losses incurred by

adversely affected groups per dollar of efﬁciency

gain are higher the lower the barrier that is

removed.

Table 1 shows some quantitative simulations

using a simple two-factor, two-goods trade model

(of the Stolper–Samuelson type): the ratio of losses

to net gains rises dramatically from a factor of 5

when tariffs are at 40 percent to more than 70

when tariffs are 3 percent!

Of course, the gains reaped by the winners, per

dollar of efﬁciency gain, also increase correspond-

ingly. But the main point is this: as globalization

advances and policy makers go after the remaining,

low barriers, trade agreements become more about

redistribution and less about expanding the overall

economic pie. This is possibly one important

reason why globalization becomes politically more

contentious in its advanced stages.

So much for theory. Does the evidence support

these sharp distributional predictions? We now

have some good empirical studies that have taken a

detailed look at the consequences of NAFTA and of

China’s entry to the WTO on labor markets in the

US. They ﬁnd that local labor-market effects in

affected communities were indeed sizeable.

The most careful analysis of NAFTA to date has

been carried out by Hakobyan & McLaren (2016).

These authors use US Census data and focus on the

1990–2000 period. They measure the NAFTA ‘‘trade

shock’’ by constructing an industry- and locality-

speciﬁc variable that measures vulnerability to

NAFTA. This indicator is a weighted average of

initial tariffs on Mexican imports, where the

weights are both local employment levels in each

industry and Mexico’s revealed comparative advan-

tage in those industries. To ensure they are captur-

ing the effects of NAFTA proper, they also control

for other trends that may be correlated with the

NAFTA shock such as the general expansion of

trade.

Hakobyan and McLaren ﬁnd that NAFTA pro-

duced modest effects for most US workers, but an

‘‘important minority’’ suffered substantial income

losses. They identify both a geographic and indus-

try effect. Regions that were most affected by tariff

reductions experienced signiﬁcantly slower wage

growth than regions that had no tariff protection

against Mexico in the ﬁrst place.

Not surprisingly,

the effect was greatest for blue-collar workers: a

high-school dropout in heavily NAFTA-impacted

locales had eight percentage points slower wage

growth over 1990–2000 compared with a similar

worker not affected by NAFTA trade. The industry

effect was even larger: wage growth in the most

protected industries that lost their protection fell

17 percentage points relative to industries that were

unprotected initially.

These are very large effects, especially when one

bears in mind that the efﬁciency gains generated by

NAFTA apparently have been tiny. Economists

have struggled to ﬁnd signiﬁcant net gains for the

US economy, largely because US tariffs vis-a

-vis

Mexico were quite low to begin with and Mexico is

so small relative to the US (less than a tenth in size).

The consensus of these studies is that there were

large trade effects, but miniscule aggregate income

effects. Romalis (2007) concludes that the overall

Populism and the economics of globalization Dani Rodrik

Journal of International Business Policy

effect for the US was approximately zero. A recently

published academic study by Caliendo & Parro

(2015) uses all the bells-and-whistles of modern

trade theory to produce the estimate that these

overall gains amount to a ‘‘welfare’’ gain of 0.08% –

800th of 1 percent – for the US trade volume

impacts were much larger: a doubling of US imports

from Mexico.

China is a much larger country, and its entry into

the WTO was a bigger deal for the US. While US

tariffs on imports from China did not change, the

uncertainty about the annual renewal of most-

favored nation status was removed and, as a result,

there was a large increase in the volume of trade. In

a well-known paper, Autor, Dorn, and Hanson have

documented the labor-market disruption caused by

the ‘‘China trade shock,’’ which was not only

massive but also very persistent (Autor, Dorn, &

Hanson, 2013; see also Autor, Dorn, & Hanson,

2016a; Autor, Dorn, Hanson, & Majlesi, 2016b).

These authors focus on the 1990–2007 period,

which covers the years before and after China’s

WTO entry. They identify the exogenous, supply-

driven increase in Chinese imports by instrument-

ing these by the contemporaneous change in

imports from China in eight other developed

countries. Their unit of analysis is the commuting

zone. Commuting zones have different composi-

tions of economic activity, and some have more

industries exposed to Chinese competition than

others. This permits an examination of the labor-

market effects of Chinese imports across different

localities.

Autor et al.’s baseline result is that a commuting

zone in the 75th percentile of exposure to Chinese

import growth had a differential fall of 4.5 percent

in the number of manufacturing employees and a

0.8 percentage point larger decline in mean log

weekly earnings, compared with a commuting zone

at the 25th percentile. They also ﬁnd a signiﬁcant

impact on overall employment and labor force

participation rates, indicating that this is an addi-

tional margin of adjustment to trade shocks. As the

authors stress, this implies that the wage reductions

are underestimated, both because of increase in

nonparticipation and the fact that the unemployed

are more likely to have lower ability and earnings.

Beyond redistributive costs, Autor et al. (2013)

point also to the adverse efﬁciency implications of

these ﬁndings. Involuntary employment and trans-

fers through the welfare system both induce non-

negligible efﬁciency losses that eat into the

standard gains from trade.

Moreover, these local labor-market effects appear

to have been highly persistent. The wage, labor-

force participation, and unemployment conse-

quences had not dissipated after a full decade of

the China trade shock (Autor et al., 2016a, b). The

offsetting employment effects in export-oriented

activities, which conventional trade models pro-

duce, had not taken place.

The studies I have just summarized look at the

local labor-market effects of speciﬁc trade shocks,

focusing on the variation across different geograph-

ical areas. There is also the question of trade’s effect

on the overall levels of earnings and employment.

There is an extensive literature on trade and wages

going back to the 1980s that ascribes some –

although not most – of the increase in wage

inequality in the advanced economies to trade.

Earlier studies tended to downplay the importance

of trade, placing much more emphasis on skill-

biased technological change as the dominant inﬂu-

ence behind the increase in the skill premium.

More recent studies have tended to give trade a

more prominent role, in part because globalization

has advanced and also because a sharp distinction

between trade and technology has become harder

to make (see Ebenstein, Harrison, McMillan, &

Phillips, 2014 and the references therein). Recent

evidence implicates import competition as the

most important factor behind the decline in labor

shares at the level of individual industries in the US

since the late 1980s (Elsby, Hobijn, & S¸ ahin, 2013).

COMPENSATION AND SAFETY NETS

In principle, the gains from trade can be redis-

tributed to compensate the losers and ensure no

identiﬁable group is left behind. Trade economists,

aware of the distributive implications of their

models, have long called for such compensation.

If successive US administrations had done a better

job at redistributing the gains from trade, could the

protectionist backlash have been avoided?

The European experience is instructive here.

Populism has certainly not spared Europe. But the

opposition to globalization has taken a different

political coloring, targeting Brussels and the EU’s

perceived intrusion in domestic policy and regula-

tions rather than trade agreements. European pop-

ulism is not antitrade or directed at speciﬁc

exporters such as China and Mexico. Some aspects

of trade agreements, such as investor-state dispute

settlement (ISDS) and regulatory harmonization,

are highly controversial in Europe as well. But

Populism and the economics of globalization Dani Rodrik

Journal of International Business Policy

neither right-wing nor left-wing populists have

pushed for trade barriers. In fact, Brexit advocates

in Britain presented free trade as an explicit objec-

tive. One of the advantages of leaving the EU, they

argued, was that it would enable Britain to pursue

policies closer to free trade!

One difference with the US that may account for

this contrast is that Europe has long had strong

social protections and a generous welfare state.

Most countries of Europe, being smaller than the

US, are much more open to trade. But openness to

trade has been accompanied by much greater

redistribution and social insurance. A number of

empirical analyses have shown that there is a direct

link between exposure to trade and expansion of

public transfers (Cameron, 1978; Rodrik, 1998a, b).

It is not an exaggeration to say that the European

welfare state is the ﬂip side of the open economy.

(Interestingly, the European backlash against immi-

grants and refugees has some of its roots in the

concern that the social beneﬁts of the welfare state

will be eroded or displaced.)

The US became a truly open economy relatively

late. The share of imports in GDP more than

doubled between the mid-1970s and the 2000s,

going from 7% in 1975 to 17% on the eve of the

ﬁnancial crisis in 2008. Much of the increase in

import penetration, especially during the 1990s,

came from low-income countries (China in partic-

ular), which created special adjustment problems.

In principle, US governments could have followed

the European model. It could have complemented

trade agreements – NAFTA, the WTO, and China’s

WTO entry – with much more robust social insur-

ance mechanisms and active labor-market pro-

grams and protections. Even though there was

much talk about such supports during the Clinton

administration, little was in fact done beyond

tinkering with the existing trade adjustment assis-

tance (TAA) mechanisms.

There are in fact two essential difﬁculties with

compensation. The economic problem is that

compensation can be very costly. Lump-sum taxes

and transfers are not practical and they are rarely

used. The taxes needed and the mechanisms used

to dispense the assistance both create behavioral

distortions – deadweight losses. These can be quite

large. Using a 40% ﬁgure for the marginal excess

burden of taxation, Autor et al. (2013) estimate that

the increase in transfers resulting from the Chinese

trade shock resulted in an annual deadweight loss

of $33 per capita. A more extensive welfare system,

along European lines, would likely produce bigger

such numbers.

Antra

s, de Gortari, & Itskhoki (2017) have recently

presented an interesting exercise to quantify the full

costs of compensation in an economy where tax-

transfer schemes are necessarily distortionary. They

construct a model in which opening up to trade

disproportionately beneﬁts the most productive

agents in the economy, exacerbating income

inequality. To counter this, an increase in the

progressivity of income taxation is needed, which

in turn produces adverse efﬁciency effects through

the labor-supply channel. They calibrate their model

using IRS tax returns data from the US and back out

the implied trade frictions from measures of trade

exposure. The authors compute both the required

adjustment to the gains from trade under a welfarist

criterion (compared to a distribution-neutral stan-

dard) and the losses incurred due distortionary

redistribution. Under their baseline parameteriza-

tion, ‘‘trade-induced increases in disposable income

inequality erode about 20% of the U.S. gains from

trade, while gains from trade would have been about

15% larger if redistribution had been carried out via

non-distortionary means.’’

Note that Antra

s et al. (2017) do not explicitly

model trade policy; their trade friction variable

takes the form of an ‘‘iceberg cost’’ and, unlike

import tariffs or quantitative restrictions, does not

create government revenue or rents. In the pres-

ence of such revenues, opening up to trade would

generally entail lower efﬁciency gains and greater

redistribution. The deadweight loss of compen-

satory distortionary taxation would be correspond-

ingly larger.

But perhaps the more serious difﬁculty with

compensation is the political one, and it relates to

credibility and time consistency. As long as revers-

ing trade agreements is costly, governments always

have the incentive to promise compensation, but

rarely to carry it out. The winners need the losers’

assent for the agreement. But once the agreement is

passed, there is little reason for the winners to

follow through. This is largely the story of TAA in

the US. Practically every trade agreement that the

US has signed has had a compensatory arrange-

ment attached to it in some form or another. Yet

there is widespread agreement that TAA and similar

measures have not proved very effective.

It is not

implausible to think that the reason is the lack of

political incentives ex post to render them

effective.

Populism and the economics of globalization Dani Rodrik

Journal of International Business Policy

Furthermore, trade agreements themselves have

played a part in the shifting bargaining power

between business and workers. They reﬂect the

decline in the political inﬂuence of organized labor.

And they reinforce that decline by changing the

rules and norms of competition in turn. Viewed

from this perspective, it is not surprising that

compensation has not featured prominently. A

labor movement strong enough to receive real

compensation would have been strong enough to

resist trade agreements with sharp redistributive

impacts – or at least to shape the agenda of

negotiations in a more worker-friendly direction.

In view of such economic and political difﬁcul-

ties, genuine compensation rarely occurs. It tends

to be an add-on to trade agreements or an after-

thought. Actual compensation works best when it

is embedded in the general social policies of a

nation, and not speciﬁcally targeted at trade

impacts. Where it has worked, as in Europe, it has

been part of a constitutive bargain between capital

and labor that couples the open economy with

generous safety nets. In these countries, there is

little need for compensatory policies targeted at

trade directly: individuals who are displaced, lose

their jobs, or need retraining can access those

broader mechanisms. European employers have

traditionally consented to the high costs of such

safety nets. And they have cemented their consent

by institutionalizing it in the form of the welfare

state.

The bargain enabled European nations to achieve

much higher levels of openness and exposure to

international competition than the US. It under-

pinned the return to high levels of globalization in

the decades following the end of the Second World

War. But this bargain too would fray eventually, as

we shall see, under the impact of capital mobility

and ﬁnancial globalization.

TRADE, REDISTRIBUTION, AND FAIRNESS

Economists understand that trade causes job dis-

placement and income losses for some groups. But

they have a harder time making sense of why trade

gets picked on so much by populists both on the

right and the left. After all, imports are only one

source of churn in labor markets, and typically not

even the most important source. Demand shocks,

technological changes, and the ordinary course of

competition with other, domestic ﬁrms produce

much greater labor displacement than increases in

import penetration. While disentangling the effects

of automation and globalization is difﬁcult, most

existing studies attribute the bulk of the decline in

US manufacturing employment to the former

rather than the latter.

Yet we do not see populists

campaign against technology or automation.

What is it that renders trade so much more salient

politically?

The usual answer is that trade is a convenient

scapegoat, since politicians can point to identiﬁ-

able foreigners – Chinese, Mexicans, or Germans –

as the source of the problem. This is no doubt true

to some extent. But there is another, deeper issue

that renders redistribution caused by trade more

contentious than other forms of competition or

technological change. Sometimes international

trade involves types of competition that are ruled

out at home because they violate widely held

domestic norms or social understandings. When

such ‘‘blocked exchanges’’ are enabled through

trade they raise difﬁcult questions of distributive

justice.

What arouses popular opposition is not

inequality per se, but perceived unfairness.

Three psychologists have recently provided inter-

esting support for this idea (Starmans, Sheskin, &

Bloom, 2017). They note that people tend to

express preference for equality in small groups,

but when asked about the ideal distribution for

their country (or large groups), they support an

unequal distribution of resources. The authors

argue that there is no evidence ‘‘people are actually

concerned with economic inequality at all.’’ What

they worry about is ‘‘something that is often

confounded with inequality: economic unfairness.’’

Fairness concerns are likely deeply embedded in our

evolutionary history as a strategy for dealing with

opportunistic behavior (i.e., to punish cheaters). At

the same time, people understand that unequal

abilities, effort, or moral deservingness imply that a

fair distribution in society would also be unequal.

As long as there is belief in social mobility, high

levels of inequality will be tolerated.

To relate these considerations to international

trade, consider the following thought experiment.

Suppose Harry and John own two ﬁrms that

compete with each other. Ask yourself how you

feel about each of the scenarios below. In each of

them, Harry outcompetes John, resulting in John

and his employees losing their jobs. Should they be

blocked or allowed to run their course?

(1) Harry works hard, saves and invests a lot, and

comes up with new techniques and products,

while John lags behind.

Populism and the economics of globalization Dani Rodrik

Journal of International Business Policy

(2) Harry ﬁnds a cheaper (or higher quality) sup-

plier in Germany.

(3) Harry outsources to a supplier in Bangladesh,

which employs workers in 12-h-a-day shifts and

under hazardous conditions.

(4) Harry brings Bangladeshi workers to the US

under temporary contracts, and puts them to

work under conditions that violate domestic

labor, environmental, and safety laws.

From the standpoint of our two competitors and

from a purely economic perspective, these scenar-

ios are isomorphic: each creates losers as well as

gainers in the process of expanding the overall size

of the economic pie for the national economy. In

each case, Harry’s gains are larger than John’s

losses.

But most audiences react very differently to

them. The manner in which the gains and losses

are generated matter. Few people have issues with

scenarios (1) or (2), which are perceived as accept-

able market outcomes even if there are losers. But

scenarios (3) and (4) appear problematic insofar as

they force John to compete with Harry under

ground rules that have been prohibited at home,

in domestic competition. Furthermore, scenario

(4) is illegal, even though its practical conse-

quences are identical to those of scenario (3).

Many economists who have no difﬁculty with the

outsourcing scenario (3) would regard scenario (4)

as unconscionable. But if we treat (3) as accept-

able, why should we not accept (4) as well?

Alternatively, if we reject (4), why would we

accept (3)?

The thought experiment clariﬁes why the general

public can view certain kinds of international

competition [represented by scenario (3) above] to

be in conﬂict with domestic norms as regards

what’s an acceptable redistribution. I take this to

be the essence of the ‘‘fair trade’’ argument. It’s one

thing to lose your job to someone who competes

under the same rules as you do. It’s a different thing

when you lose your job to someone who takes

advantage of lax labor, environmental, tax, or

safety standards in other countries.

The fairness

concern generalizes beyond the individuals who are

directly affected. Members of the broader commu-

nity are more likely to empathize with those who

lose their jobs or incur earning losses when they

believe that this is the result of ‘‘unfair’’ practices. In

the words of Rosanvallon (2016), ‘‘inequality is felt

most acutely when citizens believe that the rules

apply differently to different people.’’

The notion of fair trade is much derided by

economists, but it is already enshrined in trade laws

in the form of antidumping and countervailing

duties (used against dumped and subsidized goods,

respectively). These so-called trade remedies under-

cut the gains from trade – by blocking certain

exchanges – but they also enable political buy-in

for an open trading system. A greater appreciation

for the importance of procedural fairness may have

prepared economists for the brewing opposition to

an analogous form of dumping, one that might be

called ‘‘social’’ dumping.

Finally, it is worth noting that the nature of trade

agreements has changed over time, rendering them

more divisive in terms of value and fairness

considerations. Trade deals increasingly reach

behind the border to harmonize domestic regula-

tions. This is a form of economic integration that is

called deep integration, to distinguish it from the

shallow integration associated with tariff or quota

liberalization. Often, deep integration is pushed by

speciﬁc lobbies and special interests. For example,

pharma companies have played a major role in

bringing intellectual property rights protection into

trade agreements. Multinational companies have

pushed for special tribunals in which they can sue

host governments, so-called ISDS. Financial ﬁrms

have inserted clauses that make it difﬁcult or

impossible for signatories to manage cross-border

capital ﬂows, even though the usefulness of capital

controls is acknowledged today even by the Inter-

national Monetary Fund (IMF). Such features move

trade agreements away from efﬁciency enhance-

ment. They exacerbate clashes within nations on

the relative power of corporations or the desirabil-

ity of safeguarding regulatory autonomy.

THE PERILS OF FINANCIAL GLOBALIZATION

Economists remain on the whole in favor of free

trade, even though they recognize the distribu-

tional impacts may be sometimes adverse. The

presumption of the gains from trade – the aggregate

efﬁciency gains from eliminating barriers on cross-

border commerce – remains strong. In principle,

ﬁnancial globalization generates important eco-

nomic beneﬁts as well: it should channel savings

to countries where returns are higher, enable

intertemporal consumption for nations through

international borrowing and lending, and allow

global portfolio diversiﬁcation. Nevertheless, econ-

omists’ views on capital mobility have been more

ambiguous and prone to cycles.

Populism and the economics of globalization Dani Rodrik

Journal of International Business Policy

In the immediate aftermath of the Second World

War, the near-consensus of the economics estab-

lishment was in favor of controls on cross-border

capital ﬂows, to avoid the currency volatility of the

interwar period and leave room for domestic

macroeconomic management. By the 1990s, the

consensus was reversed. The IMF, OECD, and the

EU began to push for full capital-account convert-

ibility for the present and prospective member

states. Following the global ﬁnancial crisis of 2008–

2009, views changed yet again. The beneﬁts of

ﬁnancial globalization were downgraded, and cap-

ital controls became acceptable once more.

The economics profession’s current views on

ﬁnancial globalization can be best described as

ambivalent. Most of the skepticism is directed at

short-term ﬁnancial ﬂows, which are associated

with ﬁnancial crises and other excesses. Long-term

ﬂows and direct foreign investment in particular

are generally still viewed favorably. Direct foreign

investment tends to be more stable and growth-

promoting. But as I discuss below, there is evidence

that it has produced shifts in taxation and bargain-

ing power that are adverse to labor.

A recent paper by Broner & Ventura (2016)

captures well the mainstream perspective on the

anomalous consequences of ﬁnancial globalization.

The starting point is that opportunistic behavior by

governments and low-quality domestic institutions

make foreign ﬁnance subject to periodic debt crises.

In their model, when debt is domestic, debt con-

tracts are enforced and defaults are rare. The entry

of foreign creditors increases government incen-

tives to default. And when governments cannot

perfectly discriminate between foreign and domes-

tic creditors, foreign and domestic debt crises occur

together. This gives an incentive for domestic

residents to send their savings abroad. Financial

globalization therefore may result in the apparent

paradox that countries that should be normally

importing capital become exporters of capital

(‘‘capital ﬂight’’). Countries with deep ﬁnancial

markets where governments are loath to default

opportunistically become safe havens for foreign-

ers’ savings, and receive excessive ﬁnancial inﬂows.

The result of all of this is that the consequences of

ﬁnancial globalization for investment, growth, and

ﬁnancial stability are uncertain and cannot be

presumed beneﬁcial.

The essence of the Broner–Ventura story is that

ﬁnancial globalization accentuates the weakness of

domestic institutions and debt-enforcement mech-

anisms. Capital mobility interacts with domestic

market failures to produce adverse effects that

possibly offset its direct beneﬁcial effect.

This

type of the second-best reasoning is the reason why

many prominent economists such as Jagdish Bhag-

wati and Joe Stiglitz do not favor free capital

mobility, even though they are strong supporters

of free trade.

