View original document

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

F EDERAL R ESERVE B ANK

OF

P HILADELPHIA

First Quarter 2018
Volume 3, Issue 1

The Policy Perils of Low Interest
Rates
Banking Trends
Nontraditional Insurance and Risks
to Financial Stability

Contents
First Quarter 2018

1

18

The views expressed by the authors are not
necessarily those of the Federal Reserve.

11

A convergence of forces continues to drive down market interest rates. Lukasz Drozd explains
why this global trend may leave
central banks without their main
recession-fighting tool and why
their options may be limited.

Volume 3, Issue 1

A publication of the Research
Department of the Federal
Reserve Bank of Philadelphia

The Policy Perils
of Low Interest Rates

Nontraditional
Insurance and Risks
to Financial Stability
What role, if any, should the
Fed play in regulating insurance
companies? Yaron Leitner
explores how insurers’ expansion
into nontraditional activities
might affect financial stability.

Banking Trends:
Skin in the Game
in the CMBS Market
Does a new risk retention rule
induce originators of commercial
mortgages to make less risky
loans? James DiSalvo and Ryan
Johnston review the evidence on
whether skin in the game matters.

26

Research Update
Abstracts of the latest
working papers produced
by the Philadelphia Fed.

The Federal Reserve Bank of Philadelphia
helps formulate and implement monetary
policy, supervises banks and bank and
savings and loan holding companies, and
provides financial services to depository
institutions and the federal government. It
is one of 12 regional Reserve Banks that,
together with the U.S. Federal Reserve
Board of Governors, make up the Federal
Reserve System. The Philadelphia Fed
serves eastern and central Pennsylvania,
southern New Jersey, and Delaware.

About the Cover

Patrick T. Harker
President and
Chief Executive Officer
Michael Dotsey
Executive Vice President and
Director of Research
Colleen Gallagher
Research Publications Manager
Brendan Barry
Data Visualization Manager

ISSN 0007–7011

The $100 bill is all about Philadelphia—and the founding of our nation. On its face is
Benjamin Franklin, whose arrival in Philadelphia from Boston at age 17 helped
change the course of history. On the reverse is the engraving adapted for our cover
image of Independence Hall, where the Declaration of Independence and Constitution
were debated and signed. Two blocks north on Sixth Street is the current home
of the Federal Reserve Bank of Philadelphia, founded after the Federal Reserve Act of
1913 authorized the issuance of Federal Reserve notes such as the $100 bill. To see
how the look of the $100 bill has evolved since 1914, go to:
https://www.uscurrency.gov/denominations/100.
Photo by Rich Wood.

Connect with Us
We welcome your comments at:
PHIL.BRComments@phil.frb.org

Twitter:
@PhilFedResearch

E-mail notifications:
www.philadelphiafed.org/notifications

Facebook:
www.facebook.com/philadelphiafed/

Previous articles:
www.philadelphiafed.org/research-anddata/publications/economic-insights

LinkedIn:
www.linkedin.com/company-beta/634655/

The Policy Perils
of Low Interest Rates
Well before central banks slashed rates to fight the
Great Recession, long-term market rates began
slipping. With no reversal in sight, will policymakers
lose their main recession-fighting tool?

Lukasz Drozd is a senior economist at the Federal Reserve
Bank of Philadelphia. The views
expressed in this article are
not necessarily those of the
Federal Reserve.

BY L U K A S Z A . D RO Z D

T

he primary tool that central banks
have to fight recessions is to cut
interest rates so as to encourage
enough borrowing and spending to return
the economy to full employment. But as
we experienced during the Great Recession, there is a natural limit to how low
interest rates can go: It is known as the
zero lower bound—or the effective lower
bound. When the interest that banks pay
on deposits reaches zero, lowering rates
further means depositors earn a negative
return—in other words, they must actually
pay to deposit their money—making it
more attractive to stuff cash in a mattress.
At that point, monetary policymakers
are left without their most tested method
of stimulating demand.1
The Great Recession marked the first
time in the postwar era that the zero lower
bound became a relevant constraint for
monetary policymaking worldwide.2 Unable to lower rates any further, the Federal
Reserve and central banks in Europe
and other developed countries struggled
to deliver the additional monetary policy
stimulus needed to counteract the deepest
economic contraction since the 1930s,
finally resorting, as I will discuss, to less
proven, unconventional tools such as
forward guidance and quantitative easing.
Nine years on, economists are still debating the extent to which the lack of the
primary monetary policy instrument contributed to the severity of the recession.

Why Can’t Central Banks Simply Set Rates Below Zero?
Contrary to the common perception, central
banks do not set interest rates across
the economy. Most rates are determined
by supply and demand forces in financial
markets, and central banks influence these
forces to move rates to the desired level.
The Federal Reserve, for example, targets the
interest rate at which banks lend to each
other overnight, the so-called federal funds
rate. This market rate is determined by the
supply of and demand for funds in the
interbank market, where banks borrow and
lend overnight to ensure they are never
holding too little or too much in required
reserves based on daily fluctuations in
their assets and liabilities. The Fed targets
this particular market and this particular
rate because it can most directly influence it
and at the same time tightly control the
availability of funds in the banking sector,
which is under its supervision. Since the
current and future cost of funds is the key
determinant of how much banks charge
their customers for loans, when the fed
funds rate moves or is expected to move,
interest rates across the economy move
as well, and so does aggregate demand.3
When the zero lower bound is not an issue,
the Fed can lower the fed funds rate by, in
essence, printing money to purchase
government bonds from primary broker-

The Policy Perils of Low Interest Rates
2018 Q1

dealers, which are bank subsidiaries or deal
with banks and carry out the transaction
through their accounts with the Fed. The Fed
credits the broker-dealer’s account with
an electronic deposit equal to the payment
for the bonds. The recipient banks may
ultimately lend these funds to firms and
households, but since it takes time to
find such long-term customers, banks
typically look to earn interest in the meantime by lending the funds in the fed funds
market. Because this activity increases
the supply of funds in the interbank market,
their price—the fed funds rate—declines.
While the Fed can increase the supply of
funds at will, it cannot induce the fed funds
rate to fall much below zero. In principle,
the Fed could “tax” banks’ cash, prompting
them to lend their excess to other banks,
even at negative interest. But banks would
ultimately have to pass on the cost of such
a tax by charging their customers for making
deposits—in other words, imposing a negative rate on deposits. Since depositors can
just as well stuff money in their mattresses,
there is a limit to how much banks can
charge for deposits. Hence, there is a limit
to how low rates can go before they hurt
bank profits and credit conditions across
the economy, which would work against the
Fed’s goal during a recession of stimulating
aggregate demand by driving down rates.

Federal Reserve Bank of Philadelphia
Research Department

1

Today, the Great Recession is long over. Economic output and
employment have recovered, and the Federal Reserve has hiked
its policy target rate several times, causing market rates across the
economy to begin rising again, to the relief of savers. Yet, to policymakers, the problem of the zero lower bound remains a major
concern. What alarms them is that beyond cyclical, policyinduced fluctuations, market interest rates have been trending
downward for years, starting long before the last recession,
and bringing the zero lower bound ever closer. If this trend has
continued through the crisis and current expansion, and many
economists believe it has, then policymakers could face the next
major recession without a monetary remedy, as occurred with
dire results in the Great Depression. Even a mild recession could
compel policymakers to turn to the kind of extraordinary interventions employed during the latest crisis, only this time without
the extra margin of first responding by cutting rates.
As I will discuss, the nature of the forces behind the decline
in interest rates gives little hope for a reversal in the foreseeable
future. As I will also discuss, this outlook is prompting study
and debate over whether a low-rate environment calls for a fundamentally different approach to monetary policy and to how
central banks will fight recessions in the future.

The Global Decline in Interest Rates

Today’s exceptionally low interest rates are often blamed on the
Great Recession and the economic malaise that lingered in its
aftermath. But the picture that emerges from an analysis of the
average interest rate across countries shows that the decline very
much predates the Great Recession. Accordingly, the low rates
prevailing currently may have less to do with the crisis and more
to do with the secular global decline in long-term interest rates.
To pinpoint the beginning of this decline, economists follow
the evolution of the average inflation-adjusted yield on long-term
bonds issued by governments of major world economies that are
fiscally sound and open to international capital flows. Averaging
long-term government bonds helps filter out forces that are
expected to reverse course such as business cycle fluctuations or
monetary policy interventions to fight recessions.4 Tracking the
average world interest rate also helps identify the trend because
it focuses on movements driven by forces that are common
across countries and hence unlikely to be canceled out by international capital flows, which tend to equalize returns across
countries in the long run. In addition, tracking the average world
interest rate removes the effect of expected (real) currency exchange rate movements, which can cause interest rates to diverge
from the actual return that global investors earn after taking into
account differences in the rate of return on currencies implied
by exchange rates.5 The estimated average world interest rate
suggests that long-term rates have been declining since at least the
1990s.6 The real interest rate paid on 30-year U.S. Treasury
Inflation Protected Securities has followed a remarkably similar
downward path, highlighting the relevance of global trends for
the evolution of U.S. interest rates (Figure 1).
The global secular decline in long-term rates—alongside
aggressive interest rate cuts during the Great Recession—is a key
reason why economists believe that even though central banks

2

Federal Reserve Bank of Philadelphia
Research Department

have begun raising their policy rate targets, they might still be
operating close to the zero lower bound. What reinforces
these concerns is the fact that, as the economic recovery has
gained momentum around the globe, both inflation and interest
rates remain exceptionally low. While central banks can act to
move interest rates across the economy to stimulate spending, if
they hope to ensure low and stable inflation they must over
the long run respect the supply and demand forces generated
from within the economy that drive interest rates. Keeping rates
below the natural level implied by these forces, known as the
neutral interest rate—also called the natural rate or r-star—
eventually leads the economy to a state of full employment.7 But
at that point, further stimulation no longer prompts firms to
increase employment; rather, it makes them raise prices to meet
the excess aggregate demand implied by below-neutral rates,
ultimately leading inflation to rise out of control.8
While today’s lower rates may reflect the still-accommodative
stance of monetary policy, economists attribute much of the
secular decline in policy rates since the 1990s to the global decline
in the neutral rate (Figure 2). What leads them to this conclusion
is the long period in question—a period that includes both
economic expansions and recessions—and the remarkably stable
rate of inflation during this time.
This assessment is confirmed by econometric studies that aim
to estimate the neutral rate using empirical data. The best known
among them, by Thomas Laubach and John C. Williams, finds
the neutral rate to be below zero.
FIGURE 1

Interest Rates Have Been Falling Since the 1990s

Average inflation-adjusted interest rate in developed economies,
1985–2013, and 30-year U.S. Treasury Inflation Protected
Securities yield, 1998–2013.
5%

4%

3%

2%

1%

U.S. TIPS

0%

Developed
economy
average

−1%

1985

1990

1995

2005

2013

Note: Inflation-adjusted interest rate as estimated by Mervyn King and David Low
using inflation-protected long-term government bonds issued by G-7 countries,
minus Italy, 1985–2013. Inflation-adjusted yield earned on 30-year U.S. Treasury
Inflation Protected Securities held to maturity.

The Policy Perils of Low Interest Rates
2018 Q1

2000

Sources: King and Low (2013) and Treasury Department.

While finding slightly higher estimates,
two other studies confirm these findings
using a different methodology.9 A third,
more recent, analysis10 yields a markedly
higher estimate by allowing for investor
preferences to shift toward safe U.S. bonds,
but it also suggests the neutral rate has
fallen to an alarmingly low level, below
100 basis points.11

Forces at Play for the
Foreseeable Future

Adding to concerns that today’s low rates
may complicate the conduct of monetary
policy in the future are model-based
studies that see demographic trends—and
to a lesser extent, slower productivity
growth—in developed countries as the
main culprits. Since these changes are
persistent, the models predict that rates
will remain depressed for the foreseeable
future. How do the forces identified by
the models shape interest rates, and how
will they evolve in the future?
Three model-based studies that focus
solely on demographics forecast that

interest rates will continue to decline
until at least 2050 given projected global
demographic trends, even absent any productivity slowdown.12 A recent detailed
study for the U.S. economy confirms the
forecast for demographics, while finding
that the productivity slowdown will also
play a role in depressing interest rates.13
All of these studies seek to understand
the principal driving forces behind the
demand for and supply of funds in financial markets, where interest rates are
determined. The models derive the supply
of funds from household net saving for
retirement, while the demand comes from
firms that seek funds to invest in capital
to produce goods and services, with interest rates balancing demand with supply
by falling whenever demand rises or
supply falls.

profile of a household exhibits a remarkably similar pattern across countries
(Figure 3). That is, minors have little
income, as they supply little labor, yet they
consume out of the income of their
parents, on net subtracting from the savings of the household sector as a whole.
Similarly, retirees live off their accumulated savings, supply little labor, and also
subtract from the savings of the household
sector. The bulk of household savings
and labor supply come from working-age
individuals, so their share of the population is a crucial determinant of how much
households overall save.
Demographic forces matter not only for
households’ saving behavior but also
for firms’ demand for funds to invest in
capital, because when demographics
change, so does the supply of labor. Models assume that labor and capital are
complementary inputs in the production
of goods and services: As less labor is
employed in production, capital inputs
become less productive, meaning that
the size of the labor force affects the
amount of capital investment and hence

Demographics
What makes demographic trends play
a crucial role in the models is that saving
and working vary predictably over
a person’s lifetime. The average lifecycle

FIGURE 2

FIGURE 3

Estimated Neutral Rate Has Likewise Plummeted

Propensity to Earn and Save Varies Predictably by Age

The Laubach and Williams estimate of the neutral U.S. interest
rate, 1985–2017.

Average consumption and labor income over the lifecycle in
developed and developing economies.

5%

3.0%

4%

2.5%

3%

2.0%

2%

1.5%

1%

1.0%

0%

0.5%

−1%

0.0%

1985

1990

1995

2000

2005

2010

2016

Source: Laubach and Williams (2003), online data set as of July 28, 2017, www.
frbsf.org/economic-research/files/Laubach_Williams_updated_estimates.xlsx.

Consumption
Developed

Developing

1

15

30

45

60

75

Labor income
Developing
91 Developed

Source: Lee and Mason (2011), selected data from www.ntaccounts.org.
Note: Average earnings and consumption by age. Developed economies: Canada,
France, Germany, Italy, Japan, United Kingdom, United States. Developing
economies: Brazil, Chile, China, Philippines, Hungary, India, Indonesia, Mexico,
Peru, South Africa, South Korea, Thailand, Turkey.

