Bilateral Liner Shipping Connectivity Since 2006

AT BILATERAL LINER SHIPPING CONNECTIVITY SINCE 2006 POLICY ISSUES IN INTERNATIONAL TRADE AND COMMODITIES RESEARCH STUDY SERIES . 72 York Geneva, 2016 AT ii POLICY ISSUES IN INTERNATIONAL TRADE AND COMMODITIES purpose studies Research Study Series analyse policy issues stimulate discussions area international trade development. Series includes studies UNCTAD staff distinguished researchers organizations academia. opinions expressed research study authors official views UNCTAD secretariat member States. studies published Research Study Series read anonymously referee. Comments referees account publication studies. designations employed presentation material imply expression opinion part United Nations legal status country, territory, city area, authorities delimitation frontiers boundaries. Comments paper invited addressed author, / Publications Assistant, Trade Analysis Branch (TAB), Division International Trade Goods Services, Commodities (DITC), United Nations Conference Trade Development (UNCTAD), Palais des Nations, CH-1211 Geneva 10, Switzerland; -mail: tab@unctad.org; fax : +41 22 917 0044. Copies studies Research Study Series obtained address. Studies Research Study Series UNCTAD website http://unctad.org/tab. Series Editor: Chief Trade Analysis Branch DITC/UNCTAD UNCTAD/ITCD/TAB/74 UNITED NATIONS PUBLICATION ISSN 1607-8291 © Copyright United Nations 2016 rights reserved Bilateral Liner Shipping Connectivity 2006 iii paper presents unique analysis bilateral liner shipping connectivity evolution 9 years period 2006 2014. relevance analysis stems original contributions. , paper proposes unique bilateral liner shipping connectivity index based components capturing quality liner shipping connection countries direct liner service exists . , shows evolution bilateral liner shipping connectivity time qualifies contribution component index. Results show top 50 LSBCIs connections maximum 15 countries top 250 LSBCIs connections maximum 40 countries. highest LSBCI values obtained intra-regional routes, notably intra-Europa intra-Asia. LSBCI predominantly driven transhipment options pointing crucial importance centrality liner shipping network. Remote countries highly dependent centrality countries connected , consequence extremely vulnerable variation global set direct connections. Keywords: Liner shipping bilateral connectivity, container shipping networks JEL Classification: R49 iv POLICY ISSUES IN INTERNATIONAL TRADE AND COMMODITIES paper outcome joint project Maritime Connectivity UNCTAD' Division International Trade Goods Services, Commodities UNCTAD' Division Trade Logistics. context grateful Anne Miroux Guillermo Valles support. grateful Bismark Sitorus helping collection treatment raw data. Part results presented Trade Development Commission, seventh session held Geneva, 18-22 2015. participants interest comments. paper represents personal views authors , views UNCTAD secretariat member States. authors accept sole responsibility errors remaining. Bilateral Liner Shipping Connectivity 2006 1 INTRODUCTION .......................................................................................................................... 1 2 THE COMPONENTS OF THE PROPOSED LSBCI .................................................................... 2 2.1 Transhipments ................................................................................................................... 3 2.2 Common direct connections ............................................................................................. 5 2.3 Geometric number direct connections ............................................... 7 2.4 level competition shipping services ................................................................... 8 2.5 Ships size ......................................................................................................................... 11 3 THE LSBCI ................................................................................................................................. 12 3.1 Components normalization aggregation................................................................... 12 3.2 Inter-temporal comparison .............................................................................................. 13 3.3 Rankings .......................................................................................................................... 15 3.4 LSBCI components ....................................................................................... 17 3.5 Transhipments LSBCI ......................................................................................... 20 3.6 LSBCI maritime distance ................................................................................... 21 4 DISCUSSION ............................................................................................................................. 22 REFERENCES ......................................................................................................................................... 24 vi POLICY ISSUES IN INTERNATIONAL TRADE AND COMMODITIES List figures Figure 1. Share connections transhipments ....................................................... 4 Figure 2. Centrality 2006 centrality 2014 ........................................................................... 5 Figure 3. Kernel density Common direct Connections 2006 2013 ............................................ 6 Figure 4. Kernel density Direct Connections Geom. 2006 2014 ..................................... 8 Figure 5. Kernel density Number carriers operating competitive leg 2006 2014 ... 