A recent development in neoclassical monetary economics is the flurry of research on the Federal Reserve's (hereafter, the Fed's) daily implementation of monetary policy in the federal funds market. In the late summer of 2001, for instance, a Federal Reserve Bank of New York symposium on monetary policy offered five papers on the subject (Demiralp and Jorda 2002; Bennett and Peristiani 2002; Krieger 2002; Goodfriend 2002; Woodford 2002) while the July/August 2001 edition of the Federal Reserve Bank of St. Louis Review provided two more (Taylor 2001; Thornton 2001).
Prior to the mid 1990s, the daily implementation of monetary policy had been, relative to recent trends, a neglected area of research among neoclassicals, other than Joshua Feinman's (1993) paper on the Fed's daily reaction function. Research on the Fed's actions generally abstracted from the Open Market Desk's (hereafter, the Desk's) daily actions in the federal funds market, concentrating instead upon the strategy involved in setting intermediate targets as in the monetary policy rule and measuring monetary policy literatures. On the other hand, research on the daily federal funds market primarily examined banks as lenders and borrowers of overnight funds, while abstracting from the Desk's interventions in the market (e.g., Ho and Saunders 1985; Spindt and Hoffmeister 1988; Lasser 1992; Griffiths and Winters 1995).
During the mid to late 1990s, falling required reserves due to the proliferation of retail sweep accounts left banks at risk of overdrawing their reserve accounts during the course of routine settlement of electronic payments. The result was a substantially increased daily volatility in the federal funds rate that both complicated the implementation of monetary policy and provided impetus for new research on the supply and demand forces in the daily federal funds market (1) (Bartolini et al. 2001, 2002; Bennett and Hilton 1997; Clouse and Elmendorf 1997; Demiralp and Jorda 2000; Furfine 2000; Thornton 1999, 2001; Woodford 2002). Several leading journals have since devoted entire issues on complementary topics, such as the implementation of monetary policy in the information age (Posen 2000), monetary policy implementation when inflation and interest rates are both at or near zero (2) (Fuhrer and Sniderman 2000), and the daily operating procedures of various central banks (Underhill 2002). Clearly, analysis of the Desk's daily activities in the federal funds market now has an important role in the neoclassical monetary economics literature.
Unlike neoclassicals, Post Keynesians had already been incorporating the Fed's implementation of monetary policy in the federal funds market within the endogenous money literature. Though this is not the place for a complete treatment of this literature (see Moore 1988 or Wray 1990, 1998), a key argument is that reserves are not an exogenously controllable variable for the Fed; rather, changes to reserves must he made endogenously in response to changes in banks' demand for reserves. This argument contributes significantly to the views of adherents on several key macroeconomic issues, including the transmission of monetary policy, the proper intermediate targets for monetary policy, and the proper goals of monetary policy. Reserve endogeneity also complements recent contributions made by Post Keynesians to neo-Chartalist literature on functional finance and the inter-relationship/appropriate mix of monetary and fiscal policies (e.g., Wray 1998; Bell 2000).
This paper contributes to the monetary economics literature in general by providing a detailed analysis of the Fed's current daily tactics, while also serving as a statement on this topic from an institutionalist perspective. The purpose of the paper is the application of an institutionalist "way to go about thinking" to the implementation of monetary policy in the federal funds market through an examination of how timeliness is achieved in the Fed's daily tactics. "Timeliness" is a term introduced by John R. Commons in the Legal Foundations of Capitalism ( 1995) that refers to how institutions go about recognizing what is necessary to sustain a particular system and what are the appropriate times of implementation. "What is necessary," "sustain," and "appropriate times" are determined within the institutional context and are normally instituted through working rules. F. Gregory Hayden (1987, 1993, 1998) has shown how timeliness and the operationalization/codification of working rules into laws and regulations can be integrated into the institutionalist research program.
The paper is organized as follows: The first section explains the concept of timeliness and its relationship to analysis of time in real world systems and to Commons' concept of working rules. The second section describes the relevant categories of time or event sequences the Desk must be concerned with in its tactics and interprets recent events in monetary policy implementation from within this context. The third section illustrates how the discussion of timeliness in the Fed's daily tactics can inform research on traditional topics in monetary economics and argues that (1) the payments system, rather than reserve requirements, is the proper starting point for analysis of the Fed's daily tactics; (2) there is no liquidity effect in the federal funds market; and (3) direct control over the monetary base is not possible. That the conclusions reached in the third section support the conclusions reached by proponents of the Post Keynesian endogenous money approach is not surprising, as their research has similarly focused upon the interactions of real world institutions within the financial system.
In defining timeliness, it is useful to embed a definition within a discussion of two additional concepts--time and working rules. Institutionalists have long suggested an evolutionary, process-oriented approach to research, rather than an equilibrium-based one, as more appropriate for analysis of socioeconomic phenomena. One issue at the heart of this recommendation, one could argue, is a particular view of time and how it should be incorporated into research. As John F. Henry and L. Randall Wray have argued, "The very notion of time ... and its significance for analysis is determined by what is being examined. Simply put, there is no single meaning of time, no constant that is independent of the questions being addressed and the field of inquiry within which analysis is undertaken. What is time? It depends" (1998, 1).
Some of the more comprehensive studies of time in the institutionalist tradition have been made by Hayden, who has similarly argued that time is "not a natural phenomenon; rather it is a societal construct" (1987, 1282). As a societal construct, different societies have taken very different perspectives on time that have become more complex as societies have become more complex:
In the simplest technological societies, only a few events had to be synchronized in order to facilitate social life. Time existed only when those events had to be synchronized or when historical occurrences had to be recorded. It did not exist the remainder of the day, week, or year. There were no clocks or a sense of time sequencing. Neither was time divided into units such as weeks or hours. With the evolution of nomadic and agricultural societies, Sociotechnical processes became more complex, thus more synchronization and coordination were needed. Planting, harvesting, and warfare require more refined coordination. In addition, because the seasons became more important to when and where the tribe moved or when crops were planted, the seasons replaced events as the main time instrument. The regular rhythm of an organized society gave the sense of events passing along a time continuum. However the rate of technological change was so slow that life seemed to be repeating itself from year to year and from generation to generation. Therefore, the time construct was thought to be circular. As new technological combinations began to appear more rapidly, it became obvious that society was changing. Therefore, the time continuum ceased to be circular and began to move forward, finally becoming linear. In the industrial era the clock is not just a measure or symbol of passing time. In the minds of that era it is passing time, both operationally and as conscious proof of the passing of time. The clock's 24-hour-per-day, 60-minutes-per-hour and 60-second-per-minute has given the impression of an evenly divided flowing time to those living in an industrial society. (1304-5; emphasis in original) The integration of science (and the notion of relative time), technology, and holistic science results in another construct: real time (Hayden 1987, 1306). Real time, or system time (1311, note 42), refers to the sequential events of a system, rather than to clock time (1298). In real time, "the system determines the measurement instrument.... Real time is defined in a system context that takes account of the appearance, duration, passage, and succession of events as they are interrelated within a system" (1306). In other words, sequential deliveries themselves are the "clock" with which to measure time in modern sociotechnical processes. What exist in society are duration clocks and coordination clocks selected by society, and the sequencing of events as scheduled by societal patterns (1282). Any uniformly flowing time construct that is independent of a process will, as a consequence, be inadequate for analyzing and planning socioeconomic processes in modern societies (Hayden 1993, 105). As Alfred North Whitehead recognized, "The disassociation of time from events discloses to our immediate inspection that the attempt to set up time as an independent terminus for knowledge is like the effort to find substance in a shadow. There is time because there are happenings, and apart...