WORKING PAPERS. International Outsourcing and Labour with Sector-specific Human Capital. Kurt Kratena

Transcription

1 ÖSTERREICHISCHES INSTITT FÜR WIRTSCHAFTSFORSCHNG WORKING PAPERS International Outsourcing and Labour with Sector-speciic Human Capital Kurt Kratena 7/006

2 International Outsourcing and Labour with Sector-speciic Human Capital Kurt Kratena WIFO Working Papers, No. 7 April 006 address: Kurt.Kratena@wio.ac.at 006/090/W/0

3 INTERNATIONAL OTSORCING AND LABOR WITH SECTOR-SPECIFIC HMAN CAPITAL Kurt Kratena Austrian Institute o Economic Research P.O. Box 9, A-03 Vienna, ASTRIA Tel.: Fax: Kurt.Kratena@wio.ac.at Abstract: This paper attempts to quantiy the impact o outsourcing on production patterns and the labour market in a two sector speciic actors model with skilled labour (speciic actor) and unskilled labour (mobile actor). Outsourcing can be compared to the case, where trade liberalization leads to trade in inal goods and a change in relative prices. In the latter case a downward pressure on the wage rate or skilled labour in one sector and a wage rise or skilled labour in the other sector indicate signiicant adjustment costs, whereas with outsourcing an outcome o rising wage rates or skilled labour in both sectors is easible. The ull impact o outsourcing depends on the relative weight o the actor savings and the cost savings eect. The negative impact on the unskilled wage rate is similar in both cases and depends on the macroeconomic relevance o the respective shocks. Key words: Outsourcing, trade liberalization, speciic actors model. JEL Code: F, F5, D0 Acknowledgments: I would like to thank Michael Paermayr and Yvonne Wolmayr or their very helpul comments and suggestions. The usual disclaimer applies.

4 . Introduction Fragmentation o the value added chain and international outsourcing have been the most important recent issues in the public debate about negative consequences o globalization or labour, especially or low skilled labour. In general, in the scientiic debate about outsourcing the use o the Heckscher-Ohlin model dominates. Besides the overall welare impact also the consequences or income distribution have been discussed exhaustively. Arndt (997, 999) has emphasized the positive welare eects o outsourcing due to an outward shit o the production possibility curve, although combined with changes in the income distribution. Other recent studies clearly isolate actor intensities o ragments and o integrated production as the determining parameters or the impact on wages, as Jones and Kierzkowski (00) and Deardor (00). General results in the Heckscher-Ohlin model or outsourcing in both sectors and dierent outsourcing equilibria can be ound in other recent studies (Egger, 00, Egger and Falkinger, 003). Egger (00) has shown how outsourcing in both sectors aects the income distribution and that the magnitude o the cost savings eect together with actor intensities o ragments play the determining role. This analysis has been extended in Egger and Falkinger (003) to a Heckscher-Ohlin model with dierent diversiied and specialized equilibria. The main inding there is a potential Pareto-improving impact, which depends on the interplay o the actor substitution and the cost savings eect. Jones and Kierzkowski (00) describe the state o research on the impact o outsourcing on labour with the almost anything can happen -statement, as most studies conclude with a welare improvement and unclear distributional eects o outsourcing. One remaining open question in this ramework thereore mainly concerns the empirical determination o parameter values and variables (e.g. actor intensities o ragments). In this debate the terms ragmentation and outsourcing have oten been used synonymsly. This study ollows the terminology o Egger and Falkinger (003), where ragmentation only describes the splitting up o the value added chain and (international) outsourcing describes the dislocation o part o the production leading to arm's-length transactions. Fragmentation thereore is a precondition or international outsourcing.

5 In some recent studies, where the speciic actors model developed by Jones (97, 975) has been applied (Kohler 00, 004) the interplay between income distribution changes and the welare eect o outsourcing represent the ocus o analysis. The most important result in Kohler (004) is that net welare increases rom outsorcing are larger, when the cost savings eect (given by the wage dierential between home and abroad) is larger ( the higher the gain, the lower the pain ). Kohler (004) conronts this inding with opposite results rom the analysis o income distribution impacts o trade liberalization ( the higher the gain, the higher the pain ). He inds out that outsoucing creates a macroeconomic surplus with income distribution changes similar to the immigration surplus. This paper takes up the concept o outsourcing surplus in a speciic actors ramework and ocusses on the undamental dierence between inal goods trade (FGT) and outsourcing (OS). As the distributional changes brought about by trade translate into adjustment costs at the labour market, the speciic actors model capturing these short term impacts within a consistent long term ramework (Neary, 978) seems adequate. An important part o the literature on imperect labour markets and trade concentrates on downward wage rigidity or unskilled labour. This analysis is based on Brecher (974) and generally shows that the unemployment impact o wage rigidity might ully erode the welare gain rom trade (Davis, 998). This has generally been ormulated or FGT and has been discussed intensively or the case o Germany (see or example: Seidel, 005). Egger and Egger (003) have shown the impact o OS in a model with skilled and unskilled labour and wage rigidity. Contrary to that the perspective on imperections in the labour market chosen here is based on adjustment costs o labour reallocation. This is in line with the literature on intra-industry trade (IIT) with limited mobility o the labour orce between industries, so that IIT leads to smooth adjustment compared to inter-industry trade. This hypothesis has recently also been exposed in the ramework o the speciic actors model (Brülhart and Elliott, 00, Greenaway, Haynes and Milner, 00, Lovely and Nelson, 00). Brülhart and Elliott (00) as well as Greenaway, Haynes and Milner (00) have shown that labour market imperections due to downward wage rigidity can be easily integrated into this ramework leading to unemployment instead o distributional changes and labour reallocation. We will

6 not explicitly introduce unemployment in our analysis, but we will instead show the completely dierent distributional impacts o FGT and OS and interpret them as the necessary lexibilities in the skilled and unskilled labour market in terms o wage lexibility and labour mobility in both cases. The essence o FGT liberalization is a price decrease, whereas OS as other studies have already shown is similar to technological change. The general equilibrium impact o OS will be decomposed within the speciic actors ramework into a actor saving and a ('pure') cost saving eect similar to the analysis o Kohler (004). This analysis will be carried out applying both a theoretical model as well as a numerical example based on the stylized acts o OS rom Austria to (ormer) Eastern European countries (now new E member states). It can be clearly derived, that the interplay o the actor and the cost savings eect determines the distributional changes. In the FGT case as well as in the OS case the potential negative impacts or unskilled labour mainly depend on the 'macroeconomic signiicance' o the shock. For FGT this is measured by the magnitude o the price shock and or OS by the proportion o the unskilled labour orce used in the ragment that is sourced out. The comparison o OS with FGT reveals similarities with respect to the impact on unskilled labour, but important dierences with respect to the impact on sector speciic labour and production patterns. These dierent results can be translated into dierences in actor adjustment costs, which might be relevant or European labour markets. The paper is organized as ollows. In section a speciic actors model is set up or FGT and or OS and is solved or actor prices and output in order to analyse the comparative statics o both cases. In section 3 a numerical example is chosen according to the stylized acts or outsourcing rom Austria to Eastern European contries (new member states) ater the all o the Iron Curtain. Section 4 summarizes the main results and concludes... Final Goods Trade and Labour The speciic actors model (Jones, 97, 975) with two sectors (,), sector speciic human capital H and H and unskilled labour L as the mobile actor exhibits constant returns to scale production unctions or outputs X and X :

7 X = F (L, H ) X = F (L, H ) () As the ixed endowment or sector speciic human capital determines output in each sector or given input coeicients, the marginal product schedules or labour in both sectors determine the equilibrium wage rate or unskilled labour. Input coeicients or each actor and output levels adjust, so that or given endowments the ull employment conditions hold: h X = H () h X = H (3) l X + l X = L (4) The input coeicients h ; h ; l and l depend or a given technology on relative actor prices or unskilled labour w and or skilled labour w and w : h = h w w ; h = h w w ; l = l w w ; l = l w (5) w with h > 0, h > 0, l > 0 and l > 0. h i l i The elasticity o substitution is given by σ i = with i =, (sectors), where an ^ over w w a variable describes relative change ( ĥ = dh /h ). From the unit price equations p = l w + h w and p = l w + h w the well known equations or price changes can be derived: p ϑ Lw + ϑhw = (6) i p ϑ Lw + ϑh w = (7) where θ ji are the actor shares ( θ Li = w L i /p i X i and θ Hi = w i H i /p i X i ) and λ ji are the proprtions o the mobile actor (unskilled labour) used in the two sectors ( λ Li = l i X i / L ). The price

8 equations (6) and (7) have been derived by applying the conditions or cost minimization (s.: Jones, 97, 975): ϑ Ll + ϑhh =0 L H ϑ l + ϑ h = 0 (8) The comparative static analysis is carried out here or ixed endowment o skilled and unskilled labour, i.e.: H = 0, H = 0 and L = 0. In that case output changes are determined by changes in input coeicients o the speciic actor: X = h ; X h = (9) sing (9) as well as the deinition o the substitution elasticity we derive the ollowing expression rom the ull employment condition or unskilled labour ( L = 0) : ( w w ) + λ σ ( w w ) 0 L σ L = λ (0) This model with ixed endowments can be solved explicitly or actor prices and outputs. A deviation rom the ull employment condition could be introduced by wage rigidity concerning the unskilled as well as the two skilled wage rates. In that case ( w = 0, w = 0, w = 0 ) actors were not ully employed and the model had to be solved or unemployment instead o actor prices. The price equations (6) and (7) can be inserted into the ull employment condition (0) to derive the solution or the wage rate (s. Jones, 97, 975): w = σ σ p λ + L pλl () ϑh ϑh where is the sum σ σ λ L + λl. Thereore the change in the unskilled wage rate is a ϑh ϑh weighted average o changes in commodity prices and the term σ ϑ i Hi represents the elasticity o the marginal product (Jones, 975) o the mobile actor in each industry. Combining the wage equation () with the price equations (6) and (7) we derive the solution or each skilled wage rate:

