Im­pact Se­ries: 03–21

Buy Green not Local:
How International Trade
Can Help Save Our Planet


In this Kuehne Impact Series, we introduce the Kühne Center’s vision of environmentally sustainable globalization. Our main point is that buying local is not the same as buying green, because international trade can allow firms and households to source products from greener origins. We therefore advocate embracing international trade in the fight against climate change.

Full Article

In­ter­na­tion­al trade has a ter­ri­ble rep­u­ta­tion when it comes to its ef­fect on the en­vi­ron­ment. The main rea­son is that buy­ing lo­cal is usu­al­ly seen as equiv­a­lent to buy­ing green, giv­en that ships, trucks, and planes cause trans­port emis­sions from burn­ing fuel.

For ex­am­ple, imag­ine an en­vi­ron­men­tal­ly con­scious con­sumer from Ham­burg who is choos­ing be­tween a wine from Bor­deaux and a wine from Napa Val­ley. She would prob­a­bly choose the Bor­deaux with­out much hes­i­ta­tion, giv­en that Ger­many is clos­er to France than the Unit­ed States.

How­ev­er, this would al­most cer­tain­ly be the worse choice for the en­vi­ron­ment for two ba­sic rea­sons:

First, to­tal emis­sions not only de­pend on trans­port emis­sions but also on pro­duc­tion emis­sions, and there is huge vari­a­tion in pro­duc­tion emis­sions across coun­tries. Ac­cord­ing to Ne­me­cek and Poore (2018), vini­cul­ture caus­es around 50% more emis­sions in France than in the Unit­ed States, be­cause of the nat­ur­al ad­van­tages of the Cal­i­forn­ian cli­mate.1 Cru­cial­ly, pro­duc­tion emis­sions ac­count on av­er­age for more than 90% of to­tal emis­sions in agri­cul­ture (in­clud­ing vini­cul­ture) so that this should be the main con­sid­er­a­tion for any green oenophile.

Sec­ond, trans­port emis­sions not only de­pend on dis­tance but also the mode of trans­port, and there is again huge vari­a­tion in the trans­port emis­sions across modes. Ac­cord­ing to Ne­me­cek and Poore (2018), road trans­porta­tion caus­es 20 times more green­house gas emis­sions per ton kilo­me­ter than sea trans­porta­tion (the fac­tor is over 100 for air trans­porta­tion). And the dis­tance from San Fran­cis­co to Ham­burg by sea is only about 13 times the dis­tance from Bor­deaux to Ham­burg by road (ap­prox. 20,000 vs. 1,500 km).

Our Vision

At the Kühne Cen­ter, we there­fore be­lieve that it is cru­cial to em­brace in­ter­na­tion­al trade in the fight against cli­mate change. But what does this mean ex­act­ly?

From an eco­nom­ics per­spec­tive, the root cause of cli­mate change is that the pri­vate costs of emis­sions are low­er than the so­cial costs of emis­sions so that the mar­ket econ­o­my is ex­ces­sive­ly dirty. The text­book re­sponse is that car­bon should be priced at its so­cial costs, for ex­am­ple through an ap­pro­pri­ate car­bon tax. No­bel-prize-win­ning econ­o­mist William Nord­haus cur­rent­ly es­ti­mates the so­cial costs of car­bon to be $31 per ton.2

We there­fore de­fine en­vi­ron­men­tal­ly sus­tain­able glob­al­iza­tion as the glob­al pat­tern of trade that would

pre­vail if car­bon were priced at its so­cial costs. While this pat­tern can­not be ob­served in the data, it can be ap­prox­i­mat­ed with the help of mod­ern quan­ti­ta­tive trade mod­els. Kühne Cen­ter re­searchers Mar­cos Ri­tel and Dora Si­mon are in the process of work­ing out such a sce­nario against which cur­rent trade flows and pos­si­ble pol­i­cy in­ter­ven­tions can then be com­pared.

