CHAPTER 1

Enduring Issues

T his book interprets a nearly global history of

social spending up through the year 2020, drawing lessons for the
years up to 2050. In the long history covered here, what most enriches the
record has been harvested within the last twenty years – twenty years of
experience, and twenty years of accelerating scholarship.1 The new
knowledge extends far beyond the usual tales of the North Atlantic
community, inviting new interpretive forays into Latin America, East
Asia, and the formerly communist Eastern Europe.2 While much of the
best evidence is recent, the underlying social issues have endured for
millennia.

ALWAYS NEEDED, JUST NOW ARRIVED

Human societies have always needed safety nets to catch those who end
up in need, whether by unlucky endowments, by past mistakes, or by the
arrival of hard times. The risks are not new. They have always been there.
Yet for most of human history, we have lacked the means, or the political
will, to prevent or cushion them.
With each wave of expansion in the economic base, humans did what
little they could to cut risk at the local level, yet serious risks remained.
Prehistoric fortunes were never stabilized by hunting and gathering. The
arrival of agriculture tens of thousands of years ago initially helped to
diversify humanity’s economic portfolio, yet brought new vulnerabilities
to weather and pests once it had expanded. More recently, as the spread
of commerce diversified our sources of supply and allowed a further
expansion of population, our risks were reduced a bit further, though

3
 OVERVIEW

they remained. Still more recently – in a mere blink of an eye lasting less
than three hundred years – industrialization and the ever-growing reli-
ance on skilled services once again cut our risks somewhat yet have not
eliminated them. Our world is risky, but not increasingly so.
A common error in the way we view history, and the way it is taught in
school, is to believe that an unprecedented economic insecurity was ush-
ered in by the Industrial Revolution and by the rise of a new greedy market
mentality around 1750. Marx and Engels said so. The gentler reformists of
the Fabian Society agreed that the age of dark satanic mills brought new
urgency to finding ways of providing social insurance and social assistance
to the needy. Karl Polanyi’s The Great Transformation (1944) agreed. While
their landmark writings contained great insights, they were mistaken in
believing that the industrial era and the rise of a market mentality brought
a transformational rise of risk and a brand-new need for social security. The
risks were at least as great before the Industrial Revolution. As for the rise
of a market mentality, it did not happen. There was no modern dawn of
market exchange, nor of self-interest, nor of “greed,” because all of these
are at least as old as the human species itself.
The risk of mortality, even more than economic risk, has also been
dropping over the centuries. There has been a great convergence in
human life spans, thanks mainly to the elimination of death from child-
birth and infant mortality. In the 2020s, people will naturally share the
fear that our vulnerability to mortality shocks is greater than ever, as
witnessed by the coronavirus pandemic. Not so. Horrible as it is, the latest
pandemic will not match the introduction of smallpox and malaria into
the Western Hemisphere, which killed a majority of its population. The
Black Death of the fourteenth century killed perhaps a quarter or half of
the European population, which the coronavirus will not do. And
Chinese history recorded frequent epidemics stretching back at least
three thousand years. Life has always been filled with risks as least as
great as those we face today, underlining the point that our need for
safety nets was at least as great in the past.
Given that humanity’s exposure to risks seems eternal, it is puzzling
that societies have only very recently built effective safety nets for contain-
ing such risks.

4
 ENDURING ISSUES

Defining Some Terms

• Social spending in this book will usually refer to public (tax-

based) spending on:
• education,
• old-age support,
• incapacity and disability,
• health care and health insurance,
• family assistance,
• labor-market assistance (retraining, unemployment compensa-
tion, etc.), and
• housing.
In practice, the official measures of social spending only partially
cover tax breaks, sector subsidies, public infrastructure, etc., which
could arguably be included in the definition if there had been
sufficient data coverage of these transfers. The tax breaks and
sector subsidies excluded here (e.g. subsidies to the energy sector
or the agricultural sector) are often more pro-rich than the social
spending covered in this book.
Social spending here excludes private social spending aimed at the
same targets, even when the private spending is mandated by govern-
ment. To be sure, private social spending can make important con-
tributions toward a host of social goals. But the real controversies
center on the government’s increasingly dominant role, and this
book will stay focused accordingly.
• Safety nets are society’s many supports to keep people from
falling too low economically. They consist of both social assis-
tance expenditures for those whose needs may be life-long, and
social insurance expenditures for cushioning temporary falls.
In this book, the term “safety nets” is thus broader than just
social assistance for the poorest, or “welfare.” It covers all of
the egalitarian uplift provided by social spending, both short-

5
 OVERVIEW

(cont.)

run and long-run. Basically, safety nets are what social spend-
ing is for.
• Selfish generations: The book will use the convenient shorthand
Selfish Generations borrowed from David Thomson’s (1996) fine
book on how New Zealand governments had redistributed
resources between age groups.
The selfish is a shorthand for “advantaged by inter-generational
redistribution.” The advantaged “selfish” ones are not the whole
generation, but just those of its members who have political voice.
In many cases, the selfish generation’s poor were not helped. To
further clarify, a generation in this phrase and in Thomson’s book
refers to an age group, as in the common parlance about “the older
generation” or “the younger generation.” It does not strictly refer
to a birth cohort, as a demographer might prefer.

BASIC QUESTIONS

The recent global surge of government social spending, after millennia

without it, poses some natural questions, each of which is pursued in this
book.

(1) Why did social spending arrive so late in human history?

(2) Why has Northwest Europe always led the way?
(3) Did the task of providing safety nets to cushion us against life’s many
risks really have to fall to government, instead of to private charity
and the extended family?
(4) How does a large tax-based social budget affect our livelihoods over
our life spans?
(5) For any given size of the government social budget, which countries
have been spending it on the wrong things?
(6) What threatens social programs between now and mid-century?

6
 ENDURING ISSUES

To lighten the burden of carrying so many questions at once into the

coming chapters, I will offer two immediate spoiler alerts:

• The answer to the third question is yes. Yes, safety nets had to be
administered mainly by government. Private charity and family sup-
ports have never been up to the tasks of eliminating poverty or educat-
ing the whole population. The forces behind government social
programs came together only in the last two hundred years, with
almost all their advance coming in the last sixty years. Government
social spending has now become worldwide, absorbing around 10 per-
cent of world product.
• A partial answer to the fifth question: The tendency to get the mix
wrong is as great among the world’s low-spending governments as it is
among those who spend more. Thus, the fourth and fifth questions
can be studied separately.

A striking pattern will also emerge as to which countries got social

spending wrong, either at low levels or at high levels: The clearest
mistakes are myopic ones. The errors have tended to be errors of
“selfish generations,” in which those in office have deprived future
generations by appeasing those whose lobbying power is here and
now.

AS CONTROVERSIAL AS EVER

The fourth question, whether large “welfare-state” social budgets are

better or worse for incomes and wellbeing than smaller budgets, has
always been the main fight. The combatants in the debate over the size
of social budgets need introductions here, deferring the final verdict
until a preview in Chapter 2 and a fuller empirical update in Chapters 8
and 9.
To report on the battle between small-government free-market capit-
alism and the tax-based social spending, recently dubbed the war
between “cutthroat” and “cuddly” capitalism,3 let us begin by reviewing
the classic arguments over the economic effects of social spending versus
an imagined free-market alternative. The debate has raged for at least

7
 OVERVIEW

eight centuries. The words have changed, but the opposing positions
have not.

