Comparative Advantages in CEEC-5

Transcription

1 Comparative Advantages in CEEC-5 Tullio Gregori University of Trieste, DISPES, Italy Abstrat. The struture, geography, and organization of world trade have dramatially hanged in the last deades with the emergene of new global ompetitors and the deline of old ones. Global Value Chains are probably the most prominent feature of this new landsape and the growth of these hains is posing new hallenges to studies of international trade and ountries ompetitiveness. Some traditional measures may be of little use sine they hinge on the assumption that all ativities in the prodution of a good take plae in the loal eonomy, using domesti inputs only. Furthermore, export indexes based on gross flows beome less meaningful as part of its value is made of imported inputs. Previous applied researh uses a wide array of statistial tools to assess international speialization and trade performane but fouses is on Gross Exports. These analyses an be very misleading due to double ounting of trade in intermediate inputs. This paper would like to shed light on the export struture and ompetitiveness in five Central and Eastern European Countries during I assess whih setors enjoy a omparative advantage in using Gross Exports and Value Added in Trade data. The sets of ompetitive industries are ompared and disussed. There are some notieable differenes in these indiators in Hungary and Slovakia, while just a few in omania and Bulgaria. Keywords: international trade, omparative advantages, CEEC-5. 1 Introdution This paper examines the export performane of some Central and Eastern European ountries. It tries to shed light on their export ompetitiveness during the period using the data provided by WIOD. Under free trade these ountries should speialize in those goods and servies they have a omparative advantage. From a theoretial point of view, this issue an be easily solved if we an observe the differene between autarki and free trade relative pries (Deardorff 1980). Unfortunately, the former are usually unobservable and we must resort to ex post information. Indiators of ompetiveness should respet several riteria. First, they should address all the setors exposed to international ompetition, i.e. the ones that produe the so-alled tradeables. Then, they should been based on data that are fully omparable at the international level. Due to lak of homogeneity, studies based on export pries are often onfined to aggregate measures of manufaturing (Durand and Giorno 1987). Furthermore, for a quite large number of goods, suh as ommodities, pries tend to be determined at the world level rather than by loal produers. To overome these problems we adopt the ustomary evealed Comparative Advantage approah initiated by Liesner (1958) and Balassa (1965), who adapted the loation oeffiient introdued by Hoover (1936) to international trade. We embrae several indiators for 35 setors and deviate from standard analyses in two respets. First, we fous on setors rather than goods reduing the dimensionality of the produt spae by far. Seond, we take into aount the profound hanges in the geography of world trade with the emergene of the Global Value Chains. Hene, we move beyond standard measures based on Gross Exports to address indiators based on domesti value added generated by foreign final demand. This distintion is ruial, as within GVCs it is ustomary to offshore parts of prodution. Final goods may ontain a large share of imported intermediate and ompetitiveness indiators based on Gross Exports may be ompletely misleading as they inlude output due to other produers. To bring these issues to the fore we embrae the Value Added in Trade approah. Only net flows an reveal atual international trade 441

2 as value added, not gross output, disloses domesti prodution. Another advantage of this approah is the straightforward extension to servies. The latter are usually negleted, as these international flows are often unreorded, even if it has been widely reognized the inreasing importane of tertiary that has the largest share in GDP of all the most advaned ountries. Exluding servies from the analysis would fail to apture a fundamental ontribution to export performane in these nations as domestially produed servies are embodied even in foreign goods. Hene, it is interesting to hek export ompetitiveness through the lens of domesti value added (Ceglowski 2015). The paper is organized as follows. Next setion presents the dataset and the methodology. Then Setion 3 shows whih setors are ompetitive in five Central and Eastern European ountries: Bulgaria, the Czeh epubli, Hungary, omania, and Slovakia. Setion 4 onludes. 2 Methodology and database We adopt the world input output model with N setors and ountries. International IO tables follow the standard approah in multiregional analysis and an be represented by partitioned matries and vetors (Miller and Blair 2006): with: i x y f (1) ( I A) 1 B 1 = = x A A A B B B x = x A= A A A B= B B B 1 1 x A A A B B B, 1, 1, i= y y y 1 1 y = y + + y + + y = f + + f + + f 1 y y y where f is final demand in ountry. Country Gross Exports to ountry s are: and bilateral gross trade balane is: s s s s eg = A x + y (2) s s s s ( A x + y ) ( A x ) t =u u +y (3) s s G where u is a unit vetor. Worldwide exports and imports are easily derived by aggregation. Following Koopman et al. (2014), we an arrange the Leontief model in a different way: 442