The Policy Perils of Low Interest Rates
2018 Q1

Federal Reserve Bank of Philadelphia
Research Department

3

firms’ demand for funds. This assumed
complementarity between labor and
capital is consistent with the remarkable
stability across time and countries of the
share of an economy’s total income that
goes to wages and salaries. Although it
has declined somewhat globally for about
the past 15 years, labor’s share of income
had remained stable despite a major
decline in prices of capital goods and
interest rates, which absent the assumed
complementarity between labor and
capital would not have happened to this
extent, according to the models.14
Both key demographic trends that
according to these models have affected
interest rates—slower population growth
due to lower fertility and longer life
expectancies (Figure 4a), and declining
labor supply due to aging and the
resulting rise in the share of retirees in
the population (Figure 4b)—are projected
to continue, and at a faster pace than in
the past.
The basic mechanism that makes these
trends relevant for the evolution of interest
rates around the globe is that lower
fertility implies that the working-age
population saves more on net, as there are
fewer minors. Similarly, longer life
expectancy implies that workers in midlife
save more to afford longer retirements.
Although the growing share of retirees in
the population has the opposite effect
and tends to reduce the net savings of the
household sector, it also reduces labor
supply and hence the demand for capital
and investment due to the complementarity of capital and labor, for an attenuated
or even opposite net effect on interest
rates.

FIGURE 4

Global Demographic Trends Likely to Continue
4a: World birth rate per 1,000 population
in developed and developing regions.

4b: Old-age dependency ratio.

Percent

Number of people age 65 and older to the number
age 20–64, percent

50

50

45

45

40

40

35

35

30

30

25

25

20

20

4

Federal Reserve Bank of Philadelphia
Research Department

Developing

15

Developed

10

10

Forecast

Forecast

5

0

Productivity
Productivity in developed countries, and
also globally, has been rising more slowly
in recent years, leading to projections of
slower income growth. 15 What makes the
rate of productivity growth important
is that future income depends on labor
productivity growth; if slower growth is
expected, the working-age population
may be encouraged to save more of what
they earn so they will be able to maintain
their targeted level of consumption in
the future. Not all studies take the productivity slowdown into account, but those

Developing

15

Developed

5

1985 2000

‘15

‘30

0

‘45

1985 2000

‘15

‘30

‘50

Source: United Nations Department of Economic and Social Affairs, Population Division, World Population
Prospects: The 2017 Revision. Custom data from https://www.un.org/development/desa/publications/
world-population-prospects-the-2017-revision.html.

that do conclude that it is a force to be
reckoned with.
A comprehensive study of U.S. interest
rates that analyzed the effect of not
only demographics but also productivity
growth, government debt, and capital
goods prices found that, had it not been
for the offsetting effect of rising U.S.
government debt, the productivity slowdown would have depressed interest rates

further. Productivity growth is difficult
to forecast, and initially reported growth
rates are sometimes substantially revised. However, the analysis—by Gauti
Eggertsson, Neil Mehrotra, and Jacob
Robbins—suggests that productivity would
have to grow at trend rates not seen in
the postwar period to undo the effect of
demographic forces.

The Policy Perils of Low Interest Rates
2018 Q1

How Do Demographic Forces Affect Interest Rates?
Demographics lower interest rates by affecting both the
supply of and demand for funds in financial markets. In
Figure 5, the upward sloping blue line represents the supply
of funds. It rises as interest rates rise because higher interest
rates encourage saving. The downward sloping red line
represents the demand for funds. It falls as interest rates
rise because higher rates make loans more costly, thereby
discouraging borrowing. The equilibrium interest rate
balances out the demand for funds with the supply of funds
where the two lines intersect. 1 In recent decades, falling
birth rates and rising longevity have increased net household
savings, increasing the supply of funds in global financial
markets. 2 But rising life expectancy has also increased the
share of retirees in the population, for an attenuated effect on
the net supply of savings and interest rates. 3 Yet, it appears
that rates have continued to decline because the rising
share of retirees has also reduced the size of the labor force
and hence the demand for capital and funds.

FIGURE 5

How Demographics Have
Lowered Interest Rates
Demand for
funds

Supply of
funds

Equilibrium interest
rate

Interest rate →

Other Forces
Two other factors that modelbased studies do not take
into account might also have
affected the evolution of interest rates around the globe.17
Commentators have recently
emphasized a potential role for
rising income inequality within
countries and China’s unusually high national saving rate.
Income inequality has been
rising in most countries,
especially in developed ones,
raising the question of how
this trend might have affected
interest rates.18 Because the
rich generally save a larger fraction of their income, inequality
within a country tends to raise
the net savings of its households overall.
Lukasz Rachel and Thomas
D. Smith shed light on this
issue by showing that withincountry inequality is correlated
with cross-country net savings.
Based on this fact, they estimate that income inequality
accounts for about one-tenth of
the decline in interest rates
and argue that if inequality
rises further, interest rates will
fall further. 19
By focusing on averages,
economic models also do not
take a full account of idiosyncratic differences among
countries. For example, China’s
high saving rate has been
quite an outlier given that
nation’s level of economic
development. China’s consumption profile is well below
even lower-income countries
(Figure 7). Today, China is
actively moving toward a more
consumer-oriented rather than
export-oriented economy,
which opens up the possibility
that its saving rate may decline
in the future, alleviating the
downward pressure that its
growing economy exerts on
world interest rates. However,
there are important caveats
to this reasoning. One, this

Global funds →

A remarkable feature of the dynamic illustrated here is that
it is consistent with little change in the global saving rate.
This observation is broadly consistent with the evidence.16
While the saving rate in the U.S. and other advanced economies has been declining, for the world as a whole it has
remained remarkably constant as rates have fallen (Figure 6).
1
FIGURE 6

Lower Rates Have Not Spurred a Rise
in Global Saving

Gross savings (gross national income less total
consumption) as a fraction of GDP.
30%

25%

Global
2

Developed
20%

15%

10%

5%

0%

1985

1995

2005

2015

3

Source: World Bank, World Development Indicators; custom data
from https://data.worldbank.org/indicator/NY.GNS.ICTR.ZS

The Policy Perils of Low Interest Rates
2018 Q1

Federal Reserve Bank of Philadelphia
Research Department

5

process will take time. Two, as the gap
narrows, China’s rising share of world
income will continue to depress interest
rates until its saving rate is no longer
above the world average.

designed to work within the existing
framework is forward guidance. The basic
idea behind forward guidance is that
merely by publicly announcing how interest rates will be set in the future, a central
bank can alter the public’s expectations
about future short-term rates. Since longterm rates are tied to expectations about
short-term rates, long-term rates will
decline after such an announcement.21
Unlike traditional U.S. monetary policy,
which aims to influence rates throughout
the economy by moving overnight
interbank rates, forward guidance aims to
achieve the same goal by announcing how
future rates will be set. Banks and other
financial intermediaries price the longterm loans they make today based on their
expected cost of funds in the future, as
such loans must be financed over an extended period. So, in principle, if a central
bank’s forward guidance is credible,
lenders will raise or lower the interest they
charge on long-term loans accordingly.
And in fact, the evidence so far is quite
compelling that forward guidance was
effective in lowering long-term rates during
the crisis.22
The second type of unconventional
instrument is balance sheet policy, in
which a central bank purchases riskier
assets from banks and other lenders so

An Uncertain Future
for Monetary Policymaking

With the normalization of policy rates well
under way but market rates persistently
low, policymakers may soon face an important question: What is the best way to
respond to recessions given the projected
low interest rate environment? Two types
of options are being debated: Central
banks can accept that whenever they need
to stimulate the economy, they will have
to resort to employing unconventional
instruments within the existing monetary
policy framework to lower the effective
long-term interest rate that households
and firms pay. Or they can work around
the zero lower bound by creating a new
monetary policy framework to restore the
effectiveness of their most proven tool,
cutting rates.20

Unconventional Tools: Balance Sheet
Policy and Forward Guidance
The primary unconventional instrument

they will be in a stronger position to bear
risk and hence more willing to lend to
firms and households, thereby stimulating
the economy. To implement this tool, the
central bank either changes the composition of its own balance sheet by selling
safer assets or buying riskier assets, or
both, or it expands its balance sheet by,
in essence, printing money to buy more
risky assets.
Proponents also argue that purchasing
large amounts of long-term assets
amplifies the effect of forward guidance.
Such large-scale purchases, known as
quantitative easing, boost the credibility of
the forward guidance by signaling the
central bank’s commitment to making
good on its announcements about
reducing long-term rates and, in principle,
creating room to drive them even lower.23
Also, by making it easier for financial
intermediaries to shed their riskier assets,
these purchases lower the risk premiums
that lenders charge borrowers in the form
of higher interest rates as compensation
for bearing risk.
Opponents, however, question whether
unconventional tools would be effective
under any conditions, especially in a deep
recession with rates already at or near the
zero lower bound. They point out the
theoretical limitations of forward guidance

FIGURE 7

China’s Saving Rate Likely to Keep Down Interest Rates

Average per capita labor income and consumption rates by age, China vs. developed vs. developing countries.
3.0%
China
2.5%

Labor income

Consumption
Developed

Developed

2.0%

1.5%
Developing
1.0%
China
Developing

0.5%

0.0%

1

15

30

45

60

75

Source: Lee and Mason (2011), selected data from
www.ntaccounts.org.

6

Federal Reserve Bank of Philadelphia
Research Department

91

1

15

30

45

60

75

Note: Developing economies: Cambodia, Ethiopia, Ghana, Indonesia, Jamaica, Kenya, Mozambique, Nigeria,
Philippines, Senegal, Vietnam.

The Policy Perils of Low Interest Rates
2018 Q1

91

and the somewhat mixed evidence regarding the effectiveness of
quantitative easing.24 They also emphasize that the economy’s
reaction to these less-tested policies is less predictable, requiring
them to be deployed more slowly in order to monitor their
effectiveness and guard against unintended consequences.
A chief concern regarding forward guidance is that, at the zero
lower bound, central banks may not always be able to influence
the public’s expectations of how future policy rates will be set. At
least in theory, neutral rates may fall to zero perpetually, invalidating the effectiveness of any forward guidance that requires
the public to believe that rates will eventually lift off the zero
lower bound.25 Also, forward guidance may conflict with central
banks’ mandate to keep inflation low and stable. That is,
central banks may find themselves having to convince the public
that they will keep rates low even after the economy is expanding
again—letting it “run hot” for a while—possibly undermining
their commitment to low inflation. At the very least, policymakers
may need to rethink inflation targeting as a means to signal their
commitment to price stability and replace it with a more flexible
target that better accommodates forward guidance.26
As a stark warning of how difficult it may be to escape the zero
lower bound once it becomes binding, skeptics of unorthodox
tools cite Japan, which has remained at the zero lower bound for
decades now, despite repeated unconventional policy interventions by the Bank of Japan.27 Indeed, while the evidence suggests
that unconventional tools can stimulate demand, Japan offers
a cautionary tale. In 1999, the Bank of Japan introduced a zero
interest rate policy in which the overnight rate was targeted “as
low as possible.” In 2001, the central bank introduced quantitative
easing and an early form of forward guidance. All these efforts
largely failed to stimulate the economy and raise inflation. In 2013,
the Bank of Japan introduced its most aggressive quantitative
easing. Inflation rose briefly above the targeted rate but soon fell
back below target, where it remains today, suggesting that Japan
has been locked in a holding pattern at the zero lower bound for
almost two decades now.
It remains an open question whether the policies Japan
deployed were simply too small in scale or duration, or whether
its experience highlights the limited effectiveness of unconventional monetary policy under the kind of extreme circumstances
afflicting Japan.

A New Monetary Policy Framework
Modifying the standard monetary policy framework might well
give policymakers enough effective tools to ensure that they can
still precisely tailor policy to the state of the economy in
a recession. Among the proposals that have attracted the most
attention, increasing the inflation target tops the list. Targeting
higher inflation was first proposed soon after the Great
Recession,28 and more recently, former Federal Reserve Chair
Janet Yellen deemed the issue “one of the most important
questions facing monetary policy around the world in the future,”
and called for more research.29

Inflation can remedy the problem of the zero lower bound because nominal interest rates compensate for the expected rate of
inflation. That is, higher inflation raises people’s expectations
of more inflation, prompting them to borrow and spend more at
a given nominal interest rate, causing rates to rise from the zero
lower bound.30
The main objection to raising the inflation target is that inflation is costly. A 2011 analysis of the costs of inflation using
modern economic models suggested that even moderate inflation
may result in significant misalignments of prices,31 although
a 2017 study of actual price dispersion during a period of high
inflation in the U.S. questions this assessment,32 suggesting
rather low costs. Nonetheless, convincing the public that more
inflation is needed may prove too high a hurdle, as inflation
remains deeply unpopular. In addition, there is a practical
concern that central banks may not yet be capable of raising
inflation, which could cost them credibility if they failed to deliver
on the new higher target.
A more radical proposal than raising inflation calls for replacing
paper currency with digital currency that could be “taxed”—that
is, whenever the economy needed a monetary stimulus, the
central bank could make the virtual currency more costly to
discourage savers from hoarding cash when bank deposit rates
turn negative.33 The main advantage of such a solution is that
the inflation target could stay at the current level or even be lowered, since central banks would have no problem driving interest
rates below zero in any inflation environment.34 However,
maintaining two parallel currencies or abolishing cash would
mean entering uncharted waters for central banks, as such
a solution has never been tried, and while appealing in theory,
in practice there could be challenges.35
As a last resort, policymakers could pursue dismantling the
separation between monetary policy and fiscal policy to allow
central banks to finance government spending by, in effect,
printing money.36 Although helicopter drops of money, as they’ve
been called, could give central banks more power to stimulate
the economy, breaking the separation between monetary policy
and fiscal policy is a controversial proposal. A central bank that
ventured into fiscal policy would likely find itself under pressure
from the private sector, financial markets, and the government
to use its balance sheet to relax fiscal constraints in the short run
at the risk of undermining the stability of the monetary system.37

Concluding Remarks

Interest rates have been declining globally for years and may not
rise in the foreseeable future, according to current projections.
The experience of the Great Depression cautions that a major
recession without an adequate monetary or fiscal accommodation can have disastrous consequences for the economy. How
central banks will adapt to this “new normal” is still unclear.
What is clear, however, is that the zero lower bound will likely
remain at the top of central banks’ agendas, as sooner or later
a major recession will come along to test whatever tools are
available to fight it.

The Policy Perils of Low Interest Rates
2018 Q1

Federal Reserve Bank of Philadelphia
Research Department

7

Notes
1 Because there are costs to storing large amounts of cash, in practice
central banks may be able to drive interest rates below zero. So, for all
practical purposes, the effective lower bound occurs at whatever rate
results in cash hoarding. Negative interest rates have been implemented
in Europe and Japan to extend the scope of conventional monetary
policy. See “Why Can’t Central Banks Simply Set Rates Below Zero?” on
p. 1 for details.
2 The first country after the Great Depression to experience the zero
lower bound was Japan, as I discuss later.
3 While the actual operating procedures differ across central banks, the
basic principle of the transmission mechanism outlined here remains
similar.
4 Governments can borrow for the long term by issuing either long-term
bonds or short-term bonds in a staggered fashion so that the proceeds
from current short-term issuances can be used to pay investors who purchased prior issuances, thereby effectively borrowing for the long term.
Hence, governments will pay significantly higher yields on long-term debt
only if they expect borrowing costs for short-term bonds to rise persistently, which is not the case with transient forces.
5 Averaging interest rates across a large number of countries eliminates
the effect of expected exchange rate movements because exchange
rates are reciprocal: If one currency is expected to strengthen, others are
therefore expected to weaken, for an offsetting effect on in interest rates
around the globe and an attenuated effect on the average interest rate.