Such accounts also rationalize two sets of empir-

ical ﬁndings that pose problems for advocates of

ﬁnancial globalization. First, the association

between capital-account convertibility and eco-

nomic growth is weak at best (Rodrik, 1998a, b;

Schularick & Steger, 2010; Kose et al., 2011; Ostry,

Prati, & Spilimbergo, 2009). Typically, positive

growth effects are found either for speciﬁc periods

(e.g., during the late nineteenth century when

capital ﬂowed to the new world), speciﬁc subsam-

ples of countries (e.g., those with strong institu-

tions and macroeconomic management), or

particular types of capital (foreign direct invest-

ment). Second, there is a strong empirical associa-

tion between ﬁnancial globalization and ﬁnancial

crises over time. The canonical chart on this comes

from Reinhart & Rogoff (2009), who show that the

time trends of ﬁnancial globalization and the

incidence of banking crises coincidence coincide

almost perfectly. (Their chart is reproduced here in

Figure 2.) Banking crises and ﬁnancial globalization

rise and fall together.

The boom-and-bust cycle associated with capital

inﬂows has long been familiar to developing

nations. Prior to the global ﬁnancial crisis, there

was a presumption that such problems were largely

the province of poorer countries. Advanced econo-

mies, with their better institutions and regulation,

would be insulated from ﬁnancial crises induced by

ﬁnancial globalization. It did not quite turn out

that way. In the US, the housing bubble, excessive

risk-taking, and over-leveraging during the years

leading up to the ﬁnancial crisis were ampliﬁed by

capital inﬂows from the rest of the world. European

banks were major purchasers of US asset-backed

securities and the appetite of emerging-market

creditors for ‘‘low-risk’’ investments fueled the US

credit boom. In the words of Lane, ‘‘ﬁnancial

globalization magniﬁed the impact of underlying

distortions, such as inadequate regulation of credit

markets and banks, by allowing the scaling-up of

ﬁnancial activities that might have faced capacity

limits in autarkic ﬁnancial systems’’ (Lane, 2012:9).

In the euro zone, ﬁnancial integration, on a

regional scale, played an even larger role. Monetary

uniﬁcation and the introduction of the euro in

Populism and the economics of globalization Dani Rodrik

Journal of International Business Policy

1999 drove down risk premia in countries such as

Greece, Spain, and Portugal, and led to the conver-

gence of borrowing costs across member states. This

enabled borrowers to run large current account

deﬁcits and accumulate problematic amounts of

external debt. Construction and other nontradable

sectors were boosted in the receiving countries at

the expense of tradable activities. Such credit

booms would eventually turn into bust and sus-

tained economic collapses in Greece, Spain, Portu-

gal, and Ireland once credit dried up in the

immediate aftermath of the crisis in the US.

Financial globalization appears to have produced

adverse distributional impacts within countries as

well, in part through its effect on incidence and

severity of ﬁnancial crises. In a remarkable series of

papers, researchers at the IMF have documented

these negative inequality impacts (Jaumotte, Lall, &

Papageorgiou, 2013; Furceri & Loungani, 2015).

Most noteworthy is the recent analysis by Furceri,

Loungani, & Ostry (2017) that looks at 224 episodes

of capital-account liberalization, most of them

taking place during the last couple of decades.

Liberalization episodes are identiﬁed by big

changes in a standard measure of ﬁnancial open-

ness (the Chinn–Ito index) and large subsequent

capital ﬂows. They ﬁnd that capital-account liber-

alization leads to statistically signiﬁcant and long-

lasting declines in the labor share of income and

corresponding increases in the Gini coefﬁcient of

income inequality and in the shares of top 1, 5, and

10 percents of income (see Figure 3 for their key

results).

Furthermore, these adverse effects on

inequality are stronger in cases where de jure

liberalization was accompanied by large increase

in capital ﬂows.

Financial globalization appears to

have complemented trade in exerting downward

pressure on the labor share of income.

Why would ﬁnancial globalization increase

inequality and the capital share in particular? There

is no analogue to trade theory’s Stolper–Samuelson

theorem in international macroeconomics. So to

some extent these distributional consequences of

ﬁnancial globalization are a genuine surprise. But

there may be an obvious, bargaining-related expla-

nation (as argued in Rodrik, 1997, Chap. 2). As long

as wages are determined in part by bargaining

between employees and employers, the outside

options of each party play an important role. Capital

mobility gives employers a credible threat: accept

lower wages, or else we move abroad. Indeed, Furceri

et al. (2017) provide some evidence that the decline

in the labor share is related to the threat of relocating

production abroad. As a proxy for the potential

threat, they use layoff propensities of different

industries. They ﬁnd the effect of capital-account

liberalization on labor shares is particularly strong in

those sectors with a higher natural layoff rate. The

bargaining explanation is also consistent with the

ﬁnding in Jaumotte et al. (2013) that it is foreign

direct investment in particular that is associated

with the rise in inequality.

Differential mobility has other implications that

bear on distributional questions. Another argument

I made in Rodrik (1997) was that capital mobility

would increase volatility of labor earnings and, in

particular, shift the burden of economic shocks to

labor. This too follows from the differential

Figure 2 Capital mobility

and ﬁnancial crises. Source:

Reinhart & Rogoff (2009).

Populism and the economics of globalization Dani Rodrik

Journal of International Business Policy

mobility of labor and capital across borders. The

factor that is stuck within borders has to absorb the

costs of idiosyncratic shocks. Subsequent evidence

on the volatility of labor-market outcomes has been

largely consistent with this conjecture (Scheve &

Slaughter, 2002; OECD, 2007; Buch & Pierdzioc,

2014). Workers with the lowest skills and qualiﬁ-

cation, those least able to move across borders, are

typically the most affected by this risk shifting.

Another type of shift relates to taxation. As

capital becomes globally mobile, it becomes harder

to tax. Governments increasingly have to fund

themselves by taxing things that are less footloose –

consumption or labor. Indeed, corporate tax rates

have come down sharply in virtually all advanced

economies since the late 1980s, sometimes by half

or more. Meanwhile, the tax burden on wages

(social security charges, etc.) has remained roughly

constant, and value added tax (VAT) rates have

generally increased.

It is not implausible to think

that these global trends are related to globalization

and its effects on tax competition.

There is some systematic empirical evidence that

provides support for this hypothesis. In a panel

study of fourteen OECD countries for 1967–1996,

Bretschger & Hettich (2002) ﬁnd that proxies for

openness of capital and trade accounts are related

negatively to effective corporate tax rates and

positively to labor taxes. Similar results are reported

in a more recent study with data through 2007

(Onaran & Boesch, 2014). In a study covering the

1981–2001 period, Garretsen & Peeters (2007) ﬁnd

that foreign direct investment levels (instrumented

by an index of capital-account restrictions) exert

downward pressure on effective tax rates on capital,

but that there is considerable heterogeneity

depending on country characteristics (e.g., size,

neighborhood, density). Devereux, Lockwood, &

Redoano (2008) explicitly model governments’

selection of both the corporate tax rate and the

corporate tax base in a game theoretic framework.

They ﬁnd that strategic, downward pressures oper-

ate only in countries with open capital accounts,

and conclude ‘‘the reductions in equilibrium tax

Figure 3 The effects of capital-account liberalization on labor and top income shares (percent). Source: Furceri et al. (2017).

Populism and the economics of globalization Dani Rodrik

Journal of International Business Policy

rates can be explained almost entirely by more

intense competition generated by the relaxation of

capital controls.’’

