THE MASSIVE INVESTMENT, IMMIGRATION AND PRODUCTIVITY BOOMS IN CANADA: WHICH CAME FIRST? STUART J. WILSON

Transcription

1 UNIVERSITY OF REGINA ISSN DEPARTMENT OF ECONOMICS THE MASSIVE INVESTMENT, IMMIGRATION AND PRODUCTIVITY BOOMS IN CANADA: WHICH CAME FIRST? STUART J. WILSON Deparmen of Economics Universiy of Regina Regina, Saskachewan Canada S4S 0A2 May 27, 2003 DISCUSSION PAPER #103

2 The Massive Invesmen, Immigraion and Produciviy Booms in Canada: Which Came Firs? Suar J. Wilson * Universiy of Regina May 27, 2003 Absrac This aricle examines he links beween capial accumulaion, produciviy and immigraion o deermine which facor was he leading cause of he boom condiions in Canada over 1870 o The resuls of an economeric analysis using coinegraion echniques and causaliy ess sugges ha capial accumulaion and produciviy growh preceded he massive immigraion boom in Canada during he years before he Grea Depression. Domesic sources of invesmen played a leadership role in he accumulaion of capial during he invesmen booms, and increased he incenives for labor o migrae o Canada. Foreign capial hen chased immigrans o saisfy he addiional demand for capial and served o prolong he invesmen booms. Suggesed Running Head: Invesmen, Immigraion and Produciviy Booms in Canada. Keywords: Capial accumulaion, migraion, coinegraion, causaliy esing. JEL Classificaion: F20, N10, C32. * Deparmen of Economics, Universiy of Regina, Regina SK S4S 0A2. Tel: suar.wilson@uregina.ca. The auhor graefully acknowledges research funding in he form of a Universiy of Regina Special Projec Gran. 2

3 1. Inroducion The period from 1850 o 1914 has been characerized as one of mass migraion, wih 55 million Europeans moving o he New World counries in he Americas and Ausralia (Haon and Williamson, 1998: 3). A he same ime, massive amouns of capial were sen overseas o he New World. During a period wih no significan war aciviy and falling ransporaion coss, capial and labor were free o move. The chief reason for he migraion of boh facors of producion was o exploi he cheap resources available in he New World. The saemen ha capial chased afer he European migrans, (Taylor and Williamson, 1994: 349) implies ha labor moved firs and hen capial followed. In discussing capial accumulaion, populaion movemens have been generally regarded as exogenous, and capial as endogenous, reacing o labor flows (for example, Green and Sparks 1999; Wilson 2003). This view makes sense when we characerize he invesmen booms of he New World as being driven by populaion-sensiive invesmen wih a focus on social overhead requiremens. However, he dual scarciy problem of he New World naions firs led o he increased demand for boh capial and labor o exploi cheap resources. Once inernaional labor flows reaced o he incenives found in he New World, he capial requiremens relaed o infrasrucure and overhead inensified furher (Williamson, 1979). Green and Urquhar (1976) separaed he demand for physical capial ino hree essenial componens: (i) capial invesmen as a resul of increases in he labor force, o equip new workers for producion; (ii) populaion-sensiive invesmen o mee social overhead needs and; (iii) capial invesmen as a resul of labor-augmening echnological 2

4 change. In (i) and (ii), capial flows are a response o labor flows ha increase he reurns o capial, bu in (iii), echnological change increases he reurns o boh capial and labor in use, and hus increases he incenives for boh capial and labor o migrae and accumulae. Abramoviz (1993: 217) poined ou ha in he nineeenh cenury, massive capial accumulaion was a consequence of scale-dependen and capial-using echnological progress. Technological advance in he 19 h cenury was characerized by a bias owards physical capial and was a large source of labor produciviy. In ligh of his assessmen, much capial accumulaion in he lae 19 h cenury also fis ino caegory (iii). In all hree cases, capial accumulaes due o increased reurns. Labor also may migrae and accumulae as a resul of changes in he reurn o labor. Labor-augmening echnological change and an increase in exploiable naural resources increase real wages. In addiion, an increase in physical capial used by workers in producion also raises labor produciviy and wage raes. There is a large lieraure ha deals wih real wage differences as a moivaor for labor migraion, and real wage convergence beween wo counries as a resul of migraion. Of his lieraure, Haon and Williamson (1994) showed ha real wage differences beween sending and receiving counries were imporan deerminans of emigraion raes in eleven European counries in he lae nineeenh cenury. O Rourke, Williamson and Haon (1994) showed ha migraion was an imporan facor in reducing wage differences beween he UK and he US. Canada experienced several economic, invesmen, and immigraion booms in he years from 1870 o The purpose of his aricle is o examine he relaionships beween invesmen, immigraion and produciviy in he six decades before he Grea 3

5 Depression. This aricle presens an economeric model o empirically deermine wheher immigran labor flows, capial invesmen or produciviy improvemens occurred firs in Canada, and wheher foreign capial chased migrans o Canada or vice-versa. 2. The Canadian Experience, In he decades from Confederaion leading up o he Grea Depression, Canada experienced remendous developmen. In 1870, he populaion of Canada amouned o 3.6 million. By 1929, he populaion had increased o 10 million. Over his period, per capia real incomes more han ripled, from $1,353, o $4,144 in 1981 dollars. The raio of gross domesic capial formaion o gross naional produc increased from 16% of GNP o 22% of GNP. The domesic savings rae increased from 9% in 1870 o 17% in Figure 1 shows he ime paern of real gross naional produc on a per capia basis. The average growh rae over was 1.92%. Growh proceeded slowly in he hree decades leading up o he wenieh cenury, averaging 1.04% from Canadian developmen was dramaically alered around Real income per capia growh increased o an average of 2.61% from , he period more commonly described as he Whea Boom Era in Canada. Afer Confederaion in 1867, Canada experienced high raes of emigraion, as many Canadians moved o he Unied Saes in he pursui of greaer economic poenial. However, Canada was sill a naion wih high gross immigraion. Figure 2 shows he paern of immigran arrivals as a proporion of he oal populaion. Gross immigraion levels averaged 1.2% annually over he period, hen increased o an average of 1.9% over he period. 4

