On the Duraton of Comparatve Advantages of Top European Wne Producers Jeremás Máté BALOGH, Attla JÁMBOR Abstract Corvnus Unversty of Budapest, HU eremas.balogh@gmal.com; attla.ambor@un-corvnus.hu In the recent two decades, the mportance of the European tradtonal wne exporters have been permanently declnng n the world wne market, therefore, t s mportant to nvestgate how the wne export compettveness of these countres was changed over tme. Snce the semnal work of Balassa, a vast amount of lterature s dedcated to the analyses of revealed comparatve advantages of global trade, whle revealed comparatve advantages n the wne ndustry are scarcely nvestgated. The am of our paper s to analyse the revealed comparatve advantages of European wne trade as well as the duraton and stablty of RSCA ndces by the help of lagged ndces and Kaplan-Meer survval functon. Our paper employs European wne trade data from 2000 to 2013, for sxteen countres as the bggest producers responsble for 70% of EU total wne exports such as Austra, Bulgara, Croata, Cyprus, Czech Republc, France, Germany, Greece, Hungary, Italy, Malta, Portugal, Romana, Slovak Republc, Slovena and Span. Furthermore, the paper concentrates on the B ndex and ts transformatons, the RSCA and NRCA ndces. Our results suggest that Bulgara, Cyprus, France, Greece, Italy, Portugal and Span had revealed comparatve advantages on the world market out of the sxteen top European wne producers between 2000 and 2013. However, these advantages seem to have weakened for the maorty of countres as evdent from the dstrbuton of comparatve advantages ndces as well as from the stablty tests. The only exceptons were France and Italy wth stable comparatve advantages. Keywords: revealed comparatve advantages, Europe, top wne producers, duraton analyss JEL-Code: Q17, F13, F14 1. Introducton In the recent two decades, the mportance of the European tradtonal wne exporters have been permanently declnng n the world wne market, therefore, t s mportant to nvestgate how the wne export compettveness of these countres was changed over tme. Snce the semnal work of Balassa (1965), a vast amount of lterature s dedcated to the analyses of revealed comparatve advantages of global trade. Despte the apparent mportance of the topc, most studes have focused on ndustral products, wth agr-food sectors, partcularly the wne sector are usually neglected n emprcal works. Our paper analyses revealed comparatve advantages together wth ts duraton of leadng European wne producers from 2000 to 2013 by usng panel data econometrcs. Ths paper expands the exstng lterature n three ways. Frst, t apples the theory of revealed comparatve advantages on wne product n EU. Second, t analyses the stablty of comparatve advantages. Thrd, t seeks to draw wne specfc reasons for changng patterns n comparatve advantages. 2. Theoretcal framework There s ust a lmted amount of lterature analysng revealed comparatve advantages n the feld of agrculture. Fertő and Hubbard (2003) were among the frst to analyse revealed comparatve advantages n Hungaran agr-food sectors and dentfed eleven compettve product groups. Bonec and Fertő (2009) searched for agro-food trade compettveness of Central European and Balkan countres and showed that bulk prmary raw agrcultural commodtes had hgher and more stable relatve trade advantages compared to consumer-ready foods, mplyng compettveness shortcomngs n food processng and n nternatonal food marketng. Bonec and Fertő (2012) nvestgated the mpact of EU enlargement on the agro-food export performance of New Member States (NMS) over 1999-2007 and found longer duraton for exportng hgher value-added specalsed consumer-ready food and more compettve nche agro-food products. Bonec and Fertő (2014) analysed the agr-food compettveness of European countres and showed that most of the old EU-15 member states experenced a greater number of agr-food products havng a longer duraton of revealed comparatve export advantages than most of the new EU-12 member states have. Jámbor (2013) analysed the comparatve advantages and specalsaton of the Vsegrád 4 Countres agrfood trade and showed that comparatve advantages decreased after EU accesson n all countres, suggestng a weakenng stablty of compettve postons. Török and Jámbor (2013) analysed EU New Member States agr-food trade patterns and hghlghted that almost all countres experenced a decrease n ther comparatve advantage after the EU accesson, though t stll remaned at an acceptable level n most cases. 