Im­pact Se­ries: 01-22

The Hidden Green Sourcing Potential in European Trade

Outline

In this Kühne Impact Series, we substantiate our claim that international trade should play a central role in the fight against climate change. The key message is that “buying green” does not necessarily mean “buying local” and that a smart combination of local and foreign sourcing yields the best results. To this end, we construct an exhaustive database on the greenhouse gas emissions related to international trade flows from and to the European Union. Our first main result is that 28% of European trade flows are already green in the sense of bringing about a net reduction in greenhouse gas emissions relative to domestic sourcing. Our second main result is that a simple green sourcing rule could reduce trade-related emissions by 35%. Hence, there is a substantial green sourcing potential remaining in European trade

Full Article

At the Kühne Cen­ter for Sus­tain­able Trade and Lo­gis­tics, we be­lieve that the world should em­brace in­ter­na­tion­al trade in the fight against cli­mate change. The rea­son is that “buy­ing green” does not nec­es­sar­i­ly mean “buy­ing lo­cal”, be­cause trans­port emis­sions only ac­count for a small share of to­tal emis­sions and there is large vari­a­tion in pro­duc­tion emis­sions across coun­tries. Buy­ing green – or green sourc­ing – in­stead in­volves com­par­ing the pro­duc­tion and trans­port emis­sions as­so­ci­at­ed with all avail­able sourc­ing op­tions and choos­ing the one that min­i­mizes the sum of both.

In this Kühne Im­pact Se­ries, we sub­stan­ti­ate this view by pro­vid­ing our most sys­tem­at­ic analy­sis of green sourc­ing to date. Our analy­sis has three parts:

First, we con­struct an ex­haus­tive data­base on the green­house gas emis­sions re­lat­ed to the in­ter­na­tion­al trade flows of the EU. While this data­base is main­ly an in­put into the analy­ses dis­cussed in the sub­se­quent sec­tions, it also re­veals some ba­sic in­for­ma­tion that is in­ter­est­ing in its own right. In par­tic­u­lar, we find that trans­port emis­sions only ac­count for 16% of to­tal emis­sions and are less strong­ly re­lat­ed to dis­tance than one might think.

Sec­ond, we use this data­base to take stock of Eu­ro­pean trade as it is. In par­tic­u­lar, we com­pare the to­tal emis­sions as­so­ci­at­ed with each trade flow with the hy­po­thet­i­cal al­ter­na­tive of “buy­ing lo­cal”. Our main re­sult is that 28% of Eu­ro­pean trade flows are al­ready green in the sense of bring­ing about a net re­duc­tion in to­tal emis­sions. How­ev­er, this av­er­age masks huge vari­a­tion across ori­gins, des­ti­na­tions, and sec­tors, which we also dis­cuss.

Third, we turn to ex­plor­ing Eu­ro­pean trade as it could be. In par­tic­u­lar, we de­vel­op a sim­ple green sourc­ing rule based on a new sec­tor-lev­el met­ric that we call “for­eign green sourc­ing po­ten­tial”. A high po­ten­tial sim­ply means that trans­port emis­sions tend to be rel­a­tive­ly low and pro­duc­tion emis­sions tend to be rel­a­tive­ly dis­persed in a sec­tor so that there is a high chance of for­eign sourc­ing be­ing the green choice. Our main re­sult is that this sim­ple green sourc­ing rule could re­duce trade-re­lat­ed emis­sions by 35% – more than twice of what could be achieved by only “buy­ing lo­cal”. Hence, there is a sub­stan­tial green sourc­ing po­ten­tial re­main­ing in Eu­ro­pean trade.

Measuring trade emissions

In this sec­tion, we de­scribe the con­struc­tion of our new data­base. Read­ers who are main­ly in­ter­est­ed in our re­sults can skip for­ward to the next sec­tion.

Our data­base records the pro­duc­tion and trans­port emis­sions as­so­ci­at­ed with the in­ter­na­tion­al trade in goods and ser­vices be­tween 28 EU coun­tries and 67 trad­ing part­ners (in­clud­ing a resid­ual “Rest of the World”) dur­ing the years 2003–2018.

