Monetary policy and open market operations during 1994.

This article is adapted from a report to the Federal Open Market Committee by Peter R. Fisher, Executive Vice President of the Federal Reserve Bank of New York and Manager of the System Open Market Account. Ann-Marie Meulendyke, Adviser, Open Market Function, and Spence Hilton, Manager Open Market Trading and Analysis Staff, were primarily responsible for the preparation of this report. Other members of the Open Market Function assisting in the preparation of the report were Robert Van Wicklen, Theodore Tulpan, Eileen Steigleder, and Steve Zannetos. William May, Economist, Financial Markets and Institutions Department, also assisted.

In 1994 the operating techniques for implementing monetary policy remained similar to those of recent years; however, the Trading Desk at the Federal Reserve Bank of New York gained slightly more flexibility in its execution of open market operations after the Federal Open Market Committee began announcing its policy actions in February. As a consequence of the change in procedures, open market operations were no longer used to communicate policy shifts. Nearly all the Desk's operations added reserves because cumulative reserve shortages were substantial for the fourth consecutive year. These deficiencies reflected the continued rapid expansion of currency, which stemmed in part from heavy currency shipments abroad. Working in the other direction were declines in the demand for reserve balances arising from monetary policy tightening. Higher interest rates reined in the growth of transactions deposits and reduced the balances that banks were required to hold at the Federal Reserve. As these balances fell, banks lost some flexibility in managing their reserve positions, and by year-end the potential for operating difficulties associated with low balances had reemerged.

The next section of the report briefly reviews the course of monetary policy in 1994 and describes the responses of the fixed-income securities markets to economic and policy developments. Monetary policy moved away from the accommodative stance that had been in place for some time as the robust pace of economic growth cut into remaining excess productive capacity. With the economy expanding rapidly and the Federal Reserve acting to restrain inflationary pressures, interest rates moved sharply higher and the yield curve flattened. The extent of the rise in yields took many market participants by surprise, contributing to losses and a few bankruptcies, particularly by highly leveraged accounts.

The final section of this report discusses the Open Market Trading Desk's implementation of the objectives established by the Federal Open Market Committee (FOMC). It reviews policy techniques and factors affecting reserve supplies and demands over the year. In 1994 the Desk added a net $32 billion to its securities portfolio, the second largest annual increase. Repurchase agreements with relatively short maturities were used extensively by the Desk to manage reserves within two-week reserve maintenance periods; such transactions are well adapted to handle short-term variations in reserve levels and the frequent revisions to estimated reserve needs. In addition, pricing of daylight overdrafts, which began in April, had the potential to complicate policy implementation, but the actual effects on operations proved to be minimal.

Monetary Policy

and Financial Market Response

The Course of Monetary Policy

Monetary policy in 1994 was formulated against a background of rapid economic growth and rising resource utilization but generally modest aggregate price increases. The FOMC increased reserve pressures at five of eight meetings and once between meetings, resulting in a cumulative increase of 2 1/2 percentage points in the federal funds rate (table 1). Asymmetric directives indicating a greater likelihood that future changes in policy would be toward restraint were adopted at the three meetings at which no change was made to existing pressures. Meanwhile, the Board of Governors approved three increases in the discount rate totaling 1 3/4 percentage points. When determining the stance of policy, the FOMC continued to monitor a broad range of economic and financial indicators. Annual targets were still set for the broader monetary aggregates, but the FOMC placed limited weight on the aggregates because of the considerable uncertainty that persisted about the behavior of their velocities.(1)

[TABULAR DATA OMITTED]

Economic Background

The economic expansion remained on solid footing throughout 1994, with personal consumption, business investment, and inventory accumulation the mainstays of growth (table 2). Consumer outlays for durable goods were particularly robust, and producers' durable equipment purchases remained strong for the third consecutive year. The rate of inventory investment picked up over the first two quarters and remained at relatively high levels for the rest of the year. The pace of expansion was moderated by developments in other sectors: Residential construction activity cooled off as the year progressed, government expenditures trended lower, and the trade balance remained a modest drag. Despite these offsetting factors, by year-end the rapid pace of output expansion had brought resource utilization rates up to levels associated historically with rising inflationary pressures. The unemployment rate fell to 5.4 percent in December, and the industry operating rate stood at 85.4 percent.

[TABULAR DATA OMITTED]

Although the slack in the economy steadily diminished, aggregate price increases for final goods and services remained modest. Inflation, as measured by the fixed-weight GDP deflator and the consumer price index, showed no deterioration; increases in producer prices for finished goods remained low; and labor cost increases were restrained. Nonetheless, evidence accumulated that price pressures could be intensifying. Producer price increases at the intermediate stage of production accelerated, and manufacturers increasingly reported paying higher prices for their inputs.

Policy Initiatives

The initial monetary policy move came at the February FOMC meeting; it represented the first change in reserve conditions since September 1992 and the first move toward tightening since early 1989. The Committee adopted a limited measure, associated with a 1/4-percentage-point rise in the federal funds rate, because of the likelihood that this first step toward firming policy in some years might be magnified in the financial markets. At the same time, it was felt that this action would effectively signal the Committee's anti-inflation intentions.

