X hits on this document

PDF document

Maximizing Equity Market Sector Predictability in a Bayesian Time Varying Parameter Model* - page 2 / 46

260 views

0 shares

0 downloads

0 comments

2 / 46

1.

Introduction

As the field of finance has struggled to find a successor model to the benchmark CAPM specification, the role conditional macroeconomic factors play in determining investor risk premia and ultimately equity return predictability has come into greater focus. One of the earliest and most straightforward investigations on the role macroeconomic factors play in determining equity returns is that of Chen, Roll and Ross (1986). Using cross-sectional analysis, Chen et al. find that a number of macroeconomic risk factors are significantly priced in the stock market. In another application, Lo and MacKinlay (1997) derive predictive portfolios based on lagged macroeconomic variables that lend themselves to dynamic trading strategies. A further indication of the importance of lagged macroeconomic variables is presented in Ferson and Harvey (1999), where conditional lagged fundamental information included in a risk pricing model renders sorted portfolio attributes in the popular Fama and French (1993) three factor model insignificant.

In addition to the available empirical tests suggesting an important role for conditional macroeconomic information in asset pricing, a pricing model based strictly on prior macroeconomic information has intuitive and theoretical appeal. As pointed out by Roll (1977), any empirical examination of the standard CAPM is theoretically suspect if the chosen proxy for the market portfolio is not truly representative of the entire market. Even providing for a reasonable proxy for the market portfolio, Cochrane (1996) notes that explanations of changes in returns over the business cycle based on expected market returns are hardly useful in establishing what risk factors cause returns of individual portfolios to vary. What is surely of greater interest over the business cycle are what particular macroeconomic forces drive expected returns.

The other obvious advantage of a factor model incorporating strictly lagged information is the

potential application that up to 50% of introduced through

to return predictability. In one such exercise, Lo & MacKinlay (1997) find

the variation in what they term

returns can be explained a Maximally Predictable

by lagged Portfolio.

economic factors Any evidence of

systematic predictability naturally model based on return reaction to

lends itself

to questions

fundamental

information

of market efficiency, however, is much easier to reconcile with

a a

1

Document info
Document views260
Page views260
Page last viewedTue Dec 06 20:41:14 UTC 2016
Pages46
Paragraphs1927
Words14448

Comments