Intra-ndustry trade n agrcultural and food products: The case of Ukrane Inna Levkovych Insttut of Agrcultural Development n Central and Eastern Europe 2 Theodor-Leser Str. 06120-D, Halle/Saale levkovych@amo.de Oksana Luka Natonal agrcultural Unversty of Ukrane 15 Herojv Oborony Str. 03041 Kev, Ukrane oksana_luka@wm.nauu.kev.ua Paper prepared for presentaton at the 99 th semnar of the EAAE (European Assocaton of Agrcultural Economsts), The Future of Rural Europe n the Global Agr-Food System, Copenhagen, Denmark, August 24-27, 2005 Copyrght 2005 by Inna Levkovych and Oksana Luka. All rghts reserved. Readers may make verbatm copes of ths document for non-commercal purposes by any means, provded that ths copyrght notce appears on all such copes.
Abstract We present an analyss of agrcultural and food trade n Ukrane over the perod 1996-2002 focusng on dfferent aspects of ntra-ndustry trade. We estmated Grubel-Lloyd and margnal IIT ndexes to examne the relevance of ntra-ndustry trade between Ukrane and ts tradng partners and the changes n nature of trade flows over tme. The results ndcate that the major part of agro-food trade s of the nter-ndustry type, and thus a product of underlyng comparatve advantages. The low level of ndex seems to be typcal for prmary sectors n comparson to ndustres wth hgher product dfferentaton. Also margnal IIT appears to be low, the structure of the changes n agro-food trade flows between Ukrane and ts tradng partners durng the analyzed perod s shown to be predomnantly ntra ndustry. These results mply that the restructurng process, assocated wth trade lberalzaton of Ukrane on blateral and multlateral level, wll nduce hgher adjustment costs than n an ntra-ndustry trade envronment. Key words: JEL: ntra-ndustry trade, margnal ntra-ndustry trade, agro-food products, Ukrane F14, Q17 1 Introducton The emergence and growth of ntra-ndustry trade (IIT), defned as the smultaneous export and mport of goods wthn the same ndustry, has been one of the most mportant trends n world trade over the past few decades and has ganed ncreasng attenton n the economc lterature. A number of questons concernng IIT have been dscussed: from causes, sgnfcance, determnates of IIT, lnk to trade lberalzaton to conceptual and statstcal problems nvolved n tryng to measure IIT (Balassa, B., 1963; Grubel, H.G., Lloyd, P.J., 1975; Aquno, A., 1978; Aturupane, C. et. al., 1997; Blanes, J.V., Martn, K., 2000 etc.). The studes of Ruffn (1999), Greenaway, Mlner (2003) emphasze that wth IIT there exsts an addtonal potental source of gan ncreased varety, the exchange the scale economes and pro-competton effects. Intra-ndustry trade reduces the demands for protecton because n any ndustry there are both exports and mports, makng t dffcult to acheve unanmty among those demandng protecton (Ruffn, 1999). It s generally argued that ndustres wth hgh levels of IIT undergo less structural change - and less adjustment costs - n response to trade lberalzaton than ndustres wth low levels of IIT. The reason for ths s that t s easer to transfer and adapt resources wthn frms or ndustres than to swtch them from one ndustry to another (Kandogan, 2003a; Kösekahyaoglü, 2001; Krugman, 1981). Ths paper nvestgate the ntra-ndustry trade n Ukranan agrcultural and food trade. Our nterest to ths topc could be explaned by the followng reasons. Frst, agrculture and agrcultural trade play a sgnfcant role n the Ukranan economy. The share of agrculture n GDP s currently close to 15 percent. About 24 percent of populaton are economcally actve n ths sector. Agrcultural products and foodstuffs have accounted, on average, for 13.2 percent of the country s total exports (IER, 2003). Second, trade lberalzaton s a necessary step n the economc development of Ukrane. It presumes the possblty of a Free Trade Agreement mplementaton wth three CIS countres n the lne wth Sngle Economc Space and most mportant the expected WTO accesson. Trade theory suggests that removal of trade barrers can consderably mpact countres producton structure and ncome dstrbutons. As mentoned above adopton to nternatonal competton occurs wth more adjustment and hgher adjustment costs n ndustres characterzed by nter ndustry rather than ntra-ndustry trade. Therefore the am of ths paper s to examne the trend and extent of the ntra-ndustry trade n the agro-food sector n Ukrane over the perod 1996-2002. Ths analyss of the nature of trade provdes an nsght nto the potental consequences of further trade lberalzaton for the sector, namely expected structural adjustment costs. The paper s organzed as follows. Secton 2 consders the theoretcal framework of ntrandustry trade measurement. We present the tradtonal measure of IIT, the Grubel-Lloyd ndex, as Page 2 of 15
