Question

1. Discuss in detail what an outlier is and explain how it can be a disadvantage of using the High-Low method.
2. The CEO of Keechen & Moore has heard about some firms using regression analysis (method of least squares) to estimate costs. Briefly explain the regression method and discuss its main advantage relative to the High-Low method.

Answer 1

In iicost iiaccounting, iithe iihigh-low iimethod iiis iia iiway iiof iiattempting iito iiseparate iiout iifixed iiand iivariable iicosts iigiven iia iilimited iiamount iiof iidata. iiThe iihigh-low iimethod iiinvolves iitaking iithe iihighest iilevel iiof iiactivity iiand iithe iilowest iilevel iiof iiactivity iiand iicomparing iithe iitotal iicosts iiat iieach iilevel. iiIf iithe iivariable iicost iiis iia iifixed iicharge iiper iiunit iiand iifixed iicosts iiremain iithe iisame, iiit iiis iipossible iito iidetermine iithe iifixed iiand iivariable iicosts iiby iisolving iithe iisystem iiof iiequations.

1. The iihigh-low iimethod iiis iia iisimple iiway iito iisegregate iicosts iiwith iiminimal iiinformation.

2. The iisimplicity iiof iithe iiapproach iiassumes iithe iivariable iiand iifixed iicosts iias iiconstant, iiwhich iidoesn't iireplicate iireality.

3. Other iicost-estimating iimethods, iisuch iias iileast-squares iiregression, iimight iiprovide iibetter iiresults, iialthough iithis iimethod iirequires iimore iicomplex iicalculations.

The iicosts iiassociated iiwith iia iiproduct, iiproduct iiline, iiequipment, iistore, iigeographic iisales iiregion, iior iisubsidiary, iiconsist iiof iiboth iivariable iicosts iiand iifixed iicosts. iiTo iidetermine iiboth iicost iicomponents iiof iithe iitotal iicost, iian iianalyst iior iiaccountant iican iiuse iia iitechnique iiknown iias iithe iihigh-low iimethod.

The iihigh-low iimethod iiis iiused iito iicalculate iithe iivariable iiand iifixed iicost iiof iia iiproduct iior iientity iiwith iimixed iicosts. iiIt iitakes iitwo iifactors iiinto iiconsideration. iiIt iiconsiders iithe iitotal iidollars iiof iithe iimixed iicosts iiat iithe iihighest iivolume iiof iiactivity iiand iithe iitotal iidollars iiof iithe iimixed iicosts iiat iithe iilowest iivolume iiof iiactivity. iiThe iitotal iiamount iiof iifixed iicosts iiis iiassumed iito iibe iithe iisame iiat iiboth iipoints iiof iiactivity. iiThe iichange iiin iithe iitotal iicosts iiis iithus iithe iivariable iicost iirate iitimes iithe iichange iiin iithe iinumber iiof iiunits iiof iiactivity.

The iiDifference iibetween iithe iiHigh-Low iiMethod iiand iiRegression iiAnalysis

The iihigh-low iimethod iiis iia iisimple iianalysis iithat iitakes iiless iicalculation iiwork. iiIt iionly iirequires iithe iihigh iiand iilow iipoints iiof iithe iidata iiand iican iibe iiworked iithrough iiwith iia iisimple iicalculator. iiIt iialso iigives iianalysts iia iiway iito iiestimate iifuture iiunit iicosts. iiHowever, iithe iiformula iidoes iinot iitake iiinflation iiinto iiconsideration iiand iiprovides iia iivery iirough iiestimation iibecause iiit iionly iiconsiders iithe iiextreme iihigh iiand iilow iivalues, iiand iiexcludes iithe iiinfluence iiof iiany iioutliers.

Regression iianalysis iihelps iiforecast iicosts iias iiwell, iiby iicomparing iithe iiinfluence iiof iione iipredictive iivariable iiupon iianother iivalue iior iicriteria. iiIt iialso iiconsiders iioutlying iivalues iithat iihelp iirefine iithe iiresults. iiHowever, iiregression iianalysis iiis iionly iias iigood iias iithe iiset iiof iidata iipoints iiused, iiand iithe iiresults iisuffer iiwhen iithe iidata iiset iiis iiincomplete.

It's iialso iipossible iito iidraw iiincorrect iiconclusions iiby iiassuming iithat iijust iibecause iitwo iisets iiof iidata iicorrelate iiwith iieach iiother, iione iimust iicause iichanges iiin iithe iiother. iiRegression iianalysis iiis iialso iibest iiperformed iiusing iia iispreadsheet iiprogram iior iistatistics iiprogram.