We con­struct this data­base by es­ti­mat­ing trans­port emis­sions and com­bin­ing our es­ti­mates with pre-ex­ist­ing in­for­ma­tion on pro­duc­tion emis­sions. We es­ti­mate trans­port emis­sions us­ing a top-down ap­proach build­ing on Cristea et al (2013).1 This ap­proach re­lies on data on quan­ti­ties (rather than val­ues) in­ter­na­tion­al­ly trad­ed and bro­ken down by the broad modes of trans­port used. Com­bined with in­for­ma­tion on dis­tances be­tween coun­tries and mea­sures of av­er­age green­house gas emis­sions per ton-km for in­di­vid­ual modes of trans­port, it is pos­si­ble to es­ti­mate the CO2eq emis­sions as­so­ci­at­ed with the trans­port of in­ter­na­tion­al­ly trad­ed goods and ser­vices.

Con­crete­ly, we com­bine pub­licly avail­able data on pro­duc­tion emis­sions for a set of 67 coun­tries and 30 (trad­ed) sec­tors from the OECD2 with trade sta­tis­tics from Eu­ro­stat and IPCC tech­ni­cal co­ef­fi­cients to con­struct to­tal trade-re­lat­ed emis­sions. Eu­ro­stat pro­vides trad­ed val­ues and weights be­tween the 28 EU mem­bers and the rest of the world at the CN8 prod­uct lev­el,3 bro­ken down by sev­en modes of trans­port: sea, air, road, rail, pipeline, self-propul­sion and oth­er, be­tween 2001 and 2018. Where nec­es­sary, we im­pute the modal split of trade flows based on the ge­og­ra­phy of the ori­gin and des­ti­na­tion coun­tries, the type of good trans­port­ed, and the to­tal val­ue and weight trad­ed. A tech­ni­cal ap­pen­dix with the pre­cise pro­ce­dure of the data con­struc­tion is avail­able upon re­quest. We ob­tain av­er­age emis­sions per ton-km per mode of trans­port from the Fifth As­sess­ment Re­port of the IPCC.4 As these co­ef­fi­cients are pro­vid­ed as a range (low­er and up­per bound) we are in fact able to con­struct trans­port emis­sions un­der a HIGH emis­sions sce­nario (us­ing the up­per bound) or a LOW emis­sions sce­nario. We will most­ly fo­cus on the low sce­nario in what fol­lows.5

We close this sec­tion by record­ing two ba­sic points about trans­port emis­sions that emerge from our data­base and pro­vide some help­ful con­text to our sub­se­quent analy­sis.

First, trans­port emis­sions ac­count on av­er­age for 16% of to­tal emis­sions. The pro­duc­tion of a trad­ed good emits on av­er­age 326 g of CO2 per dol­lar pro­duced, with tremen­dous het­ero­gene­ity across coun­tries, goods, and over time.6 In com­par­i­son, the trans­port of a good emits on av­er­age be­tween 150 and 240 g of CO2 per dol­lar trans­port­ed (un­der the LOW or HIGH sce­nario, re­spec­tive­ly), with even larg­er vari­a­tion. When ac­count­ing for these two types of emis­sions in trade, we find that one dol­lar of goods trad­ed in­ter­na­tion­al­ly emits on av­er­age be­tween 476 and 566 g of CO2. To put this num­ber in per­spec­tive, a car­bon tax of $90 per ton of car­bon (ap­prox­i­mate­ly €80, which was the price of a Eu­ro­pean car­bon per­mit in Jan­u­ary 2022)7 on a good emit­ting 450 g of CO2 per dol­lar trad­ed will be equiv­a­lent to im­pos­ing a 4% ad val­orem tar­iff on the trans­ac­tion.

Sec­ond, trans­port emis­sions are less strong­ly re­lat­ed to dis­tance than one might think. This is il­lus­trat­ed in Fig­ure 1, which plots trans­port emis­sions in­ten­si­ties by trade route. For ex­am­ple, the me­di­an trans­port emis­sions in­ten­si­ty for flows be­tween Eu­rope and Ocea­nia is three times as large as the me­di­an in­ten­si­ty for in­tra-EU trade (53 g of CO2 per USD trad­ed com­pared to 17 g). Nonethe­less, the in­terquar­tile ranges sub­stan­tial­ly over­lap, sug­gest­ing that the brownest in­tra-EU trans­ac­tion has in fact a larg­er trans­port emis­sion in­ten­si­ty than the green­est EU-Ocea­nia one (40 g and 36 g, re­spec­tive­ly).