well as the more recent concept of margnal IIT. General patterns of Ukranan agro-food trade and ts developments over the concerned perod are analyzed n secton 3. In secton 4, we apply varous measures of IIT on a Ukranan trade data set and dscuss the derved emprcal results. Conclusons and possble drectons for further work n ths feld follow n the last secton. 2 Theoretcal framework of ntra-ndustry trade measurement 2.1 Standard measure of ntra-ndustry trade: the Grubel-Lloyd Index There are varous ndces for measurng ntra-ndustry trade, but the most wdely used s the GRUBEL-LLOYD (1975) ndex. In ths ndex, IIT for an ndustry s: (1) GLIIT = ( X + M) X ( X + M ) M 100, or GLIIT X M = 1 100, X + M where GLIIT s the Grubel-Lloyd ndex of IIT n ndustry, and X and M are, respectvely, the values of exports and mports n ndustry. The value of GLIIT ranges from 0 to 100. If there s no IIT (.e., ether X or M s zero) GLIIT takes the value of 0. If all trade s IIT (.e., X = M), GLIIT takes the value of 100. Grubel and Lloyd (1975) also suggested the followng formula, whch s a weghted average of the product ndces n (1) for the ndvdual product, wth the weghts beng based on the share of the specfc product n total trade j: (2) GLIIT j = (1 X ( X M + M )*100 ) The GLIIT ndex, as defned n equatons (1) for the ndvdual product and (2) for the weghted aggregate j, provdes nformaton on the composton of trade flows for each year. More specfcally, t presents the percentage of exports and mports of a smlar product as an ndcaton of the degree of external ntegraton. However, Greenaway and Mlner (2003) ponted out that the Grubel-Lloyd ndex s far from uncontroversal. One of the man pont of contenton s aggregaton. The aggregaton problem has two dmensons: geographcal and ndustral. The geographcal dmenson underscores the problem of a multlateral approach, snce the IIT measure may be upward-based at a multlateral level due to export of a product to one trade partner, and mport from another trade partner of the same product. If we seek to mnmze the bases due to geographcal aggregaton, blateral trade flows are preferred over multlateral trade flows (Fontagné and Freudenberg, 1997; Gullstrand 2002). In recent emprcal studes both approaches are used. The ndustral dmenson of the aggregaton problem underscores the mportance of calculatng IIT at a rather low aggregaton level. If we aggregate two sub-ndustres wth opposte trade-mbalance sgns at each sub-level, IIT becomes upward-based. If these mbalances are due to nter-ndustry specalzaton, a part of IIT conssts of trade that could be explaned by tradtonal trade theory and comparatve advantages. There are two ways to solve the ndustral dmenson of the aggregaton problem, thus allowng our IIT ndces to preclude flows that could be explaned by comparatve advantages. One s to reclassfy trade statstcs, and the other s to calculate a weghted average IIT ndex based on dsaggregated data (Greenaway and Mlner, 1983; 1986). The practcal problem and the lack of a Page 3 of 15
common reclassfcaton standard of the former approach lead us to the latter, whch s the tradeweghted average of sub-ndustry IIT levels. Ths wll mnmze the aggregaton problem, snce t does not cancel out the opposte trade-mbalance sgns at a sub-ndustry level. In order to mnmze the aggregaton problem, we use a 6-dgt level of the Harmonzed System nomenclature and then calculate the weghted average of sub-ndustry IIT levels across product groups for separate tradng partners, countres groups and the world as a whole over the observed perod. Trade data for measurng ntra-ndustry trade were obtaned from the COMTRADE database accordng to HS 1992. 2.2 The measures of margnal ntra-ndustry trade It s generally assumed that adjustment costs assocated wth trade lberalzaton may dffer dependng upon whether emergng trade can be classfed as nter- or ntra-ndustry. Whereas the former mples a reallocaton between ndustres, the latter mples a reallocaton wthn ndustres and, to the extent that ndustres are defned n terms of the producton space wthn whch factor substtuton can be classfed as a relatvely low cost, has a greater potental for lower adjustment costs. The lmtatons of usng changes n the standard GL ndex to capture the dynamcs of changes n IIT are wdely recognzed. Adjustment process should be analyzed usng ndcators based on margnal trade flows, because adjustment s a strctly dynamc process; knowledge of changes n trade flows s requred n order to nfer relable conclusons. By way of contrast, t would be napproprate to compare statc measures at dfferent ponts n tme. For nstance, an ncrease n the IIT measures by the GL ndex at two ponts n tme mght suggest an ntra-ndustry adjustment, although ths could be due to an ncrease n the export of an mport-orented ndustry. Some smple and now wdely-used measures of MIIT were developed by Brülhart (1994). The Brülhart A ndex s a transposton of GL ndex to trade changes: (3) MIIT = A = 1 (3) X X ( ) ( M M ) X t t X t n t n + M t t M t n t n *100 where n stands for the number of years consttutng the relevant adjustment perod. Ths s also wrtten as: (4) A = 1 ΔX ΔM ΔX + ΔM *100 The A ndex, lke the GL ndex, vares between 0 and 100, where 0 ndcates margnal trade n the partcular ndustry to be completely of the nter-ndustry type, and 1 represents margnal trade to be entrely of the nter-ndustry type. The A ndex shares most of the statstcal propertes of the GL ndex, a comprehensve descrpton of whch s provded n Greenaway and Mlner (1986). When a country s exports and mports n a partcular ndustry grow or shrnk at a smlar absolute rate (hgh A), trade-nduced adjustment s lkely to occur at the ntra-ndustry level, whle the overall performance of the ndustry s determned by factors whch tend to affect all countres symmetrcally, such as global demand or technology changes. The A ndex therefore captures the degree of crosscountry symmetry n trade changes. Conversely, where a country s exports and mports n a partcular ndustry show dvergng trends (low A), both the trade-nduced asymmetrcal forces for the geographcal nter-ndustry adjustment and the exogenous factors determnng the fate of the ndustry Page 4 of 15
across all countres are lkely to be relevant. A can be summed, as can the GL ndex, across ndustres havng the same level of statstcal dsaggregaton by applyng the followng formula for a weghted average: (5) A tot = k 1 w A, ΔX + ΔM where w ( ). = k 1 (6) = ΔX + ΔM Thus, Brülhart s dynamc ndex, A, rather than the standard Grubel-Lloyd, s the approprate ndcator of the role played by ntra-ndustry trade durng the type of adjustment process evoked by trade lberalzaton. The A ndex (lke the GL ndex) can provde results whch are relevant for multlateral studes by relatng to overall adjustment pressures. Yet t does not contan any nformaton as to the relatve trade performance of ndustres n partcular countres. In terms of net exports, nter-ndustry adjustment can reflect trade specalzaton nto or out of partcular ndustres. Hence, Brülhart (1994) suggested the followng ndex: (7) B = ΔX ΔM ΔX + ΔM *100 where B = 1 A. Ths coeffcent can take values rangng between 100 and 100. It s two-dmensonal, contanng nformaton about both the proporton of MIIT and country-specfc sectoral performance. Frst, the closer B s to zero, the hgher s MIIT, whereas at both 100 and 100 t represents margnal trade as beng entrely of the nter-ndustry type. Second, sectoral performance s defned as the change n exports and mports n relaton to each other. When B > 0, ΔX was > ΔM. The opposte holds for B < 0. Unlke the A ndex, B cannot be meanngfully aggregated across ndustres. Therefore, B cannot be used for summary statstcs resultng from calculatons on a dsaggregated level. Its applcablty s thus confned to the ndustry-by-ndustry assessment of MIIT and performance. Thus, measures of MIIT are desgned to complement the GL ndex n analyses of trade change and adjustment. A pror reasonng suggests that MIIT relates more drectly to structural adjustment than IIT, snce hgh MIIT entals relatvely low factor re-allocaton between ndustres. 3 The general pattern of Ukranan agro-food trade Agro-food products hold a sgnfcant share of Ukrane s total merchandse exports. Durng the past eght years, agrcultural products and foodstuffs have accounted, on average, for 13.2 percent of the country s total exports. The only exported commodty group that s larger s base metals and ther products. However, the share of agro-food products n total merchandse exports has fluctuated from 21.2 percent n 1996 to only 9.5 percent n 2000 (Fgure 1). The share of correspondng mports s sgnfcantly lower, and durng 1995-2002 consttuted, on average, 6.5 percent of total merchandse mports. Ukrane tradtonally has a surplus n agrcultural and food trade, as agro-food exports exceed mports. The postve balance has fluctuated between USD 0.3 bllon n 1998 and USD 1.6 bllon n Page 5 of 15
1996. A general reducton n world trade turnover, whch occurred owng to an ntensfcaton of the fnancal crss, affected the development of Ukrane s foregn trade after 1997. The 1998 fnancal crss n Russa (the man tradng partner of Ukrane) also had a substantal mpact on foregn trade. In 2001, after a steady downfall of exports from 1997-2000, Ukrane exhbted postve tendences n agro-food trade; these tendences reman (Fgure 1). Favourable prces on the world market and a comparatvely hgh domestc Value, mllon USD 3500 3000 2500 2000 1500 1000 500 0 1995 1996 1997 1998 1999 2000 2001 2002 agro-food export agro-food mport share of agro-food export n total merchandze export share of agro-food mport n total merchandze mport 25 % 20 15 10 5 0 Source: Own calculatons based on data provded by State Statstcs Commttee of Ukrane. Fgure 1. Ukranan agro-food trade, 1995-2002. supply of the man types of agrcultural products were the man reasons for augmented agro-food export. Ukranan agro-food export s characterzed by hgh concentratons on a lmted number of product groups. The leadng postons n the commodty structure of Ukranan agro-food exports consst of cereals, vegetable ols (manly sunflower ol), meat (prncpally beef), mlk products and ol