10 Figure 6. Kernel density size largest ship operating leg smallest ships 2006 2013 .............................................................................. 12 Figure 7. LSBCI distribution (Kernel density estimation) 2006 2014 ....................................... 14 Figure 8. LSBCI variation 2006 2014 ............................................................................ 15 Figure 9. LSBCI variation 2006 2010 , 2010 2014 ............................ 15 Figure 10. Direct Connections: Net creation 2008-2014 ...................................................................... 21 List tables Table 1. Number transhipments connect country pairs ............................................ 3 Table 2. Number common direct connections ................................................................................ 6 Table 3. Number direct connections: Geometric .................................................................. 7 Table 4. Largest number carriers operating competitive leg ....................................... 9 Table 5. Size TEU largest ship operating weakest leg ............................................. 11 Table 6. LSBCI: selected descriptive statistics ................................................................................. 13 Table 7. Top 20 country pairs 2006, 2010 2014 ..................................................................... 16 Table 8. Bottom 20 2006, 2010 2014 ..................................................................................... 17 Table 9. LSBCI components (raw): pairwise correlations 2014 .................................. 18 Table 10. Components (normalized) shares LSBCI ......................................................................... 18 Table 11. Decomposition LSBCI: positive negative variation 2006-2014 ......... 18 Table 12. Relative components' raw values ..................................................................... 19 Table 13. Variation number transhipments 2006-2014 ......................................................... 20 Table 14. Maritime distance LSBCI: pairwise correlations .................................................... 21 Table 15. LSBCI components: country' top 10 destinations .............................................. 23 Bilateral Liner Shipping Connectivity 2006 1 1. INTRODUCTION Maritime transport core international trade merchandises. 80% volume goods exchanged world transported sea (UNCTAD, 2008). predominance maritime transport explained large extent exponential intensification containerized transport services. decades, share general cargo containerised steadily grew -thirds total general cargo transport. terms , containerised general cargo exceeds 90% general cargo. Containerization links manufacturer producer ultimate consumer customer individual trade transaction economically justify chartering ship. network regular container shipping services transhipment operations -called hub ports, basically countries today connected . , growing participation developing countries seaborne trade,1 evidence maritime connections suggests , China, reached full potential. literature emphasized importance transportation costs infrastructure explaining trade access international markets.2 burgeoning strand trade literature attempted assess impact maritime connectivity bilateral exports concludes significantly large.3 , empirical assessments related maritime connectivity based single dimension indicators, existence direct maritime connection, bilateral indicators connectivity constructed unilateral measures lacking true bilateral nature. Hoffmann al (2014) propose bilateral index liner shipping connectivity, Liner Shipping Bilateral Connectivity Index (LSBCI). LSBCI extension UNCTAD’ country- level Liner Shipping Connectivity Index (LSCI) computed 2004 based proper bilateralization transformation. paper presents revised version LSBCI. index computed sample 155 costal countries 9 years period 2006 2014. exercise unique. comprehensive dealing bilateral liner shipping connectivity consecutive years. Results show top 50 LSBCIs connections maximum 15 countries top 250 LSBCIs connections maximum 40 countries. highest LSBCI values obtained intra-regional routes, notably intra-Europa intra-Asia. LSBCI predominantly driven transhipment options pointing crucial importance centrality liner shipping network. Remote countries highly dependent centrality countries connected , consequence extremely vulnerable variation global set direct connections. Trade impact relevance constructing bilateral index liner shipping connectivity monitor qualify evolution time strict empirical considerations. Efficient transport services contribute economic development, quality transport 1 1970 2010, developing countries´ share volume seaborne exports rose 18 cent 56 cent world´ total (UNCTAD, 2013). 2 instance, based estimation gravity model US data, Anderson van Wincoop (2003) transport costs correspond average ad valorem tax equivalent 21%. similar empirical approach, (Clark al., 2004) estimates reveal Latin American countries, transport costs greater barrier .. markets import tariffs. Arvis al. (2013) results obtained sample 178 countries 1995-2010 period maritime transport connectivity logistics performance important determinants bilateral trade costs. Fugazza, (2015) gravity model approach based dataset maritime connections sample 178 countries collected 2006-2012 period finds absence direct connection drop exports varying 42 55 cent. 2 POLICY ISSUES IN INTERNATIONAL TRADE AND COMMODITIES services markets terms service levels crucial element competitiveness regions countries globalised market place. LSBCI monitoring instrument benchmark policy making. “perception” indexes purposes, data generated polls experts obtain picture business countries, data fleet deployment hard data, vary asked, inflation rates changing perceptions. rest paper organized . section discusses components revised version LSBCI presents descriptive statistics. Stylized facts revised LSBCI commented Section 3. Section 4 dedicated illustration revised LSBCI assessing liner shipping experience specific country. Section 5 concludes. 