Exploring aggregate asset price fluctuations across countries: measurement, determinants and monetary policy implications

BIS Economic Papers No 40
April 1994

Introduction

Asset prices have undergone major medium-term fluctuations since the early 1980s, sometimes ending in sharp downward corrections. This has been true of equity prices: a generalised vigorous upswing preceded the global stock market crash of October 1987, while a more protracted cycle took place in Japan. It has also been true of real estate prices, as sizable movements in residential and commercial property prices have occurred in many industrialised countries (e.g. BIS (1990) and (1993)).

Such fluctuations are of course not new; the last similar episode took place in the early 1970s. Yet the recent cases have attracted particular attention on the part of economists and policy-makers. One reason is their impact on economic activity and on the soundness of financial institutions. It is widely believed that the boom-bust nature of asset price fluctuations has exacerbated the business cycle, fuelling the upswing, magnifying the downswing and slowing down the current recovery. And the disruption caused to balance sheets of economic agents, notably banks, has threatened or resulted in widespread financial distress: the experience of some Anglo-Saxon and Nordic countries as well as Japan are prominent examples (BIS (1991), (1992) and (1993)). A second reason is that the prolonged upswing, in contrast to that in the early 1970s, occurred against the backdrop of generally moderate and declining inflation and typically positive inflation-adjusted interest rates, often higher than output growth rates. Questions about the determinants of such asset price fluctuations and about the extent to which monetary authorities should pay attention to them in the formulation of policy have thus come to the forefront of debate.

Against this background, the objective of the present study is threefold. Firstly, it is to develop an aggregate asset price index for several of the major industrialised countries so as to summarise the information contained in the separate movements of the three asset prices exhibiting major fluctuations, viz. residential property, commercial property and share prices." Such an index facilitates the comparison of broad asset price movements over time and across countries, gives some empirical content to popular notions of general asset price "inflation" and "deflation" and may highlight patterns of behaviour that would otherwise remain undetected. Secondly, it is to begin to analyse what factors may explain the observed movements in the index. The approach tries to combine basic insights from microfinance theory and macroeconomics. Finally, it is to provide preliminary evidence on the usefulness of the aggregate asset price index as an input in the design of monetary policy.

The analysis is largely exploratory in nature. The theoretical underpinnings of the index and its construction could be refined; the development of a formal representation of the workings of the economy would permit a more unified treatment of the various issues addressed; greater attention to country-specific features and more thorough statistical testing would be needed to obtain more definitive answers. Tackling these aspects, however, would take the study well beyond its intended scope. Section I briefly reviews the main characteristics of asset price movements during the last two decades; a detailed description of the construction of the aggregate price indices for the various countries is contained in Appendix I.

Section II looks at the possible determinants of the observed fluctuations in the aggregate index. After outlining some of the conceptual underpinnings of the analysis, the section provides a stylised comparison of fluctuations in the 1980s and 1970s. On the basis of an examination of very simple relationships between asset prices, output, profits, interest rates and credit, it is argued that a distinguishing feature of the more recent episode was the role played by the relaxation of credit constraints in the wake of financial liberalisation. This hypothesis is then subjected to more formal econometric tests.

Section III, by contrast, takes aggregate asset prices as given and considers their potential usefulness as an element in the design of monetary policy. Two popular types of analysis are performed. The first subsection examines the extent to which the inclusion of the aggregate index improves the performance and stability of traditional demand for money functions; in several cases the improvement of the economic and statistical properties of the equations is substantial. The second sub-section explores the information content of the index with respect to inflation and output. Part of the analysis applies popular statistical techniques already extensively used in the assessment of other potential leading indicators, such as simple monetary aggregates, divisia indices, credit and various interest rate spreads (so-called "Granger-causality" tests). A fuller explanation of these tests is given in Appendix II. The section also examines the same question on the basis of the ability of aggregate asset price movements to explain the forecasting errors of a major international institution.

The conclusions briefly summarise the main findings of the paper. They also highlight some of the key lessons of the recent asset price fluctuations for central banks as the institutions responsible for the conduct of monetary policy but also as guarantors of the integrity of the financial system.

JEL classification: G12, E5, E58