seeds (Fgure 2). On the other hand, agro-food mports to Ukrane are more dversfed than correspondng exports. Tobacco, sugar and sugar confectonery (manly raw sugar from sugar cane), fats and ols, cocoa and cocoa products, fsh and mscellaneous edble product n recent years have accounted for the largest share of agro-food mports. A substantal share of sugar n Ukrane s agro-food mports n the reported perod can be explaned by the low compettveness of the domestc sugar refnng ndustry and a number of laws passed n 2000-2001 whch set quotas for raw sugar mports at prvleged mport duty rates. An analyss of the geographc structure of the agro-food trade ndcates that Ukrane has been gradually redrectng ts export orentaton from CIS countres to non-cis countres, manly towards EU-15 and developng countes. But n general, CIS countres account for the largest share of Ukranan aggregated exports of agrcultural and food products. Page 6 of 15
30 25 20 n % 15 10 5 0 Cereals Vegetable ols Meat Mlk products Ol seeds Sugar Cocoa Beverages and sprts Tobacco Fats and ols Fsh Msc. preparatons Frut and nuts Coffee, tea, spces Others Exports Imports Source: Own calculatons based on data provded by State Statstcs Commttee of Ukrane. Fgure 2. Commodty composton of Ukranan agro-food exports and mports, 2000-2002, on average. The value of agro-food exports to CIS countres fell sharply n 1997 and contnued to decrease n 1998 (see Fgure 3). Therefore, the share of these countres group n total Ukranan agro-food exports shrank from 74.2 percent n 1996 to 44.8 percent n 1998. In 1999-2002, the value of agrofood exports to CIS countres remaned steady, but the share of the group n total concerned wth exports decreased to 37.5 percent n 2002 due to ncreased exports to other regons. 3000 2500 mllon USD 2000 1500 1000 500 0 1996 1997 1998 1999 2000 2001 2002 CIS countres Baltc states CEE countres EU-15 OECD countres Non-OECD countres Source: Own calculatons based on data provded by State Statstcs Commttee of Ukrane. Fgure 3. Geographc structure of Ukranan agro-food export, 1996-2002. Page 7 of 15
On the other hand, durng the past few years there has been a tendency for the value of agro-food exports to EU-15 and developng countres (.e. non-oecd countres) to ncrease. In 2002, the share of Ukranan total agro-food exports to these countres was 23.8 percent and 24.2 percent, respectvely. The evoluton of the geographc structure of Ukranan agro-food exports s assocated wth changes n ts commodty structure. The CIS countres were the major consumers of processed food, but shrnkage of ths market mples a reorentaton to new tradng partners that demand manly raw agrcultural products 1400 1200 mllon USD 1000 800 600 400 200 0 1996 1997 1998 1999 2000 2001 2002 CIS countres Baltc states CEEC countres EU-15 OECD countres Non-OECD countres Source: Own calculatons based on data provded by State Statstcs Commttee of Ukrane Fgure 4. Geographc structure of Ukranan agro-food mports, 1996-2002. The geographc orgn of agro-food mports to Ukrane are n more stable (Fgure 4). The man mporters of agrcultural products and foodstuff to Ukrane are non-oecd countres and the EU, whch n 2002 accounted for 32.6 percent and 25.4 percent of total agro-food mports, respectvely. Also the sgnfcance of CIS countres (prmarly Russa) has ncreased snce 1999. 4 Ukranan ntra-ndustry trade n agro-food products: emprcal results 4.1 The tradtonal measure of IIT In ths study we focus on Ukranan agro-food trade over the perod 1996-2002. GL ndces of IIT for Ukranan agro-food trade were calculated a) by commodty groups, b) by all tradng partners (the world) and wth respect to the followng regonal specfcaton: CIS, Baltc states, CEE countres, EU- 15, OECD countres and non-oecd countres. Moreover, to examne the possble mpact of geographcal aggregaton, we have calculated GLIIT ndces for Ukranan trade n agro-food products n two ways: 1) based on blateral trade flows wth each tradng partner, whch were aggregated to the group level, and 2) based on multlateral trade flows on the group level. Results (see Table 1) confrm the senstvty of GL ndces to the choce of aggregaton level. It can be seen that IIT measures are upward-based at a multlateral level due to the export of a product to one trade partner and mport from another trade partner of the same product. Nevertheless, n the followng, we often use multlateral levels to examne IIT n more detal due to the extreme complexty of calculatons on the blateral level and subsequent aggregaton to the group level. Results presented n Table 1 reveal that generally, ntra-ndustry trade of agro-food products n Ukrane s very low. The majorty of total trade n the analyzed sector, 86.4 percent, s of the nterndustry type. The low level of the GLIIT ndex wth OECD and non-oecd countres (1.89 and 1.34 Page 8 of 15