In a departure from past practice, the Chairman of the FOMC issued a brief public statement announcing this policy decision to avoid misinterpretation of the Committee's actions by market participants. Similar brief statements were issued on a case-by-case basis to announce the other FOMC policy changes during 1994.(2)

The Committee raised reserve pressures slightly further at its March meeting, with the federal funds rate expected to rise another 1/4 percentage point. The Committee again limited the size of the move to avoid any overreaction in the financial markets. A third slight upward adjustment in reserve pressures was made between meetings in mid-April. At the May meeting, with the economy evidently expanding on a solid and self-sustaining basis, the FOMC voted to have the full 1/2-percentage-point increase in the discount rate that had been approved that day by the Board of Governors show through to reserve conditions. The Committee felt that financial markets could absorb this more aggressive policy adjustment. The Federal Reserve press release announcing these moves stated that "these actions, combined with the three adjustments initiated earlier this year by the FOMC, substantially remove the degree of monetary accommodation that prevailed throughout 1993."

At the conclusion of the July FOMC meeting, at which no policy change was initiated, a Federal Reserve press spokesperson indicated that the meeting had adjourned and that no further announcement would be made. The Committee authorized this step to avoid uncertainty about its intentions. Similar statements were authorized following the other two Committee meetings at which no rate actions were taken.

The FOMC next raised reserve pressures at its August meeting, when the full amount of a 1/2-percentage-point hike in the discount rate approved by the Board that same day was passed through to reserve markets. A Federal Reserve press statement indicated that "these measures were taken against the background of evidence of continuing strength in the economic expansion and high levels of resource utilization," and went on to add that "these actions are expected to be sufficient, at least for a time, to meet the objective of sustained, noninflationary growth."

The economy continued to display considerable forward momentum over the autumn, and there was some sense that past policy actions might be having less effect than expected, even in sectors believed to be especially sensitive to interest rate increases. At its November meeting, the Committee agreed that a substantial firming in policy was appropriate. In its final policy move of the year, the Committee voted to pass through to reserve conditions the full effect of a 3/4-percentage-point hike in the discount rate approved that day by the Board of Governors.

Financial Market Developments

Interest rates across the maturity spectrum rose sharply in 1994. Yields on Treasury coupon securities ended the year 150 to nearly 350 basis points higher than they were a year earlier, while the coupon yield curve flattened substantially.

Yields rose dramatically in the first few months after the Federal Reserve began to tighten policy in early February. By mid-May, the yield on two-year Treasury notes had risen about 180 basis points, and the thirty-year bond yield was up more than 110 basis points. Market analysts sensed that the economy retained significant forward momentum and anticipated that the Federal Reserve would respond forcefully to ward off inflationary pressures. Consequently, rates on many short- and intermediate-term securities rose, and a wide spread emerged between these yields and the federal funds rate.

Longer-term yields also rose as investors grew anxious over whether the gains made in reducing inflation in recent years might begin to erode. Market participants focused on the inflation risks posed by the shrinking degree of economic slack, and they were disturbed by information appearing in manufacturers' surveys, as well as evidence from commodity price movements, that suggested an intensification of price pressures. Rising interest rates in European countries and weakness in the dollar spilled back and reinforced the upward momentum in domestic yields. Hedging activity in the mortgage-backed-debt market, a sector particularly hard hit by the sharp rise in yields, lifted rates on intermediate-term Treasury securities.(3)

From mid-May through August, yields moved in a broad trading range. Large rate movements were often followed by abrupt reversals, a pattern that resulted in generally small net changes. Investors responded to economic data that presented a mixed picture. Episodes of dollar weakness continued to weigh on sentiment, as they did intermittently throughout the year. Meanwhile, the monetary policy adjustments in May and August were believed to have brought policy to a more neutral position, and they encouraged brief rallies in debt markets.

Driven largely by a spate of strong economic statistics, interest rates across most maturities resumed their climb from September to early November, rising 65 to 85 basis points. Measures of resource utilization notched higher, and a string of reports showing a resilient housing sector raised questions about the impact of previous interest rate hikes. Survey results of input price pressures faced by manufacturers continued to flash warning signals. By late autumn, it was widely felt that the economy was bumping up against its long-run capacity limits, and many traders began to fear that the Federal Reserve was falling behind in its efforts to rein in inflationary pressures. In late October, the yield on the most recently auctioned thirty-year Treasury bond exceeded 8 percent for the first time in more than two years.

From just before the November FOMC meeting until year-end, the Treasury coupon yield curve flattened further. Short-term Treasury coupon yields rose another 65 basis points, while long-term yields edged down about 20 basis points. The Committee's action in November, viewed by market participants as aggressive, and continued strong economic statistics convinced most analysts that further policy tightening moves were in store and put upward pressure on shorter-term rates. Selling in the front end of the yield curve was exacerbated by liquidations and hedging of portfolios made unprofitable by higher interest rates. Adding to the pressure was the disposal of the securities held by the Orange County, California, Investment Pool after its steep financial losses became known.(4) Meanwhile, the November policy action and continued favorable aggregate price statistics instilled confidence that the Federal Reserve would succeed in preventing a significant increase in inflation pressure. This expectation helped to bring down longer-term yields.

The sharp increases in interest rates in 1994 also had profound effects on investor returns, financial flows, and issuance in the fixed-income markets (table 3). Investors holding portfolios consisting of longer maturity securities sustained particularly heavy losses. The Lehman Brothers Long Treasury Bond Index fell 7 1/2 percent, the first yearly decline in this measure since 1987 and the steepest decline in the twenty-two years spanned by the index. Net returns for most categories of bond mutual funds were negative in 1994, in many cases after the funds posted strong earnings the previous year. Throughout 1994 there were reports of institutions suffering steep financial losses in domestic securities markets. In some cases, the losses were linked to exposures to derivative instruments that magnified the effect of yield movements on interest payments.

