A Tale of Two Peripheries: Real Wages in Denmark and New Zealand David Greasley School of History and Classics University of Edinburgh

Transcription

1 A Tale of Two Peripheries: Real Wages in Denmark and New Zealand David Greasley School of Hisory and Classics Universiy of Edinburgh Jakob B Madsen Deparmen of Economics Universiy of Copenhagen Absrac. Denmark achieved dramaic real wage growh afer 1870 compared o oher European economies and o hose of he New World. The ingrediens of Denmark s success are gauged by comparison wih one is major compeiors in he Briish foodproducs markes, New Zealand. Faser Danish produciviy growh explains only par of Denmark s faser real wage growh. Open economy forces, chiefly inernaional capial flows before 1913, and mos especially Danish rade union miliancy around he end of World War I, influenced income disribuion and especially favoured wage over propery income in Denmark. Denmark s GDP per capia equalled New Zealand levels beween he world wars bu her real wages surged pas hose of New Zealand as disribuional shifs favoured Denmark s wage earners. 1. Inroducion The European periphery experienced varied real wage convergence wih Europe s economic leader, Grea Briain, before World War I. Scandinavia s cach-up was he mos specacular, wih real wage growh here around wo and one-half imes faser han in Briain Ariculaing he ingrediens of Denmark s, Sweden s, and Norway s 1

2 success has informed he wider debaes on he effecs of globalizaion. The key forces are well-known, and he ou-migraion of labour, he inflow of capial, and commodiy rade flows, all raised Scandinavia s real wages compared o Briain, wih capial flows playing he dominan role. Ye, collecively, hese open economy forces accouned for only around one-half of he faser real wage growh of Scandinavia compared o Briain , and echnological cach-up and human capial formaion played equally imporan roles. 1 Wheher he chief impeus behind Scandinavia s real wage cach-up was from counry-specific produciviy forces or from cross-counry facor and commodiy flows does no emerge very clearly from he comparison wih Europe s leader, Briain. Comparisons wih he world leader, he USA offer more suppor for open economy forces narrowing gaps, principally because he mass inflow of labour ino USA reduced real wage differenials wih Scandinavia. 2 Inerpreaions of Scandinavia s real wage cach-up are hus influenced by he economic leader chosen as he comparaor. Here we adop a differen ype of economic leader as he comparaor, namely New Zealand. I is wellknown ha some New World counries had paricularly high real wages around The case of New Zealand reinforces ha he wesern off-shoos had high real wages, for example Denmark s hourly real wage (for unskilled urban labour) was half ha of New Zealand in 1873, and Sweden s and Norway s real wages were lower sill. 4 Comparing real wages in Denmark wih hose of a small, peripheral New World economy, New Zealand has he poenial o furher inform he discussion surrounding he consequences of globalizaion. The wo counries share he characerisics of being small 1 O Rourke, K. H. and J. G. Williamson, Around he European Periphery : Globalizaion, Schooling and Growh. European Review of Economic Hisory, vol. 1, 1997: 2, O Rourke and Williamson, Around he European Periphery, Williamson, J. G., The Evoluion of Global Labour Markes since 1830: Background Evidence and Hypoheses, Exploraions in Economic Hisory, vol. 32, 1995:2, Greasley, D. and Oxley L., Globalizaion and Real Wages in New Zealand , Exploraions in Economic Hisory vol. 41, 2004: 1,

3 generally open economies wih similar economic srucures, whose main expor marke was Briain. Using wo economies ha share hese feaures as comparaors has he advanage ha any convergence ha migh have been engendered by differences in srucure is lessened, and accordingly he role of inernaional forces will be highlighed. New Zealand s economic developmen resembled ha of Denmark, despie heir differen facor endowmens, once refrigeraion made possible he long disance expor of dairyproducs and mea. As in Denmark, farming efficiency in New Zealand was fosered by sae-sponsored educaion and, in he case of buer and cheese manufacure, by co-operaive enerprises. 5 The wo counries became major compeiors in he Briish markeplace. In 1891 New Zealand s buer expors o Briain were around 3% of he Danish level, bu his raio had reached 75% by Ye while New Zealand made subsanial inroads ino he Dane s share of he Briish dairy marke, Denmark achieved a dramaic real wage cach-up wih is disan compeior. By World War I New Zealand s real wage lead had been subsanially eroded, and Denmark s real wages surged ahead hereafer. Conversely, land prices fell relaively in Denmark. Shifs in he wage-renal raio have poenially powerful influence on income disribuion and hus on real wages in he wo counries. Disribuional shifs appear o have been srong boh before World War I and mos especially during is immediae afermah, and hey favoured wage earners in Denmark and landowners in New Zealand. Dairy land prices in New Zealand were around 68% of he Danish level in 1924, bu he rural land price raio had been around 14% in he 1870s. 6 Income disribuion shifs driven by he open economy forces of migraion, capial flows and rade, may hus have influenced relaive real wages. Our analysis also considers 5 Condliffe, J. B., New Zealand in he making. London: Allen and Unwin, See Figure 2, and foonoe 5 below. 3

