The full text on this page is automatically extracted from the file linked above and may contain errors and inconsistencies.
Federal Reserve Bank of Chicago
The Effects of the Saving and Banking
Glut on the U.S. Economy
Alejandro Justiniano, Giorgio E. Primiceri, and
Andrea Tambalotti
November 2013
WP 2013-17
THE EFFECTS OF THE SAVING AND BANKING GLUT
ON THE U.S. ECONOMY
ALEJANDRO JUSTINIANO, GIORGIO E. PRIMICERI, AND ANDREA TAMBALOTTI
Abstract. We use a quantitative equilibrium model with houses, collateralized debt and
foreign borrowing to study the impact of global imbalances on the U.S. economy in the
2000s. Our results suggest that the dynamics of foreign capital flows account for between
one fourth and one third of the increase in U.S. house prices and household debt that
preceded the financial crisis. The key to these findings is that the model generates the
sustained low level of interest rates observed over that period.
1. Introduction
Before the financial crisis of 2007-08, most observers saw the growing international imbalances in the trade of goods and assets as the main source of vulnerability for the U.S.
and the world economy (e.g. Roubini and Setser, 2005, Obstfeld and Rogoff, 2007, Krugman, 2007). As shown in figure 1.1, the U.S. trade deficit had been growing since the
mid 1990s, reaching 6 percent of GDP at its peak in 2006, mostly financed by emerging
Asia (especially China) and the oil-exporting countries. In a very influential speech, then
Fed Governor Ben Bernanke (2005) attributed these imbalances to a “Global Saving Glut”
(SG), which he described as an excess of saving in developing countries primarily directed
towards riskless assets in the United States. Given these lopsided patterns of international
exchange, many feared that a sudden loss of appetite for U.S. assets by international investors might precipitate an abrupt adjustment of its current account deficit, with serious
repercussions for the world economy.
Although a catastrophic financial crisis did eventually occur, its epicenter was the American housing market, rather than the international market for assets and goods. As a consequence, most of the early literature on the causes and consequences of the crisis focused on
Date: November 29, 2013. First Draft: June 2013.
We would like to thank our discussants Klaus Adam, Federico Signoretti, and Martin Uribe for their useful
suggestions, and Efrem Castelnuovo, Gianluca Benigno, Andrea Ferrero, Andrea Raffo and participants in
several conferences and seminars for comments. Aaron Kirkman and Todd Messer provided outstanding
research assistance. The views expressed in this paper are those of the authors and are not necessarily
reflective of views at the Federal Reserve Banks of Chicago, New York or the Federal Reserve System.
1
THE EFFECTS OF THE SAVING AND BANKING GLUT ON THE U.S. ECONOMY
1992
1994
1996
1998
2000
2002
2004
2006
2008
2010
2012
0
180
‐1
160
‐2
140
‐3
120
‐4
100
‐5
80
‐6
‐7
House Prices
Trade Balance‐to‐GDP Ratio
1990
2
60
Trade Balance‐to‐GDP Ratio
FHFA (2000:Q1=100)
CoreLogic (2000:Q1=100)
40
Figure 1.1. U.S. trade balance-to-GDP ratio (left axis) and real house
prices (right axis). The two measures of house prices are the FHFA (formerly
OFHEO) all transactions house price index and the CoreLogic repeated-sales
index. Source: Haver.
the features of U.S. financial markets—regulation, risk-management, securitization, funding models—that contributed to the credit and housing boom of the first half of the 2000s,
whose reversal was the proximate cause of the crisis.1
A more recent strand of literature, however, has brought back into focus the connection
between the global imbalances that preceded the crisis and the credit and housing boom
that precipitated it.2 On the empirical side, this literature points to the depressing effect
of capital inflows on U.S. interest rates and spreads, whose low levels contributed to the
boom in mortgage debt and house prices before the crisis (Bernanke et al., 2011, Bertaut
et al., 2011, Warnock and Warnock, 2009). Bertaut et al. (2011), for instance, estimate
that the purchases of Treasury and Agency debt by the SG countries over the period 2003
to 2007, amounting to roughly one trillion dollars, lowered long-term interest rates in the
U.S. by between 110 and 140 basis points.
1
See Brunnermeir (2009) for an early overview. Demyanyk and van Hemert (2011), Ashcraft and Schuermann (2008), Adrian and Shin (2010), Pozsar et al. (2010), Gorton (2009), and Mian and Sufi (2009),
among many others, provide more detailed accounts of the events of the crisis and of some of the important
mechanisms at play, with a special focus on subprime mortgages and their securitization. In fact, in this
early phase, this sequence of events was most often described as the “subprime crisis.”
2
Obstfeld and Rogoff (2009) is one of the first papers forcefully arguing that global imbalances and the
financial crisis are intimately related.
THE EFFECTS OF THE SAVING AND BANKING GLUT ON THE U.S. ECONOMY
3
Moreover, careful detective work by Bertaut et al. (2011) in tracking the flows of capital
to and from the United States has uncovered another channel through which international
capital flows might have contributed to easier financial conditions in the U.S. before the
crisis. Indeed these authors document that after 2003, European banks played an increasing
role in the market for safe U.S. assets, especially the AAA tranches of private label assetbacked securities (ABS) that turned out to be far from riskless in the crisis. Since Europe’s
current account vis-a-vis the U.S. was roughly balanced, these gross positions of European
banks in ABS were funded by direct borrowing in the dollar wholesale credit market. As
a result, these financial institutions became an integral part of the financial intermediation
sector in the U.S., in direct competition with domestic financial institutions, as also pointed
out by Acharya and Schnabl (2009).
Shin (2012) refers to these gross flows from international banks into mortgage products,
even in the absence of corresponding net imbalances, as the “Global Banking Glut” (BG),
in juxtaposition to the Global Saving Glut associated with the U.S. current account deficit.
According to his analysis, the flow of funds associated with the BG played an important
role in easing financial conditions in the United States during the boom, comparable in
magnitude to that of the purchases of Government debt by the SG countries.
In Shin’s (2012) model of global banking, spreads are negatively related to the total
amount of funds intermediated by the financial system . When risk recedes, banks expand
their balance sheet and spreads fall. This motivates Shin’s claim that higher total intermediation generated by European banks lowered the spreads between safe funding rates
and the returns on ABS, and therefore ultimately on mortgages. Consistent with this view,
Bertaut at al. (2011) estimate that the increase in the demand for ABS and similar instruments by European banks over the period 2003-2007 contributed to a decline in their yield
of between 60 and 160 basis points, depending on the instruments and the methodology.3
When the boom turned to bust, and the market for private-label ABS in which European
banks were most active disappeared, the mechanism worked in reverse, contributing to the
propagation of the U.S. financial crises around the world (Obstfeld, 2012, Acharya and
Schnabl, 2009).
3
These estimates represent an upper bound on the effect of the BG on U.S. rates, since they do not account
for the endogenous response in the supply of ABS and similar assets. The production of these assets rose
dramatically during this period, in part to satisfy the increase in demand by U.S. and foreign investors.
THE EFFECTS OF THE SAVING AND BANKING GLUT ON THE U.S. ECONOMY
4
Despite the growing empirical evidence on the effects of net and gross capital flows on U.S.
credit markets and interest rates, only a handful of papers have addressed quantitatively
the impact of the global saving and banking gluts on the U.S. macroeconomy in general,
and on the credit and house-price boom of the 2000s more specifically. In this paper, we
tackle this question using a quantitative dynamic equilibrium model, which includes most
of the ingredients of typical medium-scale DSGEs inspired by Christiano, Eichenbaum, and
Evans (2005) and Smets and Wouters (2007). In addition, our model features borrowing
and lending among heterogeneous households, as well as from abroad, with houses serving
as collateral (Iacoviello, 2005). Therefore, in our framework house prices play a key role in
determining the amount of household debt. Thanks to these features, the model is a useful
laboratory to study the macroeconomic effects of the SG and BG on the U.S. economy,
including their contribution to the credit and real estate boom of the 2000s.
To analyze the impact of the SG, we take the observed U.S. trade deficit, and the associated capital inflows, as exogenous: a given amount of goods and services that domestic
households must consume today, received from the rest of the world in exchange for a
promise of goods and services in the future. This exogenous flow of resources tilts the
intertemporal consumption profile of domestic agents towards the present, which can be
optimal only if interest rates decline. The rest of the adjustment in the domestic economy
follows from this fall in the domestic rate of return. In brief, lower interest rates stimulate
the demand for nondurable consumption, investment and housing by the lenders, who are
on their Euler equation. The resulting upward pressure on house prices then relaxes the
collateral constraint of the borrowers, who can thus also consume more. Finally, preferences
parametrized to deliver a small wealth effect prevent this expansion in consumption from
implying a sharp fall in hours worked. This chain of events is what we refer to as the Global
Saving Glut channel. Quantitatively, the SG has fairly large effects on the macroeconomy.
Consumption and investment increase by roughly 5% and 10% above the balanced growth
path, while the effect on GDP is more muted because of the deterioration of the trade
balance. Furthermore, at the peak the ratio of household debt to GDP is 8% higher, while
house prices increase by 13%.
In addition to quantifying the SG’s impact, we attempt to capture Shin’s (2012) Global
Banking Glut. We model the BG as a reduction in the spread between the interest rate
paid by (mortgage) borrowers and the funding rates of the shadow banking system, which
THE EFFECTS OF THE SAVING AND BANKING GLUT ON THE U.S. ECONOMY
5
are tied in turn to the interest rate earned by savers. In our model, this spread between
borrowing and lending rates reflects the market power of financial intermediaries, which
channel funds to the impatient households from both the domestic patient ones and the
rest of the world. More competition in this market therefore implies lower spreads. This
is the channel through which the entry of European players in the intermediation market,
which is at the heart of the Global Banking Glut hypothesis, affects credit conditions in
the model.
This simple modeling approach captures the first order effect on macroeconomic outcomes
of the entry of European banks in the American intermediation market. A reduction in
spreads is also an outcome of the more sophisticated and realistic model of intermediation
proposed by Shin (2012) to formalize the effects of the Global Banking Glut. Of course,
that model has many more implications than ours regarding the behavior of intermediaries,
but it is silent on their macroeconomic consequences.
In our simulations, the reduction in spreads due to the BG reduces the interest rate and
makes houses more valuable for the borrowers, relative to nondurable goods. This shift in
demand towards houses increases their price and relaxes the collateral constraint, allowing
the borrowers to obtain more credit. In this case without capital inflows, the increase in
credit is financed by additional lending from the domestic savers. Hence, the response
of savers to the BG tends to be opposite to that of the borrowers, leading to fairly muted
effects on aggregate variables. This is in contrast with the SG case, where the fall in interest
rates stimulates demand by both types of households. Overall, our experiments attribute
to the combined effect of the BG and SG between one fourth and one third of the increase
in house prices and household debt during the years leading up to the financial crisis.
1.1. Related literature. Among the first papers to formally address the saving glut hypothesis were Caballero, Farhi, and Gourinchas (2008), and Mendoza, Quadrini, and RiosRull (2009). These influential contributions share a focus on the differences in financial
market development across countries that made riskless U.S. assets a particularly appealing store of value for the excess saving in the rest of the world. In their detailed quantitative
exercise, Mendoza, Quadrini, and Rios-Rull (2009) also trace the dynamic implications of
global financial liberalization on international imbalances and portfolios. However, they
concentrate much less then we do on the broader macroeconomic implications in the United
States.
