1、T R A N S P O RT G LO B A L P R AC T I C EThe Container PortPERFORMANCE INDEX 2023A COMPARABLE ASSESSMENT OF PERFORMANCE BASED ON VESSEL TIME IN PORTPublic Disclosure AuthorizedPublic Disclosure AuthorizedPublic Disclosure AuthorizedPublic Disclosure Authorized 2022 International Bank for Reconstruc

2、tion and Development/International Development Association or The World Bank1818 H Street NWWashington,DC 20433Telephone:202-473-1000Internet:www.worldbank.orgThis work is a product of the staff of The World Bank,together with external contributions from S&P Global Market Intelligence.The findings,i

3、nterpretations,and conclusions expressed in this work do not necessarily reflect the views of The World Bank,its Board of Executive Directors,or the governments they represent.The World Bank does not guarantee the accuracy of the data included in this work.The boundaries,colors,denominations,and oth

4、er information shown on any map in this work do not imply any judgment on the part of The World Bank concerning the legal status of any territory or the endorsement or acceptance of such boundaries.Nothing herein shall constitute or be considered to be a limitation upon or waiver of the privileges a

5、nd immunities of The World Bank,all of which are specifically reserved.Rights and PermissionsThis work is available under the Creative Commons Attribution 3.0 IGO license(CC BY 3.0 IGO)http:/creativecommons.org/licenses/by/3.0/igo.+Under the Creative Commons Attribution license,you are free to copy,

6、distribute,transmit,and adapt this work,including for commercial purposes,under the following conditions:AttributionPlease cite the work as follows:The World Bank,2024.“The Container Port Performance Index 2023:A Comparable Assessment of Performance based on Vessel Time in Port(Fine).”World Bank,Was

7、hington,DC.License:Creative Commons Attribution CC BY 3.0 IGO.TranslationsIf you create a translation of this work,please add the following disclaimer along with the attribution:This translation was not created by The World Bank and should not be considered an official World Bank translation.The Wor

8、ld Bank shall not be liable for any content or error in this translation.AdaptationsIf you create an adaptation of this work,please add the following disclaimer along with the attribution:This is an adaptation of an original work by The World Bank.Views and opinions expressed in the adaptation are t

9、he sole responsibility of the author or authors of the adaptation and are not endorsed by The World Bank.Third-party contentThe World Bank does not necessarily own each component of the content contained within the work.The World Bank,therefore,does not warrant that the use of any third party-owned

10、individual component or part contained in the work will not infringe on the rights of those third parties.The risk of claims resulting from such infringement rests solely with you.If you wish to reuse a component of the work,it is your responsibility to determine whether permission is needed for tha

11、t reuse and to obtain permission from the copyright owner.Examples of components can include,but are not limited to,tables,figures,or images.All queries on rights and licenses should be addressed to World Bank Publications,The World Bank Group,1818 H Street NW,Washington,DC 20433,USA;fax:202-522-262

12、5;e-mail:pubrightsworldbank.org.TRANSPORT GLOBAL PRACTICEThe Container Port Performance Index 2023 A Comparable Assessment of Performance based on Vessel Time in Porti|TABLE OF CONTENTSTable of contentsAcknowledgements.iiiAbbreviations and Acronyms.ivGlossary.vForeword.viExecutive summary.11.Introdu

13、ction.92.The Port Performance Program.13Introduction.13The Port Performance Program.14The Automatic Identification System and Port Zoning.14The Anatomy of a Port Call.15Overall Port Time Distribution.17The Significance of Call Size.203.The Approach and Methodology.25The Structure of the Data.25Const

14、ructing the Index:The Administrative Approach.29Why Is Matrix Factorization Useful?.34The Statistical Methodology.35Borda-Type Approach for Index Aggregation.364.The Container Port Performance Index 2023.38Introduction.38The CPPI 2023.38Ranking by Region.40Ranking by Throughput.485.Conclusions and N

15、extSteps.55Appendix A:The CPPI 2023.56Table of conTenTs|ii TablesTable E.1 The CPPI 2023:Global Ranking of Container Ports.2Table 2.1 Average Arrival Time Development per Region and Ship Size,20222023.19Table 2.2 Average Arrival Time Performance per Ship Size Range per Region.20Table 3.1 Port Calls

16、Distribution.27Table 3.2 Ship Size Group Definitions.27Table 3.3 Call Size Sensitivity.28Table 3.4 Quantity of Ports Included per Ship Size Group.29Table 3.5 An Example of Imputing Missing Values.30Table 3.6 Port Hours Performance Appraisal.31Table 3.7 Assumptions to Determine a Fuel Consumption Ind

17、ex.32Table 3.8 Sample Port Productivity Data Structure by Ship Size.34Table 3.9 Sample Illustration of Latent Factors.34Table 3.10 An Example of Aggregated Rankings for Four Ports with Randomly Generated Administrative and Statistical Index Values.36Table 4.1 The CPPI 2023.39Table 4.2 The CPPI by Re

18、gion:North America.41Table 4.3 The CPPI by Region:Central America,South America,and the Caribbean Region.41Table 4.4 The CPPI by Region:West,Central,and South Asia(Saudi Arabia to Bangladesh).43Table 4.5 The CPPI by Region:East Asia(Myanmar to Japan).44Table 4.6 The CPPI by Region:Oceania(Australia,

19、New Zealand,and the Pacific Islands).45Table 4.7 The CPPI by Region:Sub-Saharan Africa.45Table 4.8 The CPPI by Region:Europe and North Africa.46Table 4.9 The CPPI by Throughput:Large Ports(More than 4 million TEUs per Year).49Table 4.10 The CPPI by Throughput:Medium Ports(between 0.5 million and 4 m

20、illion TEUs per Year).49Table 4.11 The CPPI by Throughput:Small Ports(Less than 0.5 million TEUs per Year).52Table A.1 Aggregated Rankings Using Borda-type Approach.56Table A.2 The CPPI 2023(the Administrative Approach).61Table A.3 The CPPI 2023(the Statistical Approach).72FiguresFigure 2.1 The Anat

21、omy of a Port Call.16Figure 2.2 In-Port Time Consumption.17Figure 2.3 Global Average Arrival Time Development 2022-2023.18Figure 2.4 The Aggregated Correlation between Ship and Call Size.21Figure 2.5 Container Moves Performed per gross Crane Hour across Various Ship Sizes.22Figure 2.6 Gross Crane Pr

22、oductivity by Call Size.22Figure 2.7 Crane Productivity by Crane Intensity.23Figure 2.8 Call Size versus Crane Intensity.23Figure 2.9 Average Moves per Crane.24Figure 3.1 The Structure of the CPPI.26Figure 3.2 Percentage of Port Calls per Ship Size Group-2023.28iii|ACkNOWLEDGEMENTSAcknowledgementsTh

23、is technical report was prepared jointly by the teams from the Transport Global Practice of the Infrastructure Vice-Presidency at the World Bank and the Maritime,Trade and Supply Chain division of S&P Global Market Intelligence.The World Bank team was led by Richard Martin Humphreys(Global Lead for

24、Connectivity and Logistics and Lead Transport Economist,ITRGk),Dominique Guillot(Associate Professor,University of Delaware),under the guidance of Binyam Reja(Global Practice Manager Transport,ITRGk)and Nicolas Peltier-Thiberge(Global Practice Director Transport,ITRGk).The S&P Global Market Intellig

25、ence team was led by Turloch Mooney(Global Head of Port Intelligence&Analytics,GIA),underthe guidance ofGuy Sear(Head of Global Risk&Maritime,GIA)and Jenny Paurys(Head ofGlobal Intelligence&Analytics).The joint team would like to extend special thanks to the following experts for their comments on t

26、he draft of the technical report:Gylfi Palsson(Lead Transport Specialist,ILTC1),Ninan Biju Oommen(Senior Transport Specialist,IEAT1),and Yin Yin Lam(Senior Transport Specialist,IEAT1).abbreviaTions and acronyms|iv abbreviations and acronymsAcronymsDescriptionAISAutomatic Identification SystemCICrane

27、 IntensityCOVID-19Coronavirus Disease 2019CPPIContainer Port Performance IndexEEZExclusive Economic ZoneFAFactor AnalysisGCIGlobal Competitiveness IndexGCMPHMoves per Gross Crane HourGDPGross Domestic ProductGRTGross Registered TonnageITUInternational Telecommunication UnionLLDCLandlocked Developing

28、 CountryLPILogistics Performance IndexSIDSSmall Island Developing StatesTEUTwenty-foot Equivalent UnitUNCTADUnited Nations Conference on Trade and Developmentv|GLOSSARYGlossaryAll fast:The point when the vessel is fully secured at berth and all mooring lines are fastArrival time/hours:The total elap

29、sed time between the vessels automatic identification system(AIS)recorded arrival at the actual port limit or anchorage(whichever recorded time is the earlier)and its all lines fast at the berthBerth hours:The time between all lines fast and all lines releasedBerth idle:The time spent on berth witho

30、ut ongoing cargo operations.The accumulated time between all fast to first move plus last move to all lines releasedCall size:The number of container moves per call,inclusive of discharge,load,and restowageCargo operations:When cargo is being exchanged,the time between first and last container moves

31、Crane intensity(CI):The quantity of cranes deployed to a ships berth call.Calculated as total accumulated gross crane hours divided by operating(first to last move)hoursFactor analysis(FA):A statistical method used to describe variability among observed,correlated variables in terms of a potentially

32、 lower number of unobserved variables called factorsFinish:Total elapsed time between last container move and all lines releasedGross crane hours:Aggregated total working time for all cranes deployed to a vessel call without any deductions.Time includes breakdowns,inclement weather,vessel inspired d

33、elays,un/lashing,gantry,boom down/up plus hatch cover and gear-box handlingGross crane productivity(GCMPH):Call size or total moves divided by total gross crane hours.Hub port:A port which is called at by deep-sea mainline container ships and serves as a transshipment point for smaller outlying,or f

34、eeder,ports within its geographical region.Typically,more than 35 percent of its total throughput would be hub and spoke or relay transshipment container activityMoves:Total container moves.Discharge+restowage moves+load.Excluding hatch covers,gearboxes,and other non-container related crane work.Bre

35、akbulk cargo lifts are excluded,however empty platform(tweendeck or flat-rack)handling moves are included.Moves per crane:Total Moves for a call divided by the crane intensityPort call:A call to a container port/terminal by a container vessel where at least one container was discharged or loadedPort

36、 hours:The number of hours a ship spends at/in port,from arrival at the port limits to sailing from the berthPort limits:Either an anchorage zone or the location where pilot embarkation or disembarkation occurs and recorded as whichever activity is the earliestPort to berth hours:The time from when

37、a ship first arrived at the port limits or anchorage zone(whichever activity occurs first)until it is all fast alongside the berth.Relay transshipment:Containers transshipped between ocean going container shipsShip size:Nominal capacity in twenty-foot equivalent units(“TEUs”)Start:The time elapsed f

38、rom berthing(all lines fast)to first container moveSteam in time:The time required to steam-in from the port limits and until all fast alongside the berthTwenty-foot equivalent unit or TEU:A standard metric for container throughput,and the physical capacity of a container terminal.A 20-foot containe

39、r is equal to 1 TEU,and a 40-foot or 45-foot container is equal to 2 TEUs.Regardless of container size(10 feet,15 feet,20 feet,30 feet,40 feet,or 45 feet),each is recorded as one move when being loaded or discharged from the vessel.Vessel capacity:Nominal capacity in twenty-foot equivalent Units(“TE

40、Us”)Waiting time:Total elapsed time from when vessel enters anchorage zone to when vessel departs anchorage zone(vessel speed must have dropped below 0.5 knots for at least 15 mins within the zone)foreword|vi forewordThe challenges caused by the COVID-19 pandemic and its aftermath on the sector ease