4 he effecs of produciviy cach-up and counry-specific disribuional shifs associaed wih labour miliancy, on relaive Denmark-New Zealand real wages. 2. Real Wages and Income Disribuion Williamson has argued ha global rade expansion afer 1870 was accompanied by ranspor-driven commodiy price convergence, and ha he concomian magnified shifs in relaive facor prices unleashed powerful and ubiquious disribuional forces. 7 Wagerenal raios fell in he land abundan counries of he periphery, and rose in land scarce Europe. Shifs in real wages were hus influenced by he income disribuional effecs of rises or falls in he wage-renal raio. For less indusrialized counries like Denmark and New Zealand, real wages were influenced by he share of income accruing o landowners. Commodiy and facor price equalizaion hus ended o augmen real wages in land scarce European economies, including Denmark. There appears lile doub ha rade, and he migraion of capial and labour, by affecing relaive facor prices, had poenially imporan consequences for income disribuion around he world economy Generally, he price of labour fell relaively in he New World, and rose in he Old, and conversely he price of land rose relaively in he New World, and fell in he Old. Denmark and New Zealand fi broadly ino his characerizaion, alhough each counry had peculiariies, which may have empered he shifs in relaive facor prices, and influenced heir real wages. Mos obviously worker-miliancy was srong in Denmark especially in he years around World War One and his may have had powerful disribuional effecs. Moreover, while Denmark experienced labour ouflows and New Zealand labour inflows, boh were recipiens of capial inflows. And while Denmark largely adhered o 7 Williamson, J. G., Land, Labor and Globalizaion in he Third World, , Journal of Economic Hisory, vol. 62, 2002: 1, 55-85, p

5 free rade for agriculural goods and reduced ariffs on manufacures, New Zealand imposed modes ariffs on manufacured impors, including on goods from Briain. Furher, Denmark remained a major farm goods exporer, despie refrigeraion evenually faciliaing compeiion from New World economies in dairy and mea producs. The myriad of forces shaping he relaive facor price shifs are furher complicaed by he uncerain effecs of echnological progress. Generally, echnological progress is anicipaed o raise he wage renal raio. 8 For New Zealand however, refrigeraion and dairy-relaed echnology appears o have principally benefied he reurns o landowners. 9 Inuiion offers only limied guidance o he magniude of likely shifs in relaive facor prices, and hus of how Denmark s and New Zealand s real wage movemens were influenced by open economy forces in he years o Subsequenly, he effecs of World War I severely disruped he facor and produc prices of boh counries, which powerfully influenced income disribuion and real wages. Denmark, wih a German land border, was neural, and coninued o expor, a rising prices (alhough volumes fell), bacon and buer o Briain and horses o Germany. 10 Disan New Zealand was a belligeren, and evenually all of is pasoral expors were commandeered, a high prices by Briain. 11 The Briish Dominion also experienced labour shorages resuling from high raes of miliary paricipaion, which may have amelioraed he fall in he wage-renal raio in New Zealand during he war. In conras, Denmark encounered rade union miliancy oward he war s end and in he immediae pos-war years, and concomianly experienced a sharp increase in is wage-renal raio. 8 O Rourke, K., Taylor, A. and J. G. Williamson, Facor Price Convergence in he lae Nineeenh Cenury, Inernaional Economic Review, vol. 37, 1996: 3, These auhors se ou a model which shows he wage-renal raio influenced by he erms of rade, facor endowmens and residual produciviy. 9 Greasley, D. and Oxley L., Refrigeraion and Disribuion: New Zealand Land Prices and Wages. Ausralian Economic Hisory Review, vol. 45, 2005: 1, Smid, H. R., Duch and Danish Agriculural Expors during World War I. Scandinavian Economic Hisory Review, vol. XLIV, 1996:2, Hawke, G. R., The Making of New Zealand. Cambridge: Cambridge Universiy Press,

6 Beween he world wars Denmark and New Zealand became fierce compeiors in he Briish food-producs marke, and boh were adversely affeced by unfavourable erms of rade, alhough New Zealand benefied from preferenial ariffs during he 1930s. Land renals ended o fall in he 1920s, especially in New Zealand, bu currency manipulaion in he Briish Dominion in1933, which favoured he reurns o landowners, probably raised he relaive prices of New Zealand land. 12 Open economy forces before 1914, he disrupion of World War I and is afermah, and he difficul rade environmen beween he world wars, had complex and varied implicaions for income disribuion and real wages in Denmark and New Zealand. To make progress in disenangling he forces shaping heir comparaive real wages we firs need o consider more fully he movemen of he relevan variables. 2.1 Esimaes of real wages. The wage series used are for urban unskilled labour, and ha for New Zealand uilizes recenly consruced daa. 13 For Denmark, he esimaes of Williamson are modified here, o adjus for a break in he series over he years of World War I. 14 Relaive Denmark-New Zealand real wages, adjused for difference in purchasing power are esimaed for 1905, when, aking Briain s real wage as 100, Denmark s level is 90 and New Zealand s New Zealand s real wage was 21.1% above Denmark s in 1905, and he raio (spread) shown in Figure 1 projec each counry s real wage from his benchmark. 12 Greasley, D. and Oxley, L., Regime shif and fas recovery on he periphery: New Zealand in he 1930s. Economic Hisory Review, vol. LV, 2002: 4, Greasley, D. and Oxley L., Globalizaion and real wages in New Zealand , Exploraions in Economic Hisory, vol. 41, 2004:1, 26-47, Greasley and Oxley Refrigeraion and Disribuion. 14 Williamson, The Evoluion. 15 These esimaes ake accoun of purchasing power differences, and in paricular he higher prices of mos consumer goods in New Zealand. The esimae of Denmark s real wage relaive o Briain is from Williamson, The Evoluion, and ha for New Zealand relaive o Briain is from Greasley and Oxley, Globalizaion and real wages. 6

