The establishment of an integrated market for central-bank money across the countries participating in the Monetary Union is an indispensable requirement for the conduct of the single monetary policy. The integration of the interbank market is necessary for the emergence of a single monetary stance, not fragmented into `local' interest rate conditions. Monetary impulses can then be transmitted throughout the Union rapidly and uniformly.
Full integration of the market for central-bank money, however, can only be attained if it is supported by the integration of national payment systems, that funds can be freely transferred across the Union within the same day. The swift mobility of liquid balances stands as a precondition for the execution of arbitrage operations which are the market mechanism bringing about a single monetary stance. Arbitrage transactions give economic body and flesh to the Walrasian auctioneer in the setting of a unique market price for central-bank money which efficiently reflects demand conditions &emdash; also influenced by idiosyncratic, national liquidity shocks (e.g. due to operations of national Treasuries) &emdash; as well as the supply policy on the part of the ESCB, irrespective of the location where the injections (withdrawals) of liquidity take place.
The TARGET Project published by the European Monetary Institute in May 1995 is the response of EU central banks to the challenge of in-tegrating (and harmonising) payment systems in the Union. The general architecture of TARGET consists of one Real Time Gross Settlement (RTGS) system in each country participating in Monetary Union; a number of technical links and procedures (the `Interlinking' network) to allow payment orders to be transferred from one RTGS sys-tem to another; decentralised settlement accounts at the NCBs. The TARGET architec-ture envisages that the ECB will not hold settlement accounts, and that it will carry out its own payments through the Interlinking network: the bulk of payments will be processed by domestic RTGS systems and exchanged, after settlement, between NCBs.
The implementation of TARGET requires the solution of a number of important issues, still open at this stage. These include the decision on the instruments and procedures for the supply of intraday liquidity, the operating hours, the pricing policies, the access requirements, the availability of queuing facilities for temporarily unfunded payment messages. In the next three Sections we explore some of these issues from a monetary policy perspective.
The switch to RTGS of the bulk of large-value payments is the key to the success of TARGET. RTGS can be seen as the response to the increased demand for payment finality brought about by the growth of both domestic and cross-border financial transactions, particularly in derivative finance.
When netting schemes are replaced by RTGS, however, higher settlement costs in terms of larger liquid balances fall on economic agents, which react by economising on costly reserves and/or delaying payments. In doing so, however, they may not take into account the externalities involved in their reserve management behaviour. An example may illustrate this kind of externalities. Suppose that bank A has to send a payment to bank B, which in turn must send a payment to a third bank. If A has sufficient reserves and makes its final payment to B, the latter needs less reserves than it would without A's incoming payment. Since holding reserves is costly, B has an incentive to wait for A to settle its transaction; if A settles, B benefits from a positive externality.
As is well known from the analysis of public goods, competitive markets generally do not take into account externalities, leading to an underprovision of the good which produces positive externalities. In the absence of a system of incentives sustaining a co-operative equilibrium which `internalises' externalities, agents will tend to hold a lower level of reserves than socially optimal.
This argument suggests that the switch from a netting to a RTGS system is not likely to be accompanied by an adequate increase in voluntary reserve balances, as individual banks can be expected to prefer delaying payments rather than bear the additional liquidity cost. Therefore, when considering the implications of RTGS, the payment pattern cannot be taken as given independently of the settlement arrangements. A structurally low level of reserves may hamper the smooth working of payment systems, with detrimental consequences on the efficiency of the conduct of the single monetary policy. Moreover, the good functioning of developed financial markets relies on the availability of sufficient liquidity. If this is missing, systemic risk might still loom large despite the higher degree of payment finality, and could possibly increase, especially in derivative markets.