13 See the work of Gauti Eggertsson, Neil Mehrotra, and Jacob Robbins,
who include the productivity slowdown among other factors.
14 For more details on the labor share, see Roc Armenter’s Business
Review article. The analysis by Eggertsson and his coauthors concludes
that although labor’s declining share of U.S. income has contributed to
interest rate declines, it has been a secondary factor.
15 For an accessible discussion, see Mike Dotsey’s Economic Insights
article. Robert Gordon has been the most prominent advocate of the hypothesis that the productivity slowdown in the U.S. and other developed
countries is here to stay. For a contrarian and more optimistic view
of what the future may bring, see the book by Erik Brynjolfsson and
Andrew McAfee.
16 An alternative hypothesis put forth by Charles Bean, Christian Broda,
Takatoshi Ito, and Randall Kroszner to explain the remarkable stability
of the world saving rate holds that either demand or supply is insensitive
to rates, implying that one of the lines in Figure 5 is vertical. Microeconomic studies suggest that both investment and saving respond to
interest rates. See the discussion in the paper by Lukasz Rachel and
Thomas Smith.
17 See, for example, the blog post by former Federal Reserve Chairman
Ben Bernanke on the global savings glut.
18 The rise in the income distribution in this case is measured by the Gini
coefficient, which ranges from 0, in which all households in a country
would have the identical income, to 1, in which a single household would
earn all the income and the rest earn nothing.

6 As estimated by Mervyn King and David Low, 2013.
7 Full employment does not imply no unemployment—as at any given
time a certain number of workers are always between jobs—only that
there is no unemployment caused by a cyclical deficiency of aggregate
demand, as occurs in a recession. The unemployment rate that occurs
at full employment is known as the natural rate of unemployment.

19 Krueger and Ludwig point out that part of the rise in inequality may be
explained by aging and show that their model in part captures the increase
in inequality measures in the data.
20 See former Federal Reserve Chair Janet Yellen’s 2016 speech outlining
the future of U.S. monetary policy. See also Bernanke’s 2017 discussion
paper.

8 Amid rising prices, workers demand higher wages, leading firms to
further increase prices to cover their rising labor costs, leading workers
to demand still higher wages, and so on.

21 See Edison Yu’s Economic Insights article for an accessible explanation
of how short-term rate expectations affect long-term rates.

9 See the work by Jens H.E. Christensen and Glenn D. Rudebusch and by
Benjamin Johannsen and Elmar Mertens.

22 See Michael Woodford’s discussion of the effectiveness of unconventional monetary policy instruments deployed during the crisis.

10 See the study by Marco Del Negro, Domenico Giannone, Marc P.
Giannoni, and Andrea Tambalotti.

23 See the work by Leonardo Melosi.

11 James Hamilton, Ethan Harris, Jan Hatzius, and Kenneth West show
that from the long-term international perspective, these estimates are
more uncertain.
12 See Dirk Krueger and Alexander Ludwig’s paper as well as the study
by Etienne Gagnon, Benjamin Johannsen, and David Lopez-Salido. The
paper by Carlos Carvalho, Andrea Ferrero, and Fernanda Nechio suggests
similar findings, although it does not offer a long-term forecast.

8

Federal Reserve Bank of Philadelphia
Research Department

24 Yu’s Economic Insights article also explores the theoretical challenges
and evidence regarding quantitative easing.
25 See the work by Eggertsson and his coauthors.
26 In this context, price-level targeting is often considered a better
alternative to inflation targeting to signal a central bank’s commitment to
price stability and to communicate its intentions without compromising
the effectiveness of forward guidance. For a recent proposal along these
lines, see the discussion paper by Bernanke.

The Policy Perils of Low Interest Rates
2018 Q1

27 An overview of the evidence on the effectiveness of long-term asset
purchases and quantitative easing can be found in, for example, Woodford’s discussion paper.

Armenter, Roc. “A Bit of a Miracle No More: The Decline of the Labor
Share,” Federal Reserve Bank of Philadelphia Business Review (Third
Quarter 2015).

28 See the work that Olivier Blanchard did with Giovanni Dell’Ariccia
and Paolo Mauro while Blanchard was research chief at the International
Monetary Fund. Also see the work by Laurence Ball as well as Stephen
Cecchetti and Kermit Schoenholtz.

Ball, Laurence. “The Case for A Long-Run Inflation Target of Four
Percent,” International Monetary Fund Working Paper 14/92 (2014).

29 See page 14 of the transcript of former Chair Yellen’s press conference
of June 14, 2017, at https://www.federalreserve.gov/mediacenter/files/
FOMCpresconf20170614.pdf. The open letter that prompted the
question can be found at https://populardemocracy.org/news-andpublications/prominent-economists-question-full-inflation-target.
30 Say you want to take out a loan and expect that by the time you need
to pay it back both wages and prices will have accelerated. The dollars
you will repay your debt with will purchase less than the dollars you borrow, and you will have to work less to repay your debt. So, at any fixed
nominal interest rate, the expectation of higher inflation will make the
lender want to lend less but will encourage you to borrow more.
31 See the work by Olivier Coibion, Yuriy Gorodnichenko, and Johannes
Wieland.
32 See the study by Emi Nakamura, Jon Steinsson, Patrick Sun, and
Daniel Villar.
33 See “Why Can’t Central Banks Simply Set Rates Below Zero?” on p. 1
for an explanation of how “taxing” cash could help break through the
zero lower bound.
34 Ruchir Agarwal and Miles Kimball argue that a partial phase-out of
paper currency to roll out a parallel electronic currency would suffice to
overcome the zero lower bound. The tax could be imposed only when the
zero lower bound became a problem.
35 See the discussion paper and book by Kenneth Rogoff for a detailed
discussion of key practical considerations underlying a complete or a
partial phaseout of paper currency.
36 For a discussion of a policy proposal along these lines, see the paper
by Adair Turner.

Bauer, Michael, and Glenn Rudebusch. “The Signaling Channel for Federal Reserve Bond Purchases,” International Journal of Central Banking,
10:3 (2014), pp. 233–289.
Bean, Charles, Christian Broda, Takatoshi Ito, and Randall Kroszner. “Low
for Long? Causes and Consequences of Persistently Low Interest Rates,”
International Center for Monetary and Banking Studies Geneva Reports
on the World Economy, 17 (October 2015).
Bernanke, Ben S. “Monetary Policy in a New Era,” Peterson Institute for
International Economics Rethinking Macroeconomic Policy Conference
(October 2017).
Bernanke, Ben S. “Why Are Interest Rates So Low, Part 3: The Global
Savings Glut,” Brookings Institution blog post (April 2015).
Blanchard, Olivier, Giovanni Dell’Ariccia, and Paolo Mauro. “Rethinking
Macroeconomic Policy,” International Monetary Fund Staff Position Note
SPN/10/03 (February 2010).
Brynjolfsson, Erik, and Andrew McAfee. “The Second Machine Age:
Work, Progress, and Prosperity in a Time of Brilliant Technologies,” New
York: W.W. Norton and Company, Inc. (2016).
Carvalho, Carlos, Andrea Ferrero, and Fernanda Nechio. “Demographics
and Real Interest Rates: Inspecting the Mechanism,” European Economic
Review, 88:C (2016), pp. 208–226.
Cecchetti, Stephen G., and Kermit L. Schoenholtz. “The Case for a Higher
Inflation Target Gets Stronger,” Money and Banking Blog Posts (April
2017), https://www.moneyandbanking.com/commentary/2017/4/2/thecase-for-a-higher-inflation-target-gets-stronger.
Christensen, Jens H.E., and Glenn D. Rudebusch. “New Evidence for a
Lower New Normal in Interest Rates,” Federal Reserve Bank of San Francisco Economic Letter 2017–17 (2017).

37 For a detailed discussion of these risks, see, for example, the essay
based on the speech by former Federal Reserve Bank of Philadelphia
President Charles Plosser.

Coibion, Olivier, Yuriy Gorodnichenko, and Johannes Wieland. “The
Optimal Inflation Rate in New Keynesian Models: Should Central Banks
Raise Their Inflation Targets in Light of the Zero Lower Bound?” Review
of Economic Studies, 79 (2012), pp. 1,371–1,406.

References
Agarwal, Ruchir, and Miles Kimball. “Breaking Through the Zero Lower
Bound,” International Monetary Fund Working Paper 15/224 (2015).

Del Negro, Marco, Domenico Giannone, Marc P. Giannoni, and Andrea
Tambalotti. “Safety, Liquidity and the Natural Rate of Interest,” Brookings
Papers on Economic Activity, 8 (Spring 2017), pp. 235–316

Ambler, Steve. “Price-level Targeting and Stabilisation Policy: A Survey,”
Journal of Economic Surveys, 23:5 (2009), pp. 974–997.

Dotsey, Michael. “Monetary Policy and the New Normal,” Federal Reserve
Bank of Philadelphia Economic Insights (First Quarter 2016).

The Policy Perils of Low Interest Rates
2018 Q1

Federal Reserve Bank of Philadelphia
Research Department

9

Eggertsson, Gauti B., Neil R. Mehrotra, and Jacob A. Robbins. “A Model
of Secular Stagnation: Theory and Quantitative Evaluation,” National
Bureau of Economic Research Working Paper 23093 (2017).
Gagnon, Etienne, Benjamin K. Johannsen, and David Lopez-Salido.
“Understanding the New Normal: The Role of Demographics,” Federal
Reserve Board Finance and Economics Discussion Series, 2016–080
(2016).

https://www.philadelphiafed.org/publications/annual-report/2012/restoring-the-boundaries.
Rachel, Lukasz, and Thomas D. Smith. “Secular Drivers of the Global Real
Interest Rate,” Bank of England Staff Working Paper 571 (December 2015).
Rogoff, Kenneth. “Costs and Benefits to Phasing Out Paper Currency,”
National Bureau of Economic Research Macroeconomics Annual Conference Series (2014).

Gordon, Robert J. “Is U.S. Economic Growth Over? Faltering Innovation
Confronts the Six Headwinds,” National Bureau of Economic Research
Working Paper 18315 (2012).

Rogoff, Kenneth. The Curse of Cash. Princeton: Princeton University
Press (2016).

Hamilton, James, D., Ethan S. Harris, Jan Hatzius, and Kenneth D. West.
“The Equilibrium Real Funds Rate: Past, Present and Future,” National
Bureau of Economic Research Working Paper 21476 (2015).

Swanson, Eric. “How Severe Has the Zero Lower Bound Constraint
Been?” VOX Policy Portal (November 8, 2014), http://voxeu.org/article/
zero-lower-bound-has-not-been-very-severe.

Holston, Kathryn, Thomas Laubach, and John C. Williams. “Measuring
the Natural Rate of Interest: International Trends and Determinants,”
Journal of International Economics, 108:S1 (2017), pp. S59–S75, https://
doi.org/10.1016/j.jinteco.2017.01.004.

Swanson, Eric T., and John C. Williams. “Measuring the Effect of the Zero
Lower Bound on Medium- and Longer-Term Interest Rates,” American
Economic Review, 104:10 (2014), pp. 3,154–3,185.

Johannsen, Benjamin K., and Elmar Mertens. “A Time Series Model of
Interest Rates with the Effective Lower Bound,” Federal Reserve Board
Finance and Economics Discussion Series 2016–033 (2016), http://dxdoi.
org/10.17016/FEDS.2016.033.
Kiley, Michael T., and John M. Roberts. “Monetary Policy in a Low Interest Rate World,” Brookings Papers on Economic Activity (Spring 2017).
King, Mervyn, and David Low. “Measuring the ‘World’ Real Interest Rate,”
National Bureau of Economic Research Working Paper 19887 (2014).
Krueger, Dirk, and Alexander Ludwig. “On the Consequences of Demographic Change for Rates of Returns to Capital, and the Distribution
of Wealth and Welfare,” Journal of Monetary Economics, 54 (2007),
pp. 47–87.
Laubach, Thomas, and John C. Williams. “Measuring the Natural Rate of
Interest,” Review of Economics and Statistics, 85:4 (2003), pp. 1,063–1,070.
Leubsdorf, Ben. “Activists in Jackson Hole Pressure Fed on Inflation,
Endorse Yellen,” Wall Street Journal (August 24, 2017).
Lee, Ronald, and Andrew Mason. Population Aging and the Generational
Economy: A Global Perspective, Cheltenham, UK: Edward Elgar Publishing Limited (2011).
Melosi, Leonardo. “Signalling Effects of Monetary Policy,” Review of
Economic Studies, 84:2 (2017), pp. 853–884.
Nakamura, Emi, Jón Steinsson, Patrick Sun, and Daniel Villar. “The Elusive Costs of Inflation: Price Dispersion during the U.S. Great Inflation,”
National Bureau of Economic Research Working Paper 22505 (2016).

Svensson, Lars E. O. “Price Level Targeting Versus Inflation Targeting:
A Free Lunch,” Journal of Money, Credit and Banking, 31 (1999), pp.
277–295.
Turner, Adair. “The Case for Monetary Finance–An Essentially Political
Issue,” International Monetary Fund Jacques Polak Research Conference
(November 2015).
United Nations Department of Economic and Social Affairs, Population Division. “World Population Prospects,” (2017 Revision), https://
www.un.org/development/desa/publications/world-population-prospects-the-2017-revision.html.
Williams, John C. “Three Questions on R-star,” Federal Reserve Bank
of San Francisco Economic Letter 2017–05 (2017). http://dx.doi.
org/10.17016/FEDS.2016.033.
Woodford, Michael. “Methods of Policy Accommodation at the Interest
Rate Lower Bound,” Federal Reserve Bank of Kansas City Economic
Policy Symposium Proceedings at Jackson Hole (2012).
Yellen, Janet L. “The Federal Reserve’s Monetary Policy Toolkit: Past,
Present, and Future,” speech at the Federal Reserve Bank of Kansas City
Economic Policy Symposium Proceedings at Jackson Hole (2016).
Yellen, Janet L. Transcript of Federal Open Market Committee press conference (June 14, 2017), https://www.federalreserve.gov/mediacenter/
files/FOMCpresconf20170614.pdf.
Yu, Edison. “Did Quantitative Easing Work?” Federal Reserve Bank of
Philadelphia Economic Insights (First Quarter 2016).

Plosser, Charles I. “Fiscal Policy and Monetary Policy—Restoring the
Boundaries,” Federal Reserve Bank of Philadelphia Annual Report (2012),

10

Federal Reserve Bank of Philadelphia
Research Department

The Policy Perils of Low Interest Rates
2018 Q1

Banking Trends:

Skin in the Game
in the CMBS Market
Issuers of commercial mortgage-backed securities
must now retain a portion on their own books. What
evidence is there that the rule will reduce risky lending?

James DiSalvo is
a banking structure
specialist and Ryan
Johnston is a
banking structure
associate in the
Research Department of the Federal
Reserve Bank of
Philadelphia. The
views expressed
in this article are
not necessarily
those of the Federal
Reserve.