2. THE COMPONENTS OF THE PROPOSED LSBCI LSBCI meant reflect specifically liner shipping connectivity pairs countries. context aspects connectivity distance excluded set components retained computation index. Distance countries, level connectivity individual countries relevant bilateral trade trade costs. , bilateral liner shipping connectivity , aim capturing stand- factor. LSBCI includes 5 components. pair countries represented sample, LSBCI based : 1) number transhipments required country country ; 2) number direct connections common country country ; 3) geometric number direct connections country country ; 4) level competition services connect country country ; 5) size largest ships weakest route connecting country country . components country-pair specific components 4 based country specific characteristic. components symmetric. instance number transhipments move container number transhipments move container . consequence, quality liner shipping connectivity country country identical quality liner shipping connectivity country country . general point view, , selecting components aimed attach weight connectivity (components 1, 2 3) "intensity" (components 4 5). essentially motivated strong impact freight costs transhipment close relationship incidence direct connections average centrality countries liner shipping network. sections briefly discuss rationale inclusion component present stylized facts. based sample 155 coastal countries (11935 country pairs) connectivity informed year period running 2006 2014. year 2007 reported due absence observations. underlying raw data obtained Lloyd’ List Intelligence, “Lloyd’ List Intelligence - Containers”. 4 instance, Wilmsmeier Hoffmann (2008) findings based sample 189 freight rates company Caribbean show trade routes indirect services (.. including transhipments) induce higher transport costs. estimates suggest transhipment equivalent impact freight rates increase distance countries 2,612 km. 5 Formally Containerisation International -line. Bilateral Liner Shipping Connectivity 2006 3 2.1 TRANSHIPMENTS component proposed LSBCI theoretical minimum number transhipments required country country . transhipment implies additional costs, time risks delays damage, LSBCI country-pair direct service higher country-pair connected direct service. small part country pairs connected , majority country pairs require transhipment order transport container . form Table 1, 2014 18 cent country pairs represented sample connected, 64 cent required transhipment 16 cent transhipments. Figures 2006 20 cent, 67 cent 12 cent . years country pairs required transhipments. country pairs require 3 transhipments – theory. share direct connections remained stable period consideration slight drop observed 2014 work 2012. spill- effect direct -direct connections definition depends . intensity effect determined direct connections constituting primary network point time. words, depending direct connections created destructed time, repercussion vary dramatically. instance, years 2006 2012 characterized precisely identical share direct connections. , share connections require single transhipment share connections require transhipments significantly . obtain dispersion measured standard deviation increases number transhipments required connect countries. analytical elements illustrate existence network externalities liner shipping company decision removing creating direct connection . kind externalities corroborated country specific experiences. instance Lithuania case observe constant group direct connections, varying 10 11 period, relative incidence single double transhipments changed dramatically year year. number connections transhipment varied 96 112 transhipments 31 47. Table 1 Number transhipments connect country pairs (shares %) Number Transhipments 2006 2008 2009 2010 2011 2012 2013 2014 0 20.05 21.08 20.29 20.82 20.26 20.05 19.6 17.69 1 66.98 67.25 64.2 64.43 63.65 64.49 64.23 63.2 2 12.81 11.66 14.93 14.68 16.02 15.4 16.09 18.98 3 0.16 0.01 0.58 0.06 0.07 0.06 0.07 0.13 Note: Statistics obtained sample 155 coastal countries 11935 country pairs. Source: Authors calculations. emerge Table 1 illustrated Figure 1 terms longer-run tendency drop direct connections translated essentially rising incidence connections necessitating transhipments. 2006 13 cent country pairs required transhipments connected. figure 2014 19 cent. 4 POLICY ISSUES IN INTERNATIONAL TRADE AND COMMODITIES Figure 1: Share connections transhipments Source: Authors calculations. observation variability terms connectivity options direct connection service cancelled observation increasing incidence - transhipment connections synonymous higher uncertainty higher costs exporting firms. Figure 2 shows relationship indicator centrality computed country sample 2006 (horizontal axis) 2014 (vertical axis). indicator ratio observed number direct connections maximum observable number , 154. diagonal corresponds 45-degree line. observation lying line represents country number direct connections increased 2006 2014. reverse true observations 45-degree line. performer appears Morocco Russian Federation Democratic Republic Congo. countries centrality deteriorated economically important countries Germany, France, Great Britain, Spain, Japan Brazil. reflect long lasting effect 2008 financial crisis rationalization network direct connections consequence crisis . Bilateral Liner Shipping Connectivity 2006 5 Figure 2: Centrality 2006 centrality 2014 Source: Authors calculations. 2.2 COMMON DIRECT CONNECTIONS component number common direct connections countries country pair. total number countries direct connection , origin country destination country pair. equivalent theoretical number options shipper faces goods shipped single transhipment. economic rationale component twofold. , countries lie coast (.. Chile, Peru, Ecuador) served services consequence connections (services connect ) common services / .. Europe, North America Asia. Empirically, shown countries lie ocean tend trade -. , common connection option trade single transhipment. common connections countries (.. Brazil Ecuador Panama, Jamaica Bahamas), countries connected trade -. Table 2 reports series basic descriptive statistics years investigation. case results obtained previous LSBCI component, indications long lasting influence effects 2008 financial crisis subsequent rationalization liner shipping network. instance, maximum dropped dramatically 2014 compared 2006 quasi-erratic path period. P90 threshold indicating number common connections country pair part top 10 connected pairs moved 22 2010 20 2014, stands 2006 threshold. Figure 3 illustrates kernel density function component years 2006 2014. shape -parameterized distribution suggests country pairs characterized small number ( 10) common direct connections. Comparing panels Figure 3 distribution shifted . larger share bilateral relationships characterized larger number common direct connections. time, , notice share country pairs large number common direct connections fallen. elements equivalent core liner shipping network concentrated 2014 compared 2006. words countries connected hubs. consistent previous observations larger number country pairs connected 2 transhipments. 6 POLICY ISSUES IN INTERNATIONAL TRADE AND COMMODITIES Table 2 Number common direct connections Median Standard Deviation Maximum p10 p90 2006 8.33 5.00 9.83 92.00 0.00 21.00 2008 8.85 5.00 10.49 95.00 0.00 22.00 2009 8.07 5.00 9.94 87.00 0.00 21.00 2010 8.57 5.00 10.33 90.00 0.00 22.00 2011 8.23 5.00 10.14 80.00 0.00 21.00 2012 8.30 5.00 10.20 87.00 0.00 21.00 2013 7.95 4.00 9.95 85.00 0.00 20.00 2014 7.92 4.00 9.99 81.00 0.00 20.00 Note: Statistics obtained sample 155 coastal countries 11935 country pairs. Source: Authors calculations. Figure 3: Kernel density Common direct Connections 2006 (upper graph) 2013 ( graph) Source: Authors computations. 0 .0 2 .0 4 .0 6 .0 8 .1 en si ty 0 20 40 60 80 100 Number Common direct Connections Kernel density estimate Normal density kernel = epanechnikov, bandwidth = 0.8908 Kernel density estimate 0 .0 5 .1 .1 5 en si ty 0 20 40 60 80 Number Common direct Connections Kernel density estimate Normal density kernel = epanechnikov, bandwidth = 0.8885 Kernel density estimate Bilateral Liner Shipping Connectivity 2006 7 2.3 THE GEOMETRIC MEAN OF THE NUMBER OF DIRECT CONNECTIONS opted inclusion measure reflect degree centrality country pair based composing country. reason twofold. , country pairs ranking obtained component proposed Hoffman al. (2014) appeared odd components rankings. , fact component bilateral essence, centrality country pair network liner shipping connections expected significantly affected centrality network country separately. , component indicator access network trade partner. context measure bilateral. geometric balanced measure bilateral access rest world. Basic descriptive statistics reported Table 3. strong link component 1 statistics. median values falling years. consequence fall share connected country pairs. Table 3 Number direct connections: Geometric Median Standard Deviation Maximum p10 p90 2006 24.12 21.35 14.73 102 8.06 43.59 2008 24.79 21.91 15.35 105 8.00 45.37 2009 23.89 21.07 14.57 98 7.94 43.27 2010 24.87 21.82 14.80 99 9.06 44.72 2011 25.08 22.25 14.82 96 9.00 44.59 2012 25.30 22.45 15.00 101 9.00 45.17 2013 24.54 21.45 14.59 100 9.17 44.00 2014 24.69 21.56 14.62 101 9.17 44.19 Note: Statistics obtained sample 155 coastal countries 11935 country pairs. Source: Authors calculations. kernel density functions component year 2006 2013 shown Figure 4. dramatic change identified. , easily distinguish appearance small bump 40. consistence tendency move network concentrated destinations. 8 POLICY ISSUES IN INTERNATIONAL TRADE AND COMMODITIES Figure 4: Kernel density Direct Connections Geom. 2006 (upper graph) 2014 ( graph) Source: Authors computations. 2.4 THE LEVEL OF COMPETITION ON SHIPPING SERVICES fourth component level competition services connect country pairs. constraining number carriers operate route country pair. higher number , higher competition. competition shipping route increased, shipping lines incentive reduce transportation costs margins routes (Hummels al., 2009, Asturias Petty, 2013) leading turn decrease transportation costs shippers route. theory, hundreds theoretical options connect countries. generation component, computed Max-Min number companies “” connection countries terms number companies. , 5 companies competing route - 8 companies competing route -, competition thinnest route option 5. option , 6 companies competing route - 7 companies competing route -, competition thinnest route option 6. Comparing options, Max-Min (.. highest number leg lowest number) 6, “6” incorporated LSBCI component. 