Further issues arise from the foreseeable coexistence of RTGS and net settlements in the EU. The risk that an insufficient number of payments be actually executed through RTGS systems cannot be ruled out. The overall degree of payment finality could then be insufficient, possibly hampering the emergence of a single monetary stance throughout the Union, which, as argued above, presupposes the possibility of executing intraday arbitrage operations. Since net systems implicitly provide intraday liquidity at no cost, `spontaneous' market forces need not lead to the wide-spread use of RTGS systems, which instead require the availability of (costly) intraday money balances. In the only two countries where RTGS systems are well developed &emdash; the United Stated and Switzerland &emdash; the introduction of this kind of payment systems has been subsidised either explicitly or implicitly: in the United States, through the gratuity, until very recently, of daylight overdrafts; in Switzerland, through the resort to a queuing mechanism which implicitly provides intraday liquidity at no cost. When no subsidy has been granted &emdash; as in Italy and Germany &emdash; the bulk of large-value payments have continued to be routed through net settlement procedures after the establishment of RTGS systems.
The promotion of the use of RTGS systems for both domestic and cross-border payments may thus be needed for the success of the TARGET system. Individual banks could otherwise lack the incentives to switch to RTGS because of the externalities involved in reserve holding, while collective action problems may hinder the adoption of RTGS as the standard way of settling transactions in organised markets. In particular, as recognized by the TARGET project, a rapid implementation of the Lamfalussy standards is crucial both in order to contain risks and to establish a `level playing field' for the competition between RTGS and net settlement systems.
Even though a number of technical devices &emdash; such as the better planning of payment flows at the individual institution's level, or the setting up of operating queues &emdash; could somewhat reduce the demand for intraday liquidity, both the very nature of RTGS systems and the experience of several countries make a strong case for central banks to supply adequate liquidity on an intraday basis. The arguments supporting this policy prescription appear all the more compelling in the EU context, as an insufficient provision of intraday liquidity could complicate the shift to RTGS.
The supply of intraday liquidity for payment-system purposes is generally thought not to have any implication for monetary control, provided that intraday central bank credit does not turn into overnight credit. This conclusion has been implicitly endorsed by European monetary authorities. In art. 4, the Council Regulation n. 3603/93, after recalling in the preamble that central-bank intraday credit may be useful for the smooth functioning of payment systems, states that intraday credit to the public sector `shall not be considered as a credit facility within the meaning of Article 104 of the Treaty, provided that they remain limited to the day and that no extension is possible'. Only if intraday liquidity does not impinge on monetary control can daylight credit to the public sector be compatible with the ban on monetary financing.
The analytical support to this view comes from an arbitrage argument which can be embedded in a broad class of models. Since it is not possible to replicate an overnight contract with a combination of intraday contracts &emdash; daylight loans have to be reimbursed fully before the end of the day &emdash; intraday liquidity conditions have no bearing on the determination of the overnight rate, which thus remains the price of the shortest-maturity contract in terms of outside money relevant for the conduct of monetary policy.
Crucial to this argument &emdash; and indeed to the intuition underlying the irrelevance of the intraday liquidity for monetary policy &emdash; is the assumption that the amount of end-of-day outside money is set by the central bank irrespective of the liquidity conditions prevailing during the day. Furthermore, market participants must be convinced that this is the case. Otherwise, their guesses about the central bank's reaction to intraday liquidity conditions would cause shifts in their demand for end-of-day overdrafts &emdash; thus affecting the overnight rate &emdash; even if the central bank had no intention whatsoever to respond to intraday fluctuations in reserves.
Whether these two assumptions are warranted depends on the information which intraday liquidity conditions can be expected to convey. In general, the demand for daylight reserves only reflects the vagaries of payment flows and hence the central bank cannot possibly extract any information relevant for monetary policy decisions. Therefore, the ESCB should, as a norm, refrain both from reacting to intraday liquidity fluctuations and from following monitoring practices which may lead market participants to surmise that movements in intraday liquidity influence the supply of end-of-day reserves.
