They assess robustness across different
sample subperiods, past volatility measurement intervals and portfolio holding intervals.
Progressions of average returns across quintiles are not systematic and results
for subperiods are very different, indicating no exploitable non-linearities.
They define the long term (relatively short intervals) as their entire sample period (rolling four -
month subperiods).
The power of aggregate discretionary accruals to predict future market returns is robust across
subperiods return / accrual measurement methods and business conditions.
Results indicate that, while the average correlation is near zero, it drifts higher and lower over relatively
short subperiods.
The modest predictive power found above for the
recent subperiod derives from strong negative momentum in stock market returns after the largest drops in margin debt.
The final chart summarizes average next - month S&P 500 Index return by ranked fifth (quintile) of monthly changes in Consumer Sentiment Index for the overall sample period and two
equal subperiods (break point at the end of 1997).
The adjustment is an approximation using a simplified assumption that returns are not serially correlated
between subperiods.
They assess robustness across different
sample subperiods, past volatility measurement intervals and portfolio holding intervals.
The second set of regressions focused on the betas» temporal stability by estimating the Market Model over thirty - four
month subperiods.
Option trader underperformance is concentrated in the (bullish) April 2003 - March 2006
subperiod, suggesting inordinate reluctance to acknowledge a change in market trend in 2003.
Overall, option traders underperform equity - only traders, with underperformance concentrated during the the second (bullish)
subperiod.
There are 96 (48) observations per quintile for the overall sample period (
subperiods).
The following table, constructed from data in the paper, summarizes average raw monthly returns for individual investors who do and do not trade equity options over the entire sample period and during two
subperiods.
To test robustness of findings, they: (1) account for one - way trading frictions ranging from 0.02 % to 0.05 % across assets; (2) consider five
subperiods to test consistency over time; and, (3) perform out - of - sample tests using the first part of each subperiod to select the best rules and roughly the last year to measure performance of these rules out - of - sample.
They test robustness of findings with data from selected international developed markets, different return variable specifications, different
subperiods and impact of transaction costs.
To compensate for data snooping bias, they specify in - sample and out - of - sample
subperiods and test whether the most successful in - sample rules work out - of - sample.
He considers two test methodologies: (1) straightforward inception - to - date in - sample rule optimization followed by out - of - sample performance measurement, with various break points between in - sample and out - of - sample
subperiods; and, (2) average performance across two sets of bootstrap simulations that preserve relevant statistical features of historical data (including serial return correlation for one set).
These excavations allowed us to refine the Aztec chronology by splitting the late Aztec period into two
subperiods — late Aztec A (13501440) and late Aztec B (14401519)-- to yield a more detailed analysis.
Results using consensus analyst forecasts rather than lagged earnings to calculate E / P over the 1977 - 2006
subperiod are similar, but not as strong.
Using return data and strategy descriptions spanning a total of 6,352 hedge funds over the period January 1970 through June 2009 and risk factor adjustment data for a January 1994 through March 2009
subperiod, he concludes that: Keep Reading
A measure of return calculated by averaging the return for
each subperiod in which a cash flow occurs into a return for a reporting period.
The 1963 - 1990 relation between BE / ME and average return is strong, and remarkably similar for the 1963 - 1976 and 1977 - 1990
subperiods.
Good point — How does one adjust a series to have a zero average during
a subperiod that has been deleted?