View original document

The full text on this page is automatically extracted from the file linked above and may contain errors and inconsistencies.

THIRD/FOURTH QUARTER 2016

FEDERALRESERVE
RESERVEBANK
BANKOF
OFRICHMOND
RICHMOND
FEDERAL

SPECIAL ISSUE

Economics Over
the Life Cycle
Millennials and Their Money
Women at Work
Life Cycle Hypothesis
Gains and Losses in Life Expectancy
Do Economists Ever Really Retire?
Interview with Jonathan Parker

VOLUME 21
NUMBER 3/4
THIRD/FOURTH
QUARTER 2016

FEATURES

11

Are the Kids All Right?
Many worry that the Great Recession and mounting student
debt have stunted millennials’ financial development

Econ Focus is the
economics magazine of the
Federal Reserve Bank of
Richmond. It covers economic
issues affecting the Fifth Federal
Reserve District and
the nation and is published
on a quarterly basis by the
Bank’s Research Department.
The Fifth District consists of the
District of Columbia,
Maryland, North Carolina,
South Carolina, Virginia,
and most of West Virginia.
DIRECTOR OF RESEARCH

Kartik Athreya

14

Why Aren’t More Women Working?
The share of American women in the labor force is slipping
even as it rises in the rest of the developed world

EDITORIAL ADVISER

Aaron Steelman
EDITOR

Renee Haltom
SENIOR EDITOR

David A. Price

18

The Mortality Gap
Life expectancy has increased dramatically over the
past century. But some people might be falling behind

MANAGING EDITOR/DESIGN LEAD

Kathy Constant
STAFF WRITERS

Helen Fessenden
Jessie Romero
Tim Sablik
EDITORIAL
ASSOCIATE
­

Lisa Kenney

DEPARTMENTS

CONTRIBUTORS

1		 President’s Message/Monetary Rules in an Independent Fed
2		 Upfront/Regional News at a Glance
3		 Around the Fed/The Changing Face of the American Family
4		 Federal Reserve/1965: The Year the Fed and LBJ Clashed
8		 Jargon Alert/Life Cycle Hypothesis
9		 Research Spotlight/Underinsuring for Old-Age Care
10		The Profession/Do Economists Ever Really Retire?
22		Interview/Jonathan A. Parker
27		 Economic History/Reaping the Benefits of the Reaper
31			Book Review/Virtual Competition: The Promise and Perils 		
			 of an Algorithm-Driven Economy
32		 District Digest/Social Networks and Economic Outcomes: 		
			 Evidence from Refugee Resettlement Programs
40 Opinion/Immigration and the Economy

TO OUR VALUED READERS:
YOUR INPUT IS IMPORTANT TO US.

Help us improve by returning the reader
survey card in this issue or go online to
harveyresearch.com/econfocus/

Charles Gerena
Santiago M. Pinto
Karl Rhodes
Michael Stanley
John A. Weinberg
DESIGN

Janin/Cliff Design, Inc.

Published quarterly by
the Federal Reserve Bank
of Richmond
P.O. Box 27622
Richmond, VA 23261
www.richmondfed.org
www.twitter.com/
RichFedResearch
Subscriptions and additional
copies: Available free of
charge through our website at
www.richmondfed.org/publications or by calling Research
Publications at (800) 322-0565.
Reprints: Text may be reprinted
with the disclaimer in italics
below. Permission from the editor
is required before reprinting
photos, charts, and tables. Credit
Econ Focus and send the editor a
copy of the publication in which
the reprinted material appears.
The views expressed in Econ Focus
are those of the contributors and not
necessarily those of the Federal Reserve Bank
of Richmond or the Federal Reserve System.
ISSN 2327-0241 (Print)
ISSN 2327-025x (Online)

PRESIDENT’SMESSAGE

Monetary Rules in an Independent Fed

A

long-running debate in central banking is whether
policymakers should follow an explicit formula
for setting monetary policy or whether they should
be allowed some leeway to exercise their best judgments.
Recently, the “rules versus discretion” debate has been reanimated by lawmakers who argue the Fed operates with too
much freedom and not enough transparency. They have proposed legislation that would require the Federal Open Market
Committee (FOMC) to establish and follow a monetary policy
rule — that is, an equation that specifies how the federal funds
rate should respond to changes in economic variables.
Perhaps the best-known rules are Taylor rules, first
developed by John Taylor of Stanford University in 1993 to
describe past central bank behavior during a time when it
was thought to be conducting policy effectively. Taylor rules
express the federal funds rate as a function of inflation and
some measure of real economic activity, such as employment. In general, Taylor rules prescribe lower interest rates
when inflation is below target or employment is falling short
and higher interest rates when inflation exceeds target or
labor markets are exceptionally tight.
Research suggests there are a number of benefits to using
such rules. For example, many economists, including some
at the Richmond Fed, have found that the Fed generally
did follow a Taylor rule during the Great Moderation, the
period from the mid-1980s to the mid-2000s when policy
was relatively successful at keeping inflation low and stable
and minimizing fluctuations in employment. A key element
of this success is that the Fed appeared to follow an aspect
of the rule known as the “Taylor principle,” which states that
the Fed should increase the federal funds rate more than
one-for-one in response to increases in inflation. In contrast,
during the 1960s and 1970s, when inflation was much more
erratic, policymakers departed from this principle.
Given that monetary policy has been fairly close to the prescriptions of a Taylor rule in recent decades, with some exceptions, and that inflation expectations have been well-anchored
over that period, departing from such behavior may erode the
public’s confidence in the Fed’s commitment to price stability. From this perspective, there might seem to be little harm
in legislating the Fed’s adherence to a Taylor-type rule.
But it’s neither reasonable nor realistic to expect monetary policymakers to unthinkingly follow a single rule. In my
view, a rigid requirement, like the one in some proposed legislation that the FOMC choose a single rule and explain any
departures after every meeting, is too draconian. (Although
the proposed legislation does give the Fed the option to
depart from the rule, the strict conditions attached to deviation would create too strong an expectation of adherence.)
One reason is that simple and strict rules might be
too inflexible for the real world, unable to accommodate

unforeseen events or changes
in financial technology, as my
colleague John Weinberg discussed in the First Quarter
2015 issue of this magazine.
In addition, there is no single
“correct” Taylor rule; multiple
versions have been proposed,
all of which rely on assumptions about unobserved variables, such as the natural rates
of unemployment or interest.
Finally, and most importantly,
there is the danger that in legislating a Taylor rule, Congress
could drift into dictating the day-to-day setting of monetary
policy instruments — and history has shown that results are
superior when the Fed sets interest rates independently in
pursuit of monetary policy goals set by Congress.
This does not mean we face an all-or-nothing choice
between blind devotion to a rule and policymakers acting
capriciously, as some would argue. Instead, I believe there is
a sensible middle course.
Policymakers should — and I do — consult the recommendations of a range of policy rules when setting monetary
policy. We should generally stay relatively close to those
recommendations and should depart only with careful consideration and good reason to believe that a departure is warranted. As we know from the pre-FOMC meeting briefing
materials released with FOMC transcripts, at least through
2011 those materials included calculations for a number of
alternative Taylor-type rules. Whether policymakers consulted rules — and if they did, which rules — in 2012 and
beyond will not be known publicly until the meeting materials are released (five years after the meeting date).
But the public deserves to know more about the rules the
committee consults. We could include the calculations for
these rules in the Board’s semiannual Monetary Policy Report
to Congress, along with a discussion of how and why policy
departed from these rules, if applicable. This is a step the Fed
could take voluntarily, without the need for legislative action.
This approach would help meet the objective of increasing
the Fed’s transparency and accountability without tying policymakers’ hands or threatening the Fed’s independence. EF

JEFFREY M. LACKER
PRESIDENT
FEDERAL RESERVE BANK OF RICHMOND

E CO N F O C U S | T H I R D/ F O U RT H Q U A RT E R | 2 0 1 6

1

UPFRONT

Regional News at a Glance

BY L I S A K E N N E Y

MARYLAND — In October, at the 2016 Innovation Showcase in Baltimore,
grants totaling $200,000 were awarded to local startups and to researchers
from Johns Hopkins and the University of Maryland, Baltimore. The Maryland
Department of Commerce awarded four grants, while the Abell Foundation
handed out two. The grants are focused on early-stage medical research and
development and aim to help with commercialization. Among the recipients were
researchers working on projects that include a kidney injury prevention system and
a 3D gamma imaging system for medical applications.
NORTH CAROLINA — In November, former Gov. Pat McCrory announced
that Mill Spring, N.C., will host the 2018 World Equestrian Games at the Tryon
International Equestrian Center — only the second time the international
championship event has been held outside of Europe. The two-year-old Tryon
Center has 12 riding arenas and hundreds of miles of equestrian trails. Based on
past events, it is estimated the games could draw 500,000 spectators to the region
over a two-week period.
SOUTH CAROLINA — Nutritional supplement manufacturer Thorne
Research will build a $35 million, 240,000-square-foot facility near Summerville,
in addition to relocating its corporate headquarters to the area. The new facility
is expected to create 330 research, manufacturing, distribution, and support
jobs and will be operational by mid-2018. Job development credits have been
approved by the Coordinating Council for Economic Development as an
incentive for the project.
VIRGINIA — Gov. Terry McAuliffe launched Cyber Vets Virginia in November.
The initiative aims to link veterans who want to live and work in Virginia with
cybersecurity training and skill development programs, including a free 12- to
15-week training course slotted to begin in April 2017. The initiative will also
provide information on financial support and career tools. Cyber Vets Virginia
is a collaboration involving the state, private-sector leaders, and the Institute for
Veterans and Military Families’ Onward to Opportunity program. McAuliffe
estimated that there are 17,000 cybersecurity job openings in Virginia.
WASHINGTON, D.C. — In late October, Vornado Realty Trust announced it
was merging its D.C. operations with Maryland-based developer JBG Companies.
The tax-free merger will create D.C.’s largest commercial real estate company –
to be called JBG SMITH Properties — and is expected to be completed by the
second quarter of 2017. The $8.4 billion deal includes 11.8 million square feet
and more than 4,400 multifamily units in the District as well as some areas in
Virginia and Maryland. The new company also has a number of projects already
under construction that could result in an additional 20 million square feet of
development.
WEST VIRGINIA — In November, West Virginia University announced
a partnership with China’s Shenhua Energy Co., one of the world’s largest
energy companies. The partnership builds on previous research collaborations
between the two organizations. The new partnership will focus on clean energy
technologies and will promote technology innovation through education and
training exchanges, with the hope that discoveries could lead to cleaner, cheaper
energy being available around the world. It will involve WVU’s colleges of
engineering, business, agriculture, arts and sciences, and law.
2

E CO N F O C U S | T H I R D/ F O U RT H Q U A RT E R | 2 0 1 6

AROUNDTHEFED

The Changing Face of the American Family
BY C H A R L E S G E R E N A

“Family Economics Writ Large.” Jeremy Greenwood,
Nezih Guner, and Guillaume Vandenbroucke, Federal
Reserve Bank of St. Louis Working Paper No. 2016-26B,
January 2017.

T

he American family looks very different than it did
50 years ago, reshaped by a multitude of changes in the
choices that people make and how they are accepted (or not)
by society. Fewer couples are getting married and more are
getting divorced, while more women are in the workforce
and fewer are having babies.
A paper published by the St. Louis Fed has examined the
use of models to better understand the macroeconomic effects
of these and other family-related decisions made at the micro
level throughout the life cycle. According to the authors,
Jeremy Greenwood of the University of Pennsylvania, Nezih
Guner of the Center for Monetary and Financial Studies,
and Guillaume Vandenbroucke at the St. Louis Fed, much
progress has been made in explaining certain trends. These
include the rise in the number of people in the same socioeconomic class marrying each other, a phenomenon known
as assortative mating, and the rise in children living with a
single mother.
Yet questions remain about other family-related decisions.
“It seems likely that the secular decline in fertility is connected
with the rise in married female labor-force participation,”
noted Greenwood, Guner, and Vandenbroucke. “Matching
these long-run facts, in addition to the cross-sectional facts
on female-labor force participation and fertility, would be an
important thing to do. The development of such a macroeconomic model is essential for understanding a host of policy
questions surrounding the family.”
As macroeconomic models incorporate these factors, the
researchers suggested, they could provide much-needed guidance for state and federal lawmakers who want to use public
policy to address societal ills. For example, should child care be
subsidized for the growing number of single-parent families?
Or, taken to an extreme, should tax policy be used to encourage marriage as it has been to encourage homeownership?
“The Role of Selective High Schools in Equalizing
Educational Outcomes: Heterogeneous Effects by
Neighborhood Socioeconomic Status.” Lisa Barrow,
Lauren Sartain, and Marisa de la Torre, Federal Reserve
Bank of Chicago Working Paper No. 2016-17, November
2016.

T

he achievement gap between low-income students
and their more affluent counterparts has proven to
be a difficult problem for policymakers to tackle. It has

widened over the last 50 years and is much larger than
the achievement gap between students of different races.
Policymakers want to break the cycle of poverty that
results from this gap, as well as from Americans’ relatively
low level of income mobility from generation to generation
compared to other developed countries.
Lisa Barrow at the Chicago Fed and Lauren Sartain and
Marisa de la Torre at the University of Chicago recently
examined the effectiveness of Chicago public high schools
with selective enrollment in bridging the achievement gap
between students of differing income levels. Selective public
schools admit students based on admission requirements
such as academic performance and entrance exam scores. In
Chicago’s case, they also consider a student’s socioeconomic
background to extend broader access to their more challenging, academically enriched environments.
Earlier research on selective high schools has suggested
mixed results. In countries where all assignments to secondary
schools are based on test scores, such as Romania and Trinidad
and Tobago, research has found that attending the most selective schools improves student scores on future high stakes
exams. But in cities such as Boston and New York, where only
a small number of schools have selective admission, the results
have been less sunny. While students may be exposed to more
rigorous course work, research has found no effect from these
schools on test scores, according to the paper.
“These findings suggest that any apparent advantages
gained by attending a selective high school are actually due to
selection and not to [the] value that the schools themselves
add for their students,” the authors noted.
Because the admissions processes of Chicago’s selective high schools give disadvantaged students a leg up, and
because those schools are academically enriched, they might
be expected to achieve better outcomes for their disadvantaged students than other Chicago schools. Based on the
paper’s findings, however, that was not the case.
In addition to a lack of an effect on test scores, selective
high schools had a large negative effect on the GPA of students from disadvantaged neighborhoods. Perhaps as a result,
these students were less likely to attend a selective college.
Overall, students at Chicago’s selective high schools did
have a more positive perception of secondary education.
“[They] are more likely to say that students get along well
and treat each other with respect, and they are similarly
more likely to report that their teachers care about them and
listen to their ideas.” They are also less likely to worry about
crime, violence, and bullying at school.
“Perhaps it is factors like these that make SEHSs highly
desirable to students and families — more so than the potential to improve test scores and college outcomes.”
EF
E CO N F O C U S | T H I R D/ F O U RT H Q U A RT E R | 2 0 1 6

3

FEDERALRESERVE
BY H E L E N F E S S E N D E N

The storied
showdown between
Fed Chairman
Bill Martin and
President Lyndon
Johnson wasn’t just
about personalities.
It was a fundamental
dispute over the Fed’s
policymaking role

President Lyndon Johnson and Fed
Chairman William McChesney
Martin Jr., in a more collegial
moment, shake hands during a bill
signing in September 1966.

4

F

ew challenges to the Federal
Reserve’s independence have
ever matched the drama of
Dec. 5, 1965. Fed Chairman William
McChesney Martin Jr. had just convinced the Board of Governors to raise
the discount rate amid signs that the
economy was starting to overheat. Fiscal
stimulus — increased spending on the
Vietnam War, expanded domestic programs for President Lyndon Johnson’s
“Great Society,” and a tax cut enacted in
1964 — had raised inflationary warning
signals for Martin and, increasingly, a
majority of the Federal Open Market
Committee (FOMC). But Johnson was
adamant that higher rates would slow
down the economy and compromise
his domestic agenda. Enraged, he called
Martin and other top economic officials
to his Texas ranch, where he was recovering from gallbladder surgery.
“You’ve got me in a position where
you can run a rapier into me and you’ve
done it,” charged Johnson, as recounted
by Robert Bremner in Chairman of the
Fed. “You took advantage of me and I
just want you to know that’s a despicable thing to do.”

E CO N F O C U S | T H I R D/ F O U RT H Q U A RT E R | 2 0 1 6

Johnson was accustomed to getting
his way — whether through bluntness
or sweet-talking, as the occasion might
require. But not this time.
“I’ve never implied that I’m right
and you’re wrong,” Martin said. “But I
do have a very strong conviction that
the Federal Reserve Act placed the
responsibility for interest rates with the
Federal Reserve Board. This is one of
those few occasions where the Federal
Reserve Board decision has to be final.”
Johnson finally relented, and Martin’s
refusal to back down is often considered
one his strongest moments as Fed chairman. His relationship with the president was sometimes strained in the
following years. But the 1965 showdown
was seen as a tough lesson to Johnson
that the Fed would flex its muscles when
needed to push back against the inflationary pressures caused, in part, by his
administration’s own policies.
What is less often remembered in
the popular mind is that the rate hike
of 1965 did not, in fact, turn a corner
on inflation. In the years that followed,
fiscal stimulus was ample, war spending
kept rising, and the deficit grew. But
FOMC members were often divided,
and their policy decisions reflected
this ambivalence. Furthermore, while
Martin saw monetary and fiscal policymakers as obligated to work together
to promote price stability and growth,
he discovered that dealing with this
particular White House and Congress
was often a one-way street. And even
though the Fed was substantially
upgrading its analytic capacity in the
1960s — hiring more Ph.D. economists, building up its research departments, and adopting forecasting — it
didn’t always translate into consistent
monetary policymaking.
What this meant for the economy
was that high inflation, so closely
associated today with the 1970s, was
already ticking upward in the 1960s.
While it averaged only 1.5 percent a

PHOTOGRAPHY: ASSOCIATED PRESS

1965: The Year the Fed and LBJ Clashed

year from 1952 to 1965, it rose to
policy decisions. As part of this
an annual average of 4.5 percent
approach, he believed, the Fed
Martin’s 19-year tenure saw
starting in 1966. In 1969, it hit
had to communicate effectively
historic changes at the Fed, and
an 18-year high of 5.75 percent.
with Treasury and Congress to
In retrospect, many scholars
achieve a common set of goals.
many scholars consider him one
now believe that the roots of
Sometimes this meant that the
of the most influential Fed
the 1970s inflationary spiral can
burden of adjustment (i.e., tightbe found in the 1960s. The ecoening policy) was on the Fed,
leaders ever.
nomic historian Allan Meltzer
since Congress, as the demohas described 1965 as a turning
cratically elected branch with
point on inflation. Robert Hetzel of the Richmond Fed,
the power of the purse, determined the course of fiscal polsimilarly, noted in his history of the Fed that “an explaicy, including whether to run deficits. “It is monetary policy
nation for the Great Inflation must deal with Martin’s
that must adapt itself to the hard facts of the budget,” is how
responsibility.” Martin himself seemed to have grasped
Martin put it in a 1965 speech. “Not the other way around.”
this, lamenting to his colleagues upon retirement in 1970,
“I’ve failed.”
Priming the Pump
Martin’s approach generally worked well during the adminThe Early Years
istrations of both Eisenhower and Kennedy, even though
Martin’s 19-year tenure saw historic changes at the Fed,
Kennedy pledged to accelerate growth and lower unemployand many scholars consider him one of the most influential
ment and hired economists who were generally supportive
Fed leaders ever. Named as chairman following the 1951
of fiscal stimulus (for example, Walter Heller as chairman
Treasury-Fed Accord — the deal that cemented the Fed’s
of the Council of Economic Advisers, or CEA). But Martin
independence from the executive branch — he presided
had to deal with a new administration in 1964. One of
over a stretch of strong economic growth, interrupted by
Johnson’s first priorities was passing Kennedy’s tax cut
a few relatively short recessions, and low inflation for the
proposal, which Congress quickly cleared that spring. At
next 14 years. During the administrations of Eisenhower
the same time, Johnson sought to ramp up domestic spendand Kennedy, he generally had good relations with a
ing. He also brought on a number of officials, including
White House that was mindful of the Fed’s authority.
Gardner Ackley at the CEA and Henry Fowler to lead the
His commitment to Fed independence and to a strong
Treasury Department, who he thought would support him
price-stability mandate was summed up in two of his most
in these efforts. This camp held that the Fed’s primary role
famous sayings: that the Fed’s role is that of the chaperone
was keeping unemployment very low, around a target of 4
who “has ordered the punch bowl removed just when the
percent, and providing stimulus through low interest rates.
party was really warming up,” and that monetary policy’s
Unlike Martin, they believed allowing a modest amount of
mandate was to “lean against the winds” of either inflation
inflation to reach low unemployment was not risky; as long
or deflation.
as the economy had not reached full employment, it would
Martin’s background was not in economics but in finance.
have enough slack to keep wage pressures in check. And if
His father, William McChesney Martin Sr., had helped draft
inflation did emerge, they believed fiscal policy, rather than
the 1913 Federal Reserve Act and later headed the St. Louis
the Fed, was the most effective tool to manage it.
Fed. Martin Jr. started after college as a bank examiner for
Martin was at odds not only with those officials in the
the St. Louis Fed and later moved to Wall Street. He got
executive branch, but also with some of his fellow FOMC
his first big professional break in 1938, when he was tapped
colleagues. The appointments of George Mitchell (1961) and
as chairman of the New York Stock Exchange at the age
Sherman Maisel (1965) as governors effectively ensured a
of 31. Steeped in Fed history and culture, Martin Jr. was
strong “dovish” plurality. Martin preferred to avoid tipping
profoundly influenced by the failure of the Federal Reserve
the scales during votes until he knew where a majority was
Banks to coordinate monetary policy effectively during the
heading, but as inflationary signs picked up, he increasingly
early years of the Depression, including the missed chance
tried to bring the Reserve Bank presidents — who generally
to prevent the 1929 crash from worsening into a recession
were more independent — to his side.
in the first place. Martin also eschewed economic theory
By spring 1965, Martin became concerned that the stimuand preferred an “intuitive” approach to monetary policy,
lus of the past year was working its way through the economy,
scouring the markets for clues on where interest rates, and
noting signs of rising demand for credit. Money market rates
the real economy, were heading. And until late in his tenure,
and bond yields were trending up. Meanwhile, the effective
he didn’t see much value in economic forecasting.
fed funds rate — what banks can charge each other for
Martin strongly believed that the Fed’s core mission
interbank loans — began to rise above the official discount
was price stability. But he also adhered to the view that
rate — what the Fed charges member banks for loans from
the Fed and the other branches of government would work
the Fed’s discount window, as determined by the Board
most effectively if they respected the interaction of their
of Governors. (At the time, the Fed’s preferred monetary
E CO N F O C U S | T H I R D/ F O U RT H Q U A RT E R | 2 0 1 6

5

policy tool was the discount rate; the fed funds rate didn’t
take on that function until the 1980s.)
To Martin, this indicated that the market was pushing short-term borrowing rates upward, and the Fed was
behind the curve. Industrial wholesale prices were also
rising after holding steady for four years, as was the money
supply, which had expanded by an annualized rate of almost
6 percent by year-end. Martin typically did not focus on
the money supply as an early indicator, but he was alarmed
about the shift in market rates, and his public comments in
the spring and summer began reflecting that. At the same
time, he worried that he didn’t have a majority of the Board
behind him.

