Sonderforschungsbereich/Transregio 15 · www.sfbtr15.de
Universität Mannheim · Freie Universität Berlin · Humboldt-Universität zu Berlin · Ludwig-Maximilians-Universität München
Rheinische Friedrich-Wilhelms-Universität Bonn · Zentrum für Europäische Wirtschaftsforschung Mannheim
Speaker: Prof. Dr. Klaus M. Schmidt · Department of Economics · University of Munich · D-80539 Munich,
Phone: +49(89)2180 2250 · Fax: +49(89)2180 3510
* / ** Federal Reserve Bank of New York
*** Universität Mannheim
Financial support from the Deutsche Forschungsgemeinschaft through SFB/TR 15 is gratefully acknowledged.
Discussion Paper No. 449
The Fragility of Short-Term
Secured Funding Markets
Antoine Martin *
David Skeie **
Ernst-Ludwig von Thadden ***
The Fragility of Short-Term Secured Funding M ark ets
Antoin e Ma rtin Dav id Skeie
Ernst-Ludwig v on Thadden
1
Journal of Ec o nomic Theory,2014
Abstract
This paper dev elops an infinite-horizon model of financial institutions
that borro w shor t-term an d in vest in long-term a ssets that can be traded in
frictionless markets. B ecause these financial in term ediaries perform maturit y
transform ation , they are su bject to poten tia l ru ns. We derive distinct liq-
uidity, collatera l, and asset liquidation constraints, which determine w hether
a run can occur as a result of chang ing mark et expectations. We sho w that
theextenttowhichborrowerscanwardoff an individual run depends on
whether it has sufficient liquidity, collateral, and asset liquidation capacit y.
These determinants depend on the borro wer’s (endogen ous) balance sheet
and on (exogenous) fundamentals. Systemic runs are possible if shocks to
the valuation of collateral held b y outside inv estors are sufficien tly strong
and uniform, and if the system as a whole is exposed to high short-term
funding risk. The theory has policy implications for prudential regulation
and lender-of-last-resort in terventions.
Keyw ords: Investment banking, securities dealers, repurc h ase agreemen ts,
runs, financia l fragility, collater al, systemic risk.
JEL classification: E44, E58, G24
1
Martin and Sk eie are at the Federal Reserve Bank of New York. Von Thad-
den is at the University of Mannheim. Author e-mails are antoine.martin@ny.frb.org,
david.skeie@ny.frb.org, and vthadden@uni-mannheim.de, respectively. We thank Viral
Acharya, Sudipto Bhattacharya, Amar Bhide, Patric k Bolton, Douglas Gale, Gary Gor-
ton, Yaron Leitner, Ed Nosal, Lasse Pedersen, Matt Pritzker, Bogdan Stacescu, Dimitri
Vay anos, an anonymous referee, as well as the participants of the Gerzensee Summer
Symposium 2011, for helpful comments, and Ali Palida for very good research assistance.
Part of this research was done while von Thadden w as at the Finance Department of the
London School of Economics. The views expressed in this paper are those of the authors
and do not necessarily reflect the views of the Federal Reserve Bank of New York or the
Federal Reserve System.
1 Introduction
This paper develops a model of financial institutions funded by short-term
borrow ing and investing in long-term marketable assets. We sho w that suc h
institutions are subject to the threat of runs similar to those faced b y com-
mercial banks and study the conditions under which runs can occur. The
analysis derives liquidity, collateral, and asset liquidation constraints for such
institutions that depend on whether the run is on an individual institution
only or is systemic. W hen these constraints are violated, runs can occur.
Institutions threatened by a run can liquidate assets through market sales
to raise cash or increase the collateral they offer to attract emergency lending.
For both these reactions, the institution’s asset base com p ared to its borro w -
ing exposure is decisiv e. In our steady-state model, these two k ey variables
are endogen ou s, but not pinned do w n uniquely. The ability to surviv e a run
depends on these t wo variables, as well as on exogenous param eters suc h as
the institution’s profitabilit y and size.
Our work builds on the theory of comm ercial bank instability developed
b y Diamond and Dyb vig (1983), Qi (1994), and others. As pointed out
b y Gorton and Metrick (2012), there are important similarities betw een the
fragilit y of comm ercial banks that borrow unsecured deposits and hold non-
mar ketable loan portfolios, and of “securitized” or “shadow” bank s, whic h
borro w in repo or other short-term funding mark ets against marketable secu-
rities as collateral.
2
In particular, repo markets perform maturit y transforma-
tion b y allowing in vestors with uncert ain liquidity needs to lend short-term
against longer term , less liquid securities. We pro v id e a formal model of
shadow banking to iden tify the dete rm inants of equilibrium profits, liquidit y,
collateral, and asset market prices that support such m atu rity transform ation
during normal times, and examine its fragility.