Notes: This fig­ure rep­re­sents the me­di­an (yel­low ver­ti­cal line), the in­terquar­tile range (box de­lim­it­ing the bot­tom 25th per­centile and the top 25th per­centile of the dis­tri­bu­tion), as well as the max­i­mum and min­i­mum val­ues (yel­low hor­i­zon­tal line) of trans­port emis­sion in­ten­si­ties un­der the LOW sce­nario by trade route for the year 2015.

This may arise be­cause of dif­fer­ences in the modal split or the com­po­si­tion of these flows. Trans­port emis­sions are high­ly de­pen­dent on the mode of trans­port used to trade goods for two rea­sons: (i) dif­fer­ences in emis­sions per ton-km while every­thing else held con­stant, and (ii) dif­fer­ences in the type of goods be­ing trans­port­ed with each mode. In our ex­am­ple from above, we find that 64% of the trad­ed val­ues be­tween Eu­rope and Ocea­nia is trans­port­ed by sea, where­as 59% of in­tra-EU trad­ed val­ue is trans­port­ed by road (Table 1). A con­tain­er ship emits on av­er­age 10 g of CO2eq per ton-km trans­port­ed, where­as a heavy-duty truck will emit 76 g of CO2eq per ton-km, tip­ping the scale in fa­vor of brown­er modes of trans­port for in­tra-EU trade. Once con­vert­ed in grams of CO2eq per dol­lar trad­ed, these val­ues be­come 15 and 37 g, re­spec­tive­ly, of CO2eq per dol­lar, re­flect­ing the high­er av­er­age val­ue per kg of goods trans­port­ed by trucks, there­by re­duc­ing trans­port emis­sion in­ten­si­ties in fa­vor of in­tra-EU trade.

Transport emissions account on average for 16% of total trade emissions.

The con­tri­bu­tion of trans­port emis­sions to to­tal trade emis­sions needs to be seen rel­a­tive to pro­duc­tion emis­sions. For ex­am­ple, the av­er­age trans­ac­tion be­tween Eu­rope and Ocea­nia emits 236 g of CO2 per dol­lar pro­duced, where­as the av­er­age trans­ac­tion with­in Eu­rope emits 250 g. Thus, trans­port emis­sions will weigh more in trade emis­sions of the Eu­rope-Ocea­nia route (59% on av­er­age) also be­cause what is pro­duced and how is in fact clean­er than in in­tra-EU trade.

Notes: This ta­ble re­ports the av­er­age (per­cent) share in 2015 of each mode of trans­port in our sam­ple in to­tal val­ues or quan­ti­ties trad­ed by trade route. Av­er­age is tak­en across years in our sam­ple. Note that the shares do not have to sum to 1 as the av­er­age is tak­en across sec­tors and coun­tries.

European trade as it is

We now use this data­base to take stock of Eu­ro­pean trade as it is. In par­tic­u­lar, we com­pare the to­tal emis­sions as­so­ci­at­ed with each trade flow with the hy­po­thet­i­cal al­ter­na­tive of “buy­ing lo­cal”. For each in­di­vid­ual trade flow of good g pro­duced in ori­gin o and con­sumed in des­ti­na­tion d, we com­pare the emis­sions from pro­duc­tion of g in o and trans­port be­tween o and d in the data in 2015, with emis­sions from pro­duc­tion of g in d with no trans­port in­volved.

This ex­er­cise is ex­treme to the ex­tent that not all im­porters have the ca­pac­i­ty of pro­duc­ing the goods that they are sourc­ing abroad, but it al­lows us nonethe­less to pin­point where trade makes a pos­i­tive con­tri­bu­tion to re­duce emis­sions. We call a trade flow “green” if the emis­sions from trade in the data are low­er than the coun­ter­fac­tu­al emis­sions un­der lo­cal pro­duc­tion.8

Our main re­sult is that 28% of Eu­ro­pean trade flows are al­ready “green” in the sense of bring­ing about a net re­duc­tion in to­tal emis­sions. This num­ber ag­gre­gates the share of green trade flows in EU ex­ports, EU im­ports, and in­tra-EU trade. In par­tic­u­lar, 34% of all EU ex­ports are green, 22% of all EU im­ports are green, and 27 % of all in­tra-EU trade flows are green, ac­cord­ing to our de­f­i­n­i­tion.