6 The paerns of invesmen, domesic savings and foreign capial inflows are depiced in Figure 3. The rae of invesmen averaged 15% over he period, hen rose above 20% during he Whea Boom Era. The figure shows he sriking increase in he rae of capial accumulaion afer The rae of invesmen rose from 12% in 1896, o a peak of 34% in 1912, hen seling o average 18% during he 1920s. The rae of domesic savings rose from an average of 8% before 1897, o 16% in 1912, hen averaged 17% in he 1920s. Foreign capial inflows averaged 7% over , hen grew o a peak of 18% in Canada experienced a massive invesmen boom in he firs decade of he wenieh cenury, fuelled by boh domesic and foreign sources of capial. The paerns of migraion and capial accumulaion coincide very closely. Canada experienced mild invesmen, immigraion, and produciviy booms in he early 1870s when he Inercolonial Railway was buil linking he Canadian Mariime Provinces o Cenral Canada, and in he mid-1880s wih he compleion of he Canadian Pacific Railway linking Cenral Canada o he wes all he way o he Pacific Ocean. Afer a recession in he mid-1890s, he Canadian economy once again experienced srong growh in invesmen, immigraion and produciviy as Canada enered he Whea Boom Era. The surge in populaion growh during he firs decade of he wenieh cenury has been aribued primarily o he emergence of he whea economy in Canada. The wesern provinces were rapidly seled during his period. The increase in wesern selemen has been aribued o a variey of facors, including he reducion of ransporaion coss due o rail expansion, he inroducion of dry-farming echniques and inensive recruimen effors of he Canadian Liberal Governmen (Norrie, 5

7 1975; Marr and Percy, 1978; and Borins, 1982). Lewis (1981) showed ha he feasible region of culivaion expanded afer 1896 due o he consrucion of rail lines and an increase in whea yields. Farmers were also able o culivae larger acreages wih new and improved equipmen (Ward, 1994). Selers migraed o he Canadian Prairies as a resul of improved economic incenives and beer prospecs for successful farming afer The period from 1897 o 1913 was one of exensive growh in various secors of he Canadian economy, and also marked he developmen of advanced manufacuring, for example, a modern iron and seel indusry (Green and Urquhar, 1994). An invesmen boom was fel hroughou he whole economy. The railways, consrucion, manufacuring, public secors and oher business secors of he economy all conribued o he massive accumulaion of physical capial. Green and Green (1993) also showed ha he immigraion boom before WWI was experienced across all secors and provinces in Canada. Immigraion and he inflow of foreign capial boh collapsed during WWI. Canada even became a ne exporer of capial a he end of he war. Immigraion, invesmen and oupu growh all recovered during he 1920s. Immigraion and foreign capial inflows boh increased in he firs half of he 1920s, bu real incomes did no significanly improve unil he lae 1920s. Economeric analysis has suggesed ha he increases in invesmen in Canada were he resul of increases in he populaion, and innovaions o expors (Green and Sparks, 1999). The rise in he domesic savings rae during his period was due in par o he dramaic inflow of working age migrans and produciviy improvemens (McLean, 1994; Wilson, 2000). However, Canada, wih a young populaion, sill required foreign 6

8 capial inflows o mee he increased capial requiremens during his period (Taylor and Williamson, 1994; Wilson, 2003). These sudies have reaed populaion movemens, and effecively, migraion, as exogenous. However, Wilson (2003) suppors he argumen ha he surge in domesic savings in Canada over canno be aribued o immigraion, bu o produciviy improvemens before he surge in immigraion. The mos likely explanaion is ha he Canadian economy pulled ou of a recession in 1897, and labor produciviy and he employmen rae boh increased a he sar of he Whea Boom Era. Unforunaely, labor and employmen saisics are only available a census daes and his possible explanaion canno be easily verified. I is also commonly believed ha capial is more mobile han labor, which would sugges ha any improvemens in echnology or resource discovery would lead o capial accumulaion and labor force growh, bu wih invesmen preceding immigraion. World capial markes have been viewed as inegraed in he years leading up o WWI, perhaps even beer inegraed han he capial markes of he 1980s (Zevin, 1992). Did capial accumulaion precede immigraion and produciviy improvemens in Canada over he period? Did capial chase migrans o Canada? Did capial and labor accumulae as a resul of produciviy improvemens ha increased reurns o boh facors of producion? The following secion provides a heoreical background for hinking abou hese quesions. 7

9 3. Theoreical Consideraions The Solow growh model and is exensions provide a rich modeling environmen in which o examine he relaionships beween labor force growh, capial accumulaion, produciviy improvemens, and he incenives for migraion and capial flows across borders Producion Consider a sandard neoclassical consan reurns o scale producion funcion incorporaing a hird facor, resources: α β 1 α β Y = A K R ( H N ). (1) In his conex, Y denoes oupu, K denoes he sock of physical capial, N denoes he populaion size, H denoes human capial per capia (a weighed average of human capial in he aggregae), R denoes a hird facor of producion, resources, and A is a scale parameer represening he effec of oal facor produciviy improvemens due o echnology. This producion funcion can be re-wrien, wih L denoing effecive labor, and equal o he populaion size imes human capial per capia: L = H N ; (2) α β 1 α β ( K, R, L ) = A K R L Y = F. (3) In his descripion of producion, effecive labor may increase wih improvemens in he skills of he exising labor force (an increase in H, all else equal), wih an increase in he 8