TRADITIONAL ARGUMENTS AGAINST GOVERNMENT SOCIAL

SPENDING. Each side of the debate sends a respectable signal along
with ideological noise. To pick up the signal worth hearing in the cri-
tiques of social spending, one must first mute out the noisy demonizing
images of welfare queens, welfare bums, and bureaucrats, and the sneers
about a nanny state.
For centuries now, the core argument against tax-based social spend-
ing, or all civilian government spending, has warned of perverse incen-
tives. The incentive argument has economic plausibility and deserves
careful testing. Basically, if the government taxes the incomes or wealth
of productive people and gives the money to less productive people in
need, the incentives work badly on both sides. The productive lose the
incentive to produce, or innovate, or take risks. The less productive are
rewarded for being in a bad state, and are likely to respond by staying
longer in that bad state and not producing.
The incentives critique has been leveled at all four main forms of social
spending – anti-poverty “welfare,” public pensions, public health, and public
education, in roughly that order of emphasis. The archetypal prediction,
since the twelfth century or earlier, is that welfare spending will kill the work
incentive. Public pensions are faulted for crowding out private savings, and
for killing the work incentive by inviting earlier retirement. Public subsidies
to health-care provision or health insurance may have subtler incentive
problems referred to as “moral hazard”: inviting people to adopt lifestyles
with greater health risk or to see the doctor too often, and inviting doctors
and hospitals to overcharge. The case against public education is the most
muted: Paying for public schools, but not for private schools, may make
parents accept lower-quality education and may lower the public schools’
incentives to become more efficient.
All of these respectable criticisms deserve careful testing.

TRADITIONAL ARGUMENTS FOR GOVERNMENT SOCIAL

SPENDING. The respectable signals in favor of social spending have
also had to contend with ideological noise from advocates of greater

8
 ENDURING ISSUES

public assistance. The noise includes demonizing images of greedy and

insensitive billionaires, robber barons, and corporations. “Privatization”
is used as a pejorative, defined as handing the people’s public assets over
to the fat cats.
The respectable signals strike notes that are replayed in today’s eco-
nomics textbooks. For basic welfare spending – that is, assistance to poor
families and unemployment compensation, the incentives critique is not
rejected altogether. Rather, it is rebutted by denying that poor people
lost work by choice, or that working more is a valid option without
government help.
Public pensions help to smooth consumption over the life cycle, and
can insure people against a bad consequence of something good: out-
living the wealth they had saved up for retirement. As for the familiar
criticism that people should self-save and purchase private old-age insur-
ance, proponents of public pensions counter that people, especially poor
people, often save little for the future because shorter-run needs look
larger to them, and they cannot borrow at reasonable rates of interest. As
for the argument that people could be forced to save in the form of
government-mandated paycheck deductions, we return to its complex-
ities in Chapter 13.
Public health care and public health insurance are defended largely
on grounds of positive externalities, or spillovers. For example, as
a taxpayer, I should help to pay for your vaccinations and good health,
lest you pass on infections and over-use my hospital’s emergency room.
Believers in tax-based health insurance deny that there is much moral
hazard, claiming that people do not take big health risks just because
taxpayers will cover some or all of their out-of-pocket costs. The alter-
native of leaving all health insurance to the private marketplace is
rejected on the grounds that voluntary private health insurance leads to
a “death spiral”: The insurers want to cover only those with lower health
risks than the premiums will cover; so the insurance will be bought only
by those who know they are likely to have health problems, not by
healthier individuals; this adverse selection threatens the insurers with
higher costs, which they then try to cover by charging still-higher pre-
miums; so even more healthy people drop the insurance; and the cycle
repeats, crippling the market for private health insurance.

9
 OVERVIEW

The argument for public education also rests on a belief in positive

externalities, in this case from the knowledge and civic responsibility that
schools are supposed to instill. The benefits from a child’s education are
not captured only by the child and the child’s family. All of society
benefits from the knowledge, therefore all of society should help pay
for it with taxes.

WHY THE CONTROVERSY ENDURES. If the debate is so ancient, why

does it persist? A partial excuse is that it takes a lot of effort to dig out the
facts, and to run convincing tests. Fortunately, an accelerating volume of
economic studies has been delivering the necessary effort, as summarized
in the chapters that follow.
More fundamentally, the conflict of self-interests is what generates
controversy forever. Any solid research finding can be challenged by
those whose self-interest it seems to threaten. Nonetheless, as has been
argued several years ago,

new facts can raise the level of the debate. They can arm all sides with an
awareness of how tax-based social spending would affect collective goods
that all profess to care about – social peace and the size of the economy.
The competitiveness of the intellectual marketplace, and of the political
marketplace in electoral democracies, allows new facts to exert pressure
toward these collective goods. At the very least, new facts can speed up
society’s rejection of bad arguments.4

Eventually, despite the noise, true signals do come through, and far-
sighted societies heed them.

10
 CHAPTER 2

Findings and Lessons

S o what do we now know about the role of social

spending in the economy and society? This chapter summarizes
ten main findings supported by global experience and fresh research in
the early twenty-first century. Documenting these findings yields three
clear policy lessons, and an international scorecard on the far-sighted-
ness or myopia of different countries’ approaches to social spending.

TEN MAIN FINDINGS

Finding #1. A country’s government social spending takes off only

after the country has both the fiscal capacity and the political will to
build safety nets. This is why the world had only negligible social
spending before 1800, and why it emerged first in Northwest
Europe after that.

The main reasons why social spending has spread all over the world is
the mirror image of what prevented social spending outside of Europe
before the last century: Having the government spend on the poor, the
sick, the elderly, and school children requires fiscal capacity plus political
pressure from below (Chapters 3 and 4).
Within the last sixty years, government social spending has grown to
absorb more than a quarter of GDP in over a dozen countries, mainly in
Europe. The demand for such tax-based programs was raised by

11
 OVERVIEW

a diversity of forces: the shocks of world wars, revolutions, hyperinfla-

tions, and the Great Depression; the continuing rise of political voice for
middle- and lower-income groups; rising trust in government; improve-
ments in medicine; and population aging (Chapters 5 and 6).

Finding #2. A major barrier to growth and equality has been the
refusal of those with power to devote taxes to universal education.
Government funding for mass education has always been the
main driver of adults’ years of schooling, building human skills
and productivity. The leading countries passed up chances to
capture these gains in the eighteenth century and part of the
nineteenth. Once tax-based schooling took off, it became more
equal. The Americans were leaders in the quantity of primary and
secondary schooling, but never leaders in its quality.

No country has ever delivered primary and secondary education to the

majority of its children without financing it mainly through taxes.
In the eighteenth century and part of the nineteenth, the leading
countries passed up chances to capture the gains from public schooling
of the masses. In Britain and the Netherlands, the payoffs from basic
schooling already promised high rates of return, privately and socially,
partly by interacting with the rise of commerce. Yet these two leaders
delayed for at least a century and a half, partly because of church-school
issues and more fundamentally because those with political voice resisted
paying taxes for educating the children of others.
Once the political will was mustered to launch universal primary
education, adults’ years of school not only grew, but became more
equal. So say the data on years of school worldwide since 1870. This
growth and leveling of schooling was, again, dominated by the rise of
tax-based public schools (Chapter 5).
Starting around 1850, the Americans were leaders in the quantity of
education, as measured by years of enrollment and of adults’ accumu-
lated education. Yet as best one can tell from indirect and circumstantial

12
 FINDINGS AND LESSONS

evidence, the United States never led the world in the quality of primary
and secondary schooling (Chapter 5).

Finding #3. Since around 1910, there has been a long mission shift
in social spending, toward support for the powerful and elderly, at
the expense of assisting the young and the poor. This mission shift
in social targeting has offset part of the pro-growth and pro-equality
effects of the overall expansion of safety-net spending.

Over the last hundred years, many countries’ social spending shifted
missions, toward public pensions for the non-poor and away from pro-
gressive policies like mass schooling and aid to the poor. This mission
shift has not been reversed since. It probably compromised both income
equality and income growth. The global mission shift toward public
pensions may have been due in large part to improvements in life
expectancy, which allowed longer life past work and contributed to
political “gray power.” The inference about gray power springs from
the fact that public pension spending rose even per elderly person, and
not just at the rate of population aging. Its per person generosity rose
faster than the rise in educational spending per child of school age.
The rising demand for government pensions was probably also linked,
in addition, to a quiet global change in the role of intra-family transfers.
Career and family developments may have raised public pressure for
more tax-based support of the elderly (Chapter 6).

Finding #4. Larger social-spending budgets have not produced any

net loss of GDP, or in skills, or in work. Without any such costs,
Europe’s welfare states have produced greater equality, cleaner
government, and even longer life. There are good economic expla-
nations for this “free lunch puzzle.”

