in Handbook of Econometrics vol. IV (eds. Engle,
McFadden), 2295-2339, Elsevier.
DATA SOURCE: DataStream, Ecowin
PREDICTABILITY OF STOCK RETURNS
SSUES:
Stock returns contain systematic elements in their dynamics. These can be caused by calendar effects (e.g.
the January effect) or by autoregressive effects (inertia in the short run, corrections in the medium run). Do
these effects persist once discovered? Are there other effects that can be uncovered?
SELECTED READING:
Campbell, Lo, MacKinlay (1997) The Econometrics of Financial Markets, Princeton University Press. [Chapter
2.]
Schwert (2003) Anomalies and market efficiency, in Handbook of the Economics of Finance vol. 1B (eds.
Constantinides, Harris, Stulz), 937-972, Elsevier.
DATA SOURCE: DataStream, Ecowin
ESTIMATING THE DEGREE OF PERSISTENCE IN VOLUMES AND VOLATILITIES
ISSUES:
There is an increasing realization that autoregressive models are too simplistic to capture the dynamics of
financial series. They imply a sharp divide between series that are integrated of order 0 (stationary) and
series that are integrated of order 1 (random walks and the like). The memory (or persistence) decays
exponentially fast in the first type, while it remains forever high in the second type. Models of fractionally-
integrated series have now been use英语论文网 【http://www.51lunwen.org】d successfully in finance, because they bridge this gap. They can be
applied here to measuring the persistence in the trading volumes &/or variability of individual stocks.
SELECTED READING:
Andersen, Bollerslev (1997) Heterogeneous information arrivals and return volatility dynamics: uncovering the
long-run in high frequency returns, Journal of Finance 52, 975-1005.
Bollerslev, Jubinski (1999) Equity trading volume and volatility: latent information arrivals and common long-
run dependencies, Journal of Business & Economic Statistics 17, 9-21.
DATA SOURCE: DataStream, Ecowin
VARIANCE RATIOS AND VARIOGRAMS
ISSUES:
Variance ratios (VRs) are used in empirical finance in an attempt to see how the variance of a series evolves
as a function of time. One implication of the analysis is to obtain a (quadratic) measure of the degree of
persistence of the time series. This has been done in the narrow context of autoregressive models and
parametric VRs. A more general statistical tool is called the variogram. Using this tool, can you confirm the
results detected by the older method, or do you find other patterns? What do your patterns mean for the
underlying series?
SELECTED READING:
Campbell, Lo, MacKinlay (1997) The Econometrics of Financial Markets, Princeton
本文来自：英语论文网 【http://www.51lunwen.org】
|
论文编号：
org200806170924442119 