0 .0 1 .0 2 .0 3 en si ty 0 20 40 60 80 100 Geometric Number direct Connections Kernel density estimate Normal density kernel = epanechnikov, bandwidth = 1.6039 Kernel density estimate 0 .0 1 .0 2 .0 3 .0 4 en si ty 0 20 40 60 80 100 Geometric Number direct Connections Kernel density estimate Normal density kernel = epanechnikov, bandwidth = 1.5811 Kernel density estimate Bilateral Liner Shipping Connectivity 2006 9 Descriptive statistics reported Table 4. average, number carriers operating maritime routes increased. 4 2006 5 2014. median 2014 median 2006. Fifty cent maritime routes served 4 carriers. time maximum number carriers operating single route diminished fourth 2006 (82 carriers) 2014 (63 carriers). 2006 2008 country pair showing maximum 82 Great Britain- Netherlands. 2009 dramatic drop maximum 58 carriers observed country pair. remained top list Belgium- Netherlands country pair. 2010, maximum number carriers operating Eastern Asian routes. Malaysia-Singapore country pair dominated years 2010 2011 58 servicing. 2012, China-South-Korea country pair occupied rank. , 72 carriers operated 2012 63 2014. high variability statistic reflection highly fluctuating trade flows strictly related fluctuating demand conditions aftermath 2008 financial crisis. Table 4 Largest number carriers operating competitive leg Median Standard Deviation Maximum p10 p90 2006 3.73 2.00 4.40 82.00 1.00 7.00 2008 3.70 2.00 4.26 82.00 1.00 8.00 2009 3.37 2.00 3.73 58.00 1.00 7.00 2010 3.35 2.00 3.82 58.00 1.00 7.00 2011 3.43 2.00 3.84 57.00 1.00 7.00 2012 4.82 3.00 5.09 72.00 1.00 10.00 2013 4.86 4.00 4.97 67.00 1.00 10.00 2014 4.90 4.00 5.00 63.00 1.00 10.00 Note: Statistics obtained sample 155 coastal countries 11935 country pairs. Source: Authors calculations. kernel density function component illustrated Figure 5. shape corroborate previous remarks. Due collapse maximum observed 2014, distribution inflated lowest values consequence dispersed 2006. effects similar redeployment carriers historical destinations strong demand containerized goods. accordance interpretation component, competition liner shipping companies increased. , redeployment occurred, results suggest ship companies focusing routes. necessarily synonymous intensified competition. 10 POLICY ISSUES IN INTERNATIONAL TRADE AND COMMODITIES Figure 5: Kernel density Number carriers operating competitive leg 2006 (upper graph) 2014 ( graph) Source: Authors computations. 0 .1 .2 .3 en si ty 0 20 40 60 80 Max-Min number carriers route Kernel density estimate Normal density kernel = epanechnikov, bandwidth = 0.2672 Kernel density estimate 0 .0 5 .1 .1 5 en si ty 0 20 40 60 Max-Min number carriers route Kernel density estimate Normal density kernel = epanechnikov, bandwidth = 0.3554 Kernel density estimate Bilateral Liner Shipping Connectivity 2006 11 2.5 SHIPS SIZE component size (expressed twenty-foot equivalent unit, TEU) largest ship thinnest route. Calculations based approach fourth component. maximum ship size considered indication level infrastructure trading countries countries transhipment occurs. vessel size indicator economies scale sea-leg. connections direct link ship deployed direct link option largest ships. , direct services west coast South America Europe deploy smaller ships deployed services Panama, Panama Europe; put differently, LSBCI include larger vessel size direct service. Descriptive statistics reported Table 5. average size increased significantly steadily 2006-2014 period. case maximum size. driven suggested inspection Figure 6 shows kernel density component. Table 5 Size TEU largest ship operating weakest leg Median Standard Deviation Maximum p10 p90 2006 1794.73 1344.00 1596.83 9742.00 485.00 3584.00 2008 1953.18 1438.00 1744.26 12508.00 519.00 4250.00 2009 2252.72 1606.00 2095.82 14770.00 519.00 5050.00 2010 2252.75 1604.00 2165.17 14770.00 518.00 4990.00 2011 2293.93 1454.00 2261.95 15550.00 518.00 5100.00 2012 2484.09 1510.00 2597.19 15550.00 500.00 5762.00 2013 2551.77 1604.00 2647.04 16020.00 600.00 6350.00 2014 2664.75 1638.00 2867.98 18270.00 418.00 6539.00 Note: Statistics obtained sample 155 coastal countries 11935 country pairs. Source: Authors calculations. Kernel density reveals intensified ships size 5000 TEU 10000 TEU 2014 compared 2006. period investigation observe intensification super-sized ships. 2006, largest existing ship sail route . 2012 2013 number routes jumped 55 size ship doubled respect 2006. years, size largest ship moved 15550 TEU 18270 TEU. change explain largest ships sail 45 routes 2014 port infrastructures time adjust size requirements. 12 POLICY ISSUES IN INTERNATIONAL TRADE AND COMMODITIES Figure 6: Kernel density size largest ship operating leg smallest ships 2006 (upper graph) 2013 ( graph) Source: Authors calculations. 3. THE LSBCI section discusses computation synthetic bilateral index. stylized facts presented. 3.1 COMPONENTS NORMALIZATION AND AGGREGATION order establish unit free index components normalized standard formula: Normalized_Value=(Raw-Min(Raw))/(Max(Raw)-Min(Raw)). opted formula (Raw/Max(Raw)) formula essentially existence minimum values differ . minimum values components formulas equivalent generate identical normalized values. 0 .0 00 1 .0 00 2 .0 00 3 .0 00 4 .0 00 5 en si ty 0 2000 4000 6000 8000 10000 Max-Min largest ship size route Kernel density estimate Normal density kernel = epanechnikov, bandwidth = 122.4872 Kernel density estimate 0 .0 00 1 .0 00 2 .0 00 3 .0 00 4 .0 00 5 en si ty 0 5000 10000 15000 20000 Max-Min largest ship size route Kernel density estimate Normal density kernel = epanechnikov, bandwidth = 152.9118 Kernel density estimate Bilateral Liner Shipping Connectivity 2006 13 LSBCI computed simple average normalized components. consequence, LSBCI values 0 1. component, simply complement unity 1-Normalized_Value respect correspondence higher values stronger connectivity. 