9 As w w σ σ σ θ L = p + λ L λl pλl () ϑh ϑh ϑh ϑh ϑh = p σ σ σ θ L λ + L λl pλl (3) ϑh ϑh ϑh ϑh ϑh ϑ > and H ϑ H >, a change in a commodity price induces a more than proportional change in the actor price o the speciic actor in that sector and we get the magniication eect or actor prices: ŵ > p ŵ p ŵ Here we assume that dierent countries specialize on dierent industries, which leads to the accumulation o sector speciic skills. A country with higher output in a sector has attained more experience and sector speciic knowledge accompanied by a higher actor reward to this sector speciic - human capital. The oreign (low wage) economy is only able to produce a ull composite unit o commodity (in terms o the home economy), i enough sector speciic capital is available. This is the prerequisite or the case o FGT, which leads to a decrease in the price o commodity. The impact o a reduction o commodity price on speciic actor rewards (in nominal terms) is straightorward rom () and (3): the speciic actor price in sector (w ) declines and the other (w ) rises. Changes in output induced by terms o trade changes can be easily derived by making use o condition (8), the elasticity o substitution deinition and the price equations (s.: Jones, 975 and Kohler, 99): X ϑ ( p w ) L = σ ϑh ϑ L ; X = σ ( p w ) ϑh (4) Equation (4) shows that expansion or contraction o commodity outputs depends on the dierence between commodity price changes and changes in the unskilled wage rate. Thereore in the case o a price decrease or commodity, sector contracts and sector expands.

10 I we assume that due to unions wage lexibility o skilled labour is restricted in the short term, the change in terms o trade (decrease o commodity price) would lead to an increase in equilibrium unemployment in sector accompanied by a decrease o the unskilled wage rate (as outlined above). This is accompanied by a decline o production in sector and an increase in sector. Obviously these results depend on the assumption o ully lexible wages or unskilled labour. I this were not the case the results in line with Davis (998) would be applicable and unemployment would also increase in the unskilled segment with consequences or the welare impact o trade liberalization. The outcome o FGT thereore clearly indicates long run adjustment costs, whereas the short term impact on unskilled labour might be positive in real terms due to the magniication eect and depending on consumption patterns. These potential gains or unskilled labour, the actual gains or sector and the actual losses or sector are directly proportional to the price decrease. Thereore or the FGT case the conclusion "the higher the gain, the higher the pain" can be drawn.. Outsourcing and Labour The model ramework has to be expanded in order to take into account ragmentation o the value added chain and international outsourcing. We start with the assumption that production in sector is ragmented without (in the beginning) assuming that ragments are actually sourced out. Production is in a irst instance organized as vertically integrated, but can be split up into two ragments. The setting is more restricted than in other recent studies that assume the easibility o outsourcing in all sectors (e.g. Egger and Falkinger, 003). Here we assume that only the more labour intensive sector (sector ) will produce in a ragmented way and might source out part o his production. The new production unctions are thereore given with ragments F and F in sector and the same unction as above in sector : X = [F (L, H ), F (L, H )] X = F (L, H ) (5)

11 It is assumed that both ragments in sector are combined by a Leontie technology o ixed input coeicients in order to produce output o commodity (assembling). Thereore substitution does not take place between ragments, but within each ragment and ull employment conditions are written as: h X + h X = H (6) h X = H (7) l X + l X + l X = L (8) The deinitions o the elasticity o substitution are the same as beore and we have σ, σ and σ. The next step is to deine the concept o OS used in this study. In the literature we ind dierent approaches. On the one hand continous OS without any indivisibility as in (Kohler, 004) is analysed, on the other hand we ind the concept o outsourcing as a discrete shit rom one technology to another and not as a marginal change in technology (Jones and Kierzkowski, 00, Egger and Falkinger, 003). In this study the second approach is chosen to represent OS and to dierentiate between vertically integrated production and production with OS. For vertically integrated production in sector the unit price equation is: p = (l + l )w + (h + h )w. Kohler (003) has clearly laid down the necessary conditions or shiting rom integrated production equilibrium to equilibrium cum outsourcing thereby deriving a margin o ragmentation and the Stolper-Samuelson disturbances brought about by OS. These results are the starting point o urther analysis in this model ramework. The dynamic price equation in sector (the OS sector) can be written by taking into account all changes caused by a shit rom integrated equilibrium to OS equilibrium. The actor shares in sector are now given with ϑl = ϑ Li and ϑh = ϑ Hi where i=, is the index o ragments. The actor shares i i o the ragments in the case o integrated production are deined with ϑ = and F ϑk, k