THE EFFECTS OF THE SAVING AND BANKING GLUT ON THE U.S. ECONOMY
6
Our work is more closely related to a recent literature connecting current account deficits
and foreign capital flows into the U.S. to the housing and credit booms that preceded the
financial crisis. This literature is far from unanimous on the importance of this international
channel for macroeconomic developments in the United States. At one extreme of this
spectrum, Favilukis et al. (2012 and 2013) find virtually no effect of the foreign purchases
of Government bonds on house prices and domestic credit. In their rich model, the reduction
in the risk-free rate associated with the flow of international resources into the U.S. economy
is offset by an increase in the risk premium on housing, due to the portfolio reallocation
forced on domestic agents by the foreign purchases of safe assets.4 In our model, on the
contrary, there is no risk, and hence no meaningful portfolio choice, since agents have
perfect foresight. As a consequence, our results highlight the intertemporal substitution
mechanisms associated with the fall in interest rates triggered by the trade deficit. On
the one hand, this stricter macroeconomic focus comes at the cost of ignoring a potentially
important portfolio channel and might bias our quantitative conclusions. On the other hand,
the irrelevance of portfolio choice in our model allows us to take a more comprehensive view
of the saving glut, without having to worry about the composition of the capital inflows.
Moreover, it enables us to calibrate the foreign impulse to the total trade deficit, rather
than just to the fraction corresponding to the purchases of Treasuries, as Favilukis et al.
(2012) do.
Our results are more in line with those of Adam, Kuang, and Marcet (2011), Garriga,
Manuelli, and Peralta-Alva (2012), and Kiyotaki, Michaelides, and Nikolov (2011), who all
find large effects of a decline in the world interest rate on house prices in small open economy
models calibrated on the U.S. experience of the 2000s, despite important differences in their
modeling approaches. In Adam et al. (2011), the large reaction of house prices is due to
the presence of agents with subjective beliefs about these prices. In Garriga et al. (2012),
the quantitative results hinge on combining lower real interest rates and looser financial
conditions, while we abstract from the latter. Furthermore, unlike these papers, we model
the domestic country as large, and populated by both borrowers and savers, which makes
interest rates endogenous and creates interesting dynamics from the interplay of the two
classes of agents. In this respect, we are closer to Kiyotaki et al. (2011), who focus on
4Caballero and Krishnamurty (2009) also highlight the increase in risk borne by domestic agents following
the absorption of safe assets by foreigners.
THE EFFECTS OF THE SAVING AND BANKING GLUT ON THE U.S. ECONOMY
7
the redistributive effects of changes in the world interest rate across net buyers and net
sellers of houses. Finally, Ferrero (2012) also addresses the correlation between current
account deficits and house prices within a two-country model, but with one representative
international saver lending to one representative borrower in the U.S.
The analysis of the BG is a distinctive feature of our paper, relative to these other
contributions. In this respect, we also make contact with the small-open economy business
cycle literature that has emphasized the role of risk premia and spreads as an important
source of business cycle fluctuations (e.g. Neumeyer and Perri 2005, Garcıa-Cicco, Pancrazi,
and Uribe, 2010).
The rest of the paper proceeds as follows. Section 2 describes the model, whose calibration
is presented in section 3. Section 4 presents the model’s dynamics in response to three
impulses: an increase in the trade deficit that captures the saving glut; a fall in the spread
between borrowing and lending rates, which captures the banking glut; and a combination
of the two. Section 5 concludes.
2. Model
This section presents the quantitative general equilibrium model we use to analyze the
impact of the SG and BG on the U.S. economy, including their effect on the recent boom in
the debt of U.S. households and in real estate prices. The model builds on Iacoviello (2005)
and Campbell and Hercowitz (2009). The key assumption is that domestic households
have heterogeneous desires to save, which generate borrowing and lending among them, as
well as from abroad. Debt, which we often call “mortgages”, is collateralized by houses, to
capture the fact that mortgages represent by far the most important component of U.S.
households’ liabilities.
The economy is populated by four main classes of domestic agents—households, house
producers, goods producers and a government. Households can borrow from each other
through a financial intermediation sector composed of monopolistic investment banks and
competitive commercial banks, which also collect funds from abroad.
2.1. Households. There are two types of households, which differ by the rate at which
they discount the future. Patient households are denoted by l, since in equilibrium they
are the ones saving and lending. They represent a share 1 − ψ of the population, and we
THE EFFECTS OF THE SAVING AND BANKING GLUT ON THE U.S. ECONOMY
8
interchangeably refer to them as the lenders or savers. Their discount factor is βl > βb ,
where βb is the discount factor of the impatient borrowers. At time 0, representative
household j = b, l maximizes the utility function
1−σ
H 1− − ϕX L1+η
∞
C
−1
X
j,t j,t
j,t j,t
E0
βjt
,
1−σ
t=0
where
µ
1−µ
1−
Xj,t = Cj,t
Hj,t
Xj,t−1
.
In these expressions, Cj,t denotes consumption of nondurable goods, Hj,t is the stock of
houses, and Lj,t is hours worked. This specification of the utility function is similar to the
one proposed by Jaimovich and Rebelo (2009). When µ = 1, it belongs to the class of
utility functions recommended by King, Plosser, and Rebelo (1988). When µ = 0, these
preferences resemble those of Greenwood, Hercowitz, and Huffman (1988), with no wealth
effect on labor supply. The difference from Jaimovich and Rebelo (2009) is that households
in this model derive utility from a Cobb-Douglas aggregator of non-durable consumption
goods and housing services (assumed to be proportional to the stock of housing). The
parameter denotes the share of nondurable consumption in this composite good.
The utility maximization problem is subject to the nominal flow budget constraint
Pt Cj,t + Pth Ξj,t + Pt Ij,t + Rj,t−1 Dj,t−1 ≤ Wj,t Lj,t + Rtk Kj,t + Πj,t − Pt Tj,t + Dj,t .
In this expression, Pt and Pth are the prices of the consumption good and of houses, while
Rtk and Wj,t are the nominal rental rates of capital and labor. The wage is indexed by j
because the labor input of the borrowers is not a perfect substitute for that of the savers.
Dj,t is the amount of one period nominal debt (or deposits, when negative) accumulated
by the end of period t, and carried into period t + 1, with gross nominal interest rate Rj,t .
The interest rate depends on j because the return paid by borrowers and that earned by
lenders differ due to the presence of a spread in the financial intermediation sector. Πj,t
is the share of profits of the intermediate firms and of investment banks accruing to each
household of type j, and Tj,t are lump-sum taxes and transfers from the government.
The stocks of houses and capital evolve according to the accumulation equations
Hj,t+1 = (1 − δh ) Hj,t + Ξj,t
THE EFFECTS OF THE SAVING AND BANKING GLUT ON THE U.S. ECONOMY
(2.1)
Kj,t+1
9
Ij,t
Ij,t ,
= (1 − δk ) Kj,t + 1 − Sk
Ij,t−1
where Ξj,t represents new houses, Ij,t is investment in productive capital, and δh and δk
are the rates of depreciation of the two stocks. The function Sk captures the presence of
adjustment costs in investment, as in Christiano, Eichenbaum, and Evans (2005), and is
parametrized as follows
1
Sk (x) = ζk (x − eγ )2 ,
2
(2.2)
so that, in steady state, Sk = Sk0 = 0 and Sk00 = ζk > 0, where eγ is the economy’s growth
rate along the balanced growth path, further described below. Finally, we assume that
the total supply of new houses (Ξt ) grows at the constant rate of technological progress,
which corresponds to imposing infinite costs for adjusting the production of new houses. In
section 4.3, we also consider a more complex specification in which new houses are produced
combining land and residential investment.
2.1.1. The borrowing limit. Households’ ability to borrow is limited by a collateral constraint, similar to Kiyotaki and Moore (1997). We model this constraint to mimic the asymmetry of mortgage contracts, as in Justiniano, Primiceri, and Tambalotti (2012). When
real estate prices increase, households can refinance their loans and therefore borrow more
against the higher value of the entire housing stock. When prices fall, however, the lower
collateral value leads to less lending against new houses, but lenders cannot require faster
repayment of the debt already outstanding.
More formally, we write the collateral constraint as
Dj,t ≤ D̄j,t =
θPth Hj,t+1
h
if Pth ≥ Pt−1
(1 − δ ) D̄
h
h
h
j,t−1 + θPt Ξj,t if Pt < Pt−1 .
h
If collateral values increase (i.e. Pth rises), households can borrow up to a fraction θ of
the current value of their entire housing stock. This is the standard formulation of the
collateral constraint, which implicitly assumes that all outstanding mortgages will be refinanced to take advantage of the new, more favorable conditions. On the contrary, if Pth
falls, households need not repay the outstanding balance on their mortgage, over and above
THE EFFECTS OF THE SAVING AND BANKING GLUT ON THE U.S. ECONOMY
10
the repayment associated with the depreciation of the housing stock (δh ).5 Therefore, the
new less favorable credit conditions only apply to the flow of new mortgages, collateralized
by the most recent house purchases.
This asymmetry of the borrowing limit makes debt in the model very similar to an
adjustable-rate mortgage, in which the interest rate changes every quarter, but the principal
can only be adjusted by the borrower, either through a cash-out refinancing, or through
pre-payment. Therefore, this modeling device captures an important feature of long-term
mortgages over real estate cycles, namely the downward stickiness of the principal. During
booms, homeowners take advantage of rising property values through refinancing, which
reduces the effective duration of outstanding mortgages. During busts, on the contrary,
existing homeowners cannot be forced to pay down their outstanding debt (over and above
the contractual amortization), even if the value of the collateral falls.
Given their low desire to save, impatient households borrow from the patient in equilibrium. In fact, local to the steady state, they borrow up to the collateral constraint.
Therefore, they choose not to hold any capital: absent the constraint, they would borrow
even more, so it is clearly not optimal for them to hold any asset other than houses. For
simplicity, we impose that borrowers do not accumulate capital also when the collateral constraint does not bind away from the steady state, even if it might occasionally be optimal
for them to do so.
2.2. Goods producers. There is a continuum of intermediate firms indexed by i ∈ [0, 1],
each producing a good Yt (i), and a competitive final good sector producing output Yt
according to
ˆ
Yt =
1
Yt (i)
1
1+λ
1+λ
di
.
0
Intermediate firms, which are owned by the lenders, operate the constant-return-to scale
production function
(2.3)
h
i1−α
Yt (i) = A1−α
Ktα (i) (ψLb,t (i))ν ((1 − ψ) Ll,t (i))1−ν
− At F .
t
5This formulation assumes that the amortization rate of the mortgage coincides with the depreciation rate
of the housing stock. In Justiniano, Primiceri, and Tambalotti (2012) we also experiment with higher
amortization rates, so that households build equity in their house over time, as in Campbell and Hercowitz
(2009).
THE EFFECTS OF THE SAVING AND BANKING GLUT ON THE U.S. ECONOMY
11
They rent labor (of the two types) and capital on competitive markets paying Wb,t , Wl,t
and Rtk respectively. F represents a fixed cost of production, and is chosen to ensure that
steady state profits are zero. The labor augmenting technology factor At grows at rate γ.
The intermediate firms operate in monopolistically competitive markets and set their
price Pt (i) subject to a nominal friction as in Calvo (1983). A random set of firms of
measure 1 − ξp optimally reset their price every period, subject to the demand for their
product, while the remaining ξp fraction of prices do not change. In section 4.1, we describe
how the presence of sticky prices amplifies the expansionary effect of the SG on labor
demand and the equilibrium level of production.
2.3. Fiscal and monetary policy. The government collects taxes, pays transfers, consumes a fraction of final output, and sets the nominal interest rate.
We assume that government spending is a constant fraction g of final output, and that
the government balances it’s budget, i.e.
Gt = gYt = ψTb,t + (1 − ψ) Tl,t .
In addition, we assume that total taxes levied on borrowers represent a constant share χ of
the financing needs of the government, i.e.
ψTb,t = χGt .
If χ = 0, the entire tax burden is on the savers, while if χ = ψ borrowers and savers pay
the same amount per-capita. Therefore, we can interpret the parameter χ as capturing the
extent of government redistribution. Given the presence of borrowing-constrained agents
and the redistributive nature of the tax rule, Ricardian equivalence does not hold in our
model. Therefore, the balanced budget assumption is not entirely innocuous, but just a
simplification that allows us to abstract from the details of how the government finances
its expenditures, as is the assumption of lump-sum taxes.
