Damodaran on Valuation_ Security Analysis for Investment and Corporate Finance ( PDFDrive )

Therearethreedecisionstheanalystmustmakeinsettingup
theregressionjustdescribed.Thefirstconcernsthelengthof
the estimation period. The trade-off is simple: A longer
estimationperiodprovidesmoredata,butthefirmitselfmight
havechangedinitsriskcharacteristicsoverthetimeperiod.
The second estimation issue relates to the return interval.
Returns on stocks are available on an annual, monthly,
weekly,daily,andevenonanintradaybasis.Usingdailyor
intradayreturnswillincreasethenumberofobservationsin
the regression, but it exposes the estimation process to a
significant bias in beta estimates related to nontrading.
26 Forinstance,thebetasestimatedforsmallfirms,whichare
morelikelytosufferfromnontrading,arebiaseddownward
whendailyreturnsareused.Usingweeklyormonthlyreturns
can reduce the nontrading bias significantly.
27 Thethirdestimationissuerelatestothechoiceofamarket
indextobeusedintheregression.Inmostcases,analystsare
facedwithamind-bogglingarrayofchoicesamongindices
whenitcomesto estimatingbetas;there aremorethan 20
broadequityindicesrangingfromtheDow 30 (DowJones
IndustrialAverage)totheWilshire 5000 intheUnitedStates
alone.Onecommonpracticeistousetheindexthatismost
appropriatefortheinvestorwhoislookingatthestock.Thus,
iftheanalysisisbeingdoneforaU.S.investor,theS&P 500
index is used. This is generally not appropriate. By this
rationale,aninvestorwhoownsonlytwostocksshoulduse
anindexcomposedofonlythosestockstoestimatebetas.The
rightindexto useinanalysisshould bedetermined bythe
holdings of the marginal investor in the company being
analyzed.Ifthemarginalinvestors inacompanyholdonly
domesticstocks,wecanusetheregressionsagainstthelocal
indices.Ifthemarginalinvestorisaglobalinvestor,amore