Ticket to trade: Belgian labour and globalization before 1914

Ticket to trade: Belgian labour andglobalization before 19141

By MICHAEL HUBERMAN

Standard trade theory, as invoked by political scientists and economists, wouldanticipate that workers in Belgium, a small Old World country, rich in labour relativeto land, were in a good position to benefit from the wave of globalization before 1914.However, wage increases remained modest and ‘labour’ moved slowly towards adopt-ing a free-trade position. Beginning in 1885, the Belgian labour party backed freetrade, but its support was conditional on more and better social legislation. Belgianworkers’ wellbeing improved in the wave of globalization, but the vehicle was labourand social legislation and not rising wages.

In 1885 at its founding congress, the Parti ouvrier belge (Belgische WerkliedenPartij; Belgian Workers’ Party; hereafter the POB) made free trade an integral

part of its programme.2 By the dawn of the twentieth century, many Europeansocialist and workers’ parties came to share the POB’s position on commercialpolicy.3 Schooled in Ricardian economics, Marx himself was an early devotee oflaissez-faire, declaring in 1848 in a celebrated address in Brussels that ‘free tradehastens the social revolution’.4 But, ever the pragmatist, Engels reasoned thatprotective tariffs would bolster jobs and membership in trade unions, and in the1880s, his opinion prevailed among Germans socialists—the leading and mostinfluential movement on the continent. The 1890s saw a radical change.5 Ayounger cohort of marxists, led by Eduard Bernstein, rejected Bismarckian social

1 Earlier versions of this paper were presented at seminars at Harvard University, the Universities of Oxford andToronto,Trinity College, Dublin, Universitat Pompeu Fabra, at the annual meetings of the Social Science HistoryAssociation (Chicago) and the Canadian Economics Association (Hamilton, Ontario), at the Canadian Instituteof Advanced Research (Montreal), and the TARGET-INE conference (Vancouver). I would like to thankparticipants at these meetings and seminars, three anonymous referees of this journal, and John Brown, LeonardDudley, Herb Emery, Serge Jaumin, Sibylle Lehmann, Peter Lindert (who suggested the title), Chris Meissner,Peter Scholliers, and Peter Solar for their comments. All errors are my own.The support of the SSHRC INE grant‘Globalization, Technological Revolutions, and Education’ is gratefully acknowledged.

2 The POB programme sought ‘le remplacement des impôts de consommation et des taxes douanières par un impôtprogressif sur le revenu’ (cited in Delsinne, Parti ouvrier, p. 71). Into the interwar years, this demand was repeatedat the annual conferences of the POB; Partie Ouvrier Belge, Rapports, various years. On the POB, see Abs,‘Statuts’; idem, Histoire du parti socialiste.

3 An exception was the French socialist party, which remained ambivalent towards free trade. On Frenchsocialists and globalization, see Berger, Première mondialisation, pp. 53–62. For contemporary views on workersand free trade across New and Old Worlds, see Hollande, Défense ouvrière, pp. 71–89; Prato, Protectionnismeouvrier, pp. 189–224.

4 Marx, ‘Speech’, