This regression has a low R-squared, but it is more a
reflectionofthenoiseinP/Eratiosthanitisontheregression
methodology. As we will see, the market regressions for
price-to-bookvalueandprice-to-salesratiostendtobebetter
behaved and have a higher R-squared than P/E ratio
regressions.Whilethecoefficientsinthisregressionallhave
the predicted signs—P/E ratios increase with growth and
payoutanddecreaseasriskincreases—thisisnotalwaysthe
case.Infact,similarregressionsrunin 2003 and 2004 hadthe
wrong sign for the beta coefficient, with higher-beta
companies havinghigherP/E ratios insteadof lower ones.
Thisoccurslargelybecausetheindependentvariablesinthis
regression are themselves correlated with each other,with
high-growthcompaniestendingtoberiskywithlowpayout
ratios.
9
PEG Ratio
When comparing PEG ratios across firms, then, it is
importantthatwecontrolfordifferencesinrisk,growth,and
payoutratioswhen makingcomparisons. Althoughwe can
attempttodothissubjectively,thecomplicatedrelationship
between PEG ratios and these fundamentals poses a
challenge. A far more promising route is the regression
approachusedforP/EratiosandtorelatethePEGratiosof
the firms being compared to measures of risk, growth
potential, and the payout ratios for these firms.
AswiththeP/Eratio,thecomparablefirmsinthisanalysis
canbedefinednarrowlyasotherfirmsinthesamebusiness,
moreexpansivelyasfirmsinthesamesector,orasallfirms
inthemarket.Inrunningtheseregressions,allthecaveatsthat