Therelationshipbetweenthestarredhotelpriceandthenumberoftouristsincities

Abstract

Basedonthe2010-2016Chinanationaltourismadministration,thekeymonitoringandstatistical50citiesasobjects,thispaperanalyzestherelationshipbetweenthecitystarhotelandthenumberofdomestictouristsbasedonapaneldataregressionmodel.Opennesstotheoutsideworld,pricelevel,tourismtrafficdevelopmentlevel,andurbanpublicservicelevelareselectedascontrolvariables,Ifindthataveragepriceofstar-ratedhotelsissignificantlycorrelatedwiththenumberofdomestictourists.Aftersolvingtheendogeneityproblem,Ifindthatfora10000increaseintouristnumbers,thehotelpricewouldincreaseby25yuanonaverage.Ialsostudyseparatelytheresponsivenessofthepricesofeachstarhotelonthetouristnumbers.Ifindthatthethree-starhotelpriceandfive-starhotelpriceexhibitasignificantandpositiverelationshipwiththenumberofdomestictourists.Incontrast,thepricesofonestar,two-star,andfour-starhotelsarenotsensitivetochangesinthenumberofdomestictourists.

Keywords:Starhotels,domestictourists,hotelprice

1.Introduction

Starhotelsareimportanttouristreceptionfacilities.Withmorethan30yearsofreformandopening-up,China'stourismindustryhasdevelopedrapidly.Moreandmorepeopleparticipateintourismactivities.Thehotelindustry,travelagencyindustryandtransportationindustry,representedbythehotelindustry,havebecomethethreepillarindustriesofthetourismindustry.Atthebeginningofthereformandopeningup,Chinesecitizenstraveled0.2timesayear,andmorethan3timesforthefirsttimein2015and3.7timesin2017.Atthebeginningofthedevelopmentoftourism,hotelsweremainlyusedforforeignaffairsreceptionwithalimitednumber.Withthegrowthoftourismdemand,aseriesoffamilyhotels,budgethotels,andhigh-endhotelshaveemerged.Toregulatethedevelopmentofthetourismhotelindustryandadapttothedevelopmenttrendofinternationalanddomestictourism,thenationaltourismadministrationimplementedthestarratingstandardofthePeople'sRepublicofChinafortouristhotelsin2010.Accordingtothescale,facilities,qualityofservice,andmanagementlevel,thehotelsaredividedfromlowtohighintofivegrades,thelowestforonestar,thehighestforfivestars.Thehigherthestarlevelis,thehigherthelevelofthehotelis.Asanecessityformanybusinesstouristsandinboundovernighttourists,starhotelscanprovideaccommodation,entertainment,fitness,andotherservices.Formostcities,especiallybigcities,thenumberandlevelofstar-ratedhotelsareanimportantwindowtoshowthecityimage,aswellasanimportantreflectionofthelevelofurbandevelopment.Atthesametime,theincomeofstar-ratedhotelsisalsoanimportantpartofurbantourismincome.

Hotelroomrateshaveasignificantimpactonthedevelopmentoftourism.Theguestroomisthecoreproductofthehotelandtheimportantplacethatthehotelprovidestheserviceforthecustomer.Inthetraditionalsense,hotelroompricereferstothepricethatconsumersarewillingtopayforacompletestayexperience.Itisareflectionofthevalueofhotelroomserviceproducts,aswellasthetemporaryuserightandservicepurchasedbycustomers.Asanecessaryexpenditureforovernighttourists,housingsharesashighas30%oftheoverallcostofthetourists’expenses,accordingtotheNationalBureauofStatistics.Alowhotelpricecouldundoubtedlystimulatepeople'sconsumptiondesire,whileahighhotelpricecouldsometimesreflectthehotel’shigh-qualityservices.Hence,itisinterestingtoexaminewhatistheroleofhotelpricesinthetourismindustry.

Regardingthefactthatthehotelpriceisoneofthemostimportantconcernsforatouristtodecidehis/herdestination,thepurposeofthispaperistoexploretherelationshipbetweentheaveragehotelpriceandthenumberofdomestictourists,andthentoexaminethesensitivityofthepricesofdifferentstar-ratedhotelstothenumberofdomestictourists.

Thispaperproceedsasfollows.Section2reviewstherelatedliterature.Section3issampleselection.Section4reportsthedatadescriptionandsummarystatisticsandpresentstheregressionresult.Section5isanextensionofthemainregression.Finally,Section6concludesthepaperwithsomepolicyrecommendationsandresearchlimitations.

2.Literaturereview

Intermsofinfluencingfactorsofthehotelroomprice,scholarsmainlyfocusonfactorssuchasstarratingandbrandchain,publicandprivateattributes,location,hotelbehavior,etc.Starratingisthefactorthathasasignificantimpactonhotelroompricesconfirmedbymostscholarsthroughresearch.

Li(2011)usedtheConsumerPriceIndex(CPI)andtourismgrowthtoexploretherelationshipbetweeninflationandtourismgrowth,andfoundthatthedevelopmentoftourisminHainanIslandhascausedhotelpricestorise.Starhotelroompriceswillbeaffectedbytourismdevelopment.Wu(2012)andotherstudiesonHainanIslandfoundthatthereisalong-termequilibriumrelationshipbetweentourismandinflation,andthedevelopmentoftourismhasapositiveeffectoninflation.Acitywithawell-developedtourismindustryreceivesarelativelylargenumberoftouristsorstaysforalongperiod.Withthesamesupplyofhotels,theinfluxoftouristswillleadtoashortageofsupplyandpromotetheoverallriseinhotelprices.Rigallitorrent(2011)andZhang(2011)findouttheenvironmentoftheareawherethehotelislocatedthroughresearch.Publicattributessuchaseconomyanddistancefromtransportationhubscansignificantlyaffecthotelroomprices.

Israeli(2002)foundthroughthestudyofthehotelindustryinIsraelthatthestarratingsystemandthehotelpriceisconsistentandcloselylinked.However,theinfluenceofthebrandchainonthehotelroompricewillbeaffectedbythehotellocation,theirownstrength,consumerspendingcapacityandotherfactorsoftheimpactofthebusinesstrip.Whenconsumerschooseahotel,theynotonlyconsiderthestarratingandbrandofthehotelbutalsopayattentiontoitspublicandprivateattributes.Becauseconsumersgetutilityfromthehotel'sprivateattributessuchasthepool,sportsfacilities,foodquality,androomservice.Atthesametime,environmentalprotection,publicsecurity,economicconditions,andotherpublicattributesofthehotelwillalsosignificantlyaffectthechoiceofconsumers.Hrane(2007)studiedthedeterminantsofhotelroompricesinOslo,thecapitalofNorway,andfoundthattheavailabilityofsmallrefrigerators,hairdryersandrabbit-feeparkingfacilitiesintheroomshadasignificantimpactontheroomprices.

Saloetal.(2014)foundthroughthestudyofhotelsonthebravacoastthattheexternalnaturalenvironment,publicsafety,distancefromculturalheritage,andotherpublicattributessignificantlyaffecttheroomprice,andthepricedifferencecausedbyitsprivateattributessuchasswimmingpoolandsportsfacilitiesalsodependonthepublicattributesofthehotel.Theauthorconductedtheresearchintwostages.Theauthorsfirstconsideredembeddingthepublicattributeintothevariable"municipalitywhereagivenaccommodationfacilityislocated".Thatis,theauthorsuselocation(municipality)asacompositeindexofpublicattributes.Inthesecondstage,theauthorsunlockthevariableshiddenbehindthelocationeffect(municipality),whichare:publicservices,privatesupplementarysupply,qualityofnaturalresources,culturalheritage,brandimage,infrastructureorspecificevents.Itisimportanttonotethattheauthordidnotusethedatainthedatabasewebsite,butthebrochureprovidedbytheintermediary.Itisbecause,in2004,thesitedidnotincludemosthotelsandsecondhomes.Asaresult,thebrochureoffersabroaderpictureofthehotelandsecondhomerentals.

Yang(2018)analyzedthedifferencesinthebenchmarkrentsofstar-ratedhotelsineachofthe50citiesthatweremonitoredandcountedbytheNationalTourismAdministrationfrom2010to2016.Thefixed-assetinvestmentisselectedtobethecontrolvariable.Theinfluenceofcitysizeandtourismdevelopmentonthestarhotelbenchmarkrentismainlystudied.Apaneldatamodelofstarhotelbenchmarkrentisestablished.Throughresearch,theauthorconcludedthatthelargerthecity,thehigherthebenchmarkhotelrentsare;thehigherthetourismdevelopment,thehigherthebenchmarkhotelrentsare.Furthermore,itisfoundthatthelevelofeconomicdevelopment,theregionwherethecityislocated,thecity'sadministrativelevel,andthefixed-assetinvestmentinstar-ratedhotelsallhaveapositiveimpactonthebenchmarkrentofstar-ratedhotels.

Wecanseefromtherelevantliteratureontheinfluencingfactorsofhotelroompricethatinthepastperiodoftime,foreigncountrieshavepaidmoreattentiontohotelroomprices,butdomesticstudiesonthisaspectarerelativelylacking.Asaserviceindustry,theroompriceofstarhotelswillbeaffectedbymanymacroandmicrofactors.Fromthemacrolevel,hotelroompricesareaffectedbycost,supplyanddemandrelationshipandresidents'consumptionability,sothedegreeofurbanopenness,tourismandtransportationdevelopmentlevel,urbanpublicservicelevelandotherfactorsarethemaininfluencingfactors.Thispaperintendstostudytherelationshipbetweentheaverageroomrateofstar-ratedhotelsandthenumberofdomestictouristsfromamacroeconomicperspective.

3.Data

Thispaperstudiestherelationshipbetweenhotelaveragepriceandthenumberofdomestictouristsfromamacroperspective.Thecontrolvariablesincludeopennesstotheoutsideworld,pricelevel,tourismtrafficdevelopmentlevel,andurbanpublicservicelevel.Theresearchobjectis50keycitiesfromChinaBureaustatistics.Theselectionofthese50citiesismainlybasedonthefollowingconsiderations:(1)mostofthese50citiesarethefirstbatch,secondbatchandthirdbatchofChina'sexcellenttourismcitiesselectedbyChinanationaltourismadministration,andmostofthemareprovincialcapitals,sub-provincialcitiesorothercitieswithgreatinfluenceinChina.(2)Relativelyrepresentative;asthetourismcitiesthatthenationaltourismadministrationfocuseson,thedevelopmentstatusofstar-ratedhotelsinthese50citiescanbasicallyreflecttheoveralldevelopmentstatusofstar-ratedhotelsinChina;(3)Theyareselectedaccordingtothereliabilityandavailabilityofdatasources.Table1showstheregionalandadministrativeleveldistributionofthe50cities

Table1.Regionalandadministrationdivisionsof50majorcities

Region

Theeasternregion

Thecentralregion

Thewesternregion

Municipalitiesdirectlyunder

thecentralgovernment

Beijing,Shanghai,Tianjin

Chongqing

Sub-provincialcity

Shenyang,Dalian,Nanjing,Hangzhou,Ningbo,Xiamen,Jinan,Qingdao,Guangzhou,Shenzhen

Changchun,Ha’erbin,Wuhan

Xi’an,Chengdu

Prefecturecity

Shijiazhuang,Fuzhou,Haikou,Sanya,Qinhuangdao,Wuxi,Suzhou,Wenzhou,Quanzhou,Dongguan,Zhuhai

Taiyuan,Hefei,Nanchang,Zhengzhou,Changsha,Huangshan,Luoyang,Yichang,Zhangjiajie

Guiyang,Kunming,Lasa,Lanzhou,Xining,Yinchuan,Wulumuqi,Nanning,Lijiang,Guilin,Huhehaote,

3.1DescriptiveStatistics

ThemaindatausedinthispaperisfromtheWindfinancialterminal,EPSGlobalStatisticalAnalysis/PredictionPlatform,andChinatourismyearbook.Consideringtheavailabilityandconsistencyofdata,thispapercollectedthepaneldataof50largeandmedium-sizedcitiesfrom2010to2016.

Thispapertakestheaveragehotelpriceasthedependentvariable.Basedonthepreviousresearch(e.g.,Wu,2019),theauthorselectedopennesstotheoutsideworld,pricelevel,tourismtrafficdevelopmentlevel,urbanpublicservicelevelascontrolvariables.

Theexplainedvariableinthisarticleistheaveragepriceofthestarhotels(ave_price).Thevariableiscalculatedbydividingthetotalroomrevenuebythenumberofroomsrentedinahotelacrossallthehotels(fromstar1tostar5)inaprovince.Hence,alltypesofroomsinahotelareincludedinthecalculation.

Theexplanatoryvariableisthenumberofdomestictourists(tour).ThenumberofdomestictouristsreferstoChinesemainlandresidents,foreigners,overseasChinese,HongKong,MacaoandTaiwancompatriotswhohaveresidedinChinaformorethanoneyear.Theyhavestayedatleastonenightintouristfacilitieselsewhereinthecountry,butnotmorethan6months.

Thecontrolvariablesareasfollow:

Opennesstotheoutsideworld(open).Thedegreeofopeningtotheoutsideworldreflectsthedegreeofopeningtotheoutsideworldandexchangesofaregion.Inthispaper,thedegreeofeconomicextroversionismeasuredbydividingtheforeigndirectinvestmentofaregion(intenthousanddollar)bythetotalsocialinvestment,(intenthousandyuan)denotedasopen.Regionswithmoreoutward-lookingeconomieswillattractmoreconferences,exhibitions,foreign-relatedbusinessevents.Itisexpectedthatthisvariablewillhaveapositiveimpactonhotelprices.

Pricelevel(CPI).CPIisanimportantmacroeconomicindicatorthatreflectsthechangesinthepricesofconsumergoodsandservicesrelatedtopeople'slives.Itisalsoanimportantindicatorofmacroeconomicanalysisanddecision-makingaswellasnationaleconomicaccounting.Theconsumerpriceindexmeasurestheaveragechangeinretailpricesovertimeformorethan200differentgoodsandservices.Thesemorethan200goodsandservicesaredividedintoeightmaincategories.Whencalculatingtheconsumerpriceindex,eachcategoryhasaweightthatindicatesitsimportance.Theseweightsaredeterminedbysurveyingthousandsoffamiliesandindividualsabouttheproductsandservicestheybuy.Considerthecloselinkagebetweenthehotelpriceandthepricesofothercommodities,itisexpectedthatthegeneralpricelevelispositivelycorrelatedwiththehotelprice.

Tourismtrafficdevelopmentlevel(road).Inthispaper,Iusehighwaymileage(km)dividedbyregionallandarea(km2)tomeasurethedevelopmentleveloftourismandtransportation,denotedasroad.Convenientandefficienttransportationoffersmorepossibilitiesfortravel.Masstourismhasarrived.Atthesametime,passengers'pursuitoftravelqualityalsoputsforwardhigherrequirementsfortransportationinfrastructureandservices.Incitieswithhighertrafficdevelopmentlevel,thecorrespondinglandpriceandrentarerelativelyhigh,sowepredictthatroadwillhaveapositiveimpactonthepriceofstar-ratedhotels.

Urbanpublicservicelevel(gov).ThispapermeasuresthelevelofurbanpublicservicesbydividingfiscalexpenditurebyGDP.Urbanpublicservicereferstotheurbanpublicsector-orientedcitiesprovidepublicproductsandpublicservices,includingurbaninfrastructureinvestmentandmaintenance,provideandstrengthenemployment,socialsecurityservicestorunandsupporteducation,scienceandtechnology,culture,health,sports,andotherpublicutilities,timelyreleaseasocialinformation,andtotheimprovementoflifequalityforthesocialpublictoparticipateinpublicaffairsprovidesstrongguaranteeandcreaterelevantconditions.Withtherapidandsteadydevelopmentoftheeconomy,China'sinvestmentinpublicculturalserviceshasbeenincreasinginrecentyears,theconstructionofpublicculturalservicesystemhasbeenincreasinglyimproved,andthefunctionsofpublicculturalservicefacilitieshavebecomemorecomplete.Hence,themassesofthelifeoftherequirementsofsupportingfacilitiesisalsorising,thehighertherequirementsofthehotelisalsohigher,thehotel'ssupportingfacilitiesarealsomoreandmoreluxuriouswiththeconsumer'srequirements,andthenthehotel'soperatingcostsbecomehigher.Therefore,wepredictthatgovhasapositiveeffectonthepriceofstarhotels.

AswillbediscussedinSection4.2,Ialsoincludethefollowingasaninstrumentalvariabletotackletheendogeneityproblem:

Greencoverage(greencover)referstotheverticalprojectionareaofallvegetationinthecity,includingpublicgreenspace,residentialgreenspace,roadgreenspace,landscapedgreenspace,andscatteredtrees.

Table2.SummaryStatistics

Variables

Unit

mean

SD

max

min

Numberofdomestictourists

(tour)

10000people

5679.12

4826.82

28115

352.9

Averagehotelprice(ave_price)

yuan

349.33

99.81

751.49

124.39

One-starhotelprice(one_price)

yuan

141.45

69.52

454.55

37.6

Two-starhotelprice(two_price)

yuan

164.49

52.68

456.52

48.13

Three-starhotelprice(three_price)

yuan

228.23

60.61

718.96

75.65

Four-starhotelprice(four_price)

yuan

351.54

73.14

633.68

145.83

Five-starhotelprice(five_price)

yuan

349.09

100.01

747.65

124.48

Tourismtrafficdevelopmentlevel(road)

km/km2

1.12

0.46

2.65

0.13

Urbanpublicservicelevel(gov)

tenthousandyuan

0.18

0.22

2.35

0.06

Pricelevel(CPI)

102.81

1.83

115.9

98.7

Opennesstotheoutsideworld

(open)

tenthousandyuan

0.0084

0.0070

0.0368

2.89327E-05

Greencoverage

(greencover)

%

40.656

4.74

58.11

21.69

Table2showsthesummarystatisticsofthemainvariablesinthearticle.Thevariablename,variabledefinitions,mean,standarddeviation,maximum,andminimumarerepresented.respectively.AsshowninTable1,theaveragenumberoftouristswas5769.12andthestandarddeviationwas4828.82.Asexpected,thehotelstarratingisincreasingwiththeaveragepricelevel.Theaveragepriceofthefive-starhotel(349.09yuan)isabout2.5timesthatoftheone-starhotel.Thelowestroomrateappearsinone-starhotels(about37.6yuan),andthehighestroomrateappearsinfive-starhotels(nearly750yuan).Besides,itcanbeseenfromthetablethatthemaximumandminimumpricesofeachstar-ratedhoteldiffergreatly.Takeone-starasanexample,themaximumandminimumpricesoftheone-starhotelare454.55and37.6yuan,respectively,andthestandarddeviationis69.52.

Also,theaveragevalueofthetrafficdevelopmentlevel(road)is1.12,themaximumvalueis2.65,andtheminimumvalueis0.13.Thisalsoimpliesalargevariationinthetrafficdevelopmentlevelamongthe50cities.Incontrast,thefiscalexpenditureratio(gov)ismuchsimilaramongthecities,withitsaveragevalue0.18,andstandarddeviation0.22.ForCPI,theaveragevalueis102.81,themaximumvalueis115.9,andtheminimumvalueis98.7,withadifferenceof17.2.Theaveragevalueofopen,whichmeasuresthedegreeofopennesstotheoutsideworld,is0.0084.Moreover,theaveragegreencoveragerateis40.656%.Chinahasmadeprogressinitsgreeningcampaigntotacklegrowingenvironmentalproblemssuchasairpollutionanddesertification,accordingtothedata.

4.Model

Thispaperattemptstousepaneldataforanalysis.Themodelisasfollows:

ave_priceit=β0+β1Tourit+β2Xit+εit(1)

Thesubscriptsiandtrepresenttheregionandtheyear,respectively,andtheexplainedvariableave_priceitistheaveragehotelprice.Xitrepresentsthecontrolvariablesincludingopennesstotheoutsideworld,generalpricelevel,tourismtrafficdevelopmentlevel,andurbanpublicservicelevel.εitistheerroritem.β1measurestheeffectofthenumberoftouristsontheaveragehotelprice.Ifthecoefficientissignificantlypositive,theincreaseinthenumberoftouristsispositivelycorrelatedwiththeaveragehotelprice;Onthecontrary,itindicatesthattheincreaseinthenumberoftouristsisnegativelycorrelatedwiththeaveragehotelprice.

Firstofall,IapplytheHausmantestmethodisadoptedtoselecttheestimationmodel.ThepurposeoftheHausmantestistodeterminethecorrelationbetweenindividualeffectsandexplanatoryvariables,soastodeterminethechoiceoffixedeffectmodelorrandom-effectmodel.IntheHausmantest,thenullhypothesisisthatthereisnocorrelationbetweenexplanatoryvariablesandindividualeffectsintherandom-effectsmodel.

4.1EmpiricalResults

TheresultsoftheHausmantestandtheregressionresultareasfollows:

Table3.Regressionresultofpanelmodel

Variables

Ave_price(1)

Ave_price(2)

Ave_price(3)

Ave_price(4)

Ave_price(5)

Tour

0.00613***

(0.0010)

0.0065***

（0.0010）

0.0064***

（0.0011）

0.0046***

（0.0013）

0.0023*

（0.0013）

CPI

2.471***

3.0698***

2.6302***

2.7791*

(1.114)

(1.1534)

(1.1625)

(1.1073)

open

-1656.032***

-1461.28***

-1207.73*

(810.5839)

(989.2454)

(943.3844)

road

88.9229***

84.8470*

(28.9874)

(27.6114)

gov

688.0933*

(147.5293)

Constant

311.03***

57.140***

15.2066***

-35.6584***

-132.1263*

(14.086)

（117.15)

(119.0982)

(123.1431)

(119.0496)

Method

re

re

fe

fe

fe

Observations

323

306

303

253

253

Standarderrorsinparentheses.reandferefertotherandomandfixedeffectmodels,respectively.***p<0.01,**p<0.05,*p<0.1.

ItcanbeseenfromTable3thatwhennocontrolvariableisadded,atthe1%significancelevel,theHausmantestresultp-valueisgreaterthan0.05,indicatingthattherandomeffectmodelcanbeselectedforthemodel.WhenthecontrolvariableCPIwasadded,theHausmantestresultshowedthatthePvalueisalsogreaterthan0.05,indicatingthattherandomeffectmodelwasselected.

Fromcolumn(3),itcanbeseenthatthecoefficientofopenisnegative,indicatingthatopenisnegativelycorrelatedwiththeaveragehotelprice.Afterallcontrolvariablesareadded,itisshownthattheaveragehotelpriceandthenumberoftouristsarestillsignificantlycorrelatedat1%.Moreover,thecoefficientoftouris0.0023,whichmeansthattheaveragepriceofthehotelincreasesby23yuanforevery10,000increaseinthenumberoftourists.Secondly,thecoefficientsofCPI,road,andgovareallgreaterthanzero,indicatingthattheyarepositivelycorrelatedwiththeaveragehotelprice,whichisconsistentwithourhypothesis.However,thecoefficientofopenisnegative,whichindicatesthatthegreaterthedegreeofopeningtotheoutsideworld,thelowertheaveragehotelprice,whichisinconsistentwithourprevioushypothesis.Itisexpectedthatthemoreopenthecountryis,thatis,themoreopenitistotheoutsideworld,themoreforeignerswillenterthecountryfortourism.Inthecaseofconstantsupplyofthehotel,thedemandofforeigntouristswillaffectthesupplyanddemandofthehotel,thusaffectingthelevelofthehotelprice.However,accordingtoourresults,thehighertheopenis,thelowertheaveragehotelpriceis,whichmaybeduetotheincreaseddemandfordomestichotelscausedbytheinboundtourismofforeigners.TouristsfromthesurroundingSoutheastAsiancountriesmightbemoresensitivetohotelprices.Moreover,manyforeigntouristsareguidedintothecountrybypackagetours,whichareoftennegotiatedatmuchlowerratesthanindividualrates.Inaperfectlycompetitivemarketenvironment,eachhotelchainreducesthehotelpriceandfinallyreachesarelativelystableprice,whichislowerthanbefore.

4.2Robustnesstest

Theendogeneityproblemreferstotheproblemthatoneormoreexplanatoryvariablesinthemodelarerelatedtorandomdisturbanceterms.Therearetworeasonsforthis.1.Missingvariablesinthemodelarerelatedtoothervariablesintroducedintothemodel.2.Explanatoryvariablesandexplainedvariablesinteractandinfluenceeachother.Inourpaper,thereisaninteractionbetweenexplanatoryvariablesandexplainedvariables,thatis,notonlydoesthenumberoftouristsaffectthehotelprice,butalsothehotelpriceaffectsthenumberoftouriststoacertainextent.

Inordertoguaranteetherobustnessofthemodelandavoidthepotentialendogeneityproblem,aninstrumentalvariableisintroduced.Wetestwhetherthesignificanceandthesignofthecoefficientschangeaftertheintroductionofinstrumentalvariables.Iselectthegreencoveragerateasaninstrumentvariable.Thegreencoveragerateisanimportantindicatortoreflectthestatusofecologicalandenvironmentalprotectioninacountryorregion.ItisalsoanimportantindicatorfortheassessmentofChina'smodelcitiesforenvironmentalprotectionandtheestablishmentofcivilizedcities.Thevariablecalculatestheproportionsofthetotalgreenareacoveredbytheresidentialareaandthetotalareaofthearea.Greencoveragereferstotheverticalprojectionareaofallvegetationinthecity,includingpublicgreenspace,residentialgreenspace,roadgreenspace,landscapedgreenspace,andscatteredtrees.Theliteraturehasfoundacloselinkagebetweenthegreencoveragerateandtourism.Forexample,Long(2016)constructedthecontributiondegree(GTCD)andcontributionrate(GTCR)modelsofgreeningtourbantourismdevelopment,andmeasuredtheannualaveragecontributionofgreeningtourbantourismdevelopmentinJiangsuprovincefrom2004to2014,withanaveragecontributiondegreeof0.29%andanaveragecontributionrateof13.26%.Onthewhole,thecontributionofJiangsuprovince'sgreenresourcestothedevelopmentofurbantourismisoutstanding.Zhao(2018)selected20citiesinHuaihaieconomiczonefrom2006to2016oftourismandgreenspaceindexofpaneldata,throughthefixedeffectsmodelwithvariablecoefficients,regressionanalysis,toexploretheurbangreeningeffectontourismeconomycondition.ResearchshowsthattherelationshipbetweenthewholeurbangreeningandtourismeconomyissignificantinHuaihaieconomiczone.Hence,Ichoosegreencoveragerateasaninstrumentalvariable,becausegreencoveragerateaffectsthenumberoftouristsaccordingtotheabovepapers,butthereisnoevidencethatitaffectstheaveragehotelprice.

AssumingthataneffectiveinstrumentalvariableGreenitoftheendogenousvariableTouritisfound,OLSregressionofGreenitonTouritcanbeperformedtoisolatetheexogenouspartofXit.

Tourit=γ+δGreenit+γXit+uit (2)

Thisregressioniscalled"firststageregression".Itisrequiredthatδ≠0;otherwise,thisseparationcannotbeachieved.NotethatthefittedvalueofthisregressionisTourit=Tourit+δGreeni,andtheresidualisuit=Tourit?Tourit

Obviously,theregressionin