10 employmen rae (an increase in H, all else equal), and wih an increase in he working populaion (hrough immigraion, an increase in N, all else equal). Given his producion funcion, he reurns o he inpus o producion are as follows, reurns per uni of physical capial (he real ineres rae, r, plus depreciaion, δ), reurns o he owners of resources on a per uni basis (η), and wages per uni of effecive labor, L (wage rae, w): α 1 β 1 α β 1 ( K, R, L ) = αa K R L r + δ = F ; (4) η = F ; (5) α β 1 1 α β 2 ( K, R, L ) = βa K R L α β α β w = F3 ( K, R, L ) = (1 α β ) A K R L. (6) I is assumed ha resources are no mobile (as examples, land, or resources ha need o be exraced locally). Wih perfecly compeiive markes and mobile capial and labor, capial and labor will move o where hey are mos producive. 3.2 Capial Accumulaion Under he assumpion of perfec capial mobiliy, and ha capial can be used in producion in wo differen economies, Economy A and Economy B, capial will flow where he reurns are highes. A a cerain poin in ime, assume ha he reurns o capial are equal in he wo economies, so ha here is no incenive for he migraion of capial: A α 1 β 1 α β A RA LA αa K δ A = r r A B 1 α β B B B = = α 1 β αa K R L δ. (7) Capial will hen have an incenive o migrae from Economy B o Economy A if he reurn o capial increases in Economy A under he following circumsances: (i) here is an increase in oal facor produciviy in Economy A (A A ) relaive o ha of Economy B B B 9

11 (A B ); (ii) here is an increase in he resource endowmen of Economy A (R A ) relaive o ha of Economy B (R B ); and (iii) here is an increase in effecive labor in Economy A (L A ) relaive o ha of Economy B (L B ). If any of he above, or any combinaion of he above occurs, capial will flow from Economy B ino Economy A, reducing he high reurns o capial in Economy A, and increasing he low reurns o capial in Economy B, and capial will migrae unil reurns are once again equal across he wo economies. 3.3 Labor Growh Similarly, under he assumpion of perfec labor mobiliy, labor will flow where real wages are highes. Assuming ha a some poin in ime, real wages are equal across he wo economies, hen here is no incenive for labor migraion: (1 α β ) A K A α A R β A L α β A = w A α β α β = w ( 1 α β )A K R L (8) B = B B B B Labor will migrae from Economy B o Economy A if (i) here is an increase in oal facor produciviy in Economy A relaive o ha of Economy B; (ii) here is an increase in he resource endowmen of Economy A relaive o ha of Economy B; and (iii) here is an increase in physical capial in Economy A relaive o ha of Economy B. If any of he above, or any combinaion of he above occurs, real wages in Economy A will be higher han hose of Economy B, and labor will migrae unil real wages are once again equaed across economies. 3.4 Per Capia Income Growh The producion funcion may be re-wrien as follows on a per capia basis: 10

12 Y N = K A N α R N β 1 α β ( H ). (9) I follows ha an increase in per capia income is a resul of: (i) an increase in he parameer A, denoing oal facor produciviy improvemens; (ii) an increase in per capia physical capial, as a resul of increased invesmen per capia; (iii) an increase in per capia resources available for producion; (iv) an increase in he average amoun of human capial per person, which may be caused by an increase in he educaion and skills of workers, and/or an increase in he proporion of employed laborers in he populaion (which may be he resul of massive migraion of workers). For muli-secor economies wih secors where resources are no a facor in producion, scale facors associaed wih an inflow of labor and capial ino some secors may increase he demand for services from oher secors of he economy, and in urn, increase he reurns o capial and labor in hese oher secors of he economy. As an example, service professionals like docors, or radesmen like consrucion workers, may experience an increased demand for heir services wih an increase in he populaion. Their wages, as well as he reurns o capial used by hese individuals will increase. Foreign professionals, radesmen, and capial will again have an incenive o migrae as a resul of a populaion boom. The discovery of new producive resources, including he opening up of land for agriculure, in he New World made i more aracive for boh labor and capial o migrae. A produciviy boom in he New World, as a resul of echnological improvemens would also encourage capial and labor o migrae. Addiionally, a massive inflow of capial o a counry, all else equal, would increase he reurns o labor in ha counry, and hereby increase he incenives for labor o immigrae. On he oher hand, a 11

13 massive inflow of labor ino an economy would increase he reurns o capial in he receiving counry, and increase he incenives for capial o migrae. Which facor moved firs? While i is generally assumed ha capial is much more mobile han labor, he lieraure sill ofen reas populaion and labor movemens as exogenous, and reas invesmen as responding o populaion and labor movemens. In he case of Canada, he massive immigraion and invesmen boom a he dawn of he wenieh cenury has been aribued o he opening up of he wes for selemen. There have been oher immigraion and invesmen booms associaed wih periods of srong growh in he early 1870s, mid-1880s, and afer WWI. The following secion presens an economeric model o examine which facor, capial or labor, was he firs-mover. This sudy will also examine he role of improved produciviy in providing incenives for invesmen and labor force growh. 4. Economeric Analysis In he economeric sudies of Taylor and Williamson (1994) and of McLean (1994), populaion changes and income growh were assumed o be exogenous facors in explaining he increase in domesic savings in Ausralia and Canada. Green and Sparks (1999) examined he massive populaion and invesmen boom in Canada a he beginning of he wenieh cenury using a vecor auoregressive (VAR) model, where populaion movemens were deermined a priori o be exogenous in relaion o he oher variables in he analysis, including invesmen, which was considered endogenous wih respec o populaion. Wilson (2000) examined he Canadian economy since 1870, using a VAR model wih coinegraed processes. In his case, populaion movemens had o be 12

14 reaed as exogenous, and income was found o be weakly exogenous, indicaing ha changes in income were no influenced by changes in domesic savings over he sample period. 2 The relaionship beween labor migraion and capial accumulaion was also examined in Wilson (2003) using a dynamic general equilibrium overlapping generaions lifecycle model. The model reaed migraion as exogenous, wih capial accumulaion responding o changes in he labor force. Changes in populaion were deemed exogenous in all of he sudies cied above. Populaion ends o change raher slowly and is less volaile relaive o mos economic variables. Under normal circumsances, changes in feriliy and moraliy occur over long periods of ime. However, an imporan componen of populaion growh is migraion, and migraion paerns can change raher abruply. During he period of sudy, migraion flows changed quickly wih respec o he populaion size, and i was migraion movemens ha dramaically shocked he Canadian economy, as opposed o he oher facors influencing populaion growh. In his sudy, migraion will be reaed as endogenously deermined a priori, along wih he oher variables of ineres: income, invesmen, domesic savings, and foreign capial inflows. Causaliy esing will also be conduced o esablish he naure of hese relaionships beween hese variables, and o deermine which variable(s) moved firs. 4.1 Daa Canadian income, invesmen, domesic savings, foreign capial inflows, gross immigraion, and populaion daa are all available annually over he period of sudy. Unforunaely, labor force, capaciy uilizaion, and human capial daa are no available, nor are daa relaing o echnological improvemens or naural resources. For 13

15 he purpose of his sudy, he series of ineres will be convered o per capia erms, in order o remove he effecs of populaion size on he variables. Given he discussion in Secion 3, in per capia erms, income, invesmen, and immigraion may be described in he following funcional forms: Y I IM = f (,, X ) ; (10) N N N I IM = f (, X ) ; (11) N N IM I = f (, X ). (12) N N In his case, X represens oher facors including echnological developmens, represened by he variable A, average human capial in he aggregae, H, available resources for use in producion, R, on a per capia basis, and he amoun of physical capial sock used in producion, K, on a per capia basis. Income will be influenced by invesmen ha increases he amoun of physical capial sock used in producion, by immigraion ha affecs he amoun of workers in he economy, and by he oher facors denoed by X. Invesmen will be influenced by immigraion ha affecs he amoun of workers in he economy, which in urn affecs he reurns o physical capial used in producion; invesmen is also influenced by changes in he variables described by X ha affec he reurns o physical capial. Immigraion is a funcion of invesmen and he variables described by X ha affec he reurns o labor. As saed earlier, daa on he variables conained in X are no available. In he empirical sudies referenced above, here is ofen feedback from measures of income o invesmen and immigraion (e.g. income growh has been found o affec capial accumulaion). Any change in echnology, in average human capial (including any 14

16 change in labor force paricipaion), and in he amoun of usable resources will ransfer o a change in produciviy and income, and so, changes in income per capia may also serve as a proxy for changes in produciviy, and influence invesmen and immigraion. So, in Eqs (11) and (12), per capia income may serve as a proxy for he oher facors described by X, facors ha influence he produciviy of he inpus o producion, and i is now clear ha we have a sysem of equaions where i is no possible o deermine a priori which of he variables are exogenous. The variables of ineres for his invesigaion are: (1) y, which will be equal o he naural logarihm of GNP divided by he GNP deflaor and he populaion size (he log of real income per capia); (2) i, equal o he naural logarihm of gross domesic capial formaion divided by he GNP deflaor and he populaion size (he log of real invesmen per capia); (3) s, equal o he naural logarihm of gross domesic capial formaion less foreign capial inflows (he log of real domesic savings per capia); (4) real ne long-erm foreign capial inflows per capia, f, in is naural logarihmic form; and (5) im, equal o he naural logarihm of he raio of immigran arrivals o he oal populaion. 3 Invesmen is spli ino he wo componens, domesic savings and foreign capial inflows, o uncover any differences in he movemens of hese variables over ime. The ne long-erm foreign capial inflow series is used insead of he ne foreign capial inflow series (including ne shor-erm flows) due o he complicaed ime-series properies of he ne foreign capial inflow series. 4 The ne long-erm foreign capial inflow series is also a beer indicaor of foreign invesmen in physical capial for use in producion. The gross immigraion series is used insead of he ne immigraion series due o similar problems. The ne immigraion series is ofen negaive from 1870 o 1899, and canno be 15

17 convered o a logarihmic form o be used in he mehodology described below. The ne immigraion series does however exhibi swings very similar o hose of he gross immigraion series, wih excepions during 1880s and from when ne immigraion remained seady while gross immigraion rose. 4.2 Mehodology Many economic ime-series, including income and invesmen, are non-saionary and exhibi a paern of growh over ime. Economeric invesigaions require ha he imeseries of ineres are saionary before proceeding wih regression analysis. Nonsaionary series are ransformed ino saionary series, eiher by removing a linear ime rend, or by aking firs differences of he logarihmic forms of he original series o examine growh raes. In many cases, series are sill non-saionary when a linear ime rend is removed, and are only saionary in he form of firs-differences. 5 As an example, income and invesmen are normally non-saionary series, bu he growh raes of income and invesmen are normally saionary series. The long-run relaionships beween nonsaionary variables canno be examined in radiional linear regressions, in he fear of misaking a spurious regression beween wo rending variables for a sable long-run relaionship. 6 As an example, he long-run income elasiciy of invesmen canno be esablished by a radiional linear regression of invesmen on income. However, he long-run relaionships beween non-saionary variables may be explored wih new economeric echniques, namely coinegraion analysis. Granger and Engle inroduced coinegraion analysis in he 1980s which has been viewed as one of he mos imporan conribuions o ime series economerics. 7 If a long- 16

