The Loan Sales Market Part 2

-24Н. Loan Sales and the Larger Context The rise of the loan sales market should not Ье seen in isolation. Banking is experiencing other changes perhaps rooted in the same transformation of the underlying technology. Particular manifestations include the growth in off-balance sheet activity and asset securitization. If loan sales represent а new contract form embodying traditional bank services, then the growth of off-balance sheet activity may represent а related phenomenon. Examples of off-balance sheet activity include loan commitments, standby letters of credit, interest rate swaps, and financial futures. While most of the off balance sheet activities have grown in volume попе have grown as fast as loan sales. Standby letters of credit have not grown nearly as fast as loan sales, growing only 36 percent between December 1983 and December 1987, and having actually declined in volume since December 1985's peak of $156.8 billion. 17 Anotheг off-balance sheet item, loan commitments, have а higher outstanding volume than loan sales, but also have not gгown as fast. Commitments have increased fгom $431 billion in Decembeг 1983 to $612 billion in December 1987. the explosive gгowth. Thus, loan sales stand out because of Inteгest гаtе swaps, howeveг, have grown vегу fast. The volume of swaps surged from $186 billion in 1985 to $715 billion in Decembeг 1987. III. Contгact Foгm and Legal Issues Secondaгy loan paгticipations, whetheг they аге straightforward sal es, ог in the foгm of "stгips'' ог "bids,'' change the bank's asset-liability stгuctuгe from the standard deposit-backed loan. Diffeгent forms of contracts control the rel ationship between the bank, the borrower, and the lender. The laws controlling each side of the participation differ from each other and from those governing standard loan contracts.18 Furthermore, since laws and -25regulations control the banks' ability to remove loans from the balance sheet, the contгacting process and form inherently involves such considerations. These contractual features and theiг inteгpretations also determine bank incentives to maintain the value of loans and fulfill theiг fudiciary oЫigations. This section defines the different participation contracts and then гeviews the legal, accounting, and гegulatoгy aspects of these contгacts. А. Participations Loan paгticipations offeг а majoг contгast with the tгaditional bank loan. In the standaгd case, th e bank makes а loan, and thus has а claim on а firm. Funding comes from depositors, who then have а debt claim on the bank, which bears the credit risk. The generic participation, in contrast, puts the credit risk on the ultimate lender, the participation buyer. The bank makes а loan to th e firm, which it sells, i.e., promises its payment stream to the Ьuуег. Thus а contract between bank and buyer replaces the deposit contract, and though th e bank remains responsiЫe to the borrowing firm, and collects payments, oveгsees the collateral and enforces financial ratios, it passes the payments on to the buyer (who usually gets а сору of the original loan agreement). Th e buyer does not hold а claim against the bank. Participations come in th ree basic types: paгticipations. 19 Th ese types differ in the degree to which th ey transfer rights and oЫigations. Assignments shift th e direct debtor-creditor relationship from the bank to the loan buyer. to the buyer. assignments, novations, and They transfer the bank's rights The assignee gets rights not only to the payment stream, but some rights to take action directly against the borrowing firm. Still, an assignment does not pass on the bank's oЫigations towards the borrower (such as loan commitments) or any advantages potentially dependent оп individual -26bank characteristics (such as tax advantages). An assignment, thus, divides the original contract into two parts; no new contract is created. А novation completely transfers all rights and oЫigations of the bank to the participation buyer. actor. The bank leaves the picture entirely, replaced Ьу а new Such а drastic shift needs the consent of the borrower unless otherwise explicitly written into the contract. А participation gives the right to receive (in whole or part) future payments to Ье made according to the underlying loan contract. accomplished Ьу creating а new contract. This is The participation contract exists only between the bank and the buyer, and thus the buyer has no rights against the firm, which may not even know а participation has been sold. Thus, а participation does not transfer any of the original rights and oЬligations of the original contract. The participation buyer must rely upon the bank to perform on the original contract, even though the bank may not now have any stake in the outcome. Should the firm enter bankruptcy, the original lender must file all claims. When the borrowing firm is notified that the loan has Ьееп sold, it is а disclosed participation, otherwise the participation is silent. Banks have used participations for а long time, but the sort most prevalent now differs from the more traditional type.20 The older sort of participation, sometimes better described as а syndication, usually involved а lead