percent, respectvely) reflects the sgnfcant dfference n the structure of ther economes compared to Ukrane. In addton, the nter-ndustry nature of trade can be explaned by the contnuaton of some trade constrants. On the mport sde, the lberalzaton of agro-food trade n Ukrane has been less substantal and more gradual than n other sectors, mplyng hgh levels of mport tarffs on food. On the export sde, Ukranan agrcultural and food products tend to lack nternatonal compettveness, partcularly to the EU. The GL ndces tend to be hgher wth CIS and CEE countres, although the tendences of IIT evoluton wth these groups are the opposte: there s an upward trend n IIT wth CIS and a downward trend for the CEEC. The exstng free trade area between CIS countres, the expanson of ther mports n Ukrane, accompaned by more stable correspondng exports n the frst case and an nterrupton n communcatons between Ukrane and the CEEC, as well as the orentaton of CEE countres to the EU market, could explan the above-mentoned tendences. Moreover the hgh level of IIT s usually attrbuted to a number of country-specfc factors, ncludng ts close geographcal proxmty, smlar level of per capta ncome, smlar level of development, smlar consumer tastes, language, culture, nsttutonal, poltcal and transport lnks. Table 1. Grubel-Lloyd ndces of ntra-ndustry trade n agro-food products between Ukrane and ts tradng partners, 1996-2002. Group 1996 1997 1998 1999 2000 2001 2002 1996-2002, n average CIS Total 2.40 4.30 7.04 5.71 11.78 14.41 14.75 8.63 Aggregated 1.24 1.25 2.74 2.90 5.26 8.54 6.91 4.12 Baltc states Total 2.62 3.14 2.98 3.57 3.15 4.28 3.56 3.33 Aggregated 1.79 1.23 1.64 2.07 1.38 2.31 2.05 1.78 CEE countres Total 11.17 10.29 9.03 12.90 8.39 5.16 5.13 8.87 Aggregated 6.35 5.60 3.02 6.61 2.16 2.53 1.75 4.00 EU-15 Total 5.53 6.42 4.69 5.09 7.21 5.75 3.33 5.43 Aggregated 3.63 3.66 2.95 2.41 2.87 1.94 1.20 2.67 OECD countres Total 2.21 1.21 1.55 1.30 2.37 2.30 2.27 1.89 Aggregated 0.68 0.84 0.54 0.62 1.77 0.72 1.00 0.88 Non-OECD countres Total 1.40 0.70 4.67 0.41 0.83 0.71 0.69 1.34 Aggregated 0.04 0.05 0.03 0.07 0.11 0.34 0.40 0.15 World Total 15.26 15.51 14.83 12.17 14.95 13.17 10.07 13.71 Aggregated 1.78 1.83 1.92 2.01 3.11 3.88 2.79 2.47 Source: Own calculatons based on COMTRADE Database. Table 2 shows the evoluton of GL ndces calculated by commodty groups. We also computed weghted average ndces across agrcultural products (HS groups 1-15), food (HS groups 16-24) and total agro-food trade (HS groups 1-24) usng as a weght the share of each ndustry s trade of total concerned trade. As expected, IIT ndces are hgher for food (22 percent on average) than for agrcultural products (about 8 percent). The relatve hgh magntude of IIT exsts n preparatons of cereals, flour, starch or mlk (HS 19), preparatons of vegetables, frut, nuts (HS 20), beverages, sprts and vnegar (HS 22), and tobacco (HS 24) and also n fsh, products of anmal orgn. Ths confrms the suggeston that IIT s more common n sectors wth sgnfcant producton dfferentaton, and s nsgnfcant n sectors wth standardzed products, such as natural resources and agrcultural products, where most trade s nter-ndustry (Kandogan 2003b). It s also mportant to note that durng the reported perod, average ndces of IIT for food were rather stable, whle at the same tme, ndces for Page 9 of 15
Table 2. Intra-ndustry trade by commodty groups between Ukrane and all tradng partners (the world), 1996-2002. HS code 1996 1997 1998 1999 2000 2001 2002 Average weghted 01- lve anmals 0.54 1.99 3.98 9.55 5.79 11.02 11.16 3.64 02- meat and edble meat offal 1.97 3.82 4.13 7.18 3.55 3.82 3.82 3.85 03- fsh and fsh products 34.79 37.84 19.69 9.79 37.15 27.11 26.07 27.57 04- dary, eggs, honey etc. 9.95 17.80 13.20 26.19 6.21 4.11 4.95 9.38 05- product of anmal orgn 35.79 45.10 28.73 48.58 20.45 11.93 10.75 27.37 06- lve trees, cut flowers etc. 18.10 8.39 13.13 5.47 4.90 5.50 3.87 7.05 07- edble vegetables etc. 11.42 4.74 5.33 10.60 9.43 10.56 6.77 8.22 08- edble fruts, nuts etc. 8.35 3.49 2.61 2.20 1.93 2.36 4.03 3.77 09- coffee, tee, mate etc. 9.18 8.92 5.24 2.11 2.23 1.88 2.94 4.12 10- cereals 5.34 3.87 1.09 1.73 32.89 12.19 0.99 6.07 11- mllng products, malt, starches 7.01 7.86 6.22 3.52 5.78 3.90 9.35 6.52 12- ol seeds, seeds etc. 13.09 13.05 12.62 14.96 11.70 11.47 39.29 14.02 13- lac, gums, resns etc. 22.68 15.73 16.40 14.66 13.00 8.09 3.97 14.72 14- vegetable platng materals 7.39 3.16 8.39 16.73 29.02 14.04 3.12 8.83 15- anmal or vegetable fats etc. 9.09 10.00 17.51 7.06 0.80 1.22 1.50 5.47 16- preparatons of meat, of fsh 14.64 7.38 24.41 27.29 18.12 57.69 49.72 21.24 17- sugars and sugar confectonery 16.51 8.77 17.82 8.63 14.20 11.08 24.61 14.92 18- cacao and cacao preparatons 26.05 12.62 16.51 12.82 12.77 13.97 7.90 13.30 19- preparatons of cereals, flour 51.23 42.03 30.28 30.06 23.86 28.77 29.03 36.45 20- preparatons of vegetables etc. 41.92 32.85 28.17 25.18 21.40 24.28 21.01 28.44 21- mscellaneous edble preparaton 16.81 18.20 11.38 8.51 10.69 21.36 18.06 15.96 22- beverages, sprts and vnegar 19.07 35.20 32.71 23.39 25.57 40.64 