3 or q q q B B B y y y q q q B B B = y y y (4) q q q B B B y y y ( ) q q q = I A f f f (5) ( ) 1 Q= I A F (5bis) where both the gross output and final demand matries are (N ). If we premultiply (5) with a diagonal matrix of diret value-added oeffiients along the main diagonal we obtain the value-added prodution matrix: vˆ 0 0 q q q 0 ˆ ˆ ˆ ˆ v v y v y vbf = = ˆ g g1 v q q q g g vˆ ˆ B y v g= 1 B y g= 1 1 1g g1 1 1g g vˆ B y vˆ B y g = 1 g = g g1 2 2g g 0 q q q B B g= 1 g= 1 (6) whose blok elements in the diagonal give eah ountry s prodution of value added absorbed at home, while off diagonal blok elements show value added produed in ountry and absorbed abroad. Value added exports from ountry to s and worldwide are: s g gs e =vˆ B y (7) VA g= 1 g gs e =vˆ B y. (8) VA s g= 1 Similarly, ountry imports value added produed in ountry s and worldwide: s s sg g e ˆ VA =v B y (9) g= 1 s sg gs m =vˆ y. (10) VA B s g= 1 This approah is applied to the well-known WIOD dataset that ontains annual time-series of world input output tables overing the period from 1995 to 2011 for 40 nations (Timmer et al. 2015). All together they over about 85% of world GDP in In addition, the remaining non-overed part is estimated too, so that these 40 ountries and the est of the World region form a omplete world IO model. They have an industry-by-industry format and provide details for 35 industries mostly at the two-digit level. Data are based on offiially published supply and make matries merged with national aounts and international trade statistis, even if it is laimed it relies more on IO tables then international trade (Weidman et al. 2011, ojas-omagosa 2015). Moreover, the alibration proedure 443

4 employed in WIOD implies the usual data disrepanies annot be fully alibrated, and the est of the World is impliitly absorbing these differenes. The literature aimed at investigating international ompetitiveness and trade performane is huge. We follow the standard evealed Comparative Advantage approah and address ex post trade data. Balassa (1965) introdued the first and the most popular measure of omparative advantage. It may trak either exports or imports, even if only the former is usually figured out. Balassa defines omparative advantage in terms of markets shares: BCA = e e (11) i i w w ei e w Where e i and ei are exports of the i-th produt in ountry and the referene area, say the s world. e = e is total exports in ountry and, similarly, w i i e is the grand total. The ritial value is one. A ountry has a omparative advantage in the prodution of the i-th good if its export share is larger than the referene area and BCA > 1. This index suffers several shortomings. First, it is a single flow index and results about exports and imports may differ. Then, the BCA index has neither an ordinal nor a ardinal property that failitate omparisons between ountries and setors: the traditional CA approah does not produe a strit ordinal index and, in ases, may not even provide a statistially signifiant ranking of industries aording to revealed omparative advantage. Conerning this problem, it is evident that the potential for bias is greatest when omparisons are made between industries whih have the widest differenes in their underlying (ountry) CA distributions (Yeats 1985, 67). Moreover, it is not additive and it tends to give a stronger advantage to small ountries with high speialization (Hoen and Oosterhaven 2006). Finally, it is asymmetri in that is bounded by zero on one end and does not have an upper bound on the other end (De Beneditis and Tamberi, 2004). The asymmetry auses problems when the Balassa index is inluded in standard eonometri models. In order to fix this issue an easy alternative has been suggested by Laursen (1998) and Dalum et al. (1998): SCA i BCAi 1 = BCA + 1. (12) i This index is symmetri and ranges from -1 to +1 with a ritial value equal to zero. Moreover, it is an approximation of the log transformation of BCA suggested by Vollrath (1991) and should provide the same rankings, sine it is an almost positive monotoni transformation. It shares other shortomings yet. It does not have a stable mean over spae and time, whih gives suspiious on its omparability between and within ountries. Some researhers argue the multipliative form of BCA and SCA auses their problemati features. Hoen and Oosterhaven (2006) suggest to adopt an additive form: ACA = e e e e (13) w i i i w where all the variables have been defined previously. The ut-off is zero again. Positive values indiate a omparative advantage and negative ones the opposite. The index is symmetri and it is not affeted by the level of setoral aggregation, but omparability in ross ountry analysis is questionable, as the sum of the ACAs with respet to a given setor is not stable (Sanidas and Shin 444