In view of this ambiguous record of ﬁnancial

globalization with respect to efﬁciency and equity,

to put it mildly, it is curious that economists did

not take a stronger stand against the cumulative

removal of restrictions on capital ﬂows. Many

prominent economists supported ﬁnancial global-

ization, believing that the promise would eventu-

ally outweigh the risks.

That in turn strengthened

the hand of policy makers in the US, Europe,

emerging markets, and multilateral institutions like

the IMF and the OECD who pushed for ﬁnancial

openness.

In his 2015 presidential address to the

American Finance Association, Zingales (2015)

would complain that economists had not distin-

guished themselves in the years leading to the

global ﬁnancial crisis. ‘‘We should acknowledge

that our view of the beneﬁts of ﬁnance is inﬂated,’’

he wrote. Economists’ attitudes toward ﬁnancial

globalization exempliﬁed this attitude.

THE POLITICAL ECONOMY OF THE BACKLASH

Globalization had a big upside. It greatly expanded

opportunities for exporters, multinational compa-

nies, investors, and international banks, as well the

managerial and professional classes who could take

advantage of larger markets. It helped some poor

countries – China in particular – rapidly transform

farmers into workers in manufacturing operations

for export markets, thereby spurring growth and

reducing poverty. But the decline in global inequal-

ity was accompanied by an increase in domestic

inequality and cleavages. Globalization drove mul-

tiple, partially overlapping wedges in society:

between capital and labor, skilled and unskilled

workers, employers and employees, globally mobile

professionals and local producers, industries/re-

gions with comparative advantage and those with-

out, cities and the countryside, cosmopolitans

versus communitarians, elites and ordinary people.

It left many countries ravaged by ﬁnancial crises

and their aftermath of austerity.

Globalization was hardly the only shock which

gutted established social contracts. By all accounts,

automation and new digital technologies played a

quantitatively greater role in de-industrialization

and in spatial and income inequalities. But global-

ization became tainted with a stigma of unfairness

that technology evaded. People thought they were

losing ground not because they had taken an

unkind draw from the lottery of market competi-

tion, but because the rules were unfair and others –

ﬁnanciers, large corporations, foreigners – were

taking advantage of a rigged playing ﬁeld.

Many of these consequences were predictable and

are not a surprise. The same can be said about the

political backlash as well. A number of empirical

papers have linked the rise of populist movements –

Trump and the right-wing Republicans in the US,

Brexit in Britain, far-right groups in Europe – to

forces associated with globalization, such as the

China trade shock, rising import penetration levels,

de-industrialization, and immigration.

Analyzing electoral results across US congres-

sional districts, Autor et al. (2016a, b) have shown

that the China trade shock aggravated political

polarization: districts affected by the shock moved

further to the right or the left, depending which

way they were leaning in the ﬁrst place. Elected

Republicans became more conservative, while

elected Democrats became more liberal. For Britain,

Becker et al. (2016) ﬁnd that austerity and immi-

gration impacts both played a role in increasing the

Brexit vote, in addition to demographic variables

and industrial composition. Also analyzing Brexit,

Colantone & Stanig (2016) ﬁnd a much more direct

role for globalization. Using an Autor et al. (2013)-

type China trade shock variable, they show regions

with larger import penetration from China had a

higher Leave vote share. They also corroborate this

ﬁnding with individual-level data from the British

Election Survey that shows individuals in regions

more affected by the import shock were more likely

to vote for Leave, conditional on education and

other characteristics.

A second paper by Colantone & Stanig (2017)

undertakes a similar analysis for 15 European

countries over the 1988–2007 period. It ﬁnds that

the China trade shock played a statistically (and

quantitatively) signiﬁcant role across regions and at

the individual level. A larger import shock is

associated with support for nationalist parties and

a shift toward radical right-wing parties. Finally

Guiso, Herrera, Morelli, & Sonno (2017) look at

European survey data on individual voting behav-

ior and ﬁnd an important role for economic

insecurity – including exposure to competition

from imports and immigrants – in driving populist

parties’ growth. The same variables also affect voter

turnout: individuals who experience greater eco-

nomic insecurity are also less likely to show up at

the polls. As Guiso et al. (2017) indicate, the latter

result suggests that studies that focus on vote shares

Populism and the economics of globalization Dani Rodrik

Journal of International Business Policy

alone underestimate the importance of these eco-

nomic drivers, including globalization shocks.

A question that has attracted little interest to date

is why the backlash has taken the particular form it

has in different countries. Most (but not all)

populist movements in the current wave are of

the right-wing variety. These emphasize a cultural

cleavage, the national, ethnic, religious, or cultural

identity of the ‘‘people’’ against outside groups who

allegedly pose a threat to the popular will. In the

US, Donald Trump has demonized at various times

the Mexicans, Chinese, and Muslims. In Europe,

right-wing populists portray Muslim immigrants,

minority groups (gypsies or Jews), and the faceless

bureaucrats of Brussels as the ‘‘other.’’

An alternative variety of populism revolves

around a largely economic cleavage, the wealthy

groups who control the economy and deﬁne its

rules versus the lower income groups without

access to power. The original American populism

of the late nineteenth century was of this variety,

focusing its opposition on the railroad barons and

the Northeastern ﬁnancial elite. Bernie Sanders’

presidential campaign in 2016 took a similar form.

In Europe, there are a few left-wing populist

movements, of which Greece’s Syriza and Spain’s

Podemos are the best known. In Latin America, by

contrast, populism has long taken mostly a left-

wing form.

In Figure 4, I provide some systematic evidence

on the dynamics of support for populist parties

around the world since the 1960s. The ﬁgure shows

the aggregate vote shares of populist parties in

countries with at least one populist party. I distin-

guish between left-wing and right-wing populists

and between Europe and Latin America. (The US

presidential election of 2016 is not included.) The

Appendix discusses data sources and parties/coun-

tries covered.

What jumps out of Figure 4 is the sharp contrast

between the patterns of populism in Europe and

Latin America. In Europe, the rise of populism is

very recent and swift – from below 5 percent of the

vote in the late 1980s to more than 20 percent by

2011–2015. Moreover, this increase is driven exclu-

sively by right-wing parties. The left-wing populist

vote share remains throughout well below 5 per-

cent of the aggregate electorate in the countries

included. By contrast, left-wing populism has

always been strong in Latin America, with vote

totals between 15 and 30 percent. It also has

experienced a recent, if less marked, rise. Right-

wing populism has remained at very low levels in

Latin America.

What explains the predominance of right-wing

populism in Europe today, compared to the pre-

dominance of its left-wing variant in Latin Amer-

ica? To shed some light on this question, it helps to

think of the rise of populism as the product of both

demand- and supply-side factors at work.

On the demand side, the distributional and other

economic fault lines created or deepened by global-

ization generate potential public support for move-

ments that position themselves outside the political

mainstream and oppose established rules of the

game. But the economic anxiety, discontent, loss of

legitimacy, fairness concerns that are generated as a

byproduct of globalization rarely come with obvious

solutions or policy perspectives. They tend to be

inchoate and need to be channeled in a particular

programmatic direction through narratives that

provide meaning and explanation to the groups in

question. That is where the supply side of politics

comes in. Populist movements supply the narratives

required for political mobilization around common

concerns. They present a story that is meant to

resonate with their base, the demand side: here is

what is happening, this is why, and these are the

people who are doing it to you.