So says the international evidence for any decade or combination of

decades back to 1880, before which there was little social spending at all.

13
 OVERVIEW

None of these decades, or any longer periods, showed any significantly

negative correlation between the share of GDP channeled into social
spending and the level (or growth) of GDP itself. The lack of any sig-
nificant negative correlation is all the more remarkable because shocks to
GDP should have biased the results in favor of seeing a negative relation-
ship (Chapter 8).
Behind this statistical outcome lie some sound reasons suggested by
recent history. One can find at least four positive features of the real-world
welfare-state bundle of policies, features that have cancelled any anti-
growth effects. One consists of those economies of scale in delivering
social insurance: the more universal the coverage, the smaller the admin-
istrative, or bureaucratic, costs of raising tax revenues and allocating
transfers. A second is that the large welfare states raise their tax revenues
through broad consumption taxes and sin taxes, the type of taxation that
conventional theory predicts will favor growth the most. Third, the wel-
fare-state policies of parental leave and pre-school child development
foster better human productivity for both the child and the career-inter-
rupting parent, usually the mother. Fourth, single-payer public health
insurance is more cost-efficient than voluntary private insurance
(Chapters 9 and 14).
The claim that greater social spending must somehow shrink the size
of the economic pie is in deep trouble, and has already retreated in the
slump of 2020.

Finding #5. Governments around the world have shifted toward

equalizing incomes since 1910. This tide has not been reversed, not
even by the conservative movements since 1980. The usual mea-
sures of fiscal redistribution understate the rise of progressivity, by
missing the delayed equalizing effects of public education
expenditures.

Over the last hundred years, government redistribution of income has

become more “progressive,” shifting income from richer households to
poorer ones in the Robin Hood tradition. Retreats toward the opposite

14
 FINDINGS AND LESSONS

“regressive” redistribution have been rare and limited. For all that has been
written about a shift of political sentiments and government policy away
from progressivity since the late 1970s, no such trend is clear yet in how
overall taxes or social spending are distributed. Among democratic welfare
states, the closest thing to a demonstrable reversal against Robin Hood is the
slight retreat in Sweden since the 1990s. Globally, the most dramatic swing
since the late 1970s has been Chile’s record-setting return toward progres-
sivity after the regressivity of Pinochet.
Finding that redistribution has continued to march slowly toward
progressivity carries a strong implication for our interpretation of the
rise in income inequality since the 1970s, so firmly established by the
World Top Incomes Project and by Thomas Piketty (2014). That rise may
owe nothing to a net shift in government redistribution toward the rich,
despite the lowering of top tax rates. If so, it is all the more important to
explore what non-fiscal forces have widened gaps in market incomes
around the world (Chapter 10).

Finding #6. A short-run threat to the future of social spending

looms from waves of refugees and the nationalist reactions to
these waves. Of four policy options facing rich destination coun-
tries, politics will probably favor the option that protects national
goals, including social programs, at the expense of international
growth and equality. The likely favorite policy, Chapter 11’s
“Option 4,” will add more cherry-picking restrictions favoring
skilled immigrants.

An open humanitarian acceptance of refugees (“Option 1”) would

serve global growth and global equality. It could also serve domestic
growth, and is fiscally sustainable. Yet it is politically unsustainable,
because it threatens domestic equality, domestic political harmony, and
social programs. The backlash against this option will be around for
a while. The opposite option, “Option 2,” would slam the immigration
door shut, catering to nativist reaction, but at a cost in terms of economic
growth. The recently discussed option of “welfare chauvinism” (“Option

15
 OVERVIEW

3”), admitting all sorts of immigrants but denying them access to social
assistance and social insurance is unsustainable. The cherry-picking
Option 4 seems to win by default (Chapter 11).

Finding #7. The most durable threat to the future of social spend-
ing and the welfare state is posed by the upward march of senior life
spans. Something has to give, for financial sustainability. Five coun-
tries are clearly keeping their total public pension costs under
control, while protecting their basic anti-poverty pensions, but at
least seven others are not.

The aging of human populations will continue, at least until 2050. By

then, the number of persons over 65 for each person of prime working
age (18–64) will have doubled.
In the face of this aging, balancing pension budgets requires that
benefits per year must grow more slowly than wages and salaries per
member of the working-age population. Something has to give. Indeed,
the same adjustment to longevity would be called for even if individuals
self-financed all of their old age privately. If the share of your adult life
spent earning income drops, with each generation living more of adult-
hood beyond work than did its ancestors, then your lifelong annual
consumption must drop relative to your annual earnings. The problem
faces us all, with or without government pensions (Chapter 12).

Finding #8. To be sustainable, average public pension spending per

elderly person should not, and need not, rise as a share of peak
GDP. The annual public support per elderly person can rise with
average income growth, but should fall behind income growth as
the population ages.

The logic governing how public pension budgets must behave in the
long run produces this necessary rule for sustainability. Some countries
have already built something like it into official policy rules. The notional

16
 FINDINGS AND LESSONS

defined contribution (NDC) system fixes the share of GDP (or earnings)
to be paid out in yearly pension benefits. The simple rule stated in
Finding #8 indexes pensions not to the latest GDP or earnings, but to
the real value of peak GDP or earnings. Pension institutions need to retain
this safeguard, to avoid having yearly pensions sink as fast as GDP in
a short-run recession or depression like the one triggered by the corona-
virus pandemic of 2020 (Chapters 12–14).1

Finding #9. The “second pillar” of government social insurance, the

job-based kind that intervenes in the workplace, has shown various
flaws. In many countries, it has redistributed in favor of top occupa-
tions. The famous second-pillar compulsory paycheck-deduction sys-
tems of Chile since 1980 and Singapore since 1965, while promoting
financial stability, do not function as their proponents claim. Far from
being “privatized,” they have become slightly regressive instruments
of state taxation and control. Universal first-pillar provision better
protects transparency and equity.

The practice of linking social security to one’s job or occupation

should be cut back (Chapters 12 and 13).
Why should social security have anything to do with one’s work-
place? The two are mismatched, aside from the narrow cases of work-
place accidents and experience-related unemployment compensation,
where the job itself is the focus. If the benefits of social insurance (e.g.
pensions, health) are shared by society as a whole and the covered
individuals, the usual principles of public goods say that the insurance
costs should be shared between the same two parties, i.e. society as
a whole and the covered individuals, job or no job.
A further violation of economic principles arises as soon as linking the
insurance to one’s current job imposes a net cost on the employment
relationship, a cost shared by employers and employees. Why should the
employment of labor bear such a cost, when employing capital bears no
such insurance cost?

17
 OVERVIEW

Worse, many countries have had top-heavy pension practices, in which

the general population bears a net fiscal burden of providing generous
pensions to those in the well-paid formal-sector occupational groups,
including civil servants, the military, and the courts. This book spotlights
such regressive practices in a global South, here meaning Latin America,
Africa, the Middle East, and South Asia (Chapters 7 and 12).
Chapter 13 reinterprets two mandatory-payroll-deduction pension
systems, those of Chile since 1980 and Singapore since 1965. Both
became famous as “privatized” pension systems, yet both are systems of
central government taxation. Thinking of them as tax systems fits both
common intuition and the official Organization for Economic
Cooperation and Development (OECD) practice of defining mandatory
payroll deductions as taxes.

Finding #10. The younger the person, the greater the long-run
social return on society’s investments in them, both on the average
and at the still-unexploited margin. For any given total public social
spending, investing in child development, not least pre-school
children, is more pro-growth and pro-equality than spending the
same amount on public pensions for the well-off (or on transfers
favoring the rich). Generations of powerful groups in many coun-
tries have selfishly failed to heed this lesson from history.