3. Methodology The varous methods elaborated around the theory of revealed comparatve advantage provde the bass for analyss. The orgnal ndex of revealed comparatve advantage was frst publshed by Balassa (1965) who defned the followng: n B = (1) t nt where means export, ndcates a gven country, s a gven product, t s a group of products and n s a group of countres. It follows that a revealed comparatve advantage (or dsadvantage) ndex of exports can be calculated by comparng a gven country s export share of ts total exports wth the export share n total exports of a reference group of countres. If B>1, a gven country has a comparatve advantage compared to the reference countres or, n contrast, a revealed comparatve dsadvantage f B<1. The Balassa-ndex s crtcsed because t neglects the dfferent effects of agrcultural polces and exhbts asymmetrc values. Trade structure s dstorted by dfferent state nterventons and trade lmtatons whle the asymmetrc value of the B ndex reveals that t extends from one to nfnty f a country enoys a comparatve advantage, but n the case of comparatve dsadvantage, t vares between zero and one, whch overestmates a sector s relatve weght. Vollrath suggested three dfferent specfcatons of revealed comparatve advantage n order to elmnate the dsadvantages of the Balassa ndex, the detaled descrpton of whch can be found n Vollrath (1991). In order to treat the asymmetrc value problem of the Balassa-ndex, Dalum et al. (1998) transformed B ndex as follows, thereby creatng the Revealed Symmetrc Comparatve Advantage (RSCA) ndex: ( B 1 )/( + 1) RSCA = B The RSCA takes values between -1 and 1, wth values between 0 and 1 ndcatng a comparatve export advantage and values between 1 and 0 a comparanve export dsadvantage. Snce the RSCA dstrbunon s symmetrc around zero, a potental bas s avoded (Dalum et al., 1998). Yu et al. (2009, 2010) adopted an alternatve measure to assess the dynamcs of comparatve advantages the Normalsed Revealed Comparatve Advantage (NRCA) ndex, defned as follows: NRCA = E E ( E )( E ) ( E E ) 2 (2) (3)
( E )( ) E where represents actual exports and stands for the comparatve-average-neutral level n exports of commodty for country. If NRCA>0, a country s comparatve advantage on the world market s revealed. The dstrbuton of NRCA values s symmetrcal, rangng from 1/4 to +1/4 wth 0 beng the comparatve-advantage-neutral pont. Besdes calculatng revealed comparatve advantages, ther stablty and duraton can be measured as well. In analysng the stablty of the RSCA ndex, a regresson was run on the dependent varable, RSCA ndex at tme t2 (for sector n country ), whch s tested aganst the ndependent varable the RSCA ndex n year t1 (3). 4. Emprcal results Fgure 1 llustrates wne trade data of EU top 16 wne exporters on the world market. It reveals ncreasng tendency of European wne export to world market from 2000 to 2013. Regardng the wne mport to European markets, t shows only a moderate rse n these countres. Fgure 1: Wne trade of EU top 16 wne exporters on world market, 2000-2013, n mllon USD RSCAt2 = α + βrsca t1 + ε (4) where α and β are standard lnear regresson parameters and ε s a resdual term. If β = 1, then ths suggests an unchanged pattern of RSCA between perods t1 and t2, meanng there s no change n the overall degree of specalzaton n European wne trade. On one hand, f β > 1, the exstng specalzaton s strengthened, meanng that a low level of specalzaton n the ntal perod leads to less specalzaton n the future, whch s called β dvergence (Bonec and Fertő, 2008). On the other hand, f 0 < β < 1, commodty groups wth low ntal B ndces grow over tme, whch s called β convergence (Bonec and Fertő, 2008). However, f β < 0 change n the sgn of the ndex s shown. However, as Dalum et al. (1998) pont out, the β > 1 s not a necessary condton for growth n the overall specalzaton pattern. They argue that suffcent condtons for specalzaton or despecalzaton need further analyses. If