In Mukand & Rodrik (2017), we provide a model

where political conﬂict can revolve around differ-

ent axes. There are three different groups in society:

the elite, the majority, and the minority. The elite

are separated from the rest of society by their

wealth. The minority is separated by particular

identity markers (ethnicity, religion, immigrant

status). Hence there are two cleavages: an ethno-

national/cultural cleavage and an income/social

class cleavage. These cleavages can be orthogonal or

overlapping, producing different patterns of alli-

ances and political outcomes.

With some simpliﬁcation, we can say that pop-

ulist politicians mobilize support by exploiting one

or the other of these two cleavages. The ‘‘enemies of

the people’’ are different in each case. Populist who

emphasize the identity cleavage target foreigners or

minorities, and this produces right-wing populism.

Those who emphasize the income cleavage target

the wealthy and large corporations, producing left-

wing populism.

It is reasonable to suppose that the relative ease

with which one or the other of these cleavages can

be targeted depends on their salience in the every-

day experience of voters. In particular, it may be

Populism and the economics of globalization Dani Rodrik

Journal of International Business Policy

easier to mobilize along the ethno-national/cul-

tural cleavage when society is experiencing an

inﬂux of immigrants and refugees with dissimilar

cultural and religious identities. Then economic

anxiety can be channeled into opposition to these

groups. Immigrants and refugees can be presented

as competing for jobs, making demands on public

services, and reducing public resources available for

natives. Indeed, a major source of support for far-

right parties in Europe has been the fear that

immigration will erode welfare state beneﬁts, a fear

that is heightened in countries experiencing aus-

terity and recession (see for example Hatton, 2016).

Cavaille & Ferwerda (2017) ﬁnd that support for

right-wing populist parties is very responsive to

perceived competition with immigrants for in-kind

beneﬁts, in their case public housing.

An important implication of this reasoning is

that even when the underlying shock is fundamen-

tally economic the political manifestations can be

cultural and nativist. What may look like a racist or

xenophobic backlash may have its roots in eco-

nomic anxieties and dislocations.

The supply side

of politics – the narrative on offer – matters a great

deal. This is a point that is often overlooked in

current diagnoses. For example, it is not easy to

know whether Trump’s victory represents an eco-

nomic or cultural phenomenon without disentan-

gling the demand and supply sides – the underlying

grievances, on the one hand, and his narrative, on

the other.

What about Latin America? The reason that

populism took a divergent path in Latin America

may be related to the fact that the salient shocks

associated with globalization took different forms

there. Latin Americans who were affected nega-

tively by globalization experienced it not as immi-

gration or rule by Brussels/Frankfurt, but as rapid

trade opening, ﬁnancial crises, IMF programs, and

entry by foreign corporations in sensitive domestic

sectors such as mining or public utilities. The anger

to be mobilized was against these forces and the

domestic groups that supported them. This lent

itself to left-wing (economic) populism rather than

right-wing (cultural) populism.

The European exceptions to right-wing populism

provide further support to this argument. The two

European countries that grew substantial left-wing

Figure 4 Contrasting patterns of populism in Europe and Latin America. Notes: See Appendix for sources and methods.

Populism and the economics of globalization Dani Rodrik

Journal of International Business Policy

populist movements – Greece and Spain – bear a

certain similarity to Latin America. They were

major recipients of capital inﬂows under the Euro-

pean model of ﬁnancial globalization, the euro.

Once the sudden stop took place, their economies

went into a tailspin and unemployment skyrock-

eted. The shock was then intensiﬁed by the pres-

ence of a common currency and austerity policies

imposed from the outside – a troika made up of the

IMF, the European Central Bank, and the European

Commission. Although all countries in Europe

were affected by the euro crisis, Greece and Spain

were among the most adversely hit. Greece has yet

to recover, and unemployment remains very high

in both countries. All this is reminiscent of Latin

American boom-and-bust cycles, going back at least

to the 1970s. So it is not surprising that the

ﬁnancial crisis and its aftermath in Spain and

Greece provided fertile ground for left-wing pop-

ulists, for similar reasons.

The relative weakness of cultural/religious cleav-

ages to be exploited may also play a part in favoring

left-wing over right-wing populist movements. In

Latin America, the bulk of immigration has been

from other Latin American countries or from

culturally similar European countries. Within Eur-

ope, Spain and Greece once again provide instruc-

tive counter-examples. Compare the immigration

experience of Spain with that of France, for exam-

ple (Table 2). Even though Spain has a somewhat

larger migrant stock in relation to its population,

the majority of Spain’s immigrants come from

either Latin America or from advanced European

countries.

In France, by contrast, the largest share

(more than 40 percent) of migrants are from

Moslem countries (Algeria, Morocco, Tunisia, Tur-

key) and an additional 10 percent come from Sub-

Saharan Africa. A right-wing populist party (i.e., the

National Front) has much more fertile ground in

France than in Spain.

Table 1 Distributive and efﬁciency consequences of trade liberalization: illustrative calculations

Initial tariff being removed (%) Change in low-skill wages Increase in real income of economy Absolute value of ratio (A)/(B)

(A,%) (B,%)

40 -19.44 4.00 4.9

30 -15.22 2.25 6.8

20 -10.61 1.00 10.6

10 -5.56 0.25 22.2

5 -2.85 0.06 45.5

3 -1.72 0.02 76.6

Notes: Column (B) is computed using the standard formula for the gains from trade (e.g., Feenstra, 2016: 220), assuming an import–GDP ratio of 25%

and an import demand elasticity of -2. Column (A) is generated using a model with two factors (low- and high-skilled labor) and two goods with

mobile factors, assuming the import-competing sector is low-skill-intensive. The cost shares of low- and high-skill labor in the import-competing sector

are taken to be 0.80 (denoted h

) and 0.20 (h

), respectively. The factor cost shares in the exportable sector are symmetric – 0.20 (h

) and 0.80 (h

). To

compute the change in real wages (

), I assume low-skilled workers spend 75 percent of their budget on the importable and 25 percent on the

exportable. The corresponding derivation yields

¼ h

 h

1

0:75



p; where

p is the percent change in the relative price of the

importable implied by the tariff reduction.

Table 2 Differences in immigrations source countries: France versus Spain

Source country (only top 25 source

countries included)

Host country

France Spain

Share of migrant

stock (%)

Share of home

population (%)

Share of migrant

stock (%)

Share of home

population (%)

Predominantly Moslem 41 5 13 2

Sub-Saharan Africa 8 1 0 0

Other developing 4 0 3 0

Eastern Europe (incl. Russia) 3 0 18 3

Latin America 2 0 33 5

Developed Europe 24 3 17 2

Total of included countries 82 9 83 12

Source: World Bank bilateral migration matrix, 2013.

Populism and the economics of globalization Dani Rodrik

Journal of International Business Policy

The US presents a mixed case, combining char-

acteristics of both of these paths. Unlike Europe

which had opened up to trade and reached a

political settlement supporting it long ago – exten-

sive safety nets in exchange for trade openness –

the US experienced increased exposure to imports

comparatively recently. And it did so without

systematic compensation. Therefore, imports (espe-

cially from China) and trade agreements (with

Mexico, Asian countries) were politically salient

issues, around which large number of voters could

be mobilized. The ﬁnancial crisis and the differing

fates of large banks versus low-income homeowners

– one bailed out, the other not – engendered anger

at the ﬁnancial elites. At the same time, immigra-

tion from Mexico, the threat of radical Muslim

terrorism, and lingering racial divides were ripe for

political manipulation. In other words, the US

presented ample ground for both types of cleavage.

Correspondingly, the 2016 presidential elections

were contested by major populist movements on

both the left and the right, led by Bernie Sanders

and Donald Trump, respectively.

CONCLUDING REMARKS

One conclusion from the preceding discussion is

that the simple economics of globalization is not

particularly auspicious with respect to its political

sustainability. This is especially true of the

advanced phases of globalization – what I have

called elsewhere ‘‘hyperglobalization’’ (Rodrik,

2011) – in which the ratio of political/distributive

costs to net economic gains is particularly unfavor-

able. Historically, the uniﬁcation of national mar-

kets has required an unequivocal political project

led by a strong central executive. Nothing compa-

rable exists globally, and the European experience

provides ample reason to be skeptical that some-

thing like that can be achieved even regionally. In a

world divided politically, markets face strong cen-

trifugal forces as well.

The global economic arrangements of the imme-

diate postwar era were built around John Maynard

Keynes’ insight that sustaining a world economy

reasonably hospitable to international trade and

investment would require carving up space for

domestic macroeconomic management. For Key-

nes, this meant capital controls in particular, which

he viewed not as a temporary expedient but as a

permanent feature of the international economic

order. The same principle was followed in other

domains as well. The GATT regime entailed a thin

model of trade integration, not reaching beyond

direct border barriers or trade in manufactured

goods among advanced economies. It left plenty of

room for countries to design their own regulations

and industrial policies – and indeed protect ‘‘sen-

sitive’’ sectors (such as agriculture or garments).

The resulting system – variably called the Bretton

Woods compromise or embedded liberalism

– was

a great success. It fostered a large increase in global

trade and investment and saw rapid economic

development in both the advanced and developing

economies. Perhaps it was too successful for its own

good. By the late 1980s, policy makers and

economists thought they could make it work even

better by pushing for deeper economic integration.

Trade agreements became more ambitious and

reached beyond the border into domestic regula-

tions. The removal of restrictions on capital mobil-

ity became the norm rather than the exception. In

the process, the ‘‘embedding’’ or ‘‘compromise’’ that

had made the earlier regime such a success was

overlooked.

The rise of populism forces a necessary reality

check. Today the big challenge facing policy mak-

ers is to rebalance globalization so as to maintain a

reasonably open world economy while curbing its

excesses. We need a rebalancing in three areas in

particular: from capital and business to labor and

the rest of society, from global governance to

national governance, and from areas where overall

economic gains are small to where they are large.

The beneﬁts of globalization are distributed

unevenly because our current model of globaliza-

tion is built on a fundamental and corrosive

asymmetry. Our trade agreements and global reg-

ulations are designed largely with the needs of

capital in mind. Trade agreements are driven

overwhelmingly by a business-led agenda. The

implicit economic model is one of trickle-down:

make investors happy and the beneﬁts will even-

tually ﬂow down to the rest of society. The interests

of labor – good pay, high labor standards, employ-

ment security, voice in the workplace, bargaining

rights – get little lip service. To move forward, labor

must be given an equal say in setting the rules of

globalization. In practical terms, this requires

reconsidering which multilateral institutions set

the agenda of the global conversation and who sits

at the bargaining table when trade agreements are

negotiated.

With respect to rebalancing governance, we

should understand that the world economy does

not really need global governance to be managed

Populism and the economics of globalization Dani Rodrik

Journal of International Business Policy

appropriately. Most failures in the world economy

are rooted in failures of domestic governance. In

particular, restrictive trade policies are due either to

complications that override the standard Ricardian

argument for the gains from trade, or to domestic

policy foul-ups (such as the neglect to redistribute

the gains appropriately).

Society may value environmental amenities or

distributional outcomes that would be undermined

by free trade and it may not be possible to address

those objectives through other policy instruments.

A developing country may beneﬁt from sheltering

infant industries from foreign competition before

they have developed some technological capabili-

ties. In such instances, the rejection of free trade

policies is not a problem, from either domestic or

international perspectives. Global institutions have

no business telling these countries to open up.

When the domestic policy process fails, there is a

genuine problem. But the problem then lies in

domestic politics – the excessive political inﬂuence

perhaps of certain rent-seeking groups – and not in

the absence of proper global rules. Empowering

global governance in such circumstances may or

may not work. Global rules sometimes can act as

counterweight to protectionist interests. But it is a

fortiori equally likely that the global rules will be

written and administered by the very special inter-

ests that dominate domestic policy making as well.

Think of the role of big banks in setting global

capital standards or pharmaceutical companies in

writing global patent rules.

None of this implies that there is no role at all for

global governance. But in rebalancing globalization

toward national governance, we must understand

that the best contribution global arrangements can

make is to make the nation state work better, not to

weaken or constrain it. Correspondingly, the

appropriate role for global institutions is to

enhance key democratic norms of representation,

participation, deliberation, rule of law, and trans-

parency – without prejudging policy outcomes or

requiring harmonization.

Finally, our approach to globalization must focus

on areas where the net gains are large. Today the

world economy is as open as it has ever been, and

the most important challenge it faces is not lack of

openness but lack of legitimacy. The traditional

approach to trade deal-making that concentrates

on exchanging market access is no longer appro-

priate. The rules that need to be developed are

those that emphasize fairness, address concerns of

social dumping, and enhance policy space in both

developed and developing nations.

ACKNOWLEDGEMENTS

I am grateful to Michael-David Mangini and Christine

Gosioco for research assistance; the Weatherhead

Center for International Affairs for research suppo rt;

and Pol Antra

s, Bart Bonikowski, Jeff Frankel, Jeffry

Frieden, Peter Hall, and Roberto Mangabeira Unger for

helpful comments and conversations. Editorial sug-

gestions from Sarianna Lundan and an anonymous

referee have improved the paper.

NOTES

Useful accounts, with varying deﬁnitions of pop-

ulism, are provided in Kazin (1998), Judis (2016),

ller (2016 ), Mudde & Kaltwasser (2017).

Judis (2016) also develops this distinction in his

recent book.

This framework can accommodate decreasing

returns readily, by introducin g ﬁctional factors of

production that soak up any surplus left over –

effectively converting the technology to constant

returns. Increasing returns are more problematic, as

they would not be compatible with perfect

competition.

In earlier work, I have called this the political cost–

beneﬁt ratio (PCBR) of liberalization and presented

some illustrative calculations (see Rodrik, 1994).

It is possible for the net efﬁciency gains of trade to

be magniﬁed if globalization generates dynamic (or

growth) effects, in addition to the static effects that

conventional trade theory emphasizes. Such growth

effects are often thought to make the redistributive

consequences less signiﬁcant. But this is not quite right

since models of endogen ous growth typically have

ambiguous implications for the growth effects of trade

liberalization. For example, the growth effects can

easily turn negative when a country has (static)

comparative disadvan tage in the dynamic industries.

In other words, dynamic effects do not as a rule

magnify the gains from trade. They simply allow for

larger effects in absolute terms – negative as well as

positive.