The Secret Surge
Another red flag to Martin was that Vietnam War spending began accelerating — and far more than the administration would let on. Johnson announced a massive troop
increase in the summer of 1965 but withheld the actual, far
higher, budget estimates from most of Congress as well as
from the Fed. Johnson got some cover from Ackley, who
said the economy could absorb the extra defense spending
without risking inflation, but Martin had his doubts.
Through secret talks that autumn with Sen. Richard
Russell, D-Ga., Martin learned that war spending was ballooning well above official numbers, by about 25 percent.
At the same time, Johnson kept telling Martin that the Fed
should hold off on any tightening until the White House
released the next year’s budget the following January.
Martin was deeply reluctant to force a confrontation, but
Johnson’s dissembling in the matter made the Fed chairman skeptical that the budget would be accurate. (Indeed,
when the White House released its budget, it asserted that
Congress didn’t need to raise taxes because the war would
end in June 1967.)
Worried that the Fed would be acting too late if it
waited until 1966, and that its independence might be
compromised, Martin decided that early December was
the time to act. On a 4-3 vote, the Board decided on Dec. 3
to lift the discount rate from 4 percent to 4.5 percent. That
also allowed it to lift the ceiling on the prime lending rate
that banks could charge to 5.5 percent (a limit known as
Regulation Q, which the Fed gradually phased out starting
in 1980). As Martin argued to his colleagues, and later to
Johnson and to Congress, if the Fed had decided to keep
short-term rates as low as 4 percent, it would have to flood
banks with more reserves, increasing the risk of inflation.
The showdown at Johnson’s ranch occurred two days
later, and Martin held his ground. He also laid out his
case in public statements after that meeting, emphasizing
that the economy was in strong enough shape — with
unemployment dropping close to 4 percent and labor costs
holding steady — that it could weather the tightening well.
He pointed out that it was a boost in credit demand, not
rising wages, that was driving inflation, and he explained
the Fed’s decision as an adjustment to meet that demand.
6

E CO N F O C U S | T H I R D/ F O U RT H Q U A RT E R | 2 0 1 6

The rate hike “is intended not to reduce the pace
of the economy’s expansion but to moderate mounting
demands for bank credit that might jeopardize that pace
by over-stimulating the economy,” he said in a speech to an
insurance conference in New York City shortly after the
Texas trip. And given that the economy was close to full
employment, he added, the risk was that “bottlenecks will
develop in strategic areas so that large new injections of
bank credit and money will serve to raise prices more than
production.”

The Tax Battle
But it wasn’t enough. Martin and others on the FOMC
soon became alarmed that inflation continued to rise
despite the December 1965 hike. It reached 2.8 percent
by March 1966, and the effective fed funds rate began to
creep over the discount rate, by around a half a percentage
point that summer. In July 1966, without the prospect of
any action on taxes, the Board asked banks to ration credit
rather than raising benchmark rates. This time, the move
had broad support.
In the following months, Martin also made progress in
another priority: getting high-level support to convince
Johnson and Congress to raise taxes to pay for Johnson’s
programs. Higher taxes, Martin believed, would relieve the
Fed of the need to tighten rates further to offset rising deficit financing. By fall 1966, both Ackley and Fowler began
siding with Martin on this point, even though both were
unhappy about the December rate hike. Still, Johnson continued to resist. Powerful fiscal conservatives in Congress
wanted domestic spending cuts in return if they were going
to raise taxes — and that was a bargain Johnson refused to
consider.
The summer tightening of 1966 did dampen inflation
temporarily but brought with it the side effect of a deep
credit crunch. By spring 1967, Martin felt that inflation
had slowed down enough to allow the Fed to dial the
discount rate back to 4 percent — on the condition
that Johnson would finally push his tax hike proposal in
Congress. Again, the president resisted. It was not until
spring 1968, when the Johnson administration and the Fed
had to scramble to address a balance-of-payments crisis
caused by destabilization in the gold market and a looming
collapse of the British pound, that Johnson and Congress
found the support to move the tax hike package. (It was
also at this point that Johnson had decided against running for re-election.) But by then both interest rates and
inflation were moving higher. In fact, starting in fall 1967,
the Board had begun raising the discount rate again, and by
July 1969 it reached 6 percent; the effective fed funds rate
topped 10 percent.
What were the drivers of this inflation? To be sure,
Johnson’s policies produced a sharp rise in deficit spending,
which Johnson failed to offset with higher taxes until the
waning days of his presidency. From 1965 to 1968, the deficit jumped from 0.2 percent of gross domestic product to

2.7 percent. But the inflation of the 1960s also can be traced
to the expansion of the money supply. From the mid-to-late
1960s, it grew at an annualized rate of 5 percent to 7 percent,
well above the average of 4 percent in the first half of the
decade. Among the newer Fed economists at the time, the
growth of money supply was getting increasing attention as
one indicator among several that merited consideration. But
in terms of policy adjustment, the Fed didn’t set targets for
money growth as an intermediate step in controlling inflation; rather, economists were still debating how to measure
it and what role it should play as an indicator.

The Changing of the Guard
Martin’s term was set to end in January 1970, but with
the election of Nixon, Martin feared his leverage would
be diminished in his remaining months. Nixon had long
resented Martin — believing that the Fed’s tightening policy
of the late 1950s caused the brief recession of 1960 and cost
Nixon the election — and settled on the economist Arthur
Burns to replace Martin. An awkward arrangement was
reached in which Burns would succeed Martin as Fed chair
once Martin served out his formal term — but until then,
Burns would work for Nixon as a White House adviser. This
close political relationship is one reason why many scholars,
in retrospect, consider Burns’ tenure to have been compromised from the start.
Many economists today view the 1970s a “lost decade” for
monetary policy, when the Fed, under Burns, failed to craft
a consistent and effective approach to address ever-rising
inflation. As the data show, however, the inflation crisis
began in the 1960s, with two important drivers in particular:
strong stimulus on the fiscal side, including deficit spending,
and the rapid growth of the money supply. Martin secured
some temporary successes — like the 1965 rate hike and the
1968 tax increase — but inflation accelerated all the same.
One constant challenge was that the increases in domestic and war spending were more substantial than initially
expected. But the Fed’s own efforts to control inflation were
not always consistent, due in part to the Board’s divisions;

one example was Martin’s decision to hold off until late 1965
to act, even though he had wanted to move earlier that year.
Finally, Martin himself later admitted he may have placed
too much emphasis on tax policy as a sufficiently powerful
tool to reach his desired outcome, after the 1968 tax hike
failed to have much impact on tamping down inflation.
Testifying before Congress in 1969, Martin addressed
the issue of consistency, suggesting he regretted the Fed’s
decision to ease in 1967 in hopes of getting the tax hike. “[A]
credibility gap has developed over our capacity and willingness to maintain restraint,” he said. “We have been unwilling
to take any real risks.”
Some scholars also note the problems with the Fed’s own
approach. As a traditionalist who preferred studying the financial markets rather than formal models, Martin had parted
company with many of the younger economists joining the
Fed, who began assessing a broader range of indicators, including the money supply. But these refinements had not been fully
incorporated into the FOMC’s own decision-making during
those critical years in the mid-1960s, as Meltzer noted in A
History of the Federal Reserve. For example, rather than take
note of the rapid rise in total reserves — the sum of all bank
deposits and cash — and other monetary aggregates in late
1965 and early 1966, Martin focused primarily on the much
smaller amount of free reserves — what a bank has on hand to
lend — and short-term market rates.
“Martin had not raised the discount rate [in 1965] to
reduce money growth,” wrote Meltzer. Martin and his backers relied “on the decline in free reserves and the rise in the
federal funds rate and other short-term rates. Once again,
these indicators misled them.”
The persistence of inflation weighed heavily on Martin
in his final days as chair — so much so that at his lavish
farewell party at the White House, he shrugged off a series
of laudatory toasts. Instead, he offered an apology for the
state of the economy. “I wish I could turn the bank over to
Arthur Burns as I would have liked,” he said. “But we are in
very deep trouble. We are in the wildest inflation since the
Civil War.” He then sat down, to uneasy applause.
EF

Readings
Bremner, Robert P. Chairman of the Fed: William McChesney
Martin Jr. and the Creation of the Modern American Financial System.
New Haven, Conn.: Yale University Press, 2004.
Collins, Robert M. “The Economic Crisis of 1968 and the
Waning of the ‘American Century.’” American Historical Review,
April 1996, vol. 101, no. 2, pp. 396-422.

Hetzel, Robert L. The Monetary Policy of the Federal Reserve:
A History. New York: Cambridge University Press, 2008.
King, Ronald F. “The President and Fiscal Policy in 1966: The
Year Taxes Were Not Raised.” Polity, Summer 1985, vol. 17,
no. 4, pp. 685-714.
Meltzer, Allan H. A History of the Federal Reserve: Vol. 2, Book 1,
1951-1969. Chicago: University of Chicago Press, 2009.

E CO N F O C U S | T H I R D/ F O U RT H Q U A RT E R | 2 0 1 6

7

JARGONALERT

Life Cycle Hypothesis

W

hat determines how individuals save and spend
their income over their lifetimes? It may seem
like simply a question of personal preference,
but the answer can have big implications for the economy as
a whole. The life cycle hypothesis, which argues that people
seek to maintain the same level of consumption throughout
their lifetimes, is one way that economists have answered
the question — but it was not the first.
An early theory of saving came from John Maynard Keynes’
General Theory of Employment, Interest and Money in 1936.
Keynes viewed saving as simply another type of good that
individuals could “purchase.” As with other goods, Keynes
reasoned that expenditures on saving would increase with
income. This posed a potential problem.
When individuals allocate income toward
saving, it means they aren’t using that
income for consumption. This reduction
in demand for goods and services could
have negative effects on economic output.
To be sure, the negative impact of
a decline in consumption is offset by
the fact that savings are often channeled into productive investments. But
what if there aren’t enough investment
opportunities to absorb people’s desire
to save? Keynes and other economists
like Alvin Hansen of Harvard University
worried that this was a very real possibility as national
incomes grew in the postwar era. Hansen coined the term
“secular stagnation” to describe the economic slowdown
that would result from a “savings glut” with too few investment opportunities.
Studies in the 1940s called Keynes’ saving theory into
question, however. In 1946, Simon Kuznets of Harvard
University examined national income in the United States
between 1869 and 1938 and found that the saving ratio in
America had barely changed across that period, despite large
increases in per capita income. And in a 1947 paper published
by the National Bureau of Economic Research, Dorothy
Brady and Rose Friedman found that the savings ratio for
families at different income levels depended on their income
relative to the mean rather than on their absolute income.
To explain these findings, in the 1950s Franco Modigliani
of MIT and his student Richard Brumberg developed a
new theory for saving. The life cycle hypothesis argued
that people seek to maintain roughly the same level of consumption throughout their lifetimes by taking on debt or
liquidating assets early and late in life (when their income
is low) and saving during their prime earning years when
their income is high. This hypothesis predicts that wealth
8

E CO N F O C U S | T H I R D/ F O U RT H Q U A RT E R | 2 0 1 6

accumulation will follow a “hump-shaped” pattern — that
is, low near the beginning of adulthood and in old age, and
peaking in the middle.
Modigliani and Brumberg’s theory has important implications for the broader economy. In contrast to the Keynesian
view that a country’s aggregate saving rate is driven by its
total level of income, the life cycle hypothesis implies that
the savings ratio depends on the growth rate of income.
When income in a country is growing, each new generation
has higher consumption expectations than the previous one.
To maintain their higher consumption when they get older,
prime-age workers in a growing economy will save more than
past cohorts of prime-age workers, and the dissaving of those
past cohorts (who are now retirees) will be
less than the current workers’ savings rate.
Over the years, empirical studies have
called into question some of the conclusions of the simple life cycle hypothesis.
Data suggest that retirees do not draw
down their wealth as quickly as the
model would predict. Moreover, studies
in the United States and the United
Kingdom find that consumption, too,
is not smooth over people’s lifetimes;
instead, it tends to rise through middle
age and fall after retirement.
There are different possible explanations for these findings. Consumption may be lower for young
people than the model predicts if they are credit constrained.
They may wish to borrow against expected higher future earnings but can do so only if lenders extend the credit to them.
Uncertainty may play a role as well. Since young individuals
don’t know exactly what their future earnings potential will
be, they may hesitate to accumulate a lot of debt for fear that
they won’t be able to pay it off.
Uncertainty plays a role at the end of life as well. Since
individuals do not know exactly how long they will live,
it is hard for them to smoothly draw down their wealth
throughout retirement. Retirees may also save more than
predicted because they wish to leave some of their wealth
to their descendants. Finally, the drop in consumption at
the end of the life cycle could be due to “hyperbolic discounting.” Behavioral economists have advanced the idea
that individuals have trouble planning for the future, which
leads them to save too little to maintain their level of consumption after retirement.
The life cycle hypothesis has evolved in the decades since
Modigliani and Brumberg first developed it, but despite
challenges to it, it remains a key part of modern economic
theory.
EF

ILLUSTRATION: TIMOTHY COOK

BY T I M S A B L I K

RESEARCH SPOTLIGHT

Underinsuring for Old-Age Care

T

BY H E L E N F E S S E N D E N

he financial cost of aging is often unexpected —
decide to buy a policy, the authors found, it would provide
but very serious — to many Americans. One of the
a median annual benefit of $33,000 with a total premium
biggest bills for seniors is late-in-life care, often in a
cost of $72,000; for men of the same age bracket, it would
nursing home, for those who can no longer meet basic needs
come to a median annual benefit of $39,000 with a total
on their own. About one in six seniors will need at least three
cost of $50,000. (The difference accounts for the fact that
years of care, with an average cost of $84,000 a year.
women usually live longer than men, and with a longer lifesDespite the price tag, only about one in five older Americans
pan comes a longer time in a care setting.) To offset these
insure themselves against this risk, according to a recent paper
premium costs, respondents typically said they would scale
published by the National Bureau of Economic Research.
back the amount they would leave to heirs.
Five researchers — John Ameriks of the Vanguard Group,
The researchers listed some possible explanations of why
Joseph Briggs of the Fed’s Board of Governors, Andrew
the current insurance market falls short from the consumers’
Caplin of New York University, Matthew Shapiro of the
perspective. One reason is that many plans don’t differentiate
University of Michigan, and Christopher Tonetti of Stanford
premiums by gender, which adversely affects male customers
University — have tried to explain this behavior. Despite the
because — as noted — they don’t live as long as women on
great expense and substantial chance of needing late-in-life
average. In addition, most of the survey respondents said that
care, why do so few buy a policy
the typical plans offered are too
to protect themselves?
expensive, and cover too little,
“Late-in-Life Risks and the Under-Insurance
The authors posed two
to be a wise insurance purchase.
potential explanations for this
The authors noted that this
Puzzle.” John Ameriks, Joseph Briggs,
puzzle at the outset. The first
perception is borne out by the
Andrew Caplin, Matthew D. Shapiro,
is that before people need this
fact that these plans do often
and Christopher Tonetti.
care, they’re either overly optihave higher overhead than other
National Bureau of Economic Research
mistic or unsure about their
forms of insurance, and these
late-in-life needs — so this
costs are passed on to customWorking Paper No. 22726, October 2016.
underinsurance reflects a risk
ers. Other research suggests that
assessment by consumers. The
these plans usually cover only
other explanation is that this particular market is riddled
about two-thirds of basic costs and often exclude conditions
with gaps: Consumers face a poor selection of insurance
that require long-term care, such as dementia. Consumers are
plans, so they decide that buying a policy is not worth the
concerned about the risk of rate hikes and of being dropped
cost. After analyzing a sample of more than 1,000 seniors
from those plans if they can’t pay those increases. And finally,
aged 55 and over, the authors concluded that much of the
seniors face a shrinking number of plan choices.
puzzle can indeed be understood this way. Far more conFor many, the fallback option is Medicaid, which insures
sumers, they found, would buy late-in-life insurance if these
most low-income seniors who need long-term care. But
policies were better priced and better designed.
many seniors and their families see this route as less than
The study polled Vanguard clients to find out how many
ideal, because the care is considered to be lower quality and
would buy insurance if they were offered well-priced, actuhealth outcomes are worse. Otherwise, seniors or their famarially fair products that they believed would meet their
ilies must either bear the substantial costs of private care or
late-in-life needs. It compared these results against a theorely on a family member for caretaking.
retical model, developed by the authors, that estimated the
The study addressed only consumer behavior and did not
highest possible percentage of seniors who would buy such
draw conclusions about the reasons why insurers did not
a policy. In contrast to the 22 percent who currently own
offer more appealing policies. The authors noted, however,
policies, the coverage rate increases to 46 percent after
that other research has pointed to concern about adverse
accounting for respondents who would buy the improved
selection (that is, the greater incentive for those who believe
product. That share comes much closer to what the model
they will be in need of long-term care to buy policies) and
estimated as the theoretical “ceiling,” which was 59 percrowding out by Medicaid, among other explanations, to
cent. In effect, this means that much, although not all, of
attempt to illuminate insurers’ behavior.
the “gap” between actual purchases and modeled demand
Americans already willingly accept the idea of insuring
can be explained by a poor offering of insurance products.
their cars and their homes, and many buy term life plans. The
What would a typical policy look like if it were better
findings of this study suggest that they might do the same in
designed and actuarially fair? For women aged 55-64 who
greater numbers for old-age care if they had better options. EF
E CO N F O C U S | T H I R D/ F O U RT H Q U A RT E R | 2 0 1 6

9

THEPROFESSION

Do Economists Ever Really Retire?
BY DAV I D A . P R I C E

N

o doubt there are some economists who retire
so they can put their profession in the rear-view
mirror. But many, it seems, never truly leave economics behind. They continue practicing economics long
after they’ve nominally retired or taken emeritus status —
and even when they eventually stop, the economist’s way of
thinking sticks with them.
For some, the compulsion to do economics in retirement
takes the form of publishing. Elmus Wicker, 90, a former
Rhodes Scholar who retired from Indiana University in
1992, turned to writing and publishing three books for
university presses on economic history. Not resting on his
laurels, he has drafted a fourth.
“It never occurred to me that retirement meant doing
something else,” he says. “And it never occurred to me that
maybe I wasn’t still qualified.”
Bruce Yandle retired from Clemson University in 2000,
then returned in 2005 to serve for two years as a dean, then
retired again for good. But he has maintained an adjunct
affiliation with another institution, the Mercatus Center at
George Mason University, where, among other activities,
he advises graduate students on their master’s and doctoral
theses. “Interaction with young people who are excited about
ideas has a contagion associated with it,” he observes. (See also
“Interview: Bruce Yandle,” Region Focus, Second Quarter 2011.)
After Leonard Schifrin retired from the College of
William and Mary in 1998, he found a lot of work coming his
way in his field of health care economics, especially contract
research for the federal government and expert-witness work
in litigation. “I was busy for eight years traveling and doing
interesting things,” he recalls. “That was a lot of fun and really
postponed my retirement from being an economist.”
Research suggests that their sentiments about economics
are widespread among their peers. A 2014 working paper
by several German and Swiss researchers, “Happiness of
Economists,” concluded on the basis of a large-scale survey
that economists are “highly happy with life”; moreover, those
in North America are the happiest (together with those from
Scandinavia and Switzerland). And while data on retirement
rates by discipline are unavailable, a 2002 article by Orley
Ashenfelter and David Card of the University of California,
Berkeley in the American Economic Review found that retirement rates for faculty in the social sciences and physical
sciences at age 70 or 71 are “significantly lower than those
for faculty in humanities or life sciences” — which may mean
that social scientists, including economists, tend to like what
they’re doing.
But if working as an economist is so much fun, why do
they retire at all? Although mandatory retirement at age
70 was once nearly universal in universities, where most
10

E CO N F O C U S | T H I R D/ F O U RT H Q U A RT E R | 2 0 1 6

research economists are employed, Congress abolished mandatory retirement for faculty starting in 1994.
Anecdotally, at least, one of the main factors luring academic economists into retirement is the lure of a zero-course
schedule. Wicker found that when he was teaching, he
couldn’t find the time to write the books he wanted to write.
Some people can do it, he says — but not him. “Even when I
taught only one course, that was distracting,” he says. “I don’t
know quite how to say this, but I think I took teaching too
seriously.”
Freedom from scheduled classes also makes it easier to
collaborate face-to-face with researchers at other institutions. “Once you retire, every day is Saturday,” says Yandle.
“So you can pull up and go somewhere and spend two or
three days with colleagues elsewhere to work on projects
and papers.”
For some academic economists, seniority can make
teaching seem less productive and enjoyable. Schifrin recalls
that the real-world examples he used in class increasingly had
taken place before his students were born. “I felt a growing
generation gap,” says Schifrin. “I lost a way of communicating through examples of economics or political economy,
many of which were at best history, at worst trivial or
unknown to students.”
Economists in retirement who want to continue to be
active in their profession have advantages over their counterparts in some other fields. For instance, unlike their colleagues
in the physical sciences, most of them don’t need a laboratory.
“I have lots of friends and colleagues from other disciplines and most of them do not seem to carry their discipline
work forward into their retirement years,” observes Yandle.
“Most of the economists I know do. I think perhaps part of
it has to do with the fact that we’re a social science. What
you need is a laptop and access to the Web.”
Another factor, Yandle suggests, is that retired economists
may simply be more in demand. “It’s a popular topic,” he says.
“Economists are typically engaged with the world — both
public and private sector — much more than individuals from
comparable professions. Their knowledge and ability to interpret data and events are in demand far beyond the classroom.”
And even if a retired economist no longer participates
in the profession in any form — no research, no writing, no
consulting, no advising students — he or she may well continue to be an economist.
“It’s a discipline where there might not be too much distinction between what we do and who we are,” Schifrin says.
“In retirement, I still think like an economist; I still view the
world from an economist’s perspective. And I think that the
field is so ever-changing that we stay interested in it, and we
want to see what happens next.”
EF

Are the Kids All Right?
Many worry that the Great Recession and mounting student
debt have stunted millennials’ financial development
BY T I M S A B L I K