R s the correlaton coeffcent of the regresson, then the pattern of a gven dstrbuton s unchanged when β = R. If β >R, then the degree of specalzaton has grown (leadng to dvergence). If β < R, then the degree of specalzaton has fallen (meanng convergence). Followng Bonec and Fertő (2008), a survval functon S(t) can also be estmated by the usng the nonparametrc Kaplan Meer product lmt estmator, whch pertans to the product level dstrbuton analyss of the RSCA ndex. Followng Bonec and Fertő (2008), a sample contans n ndependent observatons denoted (t; c), where = 1, 2,..., n, and t s the survval tme, whle c s the censorng ndcator varable C (takng on a value of 1 f falure occurred, and 0 otherwse) of observaton. Moreover, t s assumed that there are m < n recorded tmes of falure. Then, we denote the rank-ordered survval tmes as t(1) < t(2) < < t(m). Let n ndcate the number of subects at rsk of falng at t() and let d denote the number of observed falures. The Kaplan Meer estmator of the survval functon s then (wth the conventon that ˆS(t) = 1 f t < t(1)): The wne export share of EU top wne exporters confrms the relevance of 16 countres analysed. They are responsble for 70 % of world wne market; however the share of the countres was decreased below 70 % to 2013. The mport share remaned at the same level durng the perod analysed (Fgure 2). Fgure 2: wne export and mport share of EU top 16 wne exporters, 2000-2013, n percentage of world wne trade Sˆ( t) = Π t ( )< t n d n (5) The paper employs European wne trade data n HS-6 level, product code: 2204, between 2000 and 2013 for sxteen countres as the bggest producers responsble for 70% of EU total wne exports, n partcular: Austra, Bulgara, Croata, Cyprus, Czech Republc, France, Germany, Greece, Hungary, Italy, Malta, Portugal, Romana, Slovak Republc, Slovena and Span. Furthermore, the paper concentrates on the B ndex (and ts transformatons, the RSCA and NRCA ndces) as t excludes mports, whch are more lkely to be nfluenced by polcy nterventons. The phasng out of export subsdes s a further reason to choose a B- based ndex. In order to calculate the varous ndces mentoned above, the paper uses World Bank WITS database (World Bank, 2014) wne trade data at the global level.
Our results ndcate that the maorty of European wne s sold to the Unted Kngdom (21%), the Unted States (17%) and Germany (11%) wth a hgh level of concentraton by country. Furthermore, Germany and the USA can be consdered as wne producer and mporter at the same tme. Table 1: Trade values of TOP 5 EU wne exporters, 2000-2013, mllon USD and percentage Country export mport export share n proporton of world wne export France 109429 8192 34% Italy 60850 4775 18% Span 31470 2327 9% Germany 12031 35328 4% Portugal 9985 1346 3% Top 5 total 223767 51969 68% Other mnor wne mporters were Canada, Netherlands, Swtzerland, Russa and Chna durng the analysed perod, though ther wne mport share was under 10 % (Table 2). The top 10 wne mporters represented 72 % of total world wne mport. Table 2: mport of TOP 10 wne mporters, 2000-2013, mllon USD and mport share n percentage Country wne mport value mport share n proporton of world wne mport Unted Kngdom 4811 21% Unted States 4071 17% Germany 2523 11% Canada 1280 5% Belgum 1102 5% Netherlands 956 4% Swtzerland 918 4% Russan 653 3% France 585 2% Chna 488 2% Top 10 total 17388 72% Fgure 3: EU top 16 wne exporters trade balance by country, 2000-2013, n mllon USD As to the of RCA ndces, we can observe that whle France, Italy, Span, Portugal, Greece, Bulgara, Cyprus (7 countres) had postve ndces thus comparatve advantage n world market, we can fnd 9 countres wth negatve ndces (Austra, Croata, Czech Republc, Germany, Hungary, Malta, Romana, Slovaka, Slovena) suggestng comparatve dsadvantages n the EU (Fgure 4). Fgure 4: Revealed comparatve advantage of EU-16 wne trade by RCA, RSCA and NRCA ndces (mean), 2000-2013 Regardng wne trade balance, France, Italy, Span and Portugal show a surplus, whle other European wne exporters accounted for a remarkable trade defct n the wne sector. France had the hghest trade surplus (101237 mllon USD) and Germany the hghest trade defct (-23296 mllon USD) durng the analyzed perod (Fgure 3).
