allocation of credit

8/11/2019 Allocation of Credit

1/18

THE ALLOCATION OF CREDIT AND FINANCIALCOLLAPSE*

This paper examines the allocation of credit in a market in which borrowershave greater information concerning their own riskiness than do lenders. It il-lustrates that (1) the allocation of credit is inefficient and a t times can be improvedby government intervention, and 2) small changes in the exogenous risk-freeinterest rate can cause large (discontinuous) changes in the allocation of creditand the efficiency of the market equilibrium. These conclusions suggests a rolefor government as the lender of last resort.

In this paper I examine the allocation of credit in a marketin which borrowers have greater information on the probabilityof default than do lenders. My purpose is to illustrate two propo-sitions. First, the equilibrium resulting in a n unfettered marketis inefficient and can be improved by government intervention ofa sort often observed, even if the government has no informational

advantage over lenders. Second, the unfettered market equilib-rium is precarious: small changes in the exogenous risk-free in-terest rate can cause large and inefficient changes in the allo-cation of credit.

Many recent studies note the importance of asymmetric in-formation for credit markets.' The two results emphasized here,while natura l consequences of asymmetric information, often es-cape unnoticed. Understanding these conclusions, however, iscritical to evaluating the impact of various government policies.

Government intervention into the allocation of credit is sub-stantial. Federal loan guarantees to the Chrysler Corporation andto New York City are among the most publicized examples. On acontinuing basis, the government plays a central role in the mar-kets for loans to students, farmers, and homeowners. Economistsoften criticize this role on the grounds that the market can best

*I am grateful to Andy Abel, Robert Barsky, Olivier Blanchard, John Earle,Avery Katz, Mervyn Kin Deborah Mankiw, David Romer, Matthew Shapiro,Lawrence Summers, and e referees for helpful comments.


2/18

456 Q U A RT E R L Y J O U R N A L OF E C ON O M IC S

allocate red it ^ The model presented here shows that this conclu-sion is not generally correct. While credit programs are fre-

quently justified on distributional grounds, I show that a socialplanner concerned solely with economic efficiency may often en-dorse the type of policy currently effective in many creditmarkets.

The model presented here is close in spirit to those of Stiglitzand Weiss [I9811 and Ordover and Weiss [1981]. The commontheme is that changes in the interest rate alter the riskiness ofthe pool of borrowers. While these previous two papers note the

possibility that the equilibrium is socially inefficient, the policyinterventions they propose do not correct the market failure dis-cussed here. Stiglitz and Weiss suggest a usury law an interestrate ceiling) as one solution. In th e model of thi s paper a usurylaw does not improve on the market allocation; instead, i t causesthe market for these loans to disappears3 Ordover and Weiss pro-pose the policy of forcing banks to lend to all borrowers a t someinterest rate. The equilibrium in this model, however, can be

inefficient even if no borrower is credit rationed in the sense ofbeing excluded at any interest rate; even when such credit ra-tioning does occur, the Ordover-Weiss policy merely induces banksto charge a prohibitively high in terest rate. In neither case is thispolicy effective. Nonetheless, a credit subsidy, such as a loanguarantee, can a t times improve on the market allocation.

The model also has macroeconomic implications. As notedabove, in the absence of government intervention, an increase in

the exogenous risk-free interest ra te can cause the collapse of thecredit market. A market t hat was efficient at the initial interestrate can disappear, driving out socially profitable investments.In other words, the total surplus derived for a particular creditmarket can be a discontinuous function of the interest rate. Inmodels without asymmetric information, restrictive monetary policymoves the economy smoothly along the marginal efficiency of

2. For exam le, the 1982 Econo mic Report of the President [p. 941 after notingthat the Federa f~overn ment was involved in 21.4 percent of all funds advancedin U S. credit markets in fiscal year 1982, presents the standard evaluation of


3/18

L L O C T I O N O F C R E D I T N D F I N N C I L C O L L P S E 57

capital schedule; in this model restrictive policy is potentiallymore costly as it can precipitate a financial crisis.