How­ev­er, green trade flows are very dis­persed across ori­gins, des­ti­na­tions, and sec­tors. Fig­ure 2 re­ports for each ori­gin coun­try the share of green trade in the to­tal val­ue of their ex­ports. There is a small range of very green ex­porters (72% of all Irish ex­ports in val­ue are green, 62% of Ice­land’s), and a very large range of coun­tries with mo­not­o­n­i­cal­ly de­clin­ing shares (53% of Swiss ex­ports are green, 47% of Ger­many’s ex­ports, 16% for the US, 10% for Rus­sia). South­east Asian ex­ports tend to be very brown with only 2.6% of Chi­nese ex­ports to the EU la­beled green ac­cord­ing to our met­ric.

Green trade flows are very dispersed across origins, destinations, and sectors.

The pat­tern is some­what sim­i­lar when con­sid­er­ing im­porters (Fig­ure 3): 90% of Russ­ian im­ports are green­er if sourced from the EU. In­ter­est­ing­ly, 65% of US im­ports are green­er when com­ing from the EU, where­as this frac­tion falls to 50% for Chi­nese im­ports. At the end of the spec­trum for North­ern Eu­ro­pean coun­tries, less than 15% of their im­ports should be pro­duced abroad rather than lo­cal­ly (10% for Switzer­land, 8.6% for Swe­den).

Per­haps sur­pris­ing­ly, the share of green trade val­ue across sec­tors shows no cor­re­la­tion with pro­duc­tion emis­sion in­ten­si­ties. Fig­ure 4 re­veals that more than 63% of trad­ed val­ue in the sec­tor of “Min­ing of en­er­gy prod­uct” is bet­ter trad­ed than sourced lo­cal­ly, with the shares be­ing 61% in “Phar­ma­ceu­ti­cals”, 59% for “En­ter­tain­ment ser­vices”, and 53% for “Ad­min­is­tra­tive sup­port ac­tiv­i­ties”. At the bot­tom of the rank­ing, only 13% of the val­ue trad­ed in “Food prod­ucts” is green ac­cord­ing to our met­ric, sug­gest­ing that buy­ing lo­cal may have an im­pact in that sec­tor.

Notes: This fig­ure dis­plays the con­tri­bu­tion of green trade flows in the to­tal trad­ed val­ue by ori­gin. For a giv­en coun­try dis­played on the y-axis, we re­port the con­tri­bu­tion in per­cent­age of the val­ue of “green” ex­ports on the to­tal val­ue of its ex­ports.
Notes: This fig­ure dis­plays the con­tri­bu­tion of green trade flows in the to­tal trad­ed val­ue by des­ti­na­tion. For a giv­en coun­try dis­played on the y-axis, we re­port the con­tri­bu­tion in per­cent­age of the val­ue of “green” im­ports on the to­tal val­ue of its im­ports.
Notes: This fig­ure dis­plays the con­tri­bu­tion of green trade flows in the to­tal trad­ed val­ue by sec­tor. For a giv­en sec­tor on the y-axis, we dis­play the con­tri­bu­tion in per­cent­age of the val­ue of “green” trade flows on the to­tal val­ue of trans­ac­tions of goods from that sec­tor.

European trade as it could be

We now turn to ex­plor­ing Eu­ro­pean trade as it could be. In par­tic­u­lar, we cal­cu­late the emis­sions re­duc­tion as­so­ci­at­ed with a sim­ple green sourc­ing rule based on a new sec­tor-lev­el met­ric that we call “for­eign green sourc­ing po­ten­tial.” A high po­ten­tial sim­ply means that trans­port emis­sions tend to be rel­a­tive­ly low and pro­duc­tion emis­sions tend to be rel­a­tive­ly dis­persed in a sec­tor so that there is a high chance of for­eign sourc­ing be­ing the green choice.

Fig­ure 5 shows what we mean by the for­eign green sourc­ing po­ten­tial of a sec­tor. Each dot is a sec­tor. The hor­i­zon­tal axis mea­sures the dis­per­sion of the pro­duc­tion emis­sion in­ten­si­ties across coun­tries. Oth­er things equal, a high­er dis­per­sion comes along with a high­er green sourc­ing po­ten­tial, since there is then a high chance that for­eign coun­tries have ac­cess to sub­stan­tial­ly green­er pro­duc­tion tech­nolo­gies. The ver­ti­cal axis mea­sures the share of trans­port emis­sions in to­tal emis­sions. Oth­er things equal, a high­er share of trans­port emis­sions comes along with a low­er green sourc­ing po­ten­tial, sim­ply be­cause trad­ing it­self is then rel­a­tive­ly dirty.