If we take nto consderaton of the changes of RCA ndces over tme, we can conclude that France, Italy and partly Span were able to ncrease or mantan ther wne trade compettveness whle the advantages of Portugal, Cyprus, Bulgara and Greece have fallen down sgnfcantly to 2013 (Fgure 5). Fgure 5: Changes of RCA ndces of compettve wne exporters (f RCA>1) by year We have chosen the revealed symmetrc comparatve advantage ndex (RSCA) n order to analyse the duraton and stablty of ndces (equaton 4). The stablty and duraton of RSCA ndex show that the estmated β values are between 0 and 1, and decreased over tme, whch s called β convergence (Bonec and Fertő, 2008). The estmated results confrm that the degree of specalzaton has fallen (meanng convergence) snce the estmated β are lower than the correlaton coeffcent (R) for the entre perod. Table 3: Stablty of the SRCA ndex between 2000 and 2013 Lags α β p-value R2 R β/r N 1-0.012 0.987 0.000 0.991 0.995 0.992 208 2-0.023 0.977 0.000 0.981 0.991 0.986 192 3-0.033 0.963 0.000 0.968 0.984 0.979 176 4-0.042 0.949 0.000 0.955 0.977 0.972 160 5-0.050 0.932 0.000 0.938 0.968 0.963 144 6-0.057 0.913 0.000 0.923 0.961 0.951 128 7-0.066 0.894 0.000 0.902 0.949 0.942 112 8-0.075 0.876 0.000 0.879 0.938 0.935 96 9-0.087 0.862 0.000 0.853 0.923 0.933 80 10-0.102 0.844 0.000 0.827 0.909 0.929 64 11-0.116 0.825 0.000 0.799 0.894 0.923 48 12-0.131 0.807 0.000 0.777 0.882 0.916 32 13-0.146 0.776 0.000 0.747 0.864 0.898 16 In further analysng the stablty of revealed comparatve advantages n European wne trade, ts duraton was estmated by the usng the non-parametrc Kaplan Meer product lmt estmator. As descrbed n the methodology secton, equaton 5 was run on our panel dataset and results confrm that the survval tmes of revealed comparatve advantages n European wne exporters are not persstent over the perod analysed n general (Table 5). Survval chances of 96,4% at the start of the perod fell to 9,5% by 2013, suggestng that ferce competton s exstent n European wne trade. However, t s evdent that the bggest European wne exporters have mantaned ther comparatve advantages. France, Italy, Span and Portugal had stable postons n the perod analysed whle compettveness of German wne export deterorated sgnfcantly. The equalty of the survval functons across product groups can be checked usng two non-parametrc tests (Wlcoxon and log-rank tests). Results show that the hypothess of equalty can be reected at the 1% level of sgnfcance, meanng that smlartes across product groups n the duraton of comparatve advantage are absent (Table 5). Table 5: Kaplan-Meer survval rates for RCA ndex and tests for equalty of survval functons n European wne export on world market by TOP16 exporters, 2000 2013 Year Survvor Functon France Italy Span Portugal Germany 2000 0.9643 1.0000 1.0000 1.0000 1.0000 0.9286 2001 0.9226 1.0000 1.0000 1.0000 1.0000 0.8571 2002 0.8793 1.0000 1.0000 1.0000 1.0000 0.7857 2003 0.8344 1.0000 1.0000 1.0000 1.0000 0.7143 2004 0.7874 1.0000 1.0000 1.0000 1.0000 0.6429 2005 0.7382 1.0000 1.0000 1.0000 1.0000 0.5714 2006 0.6863 1.0000 1.0000 1.0000 1.0000 0.5000 2007 0.6312 1.0000 1.0000 1.0000 1.0000 0.4286 2008 0.5720 1.0000 1.0000 1.0000 1.0000 0.3571 2009 0.5076 1.0000 1.0000 1.0000 1.0000 0.2857 2010 0.4362 1.0000 1.0000 1.0000 1.0000 0.2143 2011 0.3545 1.0000 1.0000 1.0000 1.0000 0.1429 2012 0.2548 1.0000 1.0000 1.0000 1.0000 0.0714 2013 0.0955 1.0000 1.0000 1.0000 1.0000 0.0000 Log-rank test 0.000 Wlcoxon test 0.000 5. Conclusons A vast amount of lterature s dedcated to the analyses of revealed comparatve advantages of global trade. Whle most of the studes have focused on ndustral products or agr-food sectors, the wne sector s scarly nvestgated by these methods. Our paper analysed the compettveness of top 16 tradtonal wne exporters. In addton, the study nvestgated the stablty and duraton of comparatve advantages by the changes of RSCA ndces from 2000 to 2013. On the whole, our results suggest that Bulgara, Cyprus, France, Greece, Italy, Portugal and Span had revealed comparatve advantages on the world market out of the 16 top European wne producers durng the perod analysed. However, these advantages seem to have weakened for the maorty of countres as evdent from the dstrbuton of comparatve advantages ndces as well as from the stablty tests by the help of lagged ndces and Kaplan-Meer estmator of the survval functon. France, Italy and Span were the only exceptons wth stable comparatve advantages.
The contrbuton of our study s threefold. Frst, t apples the theory of revealed comparatve advantages on wne n the EU. Second, t analyses the stablty of comparatve advantages. Thrd, t seeks to draw wne specfc reasons for changng patterns n comparatve advantages. References Balassa, B. (1965): Trade Lberalzaton and Revealed Comparatve Advantage. The Manchester School, no. 33: 99-123. Bonec, Š, Fertő, I. (2012). Does EU enlargement ncrease agro-food export duraton? World Economy, 35(5): 609-631. Bonec, Š. and Fertő, I. (2008). European Enlargement and Agro-Food Trade. Canadan Journal of Agrcultural Economcs, 56(4): 563-579. Bonec, Š. and Fertő, I. (2009). Agro-food trade compettveness of Central European and Balkan countres. Food Polcy, 34(5): 417-425. Bonec, Š., I. Fertő (2009): Determnants of agro-food trade competton of Central European countres wth the European Unon. Chna Economc Revew, no. 2:327-337. Bonec, Š., I. Fertő (2012): Complementartes of trade advantage and trade compettveness measures. Appled Economcs, no. 4: 399-408 Bonec, Š., I. Fertő (2014): Agr-Food Export Compettveness n European Unon Countres. Journal of Common Market Studes, no. 3: 476 492 Dalum, B., Laursen, K. and Vllumsen, G. (1998). Structural Change n OECD Export Specalsaton Patterns: de-specalsaton and 'stckness', Internatonal Revew of Appled Economcs vol. 12(3): 423-443. Fertő, I., and Hubbard, L. J. (2003). Revealed comparatve advantage and compettveness n Hungaran agr food sectors.the World Economy, 26(2), 247-259. Fertő, I., L. J. Hubbard (2003): Revealed comparatve advantage and compettveness n Hungaran agr food sectors. The World Economy, no.2: 247-259. Hoen, A. R. and Oosterhaven, J. (2006). On the measurement of comparatve advantage. Annals of Regonal Scence 40, 677 691. Jámbor, A. (2013). Comparatve advantages and specalsaton of the Vsegrad countres agr-food trade. Acta Oeconomca et Informatca, 16(1): 22-34. Jámbor, A. (2013): Comparatve advantages and specalsaton of the Vsegrad countres agr-food trade. Acta Oeconomca et Informatca, 16(1): 22-34. Jámbor, A. (2014). Country-specfc determnants of horzontal and vertcal ntra-ndustry agr-food trade: The case of the EU New Member States Journal of Agrcultural Economcs, 65(3): pp. 663-682. Török, Á. and Jámbor, A. (2013). Agr-food trade of the New Member States snce the EU accesson. Agrcultural Economcs Czech 59(3): 101-112. Vollrath, T.L. (1991). A theoretcal evaluaton of alternatve trade ntensty measures of revealed comparatve advantage. Weltwrtschaftlches Archv 130(2): 263 279. World Bank (2014). World Integrated Trade Soluton (WITS) database: https://wts.worldbank.org/ Yu, R., Ca, J. and Leung, P.S. (2009). The normalzed revealed comparatve advantage ndex. Annals of Regonal Scence 43(1): 267 282. Yu, R., Ca, J., Loke, M.K. and Leung, P.S. (2010). Assessng the comparatve advantage of Hawa s agrcultural exports to the US manland market. Annals of Regonal Scence 45(2): 473 485. Appendx 1: RCA, RSCA and NRCA ndces (mean) of EU-16 wne exporters, ranked by RCA, 2000-2013 Country RCA SRCA NRCA * France 8,05 0,78 6,37 Portugal 7,84 0,77 0,58 Italy 5,08 0,67 3,12 Span 4,92 0,66 1,61 Cyprus 4,53 0,55 0,00 Bulgara 3,02 0,42 0,05 Greece 1,81 0,27 0,03 Croata 0,62-0,24-0,01 Hungary 0,59-0,28-0,06 Austra 0,44-0,39-0,14 Romana 0,41-0,45-0,04 Germany 0,37-0,47-1,33 Slovena 0,25-0,6-0,03 Slovak Republc 0,15-0,74-0,07 Malta 0,08-0,86-0,01 Czech Republc 0,07-0,87-0,16 *Note: because of vsblty, NRCA calculated as NRCA*10000