While the model in this paper is general I present i t in thecontext of a specific credit market. In particular I discuss loansfrom banks to students. There are two reasons that s tudent loansare a useful prototype for studying credit market imperfections.First only a limited number of financial instruments are avail-able to students. corporation can fund investments with eitherdebt or equity. It has in addition more complex options: preferredstock convertible bonds callable bonds. Imperfections in the mar-

ket for one instrument may be less important if there exist otherfinancing methods. In contrast a student faces a much simplerproblem. He must borrow if he is to invest; he cannot issue equityon his human capital. In principle we could attempt to explainthe paucity of financial instruments available to the student. Forthis paper though it is both reasonable and realistic to assumethat his only option is debt finance.

The second reason for discussing the market for student loans

is that it has evoked substantial government intervention. TheOECD reported in 1978 that Canada France Germany Japanthe Netherlands Norway Sweden and the United States all pro-vide assistance to students in the form of loans or loan guarantees.Of course there are many reasons for public support of education.Nonetheless it is instructive that this support so often takes theform of credit market intervention. The pervasiveness of publicstudent loan programs suggests a t least the perception of imper-

fections in the market for credit.

This section presents a simple model of a market for loans toa particular group of students . To the banks who provide theloans the students are indistinguishable. The students thoughdiffer by the expected return on their education and by their

probability of repaying the loan. Each student knows his ownexpected return and repayment probability even though they arenot observable by the banks or by the government.


4/18

458 Q U RT E R LY J O U R N L OF E C O N OM I C S

attributable to the underprovision of insurance to risk-averseagents.

Each potential student is considering investing in some hu-man capital, say, a college degree. The project is discrete, hasunit cost, and has expected future payment R. (All return vari-ables I use are expressed as the return factor. That is, if theexpected return is 5 percent, then R 1.05. The other charac-teristic of each student is the probability P th at he will repay theloan. The values of R and P vary across students. Each studentknows his own R and P, but these characteristics are not observ-

able by banks. These two characteristics are distributed through-out the population with the density function f(P,R), which ispublic knowledge.

The model takes each student s parameters P and R as primi-tive. One could construct a more complete model in which thestudent s default behavior is endogenous. For example, one couldmodel the students as having varying degrees of honesty; certainstudents get greater disutility from dishonest acts. Default prob-

abilities vary because a less honest student is more likely to avoidrepayment illegitimately. Alternatively, one could model all thestudents as well-meaning. A student then defaults when his re-turn leaves him unable to repay his loan ex post; the probabilityof this state occurring is then private i n f ~ r m a t i o n . ~ ither suchmodel might suggest that each student s repayment probabilitydepends on the market interest rate. I maintain the assumptionthat P is exogenous for each student to simplify the exposition.

A bank can invest in a safe asset, such as a Treasury bill,and obtain the certain future payment p. Alternatively, a bankcan lend to one of the above group of students. Let r be the interestrate the bank charges these students. It is the same for all stu-dents, since they are indistinguishable to the bank.

If a student defaults, the bank receives no payment on theloan. Including a default payment of A A p), such as collateral,is certainly possible. In particular, in such a world, one couldconsider the student as taking out a loan of Alp that is repaidwith certainty and a loan of 1 Alp that is repaid with probabilityP and fully defaulted with probability 1 P It is straightforward


5/18

ALLOCATION OF CREDIT AN D FINANCIAL COLLAPSE 459

FIGURE Area of Investment

the existence of the risk-free loan does not substantially affectthe market for the risky loan. Thus I set the default payment tozero without loss of generalityS5

Let be the average probability of repayment; that isT

isthe average of P for those potential students who in fact borrow.The expected payment to the bank on a student loan is IIr. Thisexpected payment must equal the safe payment p if the bank isto make any student loans. Hence the first equilibrium conditionis

1) IIr pThis equation describes the locus of market loan rates and re-payment probabilities th at provide the lenders the required rat eof return.