3.2 INTER-TEMPORAL COMPARISON order index values comparable time, maximum minimum values component correspond maximum minimum observed time period consideration. identical maximum minimum values years direct comparison time countries. track evolution specific country score time comparison evolution countries' scores. Table 6 reports standards descriptive statistics characterizing sample. median values show variation time. Dispersion measured standard deviation increased slightly. Table 6 LSBCI: selected descriptive statistics Median Standard Deviation Maximum p10 p90 2006 0.225 0.209 0.097 0.867 0.103 0.350 2008 0.230 0.212 0.100 0.875 0.102 0.359 2009 0.226 0.209 0.102 0.860 0.103 0.358 2010 0.230 0.211 0.104 0.845 0.105 0.362 2011 0.230 0.210 0.104 0.847 0.104 0.362 2012 0.236 0.214 0.108 0.866 0.110 0.375 2013 0.234 0.211 0.108 0.877 0.108 0.372 2014 0.235 0.213 0.110 0.863 0.109 0.374 Note: Statistics obtained sample 155 coastal countries 11935 country pairs. Source: Authors calculations fact reflected kernel probability density function shape observed 2006 2014. functions reported upper panel Figure 7 . 14 POLICY ISSUES IN INTERNATIONAL TRADE AND COMMODITIES Figure 7: LSBCI distribution (Kernel density estimation) 2006 (upper panel) 2014 ( panel) Note: dashed vertical segment , plain segment median Source: Authors calculations. Figure 8 scatters LSBCI country pair 2014 LSBCI 2006. Points 45-degree line represent country pairs LSBCI increased 2006 2014. Points 45-degree line represent country pairs LSBCI decreased. 2006 2014. majority country pairs, 67 cent, moved terms LSBCI performance. 0 2 4 6 8 en si ty 0 .2 .4 .6 .8 LSBCI kernel = epanechnikov, bandwidth = 0.0087 Kernel density estimate 0 2 4 6 en si ty 0 .2 .4 .6 .8 LSBCI kernel = epanechnikov, bandwidth = 0.0100 Kernel density estimate Bilateral Liner Shipping Connectivity 2006 15 Figure 8: LSBCI variation 2006 2014 Source: Authors calculations. decomposition variation observed 2006 2014 represented figure 9 suggest improvement occurred 2010. precise analysis LSBCI stagnated large majority country pairs aftermath 2008 world crisis 2010. Figure 9: LSBCI variation 2006 2010 , 2010 2014 Source: Authors calculations. 3.3 RANKINGS tables report top (Table 7) bottom (Table 8) twenty country pairs LSBCI. Rankings based sample 155 costal countries 11935 country pairs. Table 7 reveals developed countries essentially European countries Eastern Asian countries part top 20 country pairs. , presence marked 2014 2010 2006. top twenty country pairs constituted 16 POLICY ISSUES IN INTERNATIONAL TRADE AND COMMODITIES Eastern Asian countries country pair involves China. China-Hon Kong (Province China) pair top list 2010. moved fourth rank 2014 small progression LSBCI . Table 7 Top 20 country pairs 2006, 2010 2014 Deeper analysis reveals 2006 top 50 connections 15 countries, 9 developed (Belgium, France, Germany, Great Britain, Netherlands, Spain, Italy, Japan USA) 6 Asian (China, Taiwan (Province China), Hong Kong (Province China), Malaysia, South Korea Singapore). top 100 connections 17 countries ( previous 15 Canada Egypt) top 250 connections 36 countries. group includes Latin American countries (Argentina Mexico) Caribbean countries ( Dominican Republic Jamaica) Saharan African country, South-Africa. 2010 group making top 50 connections 2006. top 100 connections 21 countries ( previous 15 Egypt, Morocco, Panama, Portugal, Saudi Arabia United Arab Emirates). top 250 connection 33 countries. additional Latin American country (Mexico), Caribbean (Jamaica) South Africa present group. 2014, Japan exits top 50 group 100 group 18 countries . top 250 group expands include 39 countries. Bottom 20 country pairs composed essentially small remote islands (.. Cook Islands Montserrat, Nauru) poor developing countries weak centrality index. presence Latvia Albania list reflects essentially fact centrality Bilateral Liner Shipping Connectivity 2006 17 network weak remote geographical situation. poor performance terms centrality close link important hub Italy Albania Germany Latvia. Table 8 Bottom 20 2006, 2010 2014 year exporter importer LSBCI year exporter importer LSBCI year exporter importer LSBCI 2014 NRU MMR 0.07 2010 MDV COK 0.08 2006 MSR MHL 0.02 NRU MNE 0.07 MDV BMU 0.08 YEM MSR 0.02 MNE BMU 0.07 COK BMU 0.08 COK COD 0.01 NRU BMU 0.07 NRU MMR 0.07 SYC MSR 0.01 GEO COG 0.03 NRU COK 0.07 SVN MSR 0.01 COG BGR 0.03 NRU ALB 0.07 SOM MSR 0.01 COM COK 0.02 NRU BMU 0.07 MSR COD 0.01 MNE COG 0.02 SYC COK 0.01 SDN MSR 0.01 SLE COK 0.02 COK BGR 0.01 MSR KHM 0.01 COK BGR 0.02 SYC NRU 0.01 PLW MSR 0.01 GEO COK 0.02 COK COD 0.01 MSR BGD 0.01 LVA COK 0.02 NRU COD 0.01 MSR MDV 0.01 IRN COK 0.02 GEO COK 0.01 MSR BRN 0.01 COK COG 0.02 COK BHR 0.01 MSR KWT 0.01 IRQ COK 0.02 IRQ COK 0.01 MSR IRQ 0.01 COK ALB 0.01 SOM COK 0.01 MSR BHR 0.01 SOM COK 0.01 ERI COK 0.01 MSR COK 0.01 NRU COG 0.01 QAT COK 0.01 MSR MMR 0.01 ERI COK 0.01 NRU IRQ 0.01 NRU COD 0.01 MNE COK 0.01 COK ALB 0.01 NRU MSR 0.01 Source: Authors Calculations. 3.4 THE LSBCI AND ITS COMPONENTS Table 9 reports standard coefficients pairwise correlation LSBCI components raw form year 2014. coefficients significant 1 cent. Results obtained previous years similar. negative sign component due fact number transhipments, meaning implies weaker connectivity. strongest correlation component 2, number direct common connections countries share. weakest, absolute , component 4 reflecting level competition services weakest leg maritime route linking countries. 