3. Measures of performance and activity in domestic
securities markets

          Item                              1994    1993

Portfolio returns (percent
Longer-run Treasury issues                  -7.6    17.3

Mutual funds
 Intermediate-term Treasury debt            -3.3     9.8
 Short-term Treasury debt                     .0     5.8
 Intermediate-term corporate debt           -3.4     9.5
 High-yield corporate debt                  -3.9    19.0
 General municipal debt                     -6.5    12.4

Gross debt issuance (billions of dollars)
Investment-grade corporate debt            206     266
Below-investment-grade corporate debt       27      56
Municipal securities                       154     280

Financial flows (billions of dollars)
Net bound mutual fund inflows              -44     114

Sources: Returns on longer-run Treasury issues are based on the Lehman
Brothers Long Treasury Bond Index and reflect changes in principal value
and coupon income. Returns for the various categories of mutual funds are
from Lipper Analytical Services, Inc. Debt issuance data are from Securities
Data Company. Mutual fund flow data are from the Investment Company
Institute.

Efforts to reduce exposure to rising interest rates spurred huge reinvestment flows in financial markets. Redemptions from bond mutual funds soared following a year of heavy inflows, and withdrawals frequently outpaced inflows as investors reacted to reports of poor performance. The growth in noncompetitive awards at Treasury auctions suggested that many participants began to redirect their investments into securities markets. A heightened sense of uncertainty in financial markets accompanied these elevated flows. Implied price volatility in longer-term Treasury issues was substantially higher in 1994 than in 1993. Meanwhile, new issuance in major sectors dropped significantly, in part reflecting higher borrowing costs.

Implementation of Policy

Operating Procedures

In 1994, the FOMC continued to express its policy directives in terms of a desired degree of reserve pressure. Reserve pressure effectively refers to the costs and other conditions under which the Federal Reserve makes reserves available to the banking system. The FOMC has informally used the federal funds rate as a guide for evaluating conditions of reserve availability since the late 1980s.

In addition, the FOMC has continued to express reserve pressures in terms of borrowed reserves, an approach that involves using nonborrowed reserves to satisfy most, but not all, of the demand for reserves, while forcing banks to meet remaining needs at the discount window, where access is rationed. When the FOMC has increased (or reduced) reserve pressures without a change in the discount rate, expected borrowing has been adjusted upward (or downward) accordingly. The adjustments have been based on the premise that the more the banks are forced to borrow at the discount window to meet their demand for reserves, the more they will bid up the federal funds rate relative to the discount rate.

In the late 1980s, however, the relationship weakened appreciably, in part because a series of banking crises had encouraged observers to associate discount window borrowing with financial difficulties. As a result, banks became extremely reluctant to borrow. Although the banking crises have passed and the association of discount window borrowing with financial problems presumably has faded somewhat, banks apparently still have a reluctance to utilize their borrowing privileges. Consequently, if borrowing were forced to higher levels, the federal funds rate probably would rise substantially more than it had in the past. Against this background, the Desk has continued to develop objectives for nonborrowed reserves calculated as estimated demands for total reserves less the allowance for adjustment and seasonal borrowing. Whenever actual discount window borrowing has differed significantly from the allowance. however, the Trading Desk has accepted the deviation and informally modified the nonborrowed reserve objective accordingly, rather than force unwanted changes in the federal funds rate.(5)

Between February and April, the FOMC's reserve tightening actions lifted the anticipated spread between the federal funds and discount rates from zero, where it had been since September 1992, to 75 basis points. The spread remained at 75 basis points for the balance of the year because the last three policy steps involved equal changes in both rates.

With this widening of the spread, borrowing could have been expected to increase significantly. However, adjustment borrowing actually decreased slightly in 1994, averaging $65 million a day compared with $75 million a day in 1993. Although the decrease is outwardly surprising, closer examination of the data shows some indications of the expected association between borrowing and the funds rate. Adjustment borrowing did pick up on reserve-period settlement days, and it rose for most size classes of banks. Settlement-day adjustment borrowing averaged $336 million in 1994, almost double the $180 million average in 1993. Adjustment borrowing on nonsettlement days by small-and medium-sized banks also increased in 1994, although by less than would have been expected on the basis of historical relationships from the early 1980s. Some of the shortfall in borrowing likely reflected a continuing reluctance to utilize the discount window, but the strong liquidity positions of many of these banks also may have played a role. Small- and medium-sized banks usually account for a considerable portion of nonsettlement-day borrowing.

The decline in average borrowing resulted entirely from a reduction in nonsettlement-day borrowing by large money center banks. These banks have traditionally concentrated their borrowing on settlement days, and in 1994 all of their borrowing occurred on those days. By contrast, members of this group borrowed seven times on nonsettlement days in 1993, either because of operational difficulties or temporarily elevated funds rates.

In the case of seasonal borrowing, the rate incentive for stepped-up borrowing in 1994 was small because the rate charged on seasonal borrowing closely tracked federal funds and certificate of deposit rates. Nonetheless, seasonal borrowing was persistently higher than in recent years; it averaged $193 million in 1994, compared with $109 million the year before. It still followed the typical seasonal pattern, which reflected demands for agricultural loans. As a result, the Desk made ten upward technical adjustments to the formal borrowing allowance between May and August 1994 and nine downward adjustments over the remainder of the year. The increased use of the program was related in part to a marked rise in demand for farm credit at small banks. In addition, strong loan demand at midwestern correspondent banks might have constrained the correspondents' ability to provide seasonal funding to their respondent banks.(6)

The Desk's Approach to Reserve

Management(7)

Reserve Patterns over the Year

The behavior of narrowly defined money, M1, had an important influence on reserve supplies and demands over the year.(8) Currency registered another year of strong growth, and the resulting record $37 billion increase in currency in circulation was the primary factor behind the substantial need to provide reserves in 1994.(9) A decline in the deposit component, however, limited the overall growth of M1 and contributed to a fall in the demand for reserves. Consequently, required reserves, the primary source of demand, slipped about $2 billion, reducing the need to add reserves over the year.

