Open mouth operations

Abstract

This paper explains how central bank statements, rather than open market operations, can be used to implement monetary policy. In the extreme, policy instruments can be held constant, and yet interest rates will evolve along the path desired by the central bank. We show how the recent implementation of monetary policy in New Zealand works in this way. Using announcement data from New Zealand, we find that open mouth operations lead to large changes in interest rates across all maturities, and these changes cannot be explained by open market operations. Implications are drawn for monetary policy in other jurisdictions.

Introduction

Recent New Zealand monetary policy experience highlights an important channel by which a central bank can affect interest rates, which is via statements on its desired path for the short-term interest rate. New Zealand is striking because these statements – which we call open mouth operations – were exclusively used to implement monetary policy during the period we study (January 1989–September 1997), while the conventional tool of open market operations was used solely to target a level of daily settlement cash balances that was very rarely changed.¹

In this paper, we develop a model of monetary policy implementation in New Zealand which shows how the Reserve Bank of New Zealand (RBNZ) is able to control interest rates without changing conventional policy instruments. We also provide empirical evidence about the relative magnitude of interest rate adjustments arising from open mouth operations versus open market operations.

Although our interest in this topic was stimulated by the New Zealand experience, we believe that many central banks currently conduct monetary policy with open mouth operations and have done so in the past. Incorporating this informational channel into historical studies of monetary policy in other countries may therefore resolve a number of empirical puzzles, including why the liquidity effect is so hard to identify, despite the apparent ease with which monetary authorities can move overnight rates by any desired amount.

In our view the liquidity effect exists, but its use is not required to change market rates.² Open mouth operations can be used for this purpose. They work because implicit in a statement is a credible threat that, if market rates do not move to the announced level, the liquidity effect will be exploited to ensure rates do move. We argue that the central bank's threat to dry up, or flood, the market for bank reserves is sufficient to tie down the path of the overnight interest rate, as if it was using open market operations to achieve this path.

To illustrate how this works in practice, we provide a model of monetary policy implementation in New Zealand which shows exactly how the RBNZ is able to control interest rates without changing policy instruments. The key features are that the RBNZ uses open market operations solely to target a constant nominal stock of settlement cash balances, keeps deviations from this target within a constant level of forecast accuracy (in nominal terms), and pays interest on settlement cash and charges a discount rate, both of which move automatically (according to a simple formula) with market rates. This set-up, together with data on daily liquidity shocks caused by unanticipated flows between the government and private sector, allows us to measure the impact of statements on interest rates, controlling for the level of liquidity in the market. Using announcements by the RBNZ, reported by Reuters and codified by us, we find that interest rates (including the overnight rate) jump immediately after announcements in the direction desired by the Bank, and that interest rate changes are not caused by changes in open market operations.

In many countries, open mouth operations take the form of signals, which involve a change in the structure of open market operations for a single day. For instance, Feinman (1993) shows that different types of open market operations were chosen by the Federal Reserve during the period 1984–1990 to keep agents informed of the underlying stance of monetary policy. In a recent Bank for International Settlements survey on monetary policy implementation, Borio (1997, Chapter 5) details the signalling strategies used in 14 different countries. Such signals were also part of the RBNZ's communication strategy up until the early 1990s. Using data on RBNZ signals, we show that signals lead to sizeable changes in interest rates, and again we show this cannot be because of their liquidity impact. Thus for one country we are able to answer the question posed by Borio (p. 43):

Supplying, say, a somewhat larger amount [of reserves] than that targeted by banks is expected to put downward pressure on the overnight rate. It is still an open question, however, how much of the downward pressure occurs through a mechanical liquidity effect or, more fundamentally, through the signal conveyed regarding policy intentions.

The remainder of the paper proceeds as follows. Section 2 presents our theory of threat-based monetary policy, discusses the role of open mouth operations in this theory, and explains how money market equilibrium can be maintained under this policy. Section 3 explains how this theory has been put into practice in New Zealand. The data and codification of open mouth operations is described in Section 4. Section 5 provides an empirical investigation of open mouth operations. Finally, Section 6 concludes.