Each potential student must decide whether to borrow atthe market rate r and invest in additional human capital. Theexpected return on his investment is R , while the expectedcost of borrowing is Pr. Hence he borrows and invests if andonly if R Pr. It is useful to examine this investment condition

graphically. Figure I shows the area in R , P ) space for whichthe investment is made. Those students in areas A and B bor-row and invest. Those in areas C and D do not. An increase in theloan rate from ro to rl nambig o sl red ces the areas A and


6/18

46 QU ART ERL Y JO UR NAL OF ECONOMICS

B, and thus reduces the number of loans and investments. Givenany expected return R the students driven out by the increase

in the interest rate are those with relatively high repaymentprobabilities.Even a t this early stage we can show th at the market al-

location is not fully efficient. An investment should be made fromthe viewpoint of the social planner if and only if the expectedreturn R exceeds the opportunity cost p. Those investments inarea B are socially efficient and are undertaken while those inC are socially inefficient and are not undertaken. Yet those in-

vestments in area are socially efficient but not undertaken andthose in area A are socially inefficient and are undertaken. Noloan rate r can make both areas and D disappear. Thus no loanrate including the market equilibrium rate can in general reachthe first best allocation.

The assumption of asymmetric information plays a key rolehere. As already mentioned a student invests if R Pr. Usingthe equilibrium condition I ) , the investment condition is

R Plll)p. If there is no information asymmetry regarding thedefault probability then P ll , and the student invests if andonly if his return is socially profitable R p . In this case themarket reaches the fully efficient allocation even though R is notpublicly observable. If there is information asymmetry regardingP, then low P investments are overly encouraged and high Pinvestments are overly discouraged.

The repayment probability T as seen by banks is the averageof P for those students who invest tha t is for those in areas Aand B. Thus the function relating r to n is

2) l l r ) E[P Prl.

For any density f P,R), the function l l r ) is a well-defined con-ditional expectation.

Equations 1) and 2) are the two equilibrium conditions.They simultaneously determine the market loan rate r, from which


7/18

ALLOCATION OF CREDIT AN D FZNM CZA L COLLAPSE 461

FIGURE 1Market Equilibrium

Equation 2) isslabeled BB, since it is determined by theoptimizing reaction of borrowers. The shape of the BB curve ismore ambiguous. As r goes to zero, approaches the uncondi-tional expectation E P ) 1, since everyone borrows. As r goesto infinity, ll goes to zero as long as R is bounded from aboveand f P,R) is nonzero everywhere else. If R were constant acrossborrowers, the BB curve would be monotonic; as r rises, highP borrowers drop out of the market. In general, both R and P varyand the curve need not be m on ~ to n ic .~

The LL curve and the BB curve might not intersect, as inFigure 111. In this case, there is no equilibrium in which loans

are made. At any interest rate, the pool of students who seekloans is too risky to give the banks their required return. I callthis a "collapsed" credit market.

6. Even i f P and R are independently distributed, the curve can be upwardslo ing in parts. For example, suppose that R takes on the values 1.0 and 3.0eat with probability is i and P takes on the values 0.5 and 1.0 each withprobability . Then the equation for the BB curve is


8/18

QU RTERLY JOURN L OF ECONOMICS

FIGURE 11Collapsed Credit Ma rke t

The two curves may intersect more than once, as they do inFigure 11. If they do cross more than once, it seems reasonable torestrict attention to the first intersection and to rule out anyadditional equilibrium as not stable. To see why, consider point

y in Figure 11. If the economy were a t this point, both equilibriumconditions would be satisfied. But a bank would have an incentiveto lower its interest rate, say to r,. At rl, the curve lies abovethe LL curve. The repayment probability 11 is thus greater thannecessary to give the bank its required return p. A bank cantherefore make a h i g h e ~ xpected re turn by charging rl, whichis below the market ra te r . Similar reasoning shows that pointx is a stable equilibrium. At interest ra tes just above r,, lenders

can earn a rate of return above p which causes a capital inflowand lowers the interest rate. Conversely, a t interest rates belowr,, the repayment probability is too low to give banks a return ofp, which causes a capital outflow and raises the interest ra te . Forthese reasons, only the first crossing appears to be an interestingequilibrium.