bank negotiating for other banks. The contract usually took the form of an assignment, though often each participating bank would make а separate loan to the firm, governed Ьу the agreement negotiated Ьу the lead. Sometimes the participants would lend directly to the lead bank, under а correspondent relationship. Often "overlining" motivated sales when lack of funds or lending limits precluded further lending. -27Recent participations, or loan sales, appear to most commonly take the form of а straight sale of the cash flow emanating from а l oan the bank has originated. "strips." Sales of short-term loans generally take the form of "bids" or Bids are short-term loan commitment lines (sometimes called uncommitted bid lines). When the firm takes down а short-term loan (three to 90 days) the bank sells the loan. Often banks enter competing bids on all or part of the loan, resul ting in the competitive bid option discussed earlier. In some cases, the competition takes place via an explicit auction. Strips present а way to transform longer maturity loans into а series of more salaЫe short maturity loans. Banks sell shorter maturity segments, or strips, of longer maturity loans. Recently, some competitive bid options also have taken the form of revolving credit agreements, with banks competing for each segment of the longer loan. The participation agreement itself may Ье rather involved, so often partners use а master agreement, а document agreed to beforehand, that governs the buyers and sellers of participations, both bids and strips. The disposition of any collateral depends on the specifics of the loan (sometimes the buyer must take steps to perfect, sometimes not), but this is rarely important. As noted above, the resal e market for participations is thin. Some sales (mainly to subsidiaries) occur, but to avoid proЫems with the Glass-Steagall act, banks often insert clauses saying participations are not bought for resale purposes. This simple description of the basic participation conceals some subtle contractual points. The seller (or l ead bank), as l ender of record, may have oЫigations to the participation buyer, in particular to oversee the loan: col lecting and distributing payments, monitoring col lateral and financial ratios, enforcing loan covenants, and holding copies of the relevant -28documents. But these oЬligations are usually implicit since participation contracts usually involve по recourse against the selling bank, and do not transfer rights against the underlying borrower. Reliance оп the selling bank to perform сап sometimes depend оп the bank's still having а stake in the underlying loan, either because only part of the loan has been sold or because the maturity of the underlying loan is longer than the maturity of the participation. Thus, even if legally the bank need not perforrn, the incentives to maintain the value of the loan still exist if the bank's involvernent with the customer is, in the above sense, larger than the participation. В. Accounting Issues Banks which sell loans obtain maximum benefit if they сап remove the loans from their balance sheet, increasing their return оп assets, return оп equity, obtaining fee income and by-passing capital requirements. asset sales must meet two basic criteria: То d o this, (i) the loan participation and the original loan must have substantially identical terms and conditions (except for а servicing fee retained Ьу the seller); (ii) the loan participation must Ье without recourse to the transferor. The first criterion means, for example, that interest payments to the seller and the participant must occur at the same time, and а loan originally contracted оп а fixed rate basis must Ье sold оп а fixed rate basis. More importantly loans сап only Ье removed from the balance sheet if sold to maturity, so banks must match the maturity of the participation to that of the loan. These criteria are stronger than those suggested Ьу the Financial Accounting Standards Board (FASB) in Circular 77. That allowed banks to count loans as off-balance-sheet if the bank sold control of benefits, could estimate its oЫigations under recourse, and could not repurchase the asset. -29The Federal Regulatory Agencies, however, felt that recourse implied banks still b ore credit risk for the asset and thus from а regulatory perspective had not shed the loan. Thus, since banks are subject to these stricter Regulatory Accounting Practices (RAP), only non-recourse loan sales leave the b ooks. However, bank holding companies are subject to the FASB's Generally Accepted Accounting Practices (GAAP) and some loans with recourse will not appear on their balance sheet, even though they will appear on the balance sheet of the b ank they own.20 For longer-term loans, the interpretation and application of the accounting principles presents few proЫems. For the strips and bids, however, where sales of а short-term loan under а long-term credit commitment takes place, questions have been more complex. In January 1988 the Federal Accounting Standards Board ruled that loan strips could Ье recorded as sales only if the buyer of the strip assumes the full risk of loss and the lender has no contractual oЫigation to repurchase the