Monetary policy sets the short-term nominal interest rate on deposits based on the
feedback rule
!
ρR "
τy #1−ρR
1/4 τπ
Rl,t
Rl,t−1
1
Yt
(πt · πt−1 · πt−2 · πt−3 )
= max
;
,
Rl
Rl
Rl
π
Yt∗
THE EFFECTS OF THE SAVING AND BANKING GLUT ON THE U.S. ECONOMY
12
where πt is the gross rate of inflation, π is the Central Bank’s inflation target, and Yt∗ is
the trend level of output that grows at the constant rate of technological progress. The
parameters ρR , τπ and τy capture the degree of inertia, and the strength of the interest rate
reaction to the deviations of annual inflation from the target and of output from trend.
2.4. Financial intermediaries. The financial intermediation sector transfers resources to
the borrowers from savers at home and in the rest of the world. The sector consists of investment banks, which do the lending, and retail banks, which collect deposits. Investment
banks fund the purchase of mortgages by issuing securities, say asset-backed commercial
paper, on a non-competitive wholesale credit market. Retail banks (or money market mutual funds) buy these securities using the deposits that they collect. In the process, the
investment banks earn a profit in the form of a spread between the mortgage rate and their
funding rate, which is equal to the retail deposit rate paid to savers.
We assume that this spread derives from market power exercised by investment banks
in the wholesale funding market, for instance by creating differentiated credit products
that they sell to commercial banks and money market mutual funds. The existence of
this market power can be microfounded through a monopolistic competition structure as
in Gerali at al.’s (2010) DSGE model of banking. Alternatively, we could assume that the
spread earned by investment banks derives from the costs of intermediation that each of
them sustains, as in Curdia and Woodford (2009) for instance.
In either case, what is relevant for our analysis is only that the spread falls when European banks enter the wholesale credit market, as suggested by Shin’s (2012) global banking
glut hypothesis. Under monopolistic competition, this compression in the markup follows
naturally when new competitors enter the market. With intermediation costs, instead,
spreads fall if entry reduces each bank’s market footprint, and hence its volume of intermediation, under the conventional assumption that the cost is convex in the amount of funds
intermediated.6
6For our purposes, this intermediation structure is equivalent to one with a third layer of intermediation,
in which individual mortgages are securitized and sold to investment banks and other participants in the
shadow banking system, rather than purchased directly by them. This more complex structure would
be equivalent to the one presented here because the margin that is relevant for the global banking glut
hypothesis that we are trying to capture is the one between wholesale funding and mortgage products,
rather than the one between mortgages and mortgage-backed securities.
THE EFFECTS OF THE SAVING AND BANKING GLUT ON THE U.S. ECONOMY
13
2.4.1. Investment banks. Investment banks purchase mortgages with funds they raise in the
wholesale credit market, by issuing a continuum of differentiated credit instruments indexed
by n ∈ [0, Nt ] . Each of these instruments carries a net interest rate rt (n), while mortgages
accrue net interest of rb,t . Therefore, the profit earned from funding Db,t (n) dollars worth
of mortgages by selling Lt (n) dollars of instrument n is
Πt (n) = rb,t Db,t (n) − rt (n) Lt (n) ,
where one dollar in wholesale funding turns into one dollar in mortgages, so that Db,t (n) =
Lt (n) .
2.4.2. Retail banks. Retail banks collect a total amount of deposits
Dt = − (1 − ψ) Dl,t + Df,t ,
where − (1 − ψ) Dl,t come from the lenders and Df,t come from the RoW.7 The net interest
rate paid on these deposits is rl,t , and they are invested in the securities issued by investment
banks, generating total profits for the retail banking sector in the amount
ˆ Nt
rt (n) Lt (n) dn − rl,t Dt .
0
We assume that retail banks have a taste for a diversified portfolio, which implies a
downward sloping demand function for each instrument issued by the investment banks,
with elasticity ε (Nt ). For instance, under the translog specification proposed by Feenstra
(2003), this elasticity would be ε (Nt ) = − (1 + εNt ) (e.g. Bilbiie, Ghironi, and Melitz,
2012). As the number of instruments available on the market increases, they become closer
substitutes, and the elasticity of demand increases.
2.4.3. Equilibrium, spread and market clearing. The taste for portfolio diversification of
retail banks generates monopoly power in the wholesale credit market, which investment
banks exploit by selling their securities at a discount with respect to the interest rate
charged to borrowers. As a result, in the symmetric equilibrium in which all securities pay
the same interest rate, we have
(2.4)
rb,t = µ (Nt ) rl,t ,
7Recall that D denotes debt if positive and deposits if negative. Similarly, a positive D corresponds to
l,t
f,t
foreign holdings of U.S. debt.
THE EFFECTS OF THE SAVING AND BANKING GLUT ON THE U.S. ECONOMY
14
where the spread in the borrowing over the lending rate µ (Nt ) depends on the size of
the market Nt . Under the translog specification, this function is µt (Nt ) = 1 + (εNt )−1 .
Alternatively, if the portfolio purchased by retail banks were a CES aggregate of a discrete
set of Nt instruments, we would obtain µt (Nt ) = [ε (Nt − 1) + Nt ] / [ε (Nt − 1)], as in
Jaimovich and Floetotto (2008). Intuitively, both of these parametrizations imply that
markups fall as the number of competitors increases. However, in our calibration we do
not exploit any specific mapping between Nt and µt , and instead we calibrate directly the
decline in spreads due to the entry of European players.
In the symmetric equilibrium, in which one dollar of deposits turns into one dollar of
wholesale lending and one dollar of loans, with no deadweight loss from interest rate dispersion, market clearing implies
ˆ
Nt
− (1 − ψ) Dl,t + Df,t =
Lt (n) dn
0
ˆ
Nt
ψDb,t =
Lt (n) dn,
0
so that
(2.5)
ψDb,t + (1 − ψ) Dl,t = Df,t .
Moreover, the aggregate profits of the intermediation sector, which are rebated to households in period t + 1, amount to
rl,t (µ (Nt ) − 1) Dt .
2.5. The resource constraint and the current account. The economy’s resource constraint is
(2.6)
Yt + (∆Df,t − rl,t−1 Df,t−1 ) = ψCb,t + (1 − ψ) Cl,t + Ξt Pth /Pt + ψIl,t + Gt ,
where Il,t is the amount of per-capita investment undertaken by the lenders, who are the
only households accumulating capital. This expression derives from aggregating the budget
constraints of borrowers and lenders with that of the government, using the zero profit
conditions of the competitive firms and of retail banks, the definition of profits for the
intermediate firms and the investment banks, and the debt market clearing condition (2.5).
THE EFFECTS OF THE SAVING AND BANKING GLUT ON THE U.S. ECONOMY
15
Equation (2.6) clarifies the role of foreign debt in our economy. The RoW holds a stock
of domestic deposits equal to Df,t , on which it earns a return rl,t . Deposits are the only
asset traded internationally, so that Df,t also represents the net foreign debt position of
the U.S. The change in this position, ∆Df,t , corresponds to the current account deficit.
This accumulation of deposits, net of the interest paid on the outstanding debt, ∆Df,t −
rl,t−1 Df,t−1 , is the counterpart to the trade deficit in the financial account. Therefore, it
represents an addition to the resources produced domestically, Yt , which is available for
consumption (and investment) at home.8
From the perspective of the U.S. economy, the willingness of the RoW to invest its excess
savings in U.S. assets results in a flow of real resources today, to be returned with interest in
the future. This flow of resources is the exogenous force that tilts the consumption profile
of domestic agents towards the present, ”forcing” them to consume more today than what
their domestic production would allow.
As for the RoW, we do not model the source of their desire to lend to the U.S., but simply
take as given their excess of saving with respect to investment. However, this does not imply
that their desired current account surplus is insensitive to interest rates. The assumption
is instead that there exists a sequence of shocks to desired saving and/or investment in
the RoW, which generates the observed current account deficit at the prevailing interest
rate. This view that U.S. current account deficits are driven exclusively by developments
abroad is perhaps extreme, but it captures well the spirit of Bernanke’s (2005) saving glut
hypothesis. As such, however, it is most compelling for the period of growing international
imbalances leading up to the financial crisis, but less so as a description of the retrenchment
in the U.S. trade deficit that accompanied the Great Recession. This is why our quantitative
experiments stop in 2006Q3, when the current account deficit reached its peak.
8In the data, changes in the net foreign asset position can differ from the current account due to valuation
effects, which have been sizable and favorable to the U.S. in the last twenty years (e.g. Gourinchas and Rey,
2013). Capturing these effects would require a much more sophisticated model of international financial
flows than the one adopted here, which should include a distinction between net and gross asset flows, and
a rich menu of assets with different rates of return and realistic movements in their prices. Developing such
a model is well beyond the scope of this paper.
THE EFFECTS OF THE SAVING AND BANKING GLUT ON THE U.S. ECONOMY
16
Households
βl
0.998
βs
0.99
ϕ
1
η
1
σ
1
µ
0.0005
ψ
0.61
0.899
Production
γ
0.005
ν
0.525
α
0.3
λ
0.2
ξp
0.65
δh
0.003
δk
0.025
ζk
2
Other
Param.
π
1.005
ρR
0.8
τπ
2.5
τy
0.125
χ
0.45
g
0.175
Df,0 / (4Y0 )
0.15
4(Rb − Rl )
0.015
Table 1. Parameter values.
3. Calibration
We parametrize the model so that its steady state matches some key statistics for the
period of relative stability of the 1990s. Choosing a later period would be problematic, because the large swings in debt, house prices and the trade balance observed in the 2000s are
likely to represent large deviations from such a steady state. The calibration is summarized
in table 1 and is based on U.S. aggregate and micro data.
Time is in quarters. We set the Central Bank’s inflation target (π) equal to the average
gross rate of inflation (1.005, or 2% per year), and the growth rate of productivity in
steady state (γ) to match average GDP growth (0.5%) during the 1990s. In steady state,
Rl =
eγ π
βl .
Therefore, we choose a value of 0.998 for the lenders’ discount factor (βl ), to
obtain an annualized steady state nominal interest rate of 4.8%, close to the average Federal
Funds Rate. For the borrowers’ discount factor (βb ) we pick a value of 0.99, so that the
relative impatience of the two groups is similar to that in Campbell and Hercowitz (2009)
and Krusell and Smith (1998). The labor disutility parameter (ϕ) only affects the scale
of the economy, so we normalize it to 1. We also pick a Frisch elasticity of labor supply
(1/η) equal to 1. This value is a compromise between linear utility, which is typical in
the Real Business Cycle literature (Hansen, 1985), and the low elasticities of labor supply
usually estimated by labor economists and more common in the empirical DSGE literature.
We choose the standard value of 1 for the risk aversion parameter (σ) and 0.0005 for the
parameter µ, so as to minimize the strength of the wealth effect on labor supply, while still
being consistent with balanced growth.
We parametrize the degree of heterogeneity between borrowers and lenders using the
Survey of Consumer Finances (SCF), which is a triennial cross-sectional survey of the
assets and liabilities of U.S. households. We identify the borrowers as the households
θ
0.85
THE EFFECTS OF THE SAVING AND BANKING GLUT ON THE U.S. ECONOMY
17
that appear to be liquidity constrained, namely those with liquid assets worth less than
two months of their total income. Following Kaplan and Violante (2012), we define liquid
assets as the sum of money market, checking, savings and call accounts, directly held mutual
funds, stocks, bonds, and T-Bills, net of credit card debt. We apply this procedure to the
1992, 1995 and 1998 SCF, and obtain an average share of borrowers equal to 61%, which
directly pins down the parameter ψ. Given this split between borrowers and savers, we set
the production parameter ν equal to 0.525 to roughly match their relative labor income
(0.64) in the SCF. In addition, we choose the parameter controlling the progressivity of the
tax/transfer system to match the ratio of hours worked by borrowers and lenders (1.08).