It is possible that there are more than two crossing of thetwo curves. If so, at the third (and every odd) crossing the BBcurve cuts the LL curve from below. Thus, the intersection islocally stable; that is, a bank could not make a profit by a smallreduction in its interest rate. The further intersections, however,


9/18

ALLOCATION OF CREDIT AND FIN ANC IAL COLLAPSE 46

erties of the market equilibrium are impossible. As discussedabove if there is no information asymmetry regarding P , themarket reaches the first best allocation. At other times howeverthe market is grossly inefficient and government intervention canenhance efficiency. To illustrate this possibility I examine a spe-cial case.

Example: Un iform Expected Re turn

Suppose th at the expected return R is constant across stu-dents. The only unobserved heterogeneity is the repayment prob-ability P . In this case the market fails to reach the first bestallocation.

Since R is constant either all the investments are sociallyefficient or all are socially inefficient. If an equilibrium existsthen the investments are socially efficient. That is if the BB curveand the LL curve intersect then the expected return R exceedsthe opportunity cost of capital p. This proposition is easy to prove.As discussed earlier a student borrows if and only if R > Pr.Averaging this inequality over the investments undertaken showsR > l l r , Since l l r p we know R > p Thus ifR is an observablecharacteristic then the unfettered market equilibrium allows onlysocially productive investments.

On the other hand investments may be socially productivebut not undertaken in equilibrium. That is it is possible theprojects are productive in the sense that R > p but not all in-vestments are undertaken. An example most easily shows thisproposition. Suppose that P is uniformly distributed from zero toone. Then equilibrium condition 2 ) becomes for r R ,l Rl2r for r > R . The curve lies above the curve a t allr, unless R > 2p. In this example no equilibrium exists unlessthe expected payment is twice the required payment. The unfet-tered market equilibrium may leave profitable investment op-portunities unrealized.

IV. GOVER NMENTREDITPOLICYI now discuss the potential for efficient government inter-

ti I gi th t th k t b gi i th g l t d i


10/18

6 QU RTERLY JOURN L O F ECONOMICS

reduces the area D; some of those students with high returns andhigh repayment probabilities who were previously not investing

are now induced to invest. Second, it increases area A ; some morestudents with low returns and low repayment probabilities areinduced to invest. The first effect is socially beneficial, while thesecond is socially harmful.

A government loan guarantee is a special case of a subsidy.In particular, under a guarantee program, the market rate be-comes the risk-free rate r p), that is, a loan guarantee isequivalent to a government loan a t the risk-free interest rate. Atr p area D disappears, implying that all socially productiveinvestments are undertaken.

To evaluate the net social impact of such a subsidy, one needsonly to know the distribution of at tributes, f P,R). It is not nec-essary for the government to be able to distinguish the high-return students from the low-return students. As the examplebelow illustrates, it is possible that the extra investment gener-ated is on net socially optimal but is not undertaken in the marketequilibrium because i t requires that Ilr p.

Of course, a government credit subsidy has a budgetary cost.While the return from students to banks is Ilr, banks still requirereturn p. The difference is made up by the government. If thegovernment must raise money using distortionary taxes, then thedeadweight losses are an additional cost of the credit program.As with all expenditure programs, the marginal benefit mustexceed the marginal deadweight losses if the program is to besocially efficient.

