Sunteți pe pagina 1din 19

American Economic Association

Credit Rationing in Markets with Imperfect Information


Author(s): Joseph E. Stiglitz and Andrew Weiss
Source: The American Economic Review, Vol. 71, No. 3 (Jun., 1981), pp. 393-410
Published by: American Economic Association
Stable URL: http://www.jstor.org/stable/1802787 .
Accessed: 24/01/2011 13:56

Your use of the JSTOR archive indicates your acceptance of JSTOR's Terms and Conditions of Use, available at .
http://www.jstor.org/page/info/about/policies/terms.jsp. JSTOR's Terms and Conditions of Use provides, in part, that unless
you have obtained prior permission, you may not download an entire issue of a journal or multiple copies of articles, and you
may use content in the JSTOR archive only for your personal, non-commercial use.

Please contact the publisher regarding any further use of this work. Publisher contact information may be obtained at .
http://www.jstor.org/action/showPublisher?publisherCode=aea. .

Each copy of any part of a JSTOR transmission must contain the same copyright notice that appears on the screen or printed
page of such transmission.

JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of
content in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new forms
of scholarship. For more information about JSTOR, please contact support@jstor.org.

American Economic Association is collaborating with JSTOR to digitize, preserve and extend access to The
American Economic Review.

http://www.jstor.org
Credit Rationing in Marketswith
ImperfectInformation
By JOSEPH E. STIGLITZ AND ANDREW WEISS*

Why is credit rationed? Perhaps the most they receive on the loan, and the riskiness of
basic tenet of economics is that market equi- the loan. However, the interest rate a bank
librium entails supply equalling demand; that charges may itself affect the riskiness of the
if demand should exceed supply, prices will pool of loans by either: 1) sorting potential
rise, decreasing demand and/or increasing borrowers (the adverse selection effect); or 2)
supply until demand and supply are equated affecting the actions of borrowers (the incen-
at the new equilibrium price. So if prices do tive effect). Both effects derive directly from
their job, rationing should not exist. How- the residual imperfect information which is
ever, credit rationing and unemployment do present in loan markets after banks have
in fact exist. They seem to imply an excess evaluated loan applications. When the price
demand for loanable funds or an excess (interest rate) affects the nature of the trans-
supply of workers. action, it may not also clear the market.
One method of "explaining" these condi- The adverse selection aspect of interest
tions associates them with short- or long-term rates is a consequence of different borrowers
disequilibrium. In the short term they are having different probabilities of repaying
viewed as temporary disequilibriumphenom- their loan. The expected return to the bank
ena; that is, the economy has incurred an obviously depends on the probability of re-
exogenous shock, and for reasons not fully payment, so the bank would like to be able
explained, there is some stickiness in the to identify borrowers who are more likely to
prices of labor or capital (wages and interest repay. It is difficult to identify "good bor-
rates) so that there is a transitional period rowers," and to do so requires the bank to
during which rationing of jobs or credit oc- use a variety of screening devices. The inter-
curs. On the other hand, long-term un- est rate which an individual is willing to pay
employment (above some "natural rate") or may act as one such screening device: those
credit rationing is explained by governmen- who are willing to pay high interest rates
tal constraints such as usury laws or mini- may, on average, be worse risks; they are
mum wage legislation.' willing to borrow at high interest rates be-
The object of this paper is to show that cause they perceive their probability of re-
in equilibrium a loan market may be char- paying the loan to be low. As the interest
acterized by credit rationing. Banks making rate rises, the average "riskiness" of those
loans are concerned about the interest rate who borrow increases, possibly lowering the
bank's profits.
*Bell Telephone Laboratories, Inc. and Princeton Similarly, as the interest rate and other
University, and Bell Laboratories, Inc., respectively. We terms of the contract change, the behavior of
would like to thank Bruce Greenwald, Henry Landau, the borrower is likely to change. For in-
Rob Porter, and Andy Postlewaite for fruitful comments stance, raising the interest rate decreases the
and suggestions. Financial support from the National return on projects which succeed. We will
Science Foundation is gratefully acknowledged. An
earlier version of this paper was presented at the spring show that higher interest rates induce firms
1977 meetings of the Mathematics in the Social Sciences to undertake projects with lower probabili-
Board in Squam Lake, New Hampshire. ties of success but higher payoffs when suc-
'Indeed, even if markets were not competitive one cessful.
would not expect to find rationing; profit maximization
would, for instance, lead a monopolistic bank to raise
In a world with perfect and costless infor-
the interest rate it charges on loans to the point where mation, the bank would stipulate precisely
excess demand for loans was eliminated. all the actions which the borrower could
393
394 THE AMERICAN ECONOMIC REVIEW JUNE 1981

z no competitiveforcesleadingsupply to equal
m demand, and creditis rationed.
w
But the interestrate is not the only termof
1Z the contractwhichis important.The amount
0
2I-
a-
/ of the loan, and the amount of collateral or
equity the bank demandsof loan applicants,
w will also affect both the behavior of bor-
@
-
w rowers and the distributionof borrowers.In
Section III, we show that increasing the col-
INTEREST RATE
lateralrequirementsof lenders(beyond some
w r
point) may decreasethe returnsto the bank,
by either decreasingthe average degree of
FIGURE 1. THERE EXISTSAN INTERESTRATE WHICH
MAXIMIZESTHE EXPECTEDRETURN TO THE BANK
risk aversionof the pool of borrowers;or in
a multiperiodmodel inducing individual in-
vestors to undertakeriskierprojects.
Consequently,it may not be profitable to
raise the interest rate or collateral require-
undertake(which might affect the returnto ments when a bank has an excess demand
the loan). However, the bank is not able to for credit; instead, banks deny loans to bor-
directly control all the actions of the bor- rowers who are observationally indis-
rower; therefore,it will formulatethe terms tinguishablefromthose who receive loans.2
of the loan contractin a mannerdesignedto It is not our argumentthat credit rationing
induce the borrower to take actions which will always characterizecapital markets, but
are in the interest of the bank, as well as to rather that it may occur under not implausi-
attractlow-riskborrowers. ble assumptions concerning borrower and
For both these reasons, the expected re- lender behavior.
turn by the bank may increase less rapidly This paper thus providesthe first theoret-
than the interest rate; and, beyond a point, ical justificationof true credit rationing. Pre-
may actuallydecrease,as depictedin Figure vious studies have sought to explain why
1. The interest rate at which the expected each individualfaces an upward sloping in-
returnto the bank is maximized,we referto terest rate schedule.The explanationsoffered
as the "bank-optimal"rate, Pr. are (a) the probability of default for any
Both the demandfor loans and the supply particularborrowerincreasesas the amount
of funds are functions of the interest rate borrowed increases(see Stiglitz 1970, 1972;
(the latterbeing determinedby the expected Marshall Freimer and Myron Gordon;
return at r*). Clearly, it is conceivable that at Dwight Jaffee; George Stigler), or (b) the
r the demandfor funds exceeds the supply mix of borrowers changes adversely (see
of funds. Traditional analysis would argue Jaffee and ThomasRussell).In these circum-
that, in the presenceof an excess demandfor stances we would not expect loans of differ-
loans, unsatisfied borrowerswould offer to ent size to pay the same interest rate, any
pay a higher interest rate to the bank, bid- more than we would expect two borrowers,
ding up the interestrate until demandequals one of whom has a reputationfor prudence
supply. But although supply does not equal and the other a reputationas a bad credit
demand at r*, it is the equilibriuminterest risk, to be able to borrowat the same interest
rate!The bank would not lend to an individ- rate.
ual who offered to pay more than r*. In the We reserve the term credit rationing for
bank'sjudgment,such a loan is likely to be a circumstancesin whicheither(a) among loan
worse risk than the averageloan at interest applicants who appearto be identical some
rate P*, and the expected return to a loan at
an interest rate above r* is actually lower 2After this paper was completed,our attention was
than the expected return to the loans the drawnto W. Keeton'sbook.In chapter3 he developsan
bank is presently making. Hence, there are incentiveargumentfor creditrationing.
VOL. 71 NO. 3 STIGLITZAND WEISS:CREDITRATIONING 395