18 run relaionship exiss beween non-saionary variables, hen deviaions from his longrun relaionship are saionary, and hese variables are said o be coinegraed. There are a few mehods used o es for coinegraed variables. An approach pioneered by Soren Johansen uses he maximum likelihood echnique o idenify long-run equilibrium relaionships beween non-saionary variables, and uses he vecor auoregressive framework. 8 The VAR framework is applied o he full-sysem esimaion model wih he variables in levels: z k = Ω z i= 1 i i + ε, ε ~ Niid(0,Ω), (13) where z is an n 1 marix of he n variables of ineres in he sysem of equaions. The model is hen reparamerized, o accoun for he non-saionariy of he variables in levels, as follows: z k 1 = Γi z i z + + Π k ε, ε ~ Niid(0,Ω). (14) i= 1 All of he firs-differenced variables ( z) mus be saionary. Πz -k mus also be saionary. The marix Π will deermine he exen of coinegraion. If Π is a null-marix, hen no saionary long-run relaionships exis amongs he variables in z. If Π is non-zero, hen Πz -k is saionary, and he rank, r, of he marix will indicae he number of coinegraing relaionships beween variables. Π is hen unpacked as Π = αβ T, where β is an n r marix of parameers denoing he r coinegraing relaionships amongs he n variables in z, and α is an n r marix of parameers denoing he speed of adjusmen of a dependen variable owards a long-run equilibrium (coinegraing) relaionship. I is in his 17

19 framework ha he long-run relaionship beween per capia income, invesmen and immigraion will be examined. If long-run relaionships may be esablished using he VAR framework wih coinegraed processes, hese relaionships do no imply any direcion of causaliy. One of he moivaions for his paper is o idenify, where possible, a direcion of causaliy. Sims, Sock, and Wason (1990) showed ha he sandard Granger-causaliy ess can be used in VARs when he variables are in heir levels (insead of firs-differences) in cases where a single coinegraing relaionship exiss in hree-variable sysems. Accordingly, he longrun relaionships beween income, invesmen, and immigraion in per capia erms will be examined, and block exclusion ess in he sysem of equaions described by Eq. (13) will be conduced when a single coinegraing relaionship exiss in a hree-variable sysem of he variables of ineres. As an example, assuming ha one long-run relaionship exiss beween income, invesmen and immigraion, hen o es if income, y, does no Granger-cause immigraion, im, in he regression, im k k k = + φiim i + ϑi y i + i= 1 i= 1 i= 1 α γ i + ε, (15) i i he null hypohesis ha ϑ i is equal o zero for all i is esed. If he null hypohesis is rejeced, hen one may infer ha y Granger-causes im, and ha increased produciviy led o increased immigraion during he sudy period. This mehodology was used by Rousseau and Wachel (1998) and by Rousseau (1999), o examine he relaionships beween financial indicaors and income in per capia erms, for many counries in he decades leading up o he Grea Depression. The firs sep in proceeding wih his empirical invesigaion is o deermine wheher or no he series of ineres in his sudy are non-saionary in heir logarihmic 18

20 forms, and wheher or no hey are saionary in heir firs differences. The sandard Augmened Dickey-Fuller Tess were conduced for his purpose. 9 The es resuls are presened in Table 1. All series in levels appear o be non-saionary, while all series in firs differences appear o be saionary. 4.3 Applicaion: Capial Accumulaion, Income and Immigraion Equaion (14) will be esimaed, where z = [s y im] T, [i y im] T, or [s i im] T, and in each case, Π=αβ T. In his conex, β is a 3 r marix of parameers denoing he r coinegraing relaionships amongs variables in z, and α is a 3 r marix of parameers denoing he speed of adjusmen of a dependen variable owards a long-run equilibrium (coinegraing) relaionship. The unresriced VAR analysis indicaed ha a value of k se o hree for Eq. (14) generaed saisfacory residuals for he mulivariae sysems. Tess were hen conduced o idenify he rank of he marix Π. The Johansen L-max rank es resuls, presened in Table 2 suggesed ha here exiss only one coinegraing relaionship amongs he variables in wo of he hree-variable sysems. 10 A coinegraing relaionship does no exis beween i, y, and im. The parameer esimaes ha describe he long-run saionary equilibrium relaionship for wo of he hree-variable sysems are presened in Table 3. These relaionships are: cv 1 ˆ T β1 z1, 3 s im y 3 ; (16) = cv 2 ˆ T β 2 z2, 3 i im s 3. (17) = 19

21 These relaionships may be re-wrien o focus on savings and is relaionship wih immigraion and income, and o focus on invesmen, and is relaionship wih immigraion and domesic savings. I mus be noed, however, ha a direcion of causaliy canno be esablished ye in his esimaion framework, bu resrucuring hese relaionship as follows provides an easier way o hink abou hese resuls: cv 1 cv 2 s = 0.142im y ; (18) i = 0.120im s. (19) The resuls for he firs coinegraing relaionship sugges ha, in he long run, increases in he rae of immigraion and in income coincided wih increases in domesic savings, in per capia erms. The resuls for he second coinegraing relaionship sugges ha increases in invesmen coincided wih increases in immigraion and domesic savings. In ligh of he model and findings presened in Wilson (2003), an increase in immigraion can increase domesic savings as he wealh of immigrans is ransferred o he receiving counry. This auomaically increases he measured amoun of domesic savings in he naional accouns of Canada as presened in Urquhar (1993). Immigran wealh was reaed as immigran receips in Canada s balance of inernaional paymens. This reamen of immigran wealh serves o reduce he capial accoun surplus, as compared o he curren mehod of reaing immigran wealh as a foreign capial inflow. The domesic savings series is calculaed as he difference beween invesmen and foreign capial inflows, and so, an increase in immigran wealh serves o increase domesic savings, by reducing ne foreign capial inflows. I is also possible ha an increase in domesic savings can increase he reurns o labor, and provide incenives for labor o immigrae. 20