BY JA M E S D I SA LVO A N D RYA N J O H N S T O N

T

he Dodd–Frank Act imposes reforms that are designed to
prevent a repeat of the disastrous performance of
residential mortgage-backed securities and—less remarked
upon—commercial mortgage-backed securities (CMBS) during the
financial crisis. Some of these regulations are designed to force
issuers of asset-backed securities to have skin in the game—that is,
to keep on their own books a slice of the securities they sell and
thus retain some of the credit risk associated with the loans
underlying the securities. The idea is that an issuer with its own
assets at stake has a greater incentive to do its due diligence, and
that this stake signals to would-be investors that the issuer also
stands to lose money if its securities fail to pay off as promised.
Most residential mortgage securities are
exempt from the new rules because their
underlying loans already conform to the
standards stipulated by the governmentsponsored enterprises that buy them. For
commercial mortgage securities, however,
the regulations are actually binding. But
what is the evidence that skin in the game
matters? If skin in the game is so important, why don’t the securities markets insist
that issuers keep an adequate stake in
order to protect investors’ own interests?
That is, do issuers actually need a government regulation to ensure that their
commercial mortgages are safely designed and that they lend
only to creditworthy borrowers? And if such a regulation is needed, are Dodd–Frank’s mortgage securities reforms well crafted?

and they are then repaid in order of these tranches. The holders
of the senior tranches are paid off first, while those holding
the junior tranches are last in line and the most likely to suffer
losses on the securities if the underlying mortgages perform
poorly (Figure 1).
The idea behind the issuer retaining a piece of the most junior
tranche, the one that carries the most risk, is that it gives the
issuer an incentive to ensure that the security includes highquality loans. Retaining this risk is thought to send a reassuring
signal to investors, who are operating in an environment of
asymmetric information—that is, the issuer knows more than they
do about the security’s underlying loans. It would be prohibitively
expensive for the typical securities
purchaser to evaluate the characteristics
of each and every loan underlying the
security, such as the creditworthiness of
the borrower and the value of the
property.2 The inability of purchasers to
evaluate for themselves the underlying
loans can lead to agency problems.3 This
means that if an originator makes a loan
that it knows is going to be sold and
securitized, it may expend too little effort
in properly evaluating its risk of default,
creating moral hazard—that is, it can reap
higher profits without taking on the full
risk normally associated with higher potential returns because
someone else is bearing part of the cost. Issuers may also pack
a security with higher-risk loans, while retaining the higherquality loans for their own portfolios. Models show that if issuers
retain some or all of the junior tranche, purchasers can be assured that the quality of the security is good, which in turn leads
them to pay a higher price. As we will show, however, recent
theories and empirical evidence address why markets do not
necessarily conform to these models.
One might think issuers’ desire to maintain a good reputation

Do issuers actually
need a government
regulation to ensure
that their commercial
mortgages are safely
designed and that they
lend only to creditworthy borrowers?

Why Have Skin in the Game?

Most models of securitization show that issuers should retain
a share of the most junior slice of the securities that they issue,
even without a government mandate.1 Mortgage-backed securities
are generally divided into levels of seniority, called tranches,

Banking Trends: Skin in the Game in the CMBS Market

2018 Q1

Federal Reserve Bank of Philadelphia
Research Department

11

Properties
Commercial and
multi-family

Loans
Bank

Payments

Lower risk,
higher
repayment
priority

AAA

How the CMBS Market Is Structured

AA

Commercial
mortgage
payments
are pooled
into a trust.

Securities
are
divided
into
tranches
based on
risk.

Investors buy the
securities that fit their
risk tolerance.

A
BBB
BB
Higher
risk, lower
repayment
priority

B
Junior tranche

B-piece buyers buy the
riskiest securities.

More Risk
Low priority
for repayment

FIGURE 1

8ft

would be a strong enough motivation for
them to use high-quality loans in the
securities they issue. After all, they don’t
issue securities just once but many times
and would like purchasers to be repeat
customers. However, there is evidence that
reputational concerns don’t necessarily
guarantee quality. One study found that
commercial mortgage-backed securities
issued by institutions that had recently
sustained large stock losses performed
poorly.4 The study also found that troubled
issuers took poorly performing securities
from their own portfolios and packed
them into other securitized vehicles. One
interpretation of these findings is that
a firm that has suffered losses is more
likely to fail. Since a good reputation is
valuable only if the firm remains a going
concern, the value of a good reputation
falls for firms experiencing losses and facing a higher likelihood of failure.
Leading up to the financial crisis, issuers often created securities with the
intention of selling off the entire issue,
sometimes without the knowledge of
investors, as we discuss below. A substantial share of these issues proved to be
of poor quality, and many observers have
argued that the lack of skin in the game
was an important reason that the underwriting was so poor. As former Securities
and Exchange Commissioner Luis A.
Aguilar pointed out, “…since lenders were
not going to suffer if the loans were not
repaid, they no longer had the incentive
of ensuring that the loans would be of
appropriate quality.”5 The authors of Dodd–
Frank adopted the view that regulations
mandating skin in the game are necessary
to prevent securitization markets from repeating the lax underwriting that preceded
the crisis.

12

Federal Reserve Bank of Philadelphia
Research Department

An Overview of the Risk
Retention Rule

The new risk retention rule—known as
Regulation RR—requires issuers of all
types of asset-backed securities to retain
at least a 5 percent share of any security
they issue, as determined by its fair value
at the time of issuance.6 The requirement
can be met by holding a share of the
junior tranche, which is called horizontal
retention, a portion of each tranche,
known as vertical retention, or a mixture
of the two, known as L-shaped retention
(Figure 2). Issuers may not directly or
indirectly hedge or transfer the risk of the
retained share.7 However, they may sell
off all or part of the junior tranche of their
requirement to investors who are experts
at evaluating commercial real estate,
known as B-piece buyers. In the final analysis, issuers remain responsible for
compliance with the risk retention rule as
well as monitoring the B-piece buyers’
compliance with the rule.
The rule contains several exceptions for
issuers of commercial mortgage-backed

securities.8 An issuer is not required to
retain any portion of a loan that meets the
definition of a qualified commercial real
estate loan.9 The presumption here is that
a qualified loan is well documented and
has prudent terms, and that the borrower
is creditworthy, so the loan is less likely to
default. Under the allocation-to-originator
option, an issuer may allocate a portion
of its required retention to any lender
that had originated at least 20 percent of
the underlying loans in the pool. The
originator must hold at least 20 percent of
the required retention but can’t hold
a larger percentage than the percentage
of loans it originated. The rationale for
this option is that providing incentives
for the originator has essentially the same
effect as providing incentives to the issuer.10
Note that of the three options for
retention, only horizontal retention fits
the prescription from economic models
that an issuer should retain a share of
the riskiest tranche. Regulators say that
having three retention options provides
issuers with the flexibility to choose

FIGURE 2

Only Horizontal Retention Fits the Risk Prescription
Horizontal Retention

Vertical Retention

L-shaped Retention

AAA

AAA

AAA

AA

AA

AA

A

A

A

BBB

BBB

BBB

BB

BB

BB

B

B

B

Junior tranche

Junior tranche

Retention equal to no less than
5 percent of each class of abs
issued or a single vertical
security which represents an
interest in each class of abs
issued.

Retention equal to no less than
5 percent of a combination of
both horizontal and vertical
retentions.

Junior tranche
Retention of the first loss
tranche in an amount equal to
no less than 5 percent of all
abs in the securitization
transaction.

Banking Trends: Skin in the Game in the CMBS Market
2018 Q1

a structure that is compatible with the practices in a particular
securitization market. For example, if an issuer usually retains
less than 5 percent of a junior tranche in a commercial mortgagebacked security transaction in a particular segment of that
market, the rule allows the issuer to hold the rest of its requirement through a vertical slice. This flexibility permits some
variation across asset-backed securities markets, but there is
some danger that it simply ratifies inefficient market practices
by some participants.

Theory Says Unregulated Markets May Be
Inefficient

In a recent theoretical model, Gilles Chemla and Christopher
Hennessy demonstrate that unregulated markets do not necessarily provide appropriate incentives for originators to do their
due diligence, even when investors are sophisticated, by which
we mean they understand the incentives of originators and
issuers.11 Note, Chemla and Hennessy do not argue that investors
in CMBS were necessarily sophisticated during recent crisis, only
that bad market outcomes can occur even when investors are
sophisticated.
In the model, originators of loans must make some costly
effort if they want to increase the likelihood that a loan will be repaid in full and on time.12 This effort might involve carefully
examining a builder’s books and credit history and analyzing
local real estate conditions. When the originator makes such an
effort, there is a greater likelihood that the result will be a highquality loan. But even if originators make this effort, default
can still occur; for example, local real estate conditions could
deteriorate unpredictably. So, whether or not the originator
makes the effort, the loan will have either a low risk of default
(high quality) or a high risk of default (low quality). Because
there is always some risk of default, investors can’t automatically
infer that an originator had made too little effort if a loan defaults.
Originators in the model may sell a security based on the
expected cash flow from the loan. And they will retain a junior
share of the security if they expect to make a positive return
from doing so.13 Otherwise, they will sell off the junior tranche to
willing investors. It is costly for originators to keep any portion
of the loan on their books; for example, real-world originators
use the proceeds from securitizing loans they made previously to
fund new loans. In the model, originators determine how much
underwriting effort to make by estimating both how much they
expect to receive from selling the security to reinvest in new loans
and their return on their retained portion.
The model includes two types of investors. Most understand
originators’ incentives but are not sufficiently informed to fully
evaluate the riskiness of a loan. The rest, speculators, are capable
of evaluating the riskiness of a loan, but at a cost. Speculators
will bear this cost only if they expect to profit from identifying and
buying underpriced securities backed by low-risk loans.14 If
speculators are active in the market—later we discuss when they
will be active—their buying and selling raises the price of securities
backed by low-risk loans and lowers the price of securities
composed of high-risk loans. However, while speculators are more
informed than most investors, they are less well informed about

the quality of the loans than the originators are. So, even with the
benefit of superior pricing information from speculators, the
prices of the securities are noisy; that is, they don’t perfectly
reflect differences in loan quality.
How accurately prices reflect the underlying risk of default is
important to investors, who typically try to align their portfolios
according to their risk preferences. In the model world, some
investors may suffer a negative shock to their income, so they
would like to purchase insurance from other investors. The more
accurately the prices of securities reflect actual default risk, the
closer investors can come to fully insuring their income. Importantly, even though managing these risks
is important to investors, originators
are not compensated for setting prices
that convey the true risk; that is, from
the originators’ standpoint, investors’
desire for insurance is an externality.15
In the simplified world of the model,
two main types of market outcomes can
arise. One is that originators of low-risk
loans retain a junior tranche while
originators of high-risk loans retain
nothing and fully securitize their loans.
This outcome is called a separating
equilibrium because originators of loans
with different likelihoods of default—
low-risk or-high risk—retain different size
claims. The idea is that originators of
low-risk loans want to signal to investors,
via retention, that the loans are indeed of
high quality in order to get a higher price
in return. In this outcome, the prices of
the securities accurately reflect the quality
of the underlying loans, so originators
have a strong incentive to make more
effort to reduce risk. Fully informative
prices also help investors better align
their portfolios with their risk preferences.
In a separating equilibrium, speculators
have no role to play because prices are
already fully informative and nobody can
make a profit by examining securitized loans to buy cheap and
sell dear.
The other type of outcome is that all originators hold an
identical claim—a pooling equilibrium. Notably, all originators can
simply sell off their entire claim to willing investors, in which
case the pooling equilibrium could involve no skin in the game.16
Why might such a situation arise? Remember, an originator
of a low-risk loan is motivated to retain a larger share of a junior
tranche only by the prospect of fetching a sufficiently higher
price. That is, the originator of a low-risk loan wants only to
know: Will retaining a big enough share of the loan to convince
investors that the risk of default is small produce a higher total
return than selling the full share? In a pooling equilibrium,
speculators have an incentive to expend the time and effort to
examine the loan and profit from trading on their superior information. So, in equilibrium, the price of a low-risk loan will be

How
accurately
prices
reflect the
underlying
risk of
default is
important
to investors, who
typically
try to
align their
portfolios
according
to their
risk preferences.

Banking Trends: Skin in the Game in the CMBS Market

2018 Q1

Federal Reserve Bank of Philadelphia
Research Department

13

higher than the price of a high-risk loan, even if the
originator retains no exposure. Indeed, if the price
of the high-risk loan is high enough, the originator of
a low-risk loan will prefer to have no skin in the game,
just like all other originators.
This pooling equilibrium outcome has some
undesirable features. Although speculators increase
the price of low-risk loans compared with high-risk
loans, the prices are noisy because speculators are
not fully informed about loan quality. So, investors
are still unable to make informed portfolio decisions
to achieve their desired level of insurance.17 Also,
there is too much default because originators typically
exert too little effort to lower risk.
In this theory, the government can actually do
a better job of reducing losses and improving
investors’ ability to choose investments that reflect
their risk tolerance than unregulated markets can.
One way it could do so, the authors suggest, would be
to require originators to structure their claims in
a way that would ensure a separating equilibrium.
Specifically, the regulator could allow originators to
choose between two set percentages of the security,
a larger junior share and a smaller junior share,
but they would have to retain one or the other. The
regulator would size the required shares in such
a way that the low-risk originator will prefer to retain
the larger share and the high-risk originator will
prefer to retain the smaller share. Because originators
of low-risk loans know that default is less likely,
they are willing to accept the larger junior share’s
higher exposure to loss in exchange for a higher
price, and vice versa. In effect, by restricting the set of
choices available to issuers, the regulator would
lead market participants to coordinate on a separating
equilibrium.18 Furthermore, unlike in the separating equilibrium in an unregulated market, even the
originator of a high-risk loan would retain at least
a small share, thus increasing the high-risk originator’s
amount of effort.
Alternatively—and perhaps more realistically—
regulators could require that all originators hold
a share of the junior tranche above some minimum
level. That is, regulators could feasibly enforce
a pooling equilibrium in which all originators would
have skin in the game—as now required under Dodd–
Frank. Theoretically, requiring originators to retain
a share of the junior tranche would motivate originators to tighten their lending criteria, leading to fewer
losses than in an unregulated market.
Another theoretical study emphasizes an additional
factor—the importance of investors’ ability to observe
originators’ actual exposure. As we will see in the
next section, even if they initially retain some risk,
originators are disposed to find ways to minimize
their exposure, for example, secretly hedging against
the risk of loss. This ability to subsequently shed their

exposure without investors knowing it could reduce
originators’ incentive to do their due diligence. In
a model developed by Victoria Vanasco, even when
originators can’t secretly reduce their initial exposure,
outcomes similar to those identified by Chemla and
Hennessey arise.19 But if investors are unable to monitor whether an originator has retained its exposure,
particularly bad outcomes arise because originators
can no longer use their retained share to convey
information to investors. Vanasco’s model suggests
that preventing such hedging also requires regulation.