Several other factors also modestly reduced the Desk's need to provide reserves. Applied vault cash, a source of supply, increased about $3 billion, in part mirroring the currency expansion. Rising interest rates led banks to cut their required clearing balances about $2 billion as the rate at which they accumulated earned income credits rose.(10) Because the declines were not offset by higher excess reserves, the lower clearing balances lessened the overall need to provide reserves. These balances had been increased sharply in 1991 and 1992, when banks were adapting to lower required reserve levels, and had been lifted modestly in 1993.(11) On balance, cumulative changes in other supply and demand factors had smaller effects on total reserve needs over the year.(12)

Outright Transactions and Changes

in the System Portfolio

The Trading Desk met the ongoing need to add reserves by increasing the Federal Reserve System's portfolio of U.S. government securities. Altogether, the Desk purchased about $25 billion through six operations conducted in the market, four of them involving Treasury coupon issues.(13) As in the past, the market entries were arranged when available forecasts suggested that large reserve shortages would persist for at least several maintenance periods. The market purchases were supplemented by nearly $11 billion of acquisitions from foreign accounts, almost entirely Treasury bills. These purchases, typically modest in size, were arranged when orders were compatible with estimated reserve needs.

For a second consecutive year, the Desk did not sell securities, although it did redeem some. Because the Treasury no longer sells seven-year notes, the System's holdings of these notes must be redeemed early in each quarter as they mature; more than $2 billion came due in 1994. The Desk also redeemed agency securities when no suitable replacement securities were offered and when issues were called. Holdings of these issues fell for the fourteenth year in a row, declining almost $1 billion, to $3.6 billion.

As a result of the Desk's outright activity, total holdings in 1994 grew $32 billion, to $376 billion. Although somewhat less than the record rise of 1993, this increase was still the second highest ever. Slightly more than half of the increase occurred in Treasury bills, while growth in coupon holdings was strongest in the one-to-five-year sector. Consequently, the weighted-average maturity of the System's holdings was virtually unchanged in 1994.(14)

Temporary Operations

The Desk used self-reversing operations to meet the reserve shortages that developed between its outright operations arid to address reserve imbalances created by short-lived movements in other factors affecting reserves. Almost all of the temporary operations in 1994 added reserves because of the underlying growth in reserve shortages and the Desk's preference for letting deficiencies build to a sizable level before arranging outright purchases. In fact, the Desk entered only one maintenance period facing an estimated need to drain more than a very small amount of reserves, and even that surplus was subsequently erased by revisions to forecasts of operating factors. Consequently, only five matched sale-purchase agreements were arranged all year, and none exceeded one business day.

All told, the Desk arranged $362 billion of repurchase agreements (RPs) for the System and $113 billion that were customer-related. The number and average size of multiday System RPs both fell in 1994. Several factors contributed to these declines. A greater share of the year's reserve needs was met with outright operations: The Desk typically made outright purchases that left a remaining estimated need to be met with RPs, but on several occasions actual needs fell below the estimated needs. In addition, the Desk further increased its use of fixed-term operations in 1994 (discussed below), reducing the need for replacement RPs to offset early withdrawals.

Managing Reserves within a Maintenance Period

When developing strategies for each maintenance period, the Desk took into account the estimated day-to-day distribution of reserve shortages or excesses, the potential for revisions to reserve estimates, and bank reserve management strategies.(15) The Desk generally met each period's reserve needs gradually in order to accommodate sometimes uneven reserve distributions and possible revisions. It often arranged a series of multiday RPs, many of which matured in three or four days. The Desk also continued to be guided by the federal funds market. When faced with conflicting information between the funds rate and forecasts of reserve supply and demand, the Desk had to evaluate which indicator was likely to provide the more reliable information about reserve availability.

Banks' reserve management strategies can affect the funds rate because they influence reserve demands within a maintenance period. As several previous reports have explained, the cuts in reserve requirement ratios made between 1990 and 1992 reduced the level of required operating balances at the Federal Reserve.(16) These lower levels increased the likelihood that depository institutions would be unable to eliminate unwanted excess positions without running an overnight overdraft. Consequently, in the early 1990s, depositories tended to concentrate their reserve holdings late in a period, showing particular caution about holding excess reserves over the weekend, when reserves count for three days. This reluctance to hold reserves over a weekend was the main contributor to soft funds rates on Fridays.

In 1994, banks used these reserve management practices less aggressively. By the end of 1993, rapid growth in required reserves and clearing balances had restored required operating balances to the levels prevailing right before the initial round of cuts in reserve requirement ratios in late 1990. Perhaps as a result, the distribution of demands for excess reserves within a maintenance period appeared less skewed in 1994 than it had been in the preceding three years.(17) Moreover, the degree of softness on Fridays was typically slight. Nonetheless, banks were still reluctant to accumulate large excess reserve holdings early in a maintenance period. By the end of 1994, the level of required operating balances had once again fallen back to the lower levels seen in late 1991 and in 1992, thus reducing banks' reserve management flexibility. This decline reflected the drops in required reserves and clearing balances and the expansion in applied vault cash noted earlier.