This requires setting χ = 0.45, which implies a moderate level of overall redistribution.
The share of nondurable consumption in the Cobb-Douglas aggregator of the utility
function (), the depreciation of houses (δh ) and the initial loan-to-value ratio (θ) are chosen
jointly to match three targets. The first target is the real estate-to-GDP ratio, which we
estimate from Flow of Funds (FF) and NIPA data as the average ratio between the market
value of real estate owned by households and nonprofit organizations and GDP (1.2). The
second target is the debt-to-real estate ratio, for which we use FF data on the average ratio
between home mortgages and the market value of real estate (0.36). The third target is
the ratio of residential investment to GDP (4%). Hitting these targets requires choosing
δh = 0.003, which is consistent with the low end of the interval for the depreciation of houses
in the Fixed Asset Tables, and θ = 0.85, which is in line with the cumulative loan-to-value
ratio of first time home buyers estimated by Duca et al. (2011) for the 1990s.
On the production side, we follow standard practice and pick an elasticity of the production function (α) of 0.3, and a depreciation of productive capital (δk ) of 0.025. The average
net markup of intermediate firms (λ) is 20%, which is in the middle of the range of values
used in the literature and the Calvo parameter (ξp ) is 0.65, consistent with the evidence in
Nakamura and Steinsson (2008). For the second derivative of the investment adjustment
cost function (ζk ) we pick a value of 2, in line with the estimates of Eberly, Rebelo, and
Vincent (2012).
We fix the steady state ratio of G to Y equal to 0.175. For the monetary policy rule we
choose a considerable amount of interest rate inertia (ρR = 0.8), a moderate reaction to
the output gap (τy = 0.125), and a relatively strong reaction to inflation (τπ = 2.5), in line
with the typical empirical estimates of the Taylor rule in the post-1984 period.
THE EFFECTS OF THE SAVING AND BANKING GLUT ON THE U.S. ECONOMY
18
We set the initial stock of foreign lending relative to annual GDP (D0,f / (4Y0 )) equal to
15%, based on Gourinchas and Rey’s (2013) estimate of the U.S. net foreign asset position
in 1998.9 Finally, we choose the annualized steady state interest rate spread between
borrowing and lending rates (4 (Rb − Rl )) by looking at the average spreads over the 1990s
between the 30-year conventional mortgage rate and the 5-, 7- and 10-year constant maturity
Treasury rates, which were 1.79%, 1.57% and 1.45% respectively. We choose 1.5% as a round
compromise among these three numbers, since the fluctuations over time in the effective
duration of conventional mortgages make it difficult to choose Treasuries of a particular
maturity as the most appropriate benchmark.
As described in detail in the next section, we study the impact of the SG and BG
by analyzing the response of the model’s endogenous variables to changes in the forcing
process Df,t / (4Yt ) and 4 (Rb,t − Rl,t ). We compute these responses by solving the system of
nonlinear difference equations given by the first order conditions of the agents’ optimization
problems and the market clearing conditions. The algorithm used to solve this nonlinear
forward-looking model is based on Julliard, Laxton, McAdam, and Pioro (1998), but has
been modified to account for the asymmetry of the borrowing constraint, and for the fact
that this constraint always binds in steady state, but it is occasionally slack during the
transitions.
4. Results
This section presents estimates of the effects of the BG and SG on the U.S. economy,
based on the quantitative model described in sections 2 and 3. Although its open economy
dimension is extremely stylized, this model is a useful laboratory to analyze how these international forces affected the U.S. economy, including the extent to which they contributed
to the boom in house prices and household debt leading up to the financial crisis. In the
rest of the section, we start by describing the impact of the SG and BG in isolation, and
then consider their interaction.
9This ratio increases over the course of the model’s simulations, due to the accumulation of current account
deficits, as shown in figure 4.3.
THE EFFECTS OF THE SAVING AND BANKING GLUT ON THE U.S. ECONOMY
5.00
19
10‐Year Nominal Treasury minus 10‐Year Expected Inflation (SPF)
10‐Year TIPS
4.00
3.00
2.00
1.00
2011
2009
2007
2005
2003
2001
1999
0.00
‐1.00
Figure 4.1. 10-year Treasury constant maturity rate minus expected inflation (measured as the median CPI-inflation forecast over the next 10 years
from the Survey of Professional Forecasters); and continuously compounded
10-year zero-coupon yield on Treasury Inflation-Protected Securities (TIPS).
Source: Haver.
4.1. Saving glut. Bernanke (2005) coined the phrase Global Saving Glut to suggest that
the large current account deficits run by the Unites States since the 1990s were driven primarily by a high level of desired saving in some developing economies, such as EmergingAsia and the oil exporting countries. A key motivation behind this “somewhat unconventional explanation of the high and rising U.S. current account deficit” (Bernanke, 2005) was
Greenspan’s (2005) so-called “conundrum,” namely the puzzling fact that bond yields were
falling in the mid 2000s, even in the face of a monetary policy tightening. This decline is
illustrated in figure 4.1, which presents two measures of ex-ante real bond yields: the difference between the nominal yield on 10-year Treasury Notes and expected inflation from the
Survey of Professional Forecasters, and the yield of 10-year Treasury Inflation-Protected
Securities (TIPS). In light of this puzzle, Bernanke’s (2005) key observation was that an
increase in desired saving around the world could be consistent with both low interest rates
globally, and with a current account deficit in the U.S., as an endogenous reaction to those
low rates. Explanations of global imbalances based on domestic developments, on the contrary, such as a fall in desired saving by U.S. households, would be harder to square with
these observations on both prices and quantities.
One of the stylized facts associated with the saving glut is that the capital inflows generated by the surplus countries were directed mostly towards safe assets, such as Treasuries
THE EFFECTS OF THE SAVING AND BANKING GLUT ON THE U.S. ECONOMY
20
and Agency debt, and to a lesser extent towards equity (Bernanke et al., 2011). In our
model, however, this portfolio composition of the flows is irrelevant. The macroeconomic
effects of net capital flows depend instead only on the size of the associated trade imbalance.
To see more clearly why the portfolio composition of these flows can be ignored, suppose
that a foreign agent wishes to purchase U.S. bonds in exchange for a certain amount of
goods, thereby contributing to the trade deficit. Domestic lenders—the holders of these
bonds in our economy—require a reduction in interest rates to deviate from their original
consumption/saving plan, as implied by their undistorted Euler equation. Since under
perfect foresight all returns are known with certainty, no-arbitrage transmits this decline
in rates of return equally across all assets in the economy, making the domestic lender
indifferent between selling Treasuries, mortgages/MBS or shares of the capital stock to the
foreign agent. Therefore, in the experiments that follow, treat the trade deficit as the forcing
variable, and ignore the composition of flows, focusing instead on how the (net) amount
of goods shipped from abroad affects the intertemporal consumption profile of domestic
agents, and hence interest rates.
This exclusive focus on goods (as opposed to assets) may seem surprising in an economy
with collateral-constrained agents. However, the presence of the savers, who can trade all
assets and do not face a constraint, makes the previous simple intuition valid also in our
setup. In our model, international trade is ultimately an intertemporal reshuffling of consumption profiles, with effects on the interest rate dictated by the savers’ Euler equation.
The presence of borrowers, however, matters both quantitatively and qualitatively. Quantitatively, the movement in interest rates engendered by trade is different than in a model
without constrained agents, since borrowers can consume more only to the extent that their
collateral gains value. Qualitatively, the transmission mechanism of trade deficit shocks to
the domestic economy is enriched by the presence of heterogeneity among households, which
also allows us to trace the implications of these shocks for debt and house prices.
The irrelevance of the portfolio composition of the foreign capital flows in our model, at
least under the assumption of perfect foresight, is of course not a claim that this composition
does not matter in practice. In this respect, our experiments only quantify the “first-order”
effects of changes in trade flows on interest rates, and hence on the rest of the economy,
ignoring the “higher-order” considerations stemming from the presence of time-varying risk,
and the associated portfolio choices.
THE EFFECTS OF THE SAVING AND BANKING GLUT ON THE U.S. ECONOMY
21
Our calculations, therefore, can be thought of as a general equilibrium version of the
empirical exercises of McCarthy and Peach (2004) and Himmelberg, Mayer, and Sinai
(2005), who used present value models of house prices to measure the impact of low interest
rates (and other fundamentals) on real estate values in the mid 2000s. A recent and growing
literature in international finance includes a role for portfolio choice in shaping the impact
of international capital flows on macroeconomic outcomes (e.g. Mendoza, Quadrini, and
Rios-Rull, 2009, Caballero and Krishnamurty, 2009, Tille and van Wincoop, 2010, Devereux
and Sutherland, 2011, Caballero and Farhi, 2013). Extending our model along these lines
is an exciting avenue for future research.
Given that the effects of the SG in our economy depend on the size of the trade deficit,
rather than from the composition of capital inflows, we feed into the model a sequence for net
foreign debt over GDP {Df,t /Yt }2012:Q4
t=1998:Q1 that is consistent with the observed evolution of
the trade balance. As shown in figure 4.2, the U.S. trade deficit deteriorated significantly—
from approximately 1.5 to 6 percent of GDP between 1998:Q1 and 2006:Q3. We calculate
the implied path of capital inflows (i.e. the sequence of ∆Dtf ) consistent with a linear
approximation to the data over the period of rising deficits (dashed line in figure 4.2), given
f
the interest payments on foreign debt implied by the model (Rt−1 Dt−1
), and the initial level
of debt (Dtf ) observed in 1998. The solid lines in figure 4.3a and 4.3b plot the resulting
time series of the U.S. net foreign debt position as a fraction of GDP (i.e. Dtf / (4Yt )), which
we treat as the exogenous driver of the SG, and the associated trade balance.
As mentioned in section 2.5, the quantitative experiments illustrated below only cover the
period of expanding imbalances leading up to the financial crisis, since the global saving glut
hypothesis provides a much more compelling account of this period than of the retrenchment
that followed.10
Since agents in the model are forward looking, we also need to specify their expectations
regarding the future evolution of capital flows. One possibility would be to assume that
they perfectly anticipated the realized path of capital flows as of 1998, following an initial
surprise, but this seems implausible. Instead, we assume that they were a bit surprised
at every step along the way by the deterioration of the trade balance. After each of these
10In a previous version of the paper we continued the experiments for both the saving and the banking
glut through 2012. The results for the period of current account reversal following the crisis are roughly
symmetric to those illustrated below, with some differences due to the asymmetry of the collateral constraint
emphasized in section 2.1.1.
THE EFFECTS OF THE SAVING AND BANKING GLUT ON THE U.S. ECONOMY
8
22
6
Data
Model
6
4
4
2
2
0
0
1990
1992
1994
1996
1998
2000
2002
2004
2006
−2
2008
Figure 4.2. U.S. trade deficit-to-GDP ratio (solid, left axis) and linear
approximation fed through the model (dashed, right axis). Source: Haver
and authors’ calculations.
a) Foreign debt−to−GDP ratio
36
34
Realized
Expected in 2000:q1
Expected in 2003:q1
b) Trade deficit−to−GDP ratio
4.5
4
32
3.5
30
3
28
2.5
26
24
2
22
1.5
20
1
18
0.5
16
1998
0
2000
2002
2004
2006
1998
2000
2002
2004
2006
Figure 4.3. Realized paths in the simulation, and agents’ expected paths
in 2000:Q1 and 2003:Q1
surprises, they revise their expectations of the eventual size of the foreign debt-to-GDP
ratio, as shown in figure 4.3a. At each point in time, these expectations on the evolution
of the net foreign asset position are consistent with a gradual improvement in the trade
and current account deficits (figure 4.3b), as captured by an AR(1) process for ∆Dtf with
THE EFFECTS OF THE SAVING AND BANKING GLUT ON THE U.S. ECONOMY
23
persistence equal to 0.95. The pace of this improvement is roughly consistent with 1 to
4-quarter ahead expectations for real net exports from the Blue Chip survey, as well as
with the speed of current account rebalancing implied by the model of Ferrero, Gertler,
and Svensson (2010).