Yet, in some exceptional circumstances, `excessively' large intraday overdrafts might anticipate a surge in the demand for reserves stemming either from a very sizeable liquidity shock (which the central bank would presumably notice anyway) or from a sudden shift in expectations. To make a concrete example, a speculative attack on the ECU may start with a massive increase in intraday overdrafts with the central bank. If the monetary tightening in response to the attack were intended to leave the overnight rate at or below the penalty rate applied to the marginal refinancing facility &emdash; i.e. if the `Lombard' window continued to be available to banks willing to pay an interest rate higher than the market rate &emdash; the intraday supply of liquidity should not be altered. What would be the rationale, from a monetary policy point of view, in denying reserves on an intraday basis while granting them at the end of the day? Conversely, if the attack were so intense as to induce the ESCB to resort to such an extreme monetary policy measure as the closing of the Lombard facility, then also daylight overdrafts should be curtailed. Otherwise, the monetary restriction would not be as intense and effective as the ESCB desires, since the intraday supply of liquidity could lead to a slippage in monetary control.
In sum, the supply of intraday liquidity to lubricate the working of payment systems does not impinge on the effectiveness and efficiency of monetary control provided that the ESCB makes clear to market participants that, in normal circumstances, the vagaries of intraday payment flows do not influence monetary policy decisions. In exceptional circumstances, however, the principle of the irrelevance of intraday liquidity conditions for monetary policy could be forsaken. Yet, these departures should only be very rare and clearly understood by market participants. Since the distinction between normal and exceptional may turn out to be difficult in practice, the objective of transparency in the relationship between monetary control and the supply of intraday liquidity could be achieved by sticking to the rule that the granting of intraday overdrafts can be restricted only under conditions so exceptional as to require the closing of the Lombard facility.
The ESCB can supply intraday liquidity through the two basic sets of instruments which are used to regulate the monetary base: standing facilities and open market operations.
The granting of intraday overdrafts in the accounts held by commercial banks with the ESCB is a very attractive arrangement because of its flexibility. As the supply of liquidity is demand driven, it can minimise the risk of gridlocks in payment systems. Liquidity is made directly available to the banks needing it even in the absence of a well-developed market for daylight funds. Collateralisation of overdrafts shelters the ESCB from credit risk. Moreover, it attaches a cost to daylight overdrafts which prevents liquidity from being a free good, with the inefficient oversupply this would entail.
The only drawback involved in the granting of intraday overdrafts is the risk that they are turned into overnight advances, implying an undesired creation of monetary base. This risk can however be drastically reduced by imposing penalties which are stiff enough to discourage the transformation of a daylight credit into an overnight one as a source of liquidity. After all, such a transformation is analogous, from a monetary policy point of view, to the access to the Lombard facility. Also commercial banks' resort to the latter leads to a creation of outside money which was not planned by the ESCB. However, monetary control will not, in general, be affected, given that the resort to the marginal refinancing facility will typically be rare and small in size. Why should commercial banks ask the ESCB for a Lombard advance when they can get liquidity in the market at a lower rate? In exceptional circumstances, the ESCB may close (or ration) the Lombard facility as a component of an extreme monetary tightening. If this were the case, it has been argued above that the supply of intraday liquidity would have to be curtailed too.
These arguments suggest that the granting of collateralised overdrafts to supply daylight liquidity does not in general impinge on either the firmness of monetary control or the smooth functioning of the Lombard facility to accommodate unexpected liquidity shocks hitting single banks. However, especially at the beginning of Stage Three, market participants may not be accustomed to the new system and may have doubts about the resolve of the ESCB in resisting the transformation of intraday credit into overnight credit. In order to establish credibility in this aspect of monetary management, the penalty for such a transformation should be particularly stiff, in any case such as to make the overnight credit obtained by the prolongation of what had initially been announced as an intraday credit more expensive than the advances obtained by resorting directly to the Lombard facility. This penalty policy would enhance the distinction between intraday funds for payment-system purposes and overnight advances, create an additional incentive for the efficient management of liquidity, and provide an effective remedy against the moral hazard that intraday funds are not reimbursed at the end of the day.
Open market operations (OMOs) &emdash; which are bound to be the key instrument used by the ESCB to steer liquidity conditions &emdash; could in principle be used to supply intraday funds. Their main advantage lies in the fact that they are oriented to the `market' and not to single intermediaries. OMOs enhance the role of economic forces in the efficient allocation of liquidity and provide the ESCB with useful information about liquidity conditions through the auction procedure. Furthermore, they reduce the risk that intraday funds are transformed into overnight credit, as repo operations automatically provide for both the injection and the withdrawal of liquidity.
However, daylight OMOs would be viable only if the market for intraday liquidity were sufficiently developed, transparent and efficient as well as endowed with smooth procedures for the quotation of prices and the execution of transactions. So far, in no country has a market for intraday liquidity developed such features. Even in countries &emdash; such as the US, Japan or Switzerland &emdash; where intraday liquidity is valued, although for different reasons (1), there have been no signs that an intraday market with characteristics comparable to those of the interbank market is emerging. Intraday liquidity transactions have so far been relatively few and small-valued, and have taken the form of bilateral exchanges rather than of transparent market transactions with publicly-quoted prices and bid-ask spreads.
Should the ESCB promote the development of a well-functioning market for intraday funds? Competitive forces may not be sufficient to lead to the establishment of a new, welfare-improving market because benefits may not accrue to the same agents who bear the set-up costs. Indeed, this is one of the classical instances of market failure where welfare-improving public intervention is called for.
Available experience does not seem to provide a very strong case for an active promotion on the part of the ESCB. No major inefficiencies in the distribution of intraday funds have emerged despite the absence of a well-developed market for intraday liquidity. After all, the presumption that no major efficiency gains can be expected from improving the allocation of intraday liquidity may be viewed as an implication of the very argument that intraday reserves should be relatively abundant in order to `internalise' the externalities involved in the use of the monetary base and lubricate the payment system effectively. This reasoning, naturally, does not imply that the ESCB should not foster the emergence of the new market if economic forces pushing for its establishment surfaced and gained strength.
Until a fully-fledged market for intraday funds has developed, the execution of intraday OMOs would presumably encounter two kinds of problem. First, the launching of tenders may be complicated by the difficulty in estimating intraday liquidity needs - and the corresponding daylight `interest rate' conditions - if a substantial amount of intraday funds continued to be exchanged through bilateral arrangements.
Secondly, the compliance with the principles of decentralisation and equal opportunity would require that a relatively large number of banks be eligible for participation in intraday OMOs. The processing of a large number of bids - as well as the involvement of several NCBs in the injection and withdrawal of liquidity - could be a rather cumbersome and time-consuming process.
Although certainly manageable for monetary-policy-oriented OMOs, such a process could not be feasible and efficient for payment-system-oriented intraday OMOs which, by their very nature, require both legs of the repo transaction to be carried out within the same day.
On the other hand, this feature which distinguishes OMOs geared to the smooth functioning of payment systems is likely to reduce the risk that the execution of intraday OMOs might interfere with OMOs aiming at the management of the monetary stance. The execution of the latter, however, may take into account payment-system considerations, especially as regards the timing of their settlement. In particular, certain settlement arrangements - early in the day for OMOs injecting liquidity and late in the evening for those withdrawing it - could give rise to a useful source of intraday liquidity in the days when monetary-policy OMOs are carried out.
(1) In the US, intraday liquidity is valuable as a result of the recent introduction of fees on daylight overdrafts with the Federal Reserve. In Switzerland, because the fee structure of SIC penalises both late payments and prolonged permanence in the queuing mechanism. In Japan, the BOJ-Net provides two cut-off times, between which banks exchange liquidity on a continuous basis.