41、d further in 2023.Continuing or new disruptions in the form of Russias invasion of Ukraine,the attacks on shipping in the Gulf of Aden,and draught restrictions on the Panama Canal,all impacted container shipping.In addition,the glut of new capacity ordered by lines during the pandemic and falling de

42、mand meant that freight rates have fallen,after an initial slump,to pre-pandemic norms on most routes.These changes impact performance and the ranking of ports.While some problems are exogenous or systemic,some are endogenous or location specific,with the result that both impact the performance and

43、ranking of individual ports.One of the silver linings of the pandemic was greater awareness and focus on the resilience and efficiency of the maritime gateways,where any friction will result in tangible impacts on consumer choice,price,and ultimately economic development.That focus is even more impo

44、rtant now.Traditionally,one of the major challenges to stimulating improvement in the efficiency of ports has historically been the lack of a reliable,consistent,and comparable basis on which to compare operational performance across different ports.While modern ports collect data for performance pu

45、rposes,the quality,consistency,and availability of data,the definitions employed,and the capacity and willingness of the organizations to collect and transmit data to a collating body have all precluded the development of a robust comparable measure(s)to assess performance across ports and time.The

46、introduction of new technologies,increased digitalization,and the willingness on the part of industry stakeholders to work collectively toward systemwide improvements have now provided the opportunity to measure and compare container port performance in a robust and reliable manner.A partnership has

47、 resulted in this technical report,which is the fourth iteration of the Container Port Performance Index(CPPI),produced by the Transport Global Practice of the World Bank in collaboration with the Global Intelligence&Analytics division of S&P Global Market Intelligence.The CPPI is intended,as in its

48、 earlier iterations,to serve as a reference point for improvement for key stakeholders in the global economy,including national governments,port authorities and operators,development agencies,supranational organizations,various maritime interests,and other public and private stakeholders in trade,lo

49、gistics,and supply chain services.The performance of a port may be assessed based on a myriad of measurements,such as:terminal capacity or space utilization,cost,landside connectivity&services,or ship to shore interchange.The CPPI is based on available empirical objective data pertaining exclusively

50、 to time expended in a vessel stay in a port and should be interpreted as an indicative measure of container port performance,but not a definitive one.Nicolas Peltier-Thiberge Global Practice Director Transport The World BankJenny Paurys Head of Global Intelligence&Analytics S&P Global Market Intell

51、igence1|ExECUTIVE SUMMARYexecutive summaryMaritime transport forms the foundation of global trade and the manufacturing supply chain.The maritime industry provides the most cost-effective,energy-efficient,and dependable mode of transportation for long distances.More than 80 percent of global merchan

52、dise trade(by volume)is transported via sea routes.A considerable and increasing proportion of this volume,accounting for about 35 percent of total volumes and over 60 percent of commercial value,is carried in containers.The emergence of containerization brought about significant changes in how and

53、where goods are manufactured and processed,a trend that is likely to continue with digitalization.Container ports are critical nodes in global supply chains and essential to the growth strategies of many emerging economies.In numerous cases,the development of high-quality container port infrastructu

54、re operating efficiently has been a prerequisite for successful export-led growth strategies.Countries that follow such a strategy will have higher levels of economic growth than those that do not.Efficient,high quality port infrastructure can facilitate investment in production and distribution sys

55、tems,engender expansion of manufacturing and logistics,create employment opportunities,and raise income levels.However,ports and terminals,especially container terminals,can cause shipment delays,disruptions in supply chain,additional expenses,and reduced competitiveness.The negative effect of poor

56、performance in a port can extend beyond the that ports hinterland to others as container shipping services follow a fixed schedule with specific berth windows at each port of call on the route.Therefore,poor performance at one port could disrupt the entire schedule.This,in turn,increases the cost of

57、 imports and exports,reduces the competitiveness of the country and its hinterland,and hinders economic growth and poverty reduction.The consequences are particularly significant for landlocked developing countries(LLDCs)and small island developing states(SIDS).Comparing operational performance acro

58、ss ports has been a major challenge for improving global value chains due to the lack of a reliable,consistent,and comparable basis.Despite the data collected by modern ports for performance purposes,the quality,consistency,and availability of data,as well as the definitions used and the capacity an

59、d willingness of organizations to transmit data to a collating body,have hindered the development of a comparable measure(s)for assessing performance across ports and time.However,new technologies,increased digitalization,and industry interests willingness to work collectively toward systemwide impr

60、ovements now provide an opportunity to measure and compare container port performance in a robust and reliable manner.The World Banks Transport Global Practice and the Global Intelligence&Analytics division of S&P Global Market Intelligence have collaborated to produce the fourth edition of the Cont

61、ainer Port Performance Index(CPPI),presented in this technical paper.The aim of the CPPI is to pinpoint areas for enhancement that can ultimately benefit all parties involved,ranging from shipping lines to national governments and consumers.It is designed to act as a point of reference for important

62、 stakeholders in the global economy,including port authorities and operators,national governments,supranational organizations,development agencies,various maritime interests,and other public and private stakeholders in trade,logistics,and supply chain services.The development of the CPPI rests on to

63、tal container ship in port time in the manner explained in subsequent sections of the report,and as in earlier iterations of the CPPI.This fourth iteration utilizes data for the full calendar year of 2023.It continues the change introduced last year of only including execuTive summary|2 ports that h

64、ad a minimum of 24 valid port calls within the 12-month period of the study.The number of ports included in the CPPI 2023 is 405.As in earlier iterations of the CPPI,the production of the ranking employs two different methodological approaches,an administrative,or technical,approach,a pragmatic meth

65、odology reflecting expert knowledge and judgment;and a statistical approach,using factor analysis(FA),or more accurately matrix factorization.The rationale for using two approaches was to try and ensure that the ranking of container port performance reflects as closely as possible actual port perfor

66、mance,whilst also being statistically robust.As there had been a marked improvement in consistency between the rankings resulting from the two approaches since the inaugural CPPI 2020,for CPPI 2023,the same two methodological approaches were used.In addition,the rank aggregation method is employed a

67、gain to combine the results and return one aggregate ranking.The construction of the statistical and administrative approaches,the aggregation methodology and the resulting ranking is detailed in the report,while the respective rankings of the former are detailed in Appendix A.Table E.1 presents the

68、 resulting CPPI 2023.The top-ranked container ports in the CPPI 2023 are Yangshan Port(China)in first place,followed by the Port of Salalah(Oman)in second place,retaining their ranking from the CPPI 2022.Third place in the CPPI 2023 is occupied by the port of Cartagena,up from 5th place in the CPPI

69、2022,whilst Tangier-Mediterranean retains its 4th place ranking.Tanjung Pelepas improved one position to 5th,Ningbo moved up from 12th in 2022 to 7th in 2023,and Port Said moved from 16th to 10th in 2023.Ports moving in the other direction in the top ten:khalifa port falls from 3rd position in 2022

70、to 29th position in CPPI 2023.Hamad Port which fell from 8th in 2022 to 11th in 2023.TABLE E.1 The CPPI 2023:Global Ranking of Container PortsPort NameOverall RankingYANGSHAN1SALALAH2CARTAGENA(COLOMBIA)3TANGER-MEDITERRANEAN4TANJUNG PELEPAS5CHIWAN6CAI MEP7GUANGZHOU8YOkOHAMA9ALGECIRAS10HAMAD PORT11NIN

71、GBO12MAWAN13DALIAN14HONG kONG15PORT SAID16SINGAPORE17kAOHSIUNG18Port NameOverall RankingVISAkHAPATNAM19YEOSU20TIANJIN21YANTIAN22TANJUNG PRIOk23LIANYUNGANG24SHEkOU25CALLAO26MUNDRA27PORT kLANG28kHALIFA PORT29kING ABDULLAH PORT30 xIAMEN31BUSAN32GEMLIk33BARCELONA34DAMMAM35SAVONA-VADO363|ExECUTIVE SUMMAR

72、YPort NameOverall RankingPOSORJA37FUZHOU38ZEEBRUGGE39COLOMBO40PIPAVAV41RIO DE JANEIRO42kHALIFA BIN SALMAN43BUENAVENTURA44LAEM CHABANG45SHIMIZU46kAMARAJAR47INCHEON48JEBEL ALI49LAZARO CARDENAS50AARHUS51DA CHAN BAY TERMINAL ONE52CHARLESTON53TOkYO54PHILADELPHIA55NAGOYA56kATTUPALLI57JEDDAH58JUBAIL59QINZH

73、OU60kARACHI61kEELUNG62COCHIN63kOBE64PORT EVERGLADES65SOHAR66SALVADOR67HAZIRA68LONDON69HAIPHONG70kRISHNAPATNAM71WILHELMSHAVEN72BEIRUT73MIAMI74BOSTON(USA)75ANTWERP76DILISkELESI77ITAPOA78Port NameOverall RankingPUERTO LIMON79CHENNAI80WILMINGTON(USA-N CAROLINA)81MARSAxLOkk82ZHOUSHAN83SOUTHAMPTON84OSAkA8

74、5HAIFA86AQABA87BREMERHAVEN88SANTA CRUZ DE TENERIFE89MALAGA90ROTTERDAM91NEW YORk&NEW JERSEY92JOHOR93POINTE-A-PITRE94YOkkAICHI95JAWAHARLAL NEHRU PORT96CORONEL97TRIPOLI(LEBANON)98JACkSONVILLE99ALTAMIRA100TANJUNG PERAk101COLON102PARANAGUA103PIRAEUS104OSLO105BERBERA106RIO GRANDE(BRAZIL)107HALIFAx108TALLI

75、NN109SAN ANTONIO110CAT LAI111WELLINGTON112SHANTOU113FORT-DE-FRANCE114DANANG115SHANGHAI116HAkATA117IZMIR118QINGDAO119SIAM SEAPORT120execuTive summary|4 Port NameOverall RankingHAMBURG121SOkHNA122SHARJAH123VERACRUZ124PUERTO BARRIOS125TAICHUNG126MOJI127VIGO128YARIMCA129NAHA130PORT AkDENIZ131SAIGON132BA

76、TANGAS133LISBON134SINES135LAS PALMAS136SAN JUAN137CHU LAI138kLAIPEDA139OMAEZAkI140SANTA MARTA141VALENCIA142CEBU143BORUSAN144SUAPE145MUHAMMAD BIN QASIM146RIO HAINA147QUANZHOU148CORk149TANJUNG EMAS150VALPARAISO151CAGAYAN DE ORO152BARRANQUILLA153MUUGA HARBOUR154CHIBA155FREDERICIA156LIMASSOL157AL DUQM15

77、8HIBIkINADA159LIRQUEN160SHUAIBA161BURGAS162Port NameOverall RankingHELSINGBORG163PUERTO BOLIVAR(ECUADOR)164SAGUNTO165MOGADISCIO166NEW ORLEANS167kOMPONG SOM168BAR169SANTO TOMAS DE CASTILLA170DUNkIRk171ALExANDRIA(EGYPT)172MOBILE173TARRAGONA174PUERTO PROGRESO175PAPEETE176NORRkOPING177PUERTO CORTES178PE

78、CEM179BASSETERRE180GUSTAVIA181FELIxSTOWE182GIOIA TAURO183PYEONG TAEk184ARRECIFE DE LANZAROTE185PANJANG186GENERAL SAN MARTIN187QUY NHON188BALTIMORE(USA)189RAUMA190RAVENNA191HUELVA192CAUCEDO193MUARA194LA GUAIRA195LATAkIA196CONAkRY197COPENHAGEN198SHIBUSHI199CIVITAVECCHIA200BELL BAY201LARVIk202BRIDGETOW

79、N203GIJON2045|ExECUTIVE SUMMARYPort NameOverall RankingPOINT LISAS PORTS205PLOCE206TARTOUS207SHUWAIkH208CADIZ209TEESPORT210FERROL211PHILIPSBURG212CASTELLON213HELSINkI214BREST215kRISTIANSAND216BORDEAUx217SALERNO218PORT TAMPA BAY219PORT AU PRINCE220CASTRIES221OITA222HERAkLION223HONOLULU224VOLOS225FREE

80、TOWN226SUBIC BAY227SONGkHLA228PUERTO QUETZAL229BILBAO230PARAMARIBO231NGHI SON232RADES233APRA HARBOR234NEW MANGALORE235CRISTOBAL236ADEN237ALICANTE238BIG CREEk239VARNA240PALERMO241SYAMA PRASAD MOOkERJEE PORT242PAITA243MALABO244ANCONA245SEVILLE246Port NameOverall RankingMARIEL247TRABZON248GOTHENBURG249

81、YANGON250GAVLE251GRANGEMOUTH252NASSAU253GHAZAOUET254BARI255MANAUS256kOTkA257NOVOROSSIYSk258CALDERA(COSTA RICA)259BLUFF260SAINT JOHN261NANTES-ST NAZAIRE262BATUMI263TIMARU264ZARATE265PORT OF SPAIN266GENERAL SANTOS267NELSON268BUENOS AIRES269VENICE270BATA271GDYNIA272BANGkOk273TAkORADI274kUANTAN275AMBARL

82、I276RIGA277HUENEME278DAVAO279NEMRUT BAY280kOTA kINABALU281UMM QASR282SEPETIBA283SAMSUN284NOUMEA285ENSENADA286VILA DO CONDE287AGADIR288execuTive summary|6 Port NameOverall RankingPORT MORESBY289LEIxOES290kUCHING291OTAGO HARBOUR292VLISSINGEN293SANTOS294PUERTO CABELLO295LIVERPOOL(UNITED kINGDOM)296CATA

83、NIA297GEORGETOWN(GUYANA)298PENANG299TOAMASINA300PORT OF VIRGINIA301DUBLIN302NAMIBE303PORT VICTORIA304ONNE305LIVORNO306MAYOTTE307BELAWAN308LAGOS(NIGERIA)309MANILA310MELBOURNE311HOUSTON312SAN VICENTE313BALBOA314GUAYAQUIL315ARICA316kHOMS317LOME318GENOA319PORT REUNION320SAN PEDRO(COTE DIVOIRE)321MAZATLA

84、N322TURBO323PORT BOTANY324MAPUTO325LAE326THESSALONIkI327MOMBASA328LA SPEZIA329CORINTO330Port NameOverall RankingMANZANILLO(MExICO)331CASABLANCA332MEJILLONES333CHATTOGRAM334VITORIA335NAPIER336BRISBANE337GREENOCk338NAPLES339BEIRA340EL DEkHEILA341DURRES342GDANSk343MONROVIA344ADELAIDE345ALGIERS346TAURAN

85、GA347MONTREAL348POTI349AUCkLAND350SETUBAL351IQUIQUE352ABIDJAN353MARSEILLE354CONSTANTZA355VANCOUVER(CANADA)356OWENDO357NOUAkCHOTT358FREEPORT(BAHAMAS)359SEATTLE360BENGHAZI361kOPER362NACALA363TIN CAN ISLAND364BRISTOL365kRIBI DEEP SEA PORT366DAR ES SALAAM367QASR AHMED368PORT LOUIS369DOUALA370BINTULU371L

86、E HAVRE3727|ExECUTIVE SUMMARYPort NameOverall RankingLONG BEACH373FREMANTLE374LOS ANGELES375TEMA376IMBITUBA377kINGSTON(JAMAICA)378DJIBOUTI379WALVIS BAY380DAkAR381BEJAIA382ACAJUTLA383MONTEVIDEO384LYTTELTON385MATADI386DAMIETTA387PORT SUDAN388LUANDA389Port NameOverall RankingASHDOD390PORT ELIZABETH391I

87、SkENDERUN392ITAJAI393POINTE-NOIRE394SAVANNAH395TRIESTE396OAkLAND397DURBAN398PRINCE RUPERT399RIJEkA400TACOMA401COTONOU402MERSIN403NGQURA404CAPE TOWN405Source:Original table produced for this publication,based on CPPI 2021 data.There are 55 new entrants to the CPPI 2023,and several significant movers

88、since the CPPI 2022.One hundred ports improved their ranking in CPPI 2023 compared to CPPI 2022,with some of the largest movers improving their ranking by more than 200 places.execuTive summary|8 9|inTroducTion1.introductionSince the start of maritime trade,ports have played a central role in the ec

89、onomic and social development of countries.The innovation of containerization by Malcom McLean in 1958 changed the course of the shipping industry and engendered significant changes to where and how goods are manufactured.Container ports remain vital nodes in global supply chains and are crucial to

90、the growth strategies of many emerging economies.The development of high-quality port infrastructure,operated efficiently,has often been a prerequisite for successful growth strategies,particularly those driven by exports.When done correctly,it can attract investment in production and distribution s

91、ystems and eventually,support the growth of manufacturing and logistics,create employment,and increase income levels.In contrast,a poorly functioning or inefficient port can hinder trade growth,with a profound impact on LLDCs and SIDS.The port,along with the access infrastructure(inland waterways,ra

92、ilways,roads)to the hinterland,is a vital link to the global marketplace and needs to operate efficiently.Efficient performance encompasses several factors,such as the ports efficiency itself,the availability of sufficient draught,quay,and dock facilities,the quality of road and rail connections,the

93、 competitiveness of these services,and the effectiveness of the procedures utilized by public agencies for container clearance.Any inefficiencies or non-tariff barriers among these actors will result in higher costs,reduced competitiveness,and lower trade volumes(kathuria 2018).More specifically,the

94、 efficiency of port infrastructure has been identified as a key contributor to the overall port competitiveness and international trade costs.Micco et al.(2003)identified a link between port efficiency and the cost of international trade.Clark,Dollar,and Micco(2004)found a reduction inTroducTion|10

95、in country inefficiency,specifically transport cost,from the 25th to 75th percentile,resulting in an increase in bilateral trade of around 25 percent.Wilmsmeier,Hoffmann,and Sanchez(2006)confirmed the impact of port performance on international trade costs,finding that doubling port efficiency in a

96、pair of ports had the same impact on trade costs as halving the physical distance between the ports.Hoffmann,Saeed,and Sdal(2020)analyzed the short-and long-term impacts of liner shipping bilateral connectivity on South Africas trade flows,and showed that gross domestic product(GDP),the number of co

97、mmon direct connections,and the level of competition have a positive and significant effect on trade flows.However,ports and terminals,particularly for containers,can often be the main sources of shipment delays,supply chain disruptions,additional costs,and reduced competitiveness.Poorly performing

98、ports are characterized by limited spatial and operating efficiency,maritime and landside access,oversight,and coordination among the public agencies involved,which lower predictability and reliability.The result is that instead of facilitating trade,the port increases the cost of imports and export

99、s,reduces competitiveness,and inhibits economic growth and poverty reduction.The effect on national and regional economies can be severe see inter alia World Bank(2013)and has driven numerous efforts to improve performance to strengthen competitiveness.Port performance is also a key consideration fo

100、r container shipping lines that operate liner services on fixed schedules,based on agreed pro-forma berth windows.Delays at any of the scheduled ports of call on the route served by the vessel would have to be made good before the vessel arrives at the next port of call,to avoid an adverse impact on

101、 the efficient operations of the service.As such,port efficiency and port turnaround time at all the ports of call are important subjects for operators,and monitoring port performance has become an increasingly important undertaking in the competitive landscape.One of the major challenges to improvi

102、ng efficiency has been the lack of reliable measures to compare operational performance across different ports.The old management idiom,you cannot manage what you cannot measure,is reflective of the historical challenge of both managing and overseeing the sector.While modern ports collect data for p

103、erformance purposes,it is difficult to benchmark the outcomes against leading ports or ports with similar profiles due to the lack of comparative data.Unsurprisingly,there is a long history of attempts to identify a comparative set of indicators to measure port or terminal performance.A brief review

104、 of the literature was provided in The Container Port Performance Index 2020:A Comparable Assessment of Container Port Performance(World Bank 2021),CPPI 2020,which illustrated the broad approaches identified and commented on the merits and demerits of each.The measures fell into three broad categori

105、es:Firstly,measures of operational and financial performance;secondly,measures of economic efficiency;and thirdly,measures that rely,predominately,on data from sources exogenous to the port.This review has not been replicated in CPPI 2023,and interested readers are directed to CPPI 2020(World Bank 2

106、021),or the extant literature.One of the general challenges of nearly all the approaches has been the quality,consistency,and availability of data;the standardization of definitions employed;and the capacity and willingness of organizations to collect and transmit the data to a collating body.At a s

107、lightly higher level,there are several aggregate indicators that provide an indication of the comparative quality and performance of maritime gateways.The World Bank Logistics Performance Index(LPI)(Arvis et al.2018)and the World Economic Forums Global Competitiveness Index(GCI)4.0 both report on th

108、e perceived efficiency of seaport services and border clearance processes and indicate the extent to which inefficiencies at a nations sea borders can impact international trade 11|inTroducTioncompetitiveness.But the aggregate nature of the indicators,and the fact that they are perception based,mean

109、s that they offer at best an indication of comparative performance and offer little to guide spatial or operating performance improvements at the level of the individual port.This could change if the next version of the LPPI reflects the movement of the consignment from origin to destination.The Uni

110、ted Nations Conference on Trade and Developments(UNCTADs)Liner Shipping Connectivity Index(LSCI)provides an indicator of a ports position within the liner shipping network,which is partly a result of the ports performance,but does not directly measure it.Like the CPPI,the LSCI is limited to containe

111、r ports.Digitalization offers an opportunity to measure and compare container port performance in a robust and reliable manner.New technologies,increased digitalization and digitization,and growing willingness on the part of industry stakeholders to work collectively toward system-wide improvements

112、have created the capacity and opportunity to measure and compare container port performance.The data used to compile the CPPI 2023 are from S&Ps Global Port Performance Program.This program commenced in 2009 to drive efficiency improvements in container port operations and supporting programs to opt

113、imize port calls.The aim of CPPI was to utilize the existing empirical data to establish an unbiased metric for comparing container port performance among different ports,over time.The performance of container ports is most relevant in terms of customer experience,specifically the speed and efficien

114、cy with which customer assets are handled.In this fourth of CPPI,the focus remains exclusively on quayside performance,which reflects the experience of a container ship operator-the ports primary customer-and its fundamental value stream.The operational efficiency of how ports receive,and handle con

115、tainer ships is critically important in a carriers decision to choose a port over other options.The purpose of the CPPI is to help identify opportunities to improve a terminal or a port that will ultimately benefit all public and private stakeholders.The CPPI is intended to serve as a benchmark for

116、important stakeholders in the global economy,including national governments,port authorities and operators,development agencies,supranational organizations,various maritime interests,and other public and private stakeholders engaged in trade,logistics,and supply chain services.The joint team from th

117、e World Bank and S&P Global Market Intelligence intends to continue to enhance the methodology,scope,and data in future annual iterations,reflecting refinement,stakeholder feedback,and improvements in data scope and quality.ReferencesArvis,Jean-Franois,Lauri Ojala,Christina Wiederer,Ben Shepherd,Ana

118、suya Raj,karlygash Dairabayeva,and Tuomas kiiski.2018.Connecting to Compete 2018:Trade Logistics in the Global Economy.Washington DC:WorldBank.https:/openknowledge.worldbank.org/bitstream/handle/10986/29971/LPI2018.pdf.Clark,ximena,David Dollar,and Alejandro Micco.2004.“Port Efficiency,Maritime Tran

119、sport Costs,and Bilateral Trade.”Journal of Development Economics 75(2):417450.https:/doi.org/10.1016/j.jdeveco.2004.06.005.Hoffmann,Jan,Naima Saeed,and Sigbjrn Sdal.2020.“Liner Shipping Bilateral Connectivity and Its Impact on South Africas Bilateral Trade Flows.”Maritime Economics&Logistics 2020,2

120、2(3):473499.DOI:10.1057/s41278-019-00124-8.inTroducTion|12 kathuria,Sanjay.2018.A Glass Half Full:The Promise of Regional Trade in South Asia.Washington DC:WorldBank.https:/openknowledge.worldbank.org/handle/10986/30246.Levinson,Marc.2006.The Box:How the Shipping Container Made the World Smaller and

121、 the World Economy Bigger.Princeton,New Jersey,United States:Princeton University Press.Micco,Alejandro,Ricardo J.Sanchez,Georgina Pizzolitto,Jan Hoffmann,Gordon Wilmsmeier,and Martin Sgut.2003.“Port Efficiency and International Trade:Port Efficiency as a Determinant of Maritime Transport Costs.”Mar

122、itime Economics&Logistics,5(2):199218.DOI:10.1057/palgrave.mel.9100073.UNCTAD(United Nations Conference on Trade and Development).2021.Review of Maritime Transport 2021.Geneva:UNCTAD.https:/unctad.org/webflyer/review-maritime-transport-2021.Wilmsmeier,Gordon,Jan Hoffmann,and Ricardo J.Sanchez.2006.“

123、The Impact of Port Characteristics on International Maritime Trade Costs.”Research in Transportation Economics,16(1):117140.DOI:10.1016/S0739-8859(06)16006-0.World Bank.2013.“Opening the Gates:How the Port of Dar es Salaam Can Transform Tanzania.”Tanzania Economic Update 3,May 21,2013.https:/www.wor

124、ldbank.org/en/country/tanzania/publication/opening-the-gates-how-the-port-of-dar-es-salaam-can-transform-tanzania-backup#:text=US%241%2C759%20million%20%E2%80%93%20the%20total,port%20of%20Dar%20 es%20Salaam.World Bank.2022.The Container Port Performance Index 2021:A Comparable Assessment of Containe

125、r Port Performance.Washington,DC:World Bank.World Bank.2023.The Container Port Performance Index 2022:A Comparable Assessment of Container Port Performance.Washington,DC:World Bank.World Bank.2024.The Container Port Performance Index 2023:A Comparable Assessment of Container Port Performance.Washing

126、ton,DC:World Bank.13|The PorT Performance ProGram2.The Port Performance ProgramIntroductionContainer(liner)shipping services are generally highly structured service rotations.They are typically set up with weekly departure frequencies,a fixed sequence of port calls,and standard pro forma day and tim

127、e-specific berthing windows.Once a service has been defined or adjusted,it will usually remain intact for many months,or even years.The berthing windows are pre-agreed with the terminal and port operators,usually based on a slightly higher than expected average quantity of container exchange moves,a

128、nd ideally modest buffers in the sea legs between ports.The clear advantages of this model are that shippers can make long-term supply decisions and ports and terminals schedule and balance their resources to meet expected demand.With a well-planned and well-executed pro forma schedule,they can achi

129、eve higher levels of reliability and predictability.This,in turn,can lead to more effective supply chain operations and planning as container ships spend around 15 percent to 20 percent of their total full rotation time in ports,with the balance being spent at sea.Reduced port time can allow ship op

130、erators to reduce vessel speed between port calls,thereby conserving fuel,reducing emissions,and lowering costs in the process.Conversely,for every unplanned additional hour in port or at anchorage,the ships need to increase speed to maintain the schedule,resulting in increased fuel consumption,cost

131、s,and emissions.The PorT Performance ProGram|14 Inextreme cases,ships that fall many hours behind their pro forma schedule will start to arrive at ports outside of their agreed windows,causing berth availability challenges for ports and terminals,particularly those with high berth utilization rates.

132、This,in turn,causes delay to shipments and disruption to supply chains.A service recovery can involve significantly higher sailing speeds,and therefore,higher fuel consumption,emissions,and costs,or the omission of a port or ports from the service rotation.Time is valuable for stakeholders,and so it

133、 is logical to measure port performance based on the total amount of time ships are required to spend in port.The CPPI 2023 has again been developed based on the total port time in the manner explained in subsequent sections.This iteration has utilized data from the full calendar year of 2023 and ha

134、s employed the same two approaches as the earlier editions,an administrative approach,and a statistical approach.The resulting ranking of container port performance reflects as closely as possible actual port performance,while being statistically robust.The data are discussed in this section,with th

135、e methodologies discussed in Chapter 3.The results are presented in Chapter 4,and in more detail in Appendix A.The Port Performance ProgramThe data used to compile the CPPI is from S&P Globals Port Performance Program.The program was started in 2009 with the goal of supporting efficiency improvement

136、s in container port operations and to support projects to optimize container port calls.The program includes 10 of the worlds largest liner shipping companies that collectively operate close to 80 percent of global fleet capacity.The liner shipping companies provide the program with a series of data

137、 points comprising operational time stamps and other bits of information such as move counts for each individual port call undertaken globally.The data are provided monthly and cover the full global networks of each liner shipping company and their subsidiaries.In 2023,performance time stamp data we

138、re captured for 194,198 port calls involving 253.7 million container moves at 876 container terminals in 508 ports worldwide.Following receipt from the shipping lines,the port call data undergoes several validation and quality checks before mapping to historical AIS vessel movement data,which enable

139、s tracking and verification of the shipping line data.The geo-fencing of port and terminal zones within the AIS system supports the creation of several of the performance metrics tracked in the program.Most of the port performance metrics are constructed from the combined AIS and liner shipping data

140、.The combination of empirical shipping line data and AIS movement data enables the construction of more accurate and granular metrics to measure container port performance.Many of the metrics consist of a time component cross-referenced with workload achieved in that time,either in the form of move

141、counts or a specific task within the container port call process.Time stamps,definitions,and methods to calculate metrics are fully standardized in collaboration with the shipping line partners in the program.The Automatic Identification System and Port ZoningAIS technology is used to track and moni

142、tor vessels in near real time.It sends information on a vessels movement,speed,direction,and other particulars via satellite and terrestrial stations.The systems function as a localized service,and indeed global tracking,was initially considered secondary.The AIS primarily functions as a navigationa

143、l safety aid,to ensure the safety and efficiency of navigation,safety of life at sea,and maritime environmental protection.1 AIS was designed for the avoidance of vessel collision,as outlined in the Safety of Life at Sea(SOLAS)Convention.215|The PorT Performance ProGramAll ships of net tonnage of at

144、 least 300 gross register tonnage(GRT)performing international voyages,all cargo ships of at least 500 GRT not performing international voyages,and all passenger ships,regardless of size,should be equipped with AIS.This allows vessels to automatically transfer data and a plethora of navigational and

145、 identification information to other nearby ships and relevant port authorities in the form of structured messages.3 The technical requirements for AIS are specified by the International Telecommunication Union(ITU)Recommendation ITU-R M.1371-5(02/2014).4For maritime domain awareness and safety purp

146、oses,the use of continuous 24/7,near-real-time online AIS data makes it possible to monitor areas,vessels,and routes;generate shore-based alerts;and provide useful positional and navigational information in general(IALA 2005).Satellite-based AIS receivers offer coverage outside the land-based antenn

147、as range by covering the whole globe from pole to pole.Satellite AIS coverage can extend to the entire exclusive economic zone(EEZ)or globally,including remote coastal areas(IALA 2016).In the case of ports5,the usage of zones helps in recording a vessels navigational status and positioning.AIS zones

148、 offer different indicators activated automatically by the vessels signal reporting its position.Every port has at least one zone created in a way that captures the arrivals and sailings of vessels at cargo-handling facilities but avoids spurious reports being recorded from passing traffic.Where a s

149、ubject port is geographically spread out with terminals located remotely,it is likely that there will be more than one zone,with all zones linked by a standard port identification number.Ports that straddle a river or another similar body of water will often have zones along opposing shorelines with

150、 a track separating them,thus avoiding the capture of AIS reports from traffic navigating through a fairway or channel.Once again,the individual zones will be linked to their common port using the ports unique identification number.Zones also cover anchorages to record vessels arriving at a port but

151、 awaiting authority to enter,or vessels laid up awaiting orders.Additional zones cover the arrival of vessels at repair yards or those navigating locks.Anchorage zones may be created on an ad hoc basis.Not all ports have anchorage areas and among those that do,not all are shown in nautical charts.Wh

152、enever possible,S&P Global uses its own tracking and observation tools to determine where vessels anchor and create zones accordingly.Each anchorage zone is linked to the relevant port using the subject ports unique identification number.AIS is generally reliable,but it also has limitations that can

153、 impact the transmission and quality of the data captured.Some factors that may affect the signal could be the AIS transponder being turned off deliberately,problematic reception,high traffic density areas,weather conditions,or anomalous positions.The Anatomy of a Port CallEvery container ship port

154、call can be broken down into six distinct steps.These individual steps are illustrated in Figure 2.1.Total port hours is defined as the total time elapsed between when a ship reaches a port(either port limits,pilot station,or anchorage zone,whichever event occurs first)to when it departs from the be

155、rth after having completed its cargo exchange.The time spent from berth departure(All Lines Up)to the departure from the port limits is excluded.This is because any port performance loss that pertains to departure delays,such as pilot or tug availability,readiness of the mooring gang,channel access

156、and water depths,forecasting completion time,communication,and ship readiness will be incurred while the ship is still alongside the berth.The PorT Performance ProGram|16 Additional time resulting from these causes will,therefore,be captured during the period between 4.Last Lift and 5.All Lines Up(“

157、berth departure).FIGURE 2.1 The Anatomy of a Port Call125346Arrival Port LimitsAll Lines FastFirst LiftLast Lift Arrival At Anchorage and Waiting Time at Anchorage (Berth,Channel,Pilot etc.)Steam in Time Port Limit to All Lines Fast.Gangway down,authority clearence,abour available,position crenes,un

158、lash,load approval,etc All cargo operations,driven by Crane Intesity and Gross Crane Performance Lashing and checks,authority clearence,crew onboard,engine ready,repairs completed,bunkers,channel clear,tugs&pilotSteam outAll Lines UpExit Port LimitsPOINTS OF ACTIVITY123456Source:Original figure prod

159、uced for this publication.Ships may spend extra time in a port after the departure from a berth,but the time associated with these additional activities is excluded from the CPPI,as they are not influenced by the operational performance of the terminal or port.Ships may dwell within a ports limits f

160、or bunkering,repairs,or simply waiting in a safe area if they are unable to berth on arrival at their next port.Apart from bunkering being performed simultaneously with cargo operations,these causes of additional port time are not necessarily reflective of poor performance and hence,are excluded fro

161、m the CPPI.Although none of these factors necessarily indicate port inefficiency,they can contribute to additional time spent in the port.For instance,clearance authorities delays can result in delays in the first lift and idle time after cargo operations have concluded.However,the data available do

162、 not provide enough detail to identify the root causes of such delays.It is assumed that only a small percentage of ships idle at the berth after cargo operations due to factors unrelated to port performance,and their inclusion does not significantly affect the CPPI rankings.The other four component

163、s of the port call can logically be grouped into two distinct blocks of time.The first comprises elapsed time between Arrival Port Limits and All Lines Fast(steps 1 and 2 in Figure2.1);the second comprises time elapsed between All Lines Fast and All Lines Up(steps 2 to 5,also commonly referred to as

164、 berth time or berth hours).The logic behind this division is that while there will always need to be time consumed between steps 2 and 5,the bulk of time between steps 1 and 2,excluding actual sailing in time,is waiting time,which can be eliminated.17|The PorT Performance ProGramOverall Port Time D

165、istributionThe time stamps in the source data allow us to break down and summarize total port time into three categories:Arrival Time,Berth Idle,and Cargo Operations.Expressed as a percentage of total port hours recorded,the distribution of port time per ship size range and globally aggregated is sh

166、own in Figure 2.2.FIGURE 2.2 In-Port Time Consumption44.4%55.8%60.2%66.6%73.8%59.4%16.9%12.2%11.7%10.5%8.0%11.7%38.7%32.0%28.1%22.9%18.2%28.8%0%10%20%30%40%50%60%70%80%90%100%13,500OverallShip Size Range(nominal TEU)Arrival TimeBerth IdleCargo OperationsSource:Original figure produced for this publi

167、cation,based on CPPI 2023 data.As there is naturally some correlation between ship size and call size,a higher percentage of time is required for cargo operations for the larger ships,and this will be explored in detail later in this report.What is interesting,and surprising at the same time is that

168、 only 60 percent of the total port time is attributable to cargo operations,meaning there is potentially a lot of wastage in terms of excess time in the system.The average duration of a port call in 2023 was 40.5 hours,which represents a slight increase over the global average of 36.8 hours in 2022.

169、About 11.7 percent(or 3.71 hours)was idle time consumed at the berth immediately before and after cargo operations.Also known as the Start-Up and Finish sub-processes of a port call,each activity does not necessarily need to take more than 30 minutes to complete safely.There is,therefore,an opportun

170、ity to eliminate almost nearly four hours per call of port time globally simply through better planning,preparation,communication,and process streamlining.This time saved equates to more hours at sea,leading to slower sailing speeds,lower GHG emissions,and cost savings for the ship operator,which wo

171、uld be significant for each port call.In the second half of 2020,there was a rebound in the global sales of durable goods,most prominently in the US,and a sharp increase in the overall container volume demand.This coincided with continued COVID-19 restrictions and resulted in the emergence of severe

172、 port congestion.In 2021,this port congestion was still manifesting itself,reaching a peak in the third quarter of 2021 and the average The PorT Performance ProGram|18 arrival time per port call globally remained above 11 hours until the third quarter of 2022.The fourth quarter of 2022 saw reducing

173、volumes and many ports were able to clear backlogs and reduce average arrival times to close to 10 hours per port call.The expectation was that the average port arrival time globally in 2023 will continue to decline to levels prior to the start of 2021,which is what has transpired.(see Figure 2.3)FI

174、GURE 2.3 Global Average Arrival Time Development .33 11.07 11.35 10.10 9.64 8.49 8.90 9.57 8.00 8.50 9.00 9.50 10.00 10.50 11.00 11.502022Q12022Q22022Q32022Q42023Q12023Q22023Q32023Q4Average Arrival HoursSource:Original figure produced for this publication,based on CPPI 2022-23 data.At a r

175、egional level and broken down by ship size groups,the change in average arrival time per region and per ship size group over the 2022-2023 period is illustrated in Table 2.1.The column All shows the aggregate change in quantity of hours from arrival at port limits or start of anchorage time,to berth

176、ing for cargo operations to commence for each region,across all ship size groups.19|The PorT Performance ProGramTABLE 2.1 Average Arrival Time Development per Region and Ship Size,20222023CHANGE(HR)SHIP SIZE RANGEREGION1 13,500ALLAFR3.8 2.0(2.5)7.4 14.4 2.0 LAM1.4 0.3 1.3 0.5(0.0)0.6 MED2.2 1.4(0.3)

177、(0.6)(4.0)0.9 MEI4.3 2.6 1.5(0.0)(0.1)1.6 NAM(3.7)(10.0)(19.9)(28.0)(33.8)(19.1)NEA(1.6)(2.0)(1.1)(0.8)(0.3)(1.4)NEU0.2(0.3)(4.7)(6.9)(7.6)(3.1)OCE(2.1)(1.1)(2.4)(0.2)(1.3)SEA(2.8)(2.8)(1.2)(0.6)(0.7)(2.0)Global1.0(0.9)(3.6)(3.3)(3.4)(1.8)Source:Original table produced for this publication,based on

178、CPPI 2022 and 2023 data.At a global level,on average each port arrival decreased by 1.8 hours,as illustrated in Table 2.2.The largest increase in average arrival time was witnessed in North America(USA and Canada)with an average increase in time of 19.1 hours over all vessel sizes.By contrast,perfor

179、mance improved in Africa(Sub-Sahara)with an average 2.0-hour improvement in arrival time across all vessel sizes.Improvements in East Asia and Southeast Asia were also recorded.The overall improvements and reductions in average arrival hours in African ports has been driven by Dar Es Salaam,Monrovia

180、,Douala,Pointe-Noire,Tema,Luanda,Lom,Lagos,Port Victoria,Dakar,and Ngqura.The increase is slightly offset by increased average arrival time in Cape Town,San Pedro,Abidjan,and Mombasa.In East Asia,improvements were seen in Yantian and Yangshan but countered by increased time in Manila and Qingdao.The

181、re are no European ports in the top 20 improvers.Poti,La Spezia,Mersin,Trieste,Hamburg,and koper all experienced longer average arrival times.Waiting time,defined as the period between Arrival Port Limits or when the ship enters an anchorage zone,and All Lines Fast can generally be regarded as waste

182、d time.As such,in the construction of the CPPI,one possibility was to apply a penalty to waiting time.The decision was taken not to do so,as the introduction of a penalty of this type would be a normative judgement inconsistent with the overall aim of the study to create bean objective quantitative

183、index.There was consideration as to whether to apply a discount to waiting time for the smallest segment of ships.Smaller ships generally suffer less priority than larger ones,and in some hub ports might be purposely idled at anchorage waiting to load cargo which is arriving from off-schedule ocean

184、going ships.However,after reviewing average arrival time for the various ship size segments on a regional basis,the data did not support applying a discount to waiting time for the smallest segment of ships.(see Table 2.2).The PorT Performance ProGram|20 TABLE 2.2 Average Arrival Time Performance pe

185、r Ship Size Range per Region2023SHIP SIZE RANGEREGION13,500AVERAGEAFR31.729.430.527.428.129.7LAM9.37.610.08.210.48.3MED11.89.76.86.77.19.6MEI18.210.07.26.77.18.8NAM5.57.211.615.820.311.7NEA4.76.27.36.35.76.2NEU9.17.78.88.29.48.6OCE15.513.111.98.412.6SEA7.47.45.35.53.76.6Average11.09.39.38.37.29.1Sou

186、rce:Original table produced for this publication,based on CPPI 2023 data.To test the significance of purposely delayed smaller feeder vessels on the overall ranking,we conducted a simulation within the overall CPPI model.For all ports(not only the focus ports),we reduced the quantity of arrival hour

187、s by 50 percent for all ship calls where the capacity of the ship is 1,500 TEU or less in size.The quantity of berth hours for all ships was maintained at 100 percent,as was the average arrival hours for all other ship size groups.Since it is not possible to see from the data whether waiting time is

188、 voluntary or forced,it is difficult to find a suitable level at which to discount waiting time in this scenario.The port calls of ships with less than 1,500 TEUs of capacity comprise just 10 percent of the total calls in the CPPI.Therefore,the disparity in waiting times between ships with less than

189、 1,500 TEUs of nominal capacity and other segments,as simulated,has only a small impact to the overall CPPI.To keep the data pure and avoid normative judgment that is inconsistent with an objective quantitative index,the rankings published in this iteration are not influenced by adjustments made to

190、empirically recorded port hours.The Significance of Call SizeAs illustrated in Figure 2.4,over 60 percent of a port call is consumed through cargo operations,for the handling of containers.In this aspect of the call,call size is of great significance.Call size is far less significant when it comes t

191、o arrival time,which is more likely to be influenced by ship size.There have been several earlier studies,in which ships are grouped into size segments(ranges)based upon their size or capacity and port calls are ranked based on the time elapsed in port or on the berth.While these studies provide an

192、indication,the optimum outcome requires the workload for each call to be taken into consideration.In this index,workload is represented by Call Size,defined as the total quantity of containers(regardless of size),which were physically discharged,loaded,or restowed during a port call.21|The PorT Perf

193、ormance ProGramFIGURE 2.4 The Aggregated Correlation between Ship and Call Size02,0004,0006,0008,00010,00012,00014,00016,00018,0000 500 1,000 1,500 2,000 2,500 3,000 3,50013,500Average Ship Size(TEU capacity)Average Call SizeShip Size Range(nominal TEU)Call SizeShip SizeSource:Original figure produc

194、ed for this publication,based on CPPI 2023 data.Although there will be some level of correlation between the ship and call size,it is not a perfect correlation.For example,an 18,000 TEU capacity ship calling at a port in Thailand or southern Vietnam might exchange 1,000-2,000 containers per call,but

195、 that same ship in Yangshan or Singapore might exchange more than 4,000 containers.Similarly,in the Thai or southern Vietnamese ports,a 3,000 TEU(feeder ship)might exchange more than 3,000 containers,potentially twice that of an 18,000 TEU mainline ship at the same port.The 60 percent of a port call

196、,during which containers are exchanged,is influenced by two sub-factors:1.The quantity of cranes deployed2.The speed at which the cranes,especially the long crane(the crane with the highest workload in terms of cycles),operateThe PorT Performance ProGram|22 FIGURE 2.5 Container Moves Performed per g

197、ross Crane Hour across Various Ship SizesGross Cranne Moves per hr perShip Size range20.9 22.6 24.2 24.8 24.4 23.6 18.0 19.0 20.0 21.0 22.0 23.0 24.0 25.0 26.013,500Total/AverageSource:Original figure produced for this publication,based on CPPI 2023 data.The variation in containers handled per gross

198、 crane hour across all ship sizes is statistically minor.The global average for all ships is 23.5 moves per hour,so the smallest ships are 9.4 percent less efficient than the average,whereas ships in the 8,501 TEU-13,500 TEU range are 3.6 percent more efficient than the average.It is often implied t

199、hat larger ships are more difficult to work,but the data says otherwise.On the larger ships,the crane operator has higher hoists and longer trolley distances,which increases cycle time,but this is offset by more moves per bay and hatch,resulting in more containers handled per gantry or hatch-cover m

200、ove.The smaller ships can often encounter list or trim issues,making it harder for the operator to hit the cell-guides and the hatch-cover and lashing systems.FIGURE 2.6 Gross Crane Productivity by Call SizeGroos Crane ProductivityCall Size19.6 20.8 22.8 23.6 23.5 23.6 24.2 24.9 25.1 23.7 23.6 19.0

201、20.0 21.0 22.0 23.0 24.0 25.0 26.06,000All/AverageSource:Original figure produced for this publication,based on CPPI 2023 data.23|The PorT Performance ProGramFIGURE 2.7 Crane Productivity by Crane Intensity31.029.027.025.023.021.019.017.012345Rounded Crane Intensity678930.722.522.623.123.724.023.221

202、.818.0Moves per Gross Crane Hour Source:Original figure produced for this publication,based on CPPI 2022 data.A review of gross crane productivity versus call size and crane intensity reveals no strong increases or decreases through the ranges.Assessed on call size ranges,there is a 5.2 percent to 3

203、.8 percent variation to the average.Meanwhile,an assessment of crane intensity reveals that the first and last segments have extremely high and low performances,respectively,but in the mid-range,there is little difference in crane productivity across the seven ranges.This implies that crane speed(pr

204、oductivity)does not gradually increase(or decrease)as ship size,call size,or crane intensity increases.It is therefore statistically not a key determinant of operating hours.The far more significant influencer of operating time is the quantity of cranes deployed(crane intensity).FIGURE 2.8 Call Size

205、 versus Crane Intensity6,000Call Size Range1.5 1.8 2.1 2.6 3.1 3.5 3.7 4.1 4.7 4.4 1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5 5.0Crane IntensitySource:Original figure produced for this publication,based on CPPI 2023 data.The PorT Performance ProGram|24 FIGURE 2.9 Average Moves per CraneContainer Moves per Quay

206、 CraneCall Size Range6,000All/Average117 212 342 463 550 642 737 840 1,013 1,699 428 0 200 400 600 800 1,000 1,200 1,400 1,600 1,800Source:Original figure produced for this publication,based on CPPI 2023 data.As might be expected,the more container moves are to be handled,the more cranes must be dep

207、loyed.However,crane intensity lags call size growth,which means that as the call size grows,each crane is required to handle more containers.Theoretically,if a call with 1,000 moves was assigned 2 cranes,then one with 5,000 moves would require 10 cranes for a status quo,and that does not happen ofte

208、n,if at all.Since the exchange rate per crane does not increase progressively with ship size,call size,or crane intensity growth,the overall operating time increases.This makes call size differentiation the critical factor to consider when attempting port performance benchmarking and ranking.25|The

209、aPProach and meThodoloGy3.The approach and methodologyThe Structure of the DataBefore discussing the methodology employed in constructing the CPPI with matrix factorization,it is helpful to first summarize the structure of available data.The data set is segmented by the following five categories of

210、ship sizes:Feeders:13,500 TEUsFor each category,there are 10 different bands for call size.The port productivity is captured by average idle hour,which consists of two parts:port-to-berth(PB)and on-berth(B).In the previous CPPI iteration,The aPProach and meThodoloGy|26 total variables used=5 x 10 x

211、2.Of course,many of them have missing values.The objective is to build a model to summarize these variables and then construct a port productivity index for all ports under consideration.The average waiting time and average berth time is calculated for each call size.The resulting data is a table/ma

212、trix whose rows represent ports and whose columns contain the average waiting and berth times of each call size.Moving on to the construction of the dataset for the CPPI,for a port to qualify for inclusion in the CPPI it must have registered at least 24 valid port calls where port hours can be calcu

213、lated within the full calendar year.Of the 508 ports for which S&P Global received port call information,405 are included in the main index of CPPI 2023.There were 182,855 distinct port calls recorded in the data over the period at those 405 main ports.A further 103 ports registered less than 24 cal

214、ls each,these ports are excluded from the CPPI 2023.The CPPI is based solely on the average port hours per port call,with port hours being the total time elapsed from when a ship first entered a port to when it departed from the berth.Due to the large volume of data,it was possible and prudent to br

215、eak it down into ship size and call size groups or ranges.However,too much fragmentation would have diluted the data to the extent that more assumptions than actual empirical data would be present in the index.Therefore,the data were grouped into five distinct ship sizes,and then within each ship si

216、ze group by call size group,as reflected in Figure 3.1 below.FIGURE 3.1 The Structure of the CPPICONTAINER PORTPERFORMANCE INDEX 1,500 TEU 6,000 moves1,501 5,000 TEU5,001 8,500 TEU8,501 13,500 TEU13,500 TEUSHIP SIZE GROUPSCALL SIZE GROUPSSource:Original figure produced for this publication.The numbe

217、r of ship size groups was limited to five,and the number of call size groups to 10.That results in a 50(5 x 10)matrix for the qualifying ports for the main index of CPPI 2022.However,there were insufficient port calls in the larger five call size groups for the less than 1,500 TEU ship size group an

218、d similarly for the two larger call size groups for the 1,501 TEU-5,000 TEU ship size group.In total,the data was distributed into 43 ship-call size groups.27|The aPProach and meThodoloGyTABLE 3.1 Port Calls DistributionCALL SIZE GROUPSHIP SIZE GROUP60001 13,5000.0%0.2%1.5%3.6%5.8%7.9%9.2%19.6%28.9%

219、23.2%Source:Original table produced for this publication,based on CPPI 2023 data.The five ship size groups were based on where they might be deployed and the similarities of ships within each group.Although a sixth group for ships more than 18,000 TEU or 24,000 TEU could have been added,it would hav

220、e highly diluted the data in the two larger ship size groups.TABLE 3.2 Ship Size Group DefinitionsNOMINAL TEU CAPACITY RANGEDESCRIPTIONLess than 1,500Almost exclusively feeder vessels,often connecting small outlying ports with regional hub ports.Some intra-regional services will also have ships in t

221、his size range.1,500 to 5,000A significant quantity of these classic Panamax ships are deployed on intra-regional trades.They are found on North-South trades to and from Africa,Latin America,and Oceania,as well as Transatlantic services.5,000 to 8,500Vessels within this size group are mainly deploye

222、d on the North-South trade lanes.Vessel cascading and improving port capabilities has seen them start to emerge as stock vessels for Africa,Latin America,and Oceania trades.There is some presence on Transatlantic and AsiaMiddle East trades as well.8,500 to 13,500These Neo-Panamax vessels are largely

223、 deployed on East-West trades,particularly Trans-Pacific,both to North Americas west coast as well as via either the Panama or Suez Canals to North Americas east coast.They also feature on AsiaMiddle East trades,with some deployed on AsiaMediterranean rotations.Greater than 13,500These ultra-large c

224、ontainer ships(ULCS)are mainly deployed on AsiaEurope(serving both North Europe and the Mediterranean)and AsiaUnited States trades,especially on Trans-Pacific services calling at North Americas west coast ports.Source:Original table produced for this publication.The application of ship size groups i

225、s less important than call size groups,particularly since the call data is already split into 10 call size groups.However,the objective of the CPPI is to highlight through comparison the performance gaps and opportunities to save fuel and reduce emissions.The analysis should,therefore,consider that

226、the larger the ship,the more fuel it consumes,and the higher the potential to save fuel and reduce emissions.The aPProach and meThodoloGy|28 FIGURE 3.2 Percentage of Port Calls per Ship Size Group-202313,5009%Ship Size Distribution by Calls-2023 5,0018,50015%Source:Original figure produced for this

227、publication,based on CPPI 2023 data.Almost 47 percent of all ship port calls in 2023 were from the Panamax(1,501-5,000 TEU)size of ships.With just 9 percent of port calls made by ships more than 13,500 TEU,it was decided not to disaggregate these further.As the main participants of the Port Performa

228、nce Program are primarily deep-sea operators,there was a relatively small number of calls in the feeder segment(less than 1,500 TEU capacity).An attempt has been made to make the 10 call size groups as narrow as possible by grouping together calls in instances where they are most likely to have rece

229、ived similar crane intensity provisions.The analysis then compares all qualifying ports on how close(or far)the individual call size is to the average call size within each call size group.TABLE 3.3 Call Size SensitivityCALL SIZE SENSITIVITYCALL SIZE GROUP6000Average166 377 730 1,228 1,732 2,230 2,7

230、35 3,437 4,755 7,804 Median177 379 722 1,218 1,719 2,220 2,726 3,408 4,667 6,932 Lower Range166 377 730 1,228 1,732 2,230 2,735 3,437 4,755 7,804 Upper Range177 379 722 1,218 1,719 2,220 2,726 3,408 4,667 6,932 Total Ports367389369315311260Within Range254355323304259249Percenta

231、ge in Range69.2%91.3%87.5%97.1%100.0%100.0%100.0%100.0%98.2%81.7%Source:Original table produced for this publication,based on CPPI 2023 data.29|The aPProach and meThodoloGyTo assess the sensitivity within each call size group across all 405 qualifying ports,the median call size between all ports wit

232、hin a call size group was taken and a tolerance range of 15 percent above and below the median created(see Table 3.5).In the six call size groups from the 1,0011,500 to 4,0016,000 moves groups,more than 96.9 percent of ports have an average call size well within this tolerance range.Beyond the thres

233、hold of 6,000 moves per call,the call size has a much lower impact on crane intensity.This is because the number of cranes that can be deployed is limited by the overall number of cranes available or stowage splits.The quantity of ports with an average call size within the tolerance range in the thr

234、ee smallest call size groups is not as high as the quantity in the six call size groups from the 1,0011,500 to 4,0016,000 moves groups.However,for ports with an average call size above the tolerance range,it would be possible to increase crane intensity to match the slightly higher call sizes,and,th

235、erefore,the conclusion is that objective comparisons can be made within all 10 call size groups.The objective of preparing the index and the ranking is that it should reflect as closely as possible actual port performance,whilst also being statistically robust.With respect to the largest portsthe to

236、p 100 ports by annual move countthere is real empirical data present in each of the 43 distinct ship size and call size categories.However,for smaller ports there are many categories with no data,particularly those with only a few hundred calls in total.If these unpopulated categories are ignored,th

237、e appraisal of performance would be undertaken on different quantities of categories,which is likely to unduly disadvantage smaller ports that might well be quite efficient despite their modest size and throughput.Constructing the Index:The Administrative ApproachImputing missing values:The administ

238、rative approachThe handicap of missing values can be addressed in two different ways in the administrative approach and the statistical approach.The former involves assigning values to empty categories based on data that are available when a port has registered a data point within a specific ship si

239、ze range.TABLE 3.4 Quantity of Ports Included per Ship Size GroupSHIP SIZE RANGEQUANTITY OF PORTS INCLUDEDBASE CALL SIZELess than 1,500 TEUs3272515001,5005,000 TEUs3745011,0005,0008,500 TEUs2271,0011,5008,50013,500 TEUs1861,5012,000More than 13,500 TEUs1173,0014,000Source:Original table produced for

240、 this publication,based on CPPI 2023 data.For each ship size group,the call size group that has the largest quantity of data representation is selected(see Table 3.4)as the Base Call Size group.Ideally,this is a mid-range call size group because the lowest and highest groups can demonstrate some uni

241、queness.In cases where there is no actual data for the base call size group,the next highest group is examined to find an actual data set.If none is found,then the approach involves looking at the immediately lower call size band.At the end of this exercise,every port has a value assigned for the ba

242、se call size group.The aPProach and meThodoloGy|30 Imputing vessel arrival values.Where a call size group does not have an arrival hours value,it is populated using the overall average arrival time for all vessels registered at that port across all call size groups within each specific ship size gro

243、up.This is logical as call size is a less important determinant of waiting time than ship size.Imputing berth hours.From the base call size group,moving left toward the lowest group and right toward the highest group,in groups where no value exists,a value is determined on a pro rata basis given the

244、 adjacent call size group value,actual data or imputed.The rationale is that if within one call size group a port has either higher or lower berth hours than the average,the adjacent call size group too is likely to show similar trends.Table 3.5 contains an illustrative example.In this case,port A h

245、ad a higher quantity of hours in the base call size group than the group average.It is assumed that would also have been the case had the port registered actual calls in the 5011,000 and 1,5012,000 call size groups.The opposite is true for port B,which achieved a lower quantity of hours in the base

246、call size group.The calculation for port A in the 5011,000 call size group is actual hours within the group 1,0011,500(12.0)multiplied by the group average factor(0.9)for a prorated quantity of average berth hours of(10.8).TABLE 3.5 An Example of Imputing Missing ValuesPORTCALL SIZE GROUP5011,0001,0

247、011,5001,5012,000Port A10.812.014.4Port B7.28.09.6Group Average9.010.012.0Factor Multiplier0.9Base1.2Source:Original table produced for this publication,based on CPPI 2023 data.Note:The numbers in the green highlighted cells have been imputed by multiplying the base cells by the factor multiplier de

248、termined by the overall group average.The inherent risk with this approach is that poor or good performance within just one group will cascade across all call size groups.It also assumes that a ports ability to add cranes to larger call size groups exists,which might not be true in all cases.On the

249、other hand,it would be illogical to blindly assume that any port would simply achieve the average of the entire group or,possibly worse,to assume that a port performing below average in one call size group would miraculously perform much better than average in others where it did not record any actu

250、al calls.Constructing the index:the administrative approachAggregating arrival and berth hours into total port hours.This report indicated earlier that a case could be made for penalizing waiting time which is regarded as pure waste.However,as expressed earlier,this would be a normative judgment,acc

251、ordingly both arrival and berth hours are weighted as 1.0 and the two time segments are summed to form total port hours in CPPI 2021.Appraising port hours performance.Average port hours are naturally higher in the larger than smaller call size groups.This can magnify the difference in hours between

252、a subject port and the average port hours of the overall group.So,appraising on the difference between a ports average hours and average hours of the group may skew the scoring unduly toward the larger call size calls.There are also 31|The aPProach and meThodoloGyfar fewer calls within the larger th

253、an smaller call size groups,and this also needs to be reflected in the construction of the CPPI to retain maximum objectivity.The method applied to each call size group individually is that the ports average port hours is compared with the groups average port hours as a negative or positive quantity