12 ϑ F = ϑk, with k= L,H as the index o actors. I the ragments are sourced out abroad k they shall be labelled as ϑ F and ϑ F respectively and deined as: w w ϑ hi (9) p Fi = li + p with i=, is the index o ragments. Here ϑ F is the actor share o ragment i produced abroad with oreign actor prices w and w. These actor shares o ragments are used here instead o the eective ragment prices in Kohler (003). Production o outsourced ragments thereore uses the same technology as integrated home production concerning the actor intensities, but there is a cost advantage due to lower actor prices. The oreign actor prices w and w are deined in a way to include all trade and communication costs linked to arms -length transactions and there is no urther ixed cost element o OS assumed. In this analysis home is the region, where OS can take place, i oreign actor prices are below domestic actor prices: w w ; w out abroad, but no OS takes place i w. In that case one o the two ragments is sourced w = w and w w =. The oreign economy, where one o these ragments is produced, might be imagined in a position where due to scarcity o sector speciic human capital no ull composite unit o total output o commodity can be produced. In the FGT case above the oreign economy is able to produce commodity in a vertically integrated process thereby threatening the home sector, when trade liberalization takes place. In the OS case availability o sector speciic human capital in oreign is limited due to less experience in the production o these products. Thereore oreign must specialize on one ragment and the threat or the home economy is that production o one ragment will be moved abroad entirely, i there is any actor price dierence. The proportions o the mobile actor are deined as beore and we now have λ L,, λ L, and λ L. I ragment is sourced out, it is entirely shited rom domestic integrated production to production abroad, so that in the price equation or sector we have l and ĥ = - and we get the additional cost element o production abroad, namely ϑ Fi.

13 The general price equation or commodity can thereore be represented by: p L, L, H, H, F L, H, = ( ϑ + ϑ ) w + ( ϑ + ϑ ) w + ( ϑ ϑ ϑ ) (0) Equation (0) is ormulated in a way to describe the case o integrated production as well as the OS equilibrium. That can be seen by expressing the additional net cost term in (0) by dierences in actor prices: λ L, h ϑ F ϑl, ϑh, = ( w w ) + ( w w ) () π p with π as the unskilled labour productivity share o sector (π = p X / L ) and π as total unskilled productivity: π = π + π. The additional net cost term shall be also reerred to in a condensed orm as (C C). In the case o integrated production actor prices are given by w = and w = w, w thereore the additional net cost term (C C) becomes zero. Dynamics in actor prices are in that case determined as in equation (6). In the case o OS actor prices are given by w < w and w < w, thereore ragment disappears, so that ϑ L, ŵ = 0 and ϑ H, ŵ = 0. Only the actor shares o ragment play a role in price equation (0) in that case and the additional net cost term (C C) becomes relevant. The price equations in the case o OS o ragment are thereore given with: λl, h = ϑ L,w + ϑh,w + ( w w ) + ( w ) () π p p w p ϑ Lw + ϑh w = (7) Equation () takes into account the Stolper-Samuelson disturbances o OS (Kohler, 003) and reveals how they depend on actor price dierences between home and oreign. The model is complemented by output equations and ull employment conditions or given endowments in the case o outsourcing o ragment :

14 X = h ; X h = (3) Equation (3) is the speciic orm o equation (9) due to ĥ = 0, once ragment has moved abroad. The ull employment condition or unskilled labour ( L = 0) is given by: L = 0 = l λ + l λ + X ( λ + λ ) + l λ + X λ ; (4) L, L, L, L, L L As ragment moves cross the border, l = - and l λ L, = - λ L, or a given unskilled labour proportion λ L, beore OS. This is the actor savings eect o OS, which has to be compensated by a wage decrease or unskilled labour. Again we could assume wage rigidity in all labour markets ( w = 0, w = 0, w = 0 ) and solve the model or unemployment instead o actor prices. Ater cross-border OS has taken place, urther output changes in sector can lead to positive employment eects o unskilled labour, but only in ragment. Thereore X λ 0 ater OS. Applying that to equation (4) we end up with the ollowing condition: L, = λ σ w w ) + λ σ ( w w ) λ 0 (5) L, ( L L, = The model can be solved or explicit expressions o actor prices in the case o OS, where no change in terms o trade takes place (i.e. p =0 and p = 0). The solution derived or the unskilled wage rate is: w [ λ σ ( C C λ ] = * L, ) L, (6) where is the sum σ σ λ L, ϑf + λl and (C-C L, ) = ( w w ) + ( w w ) ϑh, ϑh π p λ h. For large cost advantages (C-C ) the impact o OS on the unskilled wage rate might even be positive. The most important parameter in (6) is the proportion o ragment unskilled labour input λ L,. The larger this proportion, the higher the necessary decrease in the unskilled wage rate ater OS in order to restore ull employment equilibrium. This is the

