RDP 2000-09: Consumption and Wealth 1. Introduction

A remarkable feature of the Australian economy over recent years has been sustained high growth in private consumption expenditure. During most of the 1990s, consumption growth has been strong with a steady increase in growth rates over the latter part of the decade. Coincident with the growth in consumption has been a fall in the household saving ratio to historically low levels, raising some concerns about consumer spending and the financial position of households. A potential explanation for the strong growth in consumption in recent years is an attendant rise in household wealth, driven by a steady rise in the value of both financial and non-financial household assets.

This paper examines in detail the relationship between private consumption expenditure and wealth in Australia.^[1] In doing so, we address a number of specific issues that arise. The first is whether in fact a stable relationship exists for Australia. The last two decades are characterised by considerable deregulation in financial markets and households now face a more flexible and competitive financial system (Edey and Gray 1996). A further question is whether the components of household wealth, broadly speaking financial and non-financial assets, have differing effects on consumption. In the past, with ownership of financial assets quite concentrated, variations in financial assets might have been relatively unimportant influences on aggregate consumption. Now in Australia, as in many other industrialised economies, both the direct and indirect exposure of consumers to financial markets has broadened substantially so that variations in the value of financial assets may now have more widespread effects. A related concern is whether some of the recent highly publicised demutualisations or share floats have had a discernible effect on consumption, as has been suggested by some commentators. Finally, although the focus of this paper is empirical, we also attempt to place our results within the broader empirical literature and examine whether they are consistent with standard economic theories of aggregate consumption.

Prior to any modelling, it is instructive first to examine the behaviour of consumption and wealth in Australia over the past two decades. Figure 1 neatly summarises many of the recent developments: the strong growth in consumption, the fall in household saving and the strong gains in wealth. The series reported are the ratio of total household consumption expenditure and wealth to household disposable income.^[2] From the figure, we observe very strong growth in consumption relative to household income (and hence a decline in household saving) in recent years. In addition, we observe very strong growth in wealth throughout the 1990s, particularly in recent years. The behaviour of these series has been used to explain the strong growth in consumption in recent years as a response to increased household wealth. Figure 2, however, which shows the ratio of consumption to wealth, casts some doubt upon this argument. While consumption has been increasing, the growth in wealth has exceeded the growth in consumption, especially in recent years. Our objective is to disentangle these relationships to determine what underlies these broad trends.

Much of our empirical analysis considers the components of wealth separately. Figure 3 presents the composition of household wealth (in real per capita terms). Two broad categories are presented, non-financial and financial assets, as well as the total. Non-financial assets are predominantly dwellings while financial assets consist of a variety of assets, including equities held directly and indirectly through life and pension funds. Table 1 provides a summary of the specific components. One immediate observation is the dominant influence of non-financial assets on the behaviour of total household wealth. Two episodes merit comment; the strong growth in non-financial assets in the late 1980s and in the late 1990s, both associated with strong growth in housing prices (particularly in the capital cities). These are large and persistent movements that have an important influence on the empirical analysis. To foreshadow some of our analysis, we find it necessary to focus on data from 1989 onwards, in part because of the persistent increase in wealth in the late 1980s. As the rise in wealth is not reflected in changed consumption patterns, there does not appear to be a stable relationship between consumption and wealth over the whole sample.

This fact, and our reliance on only the more recent data, does qualify the conclusions of our analysis, a point we should stress. We have two periods of strong growth in household wealth (the late 1980s and the late 1990s) and we are using only the information in the latter as the basis for our analysis and conclusions. As a consequence, our results may be specific to the episode we consider and not a stable description of consumption behaviour, a concern that should be borne in mind.

It is helpful to consider further the recent behaviour of the components of household wealth. Table 1 provides a decomposition of the two principal components of wealth, as well as identifying the sources of growth in these components in recent years. Dwellings account for sixty per cent of household assets and the growth in value has been uniformly strong in recent years. Arguably, dwellings are not very liquid assets and we might anticipate consumption to be relatively insensitive to changes in their value. Certainly, this is one possible explanation for the declining consumption to wealth ratio in Figure 2. The other feature of note is the relatively small share of wealth held directly as equities (equities are also held indirectly through life and pension funds). So, despite the very strong gains recently exhibited by this class of assets, the direct effect on consumption from changing equity values may in fact be quite small. This is also an issue we address in our empirical analysis.

The paper is structured as follows. In the following section, we provide a general description of the aggregate consumption models we estimate and relate these to standard theoretical models of consumption. In Section 3 of the paper we present the empirical analysis where we identify what effect, if any, measures of wealth have on consumption decisions. We first demonstrate that it is most sensible, for a number of reasons, to consider a relatively small sample of data, 1989–1999. For this sample, we identify a steady-state relationship between consumption, labour income and wealth. This is then used as a basis for dynamic models of consumption. Section 4 concludes.

Footnotes

There has been relatively little empirical research relating consumption and wealth for Australia. There are a number of Australian studies that examine consumption and disposable income, which implicitly includes asset income; recent examples include Moosa and Kennedy (1998), Olekalns (1997), de Brouwer (1996) and Debelle and Preston (1995). In some respects, this study follows the motivation and methods of a recent US study, Ludvigson and Steindel (1999). [1]

Details of the data are provided in Appendix A. [2]