Notes: This fig­ure dis­plays the dis­per­sion of trad­ed sec­tors in the space of trans­port and pro­duc­tion emis­sions for the year 2015. On the x-axis, we re­port the log stan­dard de­vi­a­tion of pro­duc­tion emis­sion in­ten­si­ty of a giv­en sec­tor across pro­duc­ing coun­tries (ex­clud­ing the bot­tom 5th per­centile of pro­duc­ing coun­tries with close to zero con­tri­bu­tion to glob­al out­put). On the y-axis, we rep­re­sent the con­tri­bu­tion of trans­port emis­sions to to­tal trade emis­sions for a giv­en sec­tor in per­cent­age.

Sec­tors in the north­west quad­rant are sec­tors with a low for­eign green sourc­ing po­ten­tial, sug­gest­ing that coun­tries should typ­i­cal­ly source lo­cal­ly in these sec­tors. This is the case for a ma­jor­i­ty of trad­able ser­vices (e.g. “Pro­fes­sion­al, sci­en­tif­ic and tech­ni­cal ac­tiv­i­ties” la­beled as “R&D” in Fig­ure 4, which have the sec­ond high­est trans­port emis­sion con­tri­bu­tion of 70%) that have vir­tu­al­ly no pro­duc­tion emis­sions but high trans­port emis­sions from in­di­vid­ual trav­el (in the case of R&D, 53 g of CO2 per dol­lar trad­ed, which ranks in the top 10 sec­tors with high­est trans­port emis­sion in­ten­si­ties).

Sec­tors in the south­east quad­rant are sec­tors with a high for­eign green sourc­ing po­ten­tial, sug­gest­ing that coun­tries should typ­i­cal­ly source in­ter­na­tion­al­ly in these sec­tors. It is very in­ter­est­ing to note that amongst these sec­tors we find raw ma­te­ri­als (coke and pe­tro­le­um, min­ing, non-metal­lic min­er­als) that tend in our data to rank first both in terms of av­er­age pro­duc­tion emis­sion in­ten­si­ty and pro­duc­tion emis­sion in­ten­si­ty dis­per­sion. This il­lus­trates the point made in our pre­vi­ous Kühne Im­pact Se­ries that the brownest sec­tors tend to be sourced from the brownest ori­gins, and is at the heart of our ar­gu­ment in fa­vor of green sourc­ing.

We are now ready to ask by how much emis­sions could be re­duced by go­ing from a world with Eu­ro­pean trade as it is to a world with Eu­ro­pean trade as it could be. Our main re­sult is that trade-re­lat­ed emis­sions could be re­duced by 35% if we source goods ac­cord­ing to their for­eign green sourc­ing po­ten­tial.

Con­crete­ly, we cal­cu­late what would hap­pen to to­tal emis­sions if coun­tries (i) sourced lo­cal­ly in all sec­tors from the north­west quad­rant of Fig­ure 5, (ii) sourced in­ter­na­tion­al­ly in all sec­tors from the south­east quad­rant of Fig­ure 5, and (iii) left trade un­changed in all oth­er sec­tors. In case (ii), we as­sume that each good trans­act­ed in a sec­tor g is now al­ways im­port­ed from the glob­al­ly green­est pos­si­ble ori­gin o*, that is from a coun­try with one of the low­est pro­duc­tion emis­sion in­ten­si­ty for sec­tor g, but with large enough pro­duc­tion ca­pac­i­ties. In all cas­es, we also ad­just the trade routes to ac­count for changes in trans­port emis­sions as­so­ci­at­ed with a change of ori­gin.9

Little effort is required to achieve massive emission savings.

We find that this sim­ple green sourc­ing rule would re­duce trade-re­lat­ed emis­sions by 35% rel­a­tive to their lev­el in 2015. In­ter­est­ing­ly, much of these gains could be achieved through green sourc­ing just a small share of crit­i­cal trade flows. This is il­lus­trat­ed in Fig­ure 6. To con­struct this fig­ure, we rank trade flows by their “green po­ten­tial”, that is, by the emis­sions saved when an in­di­vid­ual trans­ac­tion is be­ing sourced ac­cord­ing to our green sourc­ing rule. We then cal­cu­late how much of the 35% emis­sion re­duc­tion is ob­tained when se­quen­tial­ly re­al­iz­ing the green po­ten­tial of these ranked flows. Fig­ure 6 says, for ex­am­ple, that 95% of the pos­si­ble emis­sions re­duc­tion through green sourc­ing can al­ready be achieved by green sourc­ing only the 10% trade flows where it mat­ters most.