18 POLICY ISSUES IN INTERNATIONAL TRADE AND COMMODITIES Table 9 LSBCI components (raw): pairwise correlations 2014 Table 10 reports relative contribution LSBCI , expressed share, component. corresponds ratio normalized component' index . predominance component number transhipments connect countries consequence distribution underlying variable. values change unit raw variable translates change normalized 0.33. , country pairs normalized 0 (3 transhipments), de facto minimum 0.33. Table 10 Components (normalized)' shares LSBCI Year 1 2 3 4 5 2006 0.635 0.059 0.184 0.039 0.082 2008 0.628 0.061 0.184 0.039 0.087 2009 0.620 0.056 0.185 0.037 0.103 2010 0.617 0.059 0.189 0.035 0.100 2011 0.615 0.056 0.192 0.036 0.101 2012 0.604 0.055 0.188 0.049 0.104 2013 0.604 0.053 0.184 0.051 0.108 2014 0.602 0.052 0.185 0.051 0.110 Source: Authors Calculations. Table 11 shed light contribution component, precisely average absolute change, average absolute change LSBCI observed period 2006- 2014. positive negative variations LSBCI separately order assess precisely contribution component. Table 11 Decomposition LSBCI: positive negative variation 2006-2014 Number Transhipments Common Direct Geo. Direct Carriers Constraint Ship Size Constraint LSBCI Negative -0.101 -0.029 -0.0295 0.0068 -0.0045 -0.157 Positive 0.093 0.0176 0.042 0.0193 0.068 0.2403 Source: Authors Calculations. Bilateral Liner Shipping Connectivity 2006 19 expected, number transhipments component plays crucial role cases LSBCI variation negative. components reflecting centrality country pair liner shipping network (Common Direct Geo. Direct) play major role cases LSBCI variation negative. , contribution carriers ship size components close . cases LSBCI increases, components participate shaping variation. strongest influence Number Transhipments component Ship size component. , keeping centrality country pair liner network important factor preservation LSBCI level. relaxation carriers ship size constraints order effect. Table 12 report average relative components respective standard deviation. Statistics computed sample positive negative LSBCI separately. average number transhipments connect country pair remained constant. , increased 27 cent cases decrease LSBCI decreased 13 cent cases increase. number common direct connections, increased 7 cent , 50 cent cases positive variation LSBCI decreased 50 cent cases negative variation. similar contrast observed geometric direct connections. average, number carriers maximum ship size operating maritime itinerary increased. Interestingly, true positive negative variations LSBCI. sizeable. average number carriers increased 80 cent maximum ship size doubled. noticed , variability negligible reflected standard deviation values relative average . Table 12 Relative components' raw values LSBCI variation Standard Deviation Number Transhipments negative 0.269 0.457 positive -0.130 0.294 0.015 0.409 Common Direct negative -0.498 0.432 positive 0.494 1.158 0.072 1.044 Geo. Direct negative -0.136 0.223 positive 0.275 0.431 0.120 0.417 Carriers Constraint negative 0.754 1.907 positive 0.807 1.667 0.787 1.761 Ship Size Constraint negative 0.301 1.884 positive 1.371 3.150 0.968 2.791 Source: Authors Calculations Results presented Table 11 consistent results presented Table 12. LSBCI highly sensitive centrality indicators components. Freight costs indicators carriers maximum size components play important role driving LSBCI variations positive impact remains order. 20 POLICY ISSUES IN INTERNATIONAL TRADE AND COMMODITIES 3.5 TRANSHIPMENTS AND THE LSBCI discussed previously, choice LSBCI components "biased" favor extensivity (captured components 1 3) connectivity intensity (captured components 4 5). Previously presented figures reveal crucial role played number transhipments component, component 1. Table 13 shows component varied 2006 2014. 78 cent country pairs, number transhipment connect remained constant. 13 cent number increased (deterioration) 10 cent decreased (improvement). Table 13 Variation number transhpiments 2006-20014 Variation Country-pairs Share Total (%) -2 5 0.04 -1 1,142 9.69 0 9,134 77.53 1 1,491 12.66 2 9 0.08 Source: Authors Calculations. list countries improve connectivity terms transhipping 30 destinations ( 154) includes Bahrain, Belize, democratic Republic Congo, Jordan, Kuwait, Morocco, Qatar, Russian Federation, Slovenia. extreme, list countries connectivity terms transhipping deteriorated 30 destinations includes Aruba, Bulgaria, Congo, Fiji, Georgia, Haiti, Indonesia, Iran, Iraq, Latvia, Saint Kitts Nevis, Madagascar, Montenegro, Mauritania, Papua Guinea, Sierra Leone, Tonga Vanuatu. group countries characterized small number direct connections. consequence dependency centrality index destinations connected extremely high. Figure 10 plots progression terms net creation direct connections 2006 2014 country level. Countries ordered terms performance terms net creation. top performers Morocco, Russian Federation, Democratic Republic Congo, Panama Turkey. Net creation amounted 12 Turkey 38 Morocco. bottom performers Congo, Aruba, Indonesia, Great Britain Brazil. Net creation amounted -23 Congo Indonesia , -16 Brazil. presence Brazil Great Britain bottom performers surprising. , reflection rationalization liner shipping network implied 2008 crisis severe impact global demand. Bilateral Liner Shipping Connectivity 2006 21 Figure 10: Direct Connections: Net creation 2008-2014 Source: Authors Calculations. 