The ﬁnding that wage effects differed across local-

ities and industries is of course inconsistent with the

Stolper–Samuelson assumption that labor is fully

mobile. It suggests additional margins of redistribution

that Stolper–Samuelson overlooks. Hakobyan &

Populism and the economics of globalization Dani Rodrik

Journal of International Business Policy

McLaren (2016) do document some outmigration

from most affected commun ities, but apparently the

magnitudes were not large enough to eliminate wage

effects.

Moreover, fully half of the miniscule 0.08% gain for

US is not a n efﬁciency gain, but actually a beneﬁt due

to terms-of-trad e improvement. That is, Caliendo and

Parro estimate that the world prices of what the US

imports fell relative to what it exports. These are not

efﬁciency gains, but income transfers from other

countries (here principally Mexico and Canada). These

are gains that came at the expense of other countries.

Labor-market institutions may play an important

intermediating role in the way trade shocks affect

wages. Using French ﬁrm-level data, Carluccio, Fou-

re, & Gautier (2015) ﬁnd that offshoring has adverse

effects on blue-collar wages only in ﬁrms without

collective bargaining agreements. In ﬁrms with fre-

quent collective bargaining, offshoring does not seem

to affect low-skill wages negatively.

A 2012 study found the net effects of participation

in TAA was actually negative for affect ed workers:

unemployment and other beneﬁts did not compen-

sate for lower earning overall, compared with non-

participants (D’Amico & Schochet, 2012).

Hicks & Devaraj (2015, 2017) estimate the 13% of

the jobs lost in manufacturing during 2000–2010 were

lost due to imports and import-substitution, with the

rest due to domestic productivity growth. Acemoglu,

Autor, Dorn, Hanson, & Price (2016) attribute 10 per-

cent of the job losses over 2000– 2007 to the China

import shock.

The use of machines to replace labor was equally

controversial in an earlier era, during the early part of

the nineteenth century, when the Luddite movement

attacked and destroyed textile machinery. Since then,

the introduction of new technology has become a

regular aspect of industrial societies and does not

appear to raise the kind of fairness concerns I discuss in

the following paragraph s.

The term ‘‘blocked exchange’’ comes from Walzer

(1983) and refers to things that cannot be bought or

sold because of moral stigma or legal strictures. See

discussion and application to international trade in

Rodrik (1997: 35–36).

See Clark & D’Ambrosio (2015) for a survey of the

literature on attitudes to income inequality, including

fairness concerns.

The previous thought experiment makes clear

distinction between trade ﬂows that suffer from this

problem [scenario (3)] and those that don’t [scenario

(2)]. Similarly, low wages that are due to low produc-

tivity in a trade partner is different from low wages due

to the abuse of worker rights (the absence of collective

bargaining, freedom of association, etc.). Both may

generate distributional implications at home – but

there is a problem of unfair trade only in the second

case.

Survey evidence shows that the fairness concern

that motivates people is distinct from the self-inter-

ested ‘‘protectionist’’ aim that economists worry

about. There is widespread support for safeguarding

environmental and labor standards in trade, and the

supporters of ‘‘fair trade’’ overlap, but not are identical

to those who militate against job losses (Ehrlich,

2010).

There are many other stories of how ﬁnancial

globalization interacts, adversely, with domestic mar-

ket imperfections. In Rodrik & Subramanian (2009),

for example, we explain the poor growth experience

of capital-receiving countries by arguing that capital

inﬂows aggravates appropriability problems in trad-

able sectors by appreciating the real exchange rate

and reducing proﬁtability of investment in tradables.

Two earlier papers that document a negative

relationship between measures of capital-acco unt

liberalization and the labor share of income are

Harrison (2005) and Jadayev (2007).

Using a ‘‘kitchen-sink’’ regression that includes a

large number of potential determinants, Furceri et al.

(2017) also ﬁnd that ﬁnancial openness has an effect

on inequality that is commensurate with the effects of

trade openness and technology.

As Carluccio, Fouge

re, & Gautier (2016) argue,

the wage bargaining regime itself may be endogenous

to globalization shocks. Using administrative ﬁrm-level

data for French ﬁrms, those authors ﬁnd that positive

exogenous export shocks increase the incidence of

collective wage agreements, while offshoring shocks

have no signiﬁcant effects. The authors argue the latter

ﬁnding may be due to two co nﬂicting effects: off-

shoring increases ﬁrm-level productivity and proﬁts,

providing greater incentive for rent-sharing with

unions, but it also reduced workers’ bargaining power.

The information on tax rates here comes from the

OECD tax database (see http://www.oecd.org/tax/

tax-policy/tax-database.htm#taxBurden).

A number of other studies have found insigniﬁcant

or contradictory results to those reported here. A

meta-study that reviews two decades of empirical

analysis concludes that ‘‘study characteristics related to

[different] globalization measures give rise to totally

different ﬁndings concerning the relationship between

globalization and capital tax rates. More precisely,

studies employing: (i) international trade as percent of

GDP and (ii) the globalization index developed by

Populism and the economics of globalization Dani Rodrik

Journal of International Business Policy

Quinn (1997) are more likely to report a negative

impact of international market integration on capital

taxation, whereas studies employing the KOF index of

globalization developed by Dreher (2006) are more

likely to report a pos itive effect of globalization on

capital tax rates’’ (Adam, Kammas, & Lagou, 2013).

Stanley Fischer advanced the case in favor of

ﬁnancial globalization in the late 1990s and advocated

an amendment to the IMF’s articles the purpose of

which ‘‘would be to enable the Fund to promote the

orderly liberalization of capital movements’’ (Fischer,

1997). Despite the risks of opening up to capital ﬂows,

which he enumerated, he argued that these were

more than offset by the potential beneﬁ ts. On the

paucity of evidence on beneﬁcial effects, he argued

that the evidence in favor of capital account would

cumulate over time, just as with the evidence on the

beneﬁts of trade liberalization (Fischer, 2003: 14).

Rudiger Dornbusch declared capital controls ‘‘an idea

whose time is past’’ (Dornbu sch, 1998), having pre-

viously advocated ﬁnancial transactions taxes a short

time back (Dornbusch, 1996). Frederic Mishkin’s

(2006) book The Next Great Globalization presented

a strong argument in favor of ﬁnancial globalization,

making a case that the beneﬁts would materialize as

long as countries undertook the requisite complemen-

tary reforms.

See Abdelal (2007) for a fascinating account of the

politics of capital-account deregulation during the

1990s.

A similar question is posed by Aytac¸&O

nis¸

(2014), who compare the Kirchne rs’ left-wing pop-

ulism in Argentina to Erdogan’s right-wing populism in

Turkey. The authors link the contrastin g forms of

populist politics to the perceived causes of previous

economic crises, varying levels of dependence on

capital inﬂ ows, and the strength of organized labor.

The empirical case for cultural determinants of

populism is made by Inglehart & Norris (2016).

Gidron & Hall (2017) focus on perceived social status,

and changes therein, as a determinant of right-wing

populism.

Another instance in which the supply side of

politics may override the demand-side is presented in

the theoretical paper by Karakas & Mitra (2017). In

that paper , the outsider status of populist candidates

gives them an edge over establishment candidates

from either the right or the left, because the former are

perceived as more likely to disrupt the status quo.

Voters with left-wing preferences (e.g., traditionally

Democratic voters) may then vote for a right-wing

populist candidate (e.g., Donald Trump).

The classic book on Latin American economic

populism remains Dornbusch & Edwards (1991).

Sachs (1990) provides a parallel account of Latin

America’s populist economic policy cycle.