We can al­ready say how the com­po­si­tion of ac­tu­al trade flows dif­fers from the com­po­si­tion of sus­tain­able trade flows along a num­ber of im­por­tant di­men­sions. In par­tic­u­lar, there cur­rent­ly is (i) too much trade in rel­a­tive­ly ‘dirty’ goods, (ii) too much sourc­ing from rel­a­tive­ly ‘dirty’ coun­tries, (iii) and too much re­liance on rel­a­tive­ly ‘dirty’ trans­port modes. This is sim­ply be­cause rel­a­tive­ly dirty goods, coun­tries, and trans­port modes are still too cheap rel­a­tive to the sus­tain­able bench­mark, giv­en that car­bon is priced be­low its so­cial costs.

It is crucial to embrace international trade in the fight against climate change.

A First Look At The Data

In the re­main­der of this Se­ries, we take a first look at the data on world trade, pro­duc­tion, and emis­sions in light of our vi­sion of en­vi­ron­men­tal­ly sus­tain­able glob­al­iza­tion. Our main goal is to il­lus­trate that sus­tain­able glob­al­iza­tion can in­deed play an im­por­tant role in fight­ing cli­mate change.

We present our ev­i­dence as five facts. Facts 1–3 doc­u­ment that there is sub­stan­tial vari­a­tion in emis­sions across sec­tors, coun­tries, and trans­port modes. Facts 4–5 then show that in­ter­na­tion­al trade is cur­rent­ly tilt­ed to­wards dirt­i­er sec­tors and coun­tries. Tak­en to­geth­er, these facts sug­gest that buy­ing green­er goods, sourc­ing from green­er coun­tries, and us­ing green­er trans­port modes has the po­ten­tial to great­ly re­duce trade-em­bed­ded emis­sions.

Trade-em­bed­ded emis­sions ac­count for a sub­stan­tial share of to­tal emis­sions so that a sig­nif­i­cant re­duc­tion in trade-em­bed­ded emis­sions would go a long way in fight­ing cli­mate change. Us­ing OECD data, we es­ti­mate that the emis­sions as­so­ci­at­ed with the pro­duc­tion and trans­port of trad­ed goods ac­count for 30% of CO2 emis­sions world­wide.

Buying greener goods, sourcing from greener countries, and using greener transport modes has the potential to greatly reduce trade-embedded emissions.

Fact 1 – There is substantial heterogeneity in emissions across sectors

Fig­ure 1 plots the to­tal green­house gas emis­sions caused di­rect­ly by in­di­vid­ual eco­nom­ic sec­tors (dis­re­gard­ing the emis­sions from in­ter­me­di­ate stages of pro­duc­tion). As can be seen, there is sub­stan­tial vari­a­tion in green­house gas emis­sions across sec­tors. For ex­am­ple, pro­duc­tion of ba­sic and fab­ri­cat­ed met­als ac­counts for 7% of glob­al pro­duc­tion emis­sions, where­as pro­duc­tion of leather ap­par­el and footwear rep­re­sents less than 1%. Note that elec­tric­i­ty, gas and wa­ter sup­ply is a strong out­lier in our data. Di­rect pro­duc­tion emis­sions of this sec­tor alone make up 42% of glob­al green­house gas emis­sions.3 As one might ex­pect, man­u­fac­tur­ing sec­tors tend to be dirt­i­er than ser­vices sec­tors.

Fig­ure 1 also il­lus­trates emis­sion in­ten­si­ty per sec­tor, that is tons of emis­sions per dol­lar of out­put pro­duced. This vari­able has the ad­van­tage of be­ing in­de­pen­dent of the scale of the sec­tor. Ac­cord­ing to this mea­sure, the rank­ing of dirty sec­tors is slight­ly dif­fer­ent but still tilt­ed to­wards man­u­fac­tur­ing goods. En­er­gy, wa­ter and gas sup­ply re­mains the dirt­i­est sec­tor, fol­lowed by trans­port (in par­tic­u­lar wa­ter and air trans­port). Then come a num­ber of man­u­fac­tur­ing sec­tors, in par­tic­u­lar the man­u­fac­ture of min­er­als, met­als, fu­els, and chem­i­cals and chem­i­cal prod­ucts.