22 Produciviy improvemens can lead o increased domesic savings. If more income flows in he economy hen more is available for boh consumpion and savings, and if produciviy increases because of an increase in he employmen rae, hen more individuals in he economy are able o save ou of heir incomes. Alernaively, i is possible ha increased domesic savings can lead o increases in produciviy as savings increase domesic physical capial for use in producion, and increase he reurns o labor, providing more jobs for workers, including immigrans. The second coinegraing relaionship suppors he suggesion ha when immigraion increases, more workers need o be equipped for producion and he incenives for invesmen increase because capial becomes more producive in ligh of more labor in he economy. Increases in invesmen also corresponded wih increases in domesic savings. Again, a direcion of causaliy canno be esablished ye using hese mehods, bu he mehods have esablished ha long-run relaionships exis beween hese variables. Wih his in mind, causaliy ess may now be conduced. Since VARs in levels may be used o es for causaliy in hree-variable sysems wih one coinegraing relaionship, block exclusion ess are conduced for each dependen variable using Eq. (15). The resuls for he block exclusion ess are presened in Table 4. The Granger-causaliy es resuls sugges ha increases in boh domesic savings and produciviy caused increased immigraion, bu ha i is no possible o esablish a direcion of causaliy beween domesic savings and produciviy. The resuls also sugges ha increased invesmen caused increased immigraion, and ha increased domesic savings caused increased invesmen in Canada in he years leading up o he Grea Depression. One may infer from hese resuls ha invesmen from domesic 21

23 sources of savings, along wih produciviy improvemens, preceded immigraion booms in Canada over he period. 4.4 Applicaion: Foreign Capial Flows, Income and Immigraion Tradiionally, inernaional capial flows in he years leading o WWI, especially in he case of Briish capial flows o Canada, were geared o social infrasrucure and overhead needs. Bloomfield (1968) and Edelsein (1982) presened resuls suggesing ha capial flows from Grea Briain o Canada and o he US were posiively relaed o railroad developmen. Simon (1970: 242) showed ha over 80% of money calls were direced o governmen and governmen-backed deb, and for social overhead requiremens. I would be very useful and insrucive o use he same mehodology o invesigae he relaionships beween foreign capial flows, invesmen, income, and immigraion. Did foreign capial chase migrans o Canada? Long-erm ne foreign capial inflow daa is used in he mehodology described above o answer his quesion, by examining he rios, (f, y, im), and (f, i, im), where f denoes he logarihm of per capia long-erm foreign capial inflows. Unforunaely, in 1928, long-erm ne foreign capial inflows are negaive, and so he sample mus be shorened o he period. All relevan ime series were found o be difference saionary over his shorened period. Table 5 presens he Augmened Dickey-Fuller es resuls for he foreign capial inflow series. For each rio, one long-run equilibrium relaionship was found o exis. The Johansen Rank es saisics are presened in Table 6, and he equilibrium relaionships are presened in Table 7. These relaionships are: 22

24 cv 3 ˆ T β 3 z3, 3 f im y 3; (20) = cv 4 ˆ T β 4 z4, 3 f im i 3. (21) = Once again, hese relaionships may be re-wrien for convenience, however, a direcion of causaliy canno be esablished ye in his esimaion framework: cv 3 cv 4 f = 0.818im y ; (22) f = 0.538im i. (23) These resuls sugges ha, in he long run, increases in ne foreign capial inflows coincided wih increases in immigraion, wih increases in invesmen, and wih increases in income. Causaliy esing in hese hree-variable sysems allows us o idenify a direcion of causaliy. The es resuls are presened in Table 8 and show ha immigraion Grangercaused foreign capial inflows, and ha foreign capial inflows Granger-caused invesmen. These resuls sugges ha foreign capial was a srong conribuor o he massive invesmen booms in Canada, and ha foreign capial did indeed chase migrans o Canada in he six decades before he Grea Depression. 4.5 Discussion Canada experienced several produciviy, invesmen, and immigraion booms over he six decades before Long-run economic relaionships were found beween produciviy, immigraion, and invesmen, and beween produciviy, immigraion, and invesmen in he form of domesic savings and in he form of foreign capial inflows. During he period of sudy, increases (or decreases) in invesmen were associaed wih similar movemens in incomes and immigraion. The same may be said for he 23

25 componens of invesmen based on he sources of supply: increases (or decreases) in domesic savings were correlaed wih similar movemens in incomes and immigraion; and increases (or decreases) in foreign capial inflows were associaed wih similar movemens in incomes and immigraion. Block exclusion ess were hen used o esablish direcions of causaliy beween he variables of ineres in an aemp o idenify he leading cause or causes of hese coinciding booms. The resuls sugges ha he immigraion and invesmen booms in Canada sared as a resul of improvemens in produciviy combined wih increases in domesic savings ha were funnelled ino he invesmen in physical capial. Foreign capial flows hen followed migrans ino Canada in search of profiable invesmen opporuniies, which in urn prolonged he invesmen booms. Migraion was found o be endogenous in relaion o income and invesmen in physical capial (saisfied by domesic sources of supply firs). As a resul, populaion movemens should be reaed as an endogenous facor in Canadian economic developmen. I appears ha migrans were araced o Canada when economic and labor marke condiions improved in Canada. However, i canno be saed ha pas sudies ha reaed populaion as exogenous are compleely misleading. Immigraion was found o be moivaing facor inducing he inflow of foreign capial. Capial did indeed chase migrans ino Canada, and served o prolong he invesmen booms. Briish capial flows o Canada, were geared o social infrasrucure and overhead needs (Simon, 1970: 242). Immigrans also brough much capial along wih hem, which was reaed as immigran receips in Canada s balance of inernaional paymens in Urquhar (1993), and served o increase he measure of domesic savings. Thus, in he case of foreign capial inflows and 24

26 domesic savings, he reamen of migran flows as exogenous o capial accumulaion is no enirely misconceived, however, his sudy has shown ha surges in invesmen supplied by domesic sources, along wih produciviy improvemens, led o increased immigraion. These findings suppor he following sequence of evens: (i) he Canadian invesmen booms from 1870 o 1929 were firs he resul of increased invesmen in physical capial due o echnological change, resource developmen, and produciviy improvemens; (ii) he demand for invesmen funds was firs me by domesic sources of supply; (iii) in he early sages of he invesmen booms, he incenives for labor o migrae o Canada increased, leading o immigraion booms; (iv) as migran flows o Canada increased, foreign capial followed, o help saisfy he increased demand for capial o equip workers for producion and o saisfy social overhead needs; and (v) foreign capial inflows prolonged he invesmen booms. This sequence of evens seems naural and is no in iself surprising. A conribuion of his aricle is in presening empirical evidence, using recenly developed echniques, ha backs up his sequence of evens. This aricle also presens evidence ha populaion change as a resul of migraion, mus be viewed as endogenous o changes in per capia incomes (which serves as a proxy for produciviy) and invesmen when supplied by domesic sources, and ha foreign capial flows mus be viewed as endogenous wih respec o migraion flows. This sudy has no examined he effec of hese booms on he reurns o he facors of producion. Green (1994) and Greasley, Madsden, and Oxley (2000) examined he effec of immigraion on he reurns o labor in Canada. Green used a counerfacual exercise o show ha immigraion in he 1920s depressed real wage growh on he 25

27 Canadian Prairies. Greasley, Madsden, and Oxley, armed wih he new real wage series for Canada, provided in Williamson (1995), also concluded ha immigraion slowed down he growh rae of real wages over These resuls are expeced given he heoreical background based on he Solow model, provided in Secion 3. Unforunaely, daa on he reurns o capial in Canada are sparse. 11 The analysis of early Canadian economic developmen is consrained since labor force and capaciy uilizaion daa are no available, nor are daa ha adequaely quanify echnological improvemens, human capial or naural resources. For example, i would be very useful o have annual labor force daa o examine he effec of he business cycle on local labor markes, o explore how slack and igh labor markes affec migraion paerns, and o deermine he associaed ime lags for migrans o respond o Canadian labor marke condiions. 5. Conclusion In he six decades leading o he Grea Depression, Canada experienced several coinciden produciviy, savings, foreign invesmen, and immigraion booms. The exising lieraure on Canadian economic developmen has examined he reasons for economic growh, wesern selemen and immigraion. The lieraure has also examined he effecs of populaion on capial accumulaion and foreign capial inflows. I has been unclear, however, if growh preceded invesmen and immigraion, if invesmen preceded growh and immigraion, or if immigraion preceded growh and invesmen. In addiion, he radiional view during he mass migraions before WWI is ha capial chased afer European migrans (Taylor and Williamson, 1994). However, foreign 26

28 capial has also been radiionally seen as more mobile han labor. Did foreign capial chase afer migrans ino Canada, did hey boh move ogeher ino Canada seeking ou profiable opporuniies, or did foreign capial lead migrans o Canada? This aricle presened a heoreical framework based on he Solow model ha illusraed he relaionships beween produciviy growh, capial accumulaion, labor growh, and facor migraion. Given hese heoreical relaionships, an economeric model using coinegraion and vecor auoregression echniques was esimaed o define he long-run relaionships beween income, immigraion, invesmen, domesic savings, and foreign capial inflows for Canada over he 1870 o 1929 period. Causaliy es procedures were hen used o help sor ou he facors leading o he massive booms in Canadian developmen. The causaliy es resuls presened in his paper sugges ha changes in produciviy and domesic savings led o changes in immigraion, and ha changes in immigraion paerns led o changes in foreign capial flows o Canada. Populaion changes as a resul of migraion were endogenous wih respec o changes in per capia incomes and invesmen supplied by domesic sources. In urn, foreign capial flows were endogenous wih respec o migraion flows. These resuls suppor a sequence of evens whereby surges in domesic savings coincided wih surges in produciviy. These forces hen caused migrans o flow ino Canada looking for profiable employmen and business opporuniies. Foreign capial hen chased hese immigrans o saisfy he increased demand for capial in producion and o mee infrasrucure requiremens. 27

29 Appendix: Daa Sources The daa was aken from Urquhar (1993), Urquhar and Buckley (1965), and Dick and Floyd (1993). The number of immigran arrivals and he populaion figures are from Urquhar and Buckley (1965). The income series was consruced using he GNP series from Urquhar (1993) from 1870 o 1926, and from Urquhar and Buckley (1965) from 1926 o In 1926, he series was spliced, aking averages of he overlapping series. The invesmen series corresponds o he gross domesic capial formaion series in Urquhar (1993) from 1870 o 1926, and in Urquhar and Buckley (1965) from 1926 o In 1926, he series was spliced, using he average of he wo overlapping series. The domesic savings series was calculaed as a residual of gross domesic invesmen less he capial accoun balance (or plus he curren accoun balance). The capial accoun balance is derived from Urquhar (1993) from , and from Urquhar and Buckley (1965) from , and spliced accordingly in Long-erm foreign capial inflows were aken from Dick and Floyd (1994: 190-1) for , from Urquhar (1993: 19-23), for 1870, and , and from Urquhar and Buckley (1965), series G98, for Series were convered ino real values using he GDP deflaor provided in Urquhar (1988), wih 1981 as he base year. References Abramoviz, M. (1993), The search for economic growh: areas of ignorance, old and new. Journal of Economic Hisory 53 (2): Barro, R., and Sala-i-Marin, X. (1995), Economic Growh, McGraw-Hill, New York. 28