7 Figure 1. Real Wages DK Spread NZ Sources. New Zealand: Real Wages, Greasley and Oxley Globalizaion and real wages, Greasley and Oxley Refrigeraion and Disribuion. Denmark: Nominal wages, Johansen, H. C., Dansk hisorisk saisik , København: Gyldendal, Consumer prices, Pedersen, P. J., Arbejdslønnen i Danmark Under Skifende Konjunkurer i Perioden ca , København: Akademisk Forlag, 1930, and Johansen, Dansk hisorisk saisik. The resuls show a clear real wage caching-up by Denmark , wih New Zealand s iniial real wage advanage of around 100% being largely eliminaed by Then, World War I and is immediae afermah had dramaic effecs on he raio, principally because of sharp flucuaions in Denmark, where real wages fell in he early war years bu winessed remarkably fas growh , when hourly real wages here doubled. 16 Consequenially, Denmark s real wages surged ahead, and New Zealand s real wages were around 75% of he Danish level in he 1920s. Thereafer, he real wage 16 Denmark s wage daa are available on a consisen basis over he whole period excep for 1915 and 1916, where we have linearly inerpolaed he daa. The generaed real wage series coincide wih he real wage daa of Williamson, The Evoluion, wih he excepion in , where here is a break in Williamson s series, which corresponds o a break in Johansen s Dansk hisorisk saisik consumer price series. However, Pedersen s Arbejdslønnen i Danmark consumer price index includes In fac hree differen consumer price indexes are available before 1914 inclusive, and hey all show he same increase in consumer prices over he period from 1913 o 1914 (see Table 7.1 in Chrisensen, J. P., Lønudviklingen inden for dansk håndværk og indusri , København: Akademisk Forlag, Accordingly, he pre-1913 and pos-1914 real wage daa for Denmark can be linked. 7

8 relaively was broadly sable beween he world wars, wih Denmark susaining mos of is lead. Inspecion of he daa indicaes hree broad phases of Danish real wage movemens compared o New Zealand: caching-up , surging ahead , and susaining he lead Esimaes of Land Renals The esimaes of land renals uilize daa for land prices, using a series recenly consruced for New Zealand. 17 The Denmark series exends ha repored for occasional years by O Rourke e al o 1939, making use of he annual daa. 18 Russell and Macklin in heir comparison of dairying in New Zealand and Wisconsin, repor dairy land prices for New Zealand in 1925 a US$225 per acre (or 50% higher han dairy land prices in Wisconsin in ha year). 19 Danish dairy land prices averaged 1550 kroner or US$326 per acre in The relaive series shown in Figure 2 use hese daa as a benchmark, and show ha he New Zealand land price was 68.8% of he Danish level in Real land prices did rise in Denmark , bu only by around one-quarer, while New Zealand real land prices rose by over five imes. New Zealand land prices reached around 75% of he Danish level in he decade prior o A collapse in real land prices in Denmark occurred in he rans-war period , of around 40 per cen. In 17 Greasley and Oxley, Refrigeraion and Disribuion. 18 O Rourke, Taylor, and Williamson, Facor price convergence. 19 Russell, H. L. and T. Macklin, Inensive Dairying in New Zealand and Wisconsin. Agriculural Experimen Saion Bullein 377, Universiy of Wisconsin, Maddison, Denmark land prices are from Chrisensen, J., Landbosaisik: Håndbog I Dansk Landbohisorie, Saisik , København: Landbohisorisk Selskab, 1985 and various issues of Saisisk Årbog. The daa are lised in kroner per barrels of land (Tønder land). The daa are convered o acre, where one Tønde land = acre. 21 US$ prices are no adjused for PPP when comparing dairy land prices in Denmark and New Zealand. Such a comparison may undersae he relaive Danish price in he 1920s. For example, for 1929 L. Prados de la Escosura, Inernaional comparisons of real produc : An alernaive daa se, Exploraions in Economic Hisory, vol. 37, 1990: 1, 1-41 repors New Zealand GDP per capia based on nominal exchange raes was 128% of he Danish level, bu adjusing for PPP and aking he average of he hree repored measures shows he raio of New Zealand o Denmark GDP per capia was 110%. Applying he same average PPP adjusmen o 1925 land prices suggess he raio of New Zealand o Denmark land prices in ha year may have been as low as 59.1%. 8

9 conras, New Zealand experienced land price inflaion during and in he afermah of World War One. Temporarily, in 1919, dairy land prices in New Zealand exceeded hose in Denmark, bu he rural land marke in New Zealand collapsed in Subsequenly, Denmark s real land price rose relaive o ha of New Zealand in he 1920s and early 1930s, bu hen fell in he laer 1930s when currency manipulaion in New Zealand conribued o he improving posiion of land-owning farmers. New Zealand s real land price recovered o around 96 per cen of is 1913 level by While real wages in Denmark caugh-up and surged ahead of hose in New Zealand over he years , New Zealand land prices rose from 10% o around 80% of Denmark s level over he same period. 1.2 Figure 2: Real Land Prices Spread NZ DK Sources. New Zealand: Real land prices, Greasley and Oxley Globalizaion and real wages, and Refrigeraion and Disribuion, Russell and Macklin Inensive Dairying. Denmark: Consumer prices, Pedersen, Arbejdslønnen i Danmark and Johansen, Dansk hisorisk saisik, land prices Chrisensen, Landbosaisik and Saisisk Årbog. 9

10 2.3 Wage-Renal Raio Over he years he wage-renal raio fell in New Zealand, bu rose sharply in Denmark, see Figure 3. The low poin of New Zealand s wage-renal raio was in 1919, and he shifs in income disribuion owards land owners were powerful over he years Subsequenly, he wage-renal raio rose in New Zealand. This did no fully resore he 1875 raio, bu disribuional shifs in New Zealand over he longer period were comparaively modes. Disribuional shifs in Denmark were susained over a longer period, and srongly favoured wages. 22 Urban wage earners did relaively well in Denmark before 1914, bu rural wealh as measured by he capial value of land did no collapse. The more dramaic income disribuion shif in Denmark occurred in he afermah of World War I, when land prices fell sharply while real wages showed remarkable growh. Denmark s wage-renal raio increased, by more han 100% Mean = Figure 3. Wage/Renal Raio, NZ/DK DK NZ Sources: as for Figures 1 and Phelps Brown and Browne Phelps Brown, E. H. and Browne, M. H., A Cenury of Pay. London: Macmillan, 1968 char more generally he rise in labour s income share across he rans-war period. 10