An invest-early consensus that has emerged among micro-economists

in this century can now be backed up by international and historical
comparisons.2 The new finding gained momentum in two stages. First, in
the late twentieth century, empirical studies came to an earlier-pays-more
conclusion for the traditional levels of schooling: the average social rate
of return on education was respectable for higher education, but greater
for secondary schooling, and still greater for primary schooling, as pre-
viewed in Finding #2. Then early in this century, numerous studies based
on a variety of micro-level interventions, mostly in the United States,
concluded that the payoff to inputs into children before age 6 probably
paid even higher returns. Those returns could come from greater

18
 FINDINGS AND LESSONS

parenting time and/or from expenditures on pre-primary education

programs. The earlier the expenditure, the better.
Policymakers have been moving toward the same earlier-is-better
conclusion. First the communist countries took the lead in govern-
ment-run child care and parental leave on a modest scale. Among
democracies, substantial support arose in the late 1960s and 1970s in
response to the rise in women’s labor-force participation, with Sweden
leading the way. A pro-natal wave of supports has arrived since the 1990s,
in response to concerns about low fertility rates (Chapter 9).

THREE POLICY LESSONS FROM WORLD HISTORY

Findings #8–#10 imply three global lessons for government social-spend-

ing policies, namely
• Index annual pension benefits, both negatively to changing senior life
expectancy and positively to peak GDP per adult. That will prevent
any positive trend in the share of GDP that taxpayers spend on
public pension benefits without any absolute pension cuts.
• As much as possible, uncouple the funding of social insurance from the
workplace, reducing reliance on the second pillar of social insurance.
Shift toward funding universal safety-net pensions and universal
health care, with voluntary supplements.
• Invest more in the young, with a “cradle to career” strategy.

The final chapter returns to these three lessons, illustrating how

a country can apply them, guided by its own history and by worldwide
experience (Chapter 14).

THE MOST FAR-SIGHTED AND THE MOST MYOPIC

Which countries and which generations have shown far-sighted invest-

ment in the young, as opposed to the more myopic transfers to the
current elderly? The historical listing should start with the selfish refusal
of the whole world’s politically empowered adults to pay taxes for universal
primary schooling before the mid-nineteenth century. The generational sin
was clearest in the eighteenth and early nineteenth centuries, when

19
 OVERVIEW

investing in basic education in the richest countries could have repaid

the older generations almost immediately, as previewed in Finding #2
above.
Among the currently living generations, which countries have over-
come, and which have failed to overcome, the anti-growth selfishness
with the right mix of social spending? Several advanced countries have
got it right, with a greater educational support per child relative to
support per elderly person (Chapter 7, especially Figure 7.3). The
worst performers – among countries for which we have usable data –
have been India, Turkey, Greece, Latin America, and the global South
more generally. These countries’ underinvestment in the young shows
up in their lower international test scores, lower GDP per person, and
higher inequality.
To enrich this world geography with some history, we should ask which
recent generations have been successfully selfish in their political influence?
The generational timing of each country’s redistributions is unique, of
course. Yet, in most cases, it has been only certain powerful groups that
have reaped the inter-generational harvest. Here is a quick alphabetical
roll call of some cases in which a particular generation has recently
benefited at the expense of others.

• In Brazil, overly generous pensions have been enjoyed by every gen-

eration of pensioners between the passage of the stakeholder-defend-
ing Constitution of 1988 and the modest pension reform of 2019.
Within these generations of pensioners, the gains have been dispro-
portionately captured by the high-income civil servants, judges, and
the military. Even without the coronavirus pandemic of 2020, the
pension system was projected to run far beyond its reserves by 2022.
• In Chile, those who held formal-sector jobs in the period 1960–1973
extracted considerable social security transfers from the rest of society.
Their unearned gains were both sanctioned by, and capped by, the
military government’s pension reform of 1980–1981 (Chapter 13).
• In mainland China since the reforms launched by Deng Xiaoping, the
inter-generational favoritism has been the reverse of what has hap-
pened in most other countries. In the super-growth since the 1980s,
the government short-changed those who were born before 1925, who

20
 FINDINGS AND LESSONS

suffered through their working life before 1990, while heaping health
and education benefits on the generations born since 1970 in the east-
coast provinces.
• In Greece between 1982 and 2014, pensioners in the public and formal
private sectors received unsustainable gains paid for largely by others,
contributing to the country’s debt crisis.
• In Japan, despite a tradition of saying that the elderly should be able to
count on their adult children for support, the elderly population has
instead been supported by the taxpayers of their children’s genera-
tion, ever since a jump in taxpayer support from 1974 to 1983. The
support is only moderate per year of an elderly person’s retirement,
but becomes a huge aggregate burden on younger workers in such an
aged population (Chapters 12 and 14).
• In Singapore, the main beneficiaries will be the first future generation
to benefit from a shift in government policy away from capital accu-
mulation toward public social spending (Chapter 13).
• Turkey’s pensioners have captured increasingly unsustainable benefits
at the expense of the rest of society, at least for the period 1980–2017.
The beneficiaries are again those near the top of society, as in Greece
and several Latin American countries (Chapters 7 and 12).
• In the United States, the generation becoming 65 years old between
1967 and 2002 (born 1902–1937) benefited at the expense of other
generations from the arrival of age-restricted Medicare and the shift
toward more generous non-contributory Social Security benefits. Also
since 1967, the United States has lost its lead in public education, and
has been particularly remiss about having the nation’s taxpayers pro-
vide paid work-leave for the parents of newborns (Chapters 9, 12, and
14).

In the long run, while all public safety nets have helped to stabilize
people’s purchasing powers, the strongest safety net has been the one
woven by investing more in the young.

21
 CHAPTER 11

Do Immigration Tensions Fray the Safety Nets?

T wo future threats are likely to jeopardize the

progress of social spending and its favorable economic and social
effects. One has been developing for decades, and will continue to build
slowly. The other will appear more dramatic in the short run. Both
threats are demographic and political. The knowable trends in life
expectancy, fertility, and migration may threaten the expansion and
productivity of social spending through two politicized channels.
The longer-run channel, the one harder to block or modify, works
through human aging. At we saw in Chapters 6 and 7, population aging
has already shown signs of diverting social spending into less productive
forms.
Fortunately, the longer-run threat from aging is more predictable
than most future developments. We can more reliably forecast its impli-
cations for social spending and the welfare state in, say, the year 2050
than we can predict that year’s technology, or its weather, or its World
Cup winner. What makes the mid-century forecast clearer for the main
threats to the welfare state is that it is driven by vital rates, that is by human
survival, fertility, and migration. Demographers have been able to fore-
cast survival, fertility, and migration well enough for generations ahead.
Luckily for the purposes of this book, it just so happens that the future of
social spending is one of those things depending more heavily on those
predictable demographic vital rates than it will depend on the fuzzier
future of standard economic variables. The population will continue to
age, and Chapters 12 and 13 will concentrate on this slowly advancing
threat.

235
 CONFRONTING THREATS

This chapter addresses the other very real demographic threat – the
shorter-run threat posed by the combination of rising immigration and
the political backlash against it.

WAVES OF IMMIGRANTS – AND OF BACKLASH

Migration has come in waves. The wave-like feature is easily demonstrated by

the history of the country that has absorbed the most immigrants. For the
first century and a third of its independence, the United States was a country
in which one out of every seven residents was born abroad (Figure 11.1). Yet
anti-immigrant sentiments were evident as early as the 1880s. At the same
time that the statue of “Liberty Enlightening the World” was being built as
a French gift on Ellis Island, facing out to the Atlantic, the United States also
banned immigrants from China with the Chinese Exclusion Act of 1882.
Similarly, in 1907, in response to anti-Japanese nativism in California, the
United States and Japan came to informal agreements that Japanese would
be denied passports to the United States. Both restrictions were replaced in
the early 1920s by the even tougher Emergency Quota Act of 1921 and the
National Origins Act of 1924. These banned all immigration of Asians, while

Figure 11.1. The percentage of US residents born outside the United States, 1850–2017
Source: United States Census Bureau.