receive a loan and others do not, and the of projects; for each project 6 there is a
rejectedapplicants would not receive a loan probabilitydistributionof (gross) returnsR.
even if they offered to pay a higher interest We assume for the moment that this distri-
rate; or (b) there are identifiablegroups of bution cannot be alteredby the borrower.
individuals in the population who, with a Different firms have different probability
given supply of credit, are unable to obtain distributionsof returns.We initially assume
loans at any interestrate,even thoughwith a that-the bank is able to distinguishprojects
largersupply of credit, they would.3 with different mean returns, so we will at
In our construction of an equilibrium first confine ourselves to the decision prob-
model with credit rationing, we describe a lem of a bank facing projects having the
marketequilibriumin which there are many same mean return. However, the bank can-
banks and many potential borrowers.Both not ascertainthe riskiness of a project. For
borrowersand banks seek to maximizeprof- simplicity, we write the distribution of re-
its, the former through their choice of a turns4as F(R, 0) and the density functionas
project, the latter through the interest rate f(R, 0), and we assume that greater6 corre-
they chargeborrowersand the collateralthey sponds to greaterrisk in the sense of mean
require of borrowers (the interest rate re- preservingspreads5(see Rothschild-Stiglitz),
ceived by depositors is determinedby the i.e., for , >2,Jif
zero-profitcondition). Obviously,we are not
discussing a "price-taking"equilibrium.Our 00 0

(1) fRf(R, 01) dR= Rf(R, 2) dR


equilibrium notion is competitive in that
banks compete; one means by which they
compete is by their choice of a price(interest then for y O,
rate) which maximizes their profits. The
reader should notice that in the model pre- (2) j F(R,01)dR> jF(R,02)dR
sented below there are interestrates at which
the demand for loanable funds equals the
supply of loanable funds. However,these are If the individualborrowsthe amount B, and
not, in general,equilibriuminterestrates. If, the interstrate is r, then we say the individ-
at those interest rates, banks could increase ual defaults on his loan if the returnR plus
their profits by lowering the interest rate the collateral C is insufficient to pay back
chargedborrowers,they would do so. the promisedamount,6i.e., if
Although these resultsare presentedin the
context of credit markets,we show in Section (3) C+R<B(I +P)
V that they are applicableto a wide class of
principal-agent problems (including those
describing the landlord-tenantor employer- 4These are subjectiveprobabilitydistributions;the
employee relationship). perceptionson the part of the bank may differ from
those of the firm.
I. InterestRate as a ScreeningDevice 5MichaelRothschild and Stiglitz show that condi-
tions (I) and (2) imply that project 2 has a greater
variancethan project 1, although the converse is not
In this section we focus on the role of true. That is, the mean preservingspread criterionfor
interest rates as screening devices for dis- measuringrisk is strongerthan the increasingvariance
tinguishingbetween good and bad risks. We criterion.They also show that (I) and (2) can be in-
terpretedequallywell as: given two projectswith equal
assume that the bank has identifieda group means,every risk averterprefersproject I to project2.
6This is not the only possible definition. A firm
might be said to be in default if R< B(1 + P). Nothing
3There is another form of rationingwhich is the critical depends on the precise definition. We assume,
subjectof our 1980 paper:banksmake the provisionof however, that if the firm defaults, the bank has first
credit in later periods contingent on performancein claim on R+ C. The analysismay easily be generalized
earlierperiod; banksmay then refuseto lend even when to include bankruptcycosts. However, to simplify the
theselater periodprojectsstochasticallydominateearlier analysis,we usuallyshall ignorethese costs. Throughout
projectswhich are financed. this sectionwe assumethat the projectis the sole project
396 THE A MERICAN ECONOMIC REVIEW JUNE 1981

Thusthe net returnto the borrower7T(R,r)


can be writtenas
(4a) 7(R, r) =max(R-(1 +r)B; -C)
(1+r)B-C / --~R
The returnto the bank can be writtenas
(4b) p(R,fr)=min(R+C; B(1+r)) -C

FIGURE 2a. FIRMPROFITSAREA CONVEX


that is, the borrowermust pay back either FIJNCTIONOF THE RETURN ON THE PROJECT
the promised amount or the maximum he
can pay back (R+ C).
For simplicity,we shall assume that the
borrowerhas a givenamountof equity(which
he cannot increase), that borrowers and
lenders are risk neutral, that the supply of
loanable funds availableto a bank is unaf- C
fected by the interest rate it charges bor-
rowers,that the cost of the projectis fixed,
R
and unless the individual can borrow the (1 + r) B-C
differencebetweenhis equity and the cost of
the project, the project will not be under- FIGURE 2b. THE RETURN TO THE BANK IS A CONCAVE
FUNCTION OF THE RETURN ON THE PROJECT
taken, that is, projectsare not divisible.For
notationalsimplicity,we assumethe amount
borrowed for each project is identical, so
that the distributionfunctionsdescribingthe The value of 0 for which expectedprofits
numberof loan applicationsare identicalto are zero satisfies
those describingthe monetaryvalue of loan
applications.(In a more general model, we (5) r(IA) E
would make the amount borrowedby each
individual a function of the terms of the f max[R-(r+ 1)B; -C] dF(R, ) 0
contract; the quality mix could change not
only as a result of a change in the mix of Our argumentthat the adverseselectionof
applicants,but also because of a change in interest rates could cause the returnsto the
the relative size of applicationsof different bank to decreasewith increasinginterestrates
groups.) hinged on the conjecturethat as the interest
We shall now prove that the interest rate rate increased,the mix of applicantsbecame
acts as a screeningdevice;more preciselywe worse; or
establish
THEOREM 2: As the interest rate increases,
THEOREM 1: For a given interest rate r, the critical value of 0, below which individuals
there is a critical value 0 such that a firm do not apply for loans, increases.
borrows from the bank if and only if 0>0.
This follows immediatelyupon differenti-
This follows immediatelyupon observing ating (5):
that profitsare a convex functionof R, as in
Figure 2a. Hence expected profits increase
with risk. BJ dF(R,O)
(6) do I1+rP)B- C
>0
dr ari/ao
undertakenby the firm (individual)and that there is
limited liability. The equilibriumextent of liability is
derivedin SectionIII. For each 0, expected profits are decreased;
VOL. 71 NO. 3 STIGLITZ AND WEISS: CREDITRATIONING 397