15 consequence o the actor saving eect and can be thought o measuring the macroeconomic signiicance o the phenomenon OS. I any dierence in actor prices between home and oreign leads to total OS o one ragment, it is clearly the magnitude o this ragment in terms o the domestic labour market that matters. On the other hand we observe the pure cost saving eects in (6) determined by actor price dierences. Actually equation (6) allows to decompose the total unskilled wage impact into the (negative) actor savings impact ) and the (positive) pure cost saving eect ( [ λ σ ( C C )] ( λ * L, * L, ). The necessary decrease in the unskilled wage rate becomes smaller, when the cost savings eect becomes larger. This result is in line with Kohler s conclusion: the higher the gain, the lower the pain (Kohler, 004). The impact on the skilled wage rates can be derived rom the price equations by inserting (6): w = ( C C ϑ ) ϑ L, H, λ * L, ϑ σ + ϑ L, H, λ l, * (7) w [ λ σ ( C C + λ ] ϑ = (8) L * L, ) ϑh l, In (7) we observe that the cost advantage (C-C ) exerts a direct positive impact on the skilled wage rate in the OS sector and a negative impact as ar as it raises the unskilled wage rate. As ar as OS depresses the unskilled wage rate in order to restore ull employment equilibrium (depending on λ L, ) it also raises the skilled wage rate in the OS sector. The impact on the skilled wage rate in the other sector (equation (8)) only captures the repercussions rom the change in the unskilled wage rate: the cost advantage (C-C ) depresses the skilled wage rate in sector (as it raises the unskilled wage rate), whereas the actor savings eect (depending on λ L, ) raises it like in sector. The skilled wage rate in sector is in (8) expressed in terms o the variables o the OS version o the speciic actors model. This representation hides the aspect that changes in w are only determined by changes in w and could be directly derived

16 rom equation (7) as: w ϑ L = w. Thereore the impact o a given change in the unskilled ϑh wage rate on the skilled wage rate in sector is the same or the OS and the FGT case. I both speciic actor prices increase and the unskilled wage rate decreases, the input coeicients o skilled labour will decline due to substitution eects and allow both sectors to expand their production. As in the case o FGT we can solve the model or output changes and get: X ϑ L ϑf = σ ( C C ) w ϑf ϑh, ϑ = (9) L ; X σ w ϑh For the outsourcing sector the potential o expansion depends both on the cost advantage and on the decrease in the unskilled wage rate. For the other sector it is just a negative unction o the unskilled wage rate. Comparing equation (9) with the corresponding equation (4) o the FGT case, we observe that changes in the output o sector induced by changes in the unskilled wage rate are the same. Thereore we arrive at the conclusion that FGT with a price decrease in commodity and OS in sector exhibit the same impact on sector (in terms o wages and output) i they have the same impact on the unskilled wage rate. This impact depends on the macroeconomic dimension o the respective shocks (terms o trade and OS) The potential o expansion or both sectors generates an increase in real income (although unskilled wage incomes have decreased) and is equivalent to the result o the welare surplus o outsourcing emphasized by Arndt (997, 999) and Kohler (004). Production patterns thereore change as both sectors expand by dierent rates, but no decrease o production takes place. Downward rigid wages or skilled labour would not lead to an increase in skilled unemployment in this case and no increased labour reallocation is required or a new equilibrium, what means less adjustment costs. Anyway we would incur unskilled unemployment like in the FGT case, i the unskilled segment o the labour market were not ully competitive. Thereore we can consider OS as an attractive alternative to FGT rom a labour market-perspective, i downward wage rigidity or sector speciic labour is an important actor or unemployment. It must be emphasized however that according to the

17 assumption in this analysis OS is only easible, i the oreign economy lacks sector speciic human capital and cannot produce an entire unit o the inal good. In the opposite case inal goods can be produced in the oreign economy and compete with domestic production (the FGT case). 3. A Numerical Example A more precise comparison between the two cases depends on empirical values o variables and parameters. For this purpose a numerical example is chosen here using the stylized acts o Austrian OS to Eastern Europe ater the all o the Iron Curtain. The stylized acts o Austrian OS to Eastern Europe are well described and compiled in Egger, Paermayr and Wolmayr-Schnitzer (00) and shall be taken here as a starting point. They describe the pronounced rise o OS rom Austria to Eastern Europe during the nineties leading to a share o aggregate imported inputs in aggregate gross output o about 0 percent in 998. Obviously the share o aggregate imported inputs rom Eastern Europe is much smaller only amounting to. percent in 998. At the level o industries Egger, Paermayr and Wolmayr-Schnitzer (00) ind considerable dierences o OS (deined as imported inputs) to Eastern Europe and also a large increase concentrated on a ew sectors. They conclude with a positive (though insigniicant) correlation between the relative skill intensity ( H / L in terms o the model applied here) and the change in OS during the nineties. The data or relative skill intensity by industry or the numerical example are taken rom Egger, Paermayr and Wolmayr- Schnitzer (00) and their calculations about OS (relying on the Austrian input output tables or 990 and 995) are complemented here with data rom the Austrian input output tables or 000. In the literature we ind a discussion about the correct measure o OS, more speciically on the necessary narrowness o the OS concept. Total imported inputs might be seen as a too broad concept as they also include complementary imports (e.g. energy) that could never have been produced by the sector. An alternative thereore is to rely on the main diagonal o the import use matrix, i.e. taking only the imports o the same products as the sector produces. Applying both concepts to recent Austrian data (Table ) we ind three clusters, where OS i i

