FRBSF WEEKLY LETTER August 31, 1990 Moral Hazard in Payments Systems

Payments systems enable financial institutions Fedwire, where transfers are final at the time of and their customers to exchange funds and effect notification, this practice puts the Federal Re­ transactions quickly and efficiently. For high­ serve in the position of granting intra-day credit value transactions, there are two main payments ("daylight overdrafts"), which, in turn, exposes systems in the u.s.: Fedwire, managed by the the Fed to the risk that the sending bank could Federal Reserve System, and CHIPS (Clearing­ fail before the end of the day and not cover its house Interbank Payments System), managed by overdraft. On CHIPS, in theory, no analogous the New York Clearinghouse Association, a group credit is extended, since all payments are pro­ of private banks. "Large dollar" payments sys­ visional until settlement at the end of the day. tems can entail considerable systemic risk (that is, the risk that the failure of one transaction At the same time, however, it is common prac­ could trigger the failure of many other, related tice on all large-dollar systems for participants transactions). The success of efforts to control to make transferred funds available to recipients systemic risk depends on how effectively such soon after notification of payment. To the extent measures control the "moral hazard" that is at that CHIPS participants are doing this, they are the heart of systemic risk. The story of how sys­ making funds available before settlement, and temic risk arises, and how effectively moral are thus exposing themselves to credit risk. For hazard is controlled on Fedwire and CHIPS, example, by allowing recipients to withdraw are the topics of this Letter. funds from their account before settlement, CHIPS participants face the risk that in the Risks in payments systems event good funds to cover the withdrawals are Transactions initiated through a payments system not forthcoming, they will not be able to recover typically involve two steps: clearing and settle­ the funds from their customers. ment. Once the payor (which may be a bank or a bank's customer) initiates a transaction, the clear­ On systems such as CHIPS, where there is some inghouse or other system manager debits the sys­ risk that initiated transactions will not be settled, tem account of the payor's bank and credits the there also exists a "systemic risk" that one par­ system account of the recipient's bank; this is the ticipant's settlement default will trigger a chain clearing phase of the transaction. In the settle­ reaction, resulting in many interdependent trans­ ment phase, the payment is made "final" through actors being unable to settle. Systemic risk, as the transfer of "good" funds, which, in the U.s., distinct from the credit risk faced by individual are reserve account balances at the Federal participants, exists because of a "moral hazard" Reserve. problem; namely, individual participants do not bear all the risk that their transactions generate. On Fedwire, settlement occurs prior to clearing, The costs of systemic risk include both the costs in the sense that the Federal Reserve credits good of the increase in credit risk faced by interdepen­ funds to the receiving bank's reserve account at dent participants and the transactions costs of the time that the bank receives notification of "unwinding" defaulted payments, which, for a payment, but it does not debit the sending bank's major default, might entail substantial disruption reserve account until the end of the day. On in financial markets. CHIPS, in contrast, clearing occurs prior to settle­ ment; all payments are "provisional" until good Controlling systemic risk funds are transferred at the end of the day. To control systemic risk on CHIPS, beginning this fall; participants will be required collectively to Both Fedwire and CHIPS permit participating guarantee transactions on this system by posting banks to send payments during the day that ex­ a total of about $4 billion in collateral to cover ceed the balance in their reserve account. On defaults by system participants. The collateral FRBSF

fund technically should eliminate most of the risk. Moreover, the Federal Reserve is proposing systemic risk on CHIPS, since all but the largest that beginning in mid-1991, banks will be re­ defaults would be covered, thereby short-circuit­ quired to pay 0.25 percent interest (annual rate) ing a chain of defaults among many participants. on the amount of their average daily intra-day Fedwire overdrafts that exceeds 10 percent of On Fedwire, the Federal Reserve's offer of their risk-based capital. Such a fee presumably immediate payment finality precludes systemic woula1 Igive • reawlrer- I' paruclpams.,.. an even greater. risk. By granting immediate credits to receiving incentive to reduce their daylight overdrafts and banks' accounts and delaying debits from send­ thereby the risk exposure of the Federal Reserve. ing banks' accounts, there is no risk to parti­ cipants that a transaction will not settle; the However, the fee most likely is too far below the Federal Reserve System takes all of this risk upon true time value of money to be effective. In com­ itself. When defaults occur, the Federal Reserve parison, the interest rate on reserves banks lend usually provides the liquidity to cover them. The one another overhight in the federal funds market most well known example of this involved the currently is around eight percent. Moreover, the liquidity crisis of Continental Illinois Bank in overdraft fee is tied neither to an institution's 1984. credit-worthiness nor to the riskiness of the par­ ticular transaction. This further diminishes the Moral hazard fee's effectiveness in inducing desired changes However, when credit risk borne by the bank in banks' behavior. receiving payment is systematically reduced, banks may lose any incentive to control their Thus, a higher, risk-adjusted fee may be appro­ credit risk exposure to other payments system priate, even though some observers worry that a participants. Because each individual bank's risk high overdraft fee would induce Fedwire partici­ is perceived to be lower, its use of the payments pants to try to synchronize payment inflows and system and exposure to other participants will outflows, thereby obstructing the smooth and increase. Thus, the very measures that are efficient operation of Fedwire. If the overdraft designed to reduce systemic risk in payments fee were in fact so high that banks would prefer systems also may exacerbate the moral hazard to delay transactions, we would expect the emer­ that under! ies system ic risk. gence of a private intra-day market to compete with the F~deral Reserve Banks in providing intra­ Consequently, on CHIPS, it is possible that the day good funds credit. Such a market could be new collateral fund would not be effective in expected to set risk-based prices for credit, reducing risk, were it not for additional measures which, in itself, would be beneficial in reducing that will be put in place, as discussed below. systemic risk. Likewise, Fedwire officials have taken a number of steps to control risk, recognizing that the CHIPS' loss-sharing rule Federal Reserve's provision of daylight overdrafts On CHIPS, the proposal to reduce systemic risk and guarantee of payments finality otherwise also attempts to limit moral hazard. As discussed would tend to aggravate moral hazard. earlier, CHIPS is proposing to require each par­ ticipant to post collateral to fund a $4 billion Reducing moral hazard pool that would guarantee settlement by cover­ Both Fedwire and CHIPS are attempting to ing virtually all defaults. The amount of collateral reduce moral hazard, and thereby both systemic required of each participant will be determined risk and the risk exposure of the respective guar­ by a loss-sharing formula, up to a predetermined antors, the Federal Reserve in the case of Fedwire maximum. This loss-sharing formula stipulates and the contributors to the collateral fund in the that a given participant's obligation to cover an­ case of CHIPS. other's shortfall will be higher the higher is its allowable credit exposure to that participant. At present, Fedwire policy requires that partici­ pants set sender "net debit caps" limiting the net Each CHIPS participant is required to establish amount that they can send in excess of what they a net credit limit on its transactions with every receive. These caps limit daylight overdrafts and other participant. These "bilateral credit limits" therefore the Federal Reserve's exposure to credit determine the maximum amount of payments outstanding that a participant is willing to accept of existing private clearing and settlement sys­ from another, net of the payments it has outstand­ tems, which face systemic risk problems, sug­ ing to that other participant. gests that this concern may be unfounded, however. The proposed collateral fund for CHIPS, The CHIPS loss-sharing formula thus draws a for example, is one way to control systemicrisk. direct link between each bank's self-determined maximum exposure to a given counterparty and In addition, the private clearinghouses for the its liability for losses due to the default of that futures, options, and stock markets, such as the counterparty. In this way, the rule should give Options Clearing Corporation, effectively control banks an incentive to limit their potential bilat­ systemic risk by legally assuming the obligation eral credit risk exposure and thereby their poten­ of guaranteeing the execution of each trade. tial contribution to systemic risk. !n addition, it Guaranteeing the execution of every trade, it is should give receiving banks an incentive to set true, aggravates moral hazard, but these clearing­ lower credit limits on their transactions with less houses have taken a number of steps to reduce credit-worthy sending banks. This approach the probability that a member will default. For could prove highly effective in controlling sys­ example, they set capital requirements, position temic risk and moral hazard, provided partici­ limits, and other financial standards for mem­ pants have timely access to accurate information bers, they collect margin payments (a kind of regarding other participants' behavior. collateral), and they continuously monitor the financial strength of and portfolio positions A private market? taken by member firms. Some have argued that the best way to control the risks involved in large-dollar payments sys­ For the future••• tems is to allow a private market for intra-day The steps proposed by Fedwire and CHIPS to funds to develop. Such an interbank market, the limit systemic risk and control moral hazard will federal funds market, already operates for over­ strengthen the U.S. financial system. The pro­ night funds. The primary reason it does not pro­ posed overdraft fee for Fedwire is a first step in vide an intra-day credit facility is that the Federal reducing the potential liability of the Federal Reserve has been providing unpriced intra-day Reserve for losses due to Fedwire participant credit on Fedwire. defaults. However, more extensive efforts to control moral hazard on Fedwire may be called In a private market, credit would be priced for. In particular, risk-based overdraft fees and the according to the credit-worthiness of the bor­ type of loss-sharing rule proposed for CHIPS are rower, thereby providing incentives to control indicative of the kinds of practices that could be risk. Critics of such a proposal contend that the considered for Fedwire in the future. risk-based pricing that would arise in a private intra-day credit market would not adequately Elizabeth Laderman take into account systemic risk. The experience Economist

Opinions expressed in this newsletter do not necessarily refiect the views of the management of the Federai Reserve Bank of San Francisco, or of the Board of Governors of the Federal Reserve System. Editorial comments may be addressed to the editor (Barbara Bennett) or to the author.... Free copies of Federal Reserve publications can be obtained from the Public Information Department, Federal Reserve Bank of San Francisco, P.O. Box 7702, San Francisco 94120. Phone (415) 974-2246. Research Department Federal Reserve Bank of San Francisco

P.O. Box 7702 San Francisco, CA 94120