The Desk further increased its use of fixed-term RPs on Thursdays to run through the weekend, a strategy that avoided the risk of large early withdrawals on Fridays if the federal funds rate traded to the soft side while a large reserve need remained. The Desk believed that if withdrawable RPs had been arranged on a Thursday, dealers probably would have opted to refinance at lower rates the next day, forcing the Desk to find another opportunity to add back the reserves. The Desk also expanded the use of fixed-term RPs on the first Monday through Wednesday of each period, again to avoid unwanted withdrawals and to reduce the number of operations.(18) Withdrawable RPs were still useful at times, particularly when the Desk felt that operating factors or required reserves might turn out to be sufficiently different from estimates to sharply reduce or eliminate the estimated reserve need. Thus, withdrawable RPs continued to be used over the final few days of many maintenance periods.

Market speculation during the year that monetary policy might be tightened sometimes put upward pressure on the federal funds rate that did not seem justified by estimates of reserve imbalances. The Desk remained sensitive to these situations when formulating its operations strategy to avoid any misunderstanding by market participants, who continued to view open market operations as a possible indicator of policy shifts.(19) Consequently, on several occasions when the funds rate was very high, the Desk arranged overnight System RPs, in part to prevent any perception that it was either paving the way for a firming in policy or hinting at a Committee inclination to change policy.(20) As the year progressed and market analysts began to assume that the FOMC would indicate its policy actions through a public announcement, market participants came to feel that the Desk's open market activities were less likely to be used to communicate policy shifts. This perception gave the Desk more flexibility in selecting its operations to meet its reserve objectives.

Trading Room Automated Processing System

In 1994, the Desk began arranging its open market operations using the Trading Room Automated Processing System (TRAPS). Under TRAPS, the Desk announces reserve operations and dealers respond with their propositions through Fedline terminals. The system is also used to process operations and to notify dealers of the results. The Desk started using TRAPS for its temporary operations in July, followed in August by the first outright market purchase using the system.

Daylight Overdraft Pricing

On April 14, the Federal Reserve began charging banks a fee of 10 basis points on overdrafts incurred in their reserve accounts during the day.(21) Previously, daylight overdrafts had been subject to size limitations related to a bank's capital, but they were not subject to charges. For a few banks, such daylight overdrafts were substantial. The Trading Desk anticipated that the charges might affect its own operations by encouraging changes in the functioning of the federal funds and RP markets and in some banks' reserve management techniques. In preparation for pricing daylight overdrafts, Federal Reserve personnel had conversations with market participants and undertook some contingency planning. As it turned out, however, Desk operations were minimally affected in 1994.

Before charges were assessed for daylight overdrafts, reserve management was focused on end-of-day reserve balances rather than on intraday balances. End-of-day balances are important because they meet reserve requirements. Furthermore, banks need reserve balances at the end of the day to avoid overnight overdrafts and their associated stiff charges. In fact, total reserve balances vary considerably during the day, rising whenever the Federal Reserve or any entity maintaining an account at the Federal Reserve - the federal government, federally sponsored agencies, or foreign official institutions - makes payments and falling whenever it receives payments.(22) The most dramatic movements in intraday balances, however, have been in the distribution of reserves, with large intraday balances occurring at some banks and huge overdrafts at others during part of the day.(23)

The previous absence of fees had encouraged practices that resulted in large daylight overdrafts. For example, many financial market transactions, such as interbank federal funds and RP contracts, did not specify transaction settlement times. Yet receipt and return times do influence the intraday distribution of reserves. In federal funds transactions, the sending bank controls the timing of the reserve transfer. Under daylight overdraft pricing, it was thought that banks facing intraday reserve charges might delay sending federal funds in order to increase their intraday balances. If Fedwire traffic became concentrated near the end of the day, the funds market could lose liquidity, thus making the rate a less reliable indicator of reserve availability.

In practice, however, after daylight overdraft pricing began, the average time for sending funds transfers over Fedwire moved only slightly to later in the day. Apparently, many banks did not change their practices because they did not face large enough daylight overdrafts from their funds transactions to justify the cost of making changes. Federal funds brokers did report that some requests for transactions specified sending or returning funds during specific time periods and noted that some potential trades were rejected because the counterparty was reputed to be a "late sender." But these restrictions affected only a small portion of trades and therefore did not impede market liquidity.

For securities transactions, the sender of the securities controls the transaction time. Consequently, banks lose reserve balances when they receive securities, but they cannot control the time at which that happens.(24) Dealers, who rely heavily on RPs to finance inventories, traditionally had their clearing banks send the securities to their counterparties' custody banks between late morning and early afternoon. Then, on the maturity date, the counterparties' banks typically returned the securities at the opening of business. The prevalence of this timing pattern caused both the dealers' and their banks' accounts to be overdrawn during the morning because the dealers began the day with small working balances. In anticipation of daylight overdraft pricing, the clearing banks informed their customers that they would pass on the overdraft charges.

Dealers indicated in conversations with the Federal Reserve that they planned to speed up their negotiation and processing of RPs in the morning so that any securities being returned and then refinanced would leave their accounts more quickly. Some participants predicted that this speedup in RP operations would cause the market to be liquid only briefly early in the morning. Such a development was of particular concern to the Federal Reserve because the Desk's temporary open market operations are routinely executed around 11:30 a.m. The Federal Reserve had chosen that time because information about reserve levels is received and analyzed gradually over the morning. Only part of the data flow could be accelerated. If the Desk were forced to arrange its open market operations a couple of hours earlier, it would have to base its decisions on less reliable data.

To address these concerns, the Desk did make one change in its procedures: It delayed the return time for the collateral on its own maturing RPs from the opening of business until 11 a.m., thereby leaving reserves in the banking system for a larger part of the day. It was hoped that the later return time would encourage the dealers to participate in the late morning operations.

Once pricing began, the RP market did experience a shift toward somewhat more morning activity, but a number of customers continued to seek RP investments during the late morning and early afternoon, so market liquidity was retained. More rapid processing of trades has accounted for most of the reduction in peak and average overdrafts.(25) In addition, the volume of afternoon trades for next day delivery has increased.