18 seems important or the Austrian economy. One is the textile&clothing sector, where both total imported inputs as well as own imported inputs are important. The industry wearing apparel shows a much smaller amount o own imported inputs, but (not shown here) a large raction o textile imports. As a sharp distinction and borderline between these two industries might be diicult, we think that wearing apparel also its well in the scheme o high total as well as high own imported inputs. >Table : Outsourcing by Industries in percent o Output (Austrian Input-Output Table 000) A second cluster o high total as well as own imported inputs consists o raw material intensive sectors (basic metals and chemicals), where OS might be accompanied by oreign direct investment and location o production in dierent subsidiaries o multinational companies. A urther cluster can be identiied, where assembling plays an important role (electrical machinery, radio/television and communication equipment, mother vehicles and other transport equipment). These industries might best represent the case assumed or the theoretical analysis in this study, where the oreign economy due to scarcity o sector speciic human capital cannot produce one entire unit o the inal good, but can produce certain ragments at much lower wage costs. All these data are combined in order to design the numerical example, where the Austrian economy is split up into two sectors. The actor proportions λ Li are 0.33 or i = and 0.67 or i =. The actor shares θ ji are: given as: θ Li = 0.5 or i = and 0. or i = (the corresponding values or θ Hi are: θ Hi = 0.75 or i = and 0.8 or i = ). All substitution elasticities σ ij are in a irst step set equal to unity corresponding to the Cobb-Douglas case. In a second step some sensitivity analysis has been undertaken or the case o considerably lower elasticities. Prices and the wage rate are normalized, rom which the skilled wage rates in both sectors ollow (w =.04; w =.6). Both sectors exhibit the same input coeicients or skilled labour, namely h = h = and are only slightly dierent in terms o unskilled actor intensities: l = 0.5 and l = 0.. This assumption relects the stylized acts on the one

19 hand, but also mostly reduces the inluence o the sector, where outsourcing takes place. Foreign actor prices have been assumed at the level o 40 percent o domestic actor prices corresponding to actual data about the average per capita income o Eastern European neighbours o Austria: w = 0.4; w = 0.8. This is directly relevant or the OS case, as it determines the cost advantage (note that input coeicients or the ragment sourced out are the same abroad). For the case o FGT shocks in the price o commodity are introduced and gradually increased. The magnitude o this price decrease depends on the product o lower actor prices and lower productivity in oreign. I as lined out above - actor prices are at 40 percent o the domestic economy and (as data show) average productivity o Eastern European neighbours o Austria is about two thirds o Austrian productivity, then in our numerical example the maximum price shock or commodity would be about 40 percent. For the case o OS dierent shocks in terms o the distribution o total sector employment between both ragments are introduced, starting with λ L, = 0.05 up to λ L, = 0.3. The higher λ L, becomes, the more pronounced is the shock o OS or the domestic unskilled labour market. >>Table : Macroeconomic Impacts o Trade Liberalization (dierent shocks in terms o price decreases, (- p ) ) >>Table 3: Macroeconomic Impacts o Outsourcing (dierent shocks in terms o the unskilled labour proportion o ragment, ( λ L, ) ) All calculations have been carried out according to the corresponding ormulae o the theoretical analysis, i.e. applying constant share variables θ ji and λ ji. Obviously this is only ully correct or marginal changes in variables. The results in Table assert the main results rom the theoretical analysis. Especially the magniication eect can be observed leading to large losses or skilled labour in the import competing sector. The adjustment costs o such a