First Job, False Start
Finding a full-time job is often the first step young adults
take on the path toward financial independence. This can be
a challenge even in the best of times — and for millennials
entering the labor market in 2008 and 2009, it was hardly
the best of times. Rather than looking to hire, employers
were shedding workers at a rapid pace. From the start of
2008 to the fall of 2009, the unemployment rate doubled
from 5 percent to 10 percent. But while millennials faced
higher unemployment rates during the Great Recession,
they didn’t suffer disproportionately worse job losses relative to older workers.
Still, even those millennials who managed to land their first
job in the midst of the recession didn’t necessarily escape the
recession’s curse. Several studies have documented persistent
negative effects on wages for those who begin their careers
during economic downturns. For example, a 2012 article by
Philip Oreopoulos of the University of Toronto, Till von
Wachter of the University of California, Los Angeles, and
Andrew Heisz of Statistics Canada found that a 5 percentage
point increase in unemployment rates translated to as much
as a 9 percent initial loss in earnings for recent male college
graduates in Canada.
A 2016 article by Joseph Altonji and Lisa Kahn of Yale
University and Jamin Speer of the University of Memphis
found a similar effect for U.S. graduates. In both cases, the
authors attributed these effects to the fact that young adults
graduating in a recession have fewer job options, leading

Share of 18- to 34-Year-Olds Living at Home
40
35
30
PERCENT

25
20
15
10

Men

Women

SOURCE: Current Population Survey

E CO N F O C U S | T H I R D/ F O U RT H Q U A RT E R | 2 0 1 6

11

2016

2014

2012

2010

2008

2006

2004

2002

2000

1998

1996

1994

1992

1990

0

1988

5
1986

G

eneration Y. Echo boomers. Millennials. They’ve
been called many things, but one thing for sure
is that those born between the early 1980s and
late 1990s will shape the economy for decades to come.
According to a recent Pew Research Center report, this
group overtook baby boomers as the largest living generation in America in 2015.
But some commentators also call them the Lost
Generation, based on worries that the future doesn’t look
as bright for them as it did for previous generations. Many
millennials graduated from college and began working
just as the worst economic downturn since the Great
Depression hit. They’ve also been called the Boomerang
Generation: According to another Pew study, in 2014
roughly a third of those aged 18 to 34 lived with their parents, edging out marriage or cohabitation with a partner as
the most common living arrangement for the first time in
over a century. (See chart.)
Millennials face other longer-term challenges as well.
They are more likely to have student debt, and more of it,
than previous generations. Since 2001 alone, the median
value of student debt for those who took on loans has
nearly doubled from $6,600 to $11,100, according to the
2013 Survey of Consumer Finances. And while parents
have historically expected their children to be more prosperous than they were at the same age, there are signs
that this may no longer be the case. A recent paper by Raj
Chetty, David Grusky, and Maximilian Hell of Stanford
University, Nathaniel Hendren and Robert Manduca of
Harvard University, and Jimmy Narang of the University
of California, Berkeley found that only half of the children born in the 1980s were earning more than their
parents by age 30, compared to more than 90 percent of
30-year-olds born in 1940.
Some commentators have expressed concerns about the
long-term consequences of these trends. Conventional financial wisdom holds that the earlier one starts building wealth
and saving for retirement, the better. But if millennials are
postponing or entirely avoiding homeownership and struggling with lower wages and higher debt burdens, it may take
them much longer to achieve financial self-sufficiency — if
they ever do. In addition to individual welfare implications,
this would have repercussions for the economy as a whole.
But is the future as dire as it seems for millennials?

them to choose less desirable and lower-paying employers
than they would have in better times. Starting out on a lower
rung also negatively affected their climb up the job ladder,
meaning that these wage effects can persist for up to a
decade. They also found that the losses for recent graduates
during the Great Recession were much larger than in previous recessions going back to 1974.
And yes, one of the ways that graduates have compensated
for weaker labor market opportunities is by choosing to live
with parents longer. An analysis of data from the Current
Population Survey and Consumer Expenditure Survey in a
2015 paper by Daiji Kawaguchi of Hitotsubashi University
and Ayako Kondo of Yokohama National University found
that higher unemployment rates increase the probability that
recent graduates live at home with parents. The authors argued
that young adults use this option as a sort of “intergenerational
insurance mechanism” to smooth their consumption. As a
result, recent graduates did not reduce their consumption as
drastically as would be expected from the recession.

Building Wealth in a Recession
A weak job market was not the only effect the Great
Recession had on millennials just starting out, however. The
collapse of the housing and financial markets had a profound
effect on the wealth of young and old households alike.
On the bright side for millennials, young adults are less
likely to own assets like stocks or homes than older cohorts,
which may have insulated them somewhat from turmoil in
those markets. Indeed, a 2014 paper by Lisa Dettling and
Joanne Hsu of the Federal Reserve Board of Governors found
that, on average, young adults suffered less of a decline in net
worth than older adults during the Great Recession. Still,
those who owned a home or stocks did take a hit. Millennials’
median net worth fell by almost 40 percent, from about
$10,000 in 2004 to about $6,000 in 2013. This decline was
particularly concentrated among the college educated.
Falling asset prices weren’t completely bad for millennials, though. Young adults had the opportunity to benefit
from lower stock and house prices by buying into markets
after the crash and reaping the benefits of the recovery.
But buying stocks during a financial crisis runs counter to
most peoples’ inclinations. In a 2011 article in the Quarterly
Journal of Economics, Ulrike Malmendier of the University of
California, Berkeley and Stefan Nagel of the University of
Michigan found that individuals who have experienced low
stock market returns are less willing to take on financial risk
or participate in the stock market at all and are more pessimistic about future returns if they do participate. Young
individuals are particularly influenced by recent experiences,
since they have fewer lifetime experiences to draw from.
“If you look at the average experience that a millennial
has had with the stock market over the past 10 to 15 years,
it certainly looks different than what a young person would
have seen in, say, 1998,” says Nagel. “These cohorts have
quite different experiences, and as such they would be less
willing to take risks than these earlier cohorts.”
12

E CO N F O C U S | T H I R D/ F O U RT H Q U A RT E R | 2 0 1 6

In fact, Nagel says that young people, like many individual investors in general, tend to invest in the stock market
when returns are high and pull out of the market when
returns are low — the opposite of what would maximize
their returns and minimize the damage from recessions.
Young adults may also face constraints building other
forms of wealth. Owning a home not only provides a place for
an individual or family to live, it also represents a significant
asset. But purchasing a home typically require access to credit,
and young adults are more credit constrained than older
cohorts. Lenders might also tighten standards in response to
a market crash as they did during the Great Recession.
Credit constraints like these may limit how much young
adults can benefit from lower asset prices during a recession.
In a 2016 paper, Sewon Hur of the University of Pittsburgh
constructed a model that included borrowing constraints
for young adults. Hur estimated that young adults suffered
the largest overall welfare losses of any age group during the
Great Recession, equivalent to a 7 percent decline in lifetime
consumption.

Shouldering Student Debt
The Great Recession, while certainly a significant event in the
lives of many millennials, doesn’t seem to fully explain their
pattern of behavior. In fact, some studies, like a 2015 article
by Marianne Bitler of the University of California, Davis and
Hilary Hoynes of the University of California, Berkeley, have
found little relationship between changes in unemployment
and young adults moving in with their parents. The bigger
influence, some believe, is debt.
While it is true that student debt burdens have been
rising on average for decades, it’s not entirely clear what
impact this is having on the decisions of young adults. The
news is full of stories of recent grads struggling to pay down
five- and six-figure student loans. But those cases are more
the exception than the rule, according to economist Beth
Akers of the Manhattan Institute, co-author of the 2016
book Game of Loans.
“The median borrower is spending about 4 percent of
their monthly income on student loan repayment,” says
Akers. “If you look at the data on household expenditures,
that’s similar to the category of personal entertainment.”
Calls to reduce student debt burdens also often assume
other things are held constant. “Of course, most young
adults would like an extra $275 or so a month,” says Akers.
“But if we think about debt as allowing people to make
investments in higher education, then removing that debt
but also taking away the degree and the earning power that
comes with it would almost certainly reduce homeownership
rates and retirement savings.”
Indeed, despite rising college costs, the returns to higher
education are still substantial. A 2014 New York Fed study
estimated that for the last decade, the return from spending
on a college degree has been about 15 percent, making it still
one of the best investments an individual can make. But it is
a return that depends on finishing the degree as well as the

Different Dreams or Deferred Dreams?
How will millennials stack up to their older siblings and
parents in the long run? It is a difficult question to answer,
in large part because the story of this generation is still
being written.
If changing patterns of household formation reflect a
response to the Great Recession, then those patterns may
reverse as that event fades into memory. Still, that process
could take a long time. In terms of risk-taking and the stock
market, “even things that happened 30 years ago still play
some meaningful role,” says Nagel.
But there is some evidence to suggest that at least in the
housing market, retrenchment in response to crises won’t
last forever. A 2015 article by Renata Bottazzi and Matthew
Wakefield of the University of Bologna and Thomas
Crossley of the University of Essex studied homeownership rates in England over the past 40 years. They found
that although individuals who experienced a decline in the
housing market when young reduced their homeownership
rates, that same cohort looked largely the same as earlier
generations by the time they reached age 40. In essence,
generations that exhibit historically low homeownership
rates while young seem to “catch up” as they age.
Data suggest such a catch-up may be taking place among
millennials in the United States. Older cohorts who were in
their mid-to-late 20s when the recovery began in 2010 exhibited larger gains in homeownership by 2014 than younger

50
45
40
35
30
25
20
15
10
5
0
1990
1991
1992
1993
1994
1995
1996
1997
1998
1999
2000
2001
2002
2003
2004
2005
2006
2007
2008
2009
2010
2011
2012
2013
2014
2015

field of study, meaning that some can end up with the debt
and little to show for it. Moreover, the consistent returns to
higher education somewhat mask a trend of worsening outcomes for young adults who don’t go to college. (See chart.)
“The rate of return on higher education has held up
over time and been about constant for the past decade, but
part of what’s keeping that in place is that the alternative is
getting worse,” says Akers. “In essence, it is getting more
expensive not to go to college.”
Indeed, the changes in living arrangements for millennials don’t necessarily point to a choking effect from student
debt but rather to a growing divide between those who
finish college and those who don’t. According to a 2016
report by Richard Fry of Pew Research Center, 40 percent
of 18- to 34-year-old high school dropouts and 39 percent
of high school graduates lived with their parents in 2014,
compared to just 19 percent of college graduates.

$THOUSANDS

Median Annual Wages for Recent Graduates

Bachelor's degree

High school diploma

NOTE: Values in 2016 dollars.
SOURCES: Federal Reserve Bank of New York, U.S. Census Bureau, and U.S. Bureau of Labor Statistics

millennials. Still, there’s always the possibility that this generation will be different. Evidence from the Great Recession
suggests that both young adults and their parents have become
more accepting of living together longer. Additionally, delayed
homeownership may partly be a symptom of changing trends
in family formation. During the height of the nuclear family
era in 1960, 62 percent of young adults were married or cohabiting with a partner by their early 30s. But nearly half as many
millennials, 31.6 percent, were doing so at the same age in 2014.
What will delays in homeownership mean for millennials
on an individual level and for the economy as a whole? The
Survey of Consumer Finances provides a picture of young
adult millennials who were living independently in 2013
and suggests that they are not doing substantially worse on
average than previous cohorts. They are more likely than
young adults in 1989 — members of Generation X — to own
a bank account, a home, retirement accounts, and stocks.
And while their student debt is much higher on average,
other forms of debt like credit cards, housing, and car loans
are lower than for the median young adult in 1989. And that
higher student debt comes with a benefit. Millennials have
received more college education than any other generation
in American history. In particular, female millennials are significantly more likely to have a bachelor’s degree than their
Baby Boom or Gen X counterparts.
But what about those left behind? Non-college graduates
and some college graduates are increasingly struggling to
achieve the American dream. It may be that the data on this
generation reflect this divide. While some of the kids look
to be all right, for others, only time may tell.
EF

Readings
Akers, Beth, and Matthew M. Chingos. Game of Loans: The
Rhetoric and Reality of Student Debt. Princeton, N.J.: Princeton
University Press, 2016.
Altonji, Joseph G., Lisa B. Kahn, and Jamin D. Speer. “Cashier or
Consultant? Entry Labor Market Conditions, Field of Study, and
Career Success.” Journal of Labor Economics, January 2016, vol. 34,
no. S1, part 2, pp. S361-S401.
Dettling, Lisa J., and Joanne W. Hsu. “The State of Young
Adults’ Balance Sheets: Evidence from the Survey of Consumer

Finances.” Federal Reserve Bank of St. Louis Review, Fourth
Quarter 2014, vol. 96, no. 4, pp. 305-330.
Malmendier, Ulrike, and Stefan Nagel. “Depression Babies: Do
Macroeconomic Experiences Affect Risk Taking?” Quarterly
Journal of Economics, February 2011, vol. 126, no. 1, pp. 373-416.
Oreopoulos, Philip, Till von Wachter, and Andrew Heisz.
“The Short- and Long-Term Career Effects of Graduating in a
Recession.” American Economic Journal: Applied Economics,
January 2012, vol. 4, no. 1, pp. 1-29.
E CO N F O C U S | T H I R D/ F O U RT H Q U A RT E R | 2 0 1 6

13

Why Aren’t More
Women Working?
The share of American women in the labor
force is slipping even as it rises in the rest of
the developed world
BY H E L E N F E S S E N D E N

I

n December, the Washington, D.C., City Council took
a historic vote to require paid family leave for employees
working in the District, joining California, New Jersey,
Rhode Island, and New York as part of a movement to
expand such benefits. The debate was fierce — some business owners objected to a new payroll tax the measure would
impose — but the Council voted in the end to enact a compromise bill granting leave up to eight weeks. Lauren Kunis,
a D.C. resident and mother of a toddler, summed up the
sentiment of the bill’s supporters as she told the Washington
Post the legislation would have helped her in the “scary and
vulnerable” time right after childbirth, noting that her husband had to return to work immediately so they could make
ends meet. “It forced us into gender roles we never believed
in,” she said. “He went to work and I stayed home.”
The District and those four states are outliers in the
United States, which has no federally mandated paid leave.
To its supporters, the push for paid leave is primarily about
securing better work-life balance. But it has implications
for a surprising trend affecting the entire U.S. economy:
the declining share of women in the labor force. This drop
is prompting economists to ask just how much paid leave
and other family support policies can help women stay in
the job market over the long run.
The puzzle: American women have long been near the
top of global rankings in educational achievement, workforce participation, and career advancement. But since
2000, women who are between their student years and
retirement are increasingly dropping out of the labor force,
even as more and more complete college. Just as notable is
that the opposite is happening with working-age women
around the world, whether they’re in prosperous economies
with generous family support programs, nations hard-hit
by recession, or countries with more traditional notions of
gender roles. In terms of rank, American women now have
a middling labor participation rate among developed nations
14

E CO N F O C U S | T H I R D/ F O U RT H Q U A RT E R | 2 0 1 6

despite their gains in education — and that rate is slipping
while other nations’ rates are rising. To economists, this is a
surprise because rising education is strongly correlated with
labor force participation. Moreover, researchers are increasingly focused on the broader trend of stagnant or declining
participation by both men and women in the United States
despite the economic recovery since 2009.

The American Exception
Just how different is the United States from the rest of the
world? The Organization for Economic Co-operation and
Development (OECD), a 35-nation club of industrialized
economies, estimates that the average labor force participation rate for prime-age women (defined as aged 25-64)
among its members jumped from 62 percent to 68 percent
between 2000 and 2015. But in the United States, it fell from
73 percent to 70 percent. This may seem like a blip, but it
happened while the percent of prime-age American women
with a college degree spiked from 36 percent to 47 percent
— a jump so large that it’s now about 5 percentage points
higher than that for men. (See charts.) The labor force participation rate for American women is also striking in that it
lags the OECD leaders in female labor force participation by
10 percentage points or more.
“In many other nations, we see a rise in women’s
employment that is driven by working mothers, while in
the United States, that’s been static,” says OECD economist Olivier Thévenon. “And American women who are
highly educated aren’t participating in the labor force to
the same degree that women elsewhere are.”
To be sure, the OECD average rate masks the fact that
some countries have made a big leap from a low baseline
(Spain went from 55 percent to 75 percent), while others
with an already high rate posted a smaller gain, such as
Norway (79 percent to 81 percent). Still, taken together,
these changes cap a global historic shift of women moving

Does Paid Leave Matter?
Paid family leave is one of the major policy differences
between the United States and the rest of the OECD. The
United States is the only OECD member that hasn’t mandated this benefit at a national level, whereas almost every
other member has expanded it in the past two decades,
usually to one to three years. Even the Family and Medical

60
50
40
30
20

2015

2014

2013

2012

2011

2010

2009

2008

2007

2006

2005

2004

2003

2002

2000

0

2001

10

Share with four-year degree

Labor force participation rate

But in the United States, women’s labor force participation
is dropping despite rising educational attainment
80
70
60
50
40
30
20

Labor force participation rate

2015

2014

2013

2012

2011

2010

2009

2008

2007

2006

2005

2004

2003

0

2002

10
2001

Full-time students and retirees, as well as stay-at-home parents and disabled people who aren’t actively looking for work,
are considered out of the labor force, as are people who are so
discouraged that they stopped job hunting. In terms of women’s participation, a mother on leave is still considered in the
labor force if a return to her job is protected (whether leave is
paid or not). But if she formally quits her job to take care of
her child, she’s considered out of the labor force.
Sometimes the labor force participation rate can fall for
demographic reasons, like a rising share of retirees or of
young people who continue studies before starting work.
But if it affects people, whether men or women, in their
prime working years especially, it could have important
macroeconomic consequences. Among other things, lower
labor force participation often means slower GDP growth
(unless productivity jumps), reduced consumption, and less
Social Security and tax revenue. Long breaks from the labor
force also make it more likely that skills erode. Economists
are now focusing more research on why U.S. prime-age
labor force participation rates for both men and women
have not recovered along with the economy since 2009.

70

2000

41.73
42.82
43.40
Who’s In, Who’s Out?
43.97
To economists, labor force
44.59 participation has a very specific
45.13 full-time and part-time workers,
meaning. It includes both
46.36
as well as those who are46.90
not working but are looking for jobs.
47.41

80

PERCENT

out of the home and into the formal labor market. As
21.29
recently as 1980, for example,
the OECD average rate for
21.29
prime-age women was 22.86
only 54 percent.
24.38
What’s particularly 25.69
interesting to economists about the
U.S. decline is that it’s 26.93
concentrated among women in their
27.57
28.13participation rate for U.S. women
30s and 40s. In fact, the
29.31
aged 55-64 has jumped
since 2000, from 52 percent to
30.43
59 percent, a trend also31.90
seen among older women in many
32.31
other countries. But it’s34.14
dropped by a more than offsetting
34.83
amount for those aged 36.60
25-54, whereas it’s risen for that age
37.68that this drop is affecting U.S.
group globally. The fact
women in their childbearing and child-raising years has
led many observers to conclude that the explanation lies
in policy: The lack of paid family leave and subsidized day
care for very young children (from newborn to age 3) may
be a factor in inducing more American women to drop
out of the labor forceShare
—with
while
thedegree
expansion of those
four-year
very benefits abroad may have helped their international
36.13
counterparts stay in. Indeed,
in a 2014 report, the Pew
37.50
Research Center found38.30
an increase in the share of American
38.92
mothers who stay at 39.65
home, from 23 percent in 1999 to
40.11
29 percent in 2012.
40.60

In the OECD, higher education and labor force
participation rates of women are rising ...

PERCENT

Share with four-year degree

Share with four-year degree

NOTE: Labor force participation rate is the percentage of prime-age women, defined as 25-64, who
are working or looking for work. Percentage with four-year degrees is also for prime-age women.
SOURCE: Organization for Economic Co-operation and Development

Leave Act of 1993 (FMLA), which established the requirement of three months of unpaid leave, excludes a large
share of workers — about 40 percent — because it exempts
firms smaller than 50 employees, among other restrictions.
Economists who study the relationship between paid
leave and labor force participation have generally found
a modest, but positive, correlation. On the one hand,
employers may view paid leave as a net liability because
it may make more workers want to take a longer absence
rather than return to work soon and because it may be
financed (as in Washington, D.C.) by a tax on firms. But
international studies suggest that paid leave induces workers to stay at their jobs rather than switching employers or
moving in and out of the work force. This, in turn, builds
labor force attachment.
How has this played out in the United States? California’s
paid-leave policy, now going on 13 years, provides a break
of up to six weeks at a 55 percent wage-replacement rate,
although not all workers take it or are aware of it. But its track
record has been getting more attention from researchers. For
E CO N F O C U S | T H I R D/ F O U RT H Q U A RT E R | 2 0 1 6

15

example, a report co-authored by the Center for Economic
and Policy Research and three university institutes found that
the policy has had the largest impact on workers in low-paying
jobs, who tend to have little or no benefits; for these parents,
job retention rates rose to 83 percent compared to 74 percent
for those who took unpaid or no leave. Meanwhile, Tanya
Byker, an economist at Middlebury College, has published a
study on both the California and New Jersey laws that finds
paid family leave lifts female labor force participation by 5 to
8 percentage points in the months following birth — with a
stronger effect on women without college degrees.
In a study with a broader national sample, Claudia
Goldin of Harvard University and Joshua Mitchell of the
U.S. Census Bureau compared how long mothers in the
1990s stayed in the labor force following the birth of their
first child. Over the course of 10 years, the highest participation rate was found among those who had taken paid
leave offered by their employer, followed by those who took
unpaid leave, and last, those who quit their jobs after their
child’s birth. But Goldin and Mitchell also noted these findings aren’t clear-cut because a woman could fall into more
than one category over those 10 years.
For the time being, however, there is relatively little
U.S. data on paid leave to go on, outside of the few states
mentioned above. Only one in nine U.S. employers offer
paid family leave, so parents tend to use up savings and
vacation days to cover costs if they take time off. Many
mothers who don’t have the finances to cover an unpaid
leave return to work quickly, sometimes within days. And
women who return to work soon tend to be concentrated in
lower-paying, lower-skilled work, and they are more likely
to be single. For this reason, many advocates of paid leave
argue for it primarily on grounds of reducing inequality.

Minding the Kids
Another major policy divergence is the provision of subsidized day care for infants and toddlers. As with paid leave,
this policy has become widespread throughout the developed
world except in the United States. Proponents argue it’s
especially effective at keeping women in the labor force — especially when paired with paid leave — because it substantially
reduces the cost of working outside the home. It also provides
continuity for a woman’s career development and thereby can
make her a more valuable worker in the eyes of employers.
On average, an OECD country spends about 0.9 percent
of its gross domestic product on subsidized day care for
infants and toddlers, although in some cases, such as the
Nordic nations and France, this share rises to 2 percent. In
the United States, whether at the federal or state level, there
is almost no public money at all for day care except some
targeted programs for low-income parents, which vary from
state to state. In terms of per capita public spending on early
child care, the United States ranks near the bottom in the
OECD. (This comparison doesn’t include cash subsidies or
tax credits to offset child care costs, policies that also vary
from country to country. In the United States, parents who
16

E CO N F O C U S | T H I R D/ F O U RT H Q U A RT E R | 2 0 1 6

pay income tax are eligible for a refundable federal tax credit
for child care expenses, but only up to $1,000.)
Despite this low ranking on spending, about 30 percent
of American infants and toddlers are in day care. But this is
a market that is almost entirely private. Research by doctoral
student So Kubota of Princeton University has estimated that
inflation-adjusted hourly costs of day care have risen by 32 percent since the mid-1990s, while the hours of market-based day
care used have fallen by 27 percent (often with informal care
provided by family members making up the difference). The
effect of this cost increase, he estimated, was a drop of 5 full
percentage points in the employment rate for all women, and
a 13 percent drop for mothers with children under 5.
As for its effect, subsidized day care and early childhood
education tends to have a positive impact on women’s labor
force participation — even though it alone is not a “sufficient driver,” in Thévenon’s words. In a study of 18 OECD
countries, Thévenon has estimated that about 2.8 percent
of the total increase in prime-age women’s labor force participation from 1980 to 2007 (that is, a quarter of the total)
resulted from the expansion of those policies. Another new
paper by Claudia Olivetti of Boston College and Barbara
Petrongolo of the London School of Economics has also
found that public spending on day care and early childhood
education lifted labor force participation rates in the countries that enacted them — and generally, these measures
have had a stronger effect than paid leave policies.

The Secular Shifts
Policy debates aside, economists generally agree that even
more fundamental economic changes account for a large part
of the long-term trend of rising female labor force participation across the globe. In poor and developing countries, women’s labor force participation is actually quite high because so
many work in agriculture or in small family businesses. Then,
as economies industrialize, women drop out as men take a
lion’s share of manufacturing jobs. Later, as nations become
wealthier, education tends to become more widespread for
both boys and girls. Educated women, in turn, are much more
likely to join the labor force. They also tend to have fewer
children, and they have them later, because the opportunity
cost of each child rises as well. Another driver that brings
women back to work is the shift from manufacturing to services in advanced economies, as these jobs tend to be female
dominated. For many countries that used to have very few
women working — Southern Europe, Ireland, and Japan, for
example — these long-term changes in labor demand, rather
than modernizing cultural attitudes per se, can help explain
their rising share in the workforce.
In the case of the United States, this boost in women
working since the 1970s may also help to explain the modest decrease of married men in the labor force over that
time, from 97 percent to 93 percent, according to economists Limor Golan and Usa Kerdnunvong of the St. Louis
Fed. They found that as more married women join the
labor force, this can allow their spouses to either work part

time or take time off, whether to take over more domestic
work, spend more time looking for a better-matching job,
or go back to school.

The Part-Time Difference
The growth of the service sector also dovetails with
another trend: the rise of part-time work. Part-time jobs
are much more common in the services sector throughout
the world, and these, too, tend to be female dominated. In
some nations with high female labor force participation, a
large percentage of prime-age women also work part time.
But the OECD average — which came to 22 percent of
women aged 25-54 who work 30 hours a week or less in 2015
— masks a wide range of part-time rates. In the Nordic
countries, they are only in the teens, whereas they reach
almost 55 percent in the Netherlands. As for the United
States, a direct comparison is not quite exact, because the
Bureau of Labor Statistics definition of part time encompasses a broader pool — all women working 35 hours a week
or less. In 2015, that share was around 26 percent.
Cornell University economists Francine Blau and
Lawrence Kahn, who have studied the impact of policy
differences and the rate of part-time work on labor force
participation, believe these cross-national comparisons
are telling. In a 2013 paper, they compared an estimate
of part time incidence in the United States, harmonized
for the OECD’s definition (30 hours a week or less),
with that in 16 other OECD countries. By this measure,
they found that about 13 percent of prime-age American
women worked part time in 2010, compared to 26 percent
of their international sample, suggesting that higher labor
force participation rates outside the United States may be
inflated in part by a higher incidence of women working
part time. The paper did conclude that policy differences
— including parental leave and part-time policies, as well as
public spending on child care — could account for some of
the gap between the rate for American women and women
elsewhere, by about 29 percent. But they also noted more
than half of the employment gains for women outside the
United States came through part-time work.
Why does this matter? To be sure, some women chose
part-time work as the more suitable balance at certain
stages of their lives. But this issue is important to labor
economists because part-time jobs are less likely to lead to
career advancement and better pay.
“Part-time work is important and positive in that it
builds greater labor-force attachment” says Blau. “But it’s
not necessarily a good channel for moving up. It can keep
women trapped in secondary positions.”
Goldin and Mitchell also cite the incidence of parttime work as a factor to consider. In a recent paper, they
estimated what the drop in the international ranking of U.S.
female labor force participation would look like if it were
just confined to women aged 25-54 working in full-time jobs.
According to the OECD, the U.S. ranking fell from sixth
place to 17th from 1990 to 2014. But Goldin found that if the

rankings were adjusted to account for only full-time jobs, the
U.S. drop would be far less — from fourth to eighth.
Meanwhile, Blau and Kahn have also found that
American women are not just more likely to hold full-time
work but are twice as likely to hold managerial positions
than were women in the 16 other OECD countries compared in their sample; American women are also more
likely to work in traditionally male professions. One possible reason, they suggest, is that employers are less likely
to discriminate against female employees if they think the
risk of that employee taking a long leave or switching to
part-time work is low.

Lessons Learned Abroad
In a recent paper, the OECD’s Thévenon noted that the
question of quantifying policy impacts is a complicated
one given the great variation of approaches across countries. Equally challenging is that many of these policies’
effects tend to interact with each other. For example,
a government can offer a long or generous provision of
paid leave and a robust job protection, but if the day care
provision is modest or if the hours of day care offered per
day are limited, a mother may still be inclined to stay at
home. The comprehensiveness of day care may also affect
whether a woman chooses full-time or part-time work. In
general, though, the countries that tend to post the highest
labor force participation rates for women — the Nordic
countries and France — also tend to provide workers with
the most generous leave and day care policies, and the
effects of these two policies tend to magnify each other
in their impact on labor force participation. They also
have a higher full-time female workforce than other countries. In the United Kingdom and other English-speaking
countries, by contrast, less public money is spent on child
care, but leave policies are still generous. There tends
to be more labor market flexibility and more part-time
work. But mothers tend not to return to full-time work
until children are older. This leads to more stratification between high-paying male-dominated jobs and lowerpaying female-dominated ones.

The Disappearing “Hump”
Goldin and Mitchell have been looking at this debate from
a different angle: What if the drop-off of U.S. women in
the labor force is a temporary phenomenon? Their paper
concluded that the rise of older women working has
fundamentally changed the traditional life cycle model of
employment for women. The pattern used to be a “hump”
— more and more women would work as they entered their
30s and 40s, then they would gradually leave the labor force
as they approached retirement age. But increasingly, that
“hump” is flattening out: Among younger generations, more
women are working in their 50s and 60s than earlier generations did once they reached their later years, even if they
dropped out of the labor force in their 30s and 40s.
continued on page 30
E CO N F O C U S | T H I R D/ F O U RT H Q U A RT E R | 2 0 1 6

17

The Mortality Gap
Life expectancy has increased dramatically over the
past century. But some people might be falling behind
BY J E S S I E RO M E RO

A

t the turn of the 20th century, most babies born in
America could expect to die before age 50 — if they
survived past their first birthday, which roughly one
in 10 babies did not. But over the past century, improvements in sanitation, nutrition, and medical care — especially
the development of vaccines and antibiotics — have driven a
dramatic increase in life expectancy. Babies born today can
expect to live until they are nearly 80, and adults who reach
old age are increasingly likely to survive additional years.
But between 2014 and 2015, life expectancy for the U.S.
population as a whole declined for the first time in more
than two decades; some research suggests the decline has
been more pronounced, and perhaps has been a longer-term
trend, for whites (that is, non-Hispanic whites). In addition,
while it has long been the case that the wealthier and more
educated have had longer life expectancies and lower mortality rates than those with less money and education, the gap
appears to have grown in recent decades.
The magnitude of changes in life expectancy and “mortality inequality” is the subject of considerable debate, as
are the causes. But to the extent life expectancies have stagnated or declined, “I do think we should be concerned,” says
economist Janet Currie of Princeton University. “It suggests
that things are going wrong for certain people.”

90
80
70
60
50
40
30
20
10
0
2010

2000

1990

1980

1970

1960

1950

1940

1930

1920

All races, all genders
White women
Black women
Black men

White men

NOTE: Data on life expectancy for all races and genders are through 2015; for other categories
data are through 2014. “Black” and “white” may include persons of Hispanic origin. The drop in life
expectancy in 1918 resulted from the Spanish influenza pandemic.
SOURCE: National Center for Health Statistics, Centers for Disease Control

18

Between 1900 and 2014, life expectancy at birth in the United
States increased nearly 70 percent, from 47.3 years to 78.9
years. (See chart.) But in 2015, life expectancy declined to 78.8
years — only a slight drop, but the first since 1993, when death
rates spiked due to the AIDS pandemic and a particularly
lethal flu season. The decline was reported in a December
2016 report by the Centers for Disease Control (CDC).
Life expectancy at birth tends to understate the number
of years an individual is actually likely to live. The measure
denotes the average age infants born in a given year can
expect to reach, assuming the mortality trends prevailing at
the time of their birth prevail for their entire lives. But life
expectancy generally increases as people reach older ages.
That’s in part because, historically, mortality trends have
improved over time, and in part because life expectancy at
a given age is conditional on having reached that age. For
example, a white baby girl born in 1949 had, at that time,
a life of expectancy of about 72 years. But by the time that
baby was 65, in 2014, she could expect to live about 20 more
years, to age 85.
The 2015 decline in life expectancy was driven by increases
in mortality rates of white men and women of 1 percent and
1.6 percent, respectively, and an increase of 0.9 percent for
black men (that is, non-Hispanic black men). This decline
follows a drop in life expectancy for whites from 78.9 years
in 2013 to 78.8 years in 2014, the most recent year for which
the CDC has published life tables by race and ethnicity. The
2014 decrease was driven by a decline among white women
from 82.2 years to 81.1 years. (Life expectancy was unchanged
for white men at 76.5 years.)

Are More White People Dying?

1910

1900

YEARS

U.S. Life Expectancy at Birth

Recent Trends in Life Expectancy

E CO N F O C U S | T H I R D/ F O U RT H Q U A RT E R | 2 0 1 6

While it’s possible the recent decline for the population as
a whole is just a statistical blip, there is some evidence that
changes in mortality for whites in particular might be more
persistent.
In a recent article, Anne Case and Angus Deaton of
Princeton University analyzed mortality rates for U.S.
adults. They found that between 1978 and 1998, the mortality rate for non-Hispanic whites aged 45-54 declined about
2 percent per year on average. But between 1999 and 2013,
the mortality rate for this group increased about half a
percent per year even as mortality rates for other racial
and ethnic groups continued to decline. According to their

calculations, had the mortality rate for whites in this age
group remained at its 1998 level, nearly 100,000 fewer people would have died; had the rate continued its previous rate
of decline, nearly 500,000 fewer lives would have been lost.
Case and Deaton noted that the increase in white mortality they observed was concentrated in individuals with less
education. Similarly, in a 2012 article in the journal Health
Affairs, a team of researchers found large and unprecedented
declines in life expectancy for non-Hispanic whites who had
not graduated from high school. Between 1990 and 2008,
life expectancy at birth for white women without a high
school degree fell by more than five years; for men, the drop
was more than three years. Life expectancy for blacks and
Hispanics with less than a high school diploma continued to
increase during this period.
But it’s possible that the magnitude of this decline for
white high school dropouts was simply the result of changes
in the composition of the group of people who have not
graduated from high school. Educational attainment for the
population as a whole increased significantly over the course
of the 20th century. As a result, cohorts of people who have
not completed high school have become smaller — and perhaps more disadvantaged — over time.
“Not finishing high school when this is the norm means
that those in this group likely had some underlying background or characteristics working against them, such as a
high level of disadvantage growing up, early life poor health,
or a lack of aptitude for school,” explains Jennifer Dowd of
King’s College London and the City University of New York
School of Public Health. “It could be the case that they have
become more disadvantaged, or it could be that they are as
disadvantaged as always, but now they’re not being averaged
into the group with better outcomes.” Either way, Dowd says,
comparing groups of people who have not completed high
school decades apart is akin to “making an apples to oranges
comparison over time, but calling both fruits apples.”
One way to address the problem is to measure relative
rather than absolute educational attainment, as John Bound
and Arline Geronimus of the University of Michigan, Javier
Rodriguez of Mathematica Policy Research, and Timothy
Waidmann of the Urban Institute did in a 2015 article. They
found that life expectancy for white women in the bottom
quartile of the distribution fell 1.2 years between 1990 and
2010, and white men in this quartile experienced a slight
increase. (Black and Hispanic life expectancy might be less
affected by compositional changes because a larger proportion of blacks and Hispanics do not complete high school,
although completion rates for these groups have increased.)
Compositional changes might also have played a role in
Case and Deaton’s findings. Over the period they studied,
the baby boom generation began moving into the 45-54 age
group, and the average age of the group increased from 49.3
years to 49.7 years. While this is a relatively small increase,
mortality rates increase substantially with age, which could
bias a comparison of age-group mortality over time. Andrew
Gelman of Columbia University calculated age-adjusted

mortality rates for non-Hispanic whites in this age group and
found an increase between 1999 and 2005 and a flattening
between 2005 and 2013 — a less dramatic reversal than found
by Case and Deaton, albeit still a notable break from the
previous trend and from the mortality patterns experienced
by other groups. Gelman also found differences by gender:
Mortality for non-Hispanic white men increased until 2005,
and then began to decline again. But the mortality rate for
non-Hispanic white women increased steadily over the
period studied.
In a 2016 article in the Journal of Economic Perspectives,
Currie and Hannes Schwandt of the University of Zurich
compared life expectancies and mortality rates by race,
gender, and socioeconomic status as measured by county
poverty rates. They also found a divergence in mortality
patterns for white women. Between 1990 and 2010, mortality rates for all white women aged 20-49 were essentially
unchanged, and even increased slightly for women in the
poorest counties, compared with continued declines in
mortality for other groups. But, Currie notes, it’s not all
bad news. “Changes to life expectancy for middle-aged
white women have been very small and from a low base.
At the same time, the gains in life expectancy for young
African-American men have been huge over the past 20
years and dwarf those changes.”

Long Live the Rich and Well Educated
Although the mortality rate for non-Hispanic black men
ticked up in 2015, in general blacks have experienced large
gains in life expectancy, leading to a considerable narrowing
of the racial gap. In 1900, white life expectancy was 14 years
longer than black life expectancy; by 1970, the gap was seven
years, and in 2014, it had fallen to three years. Hispanics tend
to have longer life expectancy than both whites and blacks,
despite the fact that they tend to be of lower socioeconomic
status. (See sidebar.)
A large body of research, dating back to the seminal 1973
book Differential Mortality in the United States by Evelyn
Kitagawa and Philip Hauser, has documented lower mortality rates and longer life expectancy for people with more
income and more education. While estimates vary, studies
suggest that a 25-year-old man with a high school diploma
can expect to live between two and seven years longer than
a man without a high school diploma; the gap for women is
between two and six years.
While gender and racial gaps in life expectancy have
narrowed, socioeconomic ones have increased. Between the
1980s and 2000, the gap in life expectancy between those with
at least some college and those with a high school diploma or
less increased by about 30 percent, according to research by
Ellen Meara of Dartmouth College, Seth Richards-Shubik of
Lehigh University, and David Cutler of Harvard University.
(The authors controlled for negative selection by equalizing
the share of individuals in the high- and low-education groups.)
Bound, Geronimus, Rodriguez, and Waidmann also
found an increase in the education gap when measuring by
E CO N F O C U S | T H I R D/ F O U RT H Q U A RT E R | 2 0 1 6

19

relative educational attainment: Between 1990 and 2010,
whites in the top three quartiles of the education distribution had much larger gains in life expectancy than those in
the bottom quartile.
Similar trends are apparent when comparing life expectancy by income level. In a 2007 article, Hilary Waldron, an
economist at the Social Security Administration, found that
among men born in 1912 who survived to age 60, those in the
top half of the income distribution could expect to live 1.2
years longer than those in the bottom half. For men born in
1941, the gap had increased to 5.8 years. (Waldron’s data were
for male Social Security-covered workers.) More recently, Raj
Chetty of Stanford University and several co-authors studied
life expectancy trends between 2001 and 2014. They found
that life expectancy at age 40 increased 2.3 years during that

period for men in the top 5 percent of the income distribution but only 0.3 years for men in the bottom 5 percent. Highincome women gained 2.9 years in life expectancy, while
gains for low-income women were negligible.
Waldron’s and Chetty’s studies focused on life expectancy at older ages. But Currie and Schwandt also studied
changes in mortality inequality for children. Consistent
with other research, they found that between 1990 and
2010, mortality rates for older adults decreased more
in low-poverty counties than in high-poverty counties,
leading to greater mortality inequality. But for children,
mortality rates declined much more in poor counties than
in rich ones, resulting in less inequality. Given the large
body of research demonstrating that childhood health is a
strong predictor of adult health, this suggests that today’s

The Hispanic Paradox
Hispanics in the United States tend to have longer life
expectancy and lower mortality rates than whites (that is,
non-Hispanic whites) or blacks (non-Hispanic blacks). In
2014, the most recent year for which the CDC has published
life tables by race and ethnicity, life expectancy at birth for
Hispanics was about three years longer than for whites and
about seven years longer than for blacks. (See chart.) At the
same time, Hispanics on average have lower incomes, less
education, and are much less likely to have health insurance
than whites. Given the strong link between socioeconomic
status and health, one would expect Hispanic mortality to
resemble black mortality — and to be worse than whites’,
not better. What explains this so-called “Hispanic paradox”?
In part, it could be a statistical illusion. Mortality rates
are derived from two sources: The numerator — mortality —
comes from the National Vital Statistics System, which collects
information from local death certificates, and the denominator — population — comes from the decennial Census and
the American Community Survey. But information about
Hispanic origin was not included on death certificates in every

YEARS

U.S. Life Expectancy by Hispanic Origin
90
80
70
60
50
40
30
20
10
0

2014
Black, not Hispanic men
White, not Hispanic men
Hispanic men

Black, not Hispanic women
White, not Hispanic women
Hispanic women

SOURCE: National Center for Health Statistics, Centers for Disease Control

20

E CO N F O C U S | T H I R D/ F O U RT H Q U A RT E R | 2 0 1 6

state until 1997, and there is debate about the extent to which
death certificates still understate Hispanic origin. In addition,
the wording of the Census questions has changed over time,
potentially leading more people to identify as Hispanic. If the
denominator has become larger over time, while the numerator is underreported, the mortality rate for Hispanics could
have decreased without any actual change.
But measurement issues can’t explain all of the paradox.
Another possibility is that there are self-selection effects,
such as a tendency for healthier people to migrate in the
first place, or for less-healthy immigrants to return to their
country of origin before they die. Numerous studies have
attempted to quantify the impact of these tendencies, with
mixed results. Even in those studies that do find evidence of
selection effects, selection explains a relatively small portion
of the Hispanic paradox.
There also are social and behavioral differences. For example, some researchers have proposed that strong family and
social ties among Hispanics contribute to better health and
lower mortality. And perhaps the greatest factor is differences
in smoking rates: Hispanics are significantly less likely to
smoke than whites or blacks, and research suggests this could
account for at least half, and perhaps as much as 90 percent,
of differences in life expectancy between Hispanics and
whites, depending on gender and country of origin. Hispanics
also have lower death rates from heart disease, chronic respiratory diseases, accidents (including drug overdoses), perinatal conditions, suicide, stroke, and diabetes.
Whatever explains the Hispanic paradox, Hispanic
mortality might be less paradoxical in the future. Secondgeneration Hispanics tend to be less healthy than those who
were born outside the United States; if Hispanic immigration
rates continue to slow, the health of the population overall
could decline. In addition, rates of obesity and Type 2 diabetes have increased among Mexican Americans, which could
eventually counteract the advantage of lower smoking rates.
— Jessie Romero

children could experience less inequality in mortality and
life expectancy as adults.

A Perfect Storm
Why are today’s adults experiencing more inequality in life
expectancy? Researchers have studied multiple explanations, but establishing a causal relationship between financial
resources or education and mortality risk is a challenging task.
In part, that’s because the relationship runs in both directions; healthier people are in a better position to work and
earn higher incomes in the first place, and those in poor health
might have to stop working. In addition, a link between
wealth and mortality might exist if poor health reduces a person’s assets through high medical expenditures. In the other
direction, those with more income and wealth are able to
purchase better health care and to purchase it earlier.
With respect to education, many studies have tried to
disentangle whether increasing educational disparities are
the result of composition or causation. The answer, says
Jennifer Karas Montez of Syracuse University, is probably a
little of both. “People not graduating from high school today
certainly have more disadvantaged backgrounds than people
who didn’t graduate a hundred years ago,” she says. “At the
same time, what it means to go out into the world today
without a high school credential is much more problematic
than it was a hundred years ago. So you have a perfect storm:
a more disadvantaged group going out and achieving a level
of education that itself confers disadvantage.”
Karas Montez adds, “There’s really nothing inherently
causal about the relationship between education and mortality. The context we’re living in shapes that relationship.
Do you live in an environment where education opens the
door to getting a good job, to having health care, to living
in a safe neighborhood? Or do you have some other initial
advantages or safety net that make your own human capital
less important?”
One factor in rising mortality, particularly for whites,
could be the opioid crisis. Since 1999, overdose deaths
from opioids, including both prescription drugs and heroin, have quadrupled, according to the CDC. The increase
in opioid abuse and related deaths has been concentrated
among whites, although blacks and Hispanics also have been
affected. In addition, while the suicide rate has increased
for the population as a whole since 2000, the increase has
been much larger for whites than for blacks and Hispanics.
Case and Deaton also found that death rates from suicide

and alcohol-related liver disease increased for less-educated
whites between 1999 and 2013 but fell or remained flat for
blacks and Hispanics. Some researchers have dubbed these
“deaths of despair” and suggested that increasing economic
insecurity could be to blame. Still, says Currie, “AfricanAmericans have always had higher unemployment than
whites. So to see life expectancy continuing to improve for
African-Americans over time casts doubt on any simple
story about the health effects of economic disadvantage.”
In addition, present-day mortality patterns might reflect
decisions that were actually made decades ago, such as the
decision to start or quit smoking. After the surgeon general released a report on the hazards of smoking in 1964,
people with more education were much more likely to quit
smoking. In addition, men quit smoking more quickly than
women. Less-educated white women in particular continued
to start smoking even as other groups were quitting and were
slower to quit themselves. And because the negative effects
of smoking can manifest themselves long after a person has
stopped smoking, current mortality rates could be affected.
“If you see differences in death rates between groups
now, you shouldn’t necessarily jump to the conclusion that
it reflects what’s happening to them right now,” says Currie.
“Some of what you’re seeing are the lagged effects of things
that happened a long time ago.”
On an individual level, public health initiatives targeting
smoking, child health, and opioid abuse could lower mortality risk and increase life expectancy for certain groups.
But at the societal level, the complicated interplay between
income, education, and health makes it difficult to ascertain
how or if a given social policy will affect mortality risk. Still,
says Dowd, “There is clearly huge scope for understanding
how the malleable parts of the human social condition
can affect health. Scholars have to keep testing the health
impacts of more specific education and other social policy
changes to understand what works best to give all social
classes the best opportunities for good health.”
Policymakers could have good reason to try to reduce mortality rates and mortality inequality. Beyond basic questions of
equity and fairness, there may be implications for economic
growth. Research suggests that when people expect to live
longer, they invest more in their own human capital, making
themselves more productive. And at the most basic level, economic growth depends on how many people are working and
how productive they are: A healthier society is likely to be a
wealthier society.
EF

Readings
Case, Anne, and Angus Deaton. “Rising Morbidity and Mortality
in Midlife among White Non-Hispanic Americans in the 21st
Century.” Proceedings of the National Academy of Sciences, Dec. 8,
2015, vol. 112, no. 49, pp. 15078-15083.

Dowd, Jennifer B., and Amar Hamoudi. “Is Life Expectancy
Really Falling for Groups of Low Socio-economic Status? Lagged
Selection Bias and Artefactual Trends in Mortality.” International
Journal of Epidemiology, August 2014, vol. 43, no. 4, pp. 983-988.

Currie, Janet, and Hannes Schwandt. “Mortality Inequality: The
Good News from a County-Level Approach.” Journal of Economic
Perspectives, Spring 2016, vol. 30, no. 2, pp. 29-52.

Karas Montez, Jennifer, and Kaitlyn Barnes. “The Benefits of
Educational Attainment for U.S. Adult Mortality: Are They
Contingent on the Broader Environment?” Population Research
and Policy Review, February 2016, vol. 35, no. 1, pp. 73-100.
E CO N F O C U S | T H I R D/ F O U RT H Q U A RT E R | 2 0 1 6

21

INTERVIEW

Jonathan A. Parker
Editor’s Note: This is an abbreviated version of EF’s conversation with Jonathan Parker. For additional content, go to our
website: www.richmondfed.org/publications

22

E CO N F O C U S | T H I R D/ F O U RT H Q U A RT E R | 2 0 1 6

EF: Among your work on economic stimulus programs
is a recent paper with Daniel Green, Brian Melzer,
and Arcenis Rojas on the Car Allowance Rebate
System (CARS) of 2009, popularly known as “Cash for
Clunkers.” Could you discuss the empirical findings of
that paper as well as potential implications for structuring stimulus programs given what we know from
participation in CARS?
Parker: One of the interesting things we saw about that program was that it was massively oversubscribed. The government originally allocated $1 billion to a three-month program
and exhausted that $1 billion in about a week. It then reauthorized the program for another $2 billion and still ran out
of funds two months early. The other notable thing was that
it was a program that provided liquidity. It paid households
$3,500 or $4,500 to trade in and scrap an old vehicle. And that
means it provided liquidity — and really enough liquidity for
a down payment. So we wanted to know: Can we link these
two, the provision of liquidity and the high take-up rate? Also,
there was interesting existing research that had been done
on the program, specifically work by Atif Mian and Amir
Sufi, which produced a nice aggregate impact measure of
the program but nothing at the micro level of how individual
households were responding. And I wondered about the
reversal of the impact, which is one of their main findings: The
program generated sales, but within six to nine months afterward there was no cumulative difference in purchases for
people eligible for the program and people who weren’t.

PHOTOGRAPHY: ANDREW KUBICA, VALENCIA IMAGES/MIT SLOAN SCHOOL OF MANAGEMENT

Economists are sometimes pegged as either theorists
or empiricists. But often this dichotomy is overstated.
Many economists bring together theory and empirical analysis to study a broad range of questions. For
Jonathan Parker, this approach is perhaps the defining
characteristic of his work.
Parker, the Robert C. Merton (1970) Professor of
Finance at the Massachusetts Institute of Technology’s
Sloan School of Management, uses data in novel ways
to better understand a host of economic issues and the
theories that underpin them. For instance, the economic stimulus program of 2008 offered the potential
to examine the way households respond to an influx of
liquidity — and with it, whether people smooth their
consumption, as theory would predict. But to realize
that potential required developing some investigational
tools — in Parker’s case, designing surveys for households belonging to the Nielsen Consumer Panel to
better understand what they did with the payments
they received and why.
Parker has also looked at such issues as whether people can hold incorrect but nonetheless utility-optimizing
beliefs; which segments of the income distribution are
most affected by economic shocks and how that has
changed over time; and whether households respond
to good economic news in a proportionate manner to
bad economic news. As he says, he’s an applied microeconomist, an asset pricer, a macroeconomist, a public
finance economist, and a behavioral economist. Which
one depends on the question at hand and the methods
required to answer it.
Prior to joining the MIT faculty, where he is also the
co-director of the Golub Center for Finance and Policy,
Parker taught at Northwestern University, Princeton
University, and the University of Wisconsin, and he
was a research fellow at the University of Michigan.
He edits the National Bureau of Economic Research’s
Macroeconomics Annual, serves on the board of editors
of the American Economic Review, and is a member of
the Congressional Budget Office’s Panel of Economic
Advisers. Aaron Steelman interviewed Parker at his
office at MIT in December 2016.

We got access to the Bureau of Labor Statistics’
Consumer Expenditure Survey data and made a precise measure of eligibility of vehicles based on fuel efficiency and used
car value by make, model, and year. We then mapped the
program responses to eligibility and the economic subsidy associated with any given car. If you owned a car, the economic subsidy was the program payment minus the value you could get
for your car on the used car market. So a car worth $4,500
on the used car market would get, in effect, no subsidy from
the program, but a car worth $1,000 would get a $3,500
subsidy. We mapped from car value to the program response
to see if people with really junky old cars used the program
much more strongly. And indeed we found that to be the
case. Typically, about $1,000 of used car value reduced your
probability of participating in the program by about half a
percentage point. That suggests the government could have
gotten as large a response with slightly smaller subsidies
because the program ran out of funds and there was a lot of
response from people with moderate-valued vehicles.
EF: But how can you know people’s sensitivity to the
subsidy in advance?
Parker: Exactly the right question. Because the program
ran out of money, it’s not a program for which we observe
an unconstrained, equilibrium response. Instead, it was a
response constrained by the funding amount. So we don’t
absolutely know; what we do know is that the subsidies were
more generous than they needed to be to generate that many
sales. And what we also know is, had the subsidy been lower,
it probably would have been the people with the lousiest cars
who would have traded them in and that would have resulted
in less destruction of more valuable used cars. That’s all easy
to look at and say after the fact. But there was this massive
underestimation of the response to the program, and we
think that’s because of liquidity.
We think that an economic subsidy should generate
intertemporal substitution; it’s a temporary price subsidy to
a durable good. In this case, we figured out how the liquidity
dimension could actually be measured separately a little bit
from the economic subsidy. The economic subsidy is not
the same for everyone, but for most people it is the same
as the liquidity provided by the program. But some people
have loans on their program-eligible vehicles. If a vehicle is
securing a loan, then when it’s brought into the dealer and
scrapped as part of the program, the household has to pay
off that loan, and so they lose some of the liquidity benefit of
the program. In our study, we estimate this liquidity effect,
separate from the intertemporal substitution effect of the
economic subsidy, and we find that the effect of the program
was much smaller on vehicles that were securing loans. In
fact, it’s almost nonexistent. So we find the impact of liquidity to be very strong — it was an accelerant for the economic
subsidy in the target population.
We also find very weak evidence consistent with the
reversal effect that Mian and Sufi first discovered, which feeds

into the question: Is this a worthwhile sort of program to do?
It was a program that caused, at an annual rate, a $44 billion
increase in personal consumption expenditures on durable
goods in the third quarter of 2009, which was the quarter in
which the recession ended and in which GDP grew by about
$44 billion. And in the previous quarter GDP declined by
about $44 billion. So it looks pivotal. On the other hand, half
of the content of the vehicles purchased under the program
was imported, so that means that one has to take the number
of new purchases and divide by two to get an estimate of the
partial-equilibrium impact on demand. So really it wasn’t
pivotal at moving us from no growth to growth, and also the
program seems to have been reversed over six to nine months
because there’s no cumulative impact in sales. On the other
hand, it generated all that spending for a relatively small fiscal
cost of only $3 billion ($12 billion at an annual rate). But these
are all accounting, partial-equilibrium calculations.
For this to be optimal from a stabilization perspective,
you need to believe that the government multiplier is much
larger in the quarter in which CARS is run than six months
later. And this is a period when we are having a slow recovery. So the net benefit of the program is ultimately a general
equilibrium question that other people would need to answer,
but the hurdle is significant given that one has to see such a
significant swing in the size of the multiplier between those
two periods. If one wants to do a similar program again, and
similar programs have now been run in countries all over the
world, our results generally emphasize that the liquidity was
a crucial part of the program — not just people substituting
over time due to a temporary price subsidy — and as such, our
findings relate to the literature on investment tax credits for
firms where liquidity also seems to be important.
EF: I would like to go back to some of your earlier work
on household financial decisionmaking — in particular, your 2002 Econometrica paper with Pierre-Olivier
Gourinchas. It seems consistent with the standard life
cycle theory of saving and consumption. Would you say
that’s a fair characterization of that paper?
Parker: From the perspective of today, I think the contribution of that paper is more methodological in some sense.
We worked out a framework for taking cohort-level analysis
of microdata that had been used nicely before by Angus
Deaton and Christina Paxson, Orazio Attanasio, and others,
and combined it with a structural model of an income fluctuation problem so as to estimate the parameters governing
the behavior of households using a simulated method of
moments estimator. That said, as you noted, the model fits
the life cycle profiles of consumption and saving with a model
in which households differ solely based on their history of
income shocks and their age. So age is a major determinant
of the propensity to spend. Since then, the research has
expanded in many ways to endogenize the choices we made
exogenous in that paper or assumed away, including portfolio
choice, labor supply, illiquid retirement saving, government
E CO N F O C U S | T H I R D/ F O U RT H Q U A RT E R | 2 0 1 6

23

programs, housing, and some very
nice work by Mariacristina De Nardi
and Eric French and co-authors on
retirement. People are also considering the liquidity of different
investments now in structural models and stochastic credit constraints,
all of which we pushed away, but the
method remains a very useful one for
evaluating these models.

Jonathan A. Parker

associated with a propensity to spend
out of liquidity, as illiquidity at the
time of the payment. This same set
of people who have persistently high
propensities to consume are also the
people who characterize themselves
➤ Previous Faculty Positions
Northwestern University (2007-2013);
as the type of people who spend for
Princeton University (1999-2007);
today rather than save for tomorrow
University of Wisconsin (1997-1999)
when I asked them specifically about
their type, not their situation. They
➤ Education
are also the people who report that
Ph.D. (1996), Massachusetts Institute of
EF: You’ve revisited some questhey have not sat down and made
Technology; B.A. (1988), Yale University
tions fundamental to life cycle
financial plans.
➤ Selected Publications
theory in your recent paper,
What you end up with is that
“Why Don’t Households Smooth
“Why Don’t Households Smooth
a high propensity to consume corConsumption? Evidence from a 25
Consumption?”
relates with low liquidity, which is
Million Dollar Experiment,” American
useful for theorizing but also presEconomic Journal: Macroeconomics,
Parker: In that paper, I use Nielsen
ents a little bit of a chicken-and-egg
forthcoming; “Consumer Spending
Consumer Panel data to design and
problem. Is it different preferences,
and the Economic Stimulus Payments
run my own survey on households
objectives, or behavioral constraints
of 2008,” American Economic Review,
to measure the effect of what was
that are causing both the low liquid2013 (with co-authors); “Who Bears
then the second of these large ranity and the propensity to spend, or
Aggregate Fluctuations and How?”
American Economic Review, 2009
domized experiments run by the U.S.
is it the low liquidity that is causing
(with Annette Vissing-Jorgensen);
government, the economic stimulus
the lack of planning and high spend“Optimal Expectations,” American
program of 2008. The key feature
ing responses? So for many purposes,
Economic Review, 2005 (with Markus
of that program was that the timing
what I take my findings to mean is
Brunnermeier); “Consumption Over the
of the distribution of payments was
that the buffer-stock model is a quite
Life Cycle,” Econometrica, 2002 (with
determined by the last two digits of
reasonable model with one critical
Pierre-Olivier Gourinchas)
the Social Security number of the
ingredient. The critical difference reltaxpayer, numbers that are essentially
ative to the way I modeled households
randomly assigned. So the government effectively ran a $100
in the 2002 paper with Gourinchas is that I think there’s
billion natural experiment in 2008, distributing money ranmuch more heterogeneity in preferences across households.
domly across time to people, and this policy provides a way
While in that paper we looked at differences in preferences
to measure quite cleanly how people respond to infusions of
across occupation and industry, I think there’s just much
liquidity.
more persistence in heterogeneity in behavior, consistent in
The goal of “Why Don’t Households Smooth ...” is to
the buffer-stock model with differences in impatience. Partly
provide evidence of the structural model underlying the
I say this because I do not find a big relationship between age
observed importance of liquidity on household spending
and propensity to spend in a number of studies, and partly
behavior. And in theory, while the buffer-stock model might
from the persistence of the high-spending propensities I find
correctly match the behavior, it also might be that people
in this recent paper. But it’s also visible in some sense in even
spend expected income gains only when they arrive because
older data. Low liquidity, or low financial wealth, is a very
of problems stemming from self-control, inattention, inabilpersistent state across households, suggesting the propensity
ity to plan, some sort of rule of thumb or mental accounting
to spend is not purely situational. A lot of it is closer to an
behavior, or the like. So I designed a bunch of questions
individual-specific permanent effect than something trantrying to get at these alternative behaviors. I should clarify
sient due to temporary income shocks.
that they are not really alternatives, in the sense that they all
interact with liquidity constraints.
EF: Did people generally understand the magnitude of
The first thing I found out is that illiquidity is still a trethe 2008 stimulus program prior to receiving payments?
mendous predictor of who spends more when a predictable
And if they didn’t, did that show up in consumption
payment arrives. But it’s not only liquidity. People with low
patterns?
income have a very high propensity to spend, and not just
people who have low income today, as would be associated
Parker: In my study, one of the questions I asked people
with the standard buffer-stock model. You can imagine a
was: So you got this economic stimulus payment, did you
situation where you’ve had a bad income shock, you happen
expect it? Was it more than you expected? Was it less than
to have low liquidity, and you spend a lot. But illiquidity one
you expected? Was it a surprise? First of all, about 80 percent
or even two years prior to the payment is just as strongly
of households got basically what they expected. That means
24

➤ Present Position
Robert C. Merton (1970) Professor of
Finance, Massachusetts Institute of
Technology

E CO N F O C U S | T H I R D/ F O U RT H Q U A RT E R | 2 0 1 6

you’re never going to explain the spending response by people being surprised, as say in some versions of an inattention
model. That is a nonstarter. Expectations about the program
were reasonably accurate, with the important caveat that people may not be answering the survey truthfully. Interestingly,
there is a slightly higher propensity to spend, though not
statistically significantly so, among those who were surprised
and received more than they expected. But there is also exactly
the same response among those people who got less than they
expected. So it looks more like the people who weren’t expecting the right thing are worse at consumption smoothing.
EF: How do you define the distinction between “optimal
expectations” and “rational expectations”? What are the
differences in the ways agents with each set of expectations tend to behave? And if agents with optimal expectations may make “poorer” decisions, in some sense, how
may that ultimately be advantageous or desirable?
Parker: In some sense, the starting point for my work with
Markus Brunnermeier came from a number of observations
in the social psychology literature that people just tend to
be optimistic or overconfident, the type of behavioral biases
that lead people to believe they’re better drivers than average
— that sort of optimism. Looking at the objective functions
that we usually consider, the simplest way to maximize the
expected present discounted value of anything is to put more
probability on better outcomes — simply to be more optimistic. You can see how that can be a source of happiness
today. If we think about how good we are at many different
things, it’s nice to have confidence and believe you’re maybe
better looking or smarter than you actually are. On the other
hand, to the extent that you actually allow yourself these sorts
of enjoyable biases, you’re likely to make slightly worse decisions. You might leave insufficient time to complete a project,
for instance, which would make you worse off.
So the basic idea of that optimal expectations paper is
to think of the optimal trade-off between those two — the
idea that you will get more expected future utility today by
expecting better outcomes, but on the other hand you’re
going to make some decision errors because of that expectation. It turns out that this sort of a simple trade-off has
many interesting implications. The first is basically that
you’re always somewhat optimistic. The reason is that
moving a small amount of probability from, say, the worst
state out in the future to the best leads to a first-order gain
in expected present discounted value of utility flows of consumption. But a small change in probability causes a small
change in behavior, and a small change in behavior from the
optimal has very small — second-order — welfare costs. So,
overall, the benefits outweigh the costs.
There are also some interesting implications that come
from the fact that optimism is situational. For example,
when considering investing, one way to be optimistic is to
think the stock market’s going to go up more than everybody
else believes, and to go longer into it. But you can also be

optimistic by shorting the market and believing it’s going to
crash when everybody else thinks it’s going to go up. So when
do you short? It turns out that there are conditions under
which you will actually invest in an unfair bet if it’s positively
skewed enough. That gives you a theory that looks like people
buying lottery tickets, which are unfair gambles with a very
small probability of a very high positive payout. They provide
a very nice future state to believe in at a pretty low-dollar
cost today. So the observed unfair gambles, lottery tickets,
are exactly the sort of unfair gambles that our theory predicts
people should prefer. This type of behavior looks like the big
short. That’s a theme that runs through several of our results:
People with optimal expectations want something that has a
very high positive payoff to dream about that at the same time
isn’t very costly to invest in.
There is also a natural nonconvexity in the model,
which we didn’t expect. When I am more optimistic
about a certain outcome in the future, that means I want
to buy more consumption in that state of the world. When
I buy more consumption in that state, that means I want to
be more optimistic about it, which in turn means I want to
buy more consumption there. And this natural nonconvexity
means that people are going to do something like hold a reasonably well-diversified portfolio and then invest excessively
in a particular asset, such as one or two individual stocks.
We didn’t expect that sort of behavior to pop out, but
that’s what the model taught us. This leads to our work with
Christian Gollier that looked at the conditions under which
the model generated disagreement and could raise the return
on negatively skewed assets.
EF: How would you describe the changes we have
seen in the way high-income and high-consumption
households have become exposed to aggregate economic fluctuations over the last 30 years roughly?
Parker: Due to some difficult data issues, I have not really
been able to track the consumption of high-consumption
households, but in work with Annette Vissing-Jorgensen we
have looked at how the labor income of high-income households has changed significantly. What we zoomed in on is
that high-income households used to live a relatively quiet
life in the sense that the top 1 percent would earn a relatively
stable income, more stable than the average income. When
the average income dropped by 1 percent, the incomes of the
top 1 percent would drop by about only six-tenths of a percent. In the early 1980s that switched, so that in a recession
if aggregate income dropped by 1 percent, the incomes of the
top 1 percent dropped more like 2.5 percent — quadrupling
the previous cyclicality. So now they’re much more exposed
to aggregate fluctuations than the typical income. We
also show that decade by decade, as the top income share
increased, so did its exposure to the business cycle in the
1980s, 1990s, and 2000s. And as you go further and further
up the income distribution, that top share — not just the top
1 percent, but the top 10th of a percent, and the top 100th of
E CO N F O C U S | T H I R D/ F O U RT H Q U A RT E R | 2 0 1 6

25

a percent — there’s also been a bigger increase in inequality
and a bigger increase in the exposure to the business cycle.
EF: What’s the story for that?
Parker: First of all, we used to think the income cyclicality
was exactly the reverse, because low-income workers would
lose their jobs in recessions and high upper-income workers
would not. And so while high-income households might get
lower raises in recessions, they wouldn’t actually go down to
zero. Since job losses are concentrated among lower-earning workers, you have much greater cyclicality in overall
incomes among low-wage workers. In another paper I did
with Annette Vissing-Jorgensen, we looked at cross-country evidence in the recent decades of high inequality. The
countries with the biggest earnings inequality were also the
countries with the largest high-income cyclicality relative
to the average. So what explains these sorts of findings? We
thought there were two leading hypotheses.
First, starting around the end of the 1980s, we see the
adoption of incentive-based pay for CEOs and other highly
placed managers. Incentive compensation over this time
rises, and it happens to be that the incentive compensation
is not based on relative performance, which would therefore
difference out what goes on in the macroeconomy, but
instead is based on absolute performance. And in the U.S.
case, that could partly be due to simply what counts legally
as incentive-based compensation and so is not subject to
corporate profits tax. Pay in the form of stock options, for
example, counts as incentive-based compensation. Pure
salary does not and so is taxed as corporate profits above
$1 million.
The other possibility is that it’s purely technological.
Something like incentive-based compensation may be a
sideshow. The idea is that new information and communication technologies allow the best managers to manage
more people, to run bigger companies, and therefore to earn
more; the best investment managers to manage more money
and to make more for themselves; the best entertainers and
performers to reach more people and therefore earn a larger
share of the spending on entertainment goods. High earners
have become small businesses. While it is not universally
true that such a shift to high-volume low-markup profits for
the winners necessarily leads to greater cyclicality, it is true
for some reasonable functional forms of production.
We do know that increased cyclicality in income among
high earners can’t come simply from the financial sector.
That sector just isn’t quantitatively big enough, and you
see the increase in earnings share and in cyclicality across
industries and occupations. It’s not the case that just the
top hedge fund managers have become the high earners and
they’re very cyclical; Oprah is also.
EF: What do you think are the most important unanswered or understudied questions in household behavior and household financial decisonmaking?
26

E CO N F O C U S | T H I R D/ F O U RT H Q U A RT E R | 2 0 1 6

Parker: The big one is: Do we need a different model
than the canonical stochastic life cycle model with credit
constraints to understand consumer behavior? Do we need
to introduce inattention or hyperbolic discounting, for
instance, to make it richer? My sense is that for a lot of
questions so far, the answer is still no, but we now have a
few pieces of evidence that in a few places the answer is yes.
As we get better data, and we think about questions like
credit market equilibria and consumer financial regulation,
we have the information to evaluate rich models of behavior
and the need for models that are as complete as possible in
describing behavior.
In my work, liquidity is first order, consistent with the
buffer-stock model. But liquidity almost seems to explain too
much. In the Nielsen study that we discussed earlier, people
don’t spend the money the week before it shows up — they
spend it the week it shows up. And it seems like you’re going
to have a lot of difficulty quantitatively fitting that little foresight into a life cycle model unless people are often literally
liquidity constrained, absolutely at their debt limits.
In the Cash for Clunkers program, liquidity mattered
critically. One interpretation is that this importance is consistent with the canonical model in which some people lack
the liquidity for a down payment. But there is an alternative interpretation. Again, our main finding is that people
who have outstanding loans on their vehicles are much less
likely to participate in the program, presumably because to
buy a new car using the program, they would have to put
some cash down along with the payment in order to make
the down payment. Such people are much less likely to take
advantage of the program than people who don’t have loans
on their vehicles but instead have unsecured debt, like on
a credit card. Sounds like liquidity. But perhaps the people
who have the secured debt could walk into the dealer and
turn into that other person — that is, use their credit card
to buy the car, so they leave the dealer with unsecured debt,
just like the other person. In this case, liquidity matters,
but maybe not according to strictly the life cycle model
with liquidity constraints. Instead, such behavior sounds
more like people using heuristics or mental accounts.
The big question: In what combination do we need each
ingredient – rationality and heuristics? And where do the
heuristics come from?
The other question that I think research is really exploring
is what equilibria look like for saving and borrowing. What
equilibrium supports high-fee mutual funds, index funds, and
so on, and how does that change the flow of funds between
the corporate and household sector and the pricing of risk?
How does the market for lending to households evolve as risk
is repriced and interest rates move, and how does this feed
back into spending? The interplay between borrowers and
lenders in these markets is a very interesting and active area of
research because we’re getting a lot of the data on mortgages,
credit cards, retail investment, and financial accounts. These
data are allowing us to look at and understand the equilibria in
those markets, which is really fun.
EF

ECONOMICHISTORY

Reaping the Benefits of the Reaper
Cyrus McCormick may not have invented the reaper, but he was the
entrepreneur who made it successful
BY K A R L R H O D E S

ILLUSTRATION: MCCORMICK/IH COLLECTION, WISCONSIN HISTORICAL SOCIETY, IMAGE 39559

C

yrus McCormick spied his archrival for the first
question, says David Hounshell, professor of technology
time in the April 1834 issue of Mechanics’ Magazine,
and social change at Carnegie Mellon University. “From a
which published a drawing and description of a
Schumpeterian perspective, who was the successful entrepremechanized reaping machine patented by Obed Hussey.
neur who was innovating mechanized reaping in the United
McCormick immediately wrote a letter to the editor claimStates and Europe?”
ing that he had invented a reaper in 1831 based on the same
Joseph Schumpeter, a Harvard University economist
principle as Hussey’s machine.
who was born one year before Cyrus died, famously high“I would warn all persons against the use of the aforesaid
lighted the key role that entrepreneurs play in driving
principle,” McCormick wrote, “as I regard and treat the use
economic development. In his 1912 book, The Theory
of it, in any way, as an infringement of my right.”
of Economic Development, Schumpeter wrote: “Innovation
McCormick was staking his claim to one of the
is the market introduction of a technical or organizamost important breakthroughs in the mechanization of
tional novelty, not just its invention.” In this context, the
agriculture.
Schumpeterian entrepreneur is the innovator who replaces
“Of all the inventions during the first half of the nineold ways of doing things with better ways of doing things, a
teenth century which revolutionized agriculture, the reaper
process that Schumpeter would describe later as “creative
was probably the most important,” wrote University of
destruction.”
Chicago historian William Hutchinson in his two-volume
So regardless of who invented the reaper, Hounshell
biography of McCormick in the 1930s. The reaper broke
contends that Cyrus was the Schumpeterian entrepreneur
the harvest-labor bottleneck by allowing the farmer “to reap
whose insights and efforts led to its widespread adoption. As
as much as he could sow.” This big step toward automation
early as the 1840s, Cyrus promoted the reaper with sophistiallowed farms to become larger and more productive. In
cated use of advertising and publicity. He moved to Chicago
turn, the mechanization of agriculture accelerated indusin 1847 to better serve the emerging Midwestern market.
trialization and urbanization as displaced workers migrated
Then he assembled a large and effective sales network and
more rapidly from farms to factories.
equipped it with slick catalogs, posters, and other promoThe traditional story of the McCormick reaper begins
tional items. He capitalized on international marketing
with Cyrus’ father, Robert McCormick, who had been
opportunities, and he eventually helped bring state-of-thetrying to develop a workable reaper for several years
art manufacturing to the Midwest.
at Walnut Grove, the family’s plantation in
Rockbridge County, Va. After Robert abandoned the project in 1831, young Cyrus started
building a reaper based on a different principle. Within six weeks, he successfully demonstrated his machine by harvesting oats at nearby
Steele’s Tavern.
For many years, Cyrus was acclaimed nationally and internationally as the singular inventor
of the reaper. But some historians have said
that Hussey’s contributions may have been just
as important — perhaps more important — to
the technological evolution of the machine. And
as far back as the 1870s, some members of the
McCormick family have argued that most of
the credit for inventing the reaper should go to
Robert McCormick.
But the long-standing debate over who This illustration from 1887 is typical of McCormick Harvesting Machine Co.
invented the reaper obscures a more important advertising that claimed Cyrus invented a reaper that worked well in 1831.
E CO N F O C U S | T H I R D/ F O U RT H Q U A RT E R | 2 0 1 6

27

Slow Adoption?
Given Cyrus’ entrepreneurial prowess and the obvious utility
of the reaper, economists and historians have wondered why
farmers were slow to adopt the machine. Hussey patented
his reaper in 1833, and McCormick followed in 1834, but
farmers didn’t start purchasing the machines in large numbers until the mid-1850s.
The traditional explanation for this surge in sales was the
rapid rise of global wheat prices during the Crimean War,
which limited grain exports from Russia and other nations in
the Black Sea region. But in the 1960s, Stanford University
economist Paul David offered another primary explanation:
He argued that before the mid-1850s, most American farms
were simply too small to make reapers practical.
The average farm size was growing, however, as grain
production shifted from the East to the Midwest, where
arable land was fresh, fertile, and relatively flat. More importantly, the farm-size threshold for the reaper to be practical
was declining as the price of labor — relative to the price of
reaping machines — increased in the Midwest due to higher
demand for workers to build railroads and other infrastructure throughout the fast-growing region, David wrote.
In the 1970s, Alan Olmstead, an economist at the
University of California, Davis, agreed that factor prices and
farm sizes were important, but he argued that the breakeven analysis for purchasing a reaper should be based on the
total acreage of grain to be cut by that machine — not just
by the grain acreage on the farm of the reaper’s prospective
owner. Farmers often cooperated to use reapers on multiple
farms, a possibility that David had excluded from his model.
Olmstead also faulted David for assuming that there were
no significant advances in reaper technology between 1833
and the 1870s. This assumption that the reaper was born
fully developed grew into a “historical fact,” Olmstead wrote,
even though it ignored “extremely knowledgeable historians who emphasized how a host of technological changes
transformed an experimentally crude, heavy, unwieldy, and
unreliable prototype of the 1830s into the relatively finely
engineered machinery of the 1860s.”
The idea that the reaper was born fully developed was
promoted aggressively by the McCormick Harvesting
Machine Co. as part of a long-term branding strategy based
on the sole-inventor legend of Cyrus. Over the years, many
of the company’s distortions and exaggerations came to be
accepted as historical facts, according to Daniel Ott, a visiting professor of history at the University of Wisconsin, Eau
Claire. In particular, the company claimed that Cyrus’ invention “signaled a monumental jump forward in the progress of
civilization and the circumstances of farmers everywhere,”
Ott wrote. But in reality, the McCormick reaper of 1831 was
not a monumental jump; it was only Cyrus’ first step as the
reaper’s Schumpeterian entrepreneur.

McCormick vs. Hussey
While the McCormicks were improving their machine at
Walnut Grove, Hussey was inventing his mechanized reaper
28

E CO N F O C U S | T H I R D/ F O U RT H Q U A RT E R | 2 0 1 6

in Baltimore. He demonstrated his machine during the harvest of 1833 and patented it in December of that year.
He sold at least one reaper in 1834, and by the end of the
decade, he was producing as many as 10 per year. In sharp
contrast, the McCormicks sold no reapers in the 1830s —
except one machine they had to take back from a dissatisfied
customer.
Cyrus finally sold two reapers in 1840, but he later
admitted that they were not very useful. By then, Hussey’s
machines were operating in at least eight states, according to Hutchinson. But Hussey’s reapers encountered
problems, too. “Some farmers complained that Hussey’s
machine left too long a stubble and others that the cutter clogged in damp grain and would not reap when the
stalks were bent away from the knife,” Hutchinson wrote.
Hussey’s sales plummeted in 1840 after his attempts to
improve the machine made it worse.
The McCormicks sold two more reapers in 1841, seven in
1842, and 29 in 1843. In June of that year, Cyrus and Hussey
demonstrated their reapers in a head-to-head competition on
the plantation of Ambrose Hutcheson near Richmond, Va.
The judges of the contest wrote that “both [reapers] performed
most admirably.” They expressed “great reluctance in deciding
between them,” but they generally preferred the McCormick.
Cyrus sold more reapers than Hussey that year, but
the quality of the McCormick reaper declined after Cyrus
increased production by selling manufacturing rights. Some
of his licensees performed poorly, and the quality of the reapers made at Walnut Grove fell dramatically in 1846 and 1847,
probably due to the illness and death of Robert McCormick.
By then, Cyrus was spending most of his time in the
Midwest, where demand for reapers was growing quickly.
For the rest of the decade, Cyrus focused on the Midwest,
while Hussey concentrated on the East. But their rivalry
shifted to the U.S. Patent Office in 1848, Hutchinson wrote,
as both inventors tried to extend their rights. “The expiration of their monopolies invited new competitors to enter
the arena, and the duel of the years 1839 to 1847 rapidly
became thereafter a general melee.”
True to his word in his 1834 letter to Mechanics’ Magazine,
Cyrus sued many of those new competitors for infringing on
his various patents. He didn’t win all of those lawsuits, but he
seemed to thrive on head-to-head competition — in courtrooms, in wheat fields, and at international exhibitions. In
sharp contrast, these contests seemed to wear Hussey down.
“I never experienced half the fatigue in Rowing after a
whale in the Pacific Ocean (which I have often done) as I
experienced year after year for eighteen years in the harvest field,” Hussey wrote in an 1854 letter. “No man knows
how much I have suffered in body and mind since 1833, on
account of this thing.”
A train ran over Hussey in 1860, one year before his patent rights were extended posthumously. Cyrus’ rights were
not extended, although the reasons for this ruling may have
had more to do with politics than the merits of the case,
according to Hounshell.

McCormick vs. McCormick
Cyrus lost some legal and political battles, but he won consistently in the marketplace. By all accounts, he was tenacious,
innovative, and farsighted as an entrepreneur. Perhaps his best
strategic decision was moving to Chicago in 1847. His youngest brother, Leander, joined him there in 1848, and another
younger brother, William, followed about one year later.
The brothers manufactured and sold more than 5,000
reapers in 1859, the year when Leander and William became
minority partners in Cyrus’ company. Hutchinson notes
that Cyrus “customarily found harmony impossible with his
partners,” and his brothers were no exceptions.
After William’s death in 1865, Leander and Cyrus quarreled more frequently. Each year, they argued about how
many reapers to produce for the upcoming harvest. Cyrus,
the risk-taking marketing maven, wanted to expand as
quickly as possible in the United States and abroad. Leander,
the risk-averse factory superintendent, wanted to grow
slowly in the United States. As the company’s majority
partner, Cyrus always opted for aggressive growth with little
regard for Leander’s objections. The younger brother also
became increasingly frustrated that Cyrus was getting all
the credit for the company’s success and all the glory for
inventing the reaper. Leander started to assert — privately
at first — that their father, Robert McCormick, was the true
inventor of the machine.
Hutchinson and Hounshell attribute Leander’s reaper
reversal to jealousy, but Ott believes Leander really was
trying to set the record straight. According to Ott, Leander
probably tolerated the singular-invention legend for many
years because he viewed the story as nothing more than
harmless advertising fluff. But Leander’s tolerance waned
when he realized the company was transforming the legend
into “the concrete narrative of the invention of the reaper.”
In the 1870s, Leander started gathering statements from old
friends and relatives back in Virginia to support his claim
that Robert had invented the reaper.
Meanwhile, adulation rained upon Cyrus, particularly
in France, where he was made an officer of the Legion of
Honor and a member of the French Academy of Sciences for
having “done more than any other living man for the cause of
agriculture in the world.”
Back at the factory, Leander was struggling to keep up
with Cyrus’ aggressive expansion plans and his promises to
customize reapers for smaller European markets. Hounshell
argues that Leander could not keep up because he had failed
to adopt modern manufacturing techniques, including the
use of jigs, fixtures, gauges, and single-purpose machines
to make interchangeable parts for standardized models.
“Leander, whose only experience had been as a country
blacksmith from Rockbridge County, Virginia, operated the
reaper works as though it were a large country blacksmith
shop,” Hounshell wrote in his 1984 book, From the American
System to Mass Production, 1800-1932.
Finally, Cyrus fired Leander and hired Lewis Wilkinson,
an experienced mechanic who was well-versed in modern

manufacturing techniques. After training under Wilkinson
for one year, Cyrus McCormick Jr. took over as superintendent of the factory and implemented ambitious plans to
modernize. Capacity quickly increased to 54,000 machines
in 1884 and more than 100,000 machines in 1889.
“Had Leander and Cyrus not had an irreparable fight in
1879-80, the reaper works might not have undergone any
notable changes until Cyrus’ or Leander’s death,” Hounshell
wrote. Cyrus died in 1884, and Leander died in 1900, but the
family feud over who invented the reaper was passed down
from generation to generation.

Manufacturing History
True or not, the singular-invention legend was valuable to
the McCormick Harvesting Machine Co. — not for patent
purposes by the 1880s, but to bolster the company’s standing
with populist farmers (reaper customers) who tended to hate
big business.
To justify its higher prices, the company began to portray Cyrus Sr. as a heroic farmer whose mechanical genius
had made him a great benefactor of mankind in general and
farmers in particular. According to the ever-expanding legend, Cyrus Sr. fed the hungry around the world (by making
bread cheaper) and elevated farmers from simple sodbusters
to sophisticated managers of employees and capital.
Ott documented these exaggerations in his 2015 dissertation, Producing a Past: Cyrus McCormick’s Reaper from
Heritage to History. Ott argued that the company used
the sole-invention legend to draw parallels between the
populist “labor theory of value” and the company’s “technological surplus value ideology.” The propaganda reached
a crescendo at the 1893 World’s Columbian Exposition in
Chicago, where a large banner over the company’s exhibit
proclaimed that “all harvesters of to-day are based upon
the features C.H. McCormick invented and built in 1831.”
McCormick’s competitors quickly complained that this
claim was patently false, and the Inventors’ Congress, an
international group that was acting as the exhibition’s jury,
“forced the McCormick Harvesting Machine Company to
take down all of its placards claiming inventive priority,”
Ott wrote.
Undaunted, Cyrus Jr. lobbied the U.S. Treasury
Department to get his father’s image printed on the $10
silver certificate. Treasury Secretary John Carlisle embraced
the idea and unveiled an engraving of the proposed new
currency in 1896. But he pulled the plug on “McCormick
money” after the company’s competitors vigorously challenged the story that Cyrus alone had invented the reaper.
This time, the challenge to the singular-invention legend was more public and more damaging to the company’s
reputation, according to Ott. This embarrassing loss of
prestige came at a difficult time. Grain prices were falling,
farmers were struggling, and the company’s farm machinery
sales were dwindling. After waging a five-year price war, the
company merged with its four largest competitors in 1902 to
form International Harvester.
E CO N F O C U S | T H I R D/ F O U RT H Q U A RT E R | 2 0 1 6

29

The merger agreement called for J.P. Morgan and Co. to
manage International Harvester for 10 years, but when the
McCormick family wrested control of the company away
from the other partners in 1912, Cyrus Jr. reasserted the
legend to help fend off federal antitrust charges. The company never got Cyrus Sr.’s image printed on currency, but a
depiction of a mid-19th century reaper graced the back of
the Federal Reserve’s first $10 note in 1914.

Entrepreneurial Power
Separating fact from fiction in the Cyrus McCormick
legend is difficult — if not impossible — because there are
no contemporary accounts of what happened at Walnut
Grove during the harvest of 1831. Most of that early history
is based on the recollections of Cyrus himself and other
highly partisan participants and observers — many of them
taking sides (sometimes switching sides) in patent disputes

and the McCormick family feud.
But based on overlapping information from sources cited
by both sides of the family, it seems likely that Cyrus and
Robert both contributed to the McCormick reaper of 1831.
And so did their slave, Jo Anderson, and so did a local blacksmith, John McCown. It also seems possible that Cyrus and
Robert obtained knowledge of previous attempts to develop
a practical reaper.
“One thing we know about the evolution of technology
in general is that almost never does an important technology
come out of the blue,” Hounshell says. “There are always
precedents. There are always theories that lead up to a
breakthrough invention.”
The more important question, according to Hounshell,
is who supplied the entrepreneurial power that brought the
reaper into common use? And the answer is clearly Cyrus
McCormick.
EF

Readings
Hounshell, David. From the American System to Mass Production,
1800–1932. Baltimore: Johns Hopkins University Press, 1984.

Lyons, Norbert. The McCormick Reaper Legend: The True Story of a
Great Invention. New York: Exposition Press, 1955.

Hutchinson, William. Cyrus Hall McCormick: Seed-Time,
1809–1856. New York: Century Company, 1930.

Ott, Daniel. Producing a Past: Cyrus McCormick’s Reaper from
Heritage to History. Dissertation, Loyola University Chicago, 2015.

Hutchinson, William. Cyrus Hall McCormick: Harvest, 1856–1884.
New York: D. Appleton-Century Company, 1935.

WHY AREN’T MORE WOMEN WORKING? continued from page 14
One reason for this, they wrote, is that the most robust
predictor of whether a woman will return to work late in
life is whether she had work experience early in her career.
So the fact that labor force participation is high for young
women — and that more and more of these women are college educated — suggests that, over time, they will return
to the workforce when they are older.
Whether — or how much — diminished female labor
force participation is a drag on U.S. growth is something
economists will continue to debate. In a 2015 report, the
OECD estimated that if American women caught up
to men in this respect by 2025, this could increase GDP
growth by 0.5 percentage point a year. But many scholars

caution that, on the other side of the ledger, it’s hard to
quantify the economic contribution of unpaid work such
as care-taking and household chores that is done by people
not in the labor force. Accordingly, such estimates may
not be clear-cut. Blau is among those, and she cautions
that the question of economic impact isn’t a “strictly
mechanical” one.
“The broader question is whether people with skills
and education are contributing to the economy as much
as they can or want,” Blau adds. “You need to factor in the
reasons for nonparticipation. And here, the data suggest
the United States is not offering the fullest opportunity for
women to contribute.”
EF

Readings
Blau, Francine D., and Lawrence M. Kahn. “Female Labor
Supply: Why Is the United States Falling Behind?” American
Economic Review: Papers and Proceedings, May 2013, vol. 103, no. 3,
pp. 251-256.
Goldin, Claudia, and Lawrence F. Katz. “Women Working
Longer: Facts and Some Explanations.” National Bureau of
Economic Research Working Paper No. 22607, September 2016.
Goldin, Claudia ,and Josh Mitchell. “The New Life Cycle of
Women’s Employment: Disappearing Humps, Sagging Middles,

30

E CO N F O C U S | T H I R D/ F O U RT H Q U A RT E R | 2 0 1 6

Expanding Tops.” Journal of Economic Perspectives, Winter 2017,
vol. 31, no. 1, pp. 161-182.
Rossin-Slater, Maya, Christopher J. Ruhm, and Jane Waldfogel.
“The Effects of California’s Paid Family Leave Program
on Mothers’ Leave-Taking and Subsequent Labor Market
Outcomes.” National Bureau of Economic Research Working
Paper No. 17715, December 2011.
Thévenon, Olivier. “Drivers of Female Labour Force
Participation in the OECD.” OECD Social, Employment and
Migration Working Paper No. 145, May 2013.

BOOKREVIEW

The Web Taketh Away
VIRTUAL COMPETITION: THE
PROMISE AND PERILS OF AN
ALGORITHM-DRIVEN ECONOMY
BY ARIEL EZRACHI AND
MAURICE E. STUCKE
CAMBRIDGE, MASS.: HARVARD
UNIVERSITY PRESS, 2016, 248 PAGES
REVIEWED BY DAVID A. PRICE

F

rom its earliest days of commercial use in the 1990s
until today, the web has been almost universally
viewed as a boon to consumers — practically like
Santa Claus. It makes the competing sellers of an item easy
to find and makes comparison shopping trivial, certainly
next to the old-world alternatives of schlepping from store
to store or making phone calls. In the consensus view, the
hypercompetitive markets that result from the web’s landscape mean more choices and lower prices. From a consumer’s perspective, at least, what’s not to like?
Plenty, say Ariel Ezrachi and Maurice Stucke. Ezrachi, a
professor of antitrust law at Oxford University, and Stucke,
a law professor at the University of Tennessee, argue in their
book Virtual Competition that the web may giveth unto consumers, but it also taketh away. In particular, they contend,
our affection for online shopping has obscured a number of
latent dangers that the web poses to competitive markets
and consumer welfare. Foremost among these are invisible
collusion, price discrimination, behavioral discrimination,
and what they call the “frenemy” dynamic.
Collusion, according to Ezrachi and Stucke, can flourish
in the world of online commerce in several ways. First, intentional cartel behavior is easier among online competitors, in
part because online prices are highly visible — which means
conspiring firms can monitor each other reliably and automatically. Second, the rise of firms like Uber, which sets the
prices of numerous independent agents, means that a swath of
sellers — such as Uber’s drivers — do not compete with each
other on price. It’s a legal form of price-fixing. (To be fair,
Uber itself broke a governmentally organized cartel of sorts;
the regulated taxi drivers with whom Uber drivers share the
road don’t compete with each other on price, either.)
Ezrachi and Stucke also foresee threats to competition from the use of increasingly sophisticated pricing
algorithms, which autonomously set prices for the firm’s
offerings at their optimum levels. Even without any explicit
direction to refrain from undercutting rivals, the algorithms
might well arrive at such an outcome. “No one will be
tempted to improve their products, lower prices, or enter
new markets,” they argue, “because others will immediately
detect and punish this initiative.”

In addition, online commerce opens new frontiers for
price discrimination, in Ezrachi and Stucke’s view — that is,
charging different consumers different prices for the same
product based on their willingness to pay. Airfares that vary
with the date of purchase are an example. Online sellers can
readily use a consumer’s buying history, web behavior, and
other personal information to achieve more perfect price
discrimination. While the authors acknowledge that price
discrimination can be economically efficient, they believe it
may be unfair to consumers and may enable large, established
firms to erect barriers to entry.
Ezrachi and Stucke are also concerned by what they call
“behavioral discrimination,” by which they mean using human
biases to steer consumers’ buying behavior. (Others have
called it “nudging.”) One example they give is a travel booking site leading some users toward more expensive hotels by
placing them higher in search results. Another is that of companies artificially increasing the complexity of buying options
to make comparison shopping harder. Much as price discrimination has a venerable history in the brick-and-mortar world,
behavioral discrimination is a descendant of the “motivational
research” vilified by Vance Packard in his 1957 bestseller The
Hidden Persuaders.
Finally, the authors warn of anti-competitive behavior among frenemies, firms that cooperate in some areas
of activity and fight in others. In particular, the rise of
so-called super-platforms, companies that provide platforms for other platforms — in the way that Apple and
Google provide a platform for Uber’s ride-sharing service
within their phone operating systems, or Amazon provides
a platform for third-party sellers — may lead to suppression
of competition. For example, if Apple enters the ride-sharing business itself in some fashion, it may wish to use its
power over its phone operating system to make Uber’s life
more difficult.
While Ezrachi and Stucke’s account is highly readable
and carefully researched, one does feel some cognitive dissonance when shifting from the pages of their book to the
actual online world. Simply put, if the largest online commerce firms are in fact exercising significant market power
in the economic sense, they aren’t acting like it. There can
be little doubt that markets with major online players have
become more competitive rather than less with the advent
of the web. The financials of many of these firms also seem
inconsistent with the idea that they are exercising great
market power. Amazon, for instance, has a lower net profit
margin than Walmart.
Ezrachi and Stucke might argue that the online firms are
just engaging in temporary strategic behavior. Maybe. But
without a way to tell the difference, their argument remains
speculative — even if it’s interesting speculation.
EF
E CO N F O C U S | T H I R D/ F O U RT H Q U A RT E R | 2 0 1 6

31

DISTRICTDIGEST

Economic Trends Across the Region

Social Networks and Economic Outcomes:
Evidence from Refugee Resettlement Programs
BY S A N T I AG O M . P I N TO

M

any of an individual’s decisions are influenced by
the group of people with whom he or she interacts. Friends, neighbors, classmates, co-workers,
and other social contacts are believed to play a fundamental
role in one’s decision to study, work hard, or commit a
crime. They are also thought to play a role in outcomes
such as the likelihood of finding a job. Identifying and
quantifying such effects is challenging, however.
Economists have adopted different approaches to studying how interactions through social networks affect individual outcomes. Traditionally, the neighborhood has been
used as the unit of analysis, on the assumption that the
neighborhood is where most social interactions happen.
Recent work has studied neighborhood effects by relying on
information collected from refugee resettlement programs.
The idea is that the social and economic prospects of newly
arrived refugees, such as the probability of finding a job, can
be attributed to the neighborhood characteristics where the
refugees end up residing.
Robert McKenzie, a visiting fellow at the Brookings
Institution, has observed, “Refugees don’t just come to
nations; they move to cities.” Cities play an undeniable role
in the resettlement of refugees and on their long-run social
and economic prospects. This statement, however, can be
narrowed down even further: Refugees actually move not
only into cities, but also into neighborhoods.
Most refugee resettlement programs around the world
are intended to help refugees make a smooth economic
transition into their new communities. Understanding how
social interactions operate is, therefore, key to evaluating
the effectiveness of those programs. Insights from the
research on neighborhood effects are valuable to the extent
that they may contribute to the design and implementation
of effective immigration and refugee policies. This has
become an extremely sensitive issue considering the number
of individuals fleeing their home countries worldwide has
recently reached record numbers.

Quantifying the Effect of Social Networks
The social and economic outcomes for refugees who settle
in new locations in a country depend on a variety of forces.
Recent academic work has focused on the influence of
social interactions at the neighborhood level. A long strand
of the literature has examined how neighborhood characteristics affect labor market prospects, education and
health outcomes, and criminal activities of residents.
For researchers, identifying and quantifying the effects
of social interactions on individual behavior are made more
32

E CO N F O C U S | T H I R D/ F O U RT H Q U A RT E R | 2 0 1 6

difficult by multiple causation. Any attempt to do so must
take into account the fact that households with different
characteristics commonly sort themselves into different
types of locations. Suppose that one would like to examine
whether residing in a deprived neighborhood (for example,
a neighborhood with a high unemployment rate) affects
a resident’s labor market opportunities. To quantify the
impact of the neighborhood on individual outcomes, the
researcher has to take into account that this type of neighborhood might attract individuals with characteristics that
would make him or her less likely to find a job. For instance,
individuals who select to reside in those high unemployment neighborhoods may tend to be low-skill workers or
are already unemployed. If this is the case, poor neighborhoods and poor labor market outcomes will be positively
associated. But it is not necessarily correct to conclude
that neighborhood characteristics are the cause of the poor
outcomes. In order to assess how the neighborhood affects
individual outcomes and to determine the precise causality,
an exogenous or random allocation of individuals across
neighborhoods is required.
To overcome this problem, some novel research has
used data collected through “social experiments.” In a social
experiment, individuals or households are randomly assigned
into two groups: a group that receives the treatment or participates in the program under study (the treatment group)
and another group that does not (the control group). An
advantage of this kind of approach — for example, when
evaluating the effect of neighborhoods on outcomes — is
that the assignment of individuals is random, so the differences across neighborhoods where people reside can be
reasonably viewed as exogenous. The experiment thus minimizes the chances of observing outcomes influenced by the
fact that some types of individuals or households may prefer
a neighborhood with certain characteristics.
Two main types of social experiments have received most
of the attention. The first one is the Moving to Opportunity
(MTO) experiment. MTO is a federal housing voucher program targeted to low-income households residing in poor
neighborhoods. This program offered housing vouchers to
randomly selected households residing in poor areas to pay
for their housing rents. Those vouchers, however, could only
be used in low-poverty neighborhoods. The experiment was
conducted in five cities (Baltimore, Boston, Chicago, Los
Angeles, and New York) from 1994 to 1998, and it intended
to study the social and economic effects on low-income
households from moving to low-poverty neighborhoods.
Other research has used data collected from refugee

250
200
THOUSANDS

resettlement programs. The idea is that since locations are not selected by refugees, the assignment
of refugees across
Numberdifferent
of Admittedlocations
Refugees is exogenous.
207 of this research on refugee resettleThe conclusions
159
ment may not98only help in the design and improve61 concerning refugees, an issue that
ment of policies
70
has received a68lot of attention in the last couple of
62
years worldwide,
65 but also shed light on how social
networks and76
neighborhoods affect individuals’ out107
122
comes in general.

U.S. Annual Refugee Resettlement Ceilings and the
Number of Refugees Admitted

150
100

grams should, therefore, take neighborhood effects into
consideration.
In the United States, the Refugee Act of 1980 sets the
foundation of the federal refugee resettlement program.
This program determines eligibility for refugee status, establishes admissions procedures, defines the type of assistance
granted to refugees, and provides guidelines concerning the
resettlement process. The United States has historically led
all nations in accepting and resettling refugees. Since the
beginning of the European refugee crisis in 2015, however,
other countries have been obligated to assume a much more
important role.
A maximum number of refugees are allowed to enter the
United States every year. This ceiling is determined by the
president in consultation with Congress. The highest annual
ceiling was set at 231,700 admissions in 1980. This number
has changed through the years for a variety of reasons,
including worldwide population migration, worldwide economic conditions, and domestic political factors. From 2001
until 2015, the ceiling has fluctuated between 70,000 and
80,000. In 2016, it was raised to 85,000, and the proposed
ceiling for 2017 is 110,000. The number of actual arrivals has
generally fallen below the ceiling; since 2013, however, it has
always reached the established maximum. (See chart.)
Federal law requires that refugee resettlement locations
should be decided by the federal government in consultation
with state and local governments. The federal government
currently works with nine agencies to provide assistance to
refugees throughout the resettling process. These agencies,
jointly with their local affiliates, determine the best locations
for the newly arrived refugees. The settlement decisions are

2014

2016

2012

2010

2008

2006

2002

2004

2000

1996

1998

1992

1994

1990

1988

1984

1986

1982

1980

50
113
133
0
119
Refugee Resettlement
Programs
113
100
in the United States
76
The design and
70 implementation of refugee resetNumber of Admitted Refugees
Annual Ceiling
77
tlement programs
vary across countries. In general,
86
NOTE: Since 2013, the number of admitted refugees has always reached the annual ceiling.
73 provide temporary assistance to
programs usually
SOURCE: Migration Policy Institute
70
newly arrived 27
refugees and provide support through28
out the settlement
process. The main feature of
53
54 is that the assistance is intended to help
most programs
typically driven by factors such as the number of refugees
41
refugees achieve
self-sufficiency and become integrated
already present in the community and family reunification
48
members of 60
motives. Other indicators that describe the community’s
75the community as soon as possible. After
receiving this73 initial support from the host government,
capacity to absorb refugees are also taken into account. The
56
their economic
latter includes availability of affordable housing, health and
58 success will, among other things, be tied to
70
the characteristics
of the place where they end up residing.
educational services, and employment opportunities.
70
70 evaluation of refugee resettlement proAn appropriate
From 2001 to 2016, approximately 890,000 refugees
85

were admitted into the United States. Eight states received
almost 50 percent of the total number of refugees in that
period. California and Texas are by far the two largest refugee
hosting states, receiving 11.5 percent and 9.1 percent, respectively, of total refugees. The list continues with New York,
Florida, Minnesota, Washington, Arizona, and Michigan,
each state accounting for about 4 percent to 6 percent of the
total number of refugees. (See chart.) In the Fifth District,
North Carolina has hosted about 3 percent, while Virginia

Refugee Settlement by State, 2001-2016
California; 11.5%

Texas; 9.1%

New York; 6.0%

Other States; 50.5%

Florida; 5.2%
Minnesota; 4.6%
Washington; 4.5%
Michigan; 4.3%

Arizona; 4.4%

NOTE: The values indicate the number of refugees in the respective state as a
percentage of the number of refugees admitted in the USA throughout the period
2001-2016.
SOURCE: Refugee Processing Center 				

E CO N F O C U S | T H I R D/ F O U RT H Q U A RT E R | 2 0 1 6

33

Neighborhood Effects on Labor
Market Outcomes

Fifth District Refugee Admissions by City, 2001-2016

Part of the literature on refugee resettlement focuses on how the charNumber
City
Percent
City
Percent
City
Percent
acteristics of the community affect
labor market outcomes for newly
1
Baltimore
46.2 Charlotte
26.6 Richmond
17.0
arrived refugees. Immigrants, and
2
Silver Spring
20.9 Greensboro
17.0 Charlottesville
13.7
particularly refugees, tend to concentrate in certain areas and reside
3
Riverdale
5.0 Raleigh
15.7 Roanoke
11.7
in enclaves. Such location decisions
4
Frederick
2.6 High Point
12.5 Harrisonburg
9.4
may have both positive and negative
5
Hyattsville
1.8 Durham
7.9 Falls Church
7.4
implications regarding labor market
outcomes. On one hand, labor market
6
Rockville
1.6 New Bern
7.3 Hampton
5.9
prospects may improve because indi7
Hagerstown
1.4 Asheville
2.3 Newport News
5.5
viduals may share information about
job opportunities with other net8
Columbia
1.2 Wilmington
1.9 Alexandria
5.2
work members more effectively. On
9
Gaithersburg
1.2 Carrboro
1.7 Fredericksburg
2.9
the other hand, living in an enclave
10
Elkridge
1.2 Chapel Hill
1.2 Arlington
2.1
may reduce the incentives to acquire
certain required skills (for example,
Total
83.0
93.9
80.9
the development of language skills)
NOTE: The values indicate the number of refugees in the city as a percentage of the number of refugees who settled in
to become fully integrated into the
the state during the period.
host’s labor market.
SOURCE: Refugee Processing Center 					
Early work by Per-Anders
Edin and Olof Aslund of Uppsala
and Maryland have received about 2 percent of total refugees
University and Peter Fredriksson of Stockholm University
during the period. The percentages for South Carolina, West
examined which of these two effects tends to dominate.
Virginia, and Washington, D.C., are negligible.
In a 2003 article in the Quarterly Journal of Economics, they
The assignment of refugees across cities within each of
looked at the extent to which ethnic concentration in a
the states in the Fifth District that have hosted the largest
city affects earnings of refugees from the same country of
number of refugees widely differs during 2001-2016. (See
origin residing in those areas. They used data from a refugee
table.) In Maryland, almost half of the refugees resettled in
settlement program implemented in Sweden between 1985
the state have located in Baltimore, while the assignment of
and 1991. The conclusions of their analysis suggested that as
refugees in North Carolina and Virginia seems to be more
the size of the ethnic concentration rises, earnings increase
dispersed across cities. Charlotte (with almost 27 percent) and
as well. In fact, they showed that earnings increase more for
Richmond (with 17 percent) attract the highest proportion
low-skill individuals.
of refugees in North Carolina and Virginia, respectively, but
Yet their results indicated that the effect on earnings
they are followed in each case by Greensboro (17 percent) and
actually depends on the “quality” of the enclave: Individuals
Charlottesville (13.7 percent).
who belong to an ethnic group with higher earnings or
Occasionally, state and local officials have opposed the
higher self-employment rates have a higher return from
resettlement in their districts. For instance, a number of
residing in the enclave. Those who belong to enclaves that
state government officials have recently indicated they will
have a lower than average level of earnings may actually expenot allow the settlement of Syrian refugees in their states.
rience a negative impact on earnings.
It should be noted, however, states that have historically
More recent work by Anna Damm of Aarhus University
accepted large number of refugees (such as California, New
investigated a similar issue using data on a refugee resettleYork, Washington, and Pennsylvania, among others) have
ment program in Denmark. Her main objective was to examassured their continued participation in resettlement proine whether residing in a deprived neighborhood negatively
grams. Moreover, there seems to be conflicting opinions
affects labor market outcomes for refugees. In a 2014 article
declared by state governors and city officials. In fact, many
in the Journal of Urban Economics, she found that after accountcities in states that oppose the new admission of refugees
ing for residence sorting, such an effect is nonexistent. Her
advocate for a higher participation in resettlement programs
work concluded, along the same line as Edin, Fredriksson,
and welcome even larger number of refugees into their cities.
and Aslund, that the quality of the network, rather than its
Further opposition to refugee resettlement has recently
size, is more important for explaining individuals’ labor maremerged at the federal level. President Trump has sought to
ket outcomes. In fact, the probability that a newly arrived
suspend the admission of refugees; at press time, the legality
refugee finds a job improves as the employment rate among
of that measure is a subject of litigation.
co-nationals who reside in close proximity is higher.
Maryland

34

North Carolina

E CO N F O C U S | T H I R D/ F O U RT H Q U A RT E R | 2 0 1 6

Virginia

Lori Beaman of Northwestern University provides an
alternative view in which the effects of a larger network
might depend on the specific structure and composition
of the network. Beaman developed a model that captures
how information is transmitted through the network.
She used data from refugee programs administered by
the International Rescue Committee that assigned refugees across various cities in the United States during the
period 2001-2005. When examining the labor outcomes for
recently arrived refugees, she found that their labor market
outcomes (described mostly by the probability of employment and the level of wages) tended to be worse when the
number of network members resettled in the same year or
one year prior is larger.
Beaman found, however, that the outcomes are better for
newly arrived refugees when they interact and participate in
networks with a larger number of members with longer tenure in the United States. One possible interpretation of this
result is that newly arrived refugees compete for the same
type of jobs with other refugees that have recently relocated
into the United States. As a result, this latter group might
not find it beneficial to share and transmit information to
the newly arrived refugees about job opportunities through
the network. On the other hand, more tenured members,
typically members who already have an established job,
would feel less threatened by the arrival of new refugees, and
they would behave more cooperatively.

activities. They concluded, using data from the refugee
settlement program implemented in Denmark, that the
exposure to neighborhood crime during childhood influences the criminal behavior of individuals as adults. More
precisely, they found that as the percentage of convicted
criminals residing in a neighborhood rises, it becomes
more likely for male refugees assigned to that neighborhood to engage in crime later in life. This effect is not
observed for females, though.

Effect of Refugee Dispersal Policies on Earnings
Some countries, such as the United Kingdom, Germany,
and Sweden, follow strict settlement policies that restrict
the locations where newly arrived refugees can reside. One
of the main goals of those polices is to reduce the concentration of refugees in a small number of densely populated
cities. This objective is presumably based on the idea that
higher concentrations of refugees in an area may reduce the
level of integration and assimilation of immigrants.
Moreover, it has been claimed that refugees tend to
impose, at least initially, a heavy fiscal burden on recipient
cities. A discussion paper from the Brookings Institution
prepared by Bruce Katz, Luise Noring, and Nantke Garrelts
reviewed the recent refugee experience in Europe. The
report highlighted the fact that refugees often disproportionally locate in a small number of cities. Such a settlement pattern has created important local fiscal imbalances,
since the cities ultimately bear the cost of educating and
integrating the newly arrived refugees into their communities. Refugee dispersal policies may be viewed as a way of
spreading out and sharing the fiscal burden among several
localities.
A few papers that evaluate the effectiveness of refugee
policies suggest, however, that dispersing refugee immigrants across cities may have a detrimental effect on refugees. Edin, Fredriksson, and Aslund, in a 2004 study, found
that settling refugee immigrants away from denser areas
results in an important long-run earning loss for those immigrants. The goal of dispersing refugees, they concluded, is
attained at a significant cost for the refugees, hurting their
ability to become self-sufficient.
Many countries are making a great effort to deal with
the rising number of displaced individuals around the world.
Understanding the factors that determine the long-run outcomes of refugees, including their self-sufficiency and degree
of integration in the host country, is key to evaluating the
effectiveness of refugee resettlement programs. The academic research reviewed above may provide some guidelines
on how to design and implement these policies.
EF

Neighborhood Effects on Education
and Criminal Behavior
Other work focuses on different aspects of neighborhood
effects, such as their impact on education outcomes and the
likelihood of engaging in criminal activities. In a paper published in the American Economic Journal: Applied Economics
in 2011, Aslund, Edin, Fredriksson, and Hans Gronqvist
found that the “quality” of the network connections helps to
explain the education performance of refugees, in line with
the conclusions of their previous research that focused on
labor market outcomes. Specifically, their work showed that
education outcomes, measured by students’ school grades,
improve when the proportion of highly educated peers in
the same local ethnic group is higher. They also showed that
the positive effects are more important for those kids who
arrived in the neighborhood when they were younger (less
than 7 years old).
Research by Anna Damm and Christian Dustmann
of University College London studied the connection
between the level of crime at the neighborhood level and
the probability of individuals later engaging in criminal

u

E CO N F O C U S | T H I R D/ F O U RT H Q U A RT E R | 2 0 1 6

35

State Data, Q2:16
DC

MD

NC

SC

VA

WV

Nonfarm Employment (000s)
778.1
2,707.1
4,315.9
2,045.5
3,908.1
764.2
Q/Q Percent Change
0.4
0.6
0.3
0.6
-0.1
0.6
Y/Y Percent Change
1.5
2.0
2.1
2.6
2.0
-0.2
							
Manufacturing Employment (000s)
1.2
106.4
458.4
239.4
228.9
47.1
Q/Q Percent Change
0.0
-0.5
-0.4
0.9
-1.2
-1.1
Y/Y Percent Change
9.1
2.6
-0.4
1.6
-1.7
-1.2
					
Professional/Business Services Employment (000s) 163.9
438.3
613.0
274.8
714.0
66.2
Q/Q Percent Change
0.8
1.0
1.4
3.6
-0.6
-0.6
Y/Y Percent Change
1.3
2.2
4.8
5.6
3.0
-0.7
							
Government Employment (000s)
242.4
503.4
725.1
363.9
712.5
156.4
Q/Q Percent Change
1.0
0.2
0.2
0.3
0.1
2.8
Y/Y Percent Change
1.9
0.0
0.6
1.2
0.1
2.9
						
Civilian Labor Force (000s)
397.3
3,172.4
4,866.0
2,313.9
4,215.8
781.3
Q/Q Percent Change
1.1
0.1
0.8
1.1
-0.7
-0.6
Y/Y Percent Change
2.5
0.9
2.4
2.8
-0.2
-0.4
							
Unemployment Rate (%)
6.1
4.5
5.1
5.6
3.8
6.2
Q1:16
6.5
4.7
5.5
5.6
4.1
6.4
Q2:15
7.0
5.2
5.8
6.1
4.5
7.1
				
Real Personal Income ($Bil)
46.3
314.0
383.6
176.2
406.7
61.6
Q/Q Percent Change
0.9
1.0
0.7
0.7
0.6
-0.1
Y/Y Percent Change
3.1
2.6
3.2
3.4
2.3
-0.6
							
Building Permits
1,315
5,596
15,114
8,833
8,280
779
Q/Q Percent Change
0.0
59.7
31.6
29.6
26.7
53.0
Y/Y Percent Change
0.0
14.6
6.4
-0.6
-4.5
-12.7
							
House Price Index (1980=100)
795.4
446.1
340.0
346.9
433.6
230.7
Q/Q Percent Change
2.1
1.2
1.7
1.5
1.7
1.7
Y/Y Percent Change
8.9
2.7
5.1
5.6
3.0
1.6
NOTES:

SOURCES:

1) FRB-Richmond survey indexes are diffusion indexes representing the percentage of responding
firms reporting increase minus the percentage reporting decrease. The manufacturing composite
index is a weighted average of the shipments, new orders, and employment indexes.
2) Building permits and house prices are not seasonally adjusted; all other series are seasonally
adjusted.
3) Manufacturing employment for DC is not seasonally adjusted

Real Personal Income: Bureau of Economic Analysis/Haver Analytics.
Unemployment Rate: LAUS Program, Bureau of Labor Statistics, U.S. Department of Labor/Haver
Analytics
Employment: CES Survey, Bureau of Labor Statistics, U.S. Department of Labor/Haver Analytics
Building Permits: U.S. Census Bureau/Haver Analytics
House Prices: Federal Housing Finance Agency/Haver Analytics

For more information, contact Michael Stanley at (804) 697-8437 or e-mail michael.stanley@rich.frb.org

36

E CO N F O C U S | T H I R D/ F O U RT H Q U A RT E R | 2 0 1 6

Nonfarm Employment

Unemployment Rate

Real Personal Income

Change From Prior Year

Second Quarter 2005 - Second Quarter 2016

Change From Prior Year

Second Quarter 2005 - Second Quarter 2016

4%
3%
2%
1%
0%
-1%
-2%
-3%
-4%
-5%
-6%

Second Quarter 2005 - Second Quarter 2016

10%
9%
8%
7%
6%
5%
4%
06 07 08 09 10 11

12

13

14 15

16

3%

06 07 08 09 10 11

12

13

14 15

Fifth District

16

8%
7%
6%
5%
4%
3%
2%
1%
0%
-1%
-2%
-3%
-4%
-5%

06 07 08 09 10 11

12

13

14 15

16

14 15

16

United States

Nonfarm Employment
Major Metro Areas

Unemployment Rate
Major Metro Areas

Building Permits

Change From Prior Year

Second Quarter 2005 - Second Quarter 2016

Second Quarter 2005 - Second Quarter 2016

Change From Prior Year

Second Quarter 2005 - Second Quarter 2016

7%
6%
5%
4%
3%
2%
1%
0%
-1%
-2%
-3%
-4%
-5%
-6%
-7%
-8%

06 07 08 09 10 11
Charlotte

12

13

Baltimore

14 15

16

13%
12%
11%
10%
9%
8%
7%
6%
5%
4%
3%
2%
1%

Washington

30%
20%
10%
0%
-10%
-20%
-30%
-40%
06 07 08 09 10 11
Charlotte

12

13

Baltimore

14 15

FRB—Richmond
Manufacturing Composite Index

Second Quarter 2005 - Second Quarter 2016

Second Quarter 2005 - Second Quarter 2016

30

30

20

20

10

16

-30

-20

-40
06 07 08 09 10 11

12

13

14 15

16

-50

06 07 08 09 10 11

12

13

14 15

13

United States

Change From Prior Year
Second Quarter 2005 - Second Quarter 2016

-20

-10

12

House Prices

-10

0

06 07 08 09 10 11
Fifth District

0

10

-50%

Washington

FRB—Richmond
Services Revenues Index
40

-30

40%

16

16%
14%
12%
10%
8%
6%
4%
2%
0%
-2%
-4%
-6%
-8%

06 07 08 09 10 11
Fifth District

12

13

14 15

16

United States

E CO N F O C U S | T H I R D/ F O U RT H Q U A RT E R | 2 0 1 6

37

Metropolitan Area Data, Q2:16
Washington, DC

Baltimore, MD

Hagerstown-Martinsburg, MD-WV

Nonfarm Employment (000s)
2,653.5
1,402.2
106.1			
Q/Q Percent Change
2.0
3.0
3.5			
Y/Y Percent Change
2.4
2.3
0.9			
						
Unemployment Rate (%)
3.7
4.6
4.5			
Q1:16
4.0
4.8
4.8			
Q2:15
4.5
5.5
5.7			
					
Building Permits
7,742
2,087
253			
Q/Q Percent Change
51.0
52.6
27.8			
Y/Y Percent Change
7.9
-13.5
-18.1			
					
		
Asheville, NC
Charlotte, NC
Durham, NC
Nonfarm Employment (000s)
185.8
1,134.0
301.2			
Q/Q Percent Change
1.7
1.9
1.3			
Y/Y Percent Change
1.6
2.6
1.7			
						
Unemployment Rate (%)
4.0
4.8
4.5			
Q1:16
4.3
5.1
4.9			
Q2:15
4.8
5.6
5.1			
						
Building Permits
608
4,458
1,045			
Q/Q Percent Change
37.9
9.7
-22.5			
Y/Y Percent Change
-3.0
-10.8
37.9			
					
Greensboro-High Point, NC

Raleigh, NC

Wilmington, NC

Nonfarm Employment (000s)
361.0
598.5
121.5			
Q/Q Percent Change
1.3
1.5
3.5			
Y/Y Percent Change
1.3
3.2
0.9			
						
Unemployment Rate (%)
5.1
4.3
4.8			
Q1:16
5.6
4.6
5.3			
Q2:15
6.1
4.9
5.6			
					
Building Permits
1,017
4,201
509			
Q/Q Percent Change
79.4
99.9
28.5			
Y/Y Percent Change
80.0
17.4
57.1			
		
NOTE:

Nonfarm employment and building permits are not seasonally adjusted. Unemployment rates are seasonally adjusted.

38

E CO N F O C U S | T H I R D/ F O U RT H Q U A RT E R | 2 0 1 6

Winston-Salem, NC

Charleston, SC

Columbia, SC

Nonfarm Employment (000s)
259.9
342.4
393.1		
Q/Q Percent Change
0.8
2.3
1.1		
Y/Y Percent Change
0.7
2.4
2.6		
					
Unemployment Rate (%)
4.8
4.7
5.0		
Q1:16
5.2
4.8
5.1		
Q2:15
5.6
5.3
5.6		
					
Building Permits
635
2,013
1,232		
Q/Q Percent Change
130.1
36.7
26.4		
Y/Y Percent Change
18.0
16.4
-13.6		
					
				
Greenville, SC
Richmond, VA
Roanoke, VA
Nonfarm Employment (000s)
409.2
672.4
163.1		
Q/Q Percent Change
1.5
1.3
1.0		
Y/Y Percent Change
2.2
3.7
1.3		
					
Unemployment Rate (%)
4.8
3.8
3.6		
Q1:16
4.9
4.1
3.9		
Q2:15
5.5
4.8
4.6		
					
Building Permits
1,448
1,403
N/A		
Q/Q Percent Change
40.2
31.1
N/A		
Y/Y Percent Change
-6.2
1.7
N/A		
					
				
Virginia Beach-Norfolk, VA
Charleston, WV
Huntington, WV
Nonfarm Employment (000s)
773.2
123.0
142.4		
Q/Q Percent Change
2.3
2.0
2.3		
Y/Y Percent Change
0.6
-0.7
0.9		
					
Unemployment Rate (%)
4.4
5.7
6.1		
Q1:16
4.6
6.4
6.7		
Q2:15
5.1
6.8
6.5		
					
Building Permits
1,907
61
45		
Q/Q Percent Change
26.6
27.1
0.0		
Y/Y Percent Change
3.6
-12.9
-44.4		
					
			
For more information, contact Michael Stanley at (804) 697-8437 or e-mail michael.stanley@rich.frb.org
E CO N F O C U S | T H I R D/ F O U RT H Q U A RT E R | 2 0 1 6

39

OPINION

Immigration and the Economy
BY J O H N A . W E I N B E RG

I

n 1854, an editorial in the Philadelphia Sun worried that
“the enormous influx of foreigners will in the end prove
ruinous to American workingmen, by reducing the wages
of labor to a standard that will drive them from the farms
and workshops altogether.” Similar arguments have been
heard throughout the United States’ history, and are heard
today. Concern about the effect of immigration on native
citizens’ wages and employment, as well as the potential burden on public services, is understandable. But the extensive
economic literature on the subject suggests these concerns
may be overstated, and that immigration on net has positive
economic effects.
It’s true that the share of the population born outside
the United States has increased significantly in the past half
century, from less than 5 percent in 1970 to about 13 percent
today, including those who immigrate both legally and illegally. (The peak was nearly 15 percent in 1890.) There’s also
been a shift in immigrants’ country of origin since national
quotas were eliminated in 1965: In 1960, 84 percent of
immigrants were from Europe or Canada. Today, more than
three-quarters of immigrants are from Southeast Asia or
Latin America — 28 percent from Mexico alone.
Still, the number of unauthorized immigrants from
Mexico fell by more than a million after the 2007 peak of
6.9 million, according to estimates based on Census data.
That contributed to a drop in the total number of unauthorized immigrants from all countries, from 12.2 million in
2007 to 11.3 million in 2009. At least through 2014, the most
recent year for which data are available, the unauthorized
immigrant population was relatively flat, while the number
of authorized immigrants continued to increase.
How do unauthorized immigrants affect net spending
on public services? Looking at a large number of studies, the
answer appears to be not much. It’s estimated that between
50 percent and 75 percent of unauthorized immigrants pay
income taxes using either an Individual Tax Identification
Number or a false Social Security number. They also pay
property taxes; about one-third are homeowners, while others
pay indirectly through rent. Combined with sales taxes, these
payments help to offset federal, state, and local expenditures.
And when immigrants’ descendants are included in the analysis, the net fiscal impact may actually be positive. Researchers
also have found that immigrants pay more into Social Security
and Medicare than they receive in benefits, which may be
especially important as the U.S. population continues to age.
Unauthorized immigrants, who generally lack health
insurance, do impose costs on hospitals, which are obligated
to treat all emergency room visitors. But this arguably points
to broader flaws in our health care system, rather than to an
immigration problem per se.
40

E CO N F O C U S | T H I R D/ F O U RT H Q U A RT E R | 2 0 1 6

While many people think of immigrants working in lessskilled jobs, in fact, U.S. immigrants are over-represented
at both ends of the skill distribution. About one-third of
STEM workers with a Ph.D. are foreign born, as are about
40 percent of workers without a high school diploma. And
while there is a great deal of concern that immigrants —
less-skilled immigrants in particular — take jobs away from
natives, much empirical work shows that immigrants have
little effect on native employment. Immigrants, especially
those with less education, are more likely to compete with
other immigrants than with natives of the same skill level.
The effect on natives’ wages also is small and in some cases
slightly positive. This might seem counterintuitive — basic
supply and demand would suggest that wages go down when
there are more workers. But natives’ wages can increase to the
extent that less- and more-skilled jobs are complements. For
example, an increase in the supply of construction workers
increases the relative demand for construction managers, and
over time, natives tend to move into these higher-skill jobs. In
addition, immigrants are consumers as well as workers, which
can raise the local demand for labor.
Immigrants also increase the supply of, and lower the
prices for, some services, which boosts the real income of
natives. And those working in higher-skilled occupations contribute to long-run productivity gains and increased innovation; immigrants patent at about twice the rate of natives and
may have positive spillovers on natives’ innovation. And more
generally, faster population growth, whatever the source,
tends to be associated with faster productivity growth over
time. In addition, the aging population means that the growth
rate of the labor force is slowing, and the working-age population is declining as a share of the total population, which contributes to slower per capita GDP growth. More working-age
immigrants could help counteract this.
Of course, in the short run, there can be negative effects
on some native workers. But labor market disruptions due
to immigration are for the most part modest relative to the
disruptions that regularly occur in dynamic markets. And like
other disruptions, such as technological change, immigration
also brings long-term economic benefits. These benefits to the
host country — not to mention the benefits to the immigrants
themselves — suggest that the most efficient way to address
the distributional effects of immigration is not with barriers
but rather with workforce development policies that help
both current and future generations build up their own human
capital and expand their labor market opportunities.
EF
John A. Weinberg is senior vice president and special
advisor to the president at the Federal Reserve Bank
of Richmond.

NEXTISSUE
After the Foreclosures

Economic History

Millions of people lost their homes between 2007 and 2014.
Damage to a consumer’s credit takes seven years to clear — so
the homeowners caught up in foreclosures during the worst of
the mortgage crisis now have the damage to their credit behind
them. Observers expected a wave of “boomerang buyers” at this
point, but homeowners affected by the crisis have been slow to
return to the housing market.

“A child lives in a lead world,” wrote a
physician in 1924. Although that lead world
was known to be highly toxic by the early
1900s, it would be nearly eight decades
before the United States banned consumer
uses of lead paint. Throughout lead paint’s
history, children of lower socioeconomic
status have been at greater risk of poisoning.

Self-Driving Trucks

District Digest

Self-driving cars are getting the attention now, but self-driving
trucks are also on the way. What does this mean for labor
markets? Truck driving is perhaps the largest occupation in
which non-college-educated workers can attain middle-class
earnings. According to Bureau of Labor Statistics estimates, some
1.8 million Americans make their living as heavy and tractortrailer truck drivers.

Small Colleges

Small private colleges are under increasing strain as more
students opt for an urban learning environment — especially
where job opportunities are more plentiful. For instance, Sweet
Briar College recently teetered on the verge of closure until its
alumnae launched a successful financial rescue effort. Observers
are asking whether small colleges continue to make economic
sense. In the Fifth District, several schools are trying to buck the
trend by re-inventing their mission and their business model.

Business creation has been subdued recently,
with new startup activity remaining well
below pre-recession levels. Within the Fifth
District, which regions and sectors have
seen the greatest dampening in startup
activity, and what are the underlying factors
that may be bringing down new business
formation?

Interview
Janet Currie of Princeton University on
access to health care and safety net
programs, the economic and health effects
of pollution, and how prenatal exposures
and socioeconomic differences affect child
and adult health.

Visit us online:
www.richmondfed.org
•	To view each issue’s articles
and Web-exclusive content
• To view related Web links of
additional readings and
references
• To subscribe to our magazine
•	To request an email alert of
our online issue postings

Federal Reserve Bank
of Richmond
P.O. Box 27622
Richmond, VA 23261

Change Service Requested

To subscribe or make subscription changes, please email us at research.publications@rich.frb.org or call 800-322-0565.

NEW

Regional Report on Educational Attainment
2015

nt in Virginia

Educational Attainme

t from 1970 to 2015
Educational Attainmen

United States

Virginia

12.3
10.3

19.1

24.5

14.9

Higher
25.2
Degree or
Bachelor's
Total
Percent of

24.0

26.0

26.3

26.0

26.0

24.6

24.8

18.5

13.9

11.1

1990

2000

2010

2015

26.7
37.6

1980

1970

16.2

47.7

28.1

30.6

27.4

28.9

28.9

30.0

28.6

28.5

24.8

19.6

14.4

12.8

2000

2010

2015

33.5

1990

1980

1970

27.6

College

Some College

High School

l

24.4

24.9

34.6

27.3

Less Than High Schoo

20.3

15.7

31.1

28.4

52.2

37.0

33.8

29.5

10.7
10.6

nment in 2015
nt by Educational Attai
Earnings and Employme
Median Earnings

Educa1onal	
  

$55,799
ted	
  States	
   $50,930

by	
  Age	
  in	
  2015

AXainment	
  

Uni

Bachelor's Degree

Virginia	
  

18.4	
  

over	
  
65	
  years	
  and	
  

ate's 15.1	
  
e or Associ
23.7	
  
Some Colleg
28.4	
  
Degree
20.7	
  
or Equiva
High School26.8	
  

45	
  to	
  64	
  years	
  

10.1	
  

35	
  to	
  44	
  years	
  

8.8	
  

21.9	
  

25	
  to	
  34	
  years	
  

7.7	
  

21.8	
  

Total	
  

1.0	
  

11.0	
  

47.4	
  

17.7	
  

45	
  to	
  64	
  years	
  

11.7	
  

35	
  to	
  44	
  years	
  

10.4	
  

25	
  to	
  34	
  years	
  

8.8	
  

23.6	
  

25.0	
  
4.3

25.0	
  

2.6
2.9 19.8	
  

11.1	
  

24.2	
  

30.4	
  

9.8

24.8	
  

35	
  to	
  44	
  y

11.5	
  
13.1	
  
10.3	
  

23.8	
  
45.8	
  

30.4	
  

States
United9.5	
  
0.7	
  
14.9	
   n

Force Participatio
Labor24.5	
  

27.8	
  

27.9	
  
28.1	
  

27.1	
  

Some College or
27.6	
  
ate's Degree
11.9	
  
Associ

14.0	
  

24.8	
  
29.9	
  

21.6	
  

9.9	
  
15.2	
  

33.4	
  

High School
Less than
18.0	
  
Graduate
10.8	
  
8.7	
  

27.3	
  

21.2	
  
20.4	
  

11.5	
  

19.5	
  

11.6	
  

21.9	
  

12.2	
  

25.0	
  
31.6	
  
30.7	
  
32.9	
  

Rate (Percent)

18.7	
  
22.6	
  
25.8	
  

9.0	
  
11.6	
  

73.5
71.9
62.0
60.1

14.6	
  
12.3	
  
11.4	
   0.9	
  

49.0	
  
Other
White (not
Asian
27.3	
  
Hispanic or
Black
11.5	
  
1.4	
  
13.5	
   ic
or
Hispan
Latino
49.4	
  
131,525
9 ssional	
  
26.9	
  
Latino)
	
  or	
  Profe
361,65
000Graduate
8.8	
  
1,046,
9
's	
  
33.6
18	
  to	
  24	
  years	
  
3,753,743iate's	
   418,66Bachelor 15.6
10.2
Assoc
and Older)
28.9
24.6
College	
  or	
  8.0
15.1
Some	
  nt)
Total (Population 25
l	
   ate (Perce
29.4
	
  Schoo
Gradu
24.4
School
25.2
than
l	
   High High
24.4
Less
16.0
Schoo
32.1
Less	
  than	
  High	
  High School Graduate or Equivalent
11.8
es,	
  for	
  the	
   23.7
s1mat
27.3
e
32.7
ear	
  
y
ne	
  
	
  o
13.7
urvey	
  (ACS),
ate'sn	
  CDegree
4.7
ommunity	
  S
e or Associ
(five	
  
2010	
  13.7
23.0
for	
  
Some CollegBurea
26.0
urvey	
  
S
u's	
  2015	
  America
nity	
  
9.1
American	
  Commu
9.3
from	
  the	
  Census	
  
Notes:
and	
  from	
  the	
  
data	
  comes	
  Bachelor's Degree
17.3
all	
  
00,	
  
oted,	
  
n
970-­‐20
for	
  1
therwise	
  
Census	
  sional
Degree
to	
  
rior	
  to	
  1990.	
  	
  
• Unless	
  o
Decenn
Profes
or ial	
  
ate
from	
  the	
  
f	
  college	
  prior	
  
f	
  schooling	
  p
5	
  and	
  older.	
  
Gradu
omes	
  
ears	
  o
c
ears	
  o
y
ent	
  
ore	
  
hree	
  y
t
popula1on	
  2
m
to	
  
or	
  
Xainm

Female	
  

•

•
•
•
•
•

25	
  to	
  34	
  years	
  

6.5	
  

17.9	
  

31.5	
  

28.0	
  

onal	
  a
es).	
  	
  
nd	
  as	
  four	
  
n,	
  and	
  as	
  one	
  
Historical	
  educa1 2015	
  (one	
  year	
  es1mat higher	
  from	
  1990	
  on,	
  a
from	
  1990	
  o
es)	
  and	
  
	
  with	
  no	
  degree	
  
egree	
  or	
  
year	
  es1mat
some	
  college
	
  bachelor's	
  d
from	
  ACS	
  data.	
  
defined	
  as	
  a
te's	
  degree	
  or	
  
of	
  Richmond	
  
le.	
  
"College"	
  is	
  
as	
  an	
  associa
l	
  Reserve	
  Bank	
   hen	
  es1mates	
  are	
  availab is	
  from	
  the	
  2014	
  
"	
  is	
  defined	
  
by	
  the	
  Federa
	
  Race",	
  w
original	
  data	
  
"Some	
  College
re	
  calculated	
  
nd	
  "Some	
  Other
h	
  Service.	
  The	
  
a1on	
  rates	
  a
Alaska	
  Na1ve	
  a USDA	
  Economic	
  Researc
orce	
  par1cip
1990.	
  
n	
  Indian	
  and	
  
ent	
  and	
  labor	
  f
d	
  by	
  the	
  
ri1ng.	
  
Unemploym
"	
  includes	
  America
data	
  compile
the	
  1me	
  of	
  w
t	
  
ses	
  
a
u
Other
	
  
le	
  
"
y	
  
ounty
vailab
c
ere	
  not	
  a
The	
  race	
  categor onal	
  aXainment	
  by	
  
es	
  for	
  2015	
  w
educa1
County	
  es1mat
	
  
The	
  map	
  of	
  
urvey.
S
nity	
  
American	
  Commu

87.5
86.1

79.9
78.9

8.3	
  
7.8	
   0.6	
  

42.8	
  

15.4	
   Graduate or
School
Equivalent

10.9	
  

17.8	
  

30.3	
  

35.7	
  

19.7	
  
25.5	
  
in 2015
nment by Race
al Attai28.6	
  
18.9	
  
Education
16.0	
  
ears	
   7.3	
  
8.5	
  

16.2	
  

High
8.6	
   0.5	
  

45.5	
  

32.2	
  

9.6 19.0	
  
over	
  
65	
  years	
  and	
  
45	
  to	
  64	
  years	
  

11.6	
  

Virginia

31.6	
  

or
Bachelor's Degree
29.7	
  
13.1	
  
Higher
13.6	
  

17.2	
  

22.4	
  

5.2
6.2
6.8

17.2	
  

33.9	
  

11.5	
  

18	
  to	
  24	
  years	
  

25.3	
  

26.0	
  

25.6	
  

14.2	
  

18.3	
  
21.1	
  

16.6	
  

nt)
(Perce
ent Rate
20.0	
  
Unemploym
16.4	
  
22.3	
  

over	
  
65	
  years	
  and	
  

Male	
  

30.9	
  

Less than High School

$22,657
$21,320
13.5	
  

29.8	
  

29.4	
  

24.0	
  

10.3	
  

13.5	
  

$34,377

$29,674
24.1	
  
$29,004

12.2	
  

18.5	
  

26.1	
  

28.5	
  

12.0	
  

16.2	
  

24.7	
  

32.2	
  $35,656

17.2	
  

14.4	
  

lent
23.9	
  

30.5	
  

10.2	
  

18	
  to	
  24	
  years	
  

27.6	
  

25.4	
  

$78,049

$67,286

sional Degree
Graduate or Profes

Each year, the Census Bureau interviews more
than 2 million people for the American Community
Survey. This survey collects data on demographics,
employment, education, and other personal
characteristics from the nation’s population. A new
report from the Richmond Fed uses this information
to look at earnings, unemployment, and labor force
participation by educational attainment throughout
the Fifth District, with breakdowns across race, sex,
age, and geographic area.

Two or More
Races
117,556
9.4
18.0
33.7
22.3
16.5

Visit https://www.richmondfed.org/research/regional_
economy/reports/special_reports

TO OUR VALUED READERS: YOUR INPUT IS IMPORTANT TO US.

Help us improve by returning the reader survey card in this
issue or go online to harveyresearch.com/econfocus/