L~~~~~~~~~L

L
X LD
0
~ ~ irm '
TYPES
APPLY ONL
/ /HIG~H RISK '~ ~ ---------

/ / ~APPLY

FIGURE 4. DETERMINATIONOF THE MARKET


EQUILIBRIUM

rl ?
sive group drops -out of the market, there is a
FIGURE 3. OPTIMALINTERESTRATE r1 discrete fall in (where p(r) is the mean
return to the bank from the set of applicants at
hence using Theorem 1, the result is im- the interest rate r).
mediate.
We next show: Other conditions for nonmonotonicity of
p(r) will be established later. Theorems 5
THEOREM 3: The expected return on a loan and 6 show why nonmonotonicity is so im-
to a bank is a decreasing function of the portant:
riskiness of the loan.
THEOREM 5: Wheneverp(r) has an interior
PROOF: mode, there exist supply functions of funds
From (4b) we see that p(R, r) is a con- such that competitive equilibriumentails credit
cave function of R, hence the result is im- rationing.
mediate. The concavity of p(R, r) is il-
lustratedin Figure2b. This will be the case whenever the "Wal-
Theorems2 and 3 imply that, in addition rasian equilibrium" interest rate- the one at
to the usual direct effect of increasesin the which demand for funds equals supply-is
interestrate increasinga bank'sreturn,there such that there exists a lower interest rate for
is an indirect,adverse-selectioneffect acting which p, the return to the bank, is higher.
in the oppositedirection.We now show that In Figure 4 we illustrate a credit rationing
this adverse-selectioneffect may outweigh equilibrium. Because demand for funds de-
the directeffect. pends on r, the interest rate charged by
To see this most simply, assumethere are banks, while the supply of funds depends on
two groups; the "safe" group will borrow p, the mean return on loans, we cannot use a
only at interest rates below r,, the "risky" conventional demand/supply curve diagram.
group below r2, and r, <r2. When the inter- The demand for loans is a decreasing func-
est rate is raisedslightlyabove r,, the mix of tion of the interest rate charged borrowers;
applicantschangesdramatically:all low risk this relation LD is drawn in the upper right
applicantswithdraw.(See Figure 3.) By the quadrant. The nonmonotonic relation be-
sameargumentwe can establish tween the interest charged borrowers, and
the expected
- return to the bank per dollar
THEOREM 4: If there are a discrete number loaned is drawn in the lower right quadrant.
of potential borrowers(or types of borrowers) In the lower left- quadrant we depict the
each with a different 0, p(r) will not be a relation between and the supply of loana-
monotonicfunction of r, since as each succes- ble funds LS. (We have drawn LS as if it
398 THE A MERICAN ECONOMIC RE VIEW JUNE 1981

Figure 5 illustrates a p(r) function with


multiple modes. The nature of the equi-
libriumfor such cases is describedby Theo-
rem 6.
THEOREM 6: If the -p(r)function has several
modes, market equilibrium could either be
characterized by a single interest rate at or
I I I
I I I below the market-clearing level, or by two
interest rates, with an excess demandfor credit
r,, rm r2 r
at the lower one.
FIGURE 5. A TWO-INTERESTRATE EQUILIBRIUM
PROOF:
Denote the lowest Walrasianequilibrium
were an increasingfunctionof p. This is not interestrate by rmand denote by r the inter-
necessaryfor our analysis.)If banks
- are free est rate which maximizes p(r). If r<rm, the
to competefor depositors,then will be the analysis for Theorem5 is unaffectedby the
interest rate received by depositors. In the multiplicityof modes. There will be credit
upper right quadrantwe plot LS as a func- rationing at interest rate r. The rationed
tion of r, through the impact of r on the borrowerswill not be able to obtain credit
returnon- each loan, and hence on the inter- by offeringto pay a higherinterestrate.
est rate banks can offer to attractloanable On the other hand, if r>rm, then loans
funds. may be made at two interestrates, denoted
A creditrationingequilibriumexists given by r, and r2. r, is the interest rate which
the relationsdrawnin Figure4; the demand maximizesp(r) conditional on r<rm; r2 is
for loanable funds at r* exceeds the supply the lowest interestrate greaterthan rmsuch
of loanable funds at r* and any individual that p(r2)=p(r,). From the definition of rm,
bank increasingits interest rate beyond r* and the downwardslope of the loan demand
would lowerits returnper dollarloaned.The function,therewill be an excess demandfor
excess demand for funds is measuredby Z. loanable funds at r, (unless r, =rm, in which
Notice that there is an interest rate rm at case there is no credit rationing).Some re-
which the demandfor loanablefunds equals jected borrowers(with reservationinterest
the supply of loanablefunds; however,rmis rates greaterthan or equal to r2) will apply
not an equilibriuminterestrate.A bankcould for loans at the higher interest rate. Since
increaseits profitsby chargingr* ratherthan there would be an excess supply of loanable
rm:at the lower interestrate it would attract funds at r2 if no loans were made at r,, and
at least all the borrowersit attractedat rm an aggregateexcess demandfor funds if no
and would make larger profits from each loans weremade at r2,thereexists a distribu-
loan (or dollarloaned). tion of loanablefunds availableto borrowers
Figure 4 can also be used to illustratean at r, and r2 such that all applicantswho are
important comparativestatics property of rejectedat interestrate r, and who apply for
our marketequilibrium: loans at r2will get credit at the higherinter-
est rate. Similarly,all the funds availableat
COROLLARY 1. As the supply of funds in- p(r,) will be loaned at eitherr, or r2. (There
creases, the excess demand for funds de- is, of course,an excess demandfor loanable
creases, but the interest rate charged remains funds at r, since everyborrowerwho eventu-
unchanged, so long as there is any credit ra- ally borrowsat r2 will have first applied for
tioning. credit at r,.) Thereis clearlyno incentivefor
small deviations from r1, which is a local
Eventually,of course,Z will be reducedto maximum of p(r). A bank lending at an
zero; furtherincreasesin the supplyof funds interestrate r3 such that p(r3)<p(r,) would
then reducethe marketrate of interest. not be able to obtain credit. Thus, no bank
VOL. 71 NO. 3 STIGLITZ AND WEISS: CREDIT RATIONING 399

would switch to a loan offer betweenr, and large if (g(0)/[l - G(O)]) (dO/d?) is large,
r2. A bank offering an interest rate r4 such that is, a small change in the nominal inter-
that p(r4)>p(r,) would not be able to at- est rate induces a large change in the appli-
tract any borrowerssince by definitionr4> cant pool.
r2, and thereis no excess demandat interest
rate r2. 2. Two Outcome Projects
Here we consider the simplest kinds of
A. Alternative Sufficient Conditionsfor projects(from an analyticalpoint of view),
Credit Rationing those whicheithersucceedand yield a return
R, or fail and yield a returnD. We normalize
Theorem4 provideda sufficientcondition to let B= 1. All the projectshave the same
for adverseselection to lead to a nonmono- unsuccessfulvalue (whichcould be the value
tonic -p(r)function.In the remainderof this of the plant and equipment)while R ranges
section, we investigate other circumstances between S and K (where K> S). We also
under which for some levels of supply of assume that projectshave been screenedso
funds therewill be creditrationing. that all projectswithin a loan categoryhave
the same expected yield, T, and there is no
1. Continuumof Projects collateralrequired,that is, C= 0, and if p( R)
Let G(O)be the distributionof projectsby representsthe probabilitythat a projectwith
riskiness0, and p(O,r) be the expected re- a successfulreturnof R succeeds,then
turn to the bank of a loan of risk 0 and
interestrate r. The mean returnto the bank (9) p(R)R+ T
[1-p(R)]D=
whichlends at the interestrate r is simply
P00 In addition, the bank suffers a cost of X
P(, r) dG(O) per dollar loaned upon loans that default,
(7) (r) -G(= which could be interpretedas the difference
) betweenthe value of plant and equipmentto
the firm and the value of the plant and
FromTheorem5 we know that dp(rP)/dP<O equipmentto the bank. Again the density of
for some value of r is a sufficientcondition projectvalues is denotedby g(R), the distri-
for creditrationing.Let p(6, r)= p so that bution functionby G(R).
Therefore,the expected returnper dollar
dp lent at an interestrate r, if we let J=r+ 1, is
g(6) dO (since individualswill borrowif and only if
dr [1- G(6 ] dP R >J):
(10)
| [- F((l + r)B - C, )] dG(O)
+ r)
fKg()=A [ J Kp(R) g(R) dR
1-G(O )