To illustrate the expected paths and the realized surprises in our simulation, the dashed
line in figure 4.3a represents the projected path of foreign-lending-to-GDP in 2000:Q1.
However, in the following quarter, capital inflows are higher than originally anticipated, as
shown in figure 4.3a. In 2003:Q1, agents again forecast a gradual tapering of capital inflows,
associated with a protracted improvement of the trade balance. However, as before, inflows
and the trade deficit are unexpectedly higher the next quarter. Similar surprises happen
every quarter of the simulation until 2006:Q3, but we omit the expected paths to avoid
cluttering the figure.11
Figures 4.4 and 4.5 summarize the effects of the SG experiment in the calibrated model.
The desire of foreign agents to purchase more U.S. assets decreases real interest rates to
induce the domestic lenders—the owner of these assets—to reduce saving and increase
consumption (figure 4.4a). In our calibration, the SG reduces nominal interest rates by
approximately 150 basis points (figure 4.4b), which is broadly consistent with some recent
empirical studies on the effect of the SG on asset returns (Bertaut at al., 2011 and Warnock
and Warnock, 2009). Lower interest rates also stimulate the demand for nonresidential
investment (figure 4.4c) and real estate by the domestic lenders, by making these durable
goods relatively more desirable than nondurable consumption.
The resulting upward pressure on house prices relaxes the collateral constraint on the
borrowers. This reduces their urge to consume in the present, as measured by the multiplier
on the borrowing constraint, and makes them more willing to hold durable goods. From
their Euler equation,
1 − ξt
Rb,t
= Et
βb Λb,t+1
,
Λb,t
11If agents anticipate correctly the entire extent of the deterioration in the foreign debt position following
an initial surprise in 1998, as under standard perfect foresight simulations, their response is almost entirely
front-loaded, and implausibly large. The sequence of surprises that they experience under our simulations,
aside from being more realistic, also contributes to smoothing their reaction, thus delivering a more plausible
dynamic profile of the endogenous variables. In this respect, this assumption on expectations is related to
the learning mechanisms explored by Adam, Kuang, and Marcet (2011) and Boz and Mendoza (2010), also
in an open economy context.
THE EFFECTS OF THE SAVING AND BANKING GLUT ON THE U.S. ECONOMY
a) Real interest rate (lending)
24
b) Nominal interest rate (lending)
2.8
4.5
2.7
2.6
4
2.5
3.5
2.4
1998
2000
2002
2004
2006
1998
2000
2002
2004
2006
d) Consumption
c) Nonresidential Inv.
105
110
104
103
105
102
101
100
1998
100
2000
2002
2004
2006
1998
e) House price
2000
2002
2004
2006
f) Debt−to−GDP ratio
0.5
110
0.48
105
0.46
100
1998
0.44
2000
2002
2004
2006
1998
2000
g) GDP
2002
2004
2006
2004
2006
h) Labor
102
101.5
100.5
101
100.5
100
100
99.5
1998
2000
2002
2004
2006
99.5
1998
2000
2002
Figure 4.4. Saving glut experiment: aggregate variables
the fall in the multiplier ξt associated with the relaxation of the borrowing constraint
increases the marginal valuation of future consumption vis-a-vis current consumption (the
right-hand side of the expression), for any given level of the interest rate. Therefore, the
borrowers’ demand for houses increases more than that for nondurable consumption, and
more than that of the lenders. In equilibrium, house prices rise substantially because both
agents increase demand, and supply is fixed (figure 4.4e). Moreover, houses are sold by the
THE EFFECTS OF THE SAVING AND BANKING GLUT ON THE U.S. ECONOMY
a) Housing (borrower)
25
b) Housing (saver)
108
100
99
106
98
104
97
102
96
95
100
1998
2000
2002
2004
2006
1998
2000
2002
2004
2006
d) Consumption (saver)
c) Consumption (borrower)
106
105
104
104
103
102
102
101
100
100
1998
2000
2002
2004
2006
1998
e) Labor (borrower)
2000
2002
2004
2006
f) Labor (saver)
100.5
101
100
100.5
100
99.5
99.5
1998
2000
2002
2004
2006
1998
g) Real wage (borrower)
2000
2002
2004
2006
h) Real wage (saver)
102
101
101.5
100.5
101
100.5
100
100
99.5
1998
2000
2002
2004
2006
99.5
1998
2000
2002
2004
2006
Figure 4.5. Saving glut experiment: borrowers and lenders
lenders to the borrowers, since their demand increases more in relative terms (figures 4.5a
and 4.5b). Finally, with higher house prices, debt increases as well (figure 4.4f).
To illustrate the behavior of the housing market more clearly, consider the borrowers’
first order condition with respect to the housing stock
βb Λb,t+1
h,c
h
h
(4.1)
Pt (1 − ξt θ) = Et
M RSb,t+1 + (1 − δh ) Pt+1 ,
Λb,t
THE EFFECTS OF THE SAVING AND BANKING GLUT ON THE U.S. ECONOMY
26
where M RS h,c denotes the marginal rate of substitution between housing and consumption.
Substituting for the the discount factor in equation (4.1), and recalling that there is no
uncertainty about future realizations of returns under perfect foresight, we obtain
(4.2)
Pth =
i
1 1 − ξt h
h,c
h
M RSb,t+1
+ (1 − δh ) Pt+1
.
1 − ξt θ Rb,t
This equation can be interpreted as a downward sloping demand for houses, since the
MRS is decreasing in the housing stock, conditional on expectations about the future price.
Comparing this expression with the equivalent one for the lenders
Pth =
i
1 h
h,c
h
M RSl,t+1
+ (1 − δh ) Pt+1
,
Rl,t
we see that, as long as θ < 1, the borrowers’ valuation is higher, the more they can borrow
against the house (θ) and the slacker is the constraint (i.e. the lower is ξt ).
The net positive effect on house prices from a loosening of credit conditions depends on
two countervailing forces. First, a looser constraint makes an extra unit of housing less
valuable as collateral, as captured by the multiplier ξt at the denominator of
1 1 − ξt
.
1 − ξt θ Rb,t
Second, a lower ξt makes borrowers effectively less impatient, increasing their valuation of
future vs. current consumption, as captured by the ξt in the numerator. Since θ is less than
one, this positive effect prevails in our model, the more so the more durable is the housing
stock (i.e. the lower is δh ).
We combine these two demand functions in figure 4.6, together with the fixed supply
of houses. The demand by the borrowers is measured from left to right in the diagram,
h . A fall in the
while that of the savers from right to left, for a given future price Pt+1
interest rate shifts both schedules up, putting upward pressure on the house price.12 For a
given ξt , the demand of the lenders changes by more than that of the borrowers. However,
the relaxation of the borrowing constraint associated with this increase in collateral values
shifts the borrowers’ demand further, by lowering the multiplier ξt . In our simulations, the
borrowers’ valuation increases by more, so that some houses change hands from the savers
to the borrowers, as shown in the diagram.
12Changes in the interest rate shift and rotate the demand function at the same time.
THE EFFECTS OF THE SAVING AND BANKING GLUT ON THE U.S. ECONOMY
27
Figure 4.6. Housing demand by the borrowers and the lenders.
Iterating equation (4.2) forward
∞
s
X
Y
1
−
ξ
1
t+j−1
h,c
,
Pth =
(1 − δh ) M RSb,t+s
1 − ξt+j−1 θ Rb,t+j−1
s=1
j=1
we can also see that the borrowers’ valuation of housing reflects their expectations on
the tightness of the constraint going forward. Therefore, the longer a loosening in the
collateral constraint is expected to last, the larger the resulting shift into housing, relative
to nondurable consumption.
Turning to the implications of the SG for economic activity, figure 4.4g shows that
it generates an expansion through fairly standard channels. Lower interest rates boost
investment and therefore the capital stock, raising the marginal product of labor and labor
demand. The increase in labor demand is larger under sticky prices, because firms need to
hire to satisfy the higher demand for consumption and investment goods (i.e. markups fall).
With the standard King-Plosser-Rebelo preferences (µ = 1), labor supply would decline due
to the usual wealth effect. Under JR preferences with µ ≈ 0, instead, the position of the
labor supply schedule depends on the housing-to-consumption ratio, rather than on the
marginal utility of consumption alone. This explains why the borrowers work more, while
the savers do not (figures 4.5e and 4.5f). In sum, the boom in real activity reflects the
increase in the aggregate labor input and, particularly, in capital.
THE EFFECTS OF THE SAVING AND BANKING GLUT ON THE U.S. ECONOMY
28
4.2. Banking glut. In contrast to the saving glut, the banking glut refers to large gross
inflows of foreign capital from advanced economies, mostly Europe, which were matched by
similar outflows, resulting in small net flows and minor changes in the U.S. trade balance
(Shin, 2012). Compared to the portfolio allocation of the SG countries, which was concentrated in Treasury and agency debt, the composition of the European gross capital inflows
was more widely spread across asset classes, with a particular concentration in private-label
ABS, corporate bonds and (to a lesser extent) equity (Bernanke et al. 2011). According to
the empirical literature, the higher demand for ABS and corporate bonds from European
investors put downward pressure on the spreads of these securities relative to Treasuries,
despite significant increases in their origination (Bertaut et al. 2011).
Consistent with this description, our model captures the BG as a gross inflow of capital
into the mortgage market, which is not associated with movements in the trade balance.
More specifically, we model the demand by European investment banks for structured
mortgage products, and their associated issuance of credit instruments on the domestic
wholesale credit market, as an increase in the competitive pressure in this market. This
increase in the number of players and in the credit products they issue compresses the
spread between borrowing and lending rates, as shown in equation (2.4) when Nt increases.
Direct evidence on the quantitative implications for spreads of this increased participation
by European banks in the U.S. wholesale credit market is difficult to obtain, since spreads
between mortgage products and funding rates move for many different reasons. Therefore,
our calibration is based on the fact that the spread between mortgage rates and Treasuries
of comparable maturities declined by between 50 and 150 basis points between 2003 and
2007, at the time of the surge of European funds into the market for ABS. Moreover, this
spread compression was more pronounced in mortgage products directly connected with
the subprime boom, which featured prominently in the portfolios of European banks.13
Taking this evidence as a rough guideline, our simulation assumes that in every quarter
between 2003:Q1 and 2006:Q3, agents in the model are surprised with an unanticipated decline in the spread, which they expect to be permanent. Given the ambiguity surrounding
the appropriate choice of a spread measure, our BG experiment assumes a gradual decline
13See, for instance, figures 16 and 17 in Bertaut at al. (2011). The portfolio balance model estimated by
these authors suggests that the European purchases of ABS during this period lowered ABS yields by 60
basis points. Including the purchases of non-ABS corporate debt, the negative effect on yields reaches 160
basis points. However, these estimates do not account for the impact of the increase in supply of these
instruments, which presumably contributed to reduce these effects on rates.
THE EFFECTS OF THE SAVING AND BANKING GLUT ON THE U.S. ECONOMY
29
in the steady state spread by a total of 75 basis points. This number approximately corresponds to the decline in the spread between the 30-year mortgage rate and the 10-year
Treasury yield reported by Bertaut et al. (2011) and Bernanke et al. (2011). We also experimented with smaller (50 basis points) and larger declines (1 percent), and found roughly
proportional changes in the model’s dynamics.
