The economics of recent bond yield volatility

BIS Economic Papers No 45
July 1996

Introduction

The extreme fluctuations in yields that swept bond markets in 1994 surprised traders, portfolio managers, treasurers and central bankers. Observers from countries that were enjoying low inflation for the first time in a generation had come to hope that their financial markets had seen the last of such bouts of volatility. Those from countries with strong records of price stability saw the volatility as puzzling or even as an undeserved and undesirable intrusion from abroad. Those who had welcomed rising bond prices and low volatility in 1992-93 as evidence of increased confidence in official commitments to price stability suddenly found themselves groping for different explanations for falling bond prices and disorderly markets.

These episodes of bond market turbulence have added force to the question, raised in some quarters, of whether markets have become too powerful. One way of posing this issue is to ask whether this volatility reflects economic fundamentals of inflation, growth and policy, or whether it represents a self-generated force sweeping across markets and bearing little relation to such fundamentals.

In fact, comparatively little is generally known about the forces driving volatility. Market participants may be years ahead of the journals, as Robert Merton has argued, but they rarely gather or disseminate what they know. To be sure, managers of options trading desks feed volatilities for various markets into the routines that evaluate their positions; and risk controllers track the value of these positions under a variety of scenarios. These activities lead to a great deal of analysis of the short- term determinants of volatility. Nevertheless, much of what is learned remains proprietary. For their part, perhaps influenced by the demands of the financial marketplace, scholars have concentrated research on the very short-term estimation of the time series properties of volatility. Moreover, the October 1987 Crash's salience has channelled most of the work into the stock market, leaving some attention to the foreign exchange market. By contrast, the study of volatility in fixed income markets in general, and the bond market in particular, remains a substantial lacuna. Against this background, we aim to move the analysis of bond volatility out of the awkward corner in which it now stands.

This paper incorporates the insights of short-term analysis of volatility and then seeks to relate fluctuations in volatility over time and across markets to domestic economic factors and international influences. In particular, we set out to ground bond volatility in market participants' uncertainties regarding each country's inflation, growth, fiscal policy and money market yields. We also investigate the role of international factors, not only the spillover of volatility from one national market to another, but also the influence of increasingly mobile international capital flows. We address this work in the first instance to central bankers, who have at least two stakes in the study of bond volatility. Central banks have set the task for themselves of extracting from such derivative instruments as options information regarding market participants' expectations and balancing of risks. In addition, supervisors of financial firms at central banks and elsewhere need to understand the risks that banks are taking, including their positioning in volatility. At the same time, we address market participants, on whose insights we have drawn but who may find useful our bringing together in one place analysis of different markets and times. And we hope that scholars of volatility recognise in this work their common understanding and find our study of the link between volatility and economic factors of some interest.

Our results can be summarised briefly. Building on the well-developed short-term analysis of volatility, we confirm that when bond volatility rises, it persists at high levels for weeks. We find that volatility generally responds more markedly to sell-offs in the underlying market than to rallies and suggest novel explanations for this regularity.

As regards domestic macroeconomic determinants, we argue that an economy's inflation record and expectations increase bond volatility over long periods and across countries. Fiscal imbalances, too, may have a similar effect. Near-term revisions in inflation or growth expectations, however, do not track the short-term evolution of volatility. At least in the extreme case represented by Italy, variations in market participants' apprehensions about the state of public sector finances appear to have some influence.

Volatile money markets in general make for more volatile bond markets. Such money market volatility can in principle be related to the modus operandi; of monetary policy, but the precise nature of the link remains an important question for research.

On the international side, we find evidence of spillovers and of a powerful and hitherto unappreciated influence of foreign disinvestment. Spillovers gained in strength and geographical scope in the period of market turbulence in 1994. They tended to emanate from New York and, to a lesser extent, London. For Germany, France and Italy, we find that the rapid withdrawal of foreign investors was associated with a sharp increase in volatility.

The study is organised as follows. Section I explains the definition of volatility underlying the analysis and deals with issues of measurement. Much of the work is based on implied yield volatilities derived from options on government benchmark bonds for thirteen industrialised countries. Section II examines the time-series properties of bond yield volatility and its relationship with market movements. A key issue considered is whether volatility is directional, rising in periods of rising yields (falling prices). Section III then assesses the short-run and long-run link between bond yield volatility and fundamental domestic macroeconomic factors. The focus is on inflation, growth and fiscal policies. Section IV studies part of the link between bond volatility and monetary policy, through an evaluation of the relationship between bond yield and money market volatility. Section V then turns to the international dimension. It examines the international transmission of volatility and the specific role played by sales and purchases by non-residents. The conclusions summarise the main findings of the study and draw a number of policy implications. Inevitably, a study of this kind relies on statistical and econometric evidence to substantiate some of the claims made. We have relegated the more technical aspects to extensive annexes.