From Theorems1 and 3, the first term is


negative(representingthe changein the mix + J [1-p(R)][D-X]g(R) dR]
of applicants),while the second term (the
increase in returns, holding the applicant
pool fixed, from raisingthe interestcharges) Using l'Hopital'srule and (1), we can estab-
is positive.The first termis large,in absolute lish sufficient conditions for 1imJ,K(ap(J)/
value, if there is a large differencebetween aJ)<O (and hence for the nonmonotonicity
the mean return on loans made at interest of p): 7
rate r and the return to the bank from the
project making zero returns to the firm at 7The proofs of these propositionsare slightly com-
interest rate r (its "safest"loan). It is also plicated. Consider 1. Since p(R)=T-D/R-D, the
400 THE A MERICAN ECONOMIC REVIEW JUNE 1981

(a) if fimR-Kg(R)=O0, so then a sufficient the bank-optimalinterestrate. High interest


condition is X> K- D, or equivalently, rates may make projectswith low mean re-
limR-Kp(R)+p'(R)X<0 turns- the projectsundertakenby riskaverse
(b) if g(K) = 0, g'(K) 7 0, so then a suffi- individuals-infeasible, but leave relatively
cient condition is 2X>K-D, or equiva- unaffected the risky projects.The mean re-
lently, IimR,Kp(R)+2p (R)X<O turn to the bank, however,is lower on the
(c) if g(K)=O, g'(K)=0, g"(K)57z0,then riskierprojectsthan on the safe projects.In
a sufficient condition is 3X>K-K-D, or the following example, it is systematic dif-
equivalently,limR,Kp(R) + 3p'(R)X< 0 ferences in risk aversion which results in
Condition (a) implies that if, as 1+ rP- K, there being an optimalinterestrate.
the probabilityof an increasein the interest Assume a fraction X of the populationis
rate being repaid is outweighed by the infinitely risk averse; each such individual
deadweightloss of riskier loans, the bank undertakes the best perfectly safe project
will maximizeits returnper dollar loaned at whichis availableto him. Withinthat group,
an interestrate below the maximumrate at the distributionof returns is G(R) where
which it can loan funds (K- 1). The condi- G(K)=1. The other group is risk neutral.
tions for an interior bank optimal interest For simplicitywe shall assume that they all
rate are significantly less stringent when face the same risky projectwith probability
g(K) = 0. of success p and a return, if successful,of
R*> K; if not their return is zero. Letting
3. Differences in Attitudes TowardsRisk R =(1 + r)B the (expected) return to the
Some loan applicantsare clearlymore risk bank is
aversethan others.These differenceswill be
reflectedin project choices, and thus affect (11)

p(r) -{ X(l-G(R^))+ (I -X)p } (I +r)


expectedprofitper dollarloanedmay be rewrittenas X(l1-G(RA))+(1- X)(?)
JKg(R d
J,R-D dR
p(J)=[J-D+X][T-D] K +D-X
r[1
_ (1 -p)(l-X) 1R
j g(R)dR 1X (1-G(R))+(1-A)d B
Differentiating,and collectingterms Hence for R<K, the upper bound on re-
turns from the safe project
JK g( dR

T-D aJ rK R ) + (12) d lnj -1-


+[J-D+X]
T XD
aJ dR
dIn(1 +rP)
f Kg(R)
[ -() ) dR] (1-X)(1 -p)Ag(R)R
jfg(R)dR+g(J)
(1 -XG(R))(X(I - G(R)) +p(l -X))

A sufficientconditionfor the existenceof an


-
interior bank optimal interest rate is again
Using l'Hopital'srule and the assumptionthat g(K) that limRK K / ar<O, or from (12), X/1 -X
0,oo
fimR,Kg(R)A>p/l-p. The greateris the
( I ad1 K-D?X riskinessof the riskyproject(the lower is p),
J-K T-D aJ / ( K-D 2(K-D) ) the more likely is an interiorbank optimal
interestrate. Similarly,the higheris the rela-
or sign( lim ap ) sign (K-D-X) tive proportionof the risk averseindividuals
affected by increasesin the interest rate to
Conditions2 and 3 follow in a similarmanner. risk neutralborrowers,the moreimportantis
VOL. 71 NO. 3 STIGLITZ AND WEISS: CREDIT RATIONING 401

the self-selectioneffect, and the more likely so


is an interiorbank optimalinterestrate.

II. InterestRate as an IncentiveMechanism (14) d =-B(1-Fi((l+r')B-C))

A. Sufficient Conditions Thus, if at some r, X} =7 k, the increasein


r lowers the expectedreturnto the borrower
The second way in which the interestrate from the projectwith the higherprobability
affects the bank's expected return from a of paying back the loan by more than it
loan is by changingthe behaviorof the bor- lowers the expected returnfrom the project
rower. The interests of the lender and the with the lower probabilityof the loan being
borrowerdo not coincide. The borroweris repaid.
only concernedwith returnson the invest- On the other hand, if the firm is indiffer-
ment when the firm does not go bankrupt; ent betweentwo projectswith the samemean,
the lender is concernedwith the actions of we know from Theorem 2 that the bank
the firm only to the extent that they affect prefers to lend to the safer project. Hence
the probabilityof bankruptcy,and the re- raising the interest rate above r could so
turns in those states of nature in which the increasethe riskinessof loans as to lower the
firm does go bankrupt.Becauseof this, and expectedreturnto the bank.
because the behaviorof a borrowercannot
be perfectlyand costlesslymonitoredby the THEOREM 8: The expected return to the
lender, banks will take into account the ef- bank is lowered by an increase in the interest
fect of the interest rate on the behavior of rate at r if, at r, the firm is indifferentbetween
borrowers. two projects j and k with distributions Fj(R)
In this section,we show thatincreasingthe and Fk(R), j having a higher probability of
rate of interest increases the relative at- bankruptcythan k, and there exists a distribu-
tractivenessof riskierprojects,for which the tion F,(R) such that
return to the bank may be lower. Hence, (a) Fj(R) represents a mean preserving
raising the rate of interest may lead bor- spread of the distributionF,(R), and
rowersto take actions which are contraryto (b) Fk(R) satisfies a first-order dominance
the interestsof the lender,providinganother relation with F,(R); i.e., FI(R)>Fk(R) for
incentive for banks to ration credit rather all R.
than raise the interestrate when there is an
excess demandfor loanablefunds. PROOF:
We returnto the generalmodel presented Sincej has a higher probabilityof bank-
above,but now we assumethat each firmhas ruptcythan does k, from Theorem7 and the
a choice of projects.Considerany two pro- initial indifferenceof borrowersbetweenj
jects, denoted by superscriptsj and k. We and k, an increasein the interestrate r leads
first establish: firms to preferprojectj to k. Becauseof (a)
and Theorem3, the returnto the bank on a
THEOREM 7: If, at a given nominal interest projectwhose returnis distributedas F,(R)
rate r, a risk-neutral firm is indifferent be- is higher than on projectj, and because of
tween two projects, an increase in the interest (b) the return to the bank on project k is
rate results in the firm preferring the project higher than the return on a project distrib-
with the higher probability of bankruptcy. uted as F,(R).