254、 of hours.The result of that comparison is weighted by the ratio of port calls in each call size group for the entire group of ports Table 3.6 provides an illustration as to how it is done.7TABLE 3.6 Port Hours Performance AppraisalPORTPORT HOURSHOURS DIFFERENCECALL SIZE GROUP WEIGHTRESULTExample Po

255、rt22.5612.090.1601.9344Group Average34.65Source:Original table produced for this publication,based on CPPI 2023 data.In this illustrative example,the subject port used 12.09 fewer hours than the average of the entire group(22.56 versus 34.65).Since 16.0 percent of all port calls in this ship size gr

256、oup were in the subject call size group,the difference in hours(12.09)is multiplied by ratio 0.160 for an overall index points result of 1.9344.Where a port uses more port time than the average for all ports,the index points become negative.Aggregation to a score and rank per ship size group.The“res

257、ults”achieved per port within each of the 10 call size groups are then summed together to calculate a score within the overall ship size group(it is five and eight groups rather than 10 groups in the case of the two smaller ship size groups,respectively).Based upon these scores,there is a sub-rankin

258、g performed within each ship size group that can be reviewed in the final CPPI rankings.However,the imputation method might unfairly appraise some ports that only recorded data within a few call size groups.If,for example,the performance in a few call size groups was worse than the average for all p

259、orts within the ship size group,this would be prorated to all call size groups.This required a judgment,as the alternative of ignoring call size groups without actual data,effectively resulting in a zero score for those groups,would not necessarily result in a better outcome.In the latter case,ports

260、 with limited call size diversity would not be credited with positive scores in each and every call size group which they are likely to have achieved if they had a greater diversity of call sizes.Aggregating all ship size groups No allowance was made for ports that did not handle ships within specif

261、ic ship size groups during the period under consideration.The quantity of ports being included per ship size group was presented earlier in table 3.2.The primary reason is many of the smaller ports are not capable of handling some of the larger ship sizes and so would in effect be awarded positive(o

262、r negative)results for scenarios that are physically impossible.The omission of scores within some ship size groups would only be an issue if an attempt was made to compare the performance of major mainline ports with those of far smaller ports.But this is a comparison that is neither fair nor valua

263、ble.For the comparison between similarly sized ports,this factor will not contribute,or at least not significantly.In aggregating the scores from the various ship size groups into the overall CPPI in the administrative approach,a factor was built in to differentiate the importance and significance o

264、f better performance The aPProach and meThodoloGy|32 of larger ships over smaller ones.This was constructed based on the relative fuel consumption(and,therefore,emissions and cost)of different ship sizes in the form of an index(see Table 3.7).For each ship size group,a typical mid-range example ship

265、 was selected.Based upon the expected deployment of such ships,a range of sea legs were defined(and weighted),at a typical pro forma service speed,and the impact on fuel consumption that one hour longer(or shorter)in port would be likely to yield.TABLE 3.7 Assumptions to Determine a Fuel Consumption

266、 IndexNOMINAL TEU CAPACITY RANGEEXPECTED DEPLOYMENTSEA LEGWEIGHT(PERCENT)INDEX WEIGHTLess than 1,500 TEUsFeedersIntra-regionalSingaporeSurabayaRotterdamDublinkingstonPort-au-PrinceBusanQingdao25 25 25 25 0.461,500 to 5,000 TEUsIntra-regionalAfricaLatin AmericaOceaniaTransatlanticShanghai ManilaRotte

267、rdamGenoaAlgecirasTemaCharlestonSantosxiamenBrisbaneFelixstoweNew York30 30 10 10 10 10 1.005,000 to 8,500 TEUsAfricaLatin AmericaOceaniaTransatlanticAsiaMiddle EastHong kongTemaCharlestonSantosxiamenBrisbaneFelixstoweNew YorkShanghaiDubai20 20 20 20 20 1.548,500 to 13,500 TEUsTranspacificAsiaMiddle

268、 EastAsiaMediterraneanBusanCharleston(via Panama)Hong kongLos AngelesShanghaiDubaiSingaporePiraeus252525 25 1.97Greater than 13,500 TEUsAsiaMediterraneanAsiaNorth EuropeTranspacificSingaporePiraeusSingaporeRotterdamHong kongLos Angeles40 40 20 2.57Source:Original table produced for this publication,

269、based on CPPI 2023 data.The index weight then suggests that it is 2.57 times more costly to recover an additional hour of port time at sea for a ship with capacity in excess of 13,500 TEUs than it would be for a ship in the 1,5005,000 TEU capacity range.The total aggregated index points per port wit

270、hin each ship size group are then weighted by this“cost”factor.The sum of the weighted index points for each port across all five ship size groups are then summed and the final CPPI ranking is based upon those weighted values.The primary focus was micro-delays and it was assumed that these would be

271、recovered on long-haul ocean legs,and not between coastal ports,which would be more costly.Through simulation,if the index values are tweaked up or down by up to 10 percent the overall ranking is unaffected.If they are adjusted so that larger ship size groups have lower indices than smaller ones it

272、results in radical changes to the overall ranking.To achieve a final CPPI score and ranking in the administrative approach,accumulated results within each ship size group are multiplied by the index values per ship size group 33|The aPProach and meThodoloGyand then summed.The ranking is then based i

273、n descending order on final summed totals across all ship size groups.The resulting index using the administrative approach is presented in chapter 3 and appendixA.Constructing the index:the statistical approachImputation of Missing ValuesA major practical problem is that most idle hour variables ha

274、ve a significant number of missing values.For instance,in the port performance data set,the two smaller ship sizes contain little data for the larger call sizes.Consequently,as in the administrative approach,the call size groups with more than 2,000 moves were removed from the 1,500 TEU ship categor

275、y,and the call size groups with more than 4,000 moves were removed from the 1,501 TEU5,000 TEU ship category.A more sophisticated approach is to use likelihood-based methods to impute those missing values.For the current data set,expectationmaximization(EM)algorithm can be utilized to provide a maxi

276、mum-likelihood estimator for each missing value.It relies on two critical assumptions.The first assumption is that gaps are random,or more specifically,the gaps are not caused by sample selection bias.The second assumption is that all variables under consideration follow a normal distribution.Given

277、the data set,these two assumptions are plausible.EM computes the maximum likelihood estimator for the mean and variance of the normal distribution given the observed data.knowing the distribution that generates the missing data,we can then replace the missing values by their conditional expectation

278、given the available data.Matrix factorization can then be performed on the resulting data set,instead of the original one filled with missing values.Missing values in the resulting table/matrix are reconstructed using the EM algorithm(Dempster,Laird,and Rubin 1977).A non-negativity constraint is add

279、ed to make sure the reconstructed times are non-negative.Assuming the data has a multivariate Gaussian distribution with mean vector and covariance matrix,the EM algorithm provides an estimate of the two parameters and via maximum likelihood.Missing values are imputed using their conditional expecta

280、tion.In this approach,given a row with available values and x_a missing values x_m,the missing values are imputed by their conditional expectation E(x_m 1_(x_m)|x_a)given the available data,where the expected value is computed only over the non-negative values of to ensure the imputed values are non

281、-negative.The aPProach and meThodoloGy|34 In this iteration,arrival and berth hours are aggregated into total port hours,just like in the administrative approach.The data structure after this aggregation for a particular category k(k=1,2,3,4,5)can be summarized as shown in Table 3.8.TABLE 3.8 Sample

282、 Port Productivity Data Structure by Ship SizeSHIP SIZE(K)CALL SIZE BAND(NUMBER OF MOVES)PORTS6,000PORT-TO-BERTHBERTHTOTAL PORT HOURSPORT-TO-BERTHBERTHTOTAL PORT HOURSPORT-TO-BERTHBERTHTOTAL PORT HOURS123.Source:Original table produced for this publication.Why Is Matrix Factorization Useful?Essentia

283、lly,for each port,quite a few variables contain information about its efficiency.These include average time cost under various categories:(1)different call size bands,and(2)berth/port-to-berth.The reason matrix factorization can be helpful is that these variables are in fact determined by a small nu

284、mber of unobserved factors,which might include quality of infrastructure,expertise of staff,and so on.Depending on the data,very few of such factors can summarize almost all useful information.The challenge lies in the inability to observe those latent factors;however,a simple example could be helpf

285、ul:Imagine three ports,each with four different types of time cost,as shown in table 3.9.TABLE 3.9 Sample Illustration of Latent FactorsPORTCOST 1COST 2COST 3COST 4A1234B2468C36912Source:Original table produced for this publication.As one can observe,costs 2 to 4 are just some multiples of cost 1.Al

286、though we have four variables,to rank the efficiency of these three ports,just one variable is enough(ABC).This is an extreme case,but the idea can be generalized if these variables are somehow correlated,but to a less extreme extent.In that case,the factors are computed as some linear combination o

287、f costs 1 to 4.Of course,if costs 1 to 4 are completely independent of each other,then this method makes no sense.Fortunately,this is not the case for our data set.Thus,for each port,we can compute its score on all factors and then combine those scores together to reach a final efficiency score.35|T

288、he aPProach and meThodoloGyNote that in the statistical approach using matrix factorization,the scores are not calculated for each call size range.On the contrary,the whole data set,including the smaller ports,is used simultaneously to obtain latent factors.This is in sharp contrast to the administr

289、ative approach.The statistical approach factors in all the correlations among hours for various call size bands,which purely from a statistical perspective is more efficient.There is no right or wrong methodology,but the two different approaches are considered complementary.Hence,the decision in thi

290、s iteration of the CPPI to maintain both approaches,to try and ensure that the resulting ranking(s)of container port performance reflects as closely as possible actual port performance,whilst also being statistically robust.The Statistical MethodologyThe data are scaled and weighted as in the admini

291、strative approach.Let pij denote average port time of port i in call size j.Let pavgj denote the average of the average port time of all ports in the given call size.Let wj denote the ratio of port calls that are in the call size group jThe data are scaled by replacing pij by:xij=(pavg,j pij).wiA po

292、sitive value of xij means the port is doing better than average,whereas a negative value means it is doing worse than average.Let x=(xij)denote the resulting matrix of scaled port time.Assume x has n rows(n ports)and p columns(p call size bands).As in the previous iteration of the CPPI,the matrix x

293、is decomposed as x WH where W is a n k matrix and H is an entrywise non-negative k p matrix.The integer k(the number of columns of W)is chosen to be a small number to compress the data.The matrix W represents factors and the matrix H factor loadings that are used to explain the data x.A number of k=

294、3 factors was found to be adequate to approximate the data matrix x.Note:Traditional factor analysis(FA)used in statistical analysis produces a matrix factorization xWH as above,except that the matrix H does not need to be non-negative.This is a problem since a large positive factor does not necessa

295、rily represent a small port time if the corresponding loading is negative.In contrast,our method enforces non-negativity in the loadings matrix H.This approach produces results that are consistent with the administrative approach.The CPPI for each ship size is obtained by adding the three columns of

296、 W.The CPPI index is a weighted sum of these indices:Let CPPIi denote the CPPI index for ship size i(i=1,.,5).CPPICPPIiii=15where(1,2,3,4,5)=(0.46,1.00,1.54,1.97,2.57)The aPProach and meThodoloGy|36 Constructing the CPPI 2023 Index Using a Ranking Aggregation MethodThe CPPI has in previous iteration

297、s utilized two distinct methodologies:the administrative,or technical approach that employs expert knowledge and judgment to produce a practical methodology,and a statistical approach that utilizes factor analysis(FA).CPPI 2022 went a step further to aggregate the two rankings to produce one index t

298、hat to present the performance of ports via both methodologies,an approach that is continued in CPPI 2023.Borda-Type Approach for Index AggregationRank aggregation,that is the process of combining multiple rankings into a single ranking,is an important problem arising in many areas(Langville and Mey

299、er 2012).For example,in a ranked voting system,citizens rank candidates in their order of preference and a single winner needs to be determined.Similarly,recommender systems and search engines can produce many different rankings of items that are likely to be of interest to a given user.Such ranking

300、s can naturally be aggregated to produce a more robust list of items(Pappa et al.2020).Many strategies were proposed in the literature to combine several rankings into one that is as consistent as possible with the individual rankings(Langville and Meyer 2012,Fagin et al.2003,Dwork et al.2001,Dwork

301、et al.2012,Oliveira et al.2020)and references therein.The Borda count(Langville and Meyer 2012,Chapter 14)provides a simple and effective approach for aggregating rankings,wherein each item to rank is given points according to the number of items it outranks in its segment.These points are added and

302、 then used to produce a new ranking.Our approach to combine the administrative and the statistical rankings is inspired by the Borda count,but also considers the index values for attributing the number of points.The process is as follows:First,each index is scaled to take values into the interval 0,

303、1.This is accomplished by applying the following linear transformation:f xxMmmMm=(),where m is the minimum value of the index and M the maximum value.Observe that the port with the smallest index is always given a scaled value of 0 and the port with largest index a scaled value of 1.The other ports

304、get a scaled value between 0 and 1.Once the indices are scaled,they are added to produce a combined index.Finally,a ranking is obtained by sorting the ports according to the combined index in decreasing order.Thus,the port with the largest combined index is ranked first and the port with the smalles

305、t combined index is ranked last.TABLE 3.10 An Example of Aggregated Rankings for Four Ports with Randomly Generated Administrative and Statistical Index ValuesPORTSADMINISTRATIVE INDEXSTATISTICAL INDEXSCALED ADMINISTRATIVE INDEXSCALED STATISTICAL INDEXCOMBINED INDEXFINAL RANKINGPort 11.451.971.0001.