236
 DO IMMIGRATION TENSIONS FRAY THE SAFETY NETS?

benefiting those from Northern and Western Europe (whose migrations

declined anyway), and shutting out many from Southern and Eastern
Europe.
The anti-immigrant flames of the early 1920s were fanned by three
historical forces. First, population growth had closed the open-land frontier,
which had previously allowed separate ethnic groups based in Europe
enough space to spread out and live separately. Second, the new immigrants
of the 1900–1914 era were from Southern and Eastern Europe, perceived to
be culturally distant from those that had arrived earlier. Finally, the victory of
the Bolsheviks in the Russian Revolution of November 1917 sent a Red Scare
across America. This new fear became associated with nationalities to the
south and east, reinforcing the fear of immigrants.
For example, the New York Times of February 9, 1921 warned that immi-
gration restrictions must be imposed because “American institutions are
menaced” by “swarms of aliens whom we are importing as ‘hands’ for our
industries . . . With the diseases of Bolshevism we are importing also the most
loathsome diseases of the flesh.” In the same year Vice-President Calvin
Coolidge warned in an article in Good Housekeeping that “Biological laws tell
us that certain divergent people will not mix or blend . . . The dead weight of
alien accretion stifles national progress.”1 Such was the mood that passed
the tough immigration bills of 1921 and 1924.
The effect on the foreign-born share was deep and lasting, as Figure
11.1 shows. By the 1960s, the share of all Americans born abroad had
dropped from about 14 percent to about 5 percent. Yet after World War
II, a more prosperous and self-confident America opened the doors
again. The Immigration and Nationality Act of 1952 abolished direct
racial barriers. In 1965, the Immigration and Naturalization Act, also
known as the Hart–Celler Act, officially welcomed foreign-born refugees
and relatives of Americans. Then in 1986, the Immigration Reform and
Control Act granted amnesties to those who had entered the country
without sufficient documentation, while threatening employers of undo-
cumented workers with new fines. These liberalizations, plus postwar
peace and prosperity, allowed the share of foreign-born Americans to
return to its earlier heights by the early twenty-first century. Immigration
has continued to rise, despite attempts to close the door again.

237
 CONFRONTING THREATS

The recent rise of immigration has in fact been shared by all the
richest democracies of Western Europe and North America. Over these
years, the inflow of migrants has grown most clearly for Germany, which
accepted a particularly large number of refugees in the 2015–2016 wave
from Syria, Iraq, and Afghanistan. While this latest wave has been
impeded by the migration barriers in response to the 2020 virus pan-
demic, Europe should expect further rises in the supply of migrants, both
ordinary economic migrants and desperate refugees. The rich receiving
countries all have poorer neighbors with dysfunctional governments.
Economic breakdowns and humanitarian crises are likely. A horrific
descent into civil war like that in Syria after 2012 could easily happen
again in any of several large countries – say, in one of the large countries
of Mediterranean North Africa and the Middle East. Such a civil war
would send another wave of refugees to Europe. While some of the
African and Middle Eastern refugees will also head to Australasia and
North America, there will be an offsetting decline in migration from
Latin America. That region has entered an era of low population growth,
so there will be less demographic pressure to cross the Rio Grande, with
or without a border wall (Hanson and McIntosh 2016).
The wave of refugees that crested in 2015–2016, like earlier high
waves, raised nativist backlash against the arriving foreigners. Almost all
European countries and the United States have seen anti-immigrant
political parties capture a rising share of votes. Much of the accompany-
ing rhetoric has been raw. Viktor Orbán, prime minister of Hungary, has
repeated the phrase “the best migrant is the migrant who does not
come.”2 In Denmark in 2005, Pia Kjaersgaard, then head of the nativist
Danish People’s Party, demonized Sweden’s liberal welcoming of refu-
gees thus: “If they [Swedes] want to turn Stockholm, Gothenburg or
Malmoe into a Scandinavian Beirut, with clan wars, honour killings and
gang rapes, let them do it. We can always put a barrier on the Oeresund
Bridge [between Sweden and Denmark].”3 In the American presidential
election campaign of 2015–2016, candidate Donald Trump similarly
demonized migrants, in this case from Mexico: “When Mexico sends its
people, they’re not sending their best . . . They’re sending people that
have lots of problems, and they’re bringing those problems with us.

238
 DO IMMIGRATION TENSIONS FRAY THE SAFETY NETS?

They’re bringing drugs. They’re bringing crime. They’re rapists. And

some, I assume, are good people.”4
Such raw xenophobia seems to be accompanied by many natives’
genuine misperceptions about immigrants and their fiscal effects. To
back up this point with objective measures of people’s perceptions,
Alberto Alesina, Armando Miano, and Stefanie Stantcheva of Harvard
University (2018) surveyed people in six countries (USA, UK, France,
Italy, Germany, and Sweden) in the winter of 2017–2018. People were
asked about some magnitudes they perceived regarding immigration –
how big is it, where do the immigrants come from, and what happens to
them after they arrive? A virtue of this team’s interview design is that
people’s answers can be compared directly with the actual numbers.
Table 11.1 shows some of the results.

table 11.1. Perceptions versus reality about immigrants in six countries,

November 2017 – February 2018
Native-born persons interviewed in

USA UK France Italy Germany Sweden

What percentage of people in Actual 10.0 13.4 12.2 10.0 14.8 17.6
your country are foreign-born? Perceived 36.1 31.3 28.8 26.4 30.3 27.0
What percentage of your
foreign-born are from North Actual 4.4 6.0 38.9 13.1 18.8 25.0
Africa and the Middle East? Perceived 20.6 20.9 28.2 33.9 32.9 37.2
What percentage of your Actual 10.0 23.0 48.0 33.0 30.0 27.0
foreign-born are Muslim? Perceived 22.7 33.9 50.2 47.0 43.9 44.8
What percentage of your foreign- Actual 22.0 16.6 39.1 49.1 35.1 33.7
Born have not finished high school? Perceived 29.0 25.6 51.6 43.6 37.2 40.9
What percentage of your Actual 5.5 5.7 16.6 14.7 6.9 16.1
foreign-born are unemployed? Perceived 26.4 27.0 38.8 41.8 39.2 37.2
Ratio of government transfers
received by the average foreign- Actual 1.23 1.42 1.39 1.29 0.72 1.44
born, vs. the average native-born Perceived 1.17 1.02 1.77 1.34 1.13 1.28

Source and notes: Alesina, Miano, and Stantcheva (2018). The sample sizes are 4,500 native-
born adult interviewees for the United States, 4,001 for the UK, 4,000 for France, 4,000 for
Italy, 4,001 for Germany, and 2,004 for Sweden, for a total of 22,506 respondents.

239
 CONFRONTING THREATS

First, as shown in the top row, people in all six countries greatly over-
estimate the share of immigrants in their midst, imagining that about
30 percent of residents are foreign-born, when the true shares are only
10–18 percent. In five of the six countries, people also over-imagine the
shares of immigrants that are Muslim or that come from North Africa and
the Middle East. Strikingly, the only country that did not overestimate the
Muslim share, or the share from North Africa and the Middle East, was
France, the country where Muslims and trans-Mediterranean migrants was
the greatest. People in all six countries made the further mistake of under-
estimating the education of the immigrants; as shown in Table 11.1, they
consistently overestimated the low-education share, the share of immigrants
that had not yet finished high school.5
Regarding what happens to the immigrants after arrival, we often hear
conflicting perceptions. Some nationals think that the immigrants are
a fiscal burden because they work very little and get handouts, while others
think they work too much, taking jobs away from the native-born. The team
of Alesina, Miano, and Stantcheva came up with a clear result regarding
work by immigrants of working age. In all six countries of Table 11.1,
immigrants’ unemployment rate has in fact been much lower than people
tend to think, even in that winter of 2017–2018, when so many refugees still
had not mastered the new native language or found jobs. Finally, people
were asked whether the average adult immigrant received more transfer
payments from government than the average native-born. The truth is that
immigrants do receive more on average, because of their initial economic
hardships. On this matter, the interview responses of the native-born were
close to the truth, even though they had overestimated immigrants’ unem-
ployment and had underestimated their education.

