عدم تقارن در بازگشت و نوسانات سرایتی بین بازارهای سهام و اوراق قرضه در استرالیا

We document asymmetry in return and volatility spillover between equity and bond markets in Australia for daily returns during the period 1992–2006 using a bivariate GARCH modelling approach. Negative bond market returns spillover into lower stock market returns whereas good news originating in the equity market leads to lower bond returns. Bond market volatility spills over into the equity market but the reverse is not true. Transmission of bond volatility into equity volatility depends in a complex way upon the respective signs of the return shocks in each market.

Integration of financial markets facilitates complex interdependencies in the returns and volatilities of related markets. Spillover effects between equity and bond markets arise when information transmission between these markets is neither instantaneous nor complete. Consider the impact on equity prices of the release of macro-economic news that produces an unexpected fall in bond prices. This may result in a corresponding decline in equity prices if it signals that corporate cash flows will be lower in the future because the higher interest rates are expected to inhibit economic growth. In this case, spillover flows from the bond market into the equity market. Another scenario is when falling equity prices are associated with increased market uncertainty and rising discount rates for all assets including bonds whose prices fall in response. Here the spillover is from the equity market into the bond market. Understanding the nature of these kinds of spillovers is important for diverse purposes such as asset allocation, portfolio management, financial risk management and capital market regulation. To explore the interdependence of stock and bond markets, our empirical analysis estimates a bivariate GARCH model with spillover and asymmetric effects. It identifies asymmetries in the nature of return and volatility spillovers between the equity market and the government bond market in Australia. Negative bond market returns spillover into lower stock market returns in the subsequent trading day. In contrast, good news originating in the equity market leads to lower bond returns. We also find that while bond market volatility spills over into the equity market, the reverse is not true. Further, the impact of bond market volatility on future equity volatility is contingent upon the respective signs of the news in each market. For instance, the spillover is greatest when the equity shock is negative and the bond shock is positive. These results provide evidence relevant to the ongoing debate about the comparative importance of several competing explanations for financial market interdependence. In particular, the main views covered in the literature are: asset substitution; financial contagion (e.g., King and Wadhwani (1990)); hedging demand shifts (e.g., Fleming et al. (1998)); news specificity; news decomposition (e.g., Campbell and Vuolteenaho (2004)) and asymmetric price adjustment (e.g., Koutmos (1999)). For example, in terms of our findings, the spillover of positive equity returns into lower bond returns is compatible with the idea that good news in the equity market is interpreted as a signal that investors should substitute equities for bonds in their portfolios in order to capture improved cash flows in the corporate sector. Alternatively, while the news specificity hypothesis is able to explain our two observed significant return spillovers, it fails to predict the absence of the other potential spillover effects. Numerous studies have investigated the linkages between returns and volatilities of international equity markets.1 A smaller number of papers have recently begun to examine these relations among global bond markets.2 In contrast, comparatively little attention has been given to exploring spillover effects between stock and bond markets within individual nations.3 We find this somewhat puzzling as the debt capital markets provide an important and rich information set for equity market participants. Bond pricing dynamics have obvious implications for the pricing of equities. For instance, simple present value models capture the notion that equity prices respond to changes in discount rates which are due, inter alia, to interest rate changes. Our paper proceeds as follows. Section 2 conceptually analyses why and how spillovers in return and volatility arise within the stock and bond markets. Section 3 motivates and specifies the empirical models we estimate. Section 4 describes the data and Section 5 presents the empirical findings. Section 6 briefly concludes.

This paper documents return and volatility spillover effects between the equity and bond markets of Australia using a GARCH-based model. We find that the stock-bond spillover dynamics are strongly asymmetric. In respect of return volatility, bad news sourced from the bond market spills over into lower equity returns while good news from the equity market leads to lower bond returns. Spillovers flow from the bond market into the equity market volatility at various rates dependent upon the respective signs of the return shocks in both markets. Conversely, bond market volatility does not appear to be affected by conditions in the equity market. Although we consider several different theoretical explanations, none of these are able to fully explain the observed spillovers. While our results are insightful for the domestic Australian setting, a valid question likely posed by many readers will be: how generalizable are the findings to other markets in the Pacific Rim region? To what extent can we transport the core messages to these other markets? What caveats/cautions are appropriate? The first point to make is that it is inadvisable to apply our results to emerging markets, even in those few cases where some type of debt market does exist. However, it is by no means safe to assume broad applicability to all developed markets. One way of gaining an indicative feel for likely answers to these questions is to identify those aspects of the Australian setting that would be common and/or natural points of linkage to other regional markets. Most notably, similarity to Canada (in size/industrial composition); the US (general market development) and New Zealand (traditional close economic/cultural ties and geographic proximity) would suggest that our findings would be largely transportable to these countries. On the other hand, the contrasting settings of countries like Japan would suggest that due caution be applied with regard to spillover asymmetry characteristics that we document. Of course, these comments are purely speculative. Ultimately, the degree of generalizability is an empirical question—we commend this for future research efforts. At a theoretical level, future research is needed to develop additional explanations for the asymmetric features of spillover apparent in our data. Tested hypotheses could be further sharpened with a better understanding of the extent to which news in these markets relates to future corporate cash flows, discount rates and general economic conditions. Empirically, our tests could be strengthened by the use of intra-daily data. Since spillovers are expected to be transitory in competitive financial markets, short-lived effects can only be detected in ultra high frequency data. Such data will also help to determine the impact of different adjustment speeds on spillover effects and their asymmetries.