11 3. Uncondiional Convergence Mos empirical invesigaions of convergence have deployed cross-secional ess o consider wheher or no income or real wage growh is negaively correlaed wih iniial values of income or real wages. 23 O Rourke and Williamson repor some evidence of uncondiional real wage convergence among a group of 16 counries , alhough he negaive coefficien hey observe is no saisically significan a convenional levels. 24 Their sample of 16 counries did no include New Zealand, and osensibly, adding a counry wih high 1870 real wages bu relaively slow real wage growh in he sample would reinforce he evidence in favour of convergence. Our ineres hough concerns he specific experiences of Denmark and New Zealand, and hence we consider a bivariae ime series es for convergence since cross-secional ess do no show which, if any, of he counries among a group are converging. 25 The ess for real wage convergence uilize approach of Bernard and Durlauf. 26 Their ess of convergence are based upon he ime series properies of wo (or more) series. Time series es have generally rejeced uncondiional convergence. Greasley and Oxley hough found evidence of long run real GDP per capia convergence for Denmark and Sweden In he case of wo counries he ess consider wheher or no he differences in wo series have a uni roo and a non-zero mean, and hus wheher 23 Abramoviz, M., Caching up, forging ahead, and falling behind. Journal of Economic Hisory vol. 46, 1986:3, , Barro, R., Economic growh in a cross-secion of counries. Quarerly Journal of Economics vol. 106: 1991: The descripors uncondiional and condiional refer o wheher or no esimaes of cross-counry convergence conrol for counry specific differences in for example levels of educaion. In a ime series framework uncondiional convergence implies incomes equalize irrespecive of counry specific differences, say in educaion. 25 Bernard, A. and S. Durlauf, Inerpreing ess of he convergence hypohesis. Journal of Economerics vol. 71, 1996: 1, Bernard, A. and S. Durlauf, Convergence in inernaional oupu. Journal of Applied Economerics vol. 10, 1996: 1, Greasley, D. and L. Oxley, Time series based ess of he convergence hypohesis: Some posiive resuls, Economics Leers vol. 56, 1997:

12 movemens in income differences permanenly deviae from equaliy, which obviously precludes convergence. Uncondiional convergence requires ha long run forecass of real wages (or real GDP per capia and land prices) differences in Denmark and New Zealand end o zero as he forecasing horizon ends o infiniy. For non-saionary series his requires ha he wo series are coinegraed wih a coinegraing vecor [1, -1]. If uncondiional convergence is rejeced common forces may sill have shaped he real wages (or land prices or real GDP per capia) of he wo counries if long-erm forecass of real wages in Denmark and New Zealand moved proporionally, which implies he wo series are coinegraed, bu wih a coinegraing vecor [1, -a]. Inuiively, his implies ha he real wages (or land prices and GDP per capia) of he wo counries will be driven by common influences, bu hey will no converge uncondiionally wih a zero mean difference. Table 1. Coinegraion ess for convergence of real GDP per capia, real wages and real land prices. Dickey-Fuller ess. Real Wages Consan, No rend Consan, rend Coinegraion Vecor -3.69(-3.34) -4.24(-3.78) [1,-0.22]* Real Land Prices Consan, No rend Consan, rend -2.90(-3.34) -2.59(-3.78) Real GDP per capia Consan, No rend Consan, rend -1.84(-3.34) -2.20(-3.78) Noe. The numbers in parenheses are criical values a he 5% level. Esimaion period: * The hypohesis of a coinegraing vecor [1, -1] is rejeced a any convenional significance level ( χ 2 (1) = 1706 ). Sources. See he Daa Appendix. 12

13 The resuls in Table 1 rejec uncondiional real wage, real land price, and real GDP per capia convergence for Denmark and New Zealand. Faser Danish real wage growh herefore canno herefore be simply explained by convergence o New Zealand real wage levels. However, real wages did end o move ogeher in New Zealand and Denmark, and he coinegraion resul show ha common shocks drove real wage movemens in he wo counries. Possible common forces include echnological or insiuional shocks, for example ranspor or dairy echnology innovaions, or co-operaive enerprise, as well as hose associaed wih open economy influences. In conras, in he cases of land prices and real GDP per capia no evidence is found o show ha hese variables ended o move ogeher in he wo counries and ha hey were shaped by common forces. The findings lend some suppor o Williamson s argumen ha globalizaion forces are more likely o be observed in real wage han in real GDP per capia dispersion, bu hey also add o he growing skepicism surrounding he exisence of uncondiional real wage convergence. The nex secion considers he reasons why real wage growh was faser in Denmark han in New Zealand. 4. Explaining Real Wages Our model considers how he Denmark-New Zealand real wage raio was shaped by produciviy, and by disribuion forces. Boh he open economy forces of migraion, capial flows, and rade, and counry specific labour marke influences on income disribuion are invesigaed, wih he objecive of explaining he hree phases of he evolving real wage spread. In paricular, we consider why he real wage gap was largely eliminaed , why Denmark s real wages surged ahead , and why he Danish real wage lead was largely susained