HOW IMMIGRATION HAS AFFECTED GOVERNMENT BUDGETS

So have immigrants been a net burden or a net benefit to the native-born

population? Economists have tried to add realism to popular perceptions,
by studying the real-world effects of extra immigrants, or of measures to shut
them out. Let us next survey how these effects have played out under the
immigration policies practiced thus far, to sharpen our guesses about how
different political outcomes and different policies might affect the tax

240
 DO IMMIGRATION TENSIONS FRAY THE SAFETY NETS?

burdens and social-spending behavior of the native-born populations in

immigrant-receiving countries.
The results will depend, of course, on the type of immigration being
restricted – which skill groups would be cut, which age groups, and what
emphasis on refugees versus economic migrants, for example. To mobi-
lize the existing empirical literature, let us begin with its tendency to
summarize the effects of the observed recent mix of immigrants, without
any detailed focus on a particular proposal for immigration restriction.6

IMMIGRANTS’ EFFECTS ON GDP AND GOVERNMENT REVENUE

PER CAPITA. The effects on government revenue, via effects on GDP,
can conveniently draw on the literature estimating the effects of immigra-
tion on labor markets and productivity. That literature, while freely admit-
ting the limitations of its evidence and its econometric estimation, generally
concludes that the observed mix of immigration has probably raised GDP
per capita a bit, especially in studies that try to guesstimate externalities.7
The slight extra GDP in turn translates into a slight addition to
government revenue, slightly improving the affordability of social
programs.8 This bland and tentative positive result seems to enjoy
a consensus, despite the heated controversy over whether natives’ wage
rates are lowered, or raised, or unaffected.

NET BUDGET EFFECTS IMPLIED BY EXISTING TAX RATES AND

ENTITLEMENTS. Even if immigrants contribute slightly to GDP and
government revenue, such slight positives could be swamped by the extra
demands that immigrants place on social-expenditure programs. There is
widespread fear that immigrants are a net fiscal burden, for which the
already-arrived native population must pay. Opinion surveys have shown
that people’s fears of immigrants’ negative fiscal effects loom even larger
than their fears of effects on the labor market, such as their taking jobs away
or lowering wages.9
The fiscal burden of an individual immigrant depends very much on
the immigrant’s age, since all countries tend to support children and the
elderly, while taxing those of prime working age. The first-generation
immigrants receive net benefits when they are still children of school age,
and receive net benefits again in old age. In between, however, they are

241
 CONFRONTING THREATS

net taxpayers. The population of arriving immigrants is unevenly distrib-

uted, however. The first-generation immigrant population includes
more adults per year of working age than children or elderly, so that
the taxpaying of the middle age ranges is actually magnified relative to
the expenditures on young and old.
How does it all net out fiscally, given the age distribution of arriving
immigrants? The net fiscal effects of extra immigrants depend critically
on one’s time horizon. Instead of just asking “are immigrants a net fiscal
benefit or burden?” we must ask “over what time span are they a net
benefit or burden?” Table 11.2 summarizes this dependence concep-
tually, drawing on plausible longer-run simulations.10

table 11.2. The fiscal effects of extra immigrants depend on the question you ask
about them
Burden through public
The question you ask Burden through public pensions? schooling, aid to young?

The “pay-as-you-go” question:

Are today’s immigrants a net → No, immigrants Yes, immigrant
burden on native taxpayers pay for natives’ children subsidized
this year, “pay-as-you-go”? public pensions. by native taxpayers.

The “one lifetime” question:

Is one wave of immigrants a net → Yes, if immigrants have lower lifetime No, immigrants’
burden on native taxpayers over earnings, they get an above-average childrenrepay the rest
the life span of the immigrants and rate of return on pensions, at of society in
their descendants? expense of other taxpayers. productivity and taxes.

The “many generations” question:

Are today’s immigrants, + their No, they become No, they become
children and grandchildren to → heavy net taxpayers, like heavy net taxpayers, like
2100, a net fiscal burden? others. others.

Notes: If one could pay for pensions and other social transfers out of government debt, and
out of the reserves of the pension system, then the left-hand side of the equation should be
modified to include taxes paid in earlier and later years, not just in the current year. Yet they
have to be paid sooner or later, and the problem remains essentially as stated in this pure
same-year version of pay-as-you-go.
Note another simplification here: Taxation here refers only to those taxes that are spent
on education, pensions, and other social benefits, and not the total of all tax revenues. This
analysis ignores taxes channeled into non-transfer spending, such as national defense,
highway construction, and basic government payrolls.

242
 DO IMMIGRATION TENSIONS FRAY THE SAFETY NETS?

Suppose we take the usual short-run view, in the top row of Table 11.2,
asking the pay-as-you-go question “Do today’s extra immigrants cause a net
drain on government budgets right now, in this same year?”11 They often do
so, since extra immigrants’ families typically are a net drain through the host
country’s child-related social programs such as education. If this drain is
greater than the tax revenues collected from adult immigrants, helping to
pay for pensions and other public programs, then immigrants do cause a net
fiscal drain this year. Fiscal-demographic simulations suggest that the net
short-run fiscal effect is indeed negative. A typical mix of immigrant age
groups is so tilted toward the young that the costs of child-centered social
programs yield a negative net result for the first twelve to fifteen years after an
immigrant arrival. This short-run negative effect would also show up in
immigrants’ use of non-contributory aid to those of working age. As OECD
economists have rightly emphasized, the short-run fiscal impact depends
above all on the host country’s success or failure in helping the foreign-born
find jobs.12
Next, suppose one takes a somewhat longer view, asking
Table 11.2’s second question about the net fiscal effects over the whole
lifetimes of the first generation of new immigrants. Now the net effect is
probably positive. True, the immigrants in old age probably get a net
transfer from others, because public pension systems are typically
designed to be progressive, giving a high rate of return to lower-income
earners, such as first-generation immigrants. Yet while that first genera-
tion is aging, its children have already become productive adults, paying
positive taxes instead of needing school money. These tax contributions
should outweigh any intra-cohort redistributions toward foreign-born
pensioners.
Finally, when we consider the whole lifetimes of not only the extra
immigrants but also their children and grandchildren, the net fiscal
effects become clearly positive, as again suggested in Table 11.2. We
know that the eventual fiscal results are clearly positive, because in the
long run the immigrants and their descendants pay more in taxes than
they get in targeted transfers, just like the rest of society. So the answer to
the net-burden question is clearly “no,” not a net burden for any year
beyond about the sixteenth year after arrival. The long-run fiscal effect of
extra immigration is clearly positive.13 That makes perfect sense: In the

243
 CONFRONTING THREATS

long run, we are all descendants of immigrants, and over our lifetimes we
pay more in taxes than anybody receives in social spending – our remain-
ing taxes cover such shared public goods as national defense and public
transit.

WILL IMMIGRATION BACKLASH UNDERMINE SOCIAL

SPENDING?
FOUR OPTIONS

So under existing policies, extra immigrants will tend, on the average, to

have slightly positive effects on GDP, and effects on the government budget
balance will eventually be positive as well. These generally positive results
suggest that immigration has not endangered social spending or the welfare
state – at least not economically, and not with existing policies toward taxes
and social spending. Yet recent political backlash against immigration
threatens to cut immigration, one way or another. If it did, would the
political reaction against immigrants also bring a reaction against social
programs in general, even as they applied to those who are needy and native-
born? The positive fiscal effects of immigrants and their descendants may be
politically trumped by negative perceptions like those surveyed earlier in this
chapter, especially in the wake of a large influx of refugees. Prevailing
opinions can still be negative about the same fiscal effects, and about the
negative effects on some native workers’ earning power, not to mention
cultural phobias and fears of terrorism.
There are four options, or four political scenarios, that could be followed
in the near future. The first of the four is a baseline case in which the
government resists all calls to restrict immigration – in other words, keeps
the country’s doors as wide open as before. Recent experience and recent
surveys of public opinion tell us much about this baseline option.