3.6 THE LSBCI AND MARITIME DISTANCE discussed , LSBCI components reflect specifically liner shipping connectivity pairs countries. Maritime distance included set components represents generic indicator connectivity precisely connectability. , direct connections, maritime distance affected maritime route characteristic vary time vary. computed effective measure maritime distance shortest route pair countries. , connections direct, effective maritime distance compute coincides sea distance. correlation measures extremely high, close 0.95 year. Table 14 shows pairwise correlation coefficients measure effective maritime distance, LSBCI components. Results reported years 2006, 2010 2014. Variations time coefficients excessive. strongest correlations number transhipments component number common direct connections component. high reasonable impact components completely dictated maritime distance. Table 14 Maritime distance LSBCI: pairwise correlations Number Transhipments Common Direct Geo. Direct Carriers Constraint Ship Size Constraint LSBCI 2006 0.3859 -0.3053 -0.1376 -0.1373 -0.1213 -0.2849 2010 0.3993 -0.2911 -0.137 -0.144 -0.1326 -0.2983 2014 0.401 -0.2984 -0.1404 -0.1299 -0.1138 -0.2951 Source: Authors Calculations. conclusion holds LSBCI. words, LSBCI represent authentic liner shipping connectivity index exclusively driven maritime distance. 22 POLICY ISSUES IN INTERNATIONAL TRADE AND COMMODITIES 4. DISCUSSION paper presents revised version Liner Shipping Bilateral Connectivity Index computed set 155 coastal countries observed 9 years period 2006 2014. unique statistics trends discussed detail order shed light evolution bilateral connectivity 2006. original set information important complement UNCTAD’ country-level Liner Shipping Connectivity Index (LSCI), support unilateral analysis. definition construction LSBCI based reliable data fleet deployment empirical interest. index components expected reflect large extent freight costs. added set determinants bilateral trade containerizable goods. companion paper, Fugazza Hoffman (2015) present assessment impact LSBCI bilateral exports containerizable goods comprehensive set country pairs observed 8 years 2006-2013 period. results point significant impact index components intensive margin trade containerizable goods. definition construction LSBCI , attempt identify bilateral connectivity affects bilateral exchanges. inter-temporal dimension possibility monitor index components time extremely helpful framing practical policy orientations. LSBCI framework offers unique globalized view liner shipping network offers possibility position network specific country dimensions. Table 15 reports decomposition LSBCI top 10 destinations country year 2006 (upper panel) year 2014 ( panel). information displayed Table 15 ways. instance, drawing stylized facts top 10 connections year. composition extent country level terms geographical composition increased presence Latin American countries detriment Asian European . number transhipments connect country top ten destinations, instance, direct connections. Top 10 group countries years. case, identify precisely progression regression specific country. case Belgium (BEL), position 2006 slipped fourth rank 2014. negative identified expect loss common direct destination. "downgrading" destination due progression slower countries Colombia China. set information policy makers elect direction dig deeper order eventually policy orientation conclusions. Table 14 shows extract country' liner shipping relationships. analysis extended full set destinations 154, easily reproduce approach. consideration land-locked countries. Maritime connectivity issue land locked countries fully relying maritime connectivity transit countries. clear precise appreciation issue requires dedicated possibly systematic analysis simply assigning LSBCI transit countries land locked countries. Bilateral Liner Shipping Connectivity 2006 23 Table 15 LSBCI components: country' top 10 destinations Source: Authors calculations. Notes: column reports top 10 destinations 2006 2014 . column shows sea distance destinations. Columns 3 7 report raw values LSBCI component. Columns 8 12 report normalized LSBCI component. column reports LSBCI top 10 destinations. 24 POLICY ISSUES IN INTERNATIONAL TRADE AND COMMODITIES REFERENCES Anderson, .., Wincoop, . van, 2003. Gravity Gravitas: Solution Border Puzzle. American Economic Review 93, 170–192. Arvis, .-., Shepherd, ., Reis, .., Duval, ., Utoktham, ., 2013. Trade Costs Development: Data Set. World Bank - Economic Premise 1–4. Clark, ., Dollar, ., Micco, ., 2004. Port Efficiency, Maritime Transport Costs Bilateral Trade (NBER Working Paper . 10353). National Bureau Economic Research, . Fugazza, ., 2015. Maritime Connectivity Trade - Policy Issues International Trade Commodities, Study Series . 70. Fugazza, ., Hoffmann, ., 2015. Liner Shipping Connectivity determinant geography trade. Forthcoming Policy Issues International Trade Commodities. Hoffmann, ., Van Hoogenhuizen, .-., Wilmsmeier, ., 2014. Developing index bilateral liner shipping connectivity, : IAME 2014 Conference Proceedings. Presented International Association Maritime Economists (IAME), Norfolk, United States. UNCTAD, 2013. Review maritime transport 2012. United Nations Pubns. UNCTAD, 2008. modal split international goods transport. Transport Newsletter, UNCTAD/SDTE/TLB/MISC/2008/1. Wilmsmeier, ., Hoffmann, ., 2008. Liner Shipping Connectivity Port Infrastructure Determinants Freight Rates Caribbean. Maritime Economics & Logistics 10, 130– 151. doi:10.1057/palgrave.mel.9100195 Printed United Nations, Geneva 1600069 () – February 2016 – 246 UNCTAD/ITCD/TAB/74 United Nations publication ISSN 1607-8291