Before turning to the example, a few general propositionsregarding the optimal interest rate r* can be established. First,r* is never below the risk-free return p; charging a lower ratewould only induce inefficient investment. Second, r* is generallystrictly above the risk-free rate. To see this, note that social wel-fare ignoring the cost of raising revenue) is

3) Social Welfare R P f P,R, ) dR dP.

Th d i ti f i l lf ith t t th i t t t


11/18

A L L O C AT IO N O F C R E D I T A N D F I N A N C I A L C O L L A P S E 65

Evaluated at r p this derivative is nonnegative and is strictlypositive as long as f P,R) is everywhere nonzero. Thus, an effi-cient loan program generally charges a loan rate greater thanthe risk-free rate.

Third, depending on the density f P,R), it is possible thatthe optimal interest rate r* exceeds the unregulated equilibriuminterest rate re . In this case, the government would tax studentloans to drive out borrowers with low R and low P. While it isdifficult to derive general conditions under which r* > re, it ispossible to examine the effect on social welfare of small changesin the interest rate around re. In particular, a t r e the sign of thedSWldr is the same as the sign of d I I l d ~ - . ~ ence, if the BB curveis upward sloping at the equilibrium, then a small increase in theinterest rate is welfare-enhancing; conversely, if the BB curve isdownward sloping at the equilibrium, then a small decrease inthe interest rate is welfare-enhancing. In other words, if a smallsubsidy or tax is to increase social welfare, it must increase theaverage repayment probability.

Example, Continued

Consider again the example of uniform expected returns. Sup-pose that the government provides a loan guarantee. The verticalline r = p replaces the LL curve, as in Figure IV. This programclearly changes th e nature of the equilibrium. In particular, i t ispossible that the guarantee program creates a market, whereaswithout the program, no market existed, as in Figure 111.

As already noted, under a guarantee program, all sociallyprofitable investments are undertaken. It is possible, though, thatsocially unproductive investments are undertaken once the guar-antee is provided. That is, even if R < p those students for whomP < Rlp choose to borrow and invest. Of course, since R is knownin this example, the government can avoid this inefficiency byproviding guarantees only if R > p.

7 Suppose th at f ( P, R ) is (1,2p)with probability and (113,2p13)with prob-ability 4 The equilibrium interest rate is r 3p12, in which case both types


12/18

QU RTE RLY JOU RN L OF ECONOMICS

IGUR VEquilibrium with G overnment Loan a t Risk-free Rate

The reason the government can so effectivelycorrect th e mar-ket failure is that th e government requires a different informationset than private lenders. To make a socially profitable loan, weonly need to know that the expected return R exceeds th e required

return p. The probability of repayment P per se is irrelevant toa social planner. Remember that th e government program is notrequired to be revenue neutral.) Yet, to private lenders the ex-pected return on a project R per se is irre levant , and the repay-ment probability P is critically important. Hence, this exampleof constant R may be the case in which the government can mosteasily improve on the private allocation of credit.

Under what conditions is the unfettered market least effi-cient? To answer this question, I specialize the example furtherby supposing that P is uniformly distributed from Po to PI. From2)and straightforward calculation, we can compute the equation

for the BB curve. It is

RrI for r

2 PI

Po Rlr R R for r -.2 p1 Po

The intersection of (1) and (5)determines the interest rate in a n


13/18

L L O C T I O N OF C R E D I T N D F I N N C I L C O L L P S E 67

Note that the level of risk-Po and PI-is not relevant to thenumber of loans made in equilibrium. Instead the ratio PIIPo isthe crucial determinant. As PIIPo increases the number of loansmade decreases. The more heterogeneous are the borrowersin terms of their repayment probabilities the more severe isthe market failure and the greater is the benefit of governmentintervention.

Let us now return to the unfettered market equilibrium andconsider the effects of an increase in the required rate of returnp. This change shifts the LL curve upward and to the right as inFigure V. Not surprisingly the interest rate charged to theseborrowers increases. As shown in Section 11 the number of stu-dents taking out loans declines. The effect on T is in generalambiguous as the BB curve need not be a downward sloping.