20 shock are apparent rom these changes in the income distribution as well as rom changes in production patterns. From Table 3 we observe that OS might become a signiicant problem or the unskilled labour market, i the activities o a large part o the total unskilled labour orce can be sourced out cross the border. Starting with low proportions o the unskilled labour orce that are sourced out we conirm the result o a possible positive impact on unskilled labour rom the theoretical analysis. This result simply relects the consequence o a large cost eect o OS (given by the actor price dierences between home and oreign ) in relation to the actor savings eect. As the positive impact on sector wages and output only materializes, i unskilled wages decrease, or these cases with a low actor savings eect we ind adjustment costs or skilled labour and changes in production patterns also or OS. The numbers in Table show, that outsourcing might amount up to a maximum o 50 percent o output in a sector. Obviously no direct values or the share o OS activities in employment o a sector are available, as we cannot directly observe the actor intensity o the single ragments. I outsourcing also amounts 50 percent o unskilled employment in sector, this would yield L = 0 and λ L, = One sixth o the total unskilled labour orce could be lost in that case by cross border OS. From Table 3 we get or that case a decrease in the unskilled wage rate o 6.3 percent, whereas the skilled wage rate would increase in the outsourcing sector by 6. percent and in the other sector by.6 percent. Output would increase in both sectors indicating a net welare gain and higher aggregate income. Searching or a similar impact on the unskilled wage rate in Table ( ŵ =-6. percent), yields the case o a 7.5 percent price decrease in sector with a loss or the skilled wage rate in sector o more than 0 percent. From the theoretical analysis we know that or cases o FGT and OS with the same impact on the unskilled wage rate the results or sector skilled wages and output are also the same. This is reasserted here or both cases with ŵ =.5 (.6) percent and X =.5 (.6) percent or ŵ =-6. (-6.3) percent. For the same (negative) impact on unskilled labour FGT and OS have the same inluence on the other sector, but inal goods trade harms the sector where it occurs whereas the OS sector

21 beneits. Although this analysis does not directly treat with welare implications, this aspect can be seen as a Pareto advantage o outsourcing. The mid-term consequences expected o this dierence are higher adjustment costs o inal goods trade. Although the underlying variables or this numerical example have been derived rom sytlized acts or OS in Austria, this is not the case or the substitution elasticities. One could argue that due to the segmentation o the labour market in skilled (in our case: sector-speciic) and unskilled labour the substitution elasticity in the production process is also much lower than unity. For this reason a sensitivity analysis was carried out, where all σ ij are set equal to 0.5. For both cases (FGT and OS) we ound compensating impacts o these changes in the dierent ormulae as well as compensating impacts o the change in the elasticities in both sectors, i.e. we would get larger changes i the substitution elasticities dier in both sectors. The results or FGT in the case o σ ij = 0.5 only change concerning the output eects, which are on average ive times smaller than in the case with σ ij =. The results or OS in the case o σ ij = 0.5 only change concerning the impact on the unskilled wage rate, which turns out to be on average double o the case with σ ij =. All other impacts on income distribution are the same or FGT and OS as in the case with σ ij =. 4.Summary In this paper a speciic actors model is set up in order to compare the impacts o inal goods trade (FGT) and outsourcing (OS) on the labour market and on production patterns. One motivation o the speciic actors model is the idea o countries specialization leading to the accumulation o sector speciic skills. I a oreign low-wage economy is able to produce a ull composite unit o one commodity (in terms o the home economy), inal goods are traded and oreign imports compete with domestic production. I this is not the case due to scarcity o sector speciic human capital in oreign, OS will take place. The impacts on income distribution and on production patterns o the two cases are signiicantly dierent. In the FGT case the unskilled wage rate declines and the skilled wage

22 rate declines in the import competing sector and rises in the other sector. Production in the import competing sector declines and rises in the other sector indicating a necessary transer o resources between sectors. I due to labour market institutions skilled wages establish at a level too high or ull employment, FGT would lead to an increase in equilibrium unemployment. In the case o OS the impacts depend on the interplay o the actor saving and the cost saving eect. I actor price dierences between home and oreign are not very large, the unskilled wage rate declines and skilled wage rates in both sectors rise, indicating that in the case o rigid skilled wages no increase in unemployment would occur. Also production in both sectors rises indicating a net welare gain o OS. For cases o FGT and OS with the same impact on the unskilled wage rate the impact on the other sector is the same as well, but only with OS the directly aected sector also beneits indicating an additional Pareto advantage o outsourcing. Simulations with a numerical example capturing the stylized acts o Austrian OS to Eastern Europe clearly reassert the results rom the theoretical model. In terms o the labour market OS might be characterized as an adjustment strategy to globalization or economies with unemployment due to skilled wage rigidity and restricted mobility o skilled labour. As the impact on the unskilled wage rate is the same in both cases, these results are only valid with downward lexible wages or unskilled labour. I this condition is not ulilled and unemployment in an economy is mainly due to unskilled wage rigidity the results might dier rom this analysis. This issue should be urther investigated in a speciic actors-model with wage rigidity and unemployment.