306、0002.0001Port 21.261.210.6780.3921.0703Port 31.231.310.6270.4721.0992Port 40.860.720.0000.0000.0004Source:Original table produced for this publication.37|The aPProach and meThodoloGyFor example,the scaled administrative index value of Port 2(x=1.26)is computed as follows:the minimum and maximum valu

307、es of the administrative index are m=0.86 and M=1.45.Thus,the scaled value isf x=()1.261.450.860.861.450.860.678NOTES1 International Maritime Organization(IMO)Resolution MSC.74(69)Annex 3.2 See the International Maritime Organizations website on“International Convention for the Safety of Life at Sea

308、(SOLAS),1974,”(accessed March 2022),at https:/www.imo.org/en/About/Conventions/Pages/International-Convention-for-the-Safety-of-Life-at-Sea-(SOLAS),-1974.aspx.3 International Convention for the Safety of Life at Sea(SOLAS),under the revised SOLAS 1974 Chapter V(as amended)Safety of Navigation,sectio

309、n 19.2.415,carriage requirements for shipborne navigational systems and equipment.4 See ITUs website on“Technical Characteristics for an Automatic Identification System Using Time Division Multiple Access in the VHF Maritime Mobile Frequency Band,”(accessed November 2021),at https:/www.itu.int/dms_p

310、ubrec/itu-r/rec/m/R-REC-M.1371-5-201402-I!PDF-E.pdf.5 It may be a conventional land-based port or a stretch of water designated as an area for transferring cargo or passengers from ship to ship.6 The precise approach to produce a robust data set is detailed in appendix B.7 The actual equation is:(Gr

311、oup Average Port Hours/Example Port Hours)x Call Size Group Weight.ReferencesLangville,Amy N.,and Carl D.Meyer.Whos#1?:the Science of Rating and Ranking.Princeton University Press,2012.Fagin,Ronald,Ravi kumar,and Dakshinamurthi Sivakumar.Comparing Top k lists.SIAM Journal on Discrete Mathematics 17,

312、no.1(2003):134-160.Dwork,Cynthia,Ravi kumar,Moni Naor,and D.Sivakumar.Rank Aggregation Revisited.(2001):613-622.Dwork,Cynthia,Ravi kumar,Moni Naor,and Dandapani Sivakumar.Rank Aggregation Methods for the Web.InProceedings of the 10th International Conference on World Wide Web,pp.613-622.2001.Ali,Aln

313、ur,and Marina Meil.Experiments with kemeny Ranking:What Works When?Mathematical Social Sciences 64,no.1(2012):28-40.Oliveira,Samuel EL,Victor Diniz,Anisio Lacerda,Luiz Merschmanm,and Gisele L.Pappa.Is Rank Aggregation Effective in Recommender Systems?An Experimental Analysis.ACM Transactions on Inte

314、lligent Systems and Technology(TIST)11,no.2(2020):1-26.IALA(International Association of Marine Aids to Navigation and Lighthouse Authorities).2005.IALA Guideline 1050:The Management and Monitoring of AIS information.Edition 1.0.Saint Germain:IALA.https:/www.iala-aism.org/product/management-and-moni

315、toring-of-ais-information-1050/?download=true.IALA(International Association of Marine Aids to Navigation and Lighthouse Authorities).2016.IALA Guideline 1082:An Overview of AIS.Edition 2.0.Saint Germain:IALA.19.https:/www.iala-aism.org/product/an-overview-of-ais-1082/?download=true.The conTainer Po

316、rT Performance index 2023|38 4.The container Port Performance index 2023IntroductionThe rankings of container port performance,based on the ranking aggregation approach,are presented in this chapter.The following section presents the rankings for the top 100 best performing container ports,with the

317、full rankings of all ports by both approaches presented in Appendix A.The subsequent sections present a summary by region and port throughput(large,medium,small),so ports in the same region,or with the same throughput within broad categories,can be easily compared.The CPPI 2023Table 4.1 presents the

318、 top 100 in the rankings of container port performance in the CPPI 2023.It reflects the aggregation of the scores from the results from the administrative approach and the statistical approach in the manner described in the previous section.In the aggregate index,the two top-ranked container ports i

319、n the CPPI 2023 are Yangshan Port(China)in first place,followed by the Port of Salalah(Oman)in second place.These two ports occupy the same positions in the rankings generated by the constituent approaches.The Port of Salalah was ranked second in both approaches in CPPI 2021,while the Yangshan Port

320、ranked third and fourth in the statistical and administrative approaches,respectively,in CPPI 2021.39|The conTainer PorT Performance index 2023The top-ranked container ports in the CPPI 2023 are Yangshan Port(China)in first place,followed by the Port of Salalah(Oman)in second place,retaining their r

321、anking from the CPPI 2022.Third place in the CPPI 2023 is occupied by the port of Cartagena,up from 5th place in the CPPI 2022,whilst Tangier-Mediterranean retains it 4th ranking.Tanjung Pelepas improved one position to 5th,Ningbo moved up from 12th in 2022 to 7th in 2023,and Port Said moved from 16

322、th to 10th in 2023.Yokohama fell from 10th and 12th in CPPI 2021 to 15th place in CPPI 2022 is now back to 9th in 2023.Ports moving in the other direction in the top ten:khalifa port falls from 3rd position in 2022 to 29th position in CPPI 2023.Hamad Port which fell from 8th in 2022 to 11th in 2023.

323、There are 55 new entrants to the CPPI 2023,and several significant gainers in terms of ranking.TABLE 4.1 The CPPI 2023PORT NAMEOVERALL RANKINGYANGSHAN1SALALAH2CARTAGENA(COLOMBIA)3TANGER-MEDITERRANEAN4TANJUNG PELEPAS5CHIWAN6CAI MEP7GUANGZHOU8YOkOHAMA9ALGECIRAS10HAMAD PORT11NINGBO12MAWAN13DALIAN14HONG

324、 kONG15PORT SAID16SINGAPORE17kAOHSIUNG18VISAkHAPATNAM19YEOSU20TIANJIN21YANTIAN22TANJUNG PRIOk23LIANYUNGANG24SHEkOU25CALLAO26MUNDRA27PORT kLANG28kHALIFA PORT29kING ABDULLAH PORT30 xIAMEN31BUSAN32PORT NAMEOVERALL RANKINGGEMLIk33BARCELONA34DAMMAM35SAVONA-VADO36POSORJA37FUZHOU38ZEEBRUGGE39COLOMBO40PIPAV

325、AV41RIO DE JANEIRO42kHALIFA BIN SALMAN43BUENAVENTURA44LAEM CHABANG45SHIMIZU46kAMARAJAR47INCHEON48JEBEL ALI49LAZARO CARDENAS50AARHUS51DA CHAN BAY TERMINAL ONE52CHARLESTON53TOkYO54PHILADELPHIA55NAGOYA56kATTUPALLI57JEDDAH58JUBAIL59QINZHOU60kARACHI61kEELUNG62COCHIN63kOBE64The conTainer PorT Performance

326、index 2023|40 PORT NAMEOVERALL RANKINGPORT EVERGLADES65SOHAR66SALVADOR67HAZIRA68LONDON69HAIPHONG70kRISHNAPATNAM71WILHELMSHAVEN72BEIRUT73MIAMI74BOSTON(USA)75ANTWERP76DILISkELESI77ITAPOA78PUERTO LIMON79CHENNAI80WILMINGTON(USA-N CAROLINA)81MARSAxLOkk82PORT NAMEOVERALL RANKINGZHOUSHAN83SOUTHAMPTON84OSAk

327、A85HAIFA86AQABA87BREMERHAVEN88SANTA CRUZ DE TENERIFE89MALAGA90ROTTERDAM91NEW YORk&NEW JERSEY92JOHOR93POINTE-A-PITRE94YOkkAICHI95JAWAHARLAL NEHRU PORT96CORONEL97TRIPOLI(LEBANON)98JACkSONVILLE99ALTAMIRA100Source:Original table produced for this publication,based on CPPI 2023 data.The CPPI 2023 shows a

328、 great consistency between the two approaches,as in its 2022 edition.In CPPI 2023,more than 40 percent of all ports(162 ports)are ranked within 6 places or less from themselves in the dual rankings,whereas 50 percent of the ports are ranked within 8 places.The consistency between the two approaches

329、contributes significantly to having a well-balanced aggregated index.Ranking by RegionThis section presents an overview of the outcomes from the CPPI 2023 report.The first edition of CPPI was modified based on requests for the presentation of results and rankings by region and throughput for an impr

330、oved comparison of ports within the same region and those with similar throughput.The subsequent sections include a concise tabulation of the results and ranking(from Table 4.2)for the designated regions based on the administrative CPPI.North America(United States and Canada)Central America,South Am

331、erica,and the Caribbean Region West,Central,and South Asia(Saudi Arabia to Bangladesh)East Asia(Myanmar to Japan)Oceania(Australia,New Zealand,and the Pacific Islands)Sub-Saharan Africa Europe and North Africa41|The conTainer PorT Performance index 2023TABLE 4.2 The CPPI by Region:North AmericaPORT

332、NAMEREGIONOVERALL RANKINGPHILADELPHIANAM50CHARLESTONNAM60PORT EVERGLADESNAM63WILMINGTON(USA-N CAROLINA)NAM72BOSTON(USA)NAM73MIAMINAM77JACkSONVILLENAM83HALIFAxNAM95NEW YORk&NEW JERSEYNAM99NEW ORLEANSNAM133MOBILENAM186BALTIMORE(USA)NAM191PORT TAMPA BAYNAM214HONOLULUNAM219APRA HARBORNAM223SAINT JOHNNAM

333、265HUENEMENAM277PORT OF VIRGINIANAM306HOUSTONNAM327MONTREALNAM351SEATTLENAM356VANCOUVER(CANADA)NAM363LONG BEACHNAM376LOS ANGELESNAM378OAkLANDNAM396PRINCE RUPERTNAM397SAVANNAHNAM398TACOMANAM402Source:Original table produced for this publication,based on CPPI 2023 data.TABLE 4.3 The CPPI by Region:Central America,South America,and the Caribbean RegionPORT NAMEREGIONOVERALL RANKINGCARTAGENA(COLOMBIA)