Our findings on the implications of the banking glut for the U.S. economy are summarized
in figures 4.7 and 4.8. Faced with a lower nominal (and real) interest rate, the borrowers
reallocate their demand from nondurable consumption towards houses. This shift in demand
puts upward pressure on house prices, which rise by 4% during the first five years. Higher
collateral values allow them to borrow more, and ultimately to increase also the consumption
of nondurable goods. Moreover, the relaxation of the borrowing constraint entails an even
larger decline in the effective interest rate faced by borrowers ( Rt /(1 − ξt )), due to the
current and future expected fall in the multiplier. As a result, the borrowers’ demand for
housing increases substantially more than that for nondurable goods.
Absent any net injections of foreign capital, additional lending from the domestic savers
to satisfy the increased borrowing by the impatient households requires an increase in the
interest rate, which leads them to postpone consumption and investment. This reduction
in the savers’ demand almost exactly counteracts the expansion from the borrowers, resulting in a negligible overall increase in consumption and output. These offsetting effects
of the responses of borrowers and savers on aggregate activity are also a feature of the
closed-economy experiments in Justiniano, Primiceri, and Tambalotti (2012). As in those
experiments, we also find a sizable re-allocation of houses from the borrowers to the savers.
Our simulation of the BG is subject to at least two caveats. First, empirical evidence
suggests that capital inflows from the SG economies also contributed to the reduction in
spreads of mortgage over Treasury yields, perhaps by substantial amounts (Bertaut et al.,
2011, Warnock and Warnock, 2009). Since our model is not well equipped to capture this
effect of the SG, we might be overestimating the impact of the BG relative to the SG.
Second, foreign inflows were not the only factor behind the observed reduction in mortgage
spreads, which is why we view the magnitude of our simulations as only suggestive of the
potential effects of a compression in spreads that might be attributed to the BG channel.
4.3. Saving and banking gluts: a quantitative perspective. This subsection presents
the results of a joint simulation of the SG and the BG. From 1998:Q1 through 2002:Q4,
THE EFFECTS OF THE SAVING AND BANKING GLUT ON THE U.S. ECONOMY
a) Real interest rate (borrowing)
30
b) Nominal interest rate (borrowing)
4.3
6.3
4.2
6.2
4.1
4
6.1
3.9
6
3.8
1998
2000
2002
2004
2006
1998
c) Nonresidential Inv.
2000
2002
2004
2006
d) Consumption
100
100.5
99.5
100
99
98.5
1998
2000
2002
2004
2006
99.5
1998
e) House price
2000
2002
2004
2006
f) Debt−to−GDP ratio
104
0.5
103
0.48
102
0.46
101
100
1998
0.44
2000
2002
2004
2006
1998
2000
g) GDP
2002
2004
2006
2004
2006
h) Labor
100.6
100.6
100.4
100.4
100.2
100.2
100
100
99.8
99.8
99.6
1998
99.6
2000
2002
2004
2006
1998
2000
2002
Figure 4.7. Banking glut: aggregate variables
only the SG is active and the transition paths are identical to those described in section 4.1.
After 2003:Q1, the BG is also active. Given the focus of the previous subsections on the
qualitative mechanisms at play in the model, here we concentrate on the overall quantitative
impact of these international forces, with particular attention to their contribution to the
housing and credit boom.
In our combined experiment, real house prices increase by 17% between 1998 and 2007
(figure 4.9e). Therefore, our results attribute between one-third and one-fourth of the
THE EFFECTS OF THE SAVING AND BANKING GLUT ON THE U.S. ECONOMY
a) Housing (borrower)
31
b) Housing (saver)
100
110
98
96
105
94
92
100
1998
2000
2002
2004
2006
1998
c) Consumption (borrower)
2000
2002
2004
2006
d) Consumption (saver)
100.5
101.5
100
101
99.5
100.5
99
100
99.5
1998
2000
2002
2004
2006
98.5
1998
2000
2002
2004
2006
f) Labor (saver)
e) Labor (borrower)
101
100.4
100.2
100.5
100
99.8
100
99.6
99.4
99.5
1998
2000
2002
2004
2006
1998
g) Real wage (borrower)
2000
2002
2004
2006
h) Real wage (saver)
101
100.5
100.5
100
100
99.5
1998
2000
2002
2004
2006
99.5
1998
2000
2002
2004
2006
Figure 4.8. Banking glut: borrowers and lenders
recent house-price boom to the imbalances vis-a-vis the rest of the world. We regard the
contribution of these international forces as realistic, considering that the literature has
identified a host of domestic factors as important sources of the U.S. housing boom and
bust and of the associated financial crisis, such as changes in the availability of credit on
both the intensive and extensive margin (Mian and Sufi, 2009, Favilukis et al., 2012, Boz
and Mendoza, 2010), loose monetary policy (Taylor, 2008, Ferrero, 2012), or deviations of
house prices from fundamentals (Shiller, 2007).
THE EFFECTS OF THE SAVING AND BANKING GLUT ON THE U.S. ECONOMY
32
b) Nominal interest rate (lending)
a) Real interest rate (lending)
2.8
4.5
2.7
2.6
4
2.5
1998
2000
2002
2004
2006
3.5
1998
c) Nonresidential Inv.
2002
2004
2006
d) Consumption
110
105
108
104
106
103
104
102
102
101
100
1998
2000
100
2000
2002
2004
2006
1998
e) House price
2000
2002
2004
2006
f) Debt−to−GDP ratio
0.55
115
110
0.5
105
0.45
100
1998
2000
2002
2004
2006
1998
2000
g) GDP
2002
2004
2006
2004
2006
h) Labor
102
101.5
100.5
101
100.5
100
100
99.5
1998
2000
2002
2004
2006
99.5
1998
2000
2002
Figure 4.9. Saving and banking glut: aggregate variables
Our results suggest an even larger impact of the SG and the BG on the behavior of
household debt. The debt-to-GDP ratio rises steadily in the joint simulation to a level
of 56% (figure 4.9f), accounting for more than one-third of the observed run-up over the
same period (0.44 to 0.75). With regards to the relative contributions of the BG and SG,
our simulations attribute a greater quantitative role to the latter, a point reinforced by the
caveats on the calibration of the size of the BG impulse raised in section 4.2.
THE EFFECTS OF THE SAVING AND BANKING GLUT ON THE U.S. ECONOMY
33
The remaining panels in figure 4.9 plot the behavior of the real interest rate, GDP,
investment and aggregate consumption in the joint experiment. As for most variables in
the model, their transition paths are roughly equal to the sum of the responses in the
two experiments taken in isolation.14The effects on activity are mostly shaped by the SG,
due to the nearly offsetting responses of borrowers and savers to the BG. Investment and
consumption rise considerably above trend, but the expansionary effects on GDP are more
muted, due to the drag coming from the trade deficit.
Overall, the quantitative experiments illustrated above suggest that the sustained current
account deficits run by the United States in the last fifteen years, and to a lesser extent
the gross capital flows into mortgage-backed securities mediated by European banks, had a
sizable impact on the equilibrium of the domestic economy. These flows of goods and assets
reduced interest rates and spreads, and lifted nondurable consumption and house values.
As a result, households’ ability to borrow increased, accounting for more than one third of
the observed surge in debt.
An important qualification to this conclusion is that, given the complexity of the current
model, we took an extremely stylized view of international trade, as an exogenous flow
of imports into the United States from the RoW, with no exports going in the opposite
direction. This simplification is consistent with the two main tenets of the SG hypothesis put
forth by Bernanke (2005). First, that international imbalances are best understood through
the lens of the balance between desired saving and investment—the current account—rather
than through that of the balance between exports and imports—the trade deficit. Second,
that the main source of these imbalances since the late nineties were developments in the
RoW, rather than in the U.S., and in particular the surge in desired saving in emerging
Asia and the Middle East, looking for safe harbor in the United States.
4.3.1. Robustness: supply of new houses. Our baseline model assumes that the supply of
houses expands at the rate of technological progress. This simplifying assumption is not
crucial for our result that the dynamics of foreign capital flows account for between one
fourth and one third of the increase in U.S. house prices and household debt. To illustrate
14The most notable exception is the behavior of household debt, which increases less in the combined
experiment. The reason is the “zero lower bound” on the multiplier ξt . In the SG experiment, the collateral
constraint does not bind while the trade deficit rises (from 1998:Q1 to 2006:Q3). Since the multiplier
is already equal to zero, the reduction in the spread that starts in 2003:Q1 does not have an additional
expansionary effect on borrowing, because the borrowers are already unconstrained.
THE EFFECTS OF THE SAVING AND BANKING GLUT ON THE U.S. ECONOMY
34
this point, we experimented with a more complex version of the model, in which houses
are produced combining a fixed factor (land) and residential investment, as in Davis and
Heathcote (2005). The production of residential investment is subject to adjustment costs
similar to those in equation (2.1), parameterized to approximately match the relative increase in the price and quantity of residential investment observed between 1998 and 2006.
In this version of the model, house prices increase almost as much as in our simpler baseline specification (close to 14%). Meanwhile, the ratio of residential investment to GDP
experiences a similar increase (roughly 17%) from its steady state level.
5. Concluding Remarks
The United States has been running large trade and current account deficits for more
than twenty years. These deficits grew steadily during the 1990s and 2000s, and fell only
partially following the Great Recession. We studied the quantitative effects of this massive
transfer of resources from the Rest of the World to the U.S. economy, with a particular focus
on its impact on the credit and real estate boom that precipitated the Great Recession.
We did so with help from a dynamic, general equilibrium model that features most of the
now standard ingredients in the DSGE literature. The model also includes borrowing and
lending among heterogeneous households, a non-competitive financial intermediation sector
that charges a spread on borrowing rates, and a stylized foreign sector that lends resources
to the domestic economy in exchange for a promise of more resources in the future. With
these ingredients, we can capture two channels through which global imbalances affected
the U.S. economy. The first channel is the global saving glut popularized by Bernanke
(2005), which we represent as an influx of real resources in the U.S. economy calibrated to
the observed sequence of trade deficits.
The second channel is Shin’s (2012) global banking glut, which refers to the effect of
the gross capital flows intermediated by global (mostly European) banks into the U.S.
mortgage market. The evidence recently collected by Bertaut et al. (2011) suggests that
these gross purchases of mortgage instruments had a significant impact on interest rates
and spreads in the mortgage market, even though they were largely financed on domestic
credit markets, giving rise to virtually no current account imbalance. We capture this
second channel through a compression in the spread charged by financial intermediaries,
THE EFFECTS OF THE SAVING AND BANKING GLUT ON THE U.S. ECONOMY
35
associated with the increased competition in the wholesale funding market triggered by the
entry of European banks.
The main finding is that the saving and banking glut together were an important contributor to some of the key macroeconomic developments in the U.S. during the 2000s, and
in particular in the period preceding the financial crisis. In the calibrated model, these
two forces contribute to a large boom in consumption and investment, and account for
between one fourth and one third of the observed run-up in real house prices, and more
than a third in that of household debt. Between the two channels, the saving glut has much
larger effects, since it generates a fall in interest rates for both borrowers and savers that
stimulates their demands, especially of durable goods such as houses. The compression in
spreads with which we capture the banking glut, instead, elicits opposite responses between
borrowers and lenders, leading to smaller effects on aggregate variables.
Our modeling strategy misses at least two potentially important dimensions of the connection between global imbalances and domestic macroeconomic developments. First, the
portfolio composition of the financial flows that mirror the observed trade deficit is irrelevant in our model. This is because agents have perfect foresight, and therefore perceive no
risk in the returns of the assets they hold, making them indifferent about the composition
of their portfolio. The main advantage of this strategy is that it focuses on the overall
macroeconomic effect of trade, without having to worry about the underlying financial
flows, which can be difficult to track. The clear disadvantage is that we cannot address the
important issues raised by the strong preference of the saving glut countries, and of China
in particular, for safe U.S. dollar assets. These issues are the focus of Caballero, Farhi,
and Gourinchas (2008) and Mendoza, Quadrini, and Rios-Rull (2009), and of a growing
literature that followed them.