PROOF: B. An Example
The expected return to the ith project is
given by To illustratethe implicationsof Theorem
8, assumeall firms are identical,and have a
(13 w-E
choice of two projects,yielding,if successful,
axR'(I+-),_
returnsRa and Rb, respectively(and nothing
402 THE AMERICAN ECONOMIC REVIEW JUNE 1981

HI. TheTheoryof Collateral


and
LimitedLiability
PROJECT
b PROJEC'T
a An obvious objection to the analysis pre-
sented thus far is: When there is an excess
demand for funds, would not the bank in-
I I ..
crease its collateralrequirements(increasing
the liabilityof the borrowerin the event that
pbR0-Po Ra
Ap
r -
R'
~-1 --1
r the project fails); reducing the demand for
B(pb -p) B funds, reducingthe risk of default (or losses
to the bank in the event of default) and
FIGURE 6. AT INTERESTRATES ABOVE P*, THE increasingthe returnto the bank?
RiSKY PROJECTIS UNDERTAKENAND THE RETURN
TO THE BANK IS LOWERED
This objectionwill not in general hold. In
this section we will discuss various reasons
why banks will not decreasethe debt-equity
otherwise) where R' > Rb, and with proba- ratio of borrowers(increasingcollateral re-
bilities of success of pa and pb pa <pb. For quirements)8as a means of allocating credit.
simplicity assume that C=O. If the firm is A clear case in which reductions in the
indifferent between the projects at interest debt-equityratio of borrowersare not opti-
rate r, then mal for the bank is when smaller projects
have a higher probabilityof "failure," and
all potentialborrowershave the same amount
(15) [Ra -(1 ?P)B] pa =[Rb -(I+ P)B] pb
of equity. In those circumstances,increasing
the collateral requirements(or the required
i.e., proportion of equity finance) of loans will
imply financing smallerprojects. If projects
either succeedor fail, and yield a zero return
(16) B( ?+r)= pbRb -paRa =(I + ^*)B when they fail, then the increasein the col-
pb -pa
lateral requirementof loans will increase the
riskinessof those loans.
Thus, the expected return to the bank as a Another obvious case where increasing
functionof r appearsas in Figure6. collateral requirements may increase the
For interest rates below r*, firms choose riskiness of loans is if potential borrowers
the safe project, while for interest rates be- have differentequity, and all projectsrequire
tween r* and (Ra/B) - 1, firms choose the the sameinvestment.Wealthyborrowersmay
riskyproject.The maximuminterestrate the be those who, in the past, have succeeded at
bank could charge and still induce invest- risky endeavors.In that case they are likely
ments in projectb is r*. The highest interest to be less risk aversethan the more conserva-
rate which attractsborrowersis (Ra/B)- 1, tive individualswho havein the past invested
which would induce investmentonly in pro- in relatively safe securities, and are conse-
ject a. Thereforethe maximumexpected re- quently less able to furnishlarge amounts of
turn to a bank occurswhen the bank charges collateral.
an interest rate P* if and only if In both these examples collateral require-
ments have adverse selection effects. How-
ever, we will present a stronger result. We
pb( pbRb _paRa)
paR
pb -pa 8Increasing the fraction of the project financed by
equity and increasingthe collateralrequirementsboth
increase the expected return to the bank from any
WheneverpbRb >paRa 1+ r* >O, and p is particularproject.They have similarbut identical risk
and incentive effects. Although the analysis below
not monotonicin r, so there may be credit focuses on collateral requirements,similar arguments
rationing. apply to dept-equityratios.
VOL. 71 NO. 3 STIGLITZ AND WEISS: CREDITRATIONING 403

will show that even if there are no increasing Define


returns to scale in production and all indi-
viduals have the same utility function, the (18) VO(WO)=max{U(WOp*),JV(W0)]
sorting effect of collateral requirements can
still lead to an interior bank-optimal level of We note that
collateral requirements similar to the interior
bank-optimal interest rate derived in Sec-
tions I and II. In particular, since wealthier (9) dU(Wop*)
d WO
individuals are likely to be less risk averse,
we would expect that those who could put up
the most capital would also be willing to take (20) d(W?)=(1
the greatest risk. We show that this latter
effect is sufficiently strong that increasing
collateral requirements will, under plausible (where the subscript 1 refers to the state
conditions, lower the bank's return. "success" and the subscript 2 to the state
To see this most clearly, we assume all "failure"). We can establish that if there is
borrowers are risk averse with the same util- decreasing absolute risk aversion,9
ity function U( W), U'> 0, U" <0. Individu-
als differ, however, with respect to their ini- dU(Wop*) dV(Wo)
tial wealth, W0. Each "entrepreneur" has a dWo dWo
set of projects which he can undertake; each
project has a probability of success p(R), Hence, there exists a critical value of W0,
where R is the return if successful. If the Wo, such that if Wj > W0individuals who do
project is unsuccessful, the return is zero; not borrow undertake the project.
p'(R)<O. Each individual has an alternative For the rest of the analysis we confine
safe investment opportunity yielding the re- ourselves to the case of decreasing,absolute
turn p*. The bank cannot observe either the risk aversion and wealth less than W0.
individual's wealth or the project under- If the individual borrows, he attains a
taken. It offers the same contract, defined by utility level'0
C, the amount of collateral, and r, the inter-
est rate, to all customers. The analysis pro- (21) {max U(Wop*-(1 +P)+R)p
ceeds as earlier; we first establish:

THEOREM 9: The contract {C, r) acts as a + U((Wo- C)p*)( -p)}


screening mechanism: there exist two critical
values of WO,WO,and WO,such that if there is -VB(WO)
decreasing absolute risk aversion all individu-
The individual borrows if and only if
als with wealth WO< WO< WOapplyfor loans.