An increase in the interest rate can have far more serious

effects. It is possible that a shift in the curve can make theequilibrium disappear. Whereas at the lower interest rate theeconomy is modeled as in Figure 11; at the higher interest rateFigure I11 is the more appropriate representation. Remember thatthe investment projects may still be socially profitable at the newhigher interest rate. Nonetheless none of the investors is able toraise the necessary capital.

An inward shift in the BB curve has the same effects as anincrease in the interest rate. A small increase in the riskiness ofsome of the potential borrowers can cause the credit market for


14/18

QU RTE RLY JO UR N L OF ECONOMICS

urplus

IGUR VISu rplus as a Function of the In terest Rate

all of them to collapse, even though there may be no change inthe expected return of investment projects R. Hence, small changesin risk perception can have large effects upon the allocation of

credit.One of the previous examples can usefully illustrate the mod-

el s potential for financial collapse. Let R be constant, and let Pbe uniformly distributed from zero to one. Section 111 showed tha tno equilibrium exists when p Rl2. At p Rl2, all students bor-row in the equilibrium. Figure VI displays the surplus receivedfrom this market as a function of the safe interest rate. At p Rl2,the surplus received is R p; while a t Rl2, no surplus is

received, as no loans are made. Thus, a t p Rl2, there is a severediscontinuity. A small increase in the interest rat e can cause thedisappearance of market for loans to these borrowers. The socialcost of this sudden financial collapse is potentially great and couldreasonably motivate the government to act as the lender of lastresort.

This potential for financial collapse has important macroeco-nomic implications. In the textbook IS-LM model, restrictive mon-etary policy (or any contractionary shift in the LM curve) reducesaggregate demand by increasing the real interest rate. At thisgreater required ra te of return, some investments are no longer


15/18

A L L O C AT I O N O F C R E D I T A N D F I N A N C I A L C O L L A P S E 69

can cause the collapse of the market to some borrowers, eventhough their projects may remain socially productive a t the higher

interest rate. A monetary contraction can therefore have a largeimpact on the efficiency of the market allocation of credit. It ispossible that when the monetary authority induces or allows a

credit crisis, the government should intervene on behalf of cer-tain borrowers, even though these borrowers may require no as-sistance under normal credit market conditions.

VI. CONCLUSIONSThe Federal government has played a central role in the

allocation of credit among competing uses. This paper illustratesthat this sort of government program can under plausible con-ditions improve on the unfettered market allocation. necessarycondition for efficient government intervention is unobservableheterogeneity among would-be borrowers regarding the proba-bility of default. The greater is such heterogeneity, the greateris the potential for efficient intervention.

Historical examinations of financial markets (e.g., Kindle-berger [19781 emphasize their propensity for instability and col-lapse. Our models should therefore reflect this instability. If weare to understand the effects of alternative monetary policies, forexample, we must appreciate the potential for financial crisis. Attimes, it is necessary for the government to remove some riskfrom the private sector by guaranteeing certain financial ar-rangements or, equivalently, by acting as a lender of last resort.

Akerlof, George, The M ark et for 'Lemons': Qualita tive Unce rtainty a nd the M ar-ke t Mechanism, thi sJournal LXXXIV (19701, 488-500.Be rnan ke, Ben, N6n-M onetary Effects of th e Financ ial Crisis in th e Propagationof th e Gre at Depression, Am erican Economic Review LXX III (19831,257-76.

Blinder, Alan S ., an d Josep h E. Stiglitz, Money, Credit Con straints, and EconomicA ti it A i E i R i P d P di LXXIII (19831


16/18

470 Q U A RT E R LY J O U R N A L OF E C ON O M IC S

Ordover, Jan usz , an d Andrew W eiss, Information and the Law: Ev alu atin g LegalRestrictions on Competitive Contracts, Am erican Economic Review Papersan d P roceedings LXXI (1981), 39 9 40 4 .