strip. Controversy over whether well most strips satisfied these requirements led to further conceptual refinement. Thus, the Emerging Issues Task Force of the American Institute of Certified PuЫic Accounts allows the sale of а short-term loan under а long-term credit commitment to count as а sale if the commitment includes substantive covenants. The Board of Governors of the Federal Reserve System disagrees, and now forces b anks to keep strips on their b alance sheet for regulatory purposes. As the nature of the discussion involves the restructuring of bank loans to make them easier to sell, we provide more details in section Е, below. Banks can benefit from loan sales with recourse, especially if this creates а non-deposit liab ility not subject to reserve requirements. example, the bank may sell loans without direct recourse, but hire an For -30insurance company to guarantee payments to the purchasers. reimburses the insurance company for any losses that incur. The bank then This contingent liability must appear on the books, but it is not а deposit, being only indirectly related to the loans. Strips create their own set of ргоЫеms. Because а loan must Ье sold to maturity in order to Ье removed from the balance sheet, selling pieces of а long-term loan will not do. Instead, as described below in more detail, а series of short-term loans is negotiatied to replace the single long-term loan ог commitment. Each "advance" under the agreement is sold off to its full maturity (usually 90 days), meeting the regulatory requirements. Each advance must Ье а true refunding, so not only may interest rates change, but credit analysis may dictate that the loan not get renewed. Conversely, the borrower may discharge collateral at the end of each advance (though this may hurt the chances for renewal). As mentioned above, the bank must convince the borrowing firm that the advances are quite close to а long-term loan, but must also convince the regulators that each advance constitutes а separate agreement, with no commitment to continue lending. С. Loan Sales and Securities Laws Banks, already subject to several regulatory authorities, rarely seek further regulation. Yet to the extent that secondary participations are securities, an entirely different set of regulations applies.21 As the loan sales market develops, the instruments begin to look at least superficially, like securities, and questions about the appropriate regulations become тоге common. The criteria for determining whether or not а participation is а security also introduces economic questions about the purpose and functions of banks in participations. The relevant law defines securities quite broadly. The Securities Act of -311933 says that: . . . unless the context otherwise requires . . . the term security means any note, stock, treasury stock, bond, debenture, evidence of indebtedness, certificate of interest or participation in any profit-sharing agreement, collateral­ trust certificate, reorganization certificate or subscription, transferaЫe share, investment contract, voting-trust certificate, certificate of deposit for а security, fractional undivided interest in oil gas, or other mineral rights, any put, call, straddle, option, or privile ge on any security, certificate of deposit, or group or index of securities, including any interest therein or based on the value thereof, . . . or, in general, any interest or instrument commonly known as а "security" or any certificate of interest or participation in, temporary or interim certificate for, receipt for, guarantee of, а warrant or right to subscribe to, or purchase, any of the foregoing. Any instrument deemed а "security" is then subject to the securities laws, which for а bank may prove disadvantageous. The Securities Act of 1933 requires issuers to register and disclose information about their security. The Security Exchange Act of 1934 provides for even stricter disclosure. The anti-fraud provisions of the 1933 Act and Section 10Ь-5 of the 1934 Act hold the issuer liaЫe for fraud not only on factual misrepresentation (part of common law), but on factual omissions and half-truths as well. These statutes also make fraud easier to prove than under common law, dispensing with some concepts such as intent (or more technically, scienter). may Ье tried in Federal court.22 Finally, such cases Sellers of participations have two possiЫe ways of exempting themselves from registration. The first uses the "private placement" exemption. Securities not offered to the puЫic and sold only to sophisticated investors need not Ье registered, though they remain subject to the anti-fraud provisions. Another possiЫe exemption, currently unlikely, is for securities of maturity under nine months. The courts have held that this applies only to commercial paper, but this may change as participations more closely resemЫe commercial paper. -32А key question, if the growth of loan sales continues, is likely to Ье whetheг these contracts constitute legally defined secuгities. The courts have used fouг tests to decide if something is а security and thus subject to the security laws. These tests often hinge on the economic purpose of а tгansaction, raising economic questions