Advantages ii& iiDisadvantages iiof iiHigh-Low iiMethod ii

Fixed iiand iivariable iiexpenses iimean iidifferent iithings iito iithe iiaccountant. iiFixed iiexpenses iiremain iiunchanged iiin iitotal iithroughout iithe iiyear. iiThese iiexpenses iidecrease iias iiproduction iiincreases. iiVariable iiexpenses iiremain iiunchanged iiper iiunit. iiVariable iicosts iiincrease iias iisales iiincrease. iiAccountants iisegregate iifixed iiand iivariable iiexpenses iiwith iithe iihigh-low iimethod. iiSeveral iiadvantages iiand iidisadvantages iiare iiinherent iiin iithis iimethod.

1. Ease iiof iiUse

A iimajor iiadvantage iiof iithe iihigh-low iimethod iiof iicost iiestimation iiis iiits iiease iiof iiuse. iiBy iionly iirequiring iicost iiinformation iifrom iithe iihighest iiand iilowest iiactivity iilevel iiand iisome iisimple iialgebra, iimanagers iican iiget iiinformation iiabout iicost iibehavior iiin iijust iia iifew iiminutes. iiIn iiaddition iito iibeing iieasy iito iiuse, iibecause iithe iimethod iidoesn't iirequire iiany iikind iiof iitools iior iiprograms, iiit iiis iialso iiinexpensive iito iiimplement.

2. Accuracy iiunder iiStable iiCosts

If iicosts iiare iirelatively iistable iiover iitime, iiand iithe iihigh iiand iilow iiactivity iilevel iiare iirepresentative iiof iithe iicompany's iicost iibehavior iiover iitime, iithe iihigh-low iimethod iican iibe iiextremely iiaccurate. iiHowever, iian iiinteresting iiconundrum iioccurs iiif iithe iiendpoints iiare iinot iirepresentative. iiEven iiif iicosts iiare iivery iistable iithroughout iithe iirest iiof iithe iirange iiof iiactivity, iiif iithe iilowest iiof iihighest iilevel iiof iiactivity iiare iisystematically iidifferent, iithen iimanagers iiwill iihave iiinaccurate iiinformation. iiTo iiguard iiagainst iithis, iimanagers iimay iiwant iito iiplot iiactivity iilevels iivs. iicosts iifor iia iisubset iiof iithe iidata iithat iithe iicompany iihas. iiThis iiway, iimanagers iimay iibe iiable iito iisee iiif iithe iipoints iithey iihave iiselected iifor iithe iihigh-low iimethod iireally iiare iirepresentative iiof iinormal iicosts.

3. Inaccuracy iiwith iiVariation

A iiproblem iiwith iithe iihigh-low iimethod iiis iithat iiif iicosts iiare iirelatively iiunstable, iithis iimethod iicould iiproduce iiinaccurate iiresults. iiBecause iithe iihigh-low iimethod iiuses iionly iitwo iipoints iito iicalculate iia iicost iiestimate, iimonthly iivariation iiin iicosts iiis iinot iicaptured iiin iithe iiestimate, iieven iiif iithe iipoints iiare iirepresentative iiof iinormal iicost iibehavior. iiIf iimanagers iiplot iithe iiactivity iilevel iivs. iicost iiand iisee iihigh iivariability, iithey iihave iia iidecision iito iimake. iiIf iitimeliness iiis iimore iiimportant iithan iiaccuracy, iithen iithe iihigh-low iimethod iiis iiprobably iigood iienough. iiHowever, iiif iiaccuracy iiis iitantamount, iithen iianother iimethod iishould iibe iiconsulted.

4. Multiple iiSteps

Another iidisadvantage iiof iithe iihigh-low iimethod iiis iithe iinumber iiof iisteps iinecessary iito iiperform iithis iianalysis. iiThe iiaccountant iineeds iito iigather iimonthly iidata iiregarding iithe iiexpense iibeing iianalyzed iiand iithe iiunit iiof iiactivity. iiThe iiaccountant iilists iieach iiset iiof iidata iiand iiidentifies iithe iihigh iiand iilow iivalues. iiShe iicalculates iithe iidifference iibetween iithese iisets iiof iivalues. iiShe iidivides iithe iidifference iiin iidollars iiby iithe iidifference iiin iiactivity iito iicalculate iithe iicost iiper iiunit iiof iiactivity, iior iithe iivariable iiactivity. iiShe iimultiplies iithe iivariable iicost iiper iiunit iiby iithe iinumber iiof iiactivities iito iicalculate iithe iitotal iivariable iicost. iiShe iisubtracts iithe iitotal iivariable iicost iifrom iithe iitotal iicost iito iidetermine iithe iifixed iicost. iiEach iiadditional iistep iiincreases iithe iipotential iifor iierrors.

5. Estimate

A iidisadvantage iiof iithe iihigh-low iimethod iiis iithat iithe iiresults iiare iiestimates, iinot iiexact iinumbers. iiAn iiaccountant iiwho iineeds iito iiknow iithe iiexact iidollar iiamount iiof iifixed iiexpenses iieach iimonth iishould iicontact iia iivendor iidirectly.

6. Informal iiAnalysis

One iiadvantage iiof iithe iihigh-low iimethod iiis iithe iilack iiof iiformality iirequired. iiThe iiaccountant iican iianalyze iithese iinumbers iiusing iidata iifrom iithe iimonthly iiexpenses iiand iithe iiactivity iilevel. iiHe iidoes iinot iineed iito iicontact iianyone iioutside iiof iithe iicompany iito iidetermine iithe iifixed iiexpenses iior iithe iivariable iirate iiper iiunit.

7. Limited iiData

Another iiadvantage iiof iithis iimethod iiis iithat iiit iionly iirequires iitwo iisets iiof iinumbers iito iicalculate iithe iifixed iiand iivariable iicosts. iiThese iiinclude iithe iiactivity iilevel iiand iithe iitotal iicost. iiThe iiaccountant iireviews iithe iifinancial iitransactions iifor iithe iiaccount iiover iiseveral iimonths iito iiobtain iithe iitotal iicost iiamount. iiShe iireviews iidepartment iirecords iito iidetermine iithe iiactivity iilevels iifor iithose iisame iimonths. iiAfter iigathering iidata iifrom iithese iitwo iiplaces, iithe iiaccountant iihas iiall iithe iiinformation iishe iineeds iito iiperform iithe iianalysis.

ii ii ii ii ii ii ii ii ii ii ii ii ii ii ii ii iiThe iihigh-low iimethod iiis iirelatively iiunreliable iibecause iiit iionly iitakes iitwo iiextreme iiactivity iilevels iiinto iiconsideration. iiThe iihigh iior iilow iipoints iiused iifor iithe iicalculation iimay iinot iibe iirepresentative iiof iithe iicosts iinormally iiincurred iiat iithose iivolume iilevels iidue iito iioutlier iicosts iithat iiare iihigher iior iilower iithan iiwould iinormally iibe iiincurred. iiIn iithis iicase, iithe iihigh-low iimethod iiwill iiproduce iiinaccurate iiresults.

The iihigh-low iimethod iiis iigenerally iinot iipreferred iias iiit iican iiyield iian iiincorrect iiunderstanding iiof iithe iidata iiif iithere iiare iichanges iiin iivariable iior iifixed iicost iirates iiover iitime iior iiif iia iitiered iipricing iisystem iiis iiemployed. iiIn iimost iireal-world iicases, iiit iishould iibe iipossible iito iiobtain iimore iiinformation iiso iithe iivariable iiand iifixed iicosts iican iibe iidetermined iidirectly. iiThus, iithe iihigh-low iimethod iishould iionly iibe iiused iiwhen iiit iiis iinot iipossible iito iiobtain iiactual iibilling iidata.

i i i i i i i i i i i i i i i i i i i i i i iLeast isquares iregression imethod iis ia imethod ito isegregate ifixed icost iand ivariable icost icomponents ifrom ia imixed icost ifigure. iIt iis ialso iknown ias ilinear iregression ianalysis.

Least isquares iregression ianalysis ior ilinear iregression imethod iis ideemed ito ibe ithe imost iaccurate iand ireliable imethod ito idivide ithe icompany’s imixed icost iinto iits ifixed iand ivariable icost icomponents. iThis iis ibecause ithis imethod itakes iinto iaccount iall ithe idata ipoints iplotted ion ia igraph iat iall iactivity ilevels iwhich itheoretically idraws ia ibest ifit iline iof iregression.

Linear iregression iis ibasically ia imathematical ianalysis imethod iwhich iconsiders ithe irelationship ibetween iall ithe idata ipoints iin ia isimulation. iAll ithese ipoints iare ibased iupon itwo iunknown ivariables; ione iindependent iand ione idependent. iThe idependent ivariable iwill ibe iplotted ion ithe iy-axis iand ithe iindependent ivariable iwill ibe iplotted ito ithe ix-axis ion ithe igraph iof iregression ianalysis. iIn iliteral imanner, ileast isquare imethod iof iregression iminimizes ithe isum iof isquares iof ierrors ithat icould ibe imade ibased iupon ithe irelevant iequation.

The ileast isquares iregression imethod ifollows ithe isame icost ifunction ias ithe iother imethods iused ito isegregate ia imixed ior isemi ivariable icost iinto iits ifixed iand ivariable icomponents. iLet’s iassume ithat ithe iactivity ilevel ivaries ialong ix-axis iand ithe icost ivaries ialong iy-axis. iThe ifollowing iequation ishould irepresent ithe ithe irequired icost iline

Least-squares iiregression iiis iia iistatistical iitechnique iithat iimay iibe iiused iito iiestimate iia iilinear iitotal iicost iifunction iifor iia iimixed iicost, iibased iion iipast iicost iidata. iiThe iicost iifunction iimay iithen iibe iiused iito iipredict iithe iitotal iicost iiat iia iigiven iilevel iiof