14

Banking Trends: Skin in the Game in the CMBS Market

Federal Reserve Bank of Philadelphia
Research Department

Evidence Shows Skin in the Game
Improves Quality

There is empirical evidence that skin in the game mattered in the commercial mortgage-backed securities
market leading up to the financial crisis. Furthermore,
evidence from the CMBS market is consistent with
theoretical models such as we described above that
indicate that issuers may hold too little skin in the
game when markets are unregulated.

Evidence from the CMBS Market
In a segment of the CMBS market known as the
conduit market, before a deal is completed, the junior
claim, known as the B-piece, is typically sold to
sophisticated investors who specialize in evaluating
the quality of the underlying collateral. B-piece
investors are seen as the last underwriters of a deal
before it is issued and generally gather as much
information about the quality of the underlying loans
as the originators do. They also control which loans go
into the pool underlying the deal. During negotiations
with issuers, B-piece investors may insist on restructuring the securitization by, for example, throwing
out loans that they find are priced incorrectly. So, in
principle, the willingness of such well-informed
investors to hold a share of the junior claim should
play a key role in ensuring the quality of the issuance.
To find out if that is so, Adam Ashcraft, Kunal
Gooriah, and Amir Kermani examined this B-piece
market in the years before the crisis. They measured
the performance during the crisis of deals originated
from 2000 to 2007 in which B-piece buyers in turn
sold off their share of the most junior, lower-rated
(BBB) tranche versus deals in which they retained
their share. The rise of collateralized debt obligations
(CDOs) and other new financial instruments in the
early 2000s enabled B-piece buyers to sell off their
shares of the junior tranche and rid themselves of
the risk in the underlying loans.20 The authors argue
that investors in the BBB tranche had no way of
knowing whether a B-piece buyer had sold off its exposure. Issues in which B-piece investors had sold off
their claims performed poorly compared with issues

2018 Q1

The Conduit Market
The conduit market is
a part of the commercial mortgage-backed
securities market
that includes only
those commercial
mortgages that are
originated purely to be
securitized (Figure 1).
Conduit commercial
mortgage-backed
securities typically
consist of a diverse
pool of 25 to 100
commercial mortgages that have
higher leverage and
lower quality than
investment-grade
loans. Most conduit
transactions include
a B-piece buyer,
although this is not
a universal requirement. Lenders in the
conduit market
include life insurance
companies, pension
companies, investment banks, and large
commercial banks.

in which they retained their exposure, a finding that
supports the argument that skin in the game helps
reduce agency problems.
Also, the prices that general investors paid for
the securities were not sensitive to how much of the
junior tranche the B-piece investors had retained,
which supports the authors’ claim that investors
were unaware of the B-piece buyers’ true exposure.21
This result also supports Vanasco’s emphasis on the
benefits of regulations requiring issuers to maintain
their exposure.

Evidence from the RMBS Market
Two studies of another segment of the mortgagebacked securities market, residential mortgage-backed
securities, further support the importance of skin in
the game in securitization markets. Taylor Begley
and Amiyatosh Purnanandam show that private-label
residential mortgage-backed securities deals performed better when the issuers held a larger share of
the junior claim.22 The more opaque the security, the
stronger this positive effect. That is, if a security was
backed by home loans that only the issuer could have
useful information about—such as in so-called no-doc
mortgages that became popular leading up to the
housing crash in which there is no documentation of
borrowers’ creditworthiness—the bigger the issuer’s
retained share, the better the issue performed.
Another study also suggests that originators of residential mortgage-backed securities will make
a stronger effort to ensure that loan quality is high if
their exposure to losses on the loans is greater. Cem
Demiroglu and Christopher James found that deals
in which the originators of the underlying loans were
affiliated with the issuers of the securities experienced
fewer losses compared with deals in which the
originators and issuers were not affiliated.23 Similar
to the previous study, affiliation was more important to
the performance of securitizations with a large fraction of low-doc loans, ones for which documentation
was limited. Interestingly, the study also found that
for deals in which originators did not retain a portion
of the junior claim, yields were significantly higher
compared with those in which originators kept a portion. In other words, since investors knew that
originators had no skin in the game, they expected
the loans to be riskier and demanded higher rates of
return for taking on that risk.
Unlike the study of the CMBS market by Ashcraft
and his coauthors, both of these studies of the RMBS
market find that securities prices are sensitive to
issuers’ exposure to risk. This finding suggests that
investors were aware of RMBS issuers’ incentives
and knew how exposed to loss issuers were. Thus,
investors in the RMBS market were sophisticated, as
in Chemla and Hennessy’s model.

In contrast, investors in CMBS may have been
unaware that B-piece buyers now had wider opportunities to off-load their exposure, because the instruments such as CDOs that afforded those opportunities
were still a novelty. Armed with the knowledge of
what occurred during the crisis, CMBS investors now
may naturally be inclined to monitor for themselves
how much exposure B-piece buyers’ are retaining
without the need for an explicit regulation limiting
B-piece buyers’ ability to hedge their risk.

Effects of the Regulation

It is too soon to determine whether the risk retention
rule has improved the performance of the underlying
loans in the CMBS market or whether the restrictions
will hamper CMBS issuance in the long run. So far,
contrary to fears expressed by some market observers, issuances are up and deals have been priced
favorably. According to market data provider Trepp,
CMBS issues totaled $70.65 billion in 2016. Between
January and August 2017, they had reached almost $64
billion and were on pace to surpass their 2016 volume.
The risk retention rule appears to have changed
issuers’ behavior, perhaps in unintended ways. The
mixture of funding structures used for CMBS issues in
the first half of 2017—38.2 percent horizontal retention,
37.5 percent vertical, and 24.3 percent L-shaped24—
shows that given the choice, issuers will not always
choose horizontal, which, according to Chemla and
Hennessy and other models, is the optimal structure.
Furthermore, market participants have argued
that the new regulation has led to other changes in
the CMBS market. In some cases, issuances that
previously would have contained an entire large loan
are being replaced by multiple, smaller issuances
that each contain only a portion of a larger loan, with
each small deal having a different issuer. This allows
issuers to limit their potential losses, since the amount
an issuer is required to retain for a small security is
less than it would be on a large security.25 Note that
spreading a loan across multiple CMBS deals in this
way means more claimants if the loan defaults, which
could complicate the resolution effort. Only in the
next downturn in commercial real estate will it become clear whether this will be a significant problem.
In 2017, Treasury Secretary Steven Mnuchin
recommended expanding the definition of qualifying
exemptions based on the characteristics of the
securitized asset class and relaxing restrictions on thirdparty purchasers of the junior claim.26 The future
of Regulation RR may ultimately depend on not just
loan performance in the CMBS market but on whether
that provision in Dodd–Frank is rolled back.

Banking Trends: Skin in the Game in the CMBS Market

2018 Q1

Federal Reserve Bank of Philadelphia
Research Department

15

Notes

the issuer of the securities as a single agent. Throughout this section we
use only originators to avoid confusion.

1 See Ronel Elul’s 2012 article for a discussion of these models.
2 Unlike a residential mortgage-backed security, which is backed by
hundreds of individual home mortgages that are evaluated using a
common model, a commercial mortgage-backed security is backed by
far fewer loans that are evaluated on a case-by-case basis based on
characteristics such as location and property type.
3 Most would-be purchasers tend to rely on evaluations such as risk
ratings from a major credit ratings agency. As became clear during the
crisis, when even securities with top ratings went bad, the use of ratings
agencies does not fully overcome these agency problems.

12 In the simplified world of the model, each originator screens a single
loan.
13 We focus on the cases in which originators retain a junior tranche if
they retain any share at all. In Chemla and Hennessy’s model, cases
can arise in which originators retain a senior tranche, a particularly bad
outcome.
14 In their model, only low-risk loans are profitable to buy. However,
speculators can also profit from taking short positions in securities
backed by high-risk loans.

4 See Sheridan Titman and Sergey Tsyplakov’s article.
5 From “Skin in the Game: Aligning the Interests of Sponsors and
Investors.”
6 Although the regulation governs all asset-backed securities, we focus
on those portions that affect CMBS. The regulation defines an issuer—
otherwise known as a sponsor—as “a person who organizes and initiates
an asset-backed securities transaction by selling or transferring assets,
either directly or indirectly, including through an affiliate, to the issuing
entity.”
7 The intention of this restriction is to make sure that issuers do not
undermine the incentive effect of skin in the game by selling off the risk
without actually selling the junior security. For example, an issuer of
CMBS might hedge the risk that real estate prices will drop by buying
a credit default swap, in which the seller compensates the issuer when
real estate prices fall. The regulation does not specifically define hedging.
Whether a particular hedge is permissible will be determined in practice
over time on a case-by-case basis.
8 There are also exceptions for issuers of residential mortgage-backed
securities that are composed of conforming loans—loans that meet
the underwriting standards of the government-sponsored enterprises
(GSEs)—are exempt. In practice, this means that most of the residential
mortgage-backed securities market is exempt. The regulation may
become binding for a larger share of the RMBS market if and when the
private-label mortgage-backed securities market regains strength.
9 The regulation defines a qualifying commercial real estate loan as
a fixed-rate loan with a minimum maturity of 10 years and a maximum
amortization of 25 years (30 for loans secured by multifamily properties).
Lenders must document the income from the property for at least the
previous two years. The borrower’s debt service ratio must not exceed
1.25 percent for multifamily properties, 1.5 percent for leased properties,
and 1.7 percent for all other loans. Also, the combined loan-to-value
ratio of all loans on the property cannot exceed 70 percent, and the loanto-value ratio of the first lien loan cannot exceed 65 percent.
10 Furthermore, allowing the issuer to share risks with the originator
might reduce the cost of issuance.
11 Chemla and Hennessy simplify and treat the originator of the loan and

16

Federal Reserve Bank of Philadelphia
Research Department

15 According to Chemla and Hennessy, another externality, or failure to
price in the true cost, is the neighborhood effect of loan defaults. For
example, when the failure to screen a borrower’s creditworthiness results
in a boarded-up foreclosed property, neighboring home values may also
drop.
16 To be precise, the authors demonstrate that whenever a pooling
equilibrium is possible, a pooling equilibrium in which originators hold no
skin in the game is also possible.
17 We are simplifying Chemla and Hennessy’s analysis of the relative
efficiency of the different equilibrium outcomes. For example, in their
model, we can’t automatically conclude that a separating equilibrium
is better than a pooling equilibrium, although this is one possibility. While
the incentives are typically smaller in a pooling equilibrium for an
originator to make an effort, speculative buying and selling may lead to
a price difference large enough to prompt them to make the effort.
18 In the separating equilibrium designed by the regulator, both types of
originators hold skin in the game.
19 Note, in Chemla and Hennessey’s model, there is no issue of originators
choosing some level of retention initially and subsequently selling off the
exposure without investors’ knowledge.
20 Banks may repackage certain loans in a security into CDOs, which are
then sold to investors on the secondary market. CDOs usually consist of
a pool of loans from the lowest tranches in a securitization.
21 The insensitivity might also be consistent with investors’ not understanding the importance of skin in the game. However, the empirical
studies we discuss below cast some doubt on this explanation.
22 Private-label securitizations are those set up by firms other than
government-sponsored enterprises, such as Fannie Mae or Freddie Mac.
Begley and Purnanandam study a sample of private-label securitization
contracts in 2002 and 2004–2005.
23 Originators can also be the issuers in a securitization transaction. In
this study, originators were considered affiliated with the deal if they were
also the issuer or if they retained the servicing rights to the transaction.
An example of an unaffiliated originator is a loan broker. The broker

Banking Trends: Skin in the Game in the CMBS Market
2018 Q1

underwrites the loan and typically sells it to a bank that will assemble it
with other loans into a security.
24 From Trepp’s Q1 and Q2 2017 CMBS Issuance Recaps.
25 See the American Banker article. The extent to which this development
is actually due to regulatory changes is uncertain. While the number of
securitizations broken up in this way increased substantially in 2017, this
was an acceleration of a trend that began in 2011.

Demiroglu, Cem, and Christopher James. “How Important Is Having Skin
in the Game? Originator-Sponsor Affiliation and Losses on Mortgagebacked Securities,” Review of Financial Studies, 25 (November 2012),
pp. 3,217–3,258.
Elul, Ronel. “The Economics of Asset Securitization,” Federal Reserve
Bank of Philadelphia Business Review (Third Quarter 2005).
Elul, Ronel. “Securitization and Mortgage Default,” Journal of Financial
Services Research, forthcoming.

26 See the Treasury Department report.
Financial Stability Oversight Council, “2011 Annual Report” (2011).

References
Aguilar, Luis A. “Skin in the Game: Aligning the Interests of Sponsors and
Investors,” U.S. Securities and Exchange Commission Public Statement
(October 2014).
Ashcraft, Adam B., Kunal Gooriah, and Amir Kermani. “Does Skin-in-theGame Affect Security Performance?” working paper (October 2014).
Begley, Taylor, and Amiyatosh Purnanandam. “Design of Financial
Securities: Empirical Evidence from Private-label RMBS Deals, Review of
Financial Studies, forthcoming.

Office of the Comptroller of the Currency, Federal Reserve Board, Federal
Deposit Insurance Corporation, Federal Housing Administration,
U.S. Securities and Exchange Commission, and Housing and Urban
Development. “Final Rule on Credit Risk Retention,” Federal Register,
79:247 (December 24, 2014), pp. 77,602–77,766.
Titman, Sheridan, and Sergey Tsyplakov. “Originator Performance, CMBS
Structures, and the Risk of Commercial Mortgages,” Review of Financial
Studies, 23:9 (September 2010), pp. 3,558–3,594.
Trepp. “Q2 2017 CMBS Issuance Recap” (August 2017).

Bisbey, Allison. “Large CMBS Loans Are Being Carved Up to Spread Risk,
but…,” American Banker (May 2017).

U.S. Department of the Treasury. “A Financial System That Creates
Economic Opportunities: Capital Markets” (October 2017).

Chemla, Gilles, and Christopher Hennessy. “Skin in the Game and Moral
Hazard,” Journal of Finance, 69:4 (August 2014), pp. 1,597–1,641.

Vanasco, Victoria. “The Downside of Asset-Screening for Market Liquidity,”
Journal of Finance, forthcoming.

Banking Trends: Skin in the Game in the CMBS Market

2018 Q1

Federal Reserve Bank of Philadelphia
Research Department

17

Nontraditional Insurance
and Risks to Financial Stability
Do insurance companies pose a threat to financial
stability? Historically, the answer has been no. But
the insurance industry’s expansion into nontraditional
activities has prompted reconsideration.

Yaron Leitner is a senior
economist at the Federal
Reserve Bank of Philadelphia.
The views expressed in this
article are not necessarily
those of the Federal Reserve.