Second, there are no exports and imports explicitly in the model, just resources made
available from the rest of the world for consumption in the U.S., and no structure on
the reason why foreign agents are willing to lend those resources today. Therefore, we
have nothing to say about what happens in the rest of the world, nor about the many
channels through which imports, exports and the exchange rate affect domestic allocations.
Nevertheless, the model provides useful insights on some of the basic mechanisms through
which international imbalances affect a large economy, as well as a reasonable quantification
of their aggregate effects.
THE EFFECTS OF THE SAVING AND BANKING GLUT ON THE U.S. ECONOMY
36
References
Acharya, Viral, and Philipp Schnabl. 2009. “Do Global Banks Spread Global Imbalances? The Case of
Asset-Backed Commercial Paper during the Financial Crisis of 2007–2009.” IMF Economic Review, 58(1):
37–73.
Adam, Klaus, Pei Kuang, and Albert Marcet. 2011. “House Price Booms and the Current Account.” NBER
Working Paper no. 17224.
Adrian, Tobias, and Hyun Song Shin. 2010. “The Change in Nature of Financial Intermediation and the
Financial Crisis of 2007-09.” Annual Review of Economics, 2, 603-618.
Ashcraft, Adam B., and Til Schuermann. 2008. “Understanding the securitization of subprime mortgage
credit.” Federal Reserve Bank of New York Staff Reports, 318.
Bernanke, Ben S. 2005. "The Global Saving Glut and the US Current Account Deficit." Sandridge Lecture,
Virginia Association of Economists, April 14, 2005.
Bernanke, Ben S., Carol Bertaut, Laurie Pounder DeMarco, and Steven B. Kamin. 2011. “International
Capital Flows and the Returns to Safe Assets in the United States, 2003–2007." Banque de France Financial
Stability Review, 15: 13–26.
Bertaut, Carol C., Laurie Pounder DeMarco, Steven Kamin, and Ralph W. Tryon. 2011. "ABS Inflows to
the United States and the Global Financial Crisis." International Finance Discussion Papers 1028, Board
of Governors of the Federal Reserve System.
Bertaut, Carol, and Laurie Pounder DeMarco. 2009. “The Financial Crisis and U.S. Cross Border Financial
Flows.” Federal Reserve Bulletin, November 2009: A147–A167.
Bilbiie, Florin O., Fabio Ghironi, and Marc J. Melitz. 2012. “Endogenous Entry, Product Variety, and
Business Cycles.” Journal of Political Economy, 120(2), 304–345.
Boz, Emine, and Enrique G. Mendoza. 2010. “Financial Innovation, the Discovery of Risk, and the U.S.
Credit Crisis.” NBER Working Paper no. 16020.
Brunnermeier, Markus K. 2009. “Deciphering the Liquidity and Credit Crunch 2007-08.” Journal of Economic Perspectives, 23(1), 77–100.
Caballero, Ricardo J., and Emmanuel Farhi. 2013. “A Model of the Safe Asset Mechanism (SAM): Safety
Traps and Economic Policy.” NBER Working Paper no. 18737.
Caballero, Ricardo J., Emmanuel Farhi, and Pierre-Olivier Gourinchas. 2008. “An Equilibrium Model of
Global Imbalances and Low Interest Rates.” American Economic Review, 98(1), 358–393.
Caballero, Ricardo, and Arvind Krishnamurthy. 2009. “Global Imbalances and Financial Fragility.” American Economic Review, 99(2), 584-88.
Calvo, Guillermo A. 1983. “Staggered prices in a utility-maximizing framework.” Journal of Monetary
Economics, 12(3): 383–398.
Campbell, Jeffrey R., and Zvi Hercowitz. 2009. “Welfare implications of the transition to high household
debt.” Journal of Monetary Economics, 56(1), 1–16.
THE EFFECTS OF THE SAVING AND BANKING GLUT ON THE U.S. ECONOMY
37
Christiano, Lawrence J., Martin Eichenbaum, and Charles L. Evans. 2005. “Nominal rigidities and the
dynamic effects of a shock to monetary policy.” Journal of Political Economy, 113(1), 1–45.
Cúrdia, Vasco, and Michael Woodford. 2010. “Credit spreads and monetary policy.” Journal of Money,
Credit and Banking, 42(1), 3–35.
Davis, Morris A., and Jonathan Heathcote. 2005. "Housing and the business cycle." International Economic
Review, 46(3), 751–784.
Demyanyk, Yuliya, and Van Hemert, Otto. 2011. “Understanding the Subprime Mortgage Crisis.” Review
of Financial Studies, 24(6), 1848–1880.
Devereux, Michael B., and Alan Sutherland. 2011. "Country Portfolios In Open Economy Macro-Models."
Journal of the European Economic Association, 9(2), 337–369.
Duca, John V., John Muellbauer, and Anthony Murphy. 2011. “House Prices and Credit Constraints:
Making Sense of the US Experience.” The Economic Journal, 121(552), 533–551.
Eberly, Janice, Sergio Rebelo, and Nicolas Vincent. 2012. "What explains the lagged investment effect?"
Journal of Monetary Economics, 59(4), 370–380.
Favilukis, Jack, David Kohn, Sydney C. Ludvigson, and Stijn Van Nieuwerburgh. 2013. “International
Capital Flows and House Prices: Theory and Evidence.” Housing and the Financial Crisis, University of
Chicago Press.
Favilukis, Jack, David Kohn, Sydney C. Ludvigson, and Stijn Van Nieuwerburgh. 2012. “Foreign Ownership
of U.S. Safe Assets: Good or Bad?” NYU Working Paper no. FIN-11-057.
Feenstra, Robert C. 2003. “A homothetic utility function for monopolistic competition models, without
constant price elasticity.” Economics Letters, 78(1), 79–86.
Ferrero, Andrea. 2012. "House Price Booms, Current Account Deficits, and Low Interest Rates." Federal
Reserve Bank of New York Staff Reports, 541.
Ferrero, Andrea, Mark Gertler, and Lars EO Svensson. 2010. “Current account dynamics and monetary
policy.” International Dimensions of Monetary Policy, Jordi Gali and Mark Gertler, editors. University of
Chicago Press.
Garcia-Cicco, Javier, Roberto Pancrazi, and Martin Uribe. 2010. “Real Business Cycle in Emerging Countries? ”American Economic Review, 100(5), 2510–2531.
Garriga, Carlos, Rodolfo E. Manuelli, and Adrian Peralta-Alva. 2012. “A Model of Price Swings in the
Housing Market.” Federal Reserve Bank of St. Louis Working Paper no. 2012-022A.
Gerali, Andrea, Stefano Neri, Luca Sessa, and Federico M. Signoretti. 2010. “Credit and Banking in a
DSGE Model of the Euro Area.” Journal of Money, Credit and Banking, 42(1), 107–141.
Gorton, Gary. 2009. “The Panic of 2007.” Maintaining Stability in a Changing Financial System, Proceedings of the 2008 Jackson Hole Conference, Federal Reserve Bank of Kansas City.
Gourinchas, Pierre-Olivier, and Hélène Rey. 2013. “External Adjustment, Global Imbalances, Valuation
Effects.” Handbook of International Economics, vol. IV, ed. Gita Gopinath, Elhanan Helpman and Kenneth
Rogoff.
THE EFFECTS OF THE SAVING AND BANKING GLUT ON THE U.S. ECONOMY
38
Greenspan, Alan. 2005. “Testimony Before the Committee on Banking, Housing, and Urban Affairs,” U.S.
Senate, February 16.
Greenwood, Jeremy, Zvi Hercowitz, and Gregory W. Huffman. 1998. “Investment, capacity utilization, and
the real business cycle.” The American Economic Review, 78(3), 402–417.
Hansen, Gary D. 1985. “Indivisible labor and the business cycle.” Journal of Monetary Economics, 16(3),
309–327.
Himmelberg, Charles, Christopher Mayer and Todd Sinai. 2005. “Assessing High House Prices: Bubbles,
Fundamentals and Misperceptions.” Journal of Economic Perspectives, 19(4), 67–92.
Iacoviello, Matteo. 2005. “House prices, borrowing constraints, and monetary policy in the business cycle.”
The American Economic Review, 95(3), 739–764.
Jaimovich, Nir, and Max Floetotto. 2008. “Firm dynamics, markup variations, and the business cycle.”
Journal of Monetary Economics, 55(7), 1238–1252.
Jaimovich, Nir, and Sergio Rebelo. 2009. “Can News about the Future Drive the Business Cycle?” American
Economic Review, 99(4), 1097–1118.
Julliard, Michel, Douglas Laxton, Peter McAdam, and Hope Pioro. 1998. “An algorithm competition:
First-order iterations versus Newton-based techniques.” Journal of Economic Dynamics and Control, 22:
1291–1318.
Justiniano, Alejandro, Giorgio Primiceri, and Andrea Tambalotti. 2011. “Investment Shocks and the Relative Price of Investment.” Review of Economic Dynamics, 14(1), 102–121.
Justiniano, Alejandro, Giorgio Primiceri, and Andrea Tambalotti. 2012. “Household Leveraging and
Deleveraging.” NBER Working Paper no. 18941.
Kaplan, Greg, and Giovanni L. Violante. 2011. “A model of the consumption response to fiscal stimulus
payments.” NBER Working Paper no. 17338.
King, Robert G., Charles I. Plosser, and Sergio T. Rebelo. 1988. "Production, growth and business cycles:
I. The basic neoclassical model." Journal of monetary Economics, 21(2), 195–232.
Kiyotaki, Nobuhiro, Alexander Michaelides, and Kalin Nikolov. 2011. “Winners and Losers in Housing
Markets.” Journal of Money, Credit and Banking, 43(2-3), 255–296.
Kiyotaki, Nobuhiro, and John Moore. 1997. "Credit chains." Journal of Political Economy, 105(21), 211–
248.
Krugman, Paul. 2007. “Will There be a Dollar Crisis?” Economic Policy, 22(51), 435–467.
Krusell, Per, and Anthony A. Smith, Jr. 1998. "Income and wealth heterogeneity in the macroeconomy."
Journal of Political Economy, 106(5), 867–896.
McCarthy, Jonathan, and Richard W. Peach. 2004. “Are Home Prices the Next ’Bubble’ ?”
FRBNY
Economic Policy Review, 10(3).
Mian, Atif, and Amir Sufi. 2009. “The Consequences of Mortgage Credit Expansion: Evidence from the
U.S. Mortgage Default Crisis.” Quarterly Journal of Economics, 124(4), 1449–1496.
THE EFFECTS OF THE SAVING AND BANKING GLUT ON THE U.S. ECONOMY
39
Mendoza, Enrique G., Vincenzo Quadrini, and José-Victor Ríos-Rull. 2009. "On the welfare implications of
financial globalization without financial development." NBER International Seminar on Macroeconomics
2007, University of Chicago Press, 283–312.
Nakamura, Emi, and Jón Steinsson. "Five facts about prices: A reevaluation of menu cost models." The
Quarterly Journal of Economics, 123(4), 1415–1464.
Neumeyer, Pablo A., and Fabrizio Perri. 2005. “Business cycles in emerging economies: the role of interest
rates.” Journal of Monetary Economics, 52 (2), 345-380.
Obstfeld, Maurice, and Kenneth Rogoff. 2007. “The unsustainable US current account position revisited.”
G7 current account imbalances: Sustainability and adjustment. University of Chicago Press, 339–376.
Obstfeld, Maurice, and Kenneth Rogoff. 2009. “Global Imbalances and the Financial Crisis: Products or
Common Causes.” CEPR Working Paper DP7606.
Obstfeld, Maurice. 2012. “Financial Flows, Financial Crises, and Global Imbalances.” Journal of International Money and Finance, 31(3), 469–480.
Pozsar, Zoltan, Tobias Adrian, Adam Ashcraft, and Hayley Boesky. 2010. “Shadow Banking.” Federal
Reserve Bank of New York Staff Reports, 458.