5 2010). A final reent alternative is mixing both additive and multipliative features with the Normalized evealed Comparative Advantage index introdued by Yu et al. (2009): NCA = e ee e ee (14) w i i i w w w The referene value is still zero, while values ranges from to Sine world trade is used to normalize, figures tend to be very small. One unquestionable advantage is omparability over time and aross spae, sine the sum of NCAs is equal to zero aross both dimensions. This explains well the notion of zero sum imbedded in omparative advantage: if a ountry gains omparative advantage in one setor, then the ountry loses omparative advantage in other setors; and if one ountry gains omparative advantage in a setor, then other ountries lose omparative advantage in the setor (Sanidas and Shin 2010, 18). 3 Competitiveness in CEEC5 In this setion I present results for the five CEEC ountries under investigation, i.e. Bulgaria, the Czeh epubli, Hungary, omania, and Slovakia. This preliminary study addresses the dihotomous hoie about ompetitiveness without mathing values between and within ountries, as omparability over time and setors is questionable but NCA. Is not diffiult to prove that if an industry has a ompetitive advantage aording to the Balassa index then the other indiators onfirm it. Hene, we an restrit the analysis to the differenes between Gross Export and Value Added in Trade. For reason of larity, results are not given for eah year, but for seleted ones. We divided the sample into three periods and pik: 1995 and 2001 as the initial observation and the year before the reation of the Monetary Union, 2006 as the last year of tranquility before the subprime risis and the Lehman bankrupty, and 2011, i.e. the last year available before the omplete eruption of the Euro-risis. Setor lassifiation is provided in the Appendix. In table 1-5 I provide the lists of the industries with omparative advantages in eah ountry. I do not assess hanges over time and diversities between nations, but fous on differenes in the dihotomous lassifiation using gross and net data. These dissimilarities are highlighted in bold. For instane, in Bulgaria (table 1) aording to Value Added in Trade the 4-th setor (Textile) has a ompetitive advantage in 1995 that is not signaled by the Gross Export index. However, the former is rather questionable as BCA is 1.03 only and NCA is slightly positive ( ). Anyway, figures are muh larger in the following years suggesting an inreasing ompetitiveness. The piture is somehow different if we deem VAiT in finanial intermediation (setor 28). This industry is in the seond position in the ACA/NCA ranking, while it holds a lower plae in the BCA/SCA list (only 15 th ), but still on the ompetitive side of the eonomy. In ontrast, its performane is rather modest aording to Gross Export. This is reasonable as finanial servies may be embodied in other goods and diret sales abroad are often small. Quite interestingly, it is no longer ompetitive between 1997 and 2003, when all the indiators shift below the ritial utoff. On the opposite side, setors 9 (Chemial Produts), 11 (Other Non-Metalli Mineral), and 12 (Basi Metals and Fabriated Metal) are ompetitive in gross terms only. The very same situation refers to 21 (etail Trade) and 27 (Post and teleommuniation). The piture is slightly different in 2001 with 5 (Leather) that is ompetitive aording to VAiT only and vie versa 6 (Wood) for Gross Exports. The idential situation applies to 13 (Mahinery, ne) and 12 (Basi Metals and Fabriated Metal). The retail setor (21) is interesting too. Its performane is onordant in the first years only. Thereafter, there is a deline in value added figures. For instane, BCA is about 1.28 in 1996 and merely 0.52 two years later. In the following period, it is lose to 445

6 unity (0.98) in 2006 alone. In the last part of the sample there are just a few differenes between VAiT and Gross Exports, as the right hand olumns show. Table 1: Competitive setors in Bulgaria VA GE VA GE VA GE VA GE Table 2: Competitive setors in the Czeh epubli VA GE VA GE VA GE VA GE Let s now turn to the Czeh epubli. We an notie the number of ompetitive setors is more or less the same as in Bulgaria. At the beginning, there are some diversities between Gross Exports and VAiT indexes in primary and seondary suh as 1 (Agriulture), 3 (Food), 4 (Textile), and 10 (ubber and plasti). On the opposite, differenes mostly pertain to the tertiary in the last deade. Servie setors that deliver exports mainly via other produts are 20 (Wholesale trade), 21(etail trade), 23 (Inland transport), 26 (Other Supporting and Auxiliary Transport Ativities), 27 (Post and 446