31.62 26.85 23- resdues from food ndustry etc. 15.41 43.32 33.06 38.88 17.05 10.57 10.48 17.92 24- tobacco etc. 39.56 27.77 25.69 38.51 31.62 32.92 26.79 31.58 01-15 total agrcultural products 9.44 11.34 9.53 6.96 11.67 7.82 4.28 8.43 16-24 total food products 21.06 20.81 24.10 23.04 20.57 22.98 21.38 21.78 01-24 total agro-food products 15.26 15.51 14.83 12.17 14.95 13.17 10.07 13.68 Source: Own calculatons based on COMTRADE Database. agrcultural products fluctuated sgnfcantly, from 4.28 percent n 2002 to 11.67 percent n 2000. These results correspond to the commodty structure of Ukranan trade n agrcultural products. As mentoned above, Ukrane has a hgh concentraton of agro-food exports on a lmted number of products, wth a prevalng share of cereals, whch s caused by the country s comparatve advantage. In 2002, the share of cereals n Ukranan agro-food exports was the hghest compared to other examned years (41.8 percent of total agro-food exports). Consequently, the level of IIT was the lowest. In 2000, the stuaton was the opposte: Ukrane had the lowest level of gran exports (wth sgnfcant correspondng mports) and, as a result, the hghest level of IIT n agrcultural products. Altogether, IIT ndces for agro-food products were low and fluctuated from 10.07 n 2002 to 15.51 n 1997. Thus, there s no strct trend n the IIT evoluton, although GL ndces tend to be lower n recent years than n the frst sample years. Ths s contrary to the development of agro-food trade n CEE countres, where there s a dstnct upward trend n IIT (Bojnec 2001; Fertö and Hubbard 2001, 2002; Bojnec and Hartmann 2004). We also calculated GL ndces by commodty group based on multlateral trade flows on the specfed groups level. It s nterestng that IIT wth CIS countres, Baltc States and OECD countres corresponds to the overall tendency to a hgher level of GL ndces for food, but agro-food trade wth the EU-15 s characterzed by hgher IIT n agrcultural products. IIT trade n agrcultural and food products wth CEE countres s more or less equal, and t s nsgnfcant wth non-oecd countres. 4.2 Margnal ntra-ndustry trade So far the analyss has been based on ndces whch measure the extent of IIT as a proporton of Page 10 of 15
total trade at a gven pont of tme. But changes n the GL ndex may not capture potental adjustment costs, and measures of margnal ntra-ndustry trade (MIIT) can, therefore, be used to complement tradtonal IIT analyss. We have calculated A ndces for agro-food products from HS 6-dgt trade fgures over the perods 1996-1999 and 1999-2002 based on multlateral trade flows at the specfed groups level (Table 3). Table 3. Margnal ntra-ndustry trade n agro-food products n Ukrane, by trade partners, 1996-2002 (A ndces). Countres group 1996-1999 1999-2002 CIS countres 1.7 9.9 Baltc states 0.4 0.9 CEE countres 7.5 4.5 EU-15 4.7 1.7 OECD countres 2.0 1.4 Non-OECD countres 3.8 0.9 World total 9.6 7.6 Source: Own calculatons based on COMTRADE Database. The hghest share of margnal IIT s revealed for CIS countres over the perod 1999-2002. For other trade partners was the level of margnal IIT less relevant over both perod (exceptng CEE countres over perod 1996-1999). The generally low level of A ndces (close to zero) ndcates that most of change occurrng n trade flows has been nter-ndustry by nature and therefore very lkely have nduced hgh adjustment costs. If we look at average A ndces across HS 2-dgt sectors, we fnd that MIIT patterns resemble those of IIT n so far as food-processng ndustres exhbt consstently hgher average ndex value than prmary sectors. The hghest levels of MIIT for the perod 1996-1999 are the one s for fsh, preparatons of vegetables, fruts and nuts, and preparatons of cereals, flour, starch or mlk; for the perod 1999-2002 the hghest A levels are those for sugar and sugars confectonery, tobacco and preparatons of cereals, flour, starch or mlk sectons. Table 4 summarzes the results of the calculatons of the Brülhalt s (1994) B ndex whch measures the sectoral performance and MIIT. We classfy 6-dgt sectors of Ukranan agrcultural and food processng ndustres nto four groups accordng to the sze and the sgn of ths ndex. The frst group ncludes products where 100 B < 50 and refers to products wth bad performance, where margnal trade s manly of nter-ndustry type. The second group ( 50 B < 0) ncludes products where margnal IIT domnates, and the negatve sgn of B ndex ndcates a weak performance of these products. The thrd (0 B <50) and forth (50 B < 1) groups cover those products that reflect a good trade performance, however, whle the thrd group characterzes products where margnal IIT prevals, the vce versa holds for the forth group. Table 4. Allocaton of B ndces of margnal IIT for Ukranan agro-food trade by trade partners, 1996-2002. 1996-1999 1999-2002 I II III IV I II III IV CIS 302 16 13 147 214 20 16 199 Baltc states 180 3 6 116 107 3 6 154 CEE countres 169 7 11 276 177 11 14 198 EU-15 189 8 12 321 235 10 8 244 OECD countres 163 4 3 266 180 8 15 204 Non-OECD countres 268 15 11 251 248 2 14 268 World total 280 38 39 285 311 26 29 262 Source: Own calculatons based on COMTRADE Database. Page 11 of 15