23 Reerences Arndt, S.W. (997), Globalization and the Open Economy, North American Journal o Economics and Finance, 997, 8(), S Arndt, S.W. (998), Globalization and Economic Development, The Journal o International Trade and Economic Development, 999, 8(3), S Brecher, R., A. (974), Optimal Commercial Policy or a Minimum Wage Economy, Journal o International Economics, 4 (), Brülhart, M., Elliott, R., J. (00), Labour-Market Eects o Intra-Industry Trade: Evidence or the nited Kingdom, Review o World Economics, 38 (), Davis, D.R. (998), Does European nemployment Prop p American Wages? National Labor Markets and Global Trade, American Economic Review, 88(3), Deardor, A. (00), Fragmentation Across Cones, in: Arndt Sven W., Henryk Kierzkowski (eds.), Fragmentation. New Production Patterns in the World Economy, Oxord niversity Press, 00. Egger, H. (00), International Outsourcing in a Two-Sector Heckscher-Ohlin Model, Journal o Economic Integration, 7 (4), 00, Egger, H. and Egger, P. (003), Outsourcing and Skill-Speciic Employment in a Small Economy: Austria ater the Fall o the Iron Curtain, Oxord Economic Papers, 55 (4), 003, Egger, H. and Falkinger, J. (003), The Distributional Eects o International Outsourcing in a x-production Model, North American Journal o Economics and Finance 4 (), 003, Greenaway,D., Haynes,M. and C. Milner (00), Adjustment, Employment Characteristics and Intra-Industry Trade, Review o World Economics, 38 (),

24 Jones, R. W (97), A Three-Factor Model in Theory, Trade and History, in: Bhagwati, N.J. et.al. (eds.), Trade, Balance o Payments and Growth, Papers in International Economics in Honour o Charles P. Kindleberger (Amsterdam: North-Holland), 97. Jones, R.W. (975), Income Distribution and Eective Protection in a Multicommodity Trade Model, Journal-o-Economic Theory. 975; : -5 Jones, R.W., (996), International Trade, Real Wages, and Technical Progress: The Speciic- Factors Model, International-Review o Economics and Finance. 996; 5(): 3-4 Jones, R. W and H. Kierzkowski (00), A Framework or Fragmentation, in: Arndt Sven W., Henryk Kierzkowski (eds.), Fragmentation. New Production Patterns in the World Economy, Oxord niversity Press, 00. Kohler, W. (99), Income Distribution and Labour Market Eects o Austrian Pre- and Post Tokyo-Round Tari Protection, European Economic Review, 35 (), 99, Kohler, W. (00), A Speciic-Factors View on Outsourcing, North American Journal o Economics and Finance, (), March 00, Kohler, W. (003), The Distributional Eects o International Fragmentation, German Economic Review, 4 (), Kohler, W. (004), Aspects o International Fragmentation, Review o International Economics, (5), Lovely, M. E. and D.R. Nelson (00), Intra-Industry Trade as an Indicator o Labor Market Adjustment, Review o World Economics, 38 (), Neary, J. P (978), Short-Run Capital Speciicity and the Pure Theory o International Trade, Economic Journal, 978, (88), Seidel, T. (005), Welare Eects o Capital Mobility with Rigid Wages, Applied Economics Quarterly: Konjunkturpolitik 5 (Supplement), 8.

25 Tables Table : Outsourcing by Industries in percent o Output (Austrian Input-Output Table 000) Imported Imported Inputs (total) Inputs (own) Mining and quarrying products 8,7 0,6 Food products and beverages,4 3, Tobacco products 5,9,5 Textiles 34,3 8, Wearing apparel; urs 3,4 6,0 Leather and leather products 4, 3,4 Wood and wood products 7,8 6,0 Pulp, paper and paper products 5,7 6, Printed matter and recorded media 6,9,3 Coke, reined petroleum products 55,,4 Chemicals, chemical products 3,9 3,8 Rubber and plastic products 3,9 4,6 Other non-metallic mineral products 5,0 4,8 Basic metals 38,0 6,6 Fabricated metal products,4 6,3 Machinery and equipment 7,4 3, Oice machinery and computers 36,3 8,4 Electrical machinery and apparatus 30,6 3,3 Radio, television, communication equipment 37,0, Medical, precision and optical instruments 0,4 6,5 Motor vehicles 55,8 38,3 Other transport equipment 33,3 9,4 Furniture; other manuactured goods 0, 3,4

26 Table : Macroeconomic Impacts o Trade Liberalization (dierent shocks in terms o price decreases, (- p ) ) p ŵ ŵ ŵ X X Table 3: Macroeconomic Impacts o Outsourcing (dierent shocks in terms o the unskilled labour proportion o ragment, ( λ L, ) ) λ L, ŵ ŵ ŵ X X

27 006 Österreichisches Institut ür Wirtschatsorschung Medieninhaber (Verleger), Hersteller: Österreichisches Institut ür Wirtschatsorschung Wien 3, Arsenal, Objekt 0 A-03 Wien, Postach 9 Tel. (43 ) Fax (43 ) Verlags- und Herstellungsort: Wien Die Working Papers geben nicht notwendigerweise die Meinung des WIFO wieder Kostenloser Download: