Portal de Periódicos da CAPES

Short-term market timing using the bond–equity yield ratio

2008; Taylor & Francis; Volume: 15; Issue: 4 Linguagem: Inglês

10.1080/13518470802466097

1466-4364

Pierre Giot, Mikaël Petitjean,

Financial Reporting and Valuation Research

Abstract This paper takes a new look at the market-timing ability of the bond–equity yield ratio (BEYR). We compare the short-term profitability of a naive strategy based on the extreme values of the BEYR to the short-term profitability of a sophisticated strategy relying on regime switches. In contrast to previous studies, we do not document any major international evidence that these dynamic strategies deliver significantly higher risk-adjusted returns than the buy-and-hold portfolios. Moreover, the profitability of these active strategies is not improved when the equity yield, instead of the BEYR, is used as a criterion to time the market. Keywords: valuation ratioswitchingregimemarket timing JEL classifications : G14G15G12 Acknowledgements We gratefully acknowledge the helpful comments from seminar participants at CASS Business School, Solvay Business School, CORE, CREST, ULg-HEC, the Third European Deloitte Risk Management Conference in Antwerp, the 11th International Conference on Forecasting Financial Markets in Paris, and the 10th International Conference on Computing in Economics and Finance in Amsterdam. We also thank Dirk-Emma Baestaens, Renaud Beaupain, Carine Brasseur, Chris Brooks, Christophe Dispas, Alain Durré, Peter Hansen, Sébastien Laurent, Andrew Patton, the editor as well as an anonymous referee for their help. The usual disclaimer applies. Notes Key papers on traditional valuation ratios include Basu (Citation1977, Citation1983), Campbell and Shiller (Citation1988, Citation1989, Citation1998, Citation2001), Fama and French (Citation1988, Citation1989, Citation1992), Hodrick Citation(1992), Jaffe, Keim, and Westerfield Citation(1989), Rozeff Citation(1984), and Shiller Citation(1989). For example, US valuation ratios rose spectacularly during the 1990s. The S&P composite P/E ratio hit an all-time peak of 44.2 in December 1999 (more than double its long-term historical mean level). Berge and Ziemba Citation(2003) call it the Bond Stock Earnings Yield Differential (BSEYD). We do not consider time-varying risk premium models à la Campbell and Shiller (Citation1988, Citation1989) since the BEYR approach focuses on the contemporaneous long-run relationship between stock prices, earnings (or dividends), and long-term bond yields. Our methodology is closer to Harasty and Roulet Citation(2000) and to what some market practitioners would like to test. r f is assumed to reflect the short-term interest rates that will prevail in the future; since these rates are not observable, the current long-term yield is generally used as a proxy. It need not follow that the discount rate and cash-flow effects are independent. For instance, a positive (negative) discount rate effect on stock prices can be accompanied by a negative (positive) cash-flow effect when inflation falls (rises). Nevertheless, the cash-flow effect must not offset the discount rate effect. The concept of state dependency in stock and bond co-movements was first theoretically developed by Barsky Citation(1989). Further theoretical arguments and empirical evidence are given in Fleming, Kirby, and Ostdiek Citation(1998), David and Veronesi Citation(2004), Li Citation(2002), Ribeiro and Veronesi Citation(2002), Rigobon and Sack (2003, Citation2004), Guidolin and Timmermann Citation(2007), Scruggs and Glabadanidis Citation(2003) and Connolly, Stivers, and Sun Citation(2005), among others. Recent versions of the Fed Model include the ‘Stock Valuation Models #2’ (SVM-2) introduced by Yardeni Citation(2003) (see Asness Citation2003; Thomas Citation2005; Durré and Giot Citation2007 for a discussion of the Fed Model). The Markov switching approach is not a model in the usual context of the word; it is rather a ‘filter’ since it splits the data into two subsamples with corresponding probability estimates. Since higher degree of certainty about the forecasted regime may lead to higher trading profitability, two alternative trading rules have been followed. First, the 0.9 forecasted probability value has been used instead of the 0.5 value. Second, we have built a dynamic portfolio that is invested in the risk-free rate (resp. bonds/ stocks) when the forecasted probability lies between 0.25 and 0.75 (resp. below 0.25/ above 0.75). There appears to be no significant difference between these strategies. Futures trading costs are not easy to gauge, but Goyal and Welch Citation(2008) argue that a typical contract for a notional amount of $250,000 may cost around $10–30. A 20% movement in the underlying index about the annual volatility would correspond to $50,000, which would come to around 5 bp for a single transaction. The correlation matrices of the BEYR are not reported to save space. They are available from the authors upon request. This never happens in Brooks and Persand Citation(2001) as they (unusually) proxy the risk-free rate by the return on long-term bonds.