OPTION 1: WELCOMING IMMIGRANTS WITHOUT DISCRIMI-

NATION. As German Chancellor Angela bravely said in 2015, “wir schaf-
fen das” (We can manage this).14
A country can welcome immigrants and, after paying extra for their
initial training and language learning, give them the same kinds of need-
based entitlements as are given to the native-born. Germany and Sweden

244
 DO IMMIGRATION TENSIONS FRAY THE SAFETY NETS?

led the world in choosing this option in 2014–2016. Germany accepted

890,000 permanent-type refugees from the Middle East in 2015 alone.
Sweden accepted 163,000 that year, an even larger share of the national
population than for Germany. A sustained use of this open-door policy
would promote global growth and equality at the expense of domestic
growth, equality – and harmony.
However, both countries soon hurried to close the gate, both at their
own borders and at the Mediterranean borders of the European Union.
By 2018, Germany had accepted only 185,000 refugee applications,
a 79 percent drop from that 2015 peak of 890,000.
What kinds of countries would follow this first option, which implies
a significant rise in social-spending obligations until the immigrants have
learned the language and found jobs? The most likely candidates are
indeed countries like Germany and Sweden, who welcomed so many
refugees in the wave that peaked in 2015–2016. Initially at least, both
countries have refused to discriminate against immigrants from distant
cultures in their welfare-state system of entitlements, and have accepted
the extra initial burdens that come with such arrivals.
Is Option 1 sustainable, either as an acceptance of more immigrants
or as a commitment to universal social entitlements for all residents? As
noted, both Germany and Sweden moved within a couple of years to
restrict the acceptance of refugees for the foreseeable future. Will the
other shoe drop – that is, will these countries cut social spending even for
those that are native-born? This key question deserves careful reflection.
No immediate answer has been given by the rising anti-immigration
parties themselves. Germany’s rising right-wing party Alternative für
Deutschland (Alternative for Germany), while outspoken on many
issues, has not come out against the welfare state. Nor have their counter-
parts in Sweden. Since 2011 or earlier, Sweden’s right-wing “Sweden
Democrats” Party (not to be confused with the Social Democratic Party,
which has governed Sweden most of the time since 1932) has repeatedly
denied that they seek to cut back the welfare state, knowing that the
welfare state remains popular. Their leader, Jimmie Åkesson, was explicit
about this – “We believe in the welfare state” – while also voicing suspi-
cions about Sweden’s mosques and Islamic community groups.15

245
 CONFRONTING THREATS

The political climate can change, of course. For example, people can
update their opinions about welfare programs, and change their votes, in
response to new exposures to immigrants. We can gather two kinds of
clues to upcoming trends: public opinions about the welfare state, and
actual policy changes.
Public opinion surveys sometimes ask people whether they think that
“social benefits” should be higher, and in a subset of those surveys one
can exploit exogenous-looking variation in exposure to immigration to
see if that exposure seems to reduce support for social spending. An early
opportunity to conduct such a test occurred in Sweden in 1985–1994. At
that time, the government’s Refugee Placement Program assigned refu-
gees to cities, with preference for secondary cities. That is, the refugees
were not allowed to decide where they first lived. That suggests that which
native-born Swedes came in contact with foreigners was initially exogen-
ous, even though the refugees were soon able to change cities. Exploiting
this geographic variation within Sweden, the research team of Matz
Dahlberg, Karin Edmark, and Heléne Lundqvist (2012, 2013) estimated
the effect of locals’ exposure to immigrants on the locals’ answer to the
question “Are you in favor of decreasing the social benefits?” The authors
found that there was indeed a tendency for Swedes more exposed to
immigrants to favor cutting social benefits. And since the question about
social benefits seemed to refer to universal benefits, the implication was
that support for the welfare state was undercut by contact with newly
arrived immigrants. If true, this would suggest something like the nega-
tive effect of ethnic fractionalization on the growth of social spending, as
we described back in Chapter 6.
A more recent survey is that 2017–2018 survey conducted by Alberto
Alesina, Armando Miano, and Stefanie Stantcheva, cited earlier in con-
nection with Table 11.1. The authors found that people’s support for
generous redistribution is undermined by the perception that immi-
grants are more represented among the beneficiaries of redistribution.
At first glance, this seems to reveal direct links between their fears of
immigrants and their willingness to retreat from offering universal social
benefits. However, the link is actually not so direct here. The fears about
immigrants only take the indirect form of their being asked first about
immigrants before being asked the payoff questions about social benefits.

246
 DO IMMIGRATION TENSIONS FRAY THE SAFETY NETS?

Given this ordering of the questions, there is a strong chance that they
interpreted the questions about social benefits as referring to immigrants
as such, not to the more universal benefits of the welfare state.
A deeper test of the link between exogenous exposure to immigrants
and support for income redistribution has now been carried out by the
research team of Alberto Alesina, Elie Murard, and Hillel Rapoport
(2019). Like the Dahlberg–Edmark–Lundqvist team, they exploited var-
iation in actual exposure to immigrants among regions, in this case
among 140 regions within sixteen Western European countries between
2002 and 2016. Controlling for many other things, they explored how
regional exposure to immigrants made the interviewed individuals agree
or disagree with the statement “The government should take measures to
reduce differences in income levels.” The effects were strongly negative,
that is against progressive redistribution. Going beyond asking just about
progressive redistribution, the authors also asked questions about social
spending itself, such as whether one agrees that “social benefits place too
great strain on economy,” “social benefits cost businesses too much in
taxes and charges,” and “social benefits make people lazy.” While social
spending is indeed correlated with progressive redistribution, as shown
in Chapter 10, the authors concentrated on the demand for redistribu-
tion, not on social spending itself. Still, their overall conclusion also
strongly suggests a negative effect of exposure to immigrants on support
for social spending. Importantly, and not surprisingly, they also found
that cultural distance and low skills on the part of immigrants made the
negative effect stronger. Implication: Welcoming low-skilled immigrants
from very different cultures will weaken the universalist welfare state.
So far, all the studies are based on opinion surveys, and all suggest that
an erosion of expressed support for social spending will result either
from anti-immigrant prejudice or from actual contact with immigrants
from distant cultures.
The second kind of relevant clue comes not from expressed opinions
but from directly observed policy changes themselves. Have immigration
shocks negatively affected actual safety-net policies? One econometric
study has looked at the pool of twenty-five developed OECD countries for
the period 1980–2008, testing for effects of changes in immigration on
changes in aggregate social expenditures, controlling for time and

247
 CONFRONTING THREATS

country effects. The authors (Gaston and Rajaguru 2013) find no

negative effects on social spending, either from aggregate immi-
grants, or from lower-educated immigrants, or from migrants from
poorer countries. A similarly muted result comes from a study of the
shifting generosity of social assistance payments among provinces in
Canada between 1986 and 2001 (Green and Riddell 2019). Having
a greater proportion of immigrants did not significantly reduce the
generosity of social assistance to four kinds of household groups,
with the slight exception of assistance to couples with children after
a certain federal reform of 1996. Such studies have the virtue of
testing for actual policy outcomes instead of accepting answers to
opinion surveys. Their null results lose some value from the fact that
they test only policy behaviors from before the slump of 2008–2009
and before the refugee shock of 2015–2016.16 Perhaps the real policy
backlash is yet to come?
There remains the threat of a policy backlash against the welfare states
choosing the generous and even-handed Option 1, even though no
reduction of universal benefits has yet occurred – neither in these statis-
tical studies nor in the social-spending policies of Germany and Sweden
since 2015.

OPTION 2: THE DOOR STAYS LOCKED. At the other extreme, con-

sider the possibility of a country with a large social-spending share that
simply blocks immigration. The main candidates are five Eastern
European members of the European Union: the three Baltic republics,
Slovakia, and Hungary. To think of a country that has clearly tried to shut
the door recently, one can choose Hungary, the most outspoken in its
opposition to immigrants.
In such a case, immigrants would cease to be a short-run budget-
ary burden, aside from the cost of enforcing the barriers at the
border. Social spending for the native population could continue
as before. Equality could also be promoted within the country, if the
migrants being kept out would have competed mainly for lower-
skilled jobs.
In the longer run, however, the restrictiveness of Option 2 would take
a toll on the economy. As we have seen, after maybe twelve to fifteen

248
 DO IMMIGRATION TENSIONS FRAY THE SAFETY NETS?

years, the effects of a typical mix of immigrants on government revenues

and net fiscal surplus would have turned positive, if the migrants had
been allowed into the country. That is, the effect on government rev-
enues and surplus of keeping them out would grow increasingly
negative. So eventually, Option 2 comes with a growth cost relative to
Option 1.17

OPTION 3: “WELFARE CHAUVINISM”. A third possibility is that

immigrants are still allowed to enter, but the government discrimi-
nates against them in its provision of social services. Practicing such
discrimination, also known as “welfare chauvinism,” should in
principle make it easier to avoid dilution of benefits for natives.
This strategy harkens back to the seventeenth century, when the
European towns denied poor relief and other local services to
immigrants, as noted in Chapter 3.
To what extent have immigrant-receiving countries practiced such
discrimination recently? Thus far, countries accepting immigrants have
been unwilling to saddle themselves with immigrants who are not
entitled to basic social services. To date, the main case of discriminating
against migrants in social program entitlements has been China’s hukou
system of internal passports, which blocked rural–to–urban domestic
migrants from the better health, education, and other entitlements of
the major eastern cities between the communal era of the Great Leap
Forward and the partial relaxation of hukou in 1996. Again, there is
a clear parallel to the settlements policies of seventeenth-century
Europe.
In a federal system, discrimination in social services can gain advocates
because a state or local population resists having to pay for immigrants’
social services, yet is unable to block immigration, a policy reserved for
the central government. Another near approach to welfare chauvinism,
receiving immigrants while denying them basic services, threatened to
arise when conservative Californians passed Proposition 187 in 1994. The
proposition called for denying public K-12 education and other public
services to the families of those non-US citizens who had entered the state
without legal documentation. However, Proposition 187 was struck down
by the state’s Supreme Court, and has never been implemented. The

249
 CONFRONTING THREATS

political perils of welfare chauvinism were underlined in

California’s case: Republicans’ support for Proposition 187 is
believed to have doomed their party to long-run minority status in
California.
The real test will be the policy reactions to Europe’s heavy refugee
inflow of 2015–2016. That test is currently in progress. Among the rising
anti-immigrant political parties on the right, the support for welfare
chauvinism has shown an interesting, but logical, pattern. Laurenz
Ennser-Jedenastik (2018) has studied the policy pronouncements of
right-wing anti-immigrant parties in the Netherlands, Sweden,
Switzerland and the UK. These pronouncements have voiced strong
opposition to social assistance, but not to social insurance. That is, they
reject giving outright grants to the foreign-born on the basis of unem-
ployment or poverty, but do not oppose contributory plans in which the
foreign-born would pay into plans entitling them to later health care and
pensions.18 Thus, even the right-wing call for discrimination in social
services has been only partial. Welfare chauvinism seems unlikely to be
embodied in sustained policies.

OPTION 4: CHERRY-PICKING THE BEST EARNERS. A final option

reacts to the anti-immigrant spirit with a combination of blocking boat
people and over-border refugees, while admitting, or even actively
recruiting, the highly educated and highly skilled. The usual formal
name for cherry-picking of immigrants is a points-based immigration
system that determines a non-citizen’s eligibility to immigrate (partly or
wholly) on the basis of that non-citizen’s points in a scoring system
including such factors as education, wealth, language fluency, or an
existing job offer.
Here again, there is little threat to social programs for established
citizens, since the skilled immigrants passing through the filter need little
help and will quickly become net taxpayers.
This fourth option is now practiced widely among rich countries, with
Canada in a leading position. Figure 11.2 shows some striking interna-
tional patterns in the destinations reached by highly educated migrants.
For one thing, they end up in countries where the native-born population
is also highly educated. One reason for this is unrelated to immigration

250
 DO IMMIGRATION TENSIONS FRAY THE SAFETY NETS?

Figure 11.2. Percentage of foreign-born and native-born with post-secondary education,

among the population 25 and older, twenty countries in 2015
Source: Connor and Ruiz (2019, Appendix B).

policy: Those with high skills tend to agglomerate in places where

others have high skills, which in practice means major urban areas in
the richest countries. Figure 11.2 shows this correlation of educated
migrants with educated natives in the international snapshot taken in
2015.
The drift of the skilled toward high-education destination countries
will continue in the future. Even if there were no shift in policy toward
cherry-picking, the immigrant mix would contain more and more “cher-
ries” in the future, for a simple reason unrelated to shifts in immigration
policies.19 As we noted back in Chapter 5, the rate of growth in

251
 CONFRONTING THREATS

educational attainment is much faster today than it was in the leading

countries in the nineteenth and twentieth centuries.
While the tendency of highly educated countries like Canada to attract
the highly educated is partly a byproduct of natural agglomeration ten-
dencies, it is also a matter of explicit selectivity in immigration policy. The
history of Canada and other countries makes this clear enough. Canada
introduced a points-based immigration system back in 1967, replacing the
older bias in favor of Europeans, and has refined the system ever since.
Increasing emphasis has been given to job offers from Canadian employers
and to fluency in English and/or French. Since 2006, and especially since
the economic slump of 2008, Canada has generally tightened the skills
restrictions even further, with the exception of its leading participation,
along with Germany and Sweden, in the humanitarian acceptance of
refugees from the Middle East in 2015–2016.20
Other countries have followed suit. In the 1970s and 1980s, Australia
introduced a points system similar to Canada’s, as did New Zealand in 1991.
Between 2008 and 2010, the United Kingdom phased in a similar points
system for immigrants from outside the European Union. The United States
is drifting in the same direction. The appearance of an American lag in
implementing skill-based immigration policies needs to be qualified in two
respects, however. First, the United States already had such a skills-favoring
policy between its restrictive laws of 1921–1924 and the Immigration and
Nationality Act of 1952, before it shifted toward family reunion policies and
widening the door for Latin Americans. Second, Figure 11.2’s apparent skill
deficit for US immigrants arises primarily from its being the immediate
neighbor of Mexico and Central America. The United States receives
a much larger share of immigrants from that region than do the other
leading countries. If one excludes Mexican and Central American immi-
grants, the observable skills of America’s immigrants are similar in those of
Canada, Australia, and other top skill-importing countries.21

PROMOTING EQUALITY AND GROWTH: AT HOME OR

WORLDWIDE?

So Option 4 looks better than Options 2 and 3 on domestic-efficiency and

domestic-harmony grounds. Of course, a “cherry-picking” policy is no

252
 DO IMMIGRATION TENSIONS FRAY THE SAFETY NETS?

free lunch – with low-wage farm labor blocked from entering the country,
who will pick the real cherries off the trees? Still, Option 4’s cherry-
picking would have three favorable economic effects, relative to a more
balanced admission of immigrants. It would raise GDP per capita, gen-
erate more net revenue for government to spend, and avoid exacerbating
visible domestic inequalities. It would also help to secure social insurance
and social assistance within the country. Think of Option 4 as the
national-level version of walling off gated communities, or to restrictive
zoning laws designed to block the building of housing that the poor can
afford.
Would it really placate the divisive anti-immigrant politics? Yes, it
would help in this respect, according to a recent study by Simone
Moriconi, Giovanni Peri, and Riccardo Turati (2018). Studying elections
in twelve European countries, over the period 2007–2016, they find that
preferences change more strongly in response to the low-skilled immi-
grant share among less-educated voters than for highly educated voters.
Especially among less-educated and older native-born individuals, an
increase in low-skilled immigrants makes their votes and attitudes more
nativist, more anti-immigrant. Thus voting for nationalistic parties is
more sensitive among lower-education native-born voters, and more
sensitive in response to low-skilled immigration.
So the choice should boil down to Option 4 versus Option 1: Choose
between promoting domestic goals and promoting world goals.
For now, the political pendulum on the migration issue is swinging
toward Option 4’s preserving productivity, equality, harmony – and social
spending – within a country, at the expense of world productivity, equal-
ity, and humanitarian relief. The pendulum could swing back again in
the future. Whichever way it swings, we face the same sad dilemma we saw
in connection with Japan, Korea, and Taiwan in Chapter 10: There is an
undeniable conflict between serving these objectives inside a country and
serving them worldwide.

253