and providing scope for policy advice. The basic test, known as the Howey/Forman test, considers the "economic realities" of the object involved and sets forth four criteria for а secuгity. According to Scholl and Weaveг (1982), "the Howey test defines а secuгity as (1) the pгesence of an investment (2) in а common ventuгe (3) pгemised on а геаsоnаЫе expectation of pгofits (4) to Ье derived from the entгepreneurial ог managerial efforts of others." On the surface, а loan participation арреагs to meet these criteria, but the woгds have acquired а legal meaning that in general excludes participations. speculative," do not count as an investment. appгeciation does not count as profit. Loans, being "non­ Interest without cap ital А bank's routine collection and overseeing duties do not count as managerial effort, especially if the participation Ьuуег acquires rights to demand foreclosure or to veto changes in the loan terms. Several other tests focus on different aspects of the Howey/Forman approach. The Commercial/Investment test starts from the premise that commercial loans are not under the jurisdiction of the securities laws, and so exempts anything that looks sufficiently similar. Since participated loans usually stem from а prior relationship with the bank, are used for working capital, рау а fixed interest rate and use standard bank forms, they fall into the commercial category. The Risk Capital test similarly asks whether the investment is used to finance "risk capital" on which the repayment will Ье very dependent on the firm's entrepreneurial efforts. Even though loans may -33Ье risky, they have been judged to lack this speculative character. Finally, the Weaver test exarnines the context required Ьу the securities laws and holds that activities sufficiently regulated and protected Ьу banking laws and authorities do not need additional protection from the securities laws. Overall, court cases have split on whether ог not secondary loan participations аге securities, with most of the recent cases upholding the banks' view. This may change as banks rnake rnore loans specifically with the intent of sale and otherwise introduce contracts that make the participations look тоге and тоге like securities. In the end, though, the costs of complying with security regulations may not Ье onerous, especially for the large banks who аге the primary issuers, and the advantages of participations may make them profitaЫe even if subject to SEC regulation. Furthermore, as participations look тоге like securities, the accounting standards fог them converge toward the stricter requirements fог securities. The Federal Accounting Standards Воагd has recently pгoposed incгeased disclosure about all financial instruments, requiring data on credit risk, future payments, interest exposure and estimated market value.23 Though currently only а pгoposal (and one opposed Ьу many banks) these tougher reporting requirements would decrease the paperwork advantage of participations. D. Contract Structure and Language Sarnple contracts obtained from money center banks confirm the points made earlier in this section. further. In this subsection these contracts аге analyzed Some interesting details emerge in the агеаs of legal liability and recourse, information flow, and tailoring agreements for customers. Buyers of paгticipations might wоггу that since а bank has no stake in the loan sold, they will receive low quality loans ог loans where the bank has not thoroughly researched the credit risk of the borrower. On the other hand, -35participation buyer, either because these actions cannot Ье specified because they are too complicated, or because the bank does not want to Ье so encumbered. Indeed, rating agencies, like Moody's and Standard and Poor's, have begun to offer ratings on loans up for sale. This suggests that buyers desire more information than provided Ьу the seller.24 On top of all this, The Bank shall not, in the absence of gross negligence or willful rnisconduct, Ье under any liability to the Participant with respect to anything which it may do or refrain from doing which may seem to the Bank to Ье necessary or desiraЫe. In general, without explicitly specifying the rights and duties in the participation contract, most claims Ьу the buyer against the selling bank have not been recognized Ьу the courts (Knight, 1987). All the secondary participation contracts we examin ed contain features similar to those quoted above. Paradoxically, а market exists which at first glance looks like а "lemons market." The bank guarantees nothin g; cannot Ье forced to act against the borrower; may, in fact, have по incentive to act in the absence of such oЫigations, and yet, provides по recourse. While these features are common, there are differen ces in how the underlying contract with the borrowers is being restructured to satisfy the restrictions required for removing the loans from the balance sheet. Banks are particularly sensitive about this because of competition, and we were less successful in obtaining information. Е. Restructuring Underlying Loans The delicate issue of restructuring the underlying loan with the borrower is necessitated Ьу the requirements for removing the loan from