Organiza tion for Economic Coo erat ion and Development,Review of Student Su p-port Sch eme s i n Selected ~ E C DCountries (Paris: OECD, 1978).Stiglitz, Joseph , and Andrew W eiss, Credit Rationing in M arkets with Imperfect

Information, American Economic Review LXXI (19811, 3 9 3 4 1 0 .


17/18

You have printed the following article:

The Allocation of Credit and Financial CollapseN. Gregory MankiwThe Quarterly Journal of Economics , Vol. 101, No. 3. (Aug., 1986), pp. 455-470.Stable URL:http://links.jstor.org/sici?sici=0033-5533%28198608%29101%3A3%3C455%3ATAOCAF%3E2.0.CO%3B2-U

This article references the following linked citations. If you are trying to access articles from anoff-campus location, you may be required to first logon via your library web site to access JSTOR. Pleasevisit your library's website or contact a librarian to learn about options for remote access to JSTOR.

[Footnotes]

1 The Market for "Lemons": Quality Uncertainty and the Market MechanismGeorge A. Akerlof The Quarterly Journal of Economics , Vol. 84, No. 3. (Aug., 1970), pp. 488-500.

Stable URL:http://links.jstor.org/sici?sici=0033-5533%28197008%2984%3A3%3C488%3ATMF%22QU%3E2.0.CO%3B2-6

1 Imperfect Information, Uncertainty, and Credit RationingDwight M. Jaffee; Thomas RussellThe Quarterly Journal of Economics , Vol. 90, No. 4. (Nov., 1976), pp. 651-666.Stable URL:http://links.jstor.org/sici?sici=0033-5533%28197611%2990%3A4%3C651%3AIIUACR%3E2.0.CO%3B2-K

1 Credit Rationing in Markets with Imperfect InformationJoseph E. Stiglitz; Andrew WeissThe American Economic Review , Vol. 71, No. 3. (Jun., 1981), pp. 393-410.Stable URL:http://links.jstor.org/sici?sici=0002-8282%28198106%2971%3A3%3C393%3ACRIMWI%3E2.0.CO%3B2-0

1 Nonmonetary Effects of the Financial Crisis in the Propagation of the Great DepressionBen S. BernankeThe American Economic Review , Vol. 73, No. 3. (Jun., 1983), pp. 257-276.Stable URL:http://links.jstor.org/sici?sici=0002-8282%28198306%2973%3A3%3C257%3ANEOTFC%3E2.0.CO%3B2-0

http://www.jstor.org

LINKED CITATIONS- Page 1 of 2 -

NOTE: The reference numbering from the original has been maintained in this citation list.
http://links.jstor.org/sici?sici=0033-5533%28198608%29101%3A3%3C455%3ATAOCAF%3E2.0.CO%3B2-U&origin=JSTOR-pdfhttp://links.jstor.org/sici?sici=0033-5533%28197008%2984%3A3%3C488%3ATMF%22QU%3E2.0.CO%3B2-6&origin=JSTOR-pdfhttp://links.jstor.org/sici?sici=0033-5533%28197611%2990%3A4%3C651%3AIIUACR%3E2.0.CO%3B2-K&origin=JSTOR-pdfhttp://links.jstor.org/sici?sici=0002-8282%28198106%2971%3A3%3C393%3ACRIMWI%3E2.0.CO%3B2-0&origin=JSTOR-pdfhttp://links.jstor.org/sici?sici=0002-8282%28198306%2973%3A3%3C257%3ANEOTFC%3E2.0.CO%3B2-0&origin=JSTOR-pdfhttp://links.jstor.org/sici?sici=0002-8282%28198306%2973%3A3%3C257%3ANEOTFC%3E2.0.CO%3B2-0&origin=JSTOR-pdfhttp://links.jstor.org/sici?sici=0002-8282%28198106%2971%3A3%3C393%3ACRIMWI%3E2.0.CO%3B2-0&origin=JSTOR-pdfhttp://links.jstor.org/sici?sici=0033-5533%28197611%2990%3A4%3C651%3AIIUACR%3E2.0.CO%3B2-K&origin=JSTOR-pdfhttp://links.jstor.org/sici?sici=0033-5533%28197008%2984%3A3%3C488%3ATMF%22QU%3E2.0.CO%3B2-6&origin=JSTOR-pdfhttp://links.jstor.org/sici?sici=0033-5533%28198608%29101%3A3%3C455%3ATAOCAF%3E2.0.CO%3B2-U&origin=JSTOR-pdf