PROOF: (22) VB(WO)


2 VO(WO)
As before, we normalize so that all pro-
jects cost a dollar. If the individual does not 9To prove this, we define WOas the wealth where
borrow, he either does not undertake the undertaking the risky project is a mean-utility preserv-
ing spread (compare Peter Diamond-Stiglitz) of the safe
project, obtaining a utility of U(W0p*), or he project. But writing U'( W(U)), where W(U) is the value
finances it all himself, obtaining an expected of terminal wealth corresponding to utility level U,
utility of (assuming W0 > 1)
dU' U"l d2U' A'
(17) max{U((W0-1)p*+R)p(R) dU U' A; dU2 U-UOasA'gO

Hence with decreasing absolute risk aversion, U' is a


+ U((W0 - I)p*)(1 -p(R))} convex function of U and therefore EU' for the risky
investmentexceeds U'(p*WO).
- wo) l0ln this formulation, the collateral earns a return p*.
404 THE AMERICAN ECONOMIC REVIEW JUNE 1981

BORROWING so, using the second-order conditions for a


cx:
maximum, and (24),
0

dR U_'p + (U-U2_)p'
z SELF-FINANCED (25) dW t0as Up
RISKY INVESTMENT

SAFE INVESTMENT

H (U-Uf
-J~~~~~~~~~~EF
But
F 7ALL APPLY FOR LOANSE SE
A SCREENI E FINANCED
Li
I~~~~J
ALL ~RISKY lrn - __
-
--_ =A1
SELF-FINANCE j INVESTMENT
wI -w2 U,-U2 I
-\
~~~~~wo implying that, if WI = W2, dR/dWo = 0.
However,
FIGuRE 7. COLLATERALSERVESAS
A SCREENINGDEVICE

af -Al - U,l- U
But
awl Ul ' u;-
- ____
u;
dVB
(23) (L U( p ))p* A? U- U
dW =(Up+

=-At _ Ull + If U2 U,
Clearly, only those with W0> C can borrow. 1 Ul- U2 Ul-U2 Ul-U2
We assume there exists a value of W0>0,
denoted W0, such that VB(WO)=U(p*WO). -Al iO as Al ?0
(This will be true for some values of p*.) By
the same kind of argument used earlier, it is Hence dR/dWo >0 if A'<0.
clear that at W0, borrowing with collateral is Next we show
a mean-utility preserving spread of terminal
wealth in comparison to not borrowing and THEOREM 11: Collateral increases the
not undertaking the project. Thus using (20) bank 's returnfrom any given borrower:
and (23), dVB/dWo > dV( W0)/dWo at W0.
Hence, for W0< W<<W0 all individuals ap- dp/dC>O
ply for loans, as depicted in Figure 7.
Thus, restricting ourselves to W0<W0, we PROOF:
have established that if there is any borrow- This follows directly from the first-order
ing, it is the wealthiest in that interval who condition (24):
borrow. (The restriction W0<W0 is weaker
than the restriction that the scale of projects sign d=sign U2p*p'<0
exceeds the wealth of any individual.)
Next, we show:
and thus dp/dC>O. But
THEOREM 10: If there is decreasing abso-
lute risk aversion, wealthier individuals under- THEOREM 12: There is an adverse selection
take riskier projects: dR/dWo > 0. effect from increasing the collateral require-
ment, i.e., both the average and the marginal
PROOF: borrower who borrows is riskier," dWo/dC
From (21), we obtain the first-order condi- >0.
tion for the choice of R:
1 "Ata sufficiently high collateral, the wealthy individ-
(24) U"p+ (Ul - U2)p'=0 ual will not borrow at all.
VOL. 71 NO. 3 STIGLITZ AND WEISS: CREDITRATIONING 405

tive effect, this is not necessarilythe case.


The bank has limited control over the ac-
tions of the borrowers,as we noted earlier.
z Thus, the response of the borrowerto the
increase in lending may be to take actions
w which, in certain contingencies,will require
(n
the bank to lend more in the future. (This
argumentseemsimplicitin many discussions
z of the importanceof adequateinitial funding
4
cr~~~~~~~~~~ for projects.)Consider,for instance, the fol-
lowing simplifiedmultiperiodmodel. In the
first period,6 occurswith probabilityp,; if it
does, the return to the project (realizedthe
second period)is R,. If it does not, eitheran
additional amount M must be invested, or
FIGURE 8. INCREASINGCOLLATERALREQUIREMENT
LOWERSBANK'S RETURNS
the project fails completely(has a zero re-
turn). If the bank charges an interest rate
r2 on these additional funds, they will
PROOF: invest them in "safe" ways; if r2>?2 those
This followsimmediatelyupon differentia- funds will be investedin riskyways. Follow-
tion of (21) ing the analysisin SectionII, we assumethat
the risk differences are sufficiently strong
dVB/dC= - U2p*(1 -p)<O that the bank charges?2for additionalfunds.
Assume that there is also a set of projects
It is easy to show now that this adverse (actions)whichthe firmcan undertakein the
selection effect may more than offset the first period,but among which the bank can-
positive direct effect. Assume there are two not discriminate.The individualhas an equity
groups; for low wealth levels, increasingC of a dollar,whichhe cannot raise further,so
has no adverseselectioneffect, so returnsare the effect of a decreasein the loan is to affect
unambiguouslyincreased;but thereis a criti- the actions which the individualtakes, that
cal level of C such that requiringfurther is, it affects the parametersof the projects,
investmentsselect againstthe low wealth-low Ri, R2, and M, where M is the amount of
risk individuals, and the bank's return is second-periodfinancingneededif the project
lowered.'2(See Figure8.) fails in the first period. For simplicity, we
This simple example has demonstrated'3 take R2 as given, and let L be the size of the
that althoughcollateralmay have beneficial first-periodloan. Thus the expectedreturnto
incentive effects, it may also have counter- the firmis simply(if the additionalloan M is
vailingadverseselectioneffects. made when needed)

A. Adverse Incentive Effects


pi(Ri -(1 +P1)2L)
Although in the model presented above,
increasingcollateralhas a beneficial incen- +(R2 -[(1 +r )2L+(1 +?2)M])

1f we had not imposed the restriction WO<WO


where p P2(P1), (1+P1)2 is the amount
then there jnay exist a value of WO,WO> WO,such that
paid back(per dollarborrowed)at the end of
for WO?> WO,individuals self-finance. It is easy to show
the secondperiodon the initialloan and i2 is
that aWO/aC<O, so there is a countervailing positive
the interest on the additionalloan M; thus
selection effect. However if the density distribution of the firm chooses RI so that
wealth is decreasing fast enough, then the adverse selec-
tion effect outweighs the positive selection effect.
' 3It also shows that the results of earlier sections can - dM
be extended to the risk averse entrepreneur.
PI =P(1 +r2)dR
406 THE AMERICAN ECONOMIC REVIEW JUNE 1981

Assume that the opportunity cost of capital p


to the bank per period is p*. Then its net
expected return to the loan is
P***-

I(,( + -1)2Lp( +r)2L+l+2


2 p(r
-P*[p*L+(1 -p,)M]

r2 r3 r3 r1 r
We can show that under certain circum-
stances, it will pay the bank to extend the FIGURE 9. IF GROUPS DIFFER, THEREWILL EXIST
line of credit M. Thus, although the bank RED LINING
controls L, it does not control directly the
total (expected value) of its loans per
customer, L+ (I-p )M. THEOREM 13: For i>j, typej borrowers
But more to the point is the fact that the will only receive loans if credit is not rationed
expected return to the bank may not be to type i borrowers.
monotonically decreasing in the size of the
first-period loans. For instance, under the PROOF:
hypothesis that r', and P2are optimally cho- Assumenot. Sincethe maximumreturnon
sen and at the optimum p*>p2(l +r), the the loan to j is less than that to i, the bank
return to the bank is a decreasing function of could clearlyincreaseits returnby substitut-
M/L. Thus, if the optimal response of the ing a loan to i for a loan to j; hence the
firm to a decrease in L is an increase in M originalsituationcould not have been profit
(or a decrease in M so long as the percentage maximizing.