the balance sheet. Loans must Ье sold to maturity. In prin ciple, this mean s that, for committed facilities, the loan advances must Ье sold out to the termination of the credit agreement. But, participation buyers typically want shorter maturities. The long loan сап Ье restructured Ьу defining maturity as -36corresponding to the interest period. short loans. Thus, the long loan b ecomes а series of Practically, this means convincing the borrower to accept а sequence of short loans as being equivalent to а long loan, and convincing the participation buyer that the bank faces the right incentives to maintain the value of the loan because the bank has а long-term commitment, even if legally there only exists an implicit sequence of short-term commitments. However, this arrangement requires the decision to refund to Ье nontrivial; it cannot Ье an automatic rollover. The underlying restructured agreement requires the b orrower to satisfy certain conditions before the bank is oЫigated to advance new funds. If these conditions were not present, so that rollover was automatic, then the bank could Ье forced to lend at а time when the participant wanted to Ье paid off. For example, the participant might want out if there is а deterioration of the b orrower's condition. In this scenario, the bank' s funds would Ье used, in effect, to рау. off the participant. The participant would have а put option, а kind of recourse. In order to prevent such recourse, the decision to refund the loan must Ье nontrivial. If there is а deterioration of creditworthiness, then there must Ье conditions which would prevent the borrower from positively affirming the required clauses. The b ank would then not Ье oЫigated to refund. conditions spoken of ab ove correspond to the loan covenants. The As the quality of credit of the particular b orrower lessens, from ААА downwards, more such restrictions or "triggers'' аге included. How does this restructuring differ from the longer maturity agreement? The longer maturity agreement would have contained the same covenants and the same enforcement mechanism. That is, the takedown schedule would have Ьееп tied to performance of the covenants, explicitly ог implicitly. Under the longer contract, the bank is committed for the life of the contract. Under -37the sequence of shorter contracts, the same covenants rnust Ье enforced, in substantively the same way. Sirnilarly, under both forrns, the bank rnay have incentives to waive the restrictive covenants and continue funding even though adverse changes have occurred. lender's situation changes. А significant difference rnay occur if the Thus, buyers of strips are under no oЫigation to continue buying, and changes in interest rates, economic conditions, or portfolio decisions тау lead thern to quit part way through the revolving credit agreernent. This would put the burden of loan takedowns squarely on the bank originating the strip. Without any material change in the borrower's status, it seems unlikely that the bank could refuse the loan, even though with а reduced distribution network it rnay wish to. These potential liability funding risks for long-terrn commitrnents are at the root of claims that strips are not true sales. The Board of Governors of the Federal Resserve Systern has revised the Call Report, indicating that strips are not off-balance sheet, because it believes that the risk of having to refinance the loan creates recourse (even though the strips are made under а contract which explcicitly states that there is no recourse). regulation along these lines has stunted the strip market. Fear of Banks have switched to competitive bid options, which do not seem to have this implicit liability, though perhaps only because regulators have not had tirne to formulate а position.25 The key question then is: to what extent is explicit maturity required to insure bank performance in а long-term relationship? IV. Loan Sales and the Functions of Banking Loan sales raise very basic issues about the purpose and uniqueness of banks, about the relations between banking and capital markets, and about the forrn that banking will take in the future. With loan sales it appears that banking is changing in а fundamental way. In this section we briefly discuss -38the implications of loan sales for traditional banking roles and methods. The existence of banks depends upon the banking industry playing а unique role in the savings-investment process that cannot Ье filled Ьу markets. Recent theoretical work in banking identifies several unique services provided Ьу banks. One explanation is that banks produce information and screen out borrowers with poor projects to prevent а "lemons market" proЫem. An alternative is that banks monitor borrowers Ьу enforcing particular management actions and preventing other actions. In а securities maгket, no one peгson would want to Ьеаг the cost of pгoducing information about prospective investments ог the cost of monitoгing the fiгm because all other investors would fгee-гide on their effoгts. If information pгoduction or monitoгing could Ье done pгivately then theгe would Ье needless, and costly, duplication. firms.26 Banks aгise then as information pгoducers and as monitoгing An essential component of the гationales fог bank existence conceгns how bank depositoгs pгod uce information about banks and how depositoгs monitoг the actions of their banks. The depositors аге assuгed of the bank's performance in pгod ucing information and monitoгing pгojects because the bank owners, holding а junioг claim on the value of the bank, pгofit only after the debt holdeгs get paict.27 This aггangement is crediЬle only if the owners put up а sufficiently laгge equity stake as гesidual claimants. NotaЫy these aгguments гequire that the bank hold the loan until matuгity because only in that way will bank equity Ье at гisk, insuгing prod uction of accuгate information and monitoring. If banks could sell loans, then theгe would seem to Ье no reason for them to expend resouгces producing information and monitoгing. These roles cannot Ье fulfilled Ьу market participants, yet they buy the loans. Paradoxically, -39loan buyers know that once the bank has sold the loan, the incentives for the bank to perform are по longer present (at least in the same way), yet they still buy the loans. Thus, loan sales appear to contradict the very essence of banking. The growth of the loan sales rnarket rneans either that there is а declining need for the special services provided Ьу banks or that there is а new mechanism for insuring that banks provide such services. It may Ье that the costs to market participants of producing information and monitoring loans has fallen so much that there is по need for banks to provide these services at previous levels. Alternatively, it may Ье that banks аге continuing to provide information production and monitoring services, but ensuring their performance по longer requires risking their equity Ьу holding the loans until Which of these hypotheses is true was studied Ьу Gorton and maturity. Pennacchi (1988). Gorton and Pennacchi (1988) investigated whether the prices of loans sold contained а risk premium for the default of the selling bank. The basic idea was to test for the presence of implicit guarantees of the value of the loans sold Ьу the selling banks. follows. The possibility of implicit guarantees arises as In order to avoid securities laws banks prohibit the resale of loan sales contracts.28 Thus, there is по secondary market for loan participations. Since this makes loan sales contracts highly illiquid, banks informally offer to buy loans back before the maturity of the loan sales contract.29 The question is whether this informal offer is essentially а put option provided to the loan buyer Ьу the bank. If this put option exists then the bank's performance is ensured because there, in fact, exists а contingent liability associated with each loan sold. If the put option exists, then loan buyers would Ье concerned with the -40pгospects that the loan selling bank might default. pгiced to гeflect such pгospects. The put option would Ье Consequently, when а loan is sold the pгesence of а гisk pгemium fог the default of the selling bank would Ье the price of such а put option. If по such risk premia exist, so that the default pгospects of the selling bank аге iггelevant and only the undeгlying borrower's default risk is priced, then loan buyers must Ье doing their own infoгmation production and monitoring. Goгton and Pennacchi find weak evidence of bank гisk pгernia in loan sales pгices, suggesting that banks аге continuing to pгovide seгvices which market paгticipants cannot pгovide. The bank continues to peгform these seгvices because if it does not then the bank will Ье forced to buy the loan back. While Goгton and Pennacchi found some evidence of the pгesence of irnplicit guaгantees of bank peгfoгmance, they did not explain what enfoгces this aггangement. Why does the bank abide Ьу such an irnplicit aггangement, especially when the loan sale contгact explicitly states that the buyer has no гесоuгsе? One answeг is that th e bank's гeputation is at stake. Gorton and Haubrich (1988) investigate th e development of гeputations in the loans sales maгket as mechanisms for ensuring bank performance. Th e possibility of substituting а гeputation for bank equity depends upon а tech nological change in the transmission of information. Assuming that loan sales are feasiЫe for some reason, another important question concerns wh y banks h ave only recently started to make irnplicitly guaranteed loans sales. Is the motivation due to changes in underlying information production and monitoring technologies which competition has forced banks to adopt? Or is th e motivation due �о regulatory changes. the motivation some combination? Or is Pennacchi (1988), Flannery (1988), Greenbaum and Th akor (1987), and James (1988) all present motives for loan sales. For -41the rnost part the rnotives are regulatory based. V. Conclusion First, а раrаЫе frorn the dark ages of loan participations. А big New York bank, it was said, held а one-rnillion dollar participation оп а loan to а lease broker. When the borrower failed to perforrn, the big New York bank sent sorneone down to 0klahorna City to find out what was going оп. "What happened to the leases?" the banker asked. "I never bought'ern," said the borrower. "Wait а rninute. What do you rnean you never bought thern? Оuг agгeernent with Penn Square says that that loan is secured Ьу those leases, the unlirnited peгsonal guarantee of the boгrower, and otheг peгsonal propeгty, including а 1980 0ldsrnobile Toronado. What happened to your unlirnited guaгantee?" "I 'm Ьгоkе." "What happened to the rnoney?" "It got spent." "'It got spent!' You rnean to tel1 rne I've got а rnillion bucks tied up in а 1980 Toгonado?" "Hell, по. I sold that саг last уеаг. You haven't got а goddarn thing." [Mark Singeг, Funny Money, рр. 144-5.] This stoгy encapsulates many points about loan participations. The oil lease brokeг looks like а poor cгedit risk (it's doubtful he had access to the commeгcial рарег maгket), and no one monitored his collateral. Despite these ргоЫеms, Penn Squaгe sold the loan upstream, rneaning someone bought it. This admittedly extreme story illustrates the dangers of loan sales, namely, the seller сап Ье selling overvalued loans to unsuspecting buyers. PresumaЫy, this fеаг of buying "lemons" is one of the гeasons that rnarkets fог bank loans did not develop earlier. But, when а rnarket develops it rneans that such ргоЫеms have been overcome. Some rnechanism has arisen which allows buyers to reliaЫy assess the quality of the sellers' product. Buyers avoid the lemons. The loan sales rnarket has gгown to such an extent that it seems hard to -42гeconcile the fеаг of а lemons market with the existing reality of $240 billion of loan sales. Indeed, based soley оп the observed gгowth to date, there seems по reason to believe that such growth could not continue, douЫing in а short period of time. As the market has grown the loans being sold increasingly represent claims оп assets which are harder to value because the underlying borrowers are not puЫicly traded companies. has grown Also, as the market maturities of loan sales contracts have increased. loan sales con tracts have become increasingly risky. any market develops it develops in this way. In other words, Intuitively, whenever At first assets are sold which are not likely to Ье misvalued and then, in creasingly, тоге complicated, riskier contracts are made. Unfortunately, this process is not well understood. The implication s of loan sales for banking are so far reaching that it is hard to predict the outcome until loan sales are understood better. Certainly, the very idea that ban k loans could Ье sold in such quantities, un der such contracts, an d even outside the banking system, would have been inconceivaЫe on ly а shoгt time ago. The vегу idea contradicts the notion of а bank as an intermediary which creates securities and holds them until maturity as distinct from simply brokering claims issued Ьу firms. Loan sales suggest that the ability of banks or other firms to produce in formation and monitor borrowers has changed. The link between the production of these services and the cгeation of а loan which must Ье held until maturity Ьу the firm producing the services has, apparently, been severed, ог at least altered fundamentally. At гооt, the technology fог producing information and for monitoring has been transformed. The marketability of bank loans has significant regulatory implications since the vегу conception of what constitutes а bank changes. Previously, а -43bank was an institution which was essentially defined as а firm which financed nonmarketaЫe loans with deьt which could Ье "put" to the bank at any time. This combination created the historically incendiary possibility of banking panics since depositors could run the banking system and the banking system had no way to convince them of the value of its assets because the assets were not marketaЫe. But, looking ahead, this would no longer Ье the case if loans сап Ье sold in significant quantities. Since putaЫe deьt no longer would finance nonmarketaЫe assets, the possiЫlity of panics would Ье reduced if not eliminated. The very reason for deposit insurance and bank regulation would disappear. There are also important implications for bankers. With а well-developed loan sales market banks сап easily hold well-diversified portfolios, regardless of what comparative advantage they have in originating loans. Thus, it should not Ье а surprise that many of the largest loan sellers are also the largest loan buyers. Without such а market, small banks have limited prospects for creating diversified loan portfolios. Aside from the possiЫlity of holding diversified portfolios of loans, banks' basic functioning changes with loan sales since such sales allow the return to banks from information production and monitoring to Ье separated from the actual security itself. Loans сап Ье created and sold, earning а fee, without having to Ье held until maturity. For individual banks the strategic question arises of how best to facilitate this process. А large number of questions remain at the loan sales market continues to evolve. What will bank contracts of the future look like? Will the contracts change as this market develops? Will the underlying contracts between banks and borrowers change further? Will secondary markets for loan sales contracts develop? Is it possiЫe to sell all loans? If not, what are the limits of -44- the loan sales market?