7 teleommuniation), and 32 (Eduation). However, their ourrene is sattered through time, but the 27-th setor, that has always a ompetitive advantage as stated by the VAiT indiators and only in 1995 for Gross Exports. Table 3: Competitive setors in Hungary VA GE VA GE VA GE VA GE Table 4: Competitive setors in omania VA GE VA GE VA GE VA GE Hungary and omania look very different. Sometimes they do not seem very ompetitive in international markets, at least when we use Gross Export. The former ountry exhibits only 12 figures larger than ritial ut offs in 2006, while the latter just 13 in There are a few more before and thereafter, but ardinality of the sets is often smaller than in the previous eonomies. Conordane between the VAiT and Gross Exports is almost perfet in the first part of the sample in omania, beause only a ouple of values diverge. In the last years, there are several differenes mostly in 447

8 manufaturing. On the ontrary, several servie setors in Hungary enjoy a ompetitive advantage that is revealed only when we adopt the value added ontent. These range from 26 (Other Supporting and Auxiliary Transport Ativities) to 32 (Eduation), while 33 (Health and Soial Work) drops out after Finally, let s deem Slovakia in table 5. This ountry summarize several features disussed above. There are a few differenes onerning mainly 16 (Manufaturing, ne), 17 (Eletriity, Gas and Water Supply), 19 (Sale), 20 (Wholesale Trade), and 29 (eal Estate Ativities). Nevertheless, we must aknowledge there is an overall onordane between gross and net measures. Table 5: Competitive setors in Slovakia VA GE VA GE VA GE VA GE Conlusions This paper examines ompetitiveness in five CEEC ountries, i.e. Bulgaria, the Czeh epubli, Hungary, omania, and Slovakia. We embrae four indiators of ompetitive advantages ranging from the most popular due to Balassa (1965) to the reent one introdued by Yu et al. (2009). However, it is easy to prove that all these indiators are equivalent in term of the dihotomous hoie about ompetitive advantages, i.e. if a setor is ompetitive aording to the Balassa riterion then it do so aording to the other ones beause if BCA > 1 then SCA > 0, ACA > 0, and NCA > 0. Hene, we restrit our analysis and take into aount setoral CA indiators using Gross Exports and Value Added in Trade. The ountry net trade surplus/defiit is the same in both ases while, at the disaggregated industry level, values an differ as the VAiT inludes both value added in its diret exports and the value added it supplies as inputs to other domesti industries exports. A large literature shows global supply hains an hange the piture a lot. It is interesting to hek differenes in ompetitiveness of the ountries under srutiny. For suh a task, we embrae the WIOD dataset that provides a series of annual world input output tables from whih we an derive gross and net exports by industry. This approah is extremely useful as it allows us to onsider the tertiary as well manufaturing. The omplete inlusion of servies is important for several reasons. First, most advaned eonomies shifted their prodution towards the tertiary, whih is now the main bulk of GDP. Seond, even if servies are not diretly exported they are often embodied in goods. Hene, a Balassa index smaller than one in Gross Export but larger than unity in VAiT an reveal a setor whose output 448

9 is inluded in other goods that are sold abroad. The opposite shows an international shipment mainly made by foreign value added. Hene, the information onveyed by gross and net flows may be very different (Johnson and Noguera 2012). Koopman et al. (2014) give examples how, for some setors, a omparative advantage aording to the former an turn in a disadvantage for the latter. eently, Brakman and Marrewijk (2016), using a previous and modified version of WIOD, onlude ountries speialize dissimilarly. Consistent with the theory of omparative advantage, distributions of CA between different ountries are different. This holds for gross export CA as well as for value added CA The distributions of ross export CA and value added CA are almost always signifiantly different for a ountry. These measures thus do not onvey the same information (Brakman and Marrewijk 2016, 9-10). Their analysis drops three setors, does not onsider the last two years, and onverts urrent dollars to onstant ones using the US GDP deflator. This global deflation is questionable and I prefer to stik to urrent values using all the observations available. Nonetheless, differenes for our CEEC-5 are minimal and stylized fats are mostly onfirmed. For instane, Bulgaria appears to be a very ompetitive ountry in terms of number of setors when data about Gross Exports are used. Nineteen industries out of 35 (32 for Brakman and Marrewijk) display an export share larger than worldwide, but this number is redued when we deem net flows. However, this finding applies predominantly in the first part of the sample, when European GVCs were still building up. This phenomenon ould have been at work at the turn of the entury in Slovakia and Hungary too, as the number of ompetitive industries is inreasing in both seondary and tertiary. On the opposite, it is not so apparent in omania and the Czeh epubli. This alls for a better understating of the strength and the deepness of the links within and between supply hains and, first of all, for a quantitative assessment about the magnitude of omparative advantages. Future researh should address both issues. eferenes Balassa, B Trade Liberalization and evealed Comparative Advantage. The Manhester Shool of Eonomi and Soial Studies 33 (2): Brakman, S., and C. V. Marrewijk A loser look at revealed omparative advantage: Grossversus value-added trade flows. Papers in egional Siene Ceglowski, J Assessing Export Competitiveness through the Lens of Value Added. The World Eonomy Dalum, B., K. Laursen, and G. Villumsen Strutural hange in OECD export speialisation patterns: de-speialisation and stikiness. International eview of Applied Eonomis 12 (3): Deardorff, A. V The General Validity of the Law of Comparative Advantage. The Journal of Politial Eonomy 88 (5): De Beneditis, L. and M. Tamberi Overall Speialization Empiris: Tehniques and Appliations. Open Eonomies eview 15 (4): Durand, M., and C. Giorno Indiators of International Competitiveness: Coneptual Aspets and Evaluation. OECD Journal: Eonomi Studies 9: Hoen, A.., and J. Oosterhaven On the measurement of omparative advantage. Annals of egional Siene 40 (3) : Hoover, E. M The Measurement of Industrial Loation. eview of Eonomis and Statistis 18 (4): Johnson,. C., G. Noguera Aounting for Intermediates: Prodution Sharing and Trade in Value Added. Journal of International Eonomis 86 (2): Koopman,., Z. Wang, and Z. Wei Traing Value-Added and Double Counting in Gross Exports. Amerian Eonomi eview 104 (2):