14 4.1 The Model and he Resuls The following unresriced and resriced wage equaion is esimaed: lnw = α + ln( Y / Pop) + α lnun + α lnto + α Cif + α ln Emi + α ln Mig +ε s s DK s s DK NZ where W is consumer-price-deflaed hourly wages, Y/Pop is GDP per capia, Un is union membership per capia, To is erms-of-rade, Cif is he raio of capial inflow o GDP measured as minus he curren accoun on he balance of paymens, Emi is ou-migraion, Mig is in-migraion, and ε is a sochasic error-erm. The superscrips s, DK and NZ sand for he raio beween New Zealand and Denmark, Denmark, and New Zealand. The model is esimaed over , or in he esimaes where union membership is included as an explanaory variable. The model follows, o some exen, a sandard model of wage deerminaion where real wages depend on produciviy, unemploymen and various wage push facors, see for insance Layard and Nickell. 28 I includes roles for open economy forces influencing real wages via he effecs inernaional flows of labour and capial on facor endowmens, and by shifs in he erms of rade. Unemploymen is no included in our model because of he lack of daa. Labour produciviy is resriced o one in he preferred esimaes following convenions of he lieraure. We have imposed he resricion in he preferred esimaes because oherwise here would no be any seady sae soluion o he model. Labour s income share, and consequenly he rae of unemploymen, would oherwise increase o infiniy or decrease o minus infiniy as ime goes owards infiniy. Given he qualiy of he pre-wwii daa and ha some imporan variable are no available he esimaed coefficien of labour produciviy will generally no be one in he 28 Layard, R. and S. Nickell, Unemploymen in Briain. Economica vol. 53, 1986: S

15 esimaes. To shed some ligh on he sensiiviy of he esimaed coefficien of produciviy o inclusion of union membership and esimaion period he coinegraion esimaes of he model are presened in he able below. Table 2: Coinegraion esimaes of unresriced models REAL WAGE SPREAD s Y Ln Ln( Un ) DK ( ) s s DK NZ 2 Ln To Cif ln Emi ln Mig R DF Pop (6.08) (9.01) (2.47)(4.48) (3.47) (0.93) (8.94) (0.00)(4.92) (4.28) (0.40) (2.92) (9.09) (2.90) 3.63) (0.53) (2.47) (4.74) (1.79) (3.76)(2.62) (1.04) REAL WAGE DENMARK DK Y Ln Ln( Un ) DK ( ) DK DK DK 2 Ln To Cif Emi R DF Pop (2.07) (4.54)(2.78) (2.10) (4.54) Noes: DF = Dickey-Fuller es. Considering he esimaes in he firs row for he variables are coinegraed. The Danish union membership is included in he esimaes o accoun for he significance of he labour movemen during he wage explosions. The esimaed coefficien of produciviy is However, excluding union membership from he 15

16 esimaes increases he coefficien of produciviy o 1.46 (second row). Reducing he esimaion period o he coefficien of produciviy becomes 0.34 if union membership is included and 0.88 when union membership is excluded. Finally, he esimaes in he las row show ha he unresriced coefficien of labour produciviy is 0.44 for he Danish wage equaion. Overall, he esimaes show ha he esimaed coefficien of labour produciviy is highly sensiive o inclusion of variables and esimaion period. The mos saisfacory soluion o his problem is o follow heory and le he coefficien of produciviy be resriced o one. 29 Table 3. Resriced Parameer esimaes of real wages, New Zealand and Denmark. REAL WAGE SPREAD s Y Ln Ln( Un ) DK ( ) s s DK NZ 2 Ln To Cif ln Emi ln Mig R Pop DW (1.59) (4.07) (0.08) (0.11) (4.01) (3.06) (0.98) (3.56) (0.06) (1.18) (3.71) (3.20) REAL WAGE DENMARK DK Y Ln Ln( Un ) DK ( ) DK DK DK 2 Ln To Cif Emi R Pop DW (3.22) (1.58) (2.31) (0.54) (3.55) (2.06) Noes. The numbers in parenheses are absolue -saisics. The coefficiens of produciviy are resriced o one. A consan erm is included in he esimaes bu no shown. The Cochrane-Orcu procedure is used o correc for firs-order serial correlaion. Sources. See he Daa Appendix. 29 For New Zealand alone Greasley and Oxley, Refrigeraion and Disribuion, p. 37, repor an esimaed produciviy coefficien of for he years

17 The resuls in Table 3 show ha for he period comparaive produciviy, measured by real GDP per capia chiefly shaped he spread of Denmark-New Zealand real wages, bu ha disribuional forces were imporan during he years o 1913, and mos especially in he immediae afermah of World War I. Among he open economy forces only capial flows had a saisically significan effec on he real wage spread. In he resuls for he pos-1900 period, counry specific labour marke condiions, measured by he degree of unionizaion in Denmark, influenced real wages in Denmark and he real wage spread. Ideally he raio of union membership in New Zealand and Denmark and no jus he Danish union membership would be used in he esimaed model. However, union daa are no consisenly available for he NZ. There are wo reasons for mainaining he Danish union membership variable. Firs, he parial srike daa sugges ha union aciviy was much more significan in Denmark han New Zealand. The srike daa shows a subsanial larger srike frequency in Denmark han in New Zealand in he following key years as shown in Table 4. Table 4. Raio of srikes per employee in Denmark and New Zealand Noes. Srikes are measured as days los in srikes. For New Zealand he numbers of workers involved in srikes and lock-ous (as disinc from he number of soppages) are backdaed o 1918 and 1919 using he 1920 figure. Second, we included he Danish union daa because hey are imporan in explaining he Danish wage cach-up in The hisorical evidence, which we consider more 17

18 fully below, poins owards union aciviy as he driving force behind he Danish wage cach-up in hese years. I would, herefore, be erroneous o leave ou he addiional informaion conained in he Danish union daa from he regressions. The coinegraion esimaes above suppor he inclusion of he union daa. The variables are only coinegraed if he Danish union variable is included in he esimaes. 4.2 Real Wages and Produciviy In he longer erm real wages were chiefly shaped by GDP per capia, bu he real wage and produciviy spreads diverge wihin hese years. The chief longer erm influence diminishing relaive New Zealand produciviy is likely o have been is faser rise in he raio of labour o land. 30 The European populaion of New Zealand rose by around 500% , from around million, whereas occupied land rose by around one-hird over he same period. In conras Denmark s populaion rose less quickly by around 95% from million New Zealand s GDP per capia adjused for PPP was especially high in 1874 reflecing is naural resource abundance per capia, and was around 185% of Denmark s level, bu he raio fell o 120% in 1914 and 112% in The relaive fall in New Zealand s land abundance was no offse, a leas from he 1890s, by higher capial formaion (see Figure 5 below). Capial inflows relaive o GDP were remarkably similar in he wo counries While shifs in relaive resource endowmens shaped he longer erm movemen of New Zealand s and Denmark s GDP per capia, facor endowmens were no a singular influence on heir real wages. Mos 30 A full discussion of comparaive produciviy shis is ouside he scope of his paper. Condliffe, New Zealand in he Making, London, Allen and Unwin, 1930 provides informaion on occupied land and populaion in New Zealand. 31 These raios are projeced from Maddison s, Monioring 1938 esimaes of GDP per capia adjused for PPP. 18

19 especially i was he case ha Denmark s real wage advanage by 1939 exceeded ha of her GDP per capia. The disparae shifs in GDP per capia and real wages arose during Denmark s rapid real wage cach-up phase and mos srongly in he immediae afermah of World War One, and in hese years he real wage spread narrowed more quickly han he wo counries GDP per capia. Only wo-hirds of Denmark s real wage cach-up is explained by produciviy cach-up for he years o 1913, see Table 5 and Figure 4. Disribuional shifs hus favoured real wages in Denmark before World War I. Table 5. Simulaion resuls ( Dependen variable: real wage spread) Time period Acual change Prediced Produciviy Capial Flows % Noe. The simulaions are based on he esimaes in he second row in Table 3. Logs Figure 4. Real Wage Spread, NZ/DK Acual Vs. Prediced Acual Prediced Noe. The simulaions are based on he esimaes in he second row in Table 3. 19

20 Even more srikingly, when Denmark s real wages surged ahead in he years o 1921, New Zealand s GDP per capia growh was he faser. Disribuional shifs became even more powerful in he rans-war period , and principally favoured Danish real wages. However, over he longer period around wo-hirds of he shifs in real wage spread can be explained by produciviy shifs, which resuled from faser Danish GDP per capia growh in 1920s. In conras, in he 1930s New Zealand s GDP per capia growh was he faser, and Denmark s real wage lead narrowed a lile in ha decade. Generally, beween he world wars Denmark mainained is real wage lead ha was esablished in he afermah of World War I, and disribuional shifs were less imporan afer Disribuion Variables o 1913 According o he resuls in Table 3 inernaional capial flows, raher han labour migraion or he erms of rade, was he dominan open economy influence on he real wage spread. Boh Denmark and New Zealand were major recipiens of inernaional capial, and he exen of he inflows can be gauged from heir balance of paymens posiions, which are shown in Figure Capial inflows relaive o GDP were higher in New Zealand in earlier years, and hey averaged 10.9% , compared o 1.2% in Denmark If here is no change in gold socks i follows ha ne capial inflows are equal o he curren accoun defici of he balance of paymens. W. S. Woyinsky and E. S. Woyinsky, World Commerce and Governmens, New York: The Twenieh Cenury Fund, 1955 repor occasional direc esimaes of capial flows, bu hey provide only four and five daa poins for New Zealand and Denmark in our period. Using hese daa we obain he following OLS regression resuls, where he sandard errors are based on Whie s heeroscedasiciy consisen covariance marix: BoP = 1.07 NCI (Denmark) BoP = 1.17 NCI (New Zealand) (7.59) (8.38) where BoP is he curren accoun surplus and NCI is ne capial inflow including errors and omissions. The numbers in parenheses are absolue -saisics. The resuls show a very high correlaion beween he variables, and ha he coefficien esimaes are very close o one. Alhough we only have a few daa poins available he evidence suppors ha we use he balance of paymens as a proxy for capial flow. 33 Denmark s curren accoun daa are from Johansen Dansk hisorisk saisik. Capial inflows o New Zealand are esimaed following he procedures of W. Rosenberg, Capial impors and growh- he case of New Zealand, Economic Journal vol. 71: 281, Merchandise rade, including specie, daa are from New Zealand Official Yearbooks. Mos capial flows o New Zealand ook he form of governmen deb, alhough much of he capial was subsequenly passed o selers, for example via Sae Advances for morgages. Boh he level of governmen deb (cenral and local), and he ineres charges are repored in he 20

21 However, capial inflows became less imporan for New Zealand in he years o Her rade balance was generally in surplus from 1887 and capial inflows, necessiaed in par by deb ineres charges averaged 1.99% of GDP In conras, he raio of Danish capial inflows o GDP more han doubled o average 2.59% Srandskov and Pedersen discuss he composiion of he capial inflow o Denmark, mos especially ino he food processing secor, and highligh is ake-off around Around 13.4% of he decline in he real wage raio see Table 5, can be explained by he increasing relaive endency for inernaional capial o flow o Denmark in he years o % Figure 5. Capial Inflow relaive o GDP NZ DK Sources. Rosenberg, Capial Impors, New Zealand Official Yearbooks, Johansen Dansk hisorisk saisik. Yearbooks. These daa allow he annual capial inflow daa repored by Rosenberg o be exended for years beyond Complee daa on privae deb and privae deb ineres are no available, and use is made of Rosenberg esimaes of he raio of privae o public deb ineres o gauge New Zealand overall deb ineres paymens. The raio of privae o public deb ineres ranges from 19% for years o 1886 o 21% Srandskov, J. and K. Pedersen, Pioneering FDI ino he Danish Bacon Indusry, Scandinavian Economic Hisory Review vol. LXVIII, 2000: 1,

22 The unimporance of labour migraion for relaive real wages may be explained by he endency for in-migraion o become less imporan for New Zealand in he years o The New Zealand immigraion rae averaged 47.5 per , bu was negaive Thereafer New Zealand immigraion rae rose o average 6.5 per , which is less han 14% of he rae. 35 Denmark s ou-migraion rae was he lowes in Scandinavia, and averaged 3.67 per The saisical resuls in Table 3 show he effecs of Danish emigraion and New Zealand immigraion on heir relaive real wages was near zero. A variey of indicaors have been deployed o ry and gauge how rade influenced relaive real wages, including ariffs, impor-gdp raios, and he erms of rade. Denmark did no proec is grain farmers in he years o 1913, and gained from lower food prices, bu he ne effec on real wages is unclear. Moreover, New Zealand also gained from lower ranspor and manufacured good prices. Like Denmark, New Zealand did impose ariffs on manufacures, including on impors from Briain. Denmark s ariff on manufacures ranged from 14-20% in he years By 1913 New Zealand imposed a 20% ariff on many consumer goods, including on exiles, furniure, and bicycles, alhough machinery araced only a 5% ariff. New Zealand also axed he impor of a variey of food producs, including grains, and is ariff revenue per capia rose from 2.52 in 1893 o 3.21 in 1913 (New Zealand Official Yearbook 1914). Overall, he openness of he Danish and New Zealand economy o rade may have been broadly similar in The resuls in Table 3 show ha relaive movemens in Denmark s and in New Zealand s erms of rade did no saisically significanly influence heir real wage spread, 35 Bloomfield, G. T., New Zealand Handbook of Hisorical Saisics. Boson: G. K. Hall and Company, O Rourke and Williamson, Around he European Periphery. 37 O Rourke and Williamson Around he European Periphery. 22

23 and advanage one counry over he oher. The likely implicaion is ha boh counries gained in similar fashion from improvemens in heir erms of rade, a leas o Neiher did including heir respecive ariff raes or heir openness o impors produce saisically significan coefficiens (resuls no shown). I was he inernaional flow of capial, raher han labour migraion or he erms of rade ha reinforced produciviy cach-up, o conribue o Denmark s real wage cach-up wih New Zealand in he years o Disribuion Variables afer 1913 Open economy forces played a less imporan role in shaping he Denmark-New Zealand real wage raio afer Denmark hough did experience a dramaic rise in is wagerenal raio o 1919, bu his chiefly refleced he real wage bargaining success of Danish unions. Denmark s real wages surged ahead of New Zealand s, and he lead was largely susained hrough he years beween he world wars. The Danish real wages increased by 13% in 1918, 52% in 1919, and 9% in Much of Denmark s hisoriography poins o worker miliancy having powerful real wages effecs in he afermah of World War I. 39 Danish union membership increased from 152,000 in 1913 o 321,000 in The Danish Syndicalism movemen was founded in 1910 and fuelled by he Russian Ocober Revoluion in 1917 gained momenum over he period 1918 o In erms of members he movemen was insignifican, bu Hansen and Henriksen show i was highly effecive in influencing wages and creaing labour unres. Their sraegy was o form groups wihin he union movemens o ry and influence he direcion aken by he union. The increase in he number of srikes and lockous over he period winess an explosive 38 However Greasley and Oxley, Refrigeraion and Disrubuion, p. 39, show ha for New Zealand ha shifs in he erms of rade chiefly affeced he price of land relaive o labour. 39 Dalhoff, J.,Krigsaarernes Lønpoliik, Naionaløkonomisk Tidsskrif vol. LIX, 1921:1-36. Hansen, S. A. and I. Henriksen, Dansk Social Hisorie, København: Gyldeldal, Visser, J., In Search of Inclusive Unionism. Devener & Boson: Kluwer Law & Taxaion,

24 developmen in labour marke conflics. The days los from srikes rose six-fold from 1916 (which is he firs year srikes are recorded) o 1921 (Saisisk Årbog). The srengh of he labour movemen was parly due o shorage of labour in he pos war period. 41 The rae of unemploymen decreased from 18.1% in 1918 o 10.9% in 1919 and 6.1% in The reducion in unemploymen alone would have pushed up he real wage growh by approximaely 7% in 1919 and 5% in 1920 using he parameer esimaes of he Phillips curve during he inerwar period by Madsen. 43 Anoher facor ha conribued o he wage explosion was he reducion in working hours from 9½ hours a day o 8 hours a day over he period from 1919 o 1920, which was accompanied by an increase in unemploymen benefis. 44 The working hours reducion alone resuled in a 16% increase in real hourly wages. Togeher reducions in working hours and unemploymen resuled in a real wage increase of 28% or abou half of he wage explosion. Tha real wages rose even furher migh be conneced o workers dissaisfacion wih he earlier real wage reducion of 21% during , and o he leverage provided o union bargainers by higher unemploymen benefis when here was an absence of srong poliical opposiion o he wage increases in he wake of he Russian Revoluion and Denmark s wildca srikes of 1919 and Fully comparable labour miliancy daa are no available for New Zealand, bu ha for he numbers of workers involved in srikes does no show a dramaic increase around he end of World War I. Around 4000 New Zealand workers were involved in srikes in 1918 and 1919, compared o around 6000 and 13,000 in 1912 and 1913 respecively. In 41 Dalhof, Krigsaarernes Lønpoliik. 42 Dansk hisorisk saisik. 43 Madsen, J. B., The Lengh and he Deph of he Grea Depression: An Inernaional Comparison, Research in Economic Hisory vol. 22, 2004:2, I is no clear from Dalhoff, Krigsaarernes Lønpoliik wheher i was he replacemen rae ha increased or he real benefis rae. 45 Bang, Fagopposiionens Sammensluning De Danske Syndikaliser Aarhus: Modryk,