The Desk saw essentially no change in participation rates in its RP operations after April. Dealers reported somewhat smaller inventories of securities left to be financed at midmorning, but on most days, they were nonetheless able to submit propositions of sufficient size for the Desk to accomplish its planned operations. Furthermore, dealers' customers increased their participation in Trading Desk operations.

Appendix A: The Monetary Aggregates

Growth of the broader monetary aggregates remained subdued in 1994. The FOMC voted in February to retain the growth ranges for M2 and M3 adopted on a preliminary basis the previous summer. These ranges were consistent with the expected slowing of nominal income and the anticipated continuation of the substantial velocity increases experienced in recent years. The FOMC reaffirmed these ranges in July. For the entire year, M2 advanced a mere 1.0 percent, at the lower end of its annual growth cone, while M3 rose only 1.2 percent, within the lower half of its annual growth cone.(26) Growth in the broader aggregates was held down in 1994 by weakness in the liquid components, including savings and interest-bearing checkable deposits.(27) These deposits were relatively unattractive because depositories raised rates at a much slower pace than market rates rose.(28) The preference for market investments and the resultant increase in velocity were factors in the Committee's decisions to accept the weak aggregates.

Some components of the broader aggregates, however, did show strength. Depositories sharply increased their issuance of both overnight Eurodollars and RPs, thus lifting M2. In addition, during the second half of the year, issuance of consumer time deposits picked up, as did growth in retail money market mutual funds. M3 received some support from large time deposits and term RPs and Eurodollars, while institutional money funds were very weak early in the year but showed more robust growth later. The strength in some of these components reflected expanded bank funding needs. Total bank credit rose 6.8 percent in 1994, after having grown 5.0 percent the previous year. The increase was concentrated in bank lending; aggregate holdings of securities fell modestly on balance over the year.(29)

After three consecutive years of rapid growth, M1 rose only 2.4 percent in 1994. The slowdown in part reflected substantial increases in opportunity costs, which depressed deposits. Reduced mortgage refinancing activity also weakened demand deposits, and sweep programs initiated by several banks lowered other checkable deposits.(30) But currency, buoyed by heavy shipments overseas, registered another year of strong growth, expanding about 10 percent over the four quarters.

Finally, domestic nonfinancial debt grew 5.3 percent in 1994. The improved balance sheet condition of many borrowers supported growth of nonfederal debt. Total debt ended the year toward the lower end of its monitoring range.

Appendix B: Reserve Forecast Accuracy

This appendix reviews the accuracy of staff forecasts of the factors affecting reserve supply and demand. For the year, the accuracy of the forecasts for required reserves was similar to that for 1993 at each stage of the maintenance period (table B.1). The Desk maintained a formal allowance of $1 billion for excess reserves during each of the twenty-six maintenance periods in 1994, but it often made informal allowances when demand for excess reserves was expected to be above or below the path allowance.(31)

[TABULAR DATA OMITTED]

On average, the estimates available at the beginning of the period of the factors affecting the supply of nonborrowed reserves improved. The smaller forecast errors largely resulted from better estimates of the Treasury balance and less distortion from the treatment of premiums on RPs, while currency projections showed some deterioration.

There was a marked improvement in the first-day estimates of the Treasury's balance at the Federal Reserve in 1994, particularly around the important September and December tax payment dates. A surge in tax receipts can cause the Treasury's total cash holdings to exceed the capacities of the Treasury Tax & Loan (TT&L) note accounts at depository institutions, with any excess flowing into the Treasury's balance at the Federal Reserve. Forecasting the balance in the Federal Reserve account, therefore, can be particularly difficult around these times. In 1994, Treasury cash levels were above the capacity of the TT&L accounts on fourteen days, much less frequently than in 1993, when capacity was exceeded on thirty-two days. Two developments accounted for much of the difference: In September 1994, the capacity was about $8 billion to $10 billion higher than it was a year earlier, making room for more tax receipts. In December, approximately $35 billion of Treasury cash management bills matured without replacement, compared with $14 billion in December 1993. The enlarged maturities limited the size of the Treasury's total cash holdings.

Another factor reducing measured forecast errors was a decline in average premiums on RPs and on coupon securities purchased, elements in the "other items" category. The measured impact of any reserve transaction is based on the par value of the securities, although the actual impact depends on the market value of the securities. In practice, the Desk allows for possible net premiums (premiums less discounts) when they are expected to be large, so that the premiums do not constitute actual forecast misses. Average net premiums in 1993 had grown to 8 percent on all RPs and to 15 percent on market purchases of coupons as a result of falling interest rates. Because of rising interest rates in 1994, however, the average net premiums on securities held under RP fell back to about 2 percent of the par value, with discounts outweighing premiums on some operations. Average net premiums fell to 8 percent on coupons purchased in the market.

Currency projections at the beginning of maintenance periods deteriorated in 1994. Currency often behaved in a manner at odds with past seasonal patterns, which are used for forecasting purposes. In the first and last maintenance periods of 1994, typically times of large seasonal swings, currency drained fewer reserves than initially anticipated.

Appendix C:

Tables Summarizing 1994 Desk Activity

The tables in this appendix support the text discussion of the Trading Desk's approach to reserve management in 1994. The operating factors affecting bank reserves appear in table C.1. The Desk's outright operations are summarized in table C.2, and the operations' effects on the System portfolio are presented in tables C.3 through C.5. Temporary operations are reported in table C.6.