10 Laursen, K evealed omparative advantage and the alternatives as measures of international speialisation. DUID Working Paper Copenhagen Business Shool. Liesner, H. H The European Common Market and the British Industry. Eonomi Journal 68 (270): Miller,., and P. Blair Input-Output Analysis: Foundations and Extensions, 2nd ed. Cambridge: Cambridge University Press. oja-amagosa, H Bilateral trade balanes for the Netherlands and eight seleted ountries: Comparing gross and value added trade statistis and data soures. CPB Bakground Doument. Amsterdam. Sanidas, E., and Y. Shin Comparison of evealed Comparative Advantage Indies with Appliation to Trade Tendenies of East Asian Countries. Paper Presented at the 9 th Korea and the World Eonomy Conferene, Inheon, Timmer, M. P., E. Dietzenbaher, B. Los,. Stehrer, and G. J. de Vries An Illustrated User Guide to the World Input Output Database: the Case of Global Automotive Prodution.' eview of International Eonomis 23 (3): Vollrath, T. L A theoretial evaluation of alternative trade intensity measures of revealed omparative advantage. Weltwirtshaftlihes Arhiv 127 (2): Wiedmann, T., H. C. Wilting, M. Lenzen, S. Lutter, and V. Palm Quo Vadis MIO? Methodologial, Data and Institutional equirements for Multi-egion Input-Output Analysis. Eologial Eonomis 70 (11): Yeats, A On the Appropriate Interpretation of the evealed Comparative Advantage Index: Impliations of a Methodology Based on Industry Setor Analysis. Weltwirtshaftlihes Arhiv 121 (1): Yu,., J. Cai, and P. S. Leung The Normalized evealed Comparative Advantage Index. Annals of egional Siene, 43 (1):

11 Appendix Setoral Classifiation and NACE rev. 1 1 AtB Agriulture, Hunting, Forestry and Fishing 2 C Mining and Quarrying 3 15t16 Food, Beverages and Tobao 4 17t18 Textiles and Textile Produts 5 19 Leather, Leather and Footwear Wood and Produts of Wood and Cork 7 391t3939 Pulp, Paper, Paper, Printing and Publishing Coke, efined Petroleum and Nulear Fuel Chemials and Chemial Produts ubber and Plastis Other Non-Metalli Mineral t398 Basi Metals and Fabriated Metal Mahinery, Ne 14 30t33 Eletrial and Optial Equipment 15 34t35 Transport Equipment 16 73t37 Manufaturing, Ne; eyling 17 E Eletriity, Gas and Water Supply 18 F Constrution Sale, Maintenane and epair of Motor Vehiles and Motoryles; Wholesale Trade and Commission Trade, Exept of Motor Vehiles etail Trade, Exept of Motor Vehiles and Motoryles; 22 H Hotels and estaurants Inland Transport Water Transport Air Transport Other Supporting and Auxiliary Transport Ativities; Post and Teleommuniations 28 J Finanial Intermediation eal Estate Ativities 30 71t74 enting of M&Eq and Other Business Ativities 31 L Publi Admin and Defene; Compulsory Soial Seurity 32 M Eduation 33 N Health and Soial Work 34 O Other Community, Soial and Personal Servies 35 P Private Households with Employed Persons 451