25 conras 36,000 Danish workers were involved in srikes in 1919, he firs year such daa are available. I seems clear ha labour miliancy was a more powerful force in Denmark han in New Zealand in he afermah of World War I. The coefficiens for he Danish union variable are saisically significan in boh he real wage spread and Denmark s own real wage equaion. The inclusion of he union membership variable grealy improves he predicive power of he real wage spread equaion for he years around World War I - see Figure 6. In paricular, for 1919, he year Danish real wages surge ahead of hose in New Zealand, around one-half of he faser growh of Denmark s real wages is accouned for by he rise in union membership. Capial inflows, reflecing a sharp jump in Denmark s curren accoun defici amouning o 29% of GDP when New Zealand experienced a capial ouflow, explains around 46% of Denmark s faser real wage growh in Capial inflows o Denmark were also subsanial in In conras capial inflows o New Zealand were neural over he years , reflecing a large capial ouflow in 1919 as pen-up goods were released for expors, bu a commensurae inflow in 1920 when impors surged. The dramaic collapse in he real wage spread around he end of World War I chiefly arose from disribuional shifs on he Danish side. 46 A boom-induced surge in impors was primarily responsible for he curren accoun deficis, which climbed o 28.6% in 1919 and remained a 17.6% in The increase in he economic aciviy was significan: GDP increased by 19.9% in 1919 and by 4.7% in 1920 and he rae of unemploymen decreased from 18.1% in 1918 o 10.9% in 1919 and 6.1% in 1920, and indusrial producion increased by 40.6% in 1919 and by 18.4% in The economic upswing generaed a srong increase in impors of manufacures and capial goods, paricularly, which increased from 5.6% of GDP in 1918 o 17.0% in 1919 and remained a 14.9% in

26 Figure 6. Real Wage Spread, NZ over DK (logs) Acual Prediced Noe. The simulaions are based on he esimaes in he fourh row in Table 2. 5 Concluding remarks Denmark achieved specacular real wage growh in he half cenury afer 1870 and according o Williamson had he highes levels in he world in Her real wage cachup wih he wesern-offshoos was he mos dramaic. In 1870 urban labour s real wage had averaged around hree imes higher in Ausralia and he USA han in Denmark. Adding New Zealand o he comparisons reinforces ha here was a remarkable Danish advance. New Zealand s real wages was wice as high in he 1870s, bu averaged around 75% of Denmark s level beween he world wars. Why his happened depended upon relaively fas produciviy advance in Denmark, bu also on shifs in income disribuion ha paricularly favoured Danish workers. Over he years around wo-hirds of faser Danish real wage growh can be accouned for by her higher real GDP per capia growh. Of he conribuion made by disribuional shifs, he open economy forces leading o faser real wage growh in Denmark were mos subsanial in he years o In hese years he wage-renal raio 26

27 declined in New Zealand, bu rose in Denmark. Accordingly, he endency for faser produciviy growh o lead o high Danish real wages was reinforced, principally by inernaional capial flows. In conras, we found no evidence ha he modes in-migraion of labour o New Zealand or he modes ouflows from Denmark affeced heir real wage dispersion. Neiher do our resuls show ha shifs in he erms of rade favoured real wages in one counry over he oher. The mos dramaic shifs in income disribuion from propery owners o wage earners occurred around he end of World War I, and principally favoured real wages in Denmark. The shifs were no principally driven by open economy forces, alhough Denmark experienced subsanial capial inflows in More imporan was he rise in labour s bargaining power associaed wih rades union miliancy, and his led o a real wage surge in Denmark which peaked around The disribuional shifs were more powerful around World I han in he open world economy era o The effec was o rache Danish real wages beyond New Zealand levels afer World War I, and Denmark s real lead was susained beween he world wars. Generally, he comparisons of Denmark and New Zealand emper he conribuion of facor price equalizaion endencies and highligh he role of produciviy in explaining heir real wage dispersion. Denmark s real wages caugh-up and surged ahead hose of is new world compeior principally because of faser produciviy growh. This perspecive does no necessarily diminish he conribuion of inernaional forces, relaed for example o he spread of dairying or refrigeraion echnology, possibly simulaed by inernaional capial and rade flows, which may have raised Denmark s real GDP per capia, and hus her real wages. Neverheless, he conribuion of relaive shifs in he price of labour and land in he Old World and he New, arising from rade and facor flows, was, on he basis 27

28 of a comparison of Denmark and New Zealand a comparaively modes influence on real wages, alhough he effecs of capial flows were no inconsequenial before Daa Appendix A. Sources. Denmark Nominal GDP. Hansen, S. A., Økonomisk Væks I Danmark, København: Akademisk Forlag, Real GDP. Maddison, A., Monioring he World economy Paris: OECD, Impors. Johansen Dansk hisorisk saisik. Impor duies, Michell, B. R., European Hisorical Saisics London: Macmillan, Populaion. Johansen Dansk hisorisk saisik. Union Membership. Visser In Search. Wages. Johansen Dansk hisorisk saisik. Emigraion. Johansen Dansk hisorisk saisik. Capial inflow. Johansen Dansk hisorisk saisik. Terms of Trade. Johansen Dansk hisorisk saisik. Consumer prices. Pedersen Arbejdslønnen i Danmark, Johansen Dansk hisorisk saisik. Land Prices. Chrisensen Lønudviklingen inden, Saisisk Årbog. New Zealand Real and Nominal GDP. Greasley and Oxley Measuring New Zealand s GDP; Regime shif and fas recovery. Consumer Prices. Nesbi-Savage, A long run consumer price index. 28