C.1 Reserves measures and factors affecting reserves
                             Maintenance
    Item                       period
                                ended         Change
                              January 4,
                                1995       1994(1)   1993(2)

Bank reserves (millions
of dollars, not seasonally
adjusted)
Nonborrowed reserves           61,372      -2,133    6,100
Borrowed reserves
  Adjustment plus seasonal        246         404     -127
    Adjustment                    151          25     -131
    Seasonal                       95          79        4
  Require reserves(3)          60,451      -1,954    6,116
  Excess reserves               1,167         -74     -144

System portfolio and
operating factors
(billion of dollars)(4)
System portfolio and
    repurchase agreements
    outstanding(5)                385.3        31.5     36.2
Operating factors
  Foreign currency(6)              17.3        -2.1        7
  U.S. currency                   403.0       -37.1    -31.5
  Treasury balance                  7.1         1.4     -1.1
  Float                              .7         -.5     -1.2
  Special drawing rights            8.0          .0       .0
  Gold deposits                    11.1          .0       .0
  Foreign deposits                   .2         -.1       .1
  Applied vault cash               36.4         3.1      2.2
  Other Items                      20.8         2.1      1.3
Foreign repurchase
    agreement pool(7)               8.1         -.4      -.2

Note. Figures may not add to totals because of rounding.
(1.) Change from maintenance period ended January 5, 1994, to that ended
January 4, 1995.
(2.) Change from maintenance period ended January 6, 1993, to that ended
January 5, 1994.
(3.) Not adjusted for changes in required reserves ratios.
(4.) Indicates impact of changes in operating factors on bank reserves. All
items are biweekly averages.
(5.) Matched sale-purchase agreements with accounts are added
back in.
(6.) Acquisition value plus interest. Revaluations of foreign currency holdings
are included in "other items."
(7.) Includes customer-related repurchase agreements.

C.2 System outright operations by type of transaction and
counterparty
Billions of dollars

      Item                             1994   1993
Total outright                         38.3   38.6

By type of transaction
Purchases                              35.3   36.9
  Bills                                17.5   17.7
  Coupons                              17.8   19.2
Sales                                    .0     .0
  Bills                                  .0     .0
  Coupons                                .0     .0
Redemptions                             3.2    1.7
  Bills                                  .0     .0
  Coupons                               2.3     .8
  Agency issues                          .9     .9

By counterparty
Total outright in market               24.7   25.4
  Purchases                            24.7   25.4
    Bills                               7.7    8.6
    Coupons                            17.0   16.8
  Sales                                  .0     .0
    Bills                                .0     .0
    Coupons                              .0     .0
    Agency issues                        .0     .0

Total outright with foreign accounts   10.6   11.5
  Purchases                            10.6   11.5
    Bills                               9.8    9.1
    Coupons                              .8    2.4
  Sales                                  .0     .0
    Bills                                .0     .0
    Coupons                              .0     .0

Note. Values are on a commitments basis.

C.3 System portfolio: summary of holdings
Billions of dollars
                                Change
    Item          Year-end
                    1994     1994   1993
Total holdings     376.2     32.1   35.3
  Bills            185.4     17.5   17.7
  Coupons          187.1     15.5   18.4
  Agency issues      3.6      -.9    -.9

Note. Values are on a commitment basis. Changes in holdings are from
year-end to year-end. Figures may not add to totals because of rounding.

  [TABULAR DATA OMITTED]

C.5 Weighted-average maturity of marketable Treasury
Debt, selected years, 1960-94
Months
                         Holdings
           Federal       outside      Total
Year-end   Reserves      Federal    outstanding
           holdings(1)   Reserve
1960          19            61          55
1965          16            70          60
1970          24            45          40
1975          31            34          33
1980          55            46          48

1985          49            61          59
1986          46            64          62
1987          44            69          66
1988          42            71          67
1989          43            73          69

1990          41            71          68
1991          38            72          68
1992          36            71          67
1993          38            68          65
1994          38            66          63

(1.) The effects of all outstanding temporary transactions, including repurchase
agreements and matched sale-purchase agreements with foreign
accounts, are excluded from the calculation of the average maturity of the
portfolio.

  [TABULAR DATA OMITTED](1.) The behavior of the monetary aggregates and the Committee's targets for them are discussed in appendix A. (2.) Most announcements of policy changes were made early in the afternoon, shortly after the FOMC had completed its meeting. However, at the two-day meeting in February 1994, the announcement was made in the morning on the second day, soon after the Committee made its decision. In that instance, the Committee preferred to make the information available before the weekend and ahead of the Desk's regular 11:30 a.m. operating time. The one policy action taken between meetings was also announced in the morning.