The results ndcate that margnal trade n most Ukranan agro-food products s of the nterndustry type (the majorty of products belong to frst and fourth groups). Ths apples to the trade relatons wth the world and all tradng partners. Lookng at the trade performance measured by B ndces, we fnd that the majorty of products dsplayed negatve B values (belongng to 1 and II groups) vs-à-vs the Former Sovet Unon countres over the both examned perods ndcatng that there have been more agrcultural and food products wth a weak performance. In the trade wth European and OECD countres, however, a narrow majorty of ndustres dsplayed postve B ndces. These results document declnng trade compettveness n tradtonal markets (CIS countres and Baltc states) and a reorentaton of Ukranan agro-food exports towards new tradng partners. The trade performance of Ukranan agro-food sector deterorated over second sub-perod, f total trade wth the world s consdered. It reveals also the level of B rato, calculated as relaton between number of ndustres wth postve Bs relatve to those wth negatve Bs (Fgure 6). B rato 1.8 1.6 1.4 1.2 1 0.8 0.6 0.4 0.2 0 CIS Baltc states CEEC EU-15 OECD Non-OECD countres 1996-1999 1999-2002 World total Source: Own calculatons. Fgure 6. Ukranan agro-food trade performance, 1996-2002. Lookng at the regonal ds-aggregaton there were opposte tendences n trade performance wth dfferent groups: B ratos n 1999-2002 compared to the 1996-1999 ncreased for CIS countres, Baltc States and non-oecd countres, and decreased for the EU-14, CEEC and OECD countres. It ndcates that the Ukranan trade poston on western markets (CEE countres, European Unon and other OECD countres) mproved consderably durng the frst analyzed perod. But over 1999-2002, although trade performance remaned postve, a contnued upturn was not observed. In respect to trade wth CIS countres the Ukranan poston on the concerned market durng the second sub-perod strengthened to some extent, however, trade performance was stll negatve, as B rato stayed at a level lower than one. The reducton of B rato over 1999-2002 reveals the decrease of the Ukranan trade poston on world agrcultural markets. 4 Conclusons The process of economc transformaton n Ukrane, as n other Former Sovet Countres s characterzed by changng trade patterns n agrcultural and food products. Over the analyzed perod, Ukrane has remaned a net exporter of agro-food products, but the total volume, commodty and geographcal structure of trade flows have altered sgnfcantly. The changes n commodty structure of agro-food exports nclude a shft away from processed products towards raw materals, manly cereals and vegetable ols, followed by dary and meat products. At the same tme, the mport structure Page 12 of 15
has remaned rather steady and s more dversfed than Ukranan agro-food exports. Major changes also occurred n the drecton of trade flows, wth an ncreasng mportance of both the EU and developng countres. Despte the fact that n 2002, agro-food exports to CIS countres decreased by more than 50 percent compared to 1996, these countres (prmarly Russa) absorb the largest share of Ukranan exports. Furthermore, the Ukranan export poston on ths market has recovered n recent years, manly due to economc recovery n Russa after the crss of 1998. Thus, the economc stuaton n CIS countres has a sgnfcant mpact on Ukranan export performance. Further examnaton of Ukranan agro-food trade flows usng an ntra-ndustry trade approach enables more detaled analyss of the structural changes n trade flows and the level of external ntegraton. The analyss reveals, that ntra-ndustry trade n Ukranan agrcultural and food trade had a lttle relevance over 1996-2002. More than 86% of trade n these products s of the nter-ndustry type and thus caused by underlyng comparatve advantage of the country. The explanaton for the low level of external ntegraton n the Ukranan agro-food sector s that the protectons scale has remaned hgh over perod analyzed and thus has hndered trade n general and ntra-ndustry n partcular (von Cramon-Taubadel, S., Zorya, S., 2001). But ntra-ndustry values for dstnct tradng partners and specfed commodty groups dffer sgnfcantly and exhbt hgh varablty over tme. Among dfferent countres groups, hgher levels of IIT are found between Ukrane and such tradng partners as CIS and CEE countres. The plausble explanaton of relatve hgh ntegraton levels between Ukrane and the above-mentoned countres n terms of IIT magntude seems to be the smlartes of per capta ncome, level of economc development, taste overlap, cultural, poltcal and transport lnks, etc. Moreover, n the case of CIS countres, the efforts of lberalsaton n lne wth CIS Agreement could be a factor whch ncreased the extent of IIT (as IIT s postvely correlated wth trade ntensty