BY YA RO N L E I T N E R

W

hen we think of the U.S. insurance business, we usually
think of companies that sell life, auto, or homeowner
policies. The conventional wisdom is that these
traditional insurance activities are regulated by the states largely
to protect individual policyholders and should not be a concern
to the Federal Reserve, whose regulation of banks is intended to
protect the nation’s overall financial stability.
However, as became clear during the emergency bailout of the
insurer American International Group (AIG) during the financial
crisis in 2008, some insurance companies also engage in nontraditional activities, such as selling credit default swaps or lending
securities, that could pose a threat to financial stability. The
AIG episode has led some to suggest that the Fed should become
involved in the regulation of large insurance companies.
How could an insurer pose a threat to financial stability? While
there are many reasons that an institution could pose a threat to
financial stability, two factors seem key. First, the institution’s
activities leave it vulnerable to large losses that it cannot handle.
Second, those losses are capable of spreading to the rest of the
financial system via a domino effect, or contagion.1 As we will
see, traditional insurance activities do not satisfy these criteria,
but nontraditional activities do.
To examine more closely why they could pose a threat to the
nation’s financial system, we will explore some of the nontraditional activities that insurance companies currently engage
in and discuss what role, if any, the Fed should play in regulating these companies. But before we do that, it will help to
understand why insurers’ traditional activities do not pose such
a threat.

Traditional Insurance Risk

A traditional insurance company providing, say, auto insurance,
collects premiums from policyholders and in return promises
to pay for part or all of their loss when an accident occurs. So
the insurance company incurs the risk that accidents will occur.

18

Federal Reserve Bank of Philadelphia
Research Department

With another traditional insurance product, a life annuity, the
policyholder pays a premium in return for periodic payments
later—usually beginning in retirement and lasting until the
policyholder dies. Here, the insurance company incurs the risk
that the policyholder will live long enough to more than break
even on what he or she paid in premiums.
However, these traditional activities do not expose the insurance company to large losses that it cannot handle. From the
insurance company’s perspective, these risks are pretty much
diversifiable. When an insurance company sells many insurance
policies, losses are more predictable and are unlikely to depend
on overall economic conditions. The insurance company
can then use the premiums it collects from policyholders to make
investments that mature when it expects to need to pay claims.
To the extent that losses are not completely predictable,
insurance companies also set aside money to cover unexpected
losses. So unless insurance companies deliberately leave
themselves underprepared, they are not expected to experience
losses that they cannot handle.2
Moreover, in the traditional model, insurance companies do
not offer deposit contracts and so are not as subject as banks
are to runs, in which an unusually large number of depositors
try to withdraw their money simultaneously. Bank runs can be
triggered if depositors learn about some negative shock—say,
a news report that the dominant local industry is shuttering its
plants, which could mean deep losses for a bank heavily exposed
to commercial real estate. If the shock casts doubt on the soundness of the bank’s portfolio and undermines public confidence in
its ability to meet its obligations, a run may ensue.
While withdrawals are a normal part of the business of banking, in the case of a run the bank’s cash reserves may be
insufficient to meet the sudden demand. The fundamental reason
that nations regulate their banks is that banks’ unique mixture
of assets and liabilities is inherently unstable. Banks are in the
business of holding illiquid and long-term assets that they fund
largely with deposits and other short-term liabilities. Banks

Nontraditional Insurance and Risks to Financial Stability
2018 Q1

typically keep only a small percentage of their deposits on hand
as cash and use most of the money they take in to make loans
and invest in financial assets. And many of their assets—such as
commercial and industrial loans or commercial real estate loans—
can’t be easily sold on short notice. A surge in simultaneous
withdrawals could force the bank to sell off those assets quickly
at prices that are significantly below normal, lowering their
value to the point that they are insufficient to pay off the bank’s
liabilities, causing it to go bankrupt.
For a traditional insurance company, by contrast, even if it
does suffer losses that it cannot handle, they are unlikely to spill
over to the rest of the economy, for two reasons. First, traditional
insurance activities do not significantly expose the rest of the
financial system to insurers.3 Second, the unpredictable losses the
insurance industry might face from traditional activities (after,
say, a hurricane) are typically uncorrelated with overall economic
conditions, and so the financial system is likely to be stable
enough to absorb them. In contrast, large losses that occur when
overall economic conditions are bad and many other financial
institutions are experiencing losses at the same time are a concern
to financial stability because the financial system might not be
able to absorb them. As we will explore next, some of insurers’
nontraditional activities expose them to such losses.

large withdrawals from the fund, leading the value of a share in
the fund to fall to 97 cents per $1 invested. Such an unanticipated
drop in the value of what was supposed to be a safe investment
created panic and led investors to withdraw their money from
other money markets funds, even ones that had not invested in
Lehman or AIG.
The CDS that AIG sold also created links between it and the
large financial institutions that bought the swaps. Absent a bailout,
the failure of AIG, or even the anticipation of such a failure,
could have led to large losses
FIGURE 1
for these institutions, as they
would have lost the protection
Aid to AIG's CDS
offered by the CDS contracts.
Counterparties
$, billions.
Losses could then have spread
to other large institutions
Societe Generale
Goldman Sachs
connected to these institutions.
Deutsche Bank
One indication for the potential
Merrill Lynch
losses to AIG’s counterparties is
0
6
12
the amount of government
aid ($49.5 billion) that went to
AIG’s CDS counterparties, including Societe Generale ($11 billion),
Goldman Sachs ($8.1 billion), Deutsche Bank ($5.4 billion), and
Merrill Lynch ($4.9 billion).8

Risks from Insurers’ Nontraditional Activities

Securities Lending
Another nontraditional activity that contributed to AIG’s failure
was securities lending. In securities lending, a financial institution
such as an insurance company lends a security to another
financial institution in exchange for collateral, typically cash.
The borrower generally can return the borrowed security to the
lender and receive its collateral back on short notice, without
penalty.9 As long as the lender, in this case the insurance company, invests the cash collateral in conservative short-term assets,
there is no risk to financial stability, because the insurance
company is able to return the cash collateral to the borrower on
a short notice. However, a risk arises when the securities lender
invests the cash collateral in long-term and less-liquid assets such
as corporate bonds or mortgage-backed securities. The AIG case
illustrates this risk.
AIG loaned securities, primarily corporate bonds, to banks and
broker dealers. Between 2005 and 2007, rather than invest the
cash collateral it received from the borrowers in conservative,
short-term securities, without notice AIG changed the direction
of its investment strategy and invested a substantial portion of
the cash collateral in long-term illiquid assets such as mortgagebacked securities, other asset-backed securities, and collateralized
debt obligations, whose payoffs depended on the health of the
housing market. At the end of 2007, 65 percent of AIG’s securities
lending collateral was invested in such securities, and only 16
percent was in cash or other short-term investments.10 As the
value of these securities dropped, and as AIG’s losses on its CDS
portfolio mounted, the borrowers in AIG’s securities lending
portfolio wanted to reduce their exposure to AIG, and so they
began to return the borrowed securities to AIG and demand
the return of their cash collateral. Between just September 12
and September 30, 2008, securities lending counterparties

Credit Default Swaps
A prominent example is AIG’s credit default swaps (CDS) operations before it failed in 2008. AIG sold these financial instruments
to other financial institutions as protection against losses
resulting from mortgage defaults. So AIG was essentially betting
against a decline in real estate prices, offering protection against
risks that it could not diversify and exposing it to potentially
large losses. Indeed, as home prices started to decline, AIG was
required to post cash collateral with its CDS counterparties to
guarantee that it could fulfill its contractual obligations.4 Since AIG
was unable to come up with all the money, its credit rating was
downgraded, which required it to post even more collateral,
making its situation even more precarious.5
AIG’s losses could have spread to the rest of the financial
system. One channel through which they could have spread was
that a number of money market mutual funds had invested in
AIG’s commercial paper—short-term loans used to raise immediate
cash—exposing them to AIG.6 Money market mutual funds are
considered a safe and liquid investment, and until regulatory
changes were implemented after the crisis, the share price of
a dollar invested in the funds remained constant at $1.7 However,
as AIG’s need for cash grew, its connection with money market
investors raised concerns that if it declared bankruptcy and
defaulted on its commercial paper, the money market funds
could “break the buck,” potentially triggering runs on them and
other money funds. Indeed, following Lehman Brothers’ failure
the day before AIG was bailed out, a money market mutual fund
with more than $60 billion in assets, the Reserve Primary Fund,
broke the buck. The value of the company’s $785 million in
holdings of Lehman Brothers dropped to zero, which triggered

Nontraditional Insurance and Risks to Financial Stability
2018 Q1

Federal Reserve Bank of Philadelphia
Research Department

19

demanded that AIG return approximately $24 billion in cash.11 In
other words, AIG experienced a run.
A run triggered by securities lending is a concern to financial
stability because it forces the insurance company to sell its assets
quickly at fire-sale prices, leading to losses for the insurance company.12 Other financial institutions that hold the same class of assets may then have to mark down the value of their assets, which
could force them to sell assets quickly to ensure that their capital
does not fall below the minimum level required by the regulator.
As more companies sell assets, the reduction in asset prices is
amplified, which can affect the whole economy. For instance,
when the value of their assets drops, financial institutions may
be reluctant to make loans to businesses and consumers.13
Securities lending also creates direct links between the
insurance company and other financial institutions, which can
further spread losses to the rest of the financial system. The
borrower faces a counterparty risk that the insurance company
will be unable to return the collateral. So the failure of the
insurance company may spread to the borrower and other firms
that are connected to the borrower. Indeed, $43.7 billion of AIG’s
government aid went to AIG’s securities lending counterparties.14
Insurance companies, mostly life insurers, continue to engage
in securities lending. Moreover, life insurance companies
continue to invest a large portion of the cash collateral received
in potentially illiquid long-term assets, such as corporate bonds
and private-label asset-backed securities (Figures 2 and 3).
Empirical evidence suggests that securities lending by life insurers
is at least partially driven by a desire to take on more risk.15
Insurance companies also sell other financial products that could
expose them to runs.16

Captive Reinsurance
A final example of a nontraditional activity that could pose risks
to financial stability is captive reinsurance. In a typical captive
reinsurance transaction, the insurance company obtains
insurance from an affiliated (“captive”) company that is subject
to lower reserve and capital requirements and that in most cases
is not required to file public financial statements or follow the
same regulatory accounting practices as primary insurers. Thus,
captive reinsurance allows the insurance company as a whole
to hold less capital, even though there is no reduction in risk.
(The company that purchases reinsurance is called the ceding
company: It cedes its liabilities to the reinsurer.)
Captive reinsurance grew rapidly from $90 billion in 2002 to
$572 billion in 2012 (Figure 4).17 Initially, the growth in captive
reinsurance was mainly in life insurance products and was probably triggered by a new regulation requiring insurance companies
to hold more reserves against these products.18 New state laws after 2002 allowed life insurers to establish captives to circumvent
these new reserve requirements. Since 2007, captive reinsurance
for annuity products has also grown rapidly, even though reserve
requirements for these products were not changed.
Particularly worrisome is the rapid growth in shadow insurance, in which the captive is not supervised by the ceding
company’s state and has not been rated by an insurance rating
agency.19 Shadow insurance grew from $11 billion in 2002 to $370
billion in 2013 (Figure 5). States compete for captive business to
increase employment and tax revenue. The state where the
captive is located does not directly bear risk, because when
a captive fails, the liabilities revert to the operating company
and, ultimately, to the guarantee associations operated by the
states in which the policies were sold. Since 2009, the growth
of shadow insurance has slowed, partly because of more

FIGURE 2

FIGURE 3

FIGURE 4

Life Insurers Remain
Involved in Securities
Lending

Much Collateral Is Reinvested in Illiquid
Assets

Rapid Growth in Captive
Reinsurance

U.S. life insurance industry
securities lending.
$, billions

% of total collateral reinvested, 2015

Total life and annuity products.
$, billions

Cash and cash
equivalents

600

Corporate bonds
Other

60

500

ABS and other
structured securities

50

450

Agency RMBS

400

Short-term assets

40

U.S. governement
bonds

30
20

Captive

550

350
0%

15%

30%

300

Unaffiliated

250

Source: National Association of Insurance Commissioners,
http://www.naic.org/capital_markets_archive/160909.htm.

200

Note: Life insurance industry, as of December 31, 2015.

150
100

10

50
0

2011

0

2016

Source: SNL Financial.

20

Federal Reserve Bank of Philadelphia
Research Department

2002

2007

Source: Koijen and Yogo, 2016.

Nontraditional Insurance and Risks to Financial Stability
2018 Q1

2012

regulatory scrutiny in states such as California and
New York.20
When a captive reinsurer is unauthorized in
a state, the ceding insurer may reduce its statutory
reserves, and hence boost its capital, only if the
reinsurer posts collateral or receives a third-party
guarantee such as a letter of credit from a bank. However, as noted in a New York State Department of
Financial Services report21, in many cases the collateral was just a “contractual parental guarantee” in
which the parent company was responsible to cover
losses. So, the insurance company boosted its capital
artificially without reducing risk.22
Captive insurance, and in particular, shadow insurance, poses concerns for financial stability. First,
there is no real reduction of risk, yet the company
as a whole holds less capital. This means that the
company might be exposed to losses that it cannot
handle. Second, the use of bank letters of credit as
collateral exposes the insurance company to the
risk that the bank will not renew its letter of credit;
usually, these letters of credit have shorter maturities than the insurer’s liabilities do. So, the banks
issuing the letters of credit may run on the insurance
company. Third, these letters of credit create links
with banks, exposing banks to potential losses from
the insurance industry.

Should the Fed Help Regulate Insurers?
Under the Dodd–Frank Wall Street Reform and
Consumer Protection Act of 2010, the Fed can impose
FIGURE 5

Surge in Shadow Insurance Is Particularly
Worrisome
$ billions
400
350
300
250
200
150
100
50
0

2002

2013

Source: Koijen and Yogo, 2017.

Risks from Variable Annuities with Minimum
Guarantees
Another nontraditional activity that has received much attention from policymakers and economists is the sale of variable annuities with minimum
guarantees. This activity does not create direct links between insurers and
other financial institutions, but it could expose insurers to large losses in the
event of a deterioration in overall economic conditions.
A variable annuity is a hybrid of a traditional life annuity and a mutual
fund. Variable annuities are long-term saving products. But in contrast to
traditional annuities, policyholders’ money is invested in mutual funds that
fit their risk appetite. Their investment accounts are kept separate from
the company’s general account, and payments are drawn only from these
separate accounts. So, while this product is riskier for the annuity holder, it
poses no financial stability concerns.
However, things change when the variable annuity is joined with a minimum guarantee. A particular concern are the guaranteed living benefits,
which are optional riders that policyholders can obtain for an additional
fee, guaranteeing they will receive some minimum income (or be able to
withdraw some minimum amount) regardless of how well their mutual
fund investments actually perform. These guarantees, which are backed by
the insurance company’s general account assets, are a concern to financial
stability because the insurance company provides protection against risks
arising from worsening conditions in the overall economy. For example,
these guarantees may kick in during an economic downturn, as when equity prices drop, adding stress to an already-stressed economy.
Indeed, as Ralph Koijen and Motohiro Yogo have documented, during the
financial crisis in 2008, the variable annuity business experienced significant losses because of failing stock prices, high volatility, and low interest
rates, with two companies,
FIGURE 6
Hartford Life and Manulife
Fast
Rise in Variable Annuities
Financial, losing about half
with
Minimum Guarantees
of their capital and surplus.
Dollar
value of variable annuities with
Across the industry, life
guaranteed
living benefits.
insurers with variable annuity
1000
guarantees lost 9 percent
of their capital and surplus,
while those without guaran800
tees gained 1 percent.
Since the crisis, the estimated
total outstanding account
value of all variable annuities
with guaranteed living benefits has risen rapidly, from
$292 billion in 2008 to $843
billion in 2014 (Figure 6).
Rapid growth of an activity
is a particular source of
regulatory concern because
it suggests that risks may
not have been fully priced in.