18/18

References

The Market for "Lemons": Quality Uncertainty and the Market MechanismGeorge A. Akerlof The Quarterly Journal of Economics , Vol. 84, No. 3. (Aug., 1970), pp. 488-500.Stable URL:

http://links.jstor.org/sici?sici=0033-5533%28197008%2984%3A3%3C488%3ATMF%22QU%3E2.0.CO%3B2-6

Nonmonetary Effects of the Financial Crisis in the Propagation of the Great DepressionBen S. BernankeThe American Economic Review , Vol. 73, No. 3. (Jun., 1983), pp. 257-276.Stable URL:http://links.jstor.org/sici?sici=0002-8282%28198306%2973%3A3%3C257%3ANEOTFC%3E2.0.CO%3B2-0

Imperfect Information, Uncertainty, and Credit RationingDwight M. Jaffee; Thomas Russell

The Quarterly Journal of Economics , Vol. 90, No. 4. (Nov., 1976), pp. 651-666.Stable URL:http://links.jstor.org/sici?sici=0033-5533%28197611%2990%3A4%3C651%3AIIUACR%3E2.0.CO%3B2-K

Credit Rationing in Markets with Imperfect InformationJoseph E. Stiglitz; Andrew WeissThe American Economic Review , Vol. 71, No. 3. (Jun., 1981), pp. 393-410.Stable URL:http://links.jstor.org/sici?sici=0002-8282%28198106%2971%3A3%3C393%3ACRIMWI%3E2.0.CO%3B2-0

http://www.jstor.org

LINKED CITATIONS- Page 2 of 2 -

NOTE: The reference numbering from the original has been maintained in this citation list.
http://links.jstor.org/sici?sici=0033-5533%28197008%2984%3A3%3C488%3ATMF%22QU%3E2.0.CO%3B2-6&origin=JSTOR-pdfhttp://links.jstor.org/sici?sici=0002-8282%28198306%2973%3A3%3C257%3ANEOTFC%3E2.0.CO%3B2-0&origin=JSTOR-pdfhttp://links.jstor.org/sici?sici=0033-5533%28197611%2990%3A4%3C651%3AIIUACR%3E2.0.CO%3B2-K&origin=JSTOR-pdfhttp://links.jstor.org/sici?sici=0002-8282%28198106%2971%3A3%3C393%3ACRIMWI%3E2.0.CO%3B2-0&origin=JSTOR-pdfhttp://links.jstor.org/sici?sici=0002-8282%28198106%2971%3A3%3C393%3ACRIMWI%3E2.0.CO%3B2-0&origin=JSTOR-pdfhttp://links.jstor.org/sici?sici=0033-5533%28197611%2990%3A4%3C651%3AIIUACR%3E2.0.CO%3B2-K&origin=JSTOR-pdfhttp://links.jstor.org/sici?sici=0002-8282%28198306%2973%3A3%3C257%3ANEOTFC%3E2.0.CO%3B2-0&origin=JSTOR-pdfhttp://links.jstor.org/sici?sici=0033-5533%28197008%2984%3A3%3C488%3ATMF%22QU%3E2.0.CO%3B2-6&origin=JSTOR-pdf

allocation of credit

Documents