X hits on this document

PDF document

Maximizing Equity Market Sector Predictability in a Bayesian Time Varying Parameter Model* - page 7 / 46





7 / 46

The specification for Markov switching variance presented here closely follows that of Turner, Startz and Nelson (1989).

2.2.1. The Data: The Sectors

The financial instruments used in this study as dependent variables are the ten sector total return indices as defined by the Global Industry and Classification Standard (GICS), jointly developed by Morgan Stanley Capital International and Standard and Poor’s. The series were downloaded through FACTSET. Returns are constructed by taking the difference in the logs of two consecutive index levels. The frequency of the series is weekly and the period we examine starts on the first week of January 1990 and ends the second week of January 2003. Table 1 presents descriptive statistics for the sample period under consideration.

(Insert Table 1 here)

Figure 1 presents the evolution of an initial $100 investment in each of the 10 GICS sector indices starting on January 12, 1990. The technology sector experienced the highest return up until April 14, 2000, after which point the sector index collapsed. Over the entire sample, the highest return was in the Healthcare sector, while the lowest return was found for the Utilities sector.

2.2.2. The Data: The Factors

The factors chosen for the model are done so to maintain a parsimonious framework and to be consistent with previous advances in the literature. The chosen factors are:

  • divyield - The change in the dividend yield on the S&P 500 composite index

  • spread - The change in the spread between the 10 year treasury note yield and the 90 day treasury bill yield.

  • oil - The percent change in the near month crude oil contract.

  • junk - The change in the default spread, defined as the difference between the Moody’s Baa and Aaa corporate yield.


Document info
Document views323
Page views323
Page last viewedTue Jan 17 12:10:58 UTC 2017