Note: Shadow insurance refers to a subset of captive insurance.

Nontraditional Insurance and Risks to Financial Stability
2018 Q1

600
400
200
0

2008

2014

Source: Financial Stability Oversight Council
2015 Annual Report/ LIMRA.
Note: Data based on a survey of domestic
insurance companies.

Federal Reserve Bank of Philadelphia
Research Department

21

stricter regulations on insurance companies and other nonbank
financial institutions that the Financial Stability Oversight Council designates as systemically important (SIFI). A company can be
designated as systemically important if material financial distress
at the company, or the nature, scope, size, scale, concentration,
interconnectedness, or mix of its activities could pose a threat to
the financial stability of the United States.23
There have been calls to repeal the council’s authority to
designate firms as SIFIs.24 One concern is that the council has too
much discretion in designating an institution as a SIFI, which
could result in arbitrary and inconsistent designations. Another
concern is that market participants might interpret a SIFI
designation as a signal that the government considers the institution too big to fail and will bail it out if it gets into trouble
and the threat of contagion arises, which could create moral
hazard by undermining market discipline. That is, if everyone
expects the Fed to bail out a systemically important insurer,
it will take excessive risks, and its policyholders and counterparties will have no incentive to monitor it closely or take steps
to reduce that risk.
Another set of concerns involves the principle of state control.
Some argue that Fed involvement is unnecessary because state
regulation is adequate. Indeed, since the financial crisis, state regulators, in particular the National Association of Insurance
Commissioners (NAIC), have taken steps to reduce risks in the
insurance industry. Some proponents of state oversight see some
role for the Fed, but only insofar as nontraditional activities that
pose systemic risk, and support leaving the rest of insurers’
activities to state regulators.25
The question for policymakers is how to weigh these concerns

against the risk to financial stability from insurers’ nontraditional
activities. The Fed’s mission includes guarding the stability of
the U.S. financial system, and insurance companies are large
institutions that play a large role in the economy. As we have seen,
they engage in nontraditional activities that could pose a threat
to financial stability, and there is evidence of their having engaged
in risk-taking and regulatory arbitrage—which make their potential
threat to financial stability even larger. A quantitative measure of
systemic risk, SRISK,26 that estimates a financial institution’s
capital shortfall during a crisis, ranks insurance companies among
the most systemically risky financial institutions in the U.S.27
Interestingly, since 2008, SRISK has declined significantly for large
banks but has increased for large insurance companies except
AIG (Figure 7).
Those who argue that federal regulation
See Risk-Taking
is necessary note that an individual insurand Regulatory
ance company does not take into account
Arbitrage in
the negative consequences of its failure
the Insurance
on the rest of the economy. Likewise, an
Industry.
individual state does not take into account
the consequences of its actions for other
states. Individually or collectively, the states are not responsible
for the stability of the U.S. financial system. The aforementioned
desire to preserve states’ longstanding role in insurance regulation has led to a search for a middle ground that would feature
federal regulation of insurers’ nontraditional activities and state
regulation of traditional insurance activities. Unfortunately, statefederal regulation may prove difficult in practice, as nontraditional
and traditional insurance activities are deeply intertwined.
For example, insurers use general account assets to back both

FIGURE 7

Insurers Among Most Systemically Risky U.S. Financial Firms

SRISK capital shortfall if the S&P 500 index falls more than 40 percent over the next six months, $ billions.
SRISK measurements on August 18, 2017
Metlife
Bank of America

150
120

Prudential

60

Citigroup

0

Morgan Stanley

Principal

−60

Goldman Sachs

−120

Lincoln
Principal
0

10

20

30

150
120

40 50

Source: New York University Stern School of Business
V-Lab, https://vlab.stern.nyu.edu/en/welcome/risk/.
(Select 8 percent capital target ratio, no simulation.)
Note: More recently, V-Lab has ranked additional
firms that do not appear in the ranking above.

22

SRISK measurements January 2, 2002–August 18, 2017
Insurers
Metlife
Prudential
Lincoln

2002

2017

Banks
Bank of America

2002
Citigroup

2017

2002

2017

Morgan Stanley

2002

2017

Goldman Sachs

60
0
−60
−120

Federal Reserve Bank of Philadelphia
Research Department

2002

2017

2002

2017 2002

Nontraditional Insurance and Risks to Financial Stability
2018 Q1

2017

2002

2017

minimum guarantees for variable annuities and traditional
insurance policies. Similarly, insurance companies lend securities
from their general accounts.
Proponents of retaining the council’s authority to designate
firms as SIFIs maintain that discretion is necessary because assessing systemic risk is too complicated to be captured by fixed
rules. Indeed, one benefit of discretion is that it allows decisions
to be made based on information that applies to the case at hand.
But will SIFI designations undermine market discipline? The
concern that an inferred bailout will relax attitudes about risk
is widely shared. Yet, market participants may already expect any
large financial institution to be bailed out, regardless of whether
it is formally designated a SIFI. The best solution from a social
point of view might be to not completely rule out bailouts but
instead to monitor and regulate SIFIs closely to reduce moral hazard. The Fed could also shift more of the onus onto systemically
important institutions by taxing SIFIs to account for the risk they
pose to the economy and the costs of potential bailouts.29 An
example of such a tax is the SIFI capital surcharge rule.30
A final potential concern that is less often raised is whether
the Fed should focus only on large insurance companies. As
we saw earlier, they are not the only ones that engage in nontraditional activities that could pose a threat to financial stability.
The aggregate potential threat to financial stability from the
nontraditional activities of numerous small insurers could be of
the same magnitude as the threat from the activities of a single
large company.

Risk-Taking and Regulatory Arbitrage
in the Insurance Industry
Captive insurance is an example of regulatory arbitrage: A company
is able to hold less capital without having to actually reduce its risk.
There is other evidence that insurance companies have engaged in
risk-taking and regulatory arbitrage.
One study by Ralph Koijen and Motohiro Yogo showed that around
December 2008, insurance companies took actions that created
losses to make them look good for regulatory purposes. Life
insurers were able to make accounting profits by selling policies at
prices that were far below actuarial fair values because the amount
of reserves they had to record on their balance sheets to cover
the future liabilities created by the new policies was less than their
selling price. So, insurance companies sold policies that technically
lost money but made accounting profits.28
In another study, Bo Becker and Victoria Ivashina showed that
within a group of bonds with the same credit rating, insurance
firm portfolios tended to hold the riskier ones.
Another sign of regulatory arbitrage comes from evidence by
Becker and Markus Opp that insurance companies invested more
in mortgage-backed securities following new regulations that
substantially reduced capital requirements on such investments.

Notes
1 To learn more about some of the channels of contagion, read my
Business Review article on financial contagion and network design.
2 In a catastrophic disaster such as a hurricane, the property-casualty
insurance industry can suffer large losses that it cannot handle on its
own, and so the government might need to intervene.
3 Note, however, that insurance companies provide an important source
of funding for banks through the corporate bond market. A reduction
in their supply of funding to banks could lead to liquidity problems for
banks, at least in the short run.
4 In CDS contracts as in most derivative contracts, counterparties post
collateral, often in the form of cash. The larger one party’s obligation to
the other, the more collateral it will be required to post.
5 Robert McDonald and Anna Paulson discuss AIG’s credit default swaps
operations in more detail. They document that the amount of cash

collateral that AIG needed to post increased rapidly, from $15.8 billion at
the end of June 2008, to $33.9 billion on September 16, the day the Fed
stepped in. The difference between the amount of collateral that AIG was
required to post and the amount it actually posted increased from $2.5
billion to $11.4 billion during that same time.
6 Note that under the traditional insurance model, insurance companies
do not need to borrow short term, so there should not be much exposure
between money market funds and insurers.
7 Under new SEC regulations, in money market funds used by institutional
investors, the daily price can fluctuate along with changes in the marketbased value of the fund assets. See more details at https://www.sec.gov/
news/press-release/2014-143.
8 See the documentation of AIG’s payments to counterparties accompanying the 2009 New York Times article by Mary Williams Walsh.

Nontraditional Insurance and Risks to Financial Stability
2018 Q1

Federal Reserve Bank of Philadelphia
Research Department

23

9 The main lenders of securities are institutional investors, such as
retirement and pension plans, mutual funds, and insurance companies, the
last of which accounted for roughly 5 percent of total lending in 2014.
The main borrowers are hedge funds, broker-dealers, derivative traders,
and market makers. Borrowers may use the borrowed security as part
of a short-selling strategy or to deliver a particular security to a customer
when they do not have the security on hand.

19 Specifically, by A.M. Best Company, which is a rating agency that
focuses on the insurance industry.
20 In their 2017 book chapter, Koijen and Yogo show that in 2013, captive
insurance was $617 billion, and shadow insurance was $370 billion. We
do not have more recent data at this point.
21 See the report by Benjamin M. Lawsky.

10 Nineteen percent was invested in corporate bonds.
11 See p. 45 of the Congressional Oversight Panel Report.
12 For evidence of fire sales in the insurance industry, see the papers
by Andrew Ellul, Chotibhak Jotikasthira, and Christian Lundblad and by
Craig Merrill, Taylor Nadauld, René Stulz, and Shane Sherlund.
13 For theoretical models that analyze this issue in more depth, see the
papers by Kiyotaki and Moore and by Brunnermeier and Pederson. In
a 2017 working paper, Nathan Foley-Fisher, Stefan Gissler, and Stephane
Verani demonstrate another side effect of the collapse of AIG’s securities
lending programs in 2008: a substantial and long-lasting reduction in
the market liquidity of corporate bonds that were predominantly held
(and hence lent) by AIG.
14 See the interactive documents accompanying the 2009 New York
Times article.
15 In their 2016 paper on securities lending, Foley-Fisher, Borghan
Narajabad, and Verani show that insurers that engaged aggressively in
maturity transformation with respect to the cash collateral they received
from securities lending tended to switch to repo financing—a form of
short-term collateralized borrowing—when borrowers’ demand was low
for the securities loaned (typically corporate bonds).
16 One example is extendible funding agreement-backed notes that
insurance companies sell to institutional investors. See the 2016 paper
on self-fulfilling runs by Foley-Fisher, Narajabad, and Verani. See also
the Chicago Fed Letter by Robert McMenamin, Zain Mohey-Deen, Anna
Paulson, and Richard J. Rosen.

18 In January 2000, the National Association of Insurance Commissioners
(NAIC) adopted Model Regulation 830, commonly referred to as Regulation
XXX. This was followed by Actuarial Guideline 38 in January 2003,
commonly referred to as Regulation AXXX.

Federal Reserve Bank of Philadelphia
Research Department

23 Dodd–Frank also specifies some of the factors that the council needs
to consider, including leverage, off-balance-sheet exposure, relationships
with other significant companies, the company’s liabilities and its degree
of reliance on short-term funding, and the importance of the company as
a source of credit for households, businesses, and state and local governments, and as a source of liquidity for the United States financial system.
24 For example, the Financial Choice Act of 2017 has called for the
abolition of the SIFI designation.
25 See, for example, the paper by Scott Harrington.
26 Specifically, SRISK estimates how much capital a firm will be short,
relative to some target level of capital (8 percent) if a broad market index
falls by more than 40 percent over the next six months. SRISK is updated
weekly on New York University Stern School of Business’s V-Lab website:
https://vlab.stern.nyu.edu/en/welcome/risk/.
27 SRISK applies only to publicly traded companies and hence does not
include some large U.S. life insurers such as TIAA-CREF, New York Life,
and Northwestern Mutual Life.
28 See Koijen and Yogo’s 2015 article.
29 This idea is formulated in the paper by Viral V. Acharya, Lasse H.
Pedersen, Thomas Philippon, and Matthew Richardson.

17 See Ralph Koijen and Motohiro Yogo’s 2016 paper.

24

22 The report mentions another way in which shadow insurance allows
insurers to boost their capital artificially: The reinsurer pays a commission
to the original insurer, which can then boost its capital artificially by
recording these commissions as “retained earnings.” So the company
increases its capital by essentially paying itself.

30 For more details on the SIFI capital surcharge rule, see https://www.
federalreserve.gov/newsevents/pressreleases/bcreg20150720a.htm.

Nontraditional Insurance and Risks to Financial Stability
2018 Q1

References
Acharya, Viral V., Lasse H. Pedersen, Thomas Philippon, and Matthew
Richardson. “Measuring Systemic Risk,” Review of Financial Studies,
30:1 (2017), pp. 2–47, https://doi.org/10.1093/rfs/hhw088.
Becker, Bo, and Victoria Ivashina. “Reaching for Yield in the Bond Market,”
Journal of Finance 70:5 (2015), pp.1,863–1,902.
Becker, Bo, and Marcus Opp. “Regulatory Reform and Risk-Taking:
Replacing Ratings,” National Bureau of Economic Research Working
Paper 19257 (2013).
Brunnermeier, Markus K., and Lasse Heje Pedersen. “Market Liquidity
and Funding Liquidity,” Review of Financial Studies, 22:6 ( 2009),
pp. 2,201–2,238, https://doi.org/10.1093/rfs/hhn098.
Congressional Oversight Panel. “June Oversight Report: The AIG Rescue,
Its Impact on Markets, and the Government’s Exit Strategy” (2010),
http://cybercemetery.unt.edu/archive/cop/20110402010341/http://cop.
senate.gov/documents/cop-061010-report.pdf.
Ellul, Andrew, Chotibhak Jotikasthira, and Christian T. Lundblad.
“Regulatory Pressure and Fire Sales in the Corporate Bond Market,”
Journal of Financial Economics 101:3 (2011), pp. 596–620.
Foley-Fisher, Nathan, Borghan Narajabad, and Stephane Verani.
“Self-Fulfilling Runs: Evidence from the U.S. Life Insurance Industry,”
(February 2016), http://dx.doi.org/10.17016/FEDS.2015.032r1.
Foley-Fisher, Nathan, Borghan Narajabad, and Stephane Verani.
“Securities Lending as Wholesale Funding: Evidence from the U.S. Life
Insurance Industry,” National Bureau of Economic Research Working
Paper 22774 (2016).
Foley-Fisher, Nathan, Stefan Gissler, and Stephane Verani. “Over-theCounter Market Liquidity and Securities Lending,” (June 23, 2017),
https://ssrn.com/abstract=2869959 or http://dx.doi.org/10.2139/
ssrn.2869959.

Koijen, Ralph S.J., and Motohiro Yogo. “Shadow Insurance,” Econometrica,
84 (2016), pp. 1,265–1,287.
Koijen, Ralph, and Motohiro Yogo. “Risks of Life Insurers: Recent Trends
and Transmission Mechanisms,” in The Economics, Regulation, and
System Risk of Insurance Markets, Felix Hufeld, Ralph S. J. Koijen, and
Christian Thimann, eds., Oxford University Press (2017).
Lawsky, Benjamin M. “Shining a Light on Shadow Insurance: A Littleknown Loophole That Puts Insurance Policyholders and Taxpayers at
Greater Risk,” New York State Department of Financial Services,
Superintendent of Financial Services (2013).
Leitner, Yaron. “A Lifeline for the Weakest Link? Financial Contagion and
Network Design,” Federal Reserve Bank of Philadelphia Business Review
(Fourth Quarter 2002).
Merrill, Craig B., Taylor D. Nadauld, René M. Stulz, and Shane Sherlund.
“Did Capital Requirements and Fair Value Accounting Spark Fire Sales in
Distressed Mortgage-Backed Securities?” National Bureau of Economic
Research Working Paper 18270 (2012).
McDonald, Robert, and Anna Paulson. “AIG in Hindsight,” Journal of
Economic Perspectives, 29 (2015), pp. 81–106.
McMenamin, Robert, Zain Mohey-Deen, Anna Paulson, and Richard J.
Rosen. “How Liquid Are U.S. Life Insurance Liabilities? Chicago Fed Letter
302 (September 2012), https://www.chicagofed.org/publications/
chicago-fed-letter/2012/september-302.
Walsh, Mary Williams. “AIG Lists Banks It Paid with U.S. Bailout Funds,”
New York Times (March 15, 2009), http://www.nytimes.com/2009/03/16/
business/16rescue.html?mcubz=0. Counterparties document:
https://www.nytimes.com/interactive/projects/documents/aig-bailoutdisclosed-counterparties#p=1.

Harrington, Scott E. “Systemic Risk and Regulation: The Misguided Case of
Insurance SIFis,” (October 26, 2016), https://ssrn.com/abstract=2998646.
Kiyotaki, Nobuhiro, and John Moore. “Credit Cycles,” Journal of Political
Economy, 105:2 (1997), pp. 211–248.
Koijen, Ralph S.J., and Motohiro Yogo. “The Cost of Financial Frictions for
Life Insurers,” American Economic Review, 105.1 (2015), pp. 445–475.

Nontraditional Insurance and Risks to Financial Stability
2018 Q1

Federal Reserve Bank of Philadelphia
Research Department

25

Research Update
These papers by Philadelphia Fed economists,
analysts, and visiting scholars represent
preliminary research that is being circulated
for discussion purposes.

The views expressed in these papers are
solely those of the authors and should not
be interpreted as reflecting the views of
the Federal Reserve Bank of Philadelphia
or Federal Reserve System.

Strategic Default Among Private Student Loan
Debtors: Evidence from Bankruptcy Reform

Stuck in Subprime? Examining the Barriers
to Refinancing Mortgage Debt

Bankruptcy reform in 2005 restricted debtors’ ability to discharge
private student loan debt. The reform was motivated by the perceived
incentive of some borrowers to file bankruptcy under Chapter 7 even
if they had, or expected to have, sufficient income to service their debt.
Using a national sample of credit bureau files, the authors examine
whether private student loan borrowers distinctly adjusted their
Chapter 7 bankruptcy filing behavior in response to the reform. The
authors do not find evidence to indicate that the moral hazard
associated with dischargeability appreciably affected the behavior
of private student loan debtors prior to the policy.

Despite falling interest rates and major federal policy intervention,
many borrowers who could financially gain from refinancing have not
done so. The authors investigate the rates at which, relative to prime
borrowers, subprime borrowers seek and take out refinance loans,
conditional on not experiencing mortgage default. They find that
starting in 2009, subprime borrowers are about half as likely as prime
borrowers to refinance, although they still shop for mortgage credit,
indicating their interest in refinancing. The disparity in refinancing is
driven in part by the tightened credit environment post-financial
crisis, along with the fact that many subprime borrowers are ineligible
for the Home Affordable Refinance Program (HARP), which is the
major policy initiative designed to assist borrowers in refinancing
their mortgages. The authors argue that these barriers to refinancing
for subprime borrowers have long-term implications for social
stratification and wealth building. These concerns are exacerbated
by an additional finding of our work that refinance rates have been
significantly lower for black and Hispanic borrowers, even after
controlling for borrower credit status.

Supersedes Working Paper 15–17/R.
Working Paper 17–38. Rajeev Darolia, University of Kentucky and
Federal Reserve Bank of Philadelphia Visiting Scholar; Dubravka
Ritter, Federal Reserve Bank of Philadelphia Payment Cards Center.

The Aggregate Effects of Labor Market Frictions
Labor market frictions are able to induce sluggish aggregate employment dynamics. However, these frictions have strong implications
for the source of this propagation: They distort the path of aggregate
employment by impeding the flow of labor across firms. For a canonical class of frictions, the authors show how observable measures of
such flows can be used to assess the effect of frictions on aggregate
employment dynamics. Application of this approach to establishment
microdata for the United States reveals that the empirical flow of
labor across firms deviates markedly from the predictions of canonical
labor market frictions. Despite their ability to induce persistence in
aggregate employment, firm-size flows in these models are predicted
to respond aggressively to aggregate shocks but react sluggishly
in the data. This paper therefore concludes that the propagation
mechanism embodied in standard models of labor market frictions
fails to account for the sources of observed employment dynamics.
Working Paper 17–40. Michael W.L. Elsby, University of Edinburgh;
Ryan Michaels, Federal Reserve Bank of Philadelphia Research
Department; David Ratner, Federal Reserve Board of Governors.

26

Federal Reserve Bank of Philadelphia
Research Department

Working Paper 17–39. Lauren Lambie-Hanson, Federal Reserve Bank
of Philadelphia Supervision, Regulation and Credit; Carolina Reid,
University of California-Berkeley.

Model Secrecy and Stress Tests
Conventional wisdom holds that the models used to stress test banks
should be kept secret to prevent gaming. The authors show instead
that secrecy can be suboptimal, because although it deters gaming, it
may also deter socially desirable investment. When the regulator can
choose the minimum standard for passing the test, the authors show
that secrecy is suboptimal if the regulator is sufficiently uncertain
regarding bank characteristics. When failing the bank is socially costly,
then under some conditions, secrecy is suboptimal when the bank’s
private cost of failure is either sufficiently high or sufficiently low.
Working Paper 17–41. Yaron Leitner, Federal Reserve Bank of Philadelphia Research Department; Basil Williams, New York University.

Research Update
2018 Q1

Long-Run Trade Elasticity
and the Trade-Comovement Puzzle

The Paper Trail of Knowledge Spillovers:
Evidence from Patent Interferences

The authors show that the trade-comovement puzzle—
theory’s failure to account for the positive relation
between trade and business cycle synchronization—is
intimately related to its counterfactual implication that
short- and long-run trade elasticities are equal. Based
on this insight, the authors show that modeling the
disconnect between the low short- and the high longrun trade elasticity in consistency with the data is
promising in resolving the puzzle. In a broader context,
the authors’ findings are relevant for analyzing business cycle transmission in a large class of models and
caution against the use of static elasticity models in
cross-country studies.

The authors show evidence of localized knowledge spillovers using a new database of multiple invention from U.S. patent interferences terminated between
1998 and 2014. Patent interferences resulted when two or more independent
parties simultaneously submitted identical claims of invention to the U.S. Patent
Office. Following the idea that inventors of identical inventions share common
knowledge inputs, interferences provide a new method for measuring spillovers of
tacit knowledge compared with existing (and noisy) measures such as citation
links. Using matched pairs of inventors to control for other factors contributing
to the geography of invention and distance-based methods, the authors find that
interfering inventor pairs are 1.4 to 4 times more likely to live in the same city
or region. These results are not driven exclusively by observed social ties among
interfering inventor pairs. Interfering inventors are also more geographically
concentrated than inventors who cite the same prior patent. Our results emphasize
geographic distance as a barrier to tacit knowledge flows.

Working Paper 17–42. Lukasz A. Drozd, Federal Reserve
Bank of Philadelphia Research Department; Sergey
Kolbin, Amazon; Jaromir B. Nosal, Boston College.

Incumbency Disadvantage in U.S. National
Politics: The Role of Policy Inertia and
Prospective Voting
The authors document that postwar U.S. national
elections show a strong pattern of “incumbency disadvantage”: If the presidency has been held by a party
for some time, that party tends to lose seats in Congress.
The authors develop a model of partisan politics with
policy inertia and prospective voting to explain this finding. Positive and normative implications of the model
are explored.
Supersedes Working Paper 16–36.
Working Paper 17–43. Satyajit Chatterjee, Federal
Reserve Bank of Philadelphia Research Department;
Burcu Eyigungor, Federal Reserve Bank of Philadelphia
Research Department.

Working Paper 17–44. Ina Ganguli, University of Massachusetts–Amherst; Jeffrey
Lin, Federal Reserve Bank of Philadelphia Research Department; Nicholas Reynolds,
Brown University.

Greed as a Source of Polarization
The political process in the United States appears to be highly polarized: evidence
from voting patterns finds that the political positions of legislators have diverged
substantially, while the largest campaign contributions come from the most
extreme lobby groups and are directed to the most extreme candidates. Is the rise
in campaign contributions the cause of the growing polarity of political views? In
this paper, we show that, in standard models of lobbying and electoral competition,
a free-rider problem amongst potential contributors leads naturally to a divergence
in campaign contributors without any divergence in candidates’ policy positions.
However, we go on to show that a modest departure from standard assumptions—
allowing candidates to directly value campaign contributions (because of “ego rents”
or because lax auditing allows them to misappropriate some of these funds)—
delivers the ability of campaign contributions to cause policy divergence.
Working Paper 18–01. Igor Livshits, Federal Reserve Bank of Philadelphia Research
Department; Mark L.J. Wright, Federal Reserve Bank of Minneapolis, CAMA, NBER.

Research Update

2018 Q1

Federal Reserve Bank of Philadelphia
Research Department

27

Redefault Risk in the Aftermath of the Mortgage
Crisis: Why Did Modifications Improve More Than
Self-Cures?
This paper examines changes in the redefault rate of mortgages that
were selected for modification during 2008–2011, compared with
that of similarly situated self-cured mortgages during the same period.
We find that while the performance of both modified and self-cured
loans improved dramatically over this period, the decline in the redefault rate for modified loans was substantially larger, and the authors
attribute this difference to a few key factors. First, the modification
terms regarding repayments have become increasingly more generous,
including more principal reduction, resulting in greater financial relief
to the borrowers. Second, modifications in later vintages also benefited
from improving economic conditions. Modifications became more
effective as unemployment rates declined and home prices recovered.
Third, the authors find that the difference between redefault rate
improvement between modified loans and self-cured loans continue
to persist even after controlling for all the relevant risk and economic
factors. They attribute this difference to the servicers’ learning
process—such as data collection and information sharing among
industry participants—known as “learning-by-doing.” Early in the
mortgage crisis, many servicers had limited experience selecting the
best borrowers for modification. As modification activity increased,
lenders became more adept at screening borrowers for modification
eligibility and in selecting appropriate modification terms. The authors’
empirical findings suggest that mortgage modification effectiveness
could be enhanced through the industry’s “learning-by-doing” process.
Working Paper 18–02. Paul Calem, Federal Reserve Bank of
Philadelphia Supervision, Regulation and Credit; Julapa Jagtiani,
Federal Reserve Bank of Philadelphia Supervision, Regulation
and Credit; Raman Maingi, Federal Reserve Bank of Philadelphia
Supervision, Regulation and Credit; David Abell, Federal Reserve
Bank of Philadelphia Supervision, Regulation and Credit.

We analyze whether the passage of the Affordable Care Act’s dependent coverage mandate in 2010 reduced financial distress for young
adults. Using nationally representative, anonymized consumer
credit report information, we find that young adults covered by the
mandate lowered their past due debt, had fewer delinquencies, and
had a reduced probability of filing for bankruptcy. These effects
are stronger in geographic areas that experienced higher uninsured
rates for young adults prior to the mandate’s implementation. Our
estimates also show that some improvements are transitory because
they diminish after an individual ages out of the mandate at age 26.
Working Paper 18–03. Nathan Blascak, Federal Reserve Bank of
Philadelphia Payment Cards Center; Vyacheslav Mikhed, Federal
Reserve Bank of Philadelphia Payment Cards Center

Federal Reserve Bank of Philadelphia
Research Department

Creditors often outsource the task of obtaining repayment from
defaulting borrowers to third-party debt collectors. We argue that by
hiring third-party debt collectors, creditors can avoid competing in
terms of their debt collection practices. This explanation fits several
empirical facts about third-party debt collection and is consistent
with the evidence that third-party debt collectors use harsher debt
collection practices than original creditors. Our model shows that the
impact of third-party debt collectors on consumer welfare depends
on the riskiness of the pool of borrowers and provides insights into
which policy interventions may improve the functioning of the debt
collection market.
Supersedes Working Paper 15–43.
Working Paper 18–04. Viktar Fedaseyeu, Bocconi University; Robert
Hunt, Federal Reserve Bank of Philadelphia Consumer Finance Institute
and Payment Cards Center.

Screening on Loan Terms: Evidence from Maturity
Choice in Consumer Credit

Did the ACA’s Dependent Coverage Mandate
Reduce Financial Distress for Young Adults?

28

The Economics of Debt Collection: Enforcement
of Consumer Credit Contracts

We exploit a natural experiment in the largest online consumer lending
platform to provide the first evidence that loan terms, in particular
maturity choice, can be used to screen borrowers based on their
private information. We compare two groups of observationally equivalent borrowers who took identical unsecured 36-month loans; for
only one of the groups, a 60-month loan was also available. When
a long-maturity option is available, fewer borrowers take the shortterm loan, and those who do default less. Additional findings suggest
borrowers self-select on private information about their future ability
to repay.
Working Paper 18–05. Andrew Hertzberg, Federal Reserve Bank
of Philadelphia Research Department; Andres Liberman, New York
University; Daniel Paravisini, London School of Economics.

Research Update
2018 Q1

Forthcoming

Are Higher Capital
Requirements Worth It?
Bitcoin vs. the Buck:
Is Currency Competition
a Good Thing?

Connect with Us
Read previous articles
Economic Insights
www.philadelphiafed.org/research-and-data/publications/economic-insights

Business Review
www.philadelphiafed.org/research-and-data/publications/business-review

Sign up for e-mail notifications
www.philadelphiafed.org/notifications

PRESORTED STANDARD
U.S. POSTAGE
PAID

Federal Reserve Bank of Philadelphia

PHILADELPHIA, PA
PERMIT #583

Ten Independence Mall
Philadelphia, Pennsylvania 19106-1574

ADDRESS SERVICE REQUESTED

www.philadelphiafed.org
You can find Economic
Insights via the
Research Publications
part of our website.