decrease in M is less than the percentage We now show
decrease in L), a decrease in L actually lowers
the bank's profits.'4 THEOREM 14: The equilibriuminterest rates
are such that for all i, j receiving loans, pi('i)
r
IV. Observationally
DistinguishableBorrowers =pj(r ).

Thus far we have confined ourselves to PROOF:


situations where all borrowers appear to be Again the proof is by contradiction.Let us
identical. Let us now extend the analysis to assume that pi(r)>pj(?-); then a bank lend-
the case where there are n observationally ing to typej borrowerswould prefer to bid
distinguishable groups each with an interior type i borrowersaway from other banks. If
bank optimal interest rate denoted by ri*.15 p* is the equilibriumreturnto the banks per
The function pi(ri) denote the gross return to dollar loaned, equal to the cost of loanable
a bank charging a type i borrower interest ri. funds if banks competefreelyfor borrowers,
We can order the groups so that for i >j, then for all i, j receiving loans pi(rj)=p (r-)
maxpj(ri)>maxpj(rj). =p*. These results are illustratedfor tiree
types of borrowersin Figure9.
'4Forinstance,if some of the initialinvestmentis for If banks have a cost of loanablefunds p*
"back-up"systemsin case of variouskindsof failure,if
the reductionin initial fundingleads to a reductionin then no type 1 borrowerwill obtain a loan;
investmentin these back-up systems, when a failure all type 3 borrowerswishing to borrow at
does occur,largeamountsof additionalfundingmaybe interestrate r3(whichis less than ?3*,the rate
required. which maximizesthe bank's return)will ob-
15The analysisin this section parallelsWeiss (1980) tain loans- competitionfor those borrowers
in which it was demonstratedthat marketequilibrium
could resultin the exclusionof some groupsof workers drives their interest rate down; while some,
from the labormarket. but not necessarilyall, type 2 borrowersre-
VOL. 71 NO. 3 STIGLITZ AND WEISS: CREDIT RATIONING 407

ceive a loan at P2*.If the interest rate were to


fall to p**, then all types 2 and 3 would z
receive loans; and some (but not all) type 1 z
w
borrowers would be extended credit.
Groups such as type 1 which are excluded 0

from the credit market may be termed "red- z


lined" since there is no interest rate at which H
they would get loans if the cost of funds is w
above p**. It is possible that the investments
of type 1 borrowers are especially risky so / ~~~~R
that, although pI(rf*)<p3(P3*), the total ex- R=rent
pected return to type 1 investments (the re- FIGURE lOa
turn to the bank plus the return to the bor-
rower) exceeds the expected return to type 3
investments. It may also be true that type 1
loans are unprofitable to the bank because 0
they find it difficult to filter out risky type 1
investments. In that case it is possible that 0

the return to the bank to an investment by a z


type 1 borrower would be greater than the Sz
return to a type 3 investment if the bank
H
could exercise the same control (judgment) w
over each group of investors.
Another reason for pI(Irf)< p3(p3*) may be R
that type 1 investors have a broader range of R= rent
available projects. They can invest in all the FIGURE lOb
projects available to type 3 borrowers, but
can also invest in high-risk projects unavaila-
ble to type 3. Either because of the convexity problems is how to provide the proper incen-
of the profit function of borrowers, or be- tives for the agent. In general, revenue shar-
cause riskier investments have higher ex- ing arrangements such as equity finance, or
pected returns type 1 borrowers will choose sharecropping are inefficient. Under those
to invest in these risky projects. schemes the managers of a firm or the tenant
Thus, there is no presumption that the will equate their marginal disutility of effort
market equilibriumallocates credit to thosefor with their share of their marginal product
whom the expected return on their investments rather than with their total marginal product.
is highest. Therefore, too little effort will be forthcom-
ing from agents.
IV. Debt vs. EquityFinance,AnotherView Fixed-fee contracts (for example, rental
of the Principal-Agent
Problem agreements in agriculture, loan contracts in
credit markets) have the disadvantage that
Although we have phrased this paper in they impose a heavy risk on the agent, and
the context of credit markets, the analysis thus if agents are risk averse, they may not
could apply equally well to any one of a be desirable. But it has long been thought
number of principal-agent problems. For ex- that they have a significant advantage in not
ample, in agriculture the bank (principal) distorting incentives and thus if the agent is
corresponds to the landlord and the bor- risk neutral, fixed-fee contracts will be em-
rower (agent) to the tenant while the loan ployed.'6 These discussions have not consid-
contract corresponds to a rental agreement.
The return function for the landlord and
'6See, for instance, Stiglitz (1974). For a recent for-
tenant appears in Figures iGa and lOb. The malization of the principal-agent problem, see Steven
central concern in those principal-agent Shavell.
408 THEAMERICANECONOMICREVIEW JUNE 1981

ered the possibilitythat the agent will fail to In a multiperiodcontext, for instance, banks
pay the fixed fee. In the particularcontext of could reward"good"borrowersby offering
the bank-borrowerrelationship,the assump- to lend to them at lower interest rates, and
tion that the loan will alwaysbe repaid(with this would induce firms to undertake safer
interest) seems most peculiar. A borrower projects (just as in the labor market, the
can repaythe loan in all statesof natureonly promise of promotion and pay increases is
if the riskyproject'sreturnsplus the value of an importantpart of the incentive and sort-
the equilibriumlevel of collateralexceedsthe ing structureof firms,see Stiglitz, 1975, J. L.
safe rateof interestin all statesof nature. Guasch and Weiss, 1980, 1981). In our 1980
The consequencesof this are important. paper, we analyze the nature of equilibrium
Since the agent can by his actions affect the contracts in a dynamic context. We show
probability of bankruptcy, fixed-fee con- that such contingency contracts may char-
tracts do not eliminate the incentive prob- acterizethe dynamicequilibrium.Indeed, we
lem. establish that the bank may want to use
Moreover,they do not necessarilylead to quantity constraints- the availability of
optimalresourceallocations.For example,in credit-as an additional incentive device;
the two-projectcase discussedabove(Section thus, in the dynamiccontext there is a fur-
II, Part B), if expected returns to the safe ther argumentfor the existence of rationing
project exceed that to the risky (psRs >prRr) in a competitiveeconomy.
but the highest rate which the bank can Even after introducingall of these addi-
chargeconsistentwith the safe projectbeing tional instruments(collateral, equity, non-
chosen (r*) is too low (i.e., pS(l +r*)>prRr) linear payment schedules, contingency con-
then the bank chooses an interestrate which tracts) there may exist a contract which is
causes all its loans to be for risky projects, optimal from the point of view of the prin-
although the expected total (social) returns cipal; he will not respond,then, to an excess
on these projects are less than on the safe supply of agentsby alteringthe terms of that
projects.In this case a usury law forbidding contract;and theremay then be rationingof
interestrates in excess of r* will increasenet the form discussedin this paper, that is, an