This paper uses the model dev eloped in Martin, Skeie, and von Thadden
(2014), whic h focuses on the market microstructure of short-term funding
mar kets and comp ares the impact of differentmarketstructuresonthepos-
sibilit y of runs, but ignores asset markets. Th e present paper simplifies the
2
See P ozsar, Adrian, Ashcraft, and Boesky (2010) for a detailed discussion of the role
of shadow banking in the recent financial crisis.
1
microstru cture, but in troduces asset market activity and analyzes its impact
on market fragility. The inte rdependency bet ween the asset side and liability
side of a borrower’s balance sheet determines the borro wer’s fragilit y and in
aggregate determines market fragilit y.
In cont rast to Diamond and Dybvig (1983), we study an infinite-horizon
model. A k ey benefitisthatprofits are endogenous, so that w e can make
predictions about how the model’s structural parameters affect the stability
of the steady state via the endogenously generated liquidit y, rather than
performing comparativ e statics with respect to exogenous liquidity lev els. Qi
(1994) also considers an infinite-horizon model, but his financial institutions
are assumed to mak e zero profits.
3
In fact, we show that competition does
not necessarily driv e up interest rates to zero-profit lev els in equilibrium,
because borrowers with liquidity of their own mu st have an incentive to
borrow rather than using their ow n funds for investment. Since investing
o wn funds is profitable, so must be borro w ing. This equilibrium argume nt
for positive profits has been developed in Martin, Skeie, and von Thadden
(2014) and relies on a tradeoff between the use of external and internal funds,
whic h are endogenous in our infinite-horizon model, but would need to be
exogenously specified in a static model.
Afurtherkeydifference to Diamond-Dy bvig (1983) is the existence of
asset markets in our model. This is a feature that our model shares with
Diam on d (1997), who introduces an asset market with limited participation
in to the Diam on d-D y bvig model and sho w s that banking coexists with mar-
ket activity in equilibrium and show s how both types of activities influence
eac h other. In our model, as in several others,
4
asset market participation
is restricted to soph isticated borro wers (the “banks" in the Diamon d-D ybvig
world), which rules out the t ype of coexistence problem studied by Diamond
(1997). Differently from that paper, our focus is on the role of asset markets
in runs.
In this respect our w ork is similar to the sm all literature, led by Allen and
3
Other recent infinite-horizon models of banking instability such as He and Xiong (2012)
or Segura and Suarez (2012) also generate positive equilibrium profits, but do not consider
their interaction with asset markets and fire sales, which is the focus of the present paper.
4
Such as Acharya, Shin, and Yorulmazer (2011) or Allen and Gale (2004a).
2
Gale (2004a, 2004b), that in troduces Walrasian mar kets into the Diamon d-
Dybvig m odel.
5
Allen and Gale (2004a, 2004b) note that the liquidity ob-
tained from asset markets (or equivalen tly, perfect future ma rkets for state-
con tingent consumption) complements that obtained through in termediaries
and study its limits because of insufficient cash in the market or market in-
completen ess. Differently from that literature, we consider the use of assets
as collateral to support borrow ing. Muc h of our analysis is concerned with
expectation-based runs on individual borrowers, and our question is to what
exten t asset markets can overcome such breakdowns of short-term borro w ing
mar kets. A w ell-functio ning asset market makes it possible for a distressed
borrower to sell assets and th us obtain emergen cy liquidity. Ho wever, the
existence of these markets also in good times makes it possible for borrow ers
to bring consumption forward in time. This t ype of activit y corresponds to
securitization: the steady-state sale of assets by a finan cial institution that
could otherwise be kept on balance sheet, motivated b y the desire to gen-
erate cash before the assets mature.
6
An important result of our paper is
that securitization (net asset sales) weaken s a borro wer’s balance sheet be-
cause it reduces the assets a vailable to raise cash in case of emergency, either
through liquidation or additional collateral. The model therefore predicts
that borro wers that securitize less will be less fragile.
This view on securitization resem bles P arlour and Plantin’s (2007) argu-
ment that liquid asset markets ma y not be socially optimal, because they
facilitate securitization without pro viding incentives to monitor the qualit y
of the assets sold. In our model, liquid asset markets are beneficial because
they can pro vide liquidity when individual borrowers are under stress, but
they can create fragility if their use erodes the borrower’s asset base too
much.
The liquidit y of a borrower’s balance sheet pla ys a key role in our model
5
An excellent introduction into this literature is the broad survey by Allen and Gale
(2007).
6
Securitization typically involves both the pooling and tranching of assets and the sale
of the resulting securities. In this paper, we focus on the latter, which has been a major
factor in the financial crisis of 2007, as argued forcefully by Acharya, Schnabl, and Suarez
(2013). We abstract from issues such as pooling and tranching (see DeMarzo, 2005, Martin
and Parigi, 2013) and consider only homogenous assets.
3
