CasereportGroup12.docx

CasereportGroup12.docx

《CasereportGroup12.docx》由会员分享，可在线阅读，更多相关《CasereportGroup12.docx（12页珍藏版）》请在冰豆网上搜索。

CasereportGroup12

CaseReport

Delta’sNewSong

ByGroup12

2011/3/29

Contents

1.Background

2.Mainissue

3.Problems

4.Analysis

1）High-lowforDelta

2）SimpleRegressionforDelta

3）MultipleRegressionforDelta

4）Conditionsandassumptionsofcostfunctions

5）High-lowforJetBlue

6）SalarycostestimationforSong

5.Solutionandsuggestion

Background

TheentireUSairlinebusinessisfacingthechallengeofoperatingwithinalow-margin,high-fixed-costenvironment.Itsprofitabilityisparticularlysensitivetodecreasesinvolume,eitherfromenvironmentalfactorsorfromcompetition.Moreover,theairlinebusinessislabor-intensive.Laborcostsasapercentageofrevenuesrangesfromalowofabout25percentforthelow-fareairlinestoalmost50percentforthelarge,full-serviceairlinessuchasUnited.Furthermore,formanyairlines,laborunionsarestrong,presentinganadditionalchallengeinthemanagementofcost.Salariesarelargelyfixedintheshorttermforunionizedemployees.Inrecentyears,somemajorairlineshavewonconcessionsfromlaborunionsaboutreducingthestaffandcuttingdownsalaries.

DeltaAirlineisthethirdlargestU.S.airlineinoperatingrevenuesandrevenuepassengermilesflown.Traditionally,thecompetitioncamefromtheotherfull-serviceairlinessuchasUnitedAirlinesandAmericanAirlines.However,inrecentyears,themajorairlineshavebeenforcedtocompetewithlow-cost,no-frillairlinespioneeredbySouthwestAirlines.SalariesareasignificantcomponentofDelta’scoststructure.Deltapilotsarethehighestpaidintheindustry.Meanwhile,itistheleastunionizedofthemajorairlines.Deltapilotsaretheonlyunionizedemployeegroup.Itsflightattendantandticketagentsarenotunderunioncontract.Consequently,theirsalariespresentflexibilityandvariabilityinnature.

InNovember2002,despiteitspreviousfailureinentryintothelow-faremarket,Deltadecidedtoformanewlow-costcarrier,Song,whichistargetedtocompetewithsuccessfulnewcomerJetBlue.

Mainissue

HowcanDeltacreateadifferentcoststructureandbusinessmodelinordertosucceedinthelow-costcarriermarket?

Problems

Deltaisinapositionofevaluatingentryintothelow-costcarriermarket.ThesuccessofDelta’sSongdependsonthefollowingissues:

ØCanDeltacreateaverydifferentcoststructureandanewbusinessmodeltocompetewithJetBlue?

ØHowtopredictfuturesalariesofDeltaandJetBlue?

ØHowtodealwiththestrongworkunion?

Analysis

1.High-lowmethod

Firstly,weidentifysomepossiblecostdriverstoestimateDelta’ssalaries:

availableseatmiles;availabletonmiles;numberofdepartures;revenuepassengermiles;revenuetonmiles.Afterapplyingsimpleregressionusingeachofthepossibledriversandcomparing

andresidualerrorsofeachdriver,wechoosetwomostreasonablecostdrivers:

revenuepassengermilesandavailabletonmiles.

Thesalariesconsistofpaymentstopilots,flightattendantsandticketagents.Theyaredeterminednotonlybythenumberofpassengersandcargoesbutalsothemilesorhoursflown.Infact,milesandhoursarecorrelated.Sowechooserevenuepassengermilesandavailabletonmilesascostdrivers.Revenuepassengermilesisamajorindicatorintheairlineindustry,soit’sreasonabletobeadriver.Availabletonmiles,however,seemsnotsogood.Butaftercalculationwefindthat

oftheformeris0.1764,and

ofthelatteris0.5577.Formoreobviouscomparison,wedrawthefollowingscatterplots.

Thescatterplotbetweenrevenuepassengermilesandsalary

Thescatterplotbetweenavailabletonmilesandsalary

Wecanseethatthelatterscatterplotshowsamorelinearrelationshipbetweenthetwovariables.Fromtheperspectiveofeithernumericalanalysisorvisualjudgment,availabletonmilesismoreaccurateforestimation.Sowechooseavailabletonmilesasthecostdriver.

Lowpoint（3132,1145）,highpoint（4029,1514）

Salary=0.4114×availabletonmiles-143.50

Thistechniquehasadvantages:

onlytwodataisneeded,soit’squiteconvenient.It’seasytoapplyandillustratemathematicallyhowachangeinacostdrivercanchangetotalcost.

However,notalltheinformationisused,whichisregardedasinefficient.Becauseitbasescostfunctionononlytwoperiods’costexperience,regardlessofhowmanyrelevantdatapointshavebeenallocated.Inaword,it’slessaccurate.

2.SimpleRegression

Weusesimpleregressiontoestimatethesalarycostwithavailabletonmilesasthecostdriver.Resultsareasfollows:

Coefficients

Standarddeviation

Intercept

-682.643

282.6033

XVariable1

0.551693

0.079698

Salary=0.5517×availabletonmiles-682.64

=0.5577,andstandarddeviationsaremuchsmallerthancoefficients,soit’sstatisticallyvalidandsignificant.

Regressionanalysismeasurescostbehaviormorereliablythanothercostmeasurementmethods,sincethistechniqueusesstatisticstofitacostfunctiontoallthehistoricaldata.Comparedwiththehigh-lowmethodwhichonlycontainstwogroupsofdata,it’sanimprovement.What’smore,regressionanalysisyieldsimportantstatisticalinformationaboutthereliabilityofthecostestimates,whichallowsanalysttoassessconfidenceinthecostmeasuresandselectbestcostdriver.Butweshouldnoticethatonlyonecostdriverisconsidered,soitcan’texplainthevariationofsalariescompletely.

3.MultipleRegression

BecauseofthereasonsmentionedinQuestion1,wechooserevenuepassengermilesandavailabletonmilesascostdriversandusemultipleregressiontoestimatethesalarycost.Resultsareasfollows:

Coefficients

Standarddeviation

Intercept

-1144.55

243.2101

XVariable1

1.051937

0.120829

XVariable2

-72.2955

14.88974

Salary=-1144.55+1.05×availabletonmiles-72.30×revenuepassengermiles

=0.5577,andstandarddeviationsaremuchsmallerthancoefficients,soit’sstatisticallyvalidandsignificant.

Thistechniquetakesmorecostdriversintoconsideration,andtheresultscalculatedaremoreclosetothedatagiven,soit’sanimprovementoverthemodelestimatedinQuestion2.Theaccuracyofestimationthusservesasitstopadvantage.However,itmightbemorecostly,time-consumingandcomplicatedtobeimplementedthanothertechniques.

4.TheusefulnessofanalysisinQuestion1-3

ThecostfunctionsestimatedinQuestion1-3arebasedontheassumptionthatthewagesperhourremainthesameandthereisnoadditionallaborneeded,soit’susefulonlyundercertainconditions.Takingthebackgroundoftheindustryandthecompany’scircumstanceintoconsideration,wethinktheseareimportant:

ØThefirstoneisthatthepresentequilibriumbetweenDeltaandthelaborunionsisnotinterfered.Thatis,endeavorsconcerningloweringpilots’salariestoindustrylevelwillnotbeobstructedbyunionforces,andemployeesotherthanpilotswillnotjoinlaborunionstorequirehigherpayments.Ifnot,adjustmentsoncoststructurewillbefutileandSongwillonlyturnouttobeanotherDeltaExpress.

ØSecondly,noharshregulationsregardingreducingstaffsorcuttingsalariesaretobeformulated.Restrictionsaboutlayoffswilldirectlyleadtoweakcontroloverbudgets,andinturncreatessimilarproblemsashighsalariesdo.Nevertheless,regulationsofthiskindarealmostinevitable.Accordingtoprecedents,largescalefurloughshavealreadybeenblockedoncebyALPA,thuswehavenoreasontoremainoptimisticaslongastherecessionofthegeneraleconomicsstayasafact.

ØThelastpointwecomeupwithisthenewfixedcostcausedbynewsecuritydirectivesaftertheSeptember11terroristattacks.However,sincesecuritycostscanbeexpectedamidthewholeindustry,itshallnotbecomeamajorconcernforSong,althoughourpredictionmodelmayoverallshiftupwards.

Iftheconditionsarenotmet,thecostfunctionswillbelessuseful.

5.EstimatethesalarycostforJetBlue

AccordingtoQuestion1,availabletonmilesshouldbeusedtoestimatethesalarycost.However,availabletonmilesof2002Q3iseccentricallylow.Sowedrawascatterplot:

Inthissituation,availabletonmilesandsalariesarenotlinear.Thescatterplotofrevenuepassengermilesandsalariesisasfollows:

Revenuepassengermilesandsalariesarequitelinear.Weusethehigh-lowtechniquetoestimatethesalarycostwithrevenuepassengermilesascostdriver.

Lowpoint（599.4,16000）,highpoint（2016.2,49000）

Salary=23.29×revenuepassengermiles+2038.83

6.EstimatethesalariescostforSonginitsfirstyear

ToestimatethesalariescostofSongisquitedifficultbecausethereisnohistoricaldataforreference.BecauseJetBlueisasuccessfulexampleinthelow-costmarket,weuseitshistoricalsalariescosttopredictSong’ssalary.Forsimplicity,wemakesomeassumptions:

ØSongcanachievethesamerevenuepassengermilesasJetBlueineveryquarter

ØJetBlue’ssalariesarelinearwithtimeseries

Thoughthefirstassumptionisverystrong,thesecondonecanbeeasilyverified.Wenumbereachquarterin2001and2002from1to8,andmakesimpleregressionbetweenthetimeseriesandsalary.

Quarter

Numberoftimeseries

JetBlue’sSalary

2001Q1

1

16000

2001Q2

2

19000

2001Q3

3

22000

2001Q4

4

28000

2002Q1

5

34000

2002Q2

6

38000

2002Q3

7

42000

2002Q4

8

49000

Simpleregressionbetweenthetimeseriesandsalary:

Coefficients

Standarddeviation

Intercept

9571.429

986.5881

XVariable

4761.905

195.3735

JetBlue’ssalary=$4761.91×numberoftimeseries+$9571.43

Thescatterplotofnumberoftimeseriesandsalaryisasfollows:

=0.9900andstandarddeviationsaremuchsmallerthan