national output. Our 1980 paper and Janusz excess demandfor loans (capital,land) at the
Ordover and Weiss show that government "competitive"contract.
interventionsof variousformslead to Pareto
improvementsin the allocationof credit. VI. Conclusions
Because neither equity finance nor debt
finance lead to efficient resourceallocations, We have presenteda model of credit ra-
we would not expect to see the exclusiveuse tioning in which amongobservationallyiden-
of either method of financing (even with tical borrowerssome receiveloans and others
risk-neutralagentsand principals).Similarly, do not. Potentialborrowerswho are denied
in agriculture,we would not expectto see the loans would not be able to borrow even if
exclusive use of rental or sharecropping they indicated a willingness to pay more
tenancy arrangements. In general, where than the marketinterest rate, or to put up
feasible,the payoff will be a non-linearfunc- more collateralthan is demanded of recipi-
tion of output (profits). The terms of these ents of loans. Increasing interest rates or
contractswill depend on the risk preferences increasing collateralrequirementscould in-
of the principal and agent, the extent to crease the riskinessof the bank's loan port-
which their actions (both the level of effort folio, either by discouragingsafer investors,
and riskiness of outcomes) can affect the or by inducingborrowersto invest in riskier
probabilityof bankruptcy,and actions can projects, and therefore could decrease the
be specifiedwithin the contractor controlled bank's profits.Hence neitherinstrumentwill
directlyby the principal. necessarilybe used to equate the supply of
One possiblecriticismof this paperis that loanable funds with the demandfor loanable
the single period analysis presented above funds. Under those circumstancescredit re-
artificiallylimitsthe strategyspaceof lenders. strictions take the form of limiting the num-
VOL. 71 NO. 3 STIGLITZ AND WEISS. CREDIT RATIONING 409

ber of loans the bank will make,ratherthan directly affects the quality of the loan in a
limitingthe size of each loan, or makingthe manner which matters to the bank. Other
interestrate chargedan increasingfunction models in which prices are set competitive-
of the magnitude of the loan, as in most ly and non-market-clearingequilibria exist
previousdiscussionsof creditrationing. share the propertythat the expected quality
Note that in a rationing equilibrium,to of a commodityis a functionof its price (see
the extent that monetarypolicy succeedsin Weiss, 1976, 1980,or Stiglitz,1976a,b for the
shiftingthe supplyof funds, it will affect the labor marketand C. Wilson for the used car
level of investment,not throughthe interest market).
rate mechanism, but rather through the In any of these models in which, for in-
availability of credit. Although this is a stance, the wage affects the qualityof labor,
"monetarist"result, it should be apparent if thereis an excess supply of workersat the
that the mechanismis different from that wage which minimizeslabor costs, there is
usuallyput forth in the monetaristliterature. not necessarilyan inducementfor firms to
Although we have focused on analyzing lower wages.
the existenceof excess demandequilibriain The Law of Supplyand Demand is not in
credit markets, imperfect information can fact a law, nor should it be viewed as an
lead to excess supply equilibriaas well. We assumptionneeded for competitiveanalysis.
will sketchan outline of an argumenthere (a It is rathera resultgeneratedby the underly-
fuller discussion of the issue and of the ing assumptionsthat priceshave neithersort-
macro-economicimplicationsof this paper ing nor incentive effects. The usual result
will appearin futurework by the authorsin of economic theorizing: that prices clear
conjunctionwith BruceGreenwald)."7 Let us markets,is model specific and is not a gen-
assumethat banks make higherexpectedre- eral property of markets- unemployment
turns on some of their borrowersthan on and creditrationingare not phantasms.
others: they know who their most credit
worthy customersare, but competingbanks
do not. If a banktriesto attractthe customers REFERENCES
of its competitorsby offeringa lowerinterest
rate, it will find that its offer is counteredby P. Diamond and J. E. Stiglitz, "Increases in
an equally low interest rate when the Risk and in Risk Aversion," J. Econ. The-
customer being competed for is a "good" ory, July 1974, 8, 337-60.
credit risk, and will not be matched if the M. Freimerand M. J. Gordon,"Why Bankers
borroweris not a profitablecustomerof the Ration Credit," Quart.J. Econ., Aug. 1965,
bank. Consequently,banks will seldom seek 79, 397-416.
to steal the customersof their competitors, Bruce Greenwald,Adverse Selection in the
since they will only succeedin attractingthe Labor Market, New York: Garland Press
least profitableof those customers(introduc- 1979.
ing some noise in the system enables the J. L. Guaschand A. Weiss, "Wages as Sorting
developmentof an equilibrium).A bankwith Mechanisms: A Theory of Testing," Rev.
an excess supply of loanable funds must Econ. Studies, July 1980, 47, 653-65.
assess the profitabilityof the loans a lower and _ , "Self-Selection in the
interest rate would attract. In equilibrium Labor Market," Amer. Econ. Rev., forth-
each bank may have an excess supply of coming.
loanable funds, but no bank will lower its DwightJaffee, Credit Rationing and the Com-
interestrate. mercial Loan Market, New York: John
The reason we have been able to model Wiley & Sons 1971.
excess demand and excess supply equilibria and T. Russell, "Imperfect Informa-
in credit markets is that the interest rate tion and Credit Rationing," Quart.J. Econ.
Nov. 1976, 90, 651-66.
17A similar argument to that presented here appears W. Keeton,EquilibriumCredit Rationing, New
in Greenwald in the context of labor markets. York: Garland Press 1979.
410 THE AMERICAN ECONOMIC REVIEW JUNE 1981

J. OrdoverandA. Weiss, "Information and the come in L.D.C.'s," Oxford Econ. Papers,
Law: Evaluating Legal Restrictions on July 1976, 28, 185-207.
Competitive Contracts," Amer. Econ. Rev. ,9___ "Perfect and Imperfect Capital
Proc., May 1981, 71, 399-404. Markets," paper presented to the New
M. Rothschildand J. E. Stiglitz, "Increasing Orleans meeting of the Econometric
Risk: I, A Definition," J. Econ. Theory, Society,Dec. 1970.
Sept. 1970, 2, 225-43. , "Some Aspects of the Pure Theory
S. Shavell, "Risk Sharing and Incentives in of CorporateFinance: Bankruptciesand
the Principal and Agent Problem," Bell J. Take-Overs,"Bell J. Econ., Autumn 1972,
Econ., Spring 1979, 10, 55-73. 3, 458-82.
G. Stigler, "Imperfections in the Capital and A. Weiss, "Credit Rationing in
Market," J. Polit. Econ., June 1967, 85, Marketswith ImperfectInformation,Part
287-92. II: A Theory of ContingencyContracts,"
J. E. Stiglitz,"Incentives and Risk Sharing in mimeo. Bell Laboratoriesand Princeton
Sharecropping," Rev. Econ. Studies, Apr. Univ. 1980.
1974, 41, 219-55. A. Weiss, "A Theory of Limited Labor
, "Incentives, Risk, and Information: Markets,"unpublisheddoctoral disserta-
Notes Towards a Theory of Hierarchy," tion, StanfordUniv. 1976.
Bell J. Econ., Autumn 1975, 6, 552-79. , "Job Queues and Layoffs in Labor
, "Prices and Queues as Screening Markets with Flexible Wages," J. Polit.
Devices in Competitive Markets," IMSSS Econ.,June 1980,88, 526-38.
tech. report no. 212, Stanford Univ. C. Wilson,"The Nature of Equilibriumin
, "The Efficiency Wage Hypothesis, Marketswith Adverse Selection,"Bell J.
Surplus Labor and the Distribution of In- Econ., Spring 1980, 11, 108-30.

S-ar putea să vă placă și