30 Bloomfield, A.I., (1968), Paerns of Flucuaions in Inernaional Invesmen Before 1914, Princeon Sudies in Inernaional Finance No.11, Princeon: Princeon Universiy, Deparmen of Economics, Inernaional Finance Secion. Borins, S. (1982), Wesern Canadian homeseading in ime and space. Canadian Journal of Economics 15: Charemza, W.W., and Deadman, D.F. (1992), New Direcions in Economeric Pracice. Edward Elgar Publishing Limied, Vermon. Davidson, R., and McKinnon, J.G. (1993), Esimaion and Inference in Economerics. Oxford Universiy Press, New York. Dick, T., and Floyd, J. (1992), Canada and he Gold Sandard, Cambridge: Cambridge Universiy Press. Edelsein, M. (1982), Overseas Invesmen in he Age of High Imperialism: The Unied Kingdom, , New York: Columbia Universiy Press. Engle, R.F., and Granger, C.W.J. (1987), Co-inegraion and error-correcion: represenaion, esimaion and esing. Economerica, 55: Fuller, W. (1976), Inroducion o Saisical Time series, John Wiley, New York. Granger, C.W.J. (1981), Some properies of ime series daa and heir use in economeric model specificaion. Journal of Economerics, 16: Greasley, D., Madsden, J.B., and Oxley, L. (2000), Real wages in Ausralia and Canada, : globalizaion versus produciviy. Ausralian Economic Hisory Review 40 (2): Green, A.G. (1994), Inernaional migraion and he evoluion of prairie labor markes in Canada, In Migraion and he Inernaional Labor Marke,

31 1939. T.J. Haon and J.G. Williamson, ediors. New York: Oxford Universiy Press. Green, A.G. and Green, D.A. (1993), Balanced growh and he geographic disribuion of European immigran arrivals o Canada, Exploraions in Economic Hisory 30: Green, A.G. and Sparks, G.R. (1999), Populaion growh and he dynamics of Canadian developmen: a mulivariae ime series approach. Exploraions in Economic Hisory, 36: Green, A.G. and Urquhar, M.C. (1976), Facor and commodiy flows in he inernaional economy of : a muli-counry view. Journal of Economic Hisory : Green, A.G. and Urquhar, M.C. (1994), New esimaes of oupu growh in Canada: measuremen and inerpreaion. In Perspecives on Canadian Economic Hisory. D. McCalla and M. Huberman, Ediors. Torono: Copp Clark Longman Ld. Haon, T., and Williamson, J. (1994), Wha drove he mass migraions from Europe in he lae nineeenh cenury? Populaion and Developmen Review 20 (3): Haon, T., and Williamson, J. (1998), The Age of Mass Migraion, New York; Oxford Universiy Press. Johansen, S. (1988), Saisical analysis of coinegraion vecors. Journal of Economic Dynamics and Conrol, 12: Johansen, S and Juselius, K. (1994), Idenificaion of he long-run and shor-run srucure: an applicaion o he ISLM model. Journal of Economerics 63:

32 Juselius, K. (1991), CATS in RATS: A manual o coinegraion analysis. Insiue of Economics, Universiy of Copenhagen, Copenhagen. Lewis, F. (1981), Farm selemen on he Canadian Prairies, 1898 o Journal of Economic Hisory 41, 3: Lucas, R. Jr. (1988), On he mechanics of economic developmen. Journal of Moneary Economics 22: Marr, W. and Percy, M. (1978), The governmen and he rae of prairie selemen. Canadian Journal of Economics 11: McLean, I. (1994), Saving in seler economies: Ausralian and Norh American comparisons. Exploraions in Economic Hisory 31: Neufeld, E.P. (1972), The Financial Sysem of Canada. Torono: Macmillan Company of Canada. Norrie, K. (1975), The rae of selemen of he Canadian Prairies, Journal of Economic Hisory 25: O Rourke, K., Williamson, J., and Haon, T., (1994), Mass migraion, commodiy marke inegraion and real wage convergence: he lae-nineeenh cenury Alanic economy. in Migraion and he Inernaional Labor Marke, T.J. Haon and J.G. Williamson, ediors. New York: Oxford Universiy Press. Rebelo, S. (1991), Long-run policy analysis and long-run growh. Journal of Poliical Economy 99: Rich, G. (1988), The Cross of Gold: Money and he Canadian Business Cycle, , Oawa: Carleon Universiy Press. Romer, D. (1996), Advanced Macroeconomics, McGraw-Hill, New York. 31

33 Romer, P. (1986), Increasing reurns and long-run growh. Journal of Poliical Economy 94: Rousseau, P. and Wachel, P. (1998), Financial inermediaion and economic performance: hisorical evidence from five indusrialized counries. Journal of Money, Credi and Banking 30 (4): Rousseau, P. (1999), Finance, invesmen and growh in Meiji-era Japan. Japan and he World Economy 11: Simon, M. (1970), New Briish invesmens in Canada, Canadian Journal of Economics 3 (2): Sims, C.A., Sock, J.H., and Wason, M.W., (1990), Inference in ime series models wih some uni roos. Economerica 58: Solow, Rober M. (1957), Technical change and he aggregae producion funcion. Review of Economics and Saisics, Vol. 39, pp Taylor, A., and Williamson, J. (1994), Capial flows o he New World as an inergeneraional ransfer. Journal of Poliical Economy 102: Urquhar, M.C., and Buckley, K. (1965), Hisorical Saisics of Canada, Cambridge: Cambridge Universiy Press. Urquhar, M.C. (1988), Canadian Economic Growh Discussion Paper No Insiue for Economic Research, Queen s Universiy. Urquhar, M.C.(1993), Gross Naional Produc, Canada : The Derivaion of he Esimaes. McGill-Queen s Universiy Press, Kingson. Ward, T. (1994), The origins of he Canadian whea boom, Canadian Journal of Economics 27:

34 Williamson, J. (1979), Inequaliy, accumulaion, and echnological imbalance: a growhequiy conflic in American hisory? Economic Developmen and Culural Change 27: Williamson, J. (1995), The evoluion of global labor markes since 1830: background evidence and hypoheses. Exploraions in Economic Hisory 32: Wilson, S.J. (2000), The savings rae debae: does he dependency hypohesis hold for Ausralia and Canada? Ausralian Economic Hisory Review 40 (2): Wilson, S.J. (2003), A dynamic equilibrium analysis of migraion and capial accumulaion: he case of Canada. Review of Economic Dynamics 6: Zevin, R. (1992), Are world financial markes more open? If so, why and wih wha effecs? In Financial Openness and Naional Auonomy, Oxford: Clarendon Press. 33