In February 1995, the Committee formally adopted new procedures for conveying information to the public. The procedures include the announcement of all changes in the stance of monetary policy. on the day the changes are made. (3.) Higher interest rates extended the expected durations of mortgage-backed securities, thereby compounding the downward pressure on prices for this debt. Holders of mortgage-backed securities often hedge their exposures by selling intermediate-term Treasury debt. (4.) Roughly $20 billion of securities held by the highly leveraged Orange County fund were sold. Most of these securities were government agency notes, many of them derivative instruments that paid interest according to formulas based on movements in market yields. (5.) The borrowing relationship has been discussed more extensively in previous annual reports of the Open Market Function. (6.) Only small banks are eligible for the seasonal credit program. (7.) Many of the statistics cited in this section appear in tables in appendix C. (8.) Changes in the components of MI and the reasons for the components' behavior are described in appendix A. (9.) Currency in circulation, which is the factor that affects reserve balances, includes cash held by depository institutions; for money supply calculations, however, this vault cash is subtracted. (10.) Earned income credits accumulate at a rate linked to the federal funds rate. The credits may be used only to pay for certain priced services provided by the Federal Reserve, and many large banks hold clearing balances sufficient to generate credits to pay for all the services they use. As the rate at which the credits are earned increases, the maximum useful level of a bank's clearing balance decreases. (11.) Technically, clearing balances are treated as a factor reducing the supply of reserves, although they are actually a source of demand for reserves. (12.) The various foreign-exchange-related activities on the System's balance sheet drained less than $0.5 billion. The historical value of the foreign currency sold was $3.0 billion, about $0.7 billion below the market value. The value of the System's foreign exchange holdings was increased $2.4 billion as a result of upward revaluations, while interest earnings totaled $0.9 billion. In the reserve factor categories, interest earnings and the historical value of foreign currency transactions appear under "foreign currency," while revaluations and the profit or loss on foreign currency transactions appear in the "other items" category. (13.) The Desk bought, in par values, $3.3 billion of Treasury coupon securities on March 15, $5.0 billion of coupons on April 12 (a record volume), $3.8 billion of bills on June 1, $4.5 billion of coupons on August 30, $3.9 billion of bills on November 9, and $4.2 billion of coupons on November 29. (14.) The average maturity of the portfolio is also affected by the reinvestment choices made for maturing securities at auctions. (15.) The accuracy of the staff forecasts for reserve supply and demand is reviewed in appendix B. (16.) Required operating balances are defined as required reserves plus required clearing balances less applied vault cash; they represent the working balances held by depository institutions at the Federal Reserve for supporting payment transactions. (17.) The average levels of excess reserves in the first and second weeks of a maintenance period in 1994 were $725 million and $1,375 million respectively. During 1993, the corresponding figures were $170 million and $1,980 million, and a similar distribution characterized 1992 after the round of reserve requirement cuts made in April of that year. Before December 1990, the distribution of excess reserves within the maintenance period was, on average, fairly even. Of course, Desk reserve provision strategies, which may not match ex ante demands, also contribute to the actual pattern of excess reserves. (18.) A total of forty-four fixed-term RPs were arranged in 1994 (thirty of which were in place on Fridays), compared with thirty-one in the previous year (twenty-three covering Fridays). By contrast, just nine fixed-term operations had been arranged in 1992. (19.) Misinterpretations did in fact arise. On February 3, with fed funds trading just 1/16 of a percentage point above the level associated with the desired degree of reserve pressures, the Desk took no market action to affect reserves because a shortage was not seen. With an FOMC meeting scheduled to start later that day and with expectations of a policy shift running high, some participants interpreted the Desk's inaction as indicating such a shift. In fact, this was not the case, although the FOMC did decide to firm pressures the following day. This episode occurred before the FOMC began to announce policy changes. (20.) With expectations of an easing in policy almost entirely absent in 1994, the Desk felt freer to add reserves when called for by its reserve projections, even when the funds rate was slightly soft. It did so on numerous occasions. (21.) The fee reflects an annual rate of 24 basis points using a standard ten-hour day for Fedwire operations. The charge is made on all end-of-minute overdrafts in excess of a deductible based on 10 percent of the bank's capital. The "Overview of the Federal Reserve's Payments System Risk Policy," published by the Federal Reserve System in October 1993, describes the calculations in detail. (22.) Differences in posting times for check credits and debits also influence aggregate intraday reserve levels. (23.) In the six months before daylight overdraft charges took effect, peak overdraft levels averaged $124 billion. From mid-April through year-end, they averaged $70 billion. To put the overdraft figures in perspective, total end-of-day reserve balances averaged $34.5 billion and $31 billion respectively, over those two periods. (24.) Under the delivery-versus-payment system used for the transfer of government securities, reserve balances are automatically moved from the account of the bank receiving the securities to that of the bank sending them when the transfer is processed. (25.) Average daylight overdrafts fell from $70 billion in the six months before pricing to $43 billion over the balance of 1994. (26.) The data on all the monetary aggregates are as of January 26, 1995, and do not reflect the annual seasonal factor and benchmark revisions of February 2. The earlier data are used because they more closely approximate the information the Committee had when it made its policy decisions. The revisions generally had a minimal effect on total growth over the year. On balance, the revisions redistributed a little more of the net increases in M1 and M2 into the first half of the year and shifted more of the growth in M3 into the second half of the year. The annual changes of the monetary aggregates are measured from the fourth quarter of 1993 to the fourth quarter of 1994. Data on nonfinancial debt reported in this section are as of March 3, 1995. (27.) The behavior of the monetary aggregates is described in more detail in the "Monetary Policy Report to the Congress Pursuant to the Full Employment and Balanced Growth Act of 1978" (Board of Governors of the Federal Reserve System), July 20, 1994, and February 21, 1995. (28.) Investors moving out of mutual funds favored instruments not included in the aggregates, such as the direct purchase of Treasury debt. For this reason, and because of capital losses suffered by many funds, M2 plus bond and stock mutual funds rose less than 1 percent in 1994, an increase similar to that for M2 and well below the nearly 7 percent gain of the previous year. (29.) Credit expansion was partially funded by bank borrowings from abroad, which nearly doubled over the year. (30.) In January, one large regional bank initiated a sweep program that transferred funds from other checkable deposits into money market deposit accounts. Another large regional bank phased in a similar program during September and October. Altogether, these programs lowered M1 growth about 1 percentage point in 1994. The sweep programs shifted funds between accounts included in M2 and therefore had no effect on the broader aggregates. (31.) Excess reserves are estimated from a combination of models and observed behavior during maintenance periods. Any analysis of the accuracy of these estimates would be misleading because it would not take account of the informal revisions.