This re­sult re­flects the skew­ness of the dis­tri­bu­tion in trade emis­sion in­ten­si­ties: a few flows in our data are ex­treme­ly brown com­pared to the rest, for ex­am­ple im­ports of ma­chin­ery and equip­ment from Chi­na that could in­stead be sourced from Den­mark. This means that lit­tle ef­fort is re­quired to achieve mas­sive emis­sion sav­ings.

We can com­pare this 35% re­duc­tion in emis­sion with the gain achieved if one were to “buy lo­cal” only. If we im­pose that any trans­ac­tion of good g be­tween ori­gin o and des­ti­na­tion d is pro­duced lo­cal­ly – by coun­try d rather than coun­try o – we find that trade-re­lat­ed emis­sions would fall by 14% rel­a­tive to 2015, less than half of what our green sourc­ing rule could achieve.10 This con­firms the tremen­dous po­ten­tial of “buy­ing green” rather than “lo­cal” to re­duce emis­sions.

Notes: This fig­ure dis­plays the path to emis­sion re­duc­tion when pro­gres­sive­ly go­ing from the base­line (real) data to the coun­ter­fac­tu­al sce­nario with op­ti­miza­tion of sourc­ing ori­gins. On the x-axis we re­port the pro­por­tion of trade flows in our data that we mod­i­fied to sat­is­fy the coun­ter­fac­tu­al rules. We start from the real data, and only change emis­sions (ac­cord­ing to our coun­ter­fac­tu­al sce­nario) for the brownest flow. We pro­gres­sive­ly mod­i­fy more trade flows, go­ing from the ones with the largest emis­sion-sav­ing po­ten­tial to the ones gen­er­at­ing lit­tle change. On the y-axis, we re­port the frac­tion of the to­tal emis­sion re­duc­tion we get with a giv­en frac­tion of flows mod­i­fied.

Conclusion

In this Kühne Im­pact Se­ries, we sub­stan­ti­ate our claim that in­ter­na­tion­al trade should play a cen­tral role in the fight against cli­mate change by pro­vid­ing our most sys­tem­at­ic analy­sis of green sourc­ing to date. Our main point is that trade can con­tribute sub­stan­tial­ly to a re­duc­tion in green­house gas emis­sions if sourc­ing de­ci­sions are made by keep­ing trans­port and pro­duc­tion emis­sions in mind. Our quan­ti­ta­tive as­sess­ment is that a sim­ple green sourc­ing rule im­posed on Eu­ro­pean trade could re­duce trade-re­lat­ed emis­sions by 35% – more than twice of what could be achieved by only “buy­ing lo­cal”. Hence, there is a sub­stan­tial green sourc­ing po­ten­tial re­main­ing in Eu­ro­pean trade.

While we have done our best to keep our quan­ti­ta­tive analy­ses re­al­is­tic, there is clear­ly much scope for fu­ture work. An im­por­tant short­com­ing of our ap­proach is that we sim­ply im­pose a giv­en green sourc­ing rule, in­stead of study­ing how the world econ­o­my would re­act to a more re­al­is­tic pol­i­cy in­ter­ven­tion such as a car­bon tax. This is ex­act­ly what we tack­le in our on­go­ing Glob­al Sus­tain­abil­i­ty In­dex project, and we are look­ing for­ward to re­leas­ing our first re­sults soon.

  1. “Trade and the green­house gas emis­sions from in­ter­na­tion­al freight trans­port”. 2013. Cristea, A., Hum­mels, D., Puzel­lo, L. and Avetisyan, M. Jour­nal of En­vi­ron­men­tal Eco­nom­ics and Man­age­ment 65, pp. 153–173.
  2. https://www.oecd.org/sti/ind/car­bon­diox­idee­mis­sion­sem­bod­ie­din­in­ter­na­tion­al­trade.htm
  3. The Com­bined Nomen­cla­ture (CN) is the pri­ma­ry nomen­cla­ture used by the EU mem­ber states to col­lect de­tailed data on their trad­ing of goods since 1988. It is based on the Har­mo­nized Sys­tem: The CN cor­re­sponds to the HS6 code plus a fur­ther break­down at the eight-dig­it lev­el de­fined to meet EU needs. We use the HS6 part. The Har­mo­nized Sys­tem (HS) clas­si­fi­ca­tion is an in­ter­na­tion­al nomen­cla­ture de­vel­oped by the World Cus­tom Or­ga­ni­za­tion for the clas­si­fi­ca­tion of goods.
  4. An­nex III: Tech­nol­o­gy-spe­cif­ic Cost and Per­for­mance Pa­ra­me­ters. 2014. Work­ing Group III Con­tri­bu­tion to the Fifth As­sess­ment Re­port of the In­ter­gov­ern­men­tal Pan­el on Cli­mate Change, pp. 1329–1357.
  5. The orig­i­nal data also in­cludes in­ter­na­tion­al trade flows from and to the US that we do not use in this Se­ries. It should be em­pha­sized that the data is still in con­struc­tion and sub­ject to caveats that we are cur­rent­ly ad­dress­ing. One of the key lim­i­ta­tions to the cur­rent data lies in the ac­cu­ra­cy of the weights trad­ed. Trade data are not nec­es­sar­i­ly re­port­ed in stan­dard units (tons or kg), and some records were in­com­plete. When pos­si­ble, weights were ap­prox­i­mat­ed us­ing weight-to­val­ue ra­tios at the year-ori­gin-prod­uct lev­el (across trans­port modes). Such ra­tios are strong­ly de­pen­dent on whether val­ues are mea­sured Free on Board (FOB) or in­clud­ing “Cost, in­sur­ance and freight” (CIF). Eu­ro­pean im­ports, for ex­am­ple, are all re­port­ed in CIF. This means that the mon­e­tary val­ue of these flows in­cludes the cost of the trans­port, and as such is over­es­ti­mat­ed. As a re­sult, im­put­ed weights will be un­der­es­ti­mat­ed. The emis­sions re­port­ed in this Se­ries should there­fore be viewed as a con­ser­v­a­tive es­ti­mate of trans­port-re­lat­ed emis­sions.
  6. See our Kühne Im­pact Se­ries “Buy Green not Lo­cal: How In­ter­na­tion­al Trade Can Help Save Our Plan­et” (03/2021) for a de­tailed dis­cus­sion of het­ero­gene­ity in emis­sion in­ten­si­ties
    across pro­duc­ing coun­tries and sec­tors.
  7. See our Kühne Im­pact Se­ries “The Eu­ro­pean Green Deal: Trans­form­ing In­ter­na­tion­al Trade and Trans­porta­tion” (12/2021) for a dis­cus­sion of the Eu­ro­pean propo­si­tions on car­bon tax­a­tion.
  8. Note that in the cas­es where a coun­try sim­ply can­not pro­duce a good lo­cal­ly (e.g. for raw ma­te­ri­als) it will be the case that in fact we don’t have a pro­duc­tion emis­sion
    in­ten­si­ty for that coun­try (since there is no out­put). These in­stances are rare giv­en the coarse­ness of the data (0.7% of the coun­try-sec­tor pairs in 2015). When that is the case, both the orig­i­nal trade flow and the coun­ter­fac­tu­al emis­sions are dropped from the com­par­i­son so as not to bias the main mes­sage.
  9. To keep this ex­er­cise as re­al­is­tic as pos­si­ble, we adopt a cer­tain num­ber of rules to al­low a coun­try to be con­sid­ered as po­ten­tial sourc­ing ori­gin (main­ly that they al­ready have a large out­put). We also need to ad­just trans­port modal split to re­flect the new routes. See the tech­ni­cal ap­pen­dix for an ex­act de­scrip­tion of our sourc­ing rules, and the method­ol­o­gy adopt­ed to con­struct coun­ter­fac­tu­al trade flows and emis­sions.
  10. To avoid cas­es where a coun­try would be forced to pro­duce some­thing it nev­er has be­fore, we im­pose again some re­stric­tions on the sourc­ing pat­terns. If a coun­try has no
    ca­pac­i­ty to pro­duce a sec­tor, be it be­cause ob­served gross out­put is too low or be­cause of nat­ur­al re­sources con­straints, we im­pose that sec­tor g be sourced from a close neigh­bor in terms of ge­o­graph­ic dis­tance. See de­tails in the tech­ni­cal ap­pen­dix.

Author

Mathilde Le Moigne

Senior Research Fellow at the Kühne Center for Sustainable Trade and Logistics at the University of Zurich

Author

Luca Poll

Junior Research Fellow at the Kühne Center for Sustainable Trade and Logistics at the University of Zurich

Back

More Issues

Twitter logo