and lberalzaton agreements). Among specfc product groups, the values of the GLIIT ndex are hghest for sub-sectors wth hgher product dfferentaton, such as processed foods, whch corresponds to the theory and prevous emprcal studes. The same tendency was found by analyzng margnal IIT: food products nvolvng a greater degree of processng show hgher ndex value than prmary sub-sectors. The low shares of GLIIT and margnal IIT ndexes ndcate that trade-nduced reallocaton of producton factors has occurred between sectors rather than wthn sectors, mples hgh adjustment costs. The Ukranan agro-food sector faces hgh pressure for adjustment n the course of further economcal ntegraton the access of Ukrane n WTO. But t s a necessary step and the success wll depend on the consstency and depth of reform n order to mprove economcal envronment. References Aquno, A. (1978): Intra-Industry Trade and Inter-Industry Specalzaton as Concurrent Source of Internatonal Trade n Manufactures. Weltwrtschaftlches Archv 114: 275-296. Aturupane, C., Djankow, S., Hoekman, B. (1997): Determnants of Intra-Industry Trade between East and West Europe. CEPR Dscusson Paper 1721, Centre for Economc Polcy Research (CEPR). London. Balassa, B. (1963): An Emprcal Demonstraton of Classcal Comparatve Cost Theory. Revew of Economcs and statstcs 4: 231-238. Blanes, J.V., Martn, K. (2000): The Nature and Causes of Intra-Industry Trade: Back to the Comparatve Advantage Explanaton? Weltwrtschaftlches Archv 136 (3): 423-441. Bojnec, Š. (2001): Patterns of Intra-Industry Trade n Agrcultural and Food Products Durng Transton. Eastern European Economcs 39 (1): 61-89. Page 13 of 15
Bojnec, Š. Hartmann, M. (2004): Agrcultural and Food Trade n Central and Eastern Europe: The Case of Slovenan Intra-Industry Trade and Induced Structural Adjustment Costs. IAMO Dscusson Paper 65, Insttute of Agrcultural Development n Central and Eastern Europe (IAMO). Halle (Saale). Brülhart, M. (1994): Margnal Intra-Industry Trade: Measurement and the Relevance for the Pattern of Industral Adjustment. Weltwrtschaftlches Archv 130 (3): 600-613. Brülhart, M. (1999): Margnal Intra-ndustry trade and Trade-Induced Adjustment: A Survey. In Brülhart, Hne, M.R.C. (eds.), Intra-Industry Trade and Adjustment. The European Experence. Basngstoke, Hampshre: Macmllan, 36-69. Burakovsky, I. (ed.) (2004): Ukrane s WTO Accesson: Challenge for Domestc Economc Reforms. Helderberg, Physca-Verlag. COMTRADE: Unted Natons Commodty Trade Statstcs Database, http://unstats.un.org/unsd/comtrade/default.aspx, July 2004. Djankow, S., Hoekman, B. (1996): Intra-Industry Trade, Foregn Drect Investment and Reorentaton of East European Exports. CEPR Dscusson Paper 1377, Centre for Economc Polcy Research (CEPR). London. Fertö, I, Hubbard, L.J. (2001): Intra-Industry Trade n Agr-Food Products between Hungary and the EU. Dscusson Paper 2001/6, Insttute of Economcs Hungaran Academy of Scences. Budapest. Fontagné, L., Freudenberg, M. (1997): Intra-Industry Trade: Methodologcal Issues Reconsdered. Document de Traval 97-01, Centre d Etudes Prospectves et d Informatons Internatonales (CEPII). Pars. Gray, P. (1979): Intra-ndustry Trade: The Effects of Dfferent Levels of Data Aggregaton. In Gersch, H. (ed.): On the economcs of ntra-ndustry trade: Symposum 1978, Insttut für Weltwrtschaft an der Unverstät Kel. Tübngen: Mohr: 87-110. Greenaway, D., Mlner, C. (1981): Trade Imbalance Effects and the Measurement of Intra-Industry Trade. Weltwrtschaftlches Archv 117 (4): 756-762. Greenaway, D., Mlner, C. (2003): What Have We Learned from a Generaton s Research on Intra- Industry Trade? GEP Research Paper 2002/44, Leverhulme Centre for Research on Globalzaton and Economc Polcy. Greenaway, D., Mlner, C. (1986): The Economcs of Intra-Industry Trade. Oxford: Basl Blackwell. Grubel, H.G., Lloyd, P.J. (1975): Intra-Industry Trade: The Theory and Measurement of Internatonal Trade n Dfferentated Products. London and New York: Wley. Gullstrand J. (2002): Does the Measurement of ntra-industry Trade Matter? Weltwrtschaftlches Archv 138 (2): 317-339. Hamlton, C., Knest, P. (1991): Trade Lberalzaton, Structural Adjustment and Intra-Industry Trade. Weltwrtschaftlches Archv 127 (2): 356-367. IER (2003): Development of Domestc Markets n Ukrane: Welfare through Competton. Insttute for Economc Research and Polcy Consultng, Kev. Kandogan, Y. (2003a): Reconsderng the adjustment Costs of the Europe Agreements. Appled Economcs Letters 10 (2): 63-68. Page 14 of 15
Kandogan, Y. (2003b): Intra-Industry Trade of Transton Countres: Trends and Determnants. Wllam Davdson Workng Paper No. 566, May. Krugman, P. (1981): Intra-ndustry specalsaton and the gans from trade, Journal of Poltcal Economy, 89, pp 959-73. von Cramon-Taubadel, S., Zorya, S. (2001): WTO Accesson and Agrcultural Polcy n Ukrane. In von Cramon-Taubadel, S., Zorya, S., Streve, L. (eds.): Polces and Agrcultural Development n Ukrane. Aachen: Schaker, 155-176. Page 15 of 15