Roubini, Nouriel, and Brad Setser. 2005. “Will the Bretton Woods 2 Regime Unravel Soon? The Risk of
a Hard Landing in 2005–-2006.” Manuscript, New York University and Oxford University.
Shiller, Robert J. 2007. “Understanding recent trends in house prices and homeownership.” Proceedings,
Federal Reserve Bank of Kansas City: 89–123.
Shin, Hyun Song. 2012. “Global Banking Glut and Loan Risk Premium.” IMF Economic Review, 60(2),
155–192.
Smets, Frank, and Rafael Wouters. 2007. “Shocks and Frictions in US Business Cycles: A Bayesian DSGE
Approach.” American Economic Review, 97(3), 586–606.
Taylor, John B. 2008. “The financial crisis and the policy responses: An empirical analysis of what went
wrong.” A Festschrift in Honour of David Dodge’s Contributions to Canadian Public Policy, Bank of
Canada, 1-18.
Tille, Cédric, and Eric van Wincoop. 2010. "International capital flows." Journal of International Economics, 80(2), 157–175.
Warnock, Frank, and Veronica Warnock. 2009. “International Capital Flows and U.S. Interest Rates.”
Journal of International Money and Finance, 28, 903–919.
THE EFFECTS OF THE SAVING AND BANKING GLUT ON THE U.S. ECONOMY
Federal Reserve Bank of Chicago
E-mail address: ajustiniano@frbchi.org
Northwestern University, CEPR, and NBER
E-mail address: g-primiceri@northwestern.edu
Federal Reserve Bank of New York
E-mail address: a.tambalotti@gmail.com
40
Working Paper Series
A series of research studies on regional economic issues relating to the Seventh Federal
Reserve District, and on financial and economic topics.
Comment on “Letting Different Views about Business Cycles Compete”
Jonas D.M. Fisher
WP-10-01
Macroeconomic Implications of Agglomeration
Morris A. Davis, Jonas D.M. Fisher and Toni M. Whited
WP-10-02
Accounting for non-annuitization
Svetlana Pashchenko
WP-10-03
Robustness and Macroeconomic Policy
Gadi Barlevy
WP-10-04
Benefits of Relationship Banking: Evidence from Consumer Credit Markets
Sumit Agarwal, Souphala Chomsisengphet, Chunlin Liu, and Nicholas S. Souleles
WP-10-05
The Effect of Sales Tax Holidays on Household Consumption Patterns
Nathan Marwell and Leslie McGranahan
WP-10-06
Gathering Insights on the Forest from the Trees: A New Metric for Financial Conditions
Scott Brave and R. Andrew Butters
WP-10-07
Identification of Models of the Labor Market
Eric French and Christopher Taber
WP-10-08
Public Pensions and Labor Supply Over the Life Cycle
Eric French and John Jones
WP-10-09
Explaining Asset Pricing Puzzles Associated with the 1987 Market Crash
Luca Benzoni, Pierre Collin-Dufresne, and Robert S. Goldstein
WP-10-10
Prenatal Sex Selection and Girls’ Well‐Being: Evidence from India
Luojia Hu and Analía Schlosser
WP-10-11
Mortgage Choices and Housing Speculation
Gadi Barlevy and Jonas D.M. Fisher
WP-10-12
Did Adhering to the Gold Standard Reduce the Cost of Capital?
Ron Alquist and Benjamin Chabot
WP-10-13
Introduction to the Macroeconomic Dynamics:
Special issues on money, credit, and liquidity
Ed Nosal, Christopher Waller, and Randall Wright
WP-10-14
Summer Workshop on Money, Banking, Payments and Finance: An Overview
Ed Nosal and Randall Wright
WP-10-15
Cognitive Abilities and Household Financial Decision Making
Sumit Agarwal and Bhashkar Mazumder
WP-10-16
1
Working Paper Series (continued)
Complex Mortgages
Gene Amromin, Jennifer Huang, Clemens Sialm, and Edward Zhong
WP-10-17
The Role of Housing in Labor Reallocation
Morris Davis, Jonas Fisher, and Marcelo Veracierto
WP-10-18
Why Do Banks Reward their Customers to Use their Credit Cards?
Sumit Agarwal, Sujit Chakravorti, and Anna Lunn
WP-10-19
The impact of the originate-to-distribute model on banks
before and during the financial crisis
Richard J. Rosen
WP-10-20
Simple Markov-Perfect Industry Dynamics
Jaap H. Abbring, Jeffrey R. Campbell, and Nan Yang
WP-10-21
Commodity Money with Frequent Search
Ezra Oberfield and Nicholas Trachter
WP-10-22
Corporate Average Fuel Economy Standards and the Market for New Vehicles
Thomas Klier and Joshua Linn
WP-11-01
The Role of Securitization in Mortgage Renegotiation
Sumit Agarwal, Gene Amromin, Itzhak Ben-David, Souphala Chomsisengphet,
and Douglas D. Evanoff
WP-11-02
Market-Based Loss Mitigation Practices for Troubled Mortgages
Following the Financial Crisis
Sumit Agarwal, Gene Amromin, Itzhak Ben-David, Souphala Chomsisengphet,
and Douglas D. Evanoff
WP-11-03
Federal Reserve Policies and Financial Market Conditions During the Crisis
Scott A. Brave and Hesna Genay
WP-11-04
The Financial Labor Supply Accelerator
Jeffrey R. Campbell and Zvi Hercowitz
WP-11-05
Survival and long-run dynamics with heterogeneous beliefs under recursive preferences
Jaroslav Borovička
WP-11-06
A Leverage-based Model of Speculative Bubbles (Revised)
Gadi Barlevy
WP-11-07
Estimation of Panel Data Regression Models with Two-Sided Censoring or Truncation
Sule Alan, Bo E. Honoré, Luojia Hu, and Søren Leth–Petersen
WP-11-08
Fertility Transitions Along the Extensive and Intensive Margins
Daniel Aaronson, Fabian Lange, and Bhashkar Mazumder
WP-11-09
Black-White Differences in Intergenerational Economic Mobility in the US
Bhashkar Mazumder
WP-11-10
2
Working Paper Series (continued)
Can Standard Preferences Explain the Prices of Out-of-the-Money S&P 500 Put Options?
Luca Benzoni, Pierre Collin-Dufresne, and Robert S. Goldstein
Business Networks, Production Chains, and Productivity:
A Theory of Input-Output Architecture
Ezra Oberfield
WP-11-11
WP-11-12
Equilibrium Bank Runs Revisited
Ed Nosal
WP-11-13
Are Covered Bonds a Substitute for Mortgage-Backed Securities?
Santiago Carbó-Valverde, Richard J. Rosen, and Francisco Rodríguez-Fernández
WP-11-14
The Cost of Banking Panics in an Age before “Too Big to Fail”
Benjamin Chabot
WP-11-15
Import Protection, Business Cycles, and Exchange Rates:
Evidence from the Great Recession
Chad P. Bown and Meredith A. Crowley
WP-11-16
Examining Macroeconomic Models through the Lens of Asset Pricing
Jaroslav Borovička and Lars Peter Hansen
WP-12-01
The Chicago Fed DSGE Model
Scott A. Brave, Jeffrey R. Campbell, Jonas D.M. Fisher, and Alejandro Justiniano
WP-12-02
Macroeconomic Effects of Federal Reserve Forward Guidance
Jeffrey R. Campbell, Charles L. Evans, Jonas D.M. Fisher, and Alejandro Justiniano
WP-12-03
Modeling Credit Contagion via the Updating of Fragile Beliefs
Luca Benzoni, Pierre Collin-Dufresne, Robert S. Goldstein, and Jean Helwege
WP-12-04
Signaling Effects of Monetary Policy
Leonardo Melosi
WP-12-05
Empirical Research on Sovereign Debt and Default
Michael Tomz and Mark L. J. Wright
WP-12-06
Credit Risk and Disaster Risk
François Gourio
WP-12-07
From the Horse’s Mouth: How do Investor Expectations of Risk and Return
Vary with Economic Conditions?
Gene Amromin and Steven A. Sharpe
WP-12-08
Using Vehicle Taxes To Reduce Carbon Dioxide Emissions Rates of
New Passenger Vehicles: Evidence from France, Germany, and Sweden
Thomas Klier and Joshua Linn
WP-12-09
Spending Responses to State Sales Tax Holidays
Sumit Agarwal and Leslie McGranahan
WP-12-10
3
Working Paper Series (continued)
Micro Data and Macro Technology
Ezra Oberfield and Devesh Raval
WP-12-11
The Effect of Disability Insurance Receipt on Labor Supply: A Dynamic Analysis
Eric French and Jae Song
WP-12-12
Medicaid Insurance in Old Age
Mariacristina De Nardi, Eric French, and John Bailey Jones
WP-12-13
Fetal Origins and Parental Responses
Douglas Almond and Bhashkar Mazumder
WP-12-14
Repos, Fire Sales, and Bankruptcy Policy
Gaetano Antinolfi, Francesca Carapella, Charles Kahn, Antoine Martin,
David Mills, and Ed Nosal
WP-12-15
Speculative Runs on Interest Rate Pegs
The Frictionless Case
Marco Bassetto and Christopher Phelan
WP-12-16
Institutions, the Cost of Capital, and Long-Run Economic Growth:
Evidence from the 19th Century Capital Market
Ron Alquist and Ben Chabot
WP-12-17
Emerging Economies, Trade Policy, and Macroeconomic Shocks
Chad P. Bown and Meredith A. Crowley
WP-12-18
The Urban Density Premium across Establishments
R. Jason Faberman and Matthew Freedman
WP-13-01
Why Do Borrowers Make Mortgage Refinancing Mistakes?
Sumit Agarwal, Richard J. Rosen, and Vincent Yao
WP-13-02
Bank Panics, Government Guarantees, and the Long-Run Size of the Financial Sector:
Evidence from Free-Banking America
Benjamin Chabot and Charles C. Moul
WP-13-03
Fiscal Consequences of Paying Interest on Reserves
Marco Bassetto and Todd Messer
WP-13-04
Properties of the Vacancy Statistic in the Discrete Circle Covering Problem
Gadi Barlevy and H. N. Nagaraja
WP-13-05
Credit Crunches and Credit Allocation in a Model of Entrepreneurship
Marco Bassetto, Marco Cagetti, and Mariacristina De Nardi
WP-13-06
4
Working Paper Series (continued)
Financial Incentives and Educational Investment:
The Impact of Performance-Based Scholarships on Student Time Use
Lisa Barrow and Cecilia Elena Rouse
WP-13-07
The Global Welfare Impact of China: Trade Integration and Technological Change
Julian di Giovanni, Andrei A. Levchenko, and Jing Zhang
WP-13-08
Structural Change in an Open Economy
Timothy Uy, Kei-Mu Yi, and Jing Zhang
WP-13-09
The Global Labor Market Impact of Emerging Giants: a Quantitative Assessment
Andrei A. Levchenko and Jing Zhang
WP-13-10
Size-Dependent Regulations, Firm Size Distribution, and Reallocation
François Gourio and Nicolas Roys
WP-13-11
Modeling the Evolution of Expectations and Uncertainty in General Equilibrium
Francesco Bianchi and Leonardo Melosi
WP-13-12
Rushing into American Dream? House Prices, Timing of Homeownership,
and Adjustment of Consumer Credit
Sumit Agarwal, Luojia Hu, and Xing Huang
WP-13-13
The Earned Income Tax Credit and Food Consumption Patterns
Leslie McGranahan and Diane W. Schanzenbach
WP-13-14
Agglomeration in the European automobile supplier industry
Thomas Klier and Dan McMillen
WP-13-15
Human Capital and Long-Run Labor Income Risk
Luca Benzoni and Olena Chyruk
WP-13-16
The Effects of the Saving and Banking Glut on the U.S. Economy
Alejandro Justiniano, Giorgio E. Primiceri, and Andrea Tambalotti
WP-13-17
5
@FedFRASER