Fact 2 – There is substantial heterogeneity in emissions across countries

Fig­ure 2 plots the to­tal green­house gas emis­sions caused di­rect­ly by pro­duc­tion ac­tiv­i­ties tak­ing place in in­di­vid­ual coun­tries. As can be seen, there is sub­stan­tial vari­a­tion in green­house gas emis­sions across coun­tries. Much of this vari­a­tion can be ex­plained by vari­a­tion in (i) the scale of pro­duc­tion, (ii) the sec­tor com­po­si­tion, and (iii) the en­er­gy mix across coun­tries. For ex­am­ple, Chi­na is by far the largest emit­ter in the world, be­cause it pro­duces vast amounts, spe­cial­izes in some of the dirt­i­est sec­tors such as ba­sic and fab­ri­cat­ed met­als or chem­i­cals and chem­i­cal prod­ucts, and still re­lies heav­i­ly on brown en­er­gy sources such as coal.4 As one might ex­pect, coun­tries with a high GDP tend to cause more green­house gas emis­sions than coun­tries with a low GDP.

In terms of emis­sion in­ten­si­ty (emis­sions per dol­lar of gross do­mes­tic prod­uct), we see that some of the small­er coun­tries that ap­peared as low con­trib­u­tors of to­tal emis­sions (be­cause of their small scale) are in fact quite dirty. While Chi­na re­mains the brownest coun­try, it is close­ly fol­lowed by Cen­tral and East­ern Eu­ro­pean coun­tries (Es­to­nia, Rus­sia, Bul­gar­ia, Poland and the Czech Re­pub­lic rank among the top-10 more emis­sion in­ten­sive coun­tries) and Cen­tral and South-East Asian coun­tries (e.g., Ko­rea, In­dia, and Sin­ga­pore).5 Note that in terms of emis­sion in­ten­si­ty, the EU’s emis­sions are only about one-third of Chi­na’s.

In terms of emission intensity, some of the smaller countries that appeared as low contributors of total emissions are in fact quite dirty.

Fact 3 – There is substantial heterogeneity in emissions across transportation modes

Trans­porta­tion is one of the dirt­i­est sec­tors in the econ­o­my, ac­count­ing for al­most 10% of glob­al di­rect emis­sions (and rank­ing sec­ond in Fig­ure 1). No­tice, how­ev­er, that the trans­porta­tion sec­tor moves goods and peo­ple do­mes­ti­cal­ly and in­ter­na­tion­al­ly so that only part of these emis­sions are re­lat­ed to in­ter­na­tion­al trade.

There is sub­stan­tial het­ero­gene­ity across trans­port modes. In Table 1, we de­com­pose trans­port emis­sions into the con­tri­bu­tions of in­land trans­port, wa­ter trans­port, and air trans­port.

In­land trans­port alone ac­counts for more than 72% of to­tal emis­sions pro­duced by the trans­port sec­tor, fol­lowed by air trans­port (about 18%) and wa­ter trans­port (about 10%). But in­ter­na­tion­al trade ac­counts for only 8% of in­land trans­port, com­pared to 29% of air trans­port and 45% of wa­ter trans­port. Over­all, David Hum­mels (Pur­due Uni­ver­si­ty) and co-au­thors es­ti­mate that in­ter­na­tion­al trans­port ac­counts for 33% of trade-re­lat­ed emis­sions world­wide.6

Source: WIOD 2013 Re­lease. This ta­ble de­com­pos­es to­tal trans­port emis­sions into in­land, wa­ter and air trans­port. It also re­ports emis­sions per ton kilo­me­ter and per dol­lar of out­put. Emis­sions per ton kilo­me­ter come from Ne­me­cek and Poore (2018).

Air trans­port is the dirt­i­est trans­port mode (in terms of CO2 emit­ted per ton-kilo­me­ter), where­as wa­ter trans­port is the clean­est. Even with­in these groups, there is again large vari­a­tion: for ex­am­ple, con­tain­er ships emit about three times more tons of CO2 per ton-kilo­me­ter than bulk ships, and car­go fleets are twice as pol­lut­ing as bel­ly freight on pas­sen­ger planes. But the way trans­port modes are priced shifts the per-dol­lar con­tri­bu­tion of each mode. Be­cause air trans­port is so ex­pen­sive, its emis­sion rate per dol­lar of out­put is slight­ly low­er than wa­ter trans­port, but still high­er than in­land trans­port.

Air transport is the dirtiest transport mode in terms of CO₂ emitted per ton-kilometer, whereas water transport is the cleanest.

Fact 4 – International trade is tilted towards dirtier goods

Fig­ure 3 shows that dirt­i­er goods tend to be trad­ed more than clean­er goods, where a “good” is a coun­try­sec­tor pair such as Chi­nese chem­i­cals and chem­i­cal prod­ucts. No­tice that a good can there­fore be dirty ei­ther be­cause it is pro­duced by a high-emis­sions coun­try, or be­cause it be­longs to a high-emis­sions sec­tor, or both. The vari­a­tion in trade shares across sec­tors alone is al­ready re­mark­able. While 24% of the dirt­i­est sec­tors’ out­put is trad­ed, only 12% of the clean­est sec­tors’ out­put is trad­ed. As one would ex­pect, the trade-em­bed­ded emis­sions come most­ly from trade in man­u­fac­tur­ing goods. All things con­sid­ered, the man­u­fac­ture of chem­i­cals and non-metal­lic min­er­al prod­ucts, ba­sic met­als and fab­ri­cat­ed met­al prod­ucts, and com­put­ers, elec­tron­ic and elec­tri­cal equip­ment are the top-3 con­trib­u­tors of trade-em­bed­ded emis­sions.

Source: WIOD 2013 Re­lease. This fig­ure dis­plays the dis­per­sion of sec­tor-coun­try pairs (ex­clud­ing all coun­try-en­er­gy, gas and wa­ter sup­ply pairs). On the x-axis, we re­port the share of out­put that is trad­ed in­ter­na­tion­al­ly for each sec­tor-coun­try pair. On the y-axis, we re­port the to­tal emis­sions (pro­duc­tion emis­sions in­clud­ing in­ter­me­di­ate stages) in tons of CO₂ of that good. The pos­i­tive slope of the fit­ted line shows that there is a pos­i­tive cor­re­la­tion be­tween how much a good is trad­ed and how dirty it is in terms of CO₂-equiv­a­lent green­house gas pro­duc­tion emis­sions.

In­ter­est­ing­ly, Joseph Shapiro (Uni­ver­si­ty of Cal­i­for­nia – Berke­ley) has re­cent­ly doc­u­ment­ed that im­port tar­iffs and non-tar­iff bar­ri­ers tend to be low­er for dirt­i­er in­dus­tries. He es­ti­mates that this trade pol­i­cy bias amounts to an im­plic­it sub­sidy for dirty sec­tors of $85 to $120 per ton of CO2 emit­ted.7 This re­sult is es­sen­tial­ly a com­bi­na­tion of the ob­ser­va­tions that up­stream in­dus­tries tend to be dirt­i­er and coun­tries tend to im­pose low­er tar­iffs on up­stream im­ports.

Fact 5 – International trade is tilted towards dirtier countries

Fig­ure 4 plots the CO2 emis­sions em­bed­ded in net im­ports by coun­try. A pos­i­tive val­ue means that the im­ports are dirt­i­er than the ex­ports and vice ver­sa.

In lev­els, the U.S., Japan, France, the U.K., Ger­many, and Italy are the largest net im­porters of CO2 emis­sions. In 2005, net im­ports of CO2 by these coun­tries ac­count­ed for 6.75% of to­tal pro­duc­tion emis­sions, or a third of the to­tal en­vi­ron­men­tal im­pact of trade. These emis­sions are over­whelm­ing­ly im­port­ed from Chi­na (al­most five times more than from any oth­er coun­try), Rus­sia, and In­dia, as well as Asian coun­tries (e.g. Sin­ga­pore, Thai­land, Ko­rea, Viet­nam), and Cen­tral and East­ern Eu­ro­pean coun­tries (e.g. the Czech Re­pub­lic, Poland, Bul­gar­ia, or Es­to­nia), which ranked as some of the dirt­i­est coun­tries in Fig­ure 2 in terms of emis­sion in­ten­si­ty. This pat­tern also emerges from our analy­sis of the World In­put Out­put Table, which pro­vides in­for­ma­tion on the glob­al sourc­ing choic­es of each sec­tor in each coun­try.

We il­lus­trate this fur­ther with a coarse but in­struc­tive thought ex­per­i­ment. We ask by how much glob­al pro­duc­tion emis­sions would de­cline if coun­tries were sourc­ing from green­er ori­gins, such as the EU, rather than from brown­er coun­tries, such as Chi­na? To this end, we sim­ply re­place the emis­sion in­ten­si­ties of Chi­nese sec­tors by the cor­re­spond­ing emis­sions in­ten­si­ties of EU coun­tries and com­pute the re­sult­ing de­crease in glob­al pro­duc­tion emis­sions.8 In this hy­po­thet­i­cal world, buy­ing green (and not nec­es­sar­i­ly lo­cal) gen­er­ates a de­cline of 6.3% of glob­al pro­duc­tion emis­sions.


In this Kühne Im­pact Se­ries, we have in­tro­duced the Kühne Cen­ter’s vi­sion of en­vi­ron­men­tal­ly sus­tain­able glob­al­iza­tion. Our main point is that buy­ing lo­cal is not the same as buy­ing green, be­cause in­ter­na­tion­al trade can al­low firms and house­holds to source prod­ucts from green­er ori­gins. We have ar­gued that the com­po­si­tion of trade is dis­tort­ed away from its ide­al. In par­tic­u­lar, there is (i) too much trade in rel­a­tive­ly ‘dirty’ goods, (ii) too much sourc­ing from rel­a­tive­ly ‘dirty’ coun­tries, and (iii) too much re­liance on rel­a­tive­ly ‘dirty’ modes of trans­port.

For us, this Kühne Im­pact Se­ries is only the first step in a re­search agen­da on sus­tain­able glob­al­iza­tion. Cli­mate change is one of the most press­ing chal­lenges fac­ing hu­man­i­ty and we be­lieve in­ter­na­tion­al trade can play a key role in fight­ing it. In fu­ture work, we aim to sub­stan­ti­ate this point and pro­pose new ways of op­er­a­tional­iz­ing it.

  1. Thomas Ne­me­cek and Joseph Poore (2018): “Re­duc­ing food’s en­vi­ron­men­tal im­pact through pro­duc­ers and con­sumers”. Sci­ence Vol. 360(6392) pp. 987– 992
  2. William D. Nord­haus (2017): “Re­vis­it­ing the So­cial Costs of Car­bon”. Pro­ceed­ings of the Na­tion­al Acad­e­my of Sci­ences of the Unit­ed States of Amer­i­ca.
  3. Be­cause elec­tric­i­ty, gas and wa­ter sup­ply is such a strong out­lier in our data, which sim­ply comes from the fact that it is the sec­tor in­volved in pro­duc­ing en­er­gy for the rest of our sec­tors, we will ex­clude it in most of our sub­se­quent analy­sis.
  4. Two things must be men­tioned about Chi­na: First, Chi­na’s ter­ri­to­r­i­al emis­sions only sur­passed those of the U.S. in 2006. Sec­ond, Chi­na is among the coun­tries that have made the largest progress in terms of en­er­gy mix in the last decade. The na­tion­al ob­jec­tive of 25% of to­tal en­er­gy use from so­lar and wind en­er­gy was reached in 2019.
  5. Un­for­tu­nate­ly, there are a lot of miss­ing val­ues for Cen­tral and South Asian coun­tries which lat­er ap­pear as large net ex­porters of emis­sions, such as Viet­nam or Thai­land.
  6. A. Cristea and D. Hum­mels and L. Puzel­lo and M. Avetysian (2013): “Trade and the Green­house Gas Emis­sions from In­ter­na­tion­al Freight Trans­port”. Jour­nal of En­vi­ron­men­tal Eco­nom­ics and Man­age­ment Vol. 65 pp. 153 – 173
  7. Joseph S. Shapiro (forth­com­ing): “The en­vi­ron­men­tal Bias of Trade Pol­i­cy”. Quar­ter­ly Jour­nal of Eco­nom­ics
  8. In this ex­er­cise, trans­porta­tion en­ters the in­put-out­put ta­ble as an in­put, but we do not ex­plic­it­ly ac­count for dif­fer­ent trans­porta­tion routes and short­er dis­tances, which would also change trans­porta­tion modes. In oth­er words, we ab­stract from the “trans­porta­tion part” of trade.

Ac­knowl­edge­ments: We thank Ja­cob Pichel­mann and Luca Poll for ex­cel­lent re­search as­sis­tance.


Mathilde Le Moigne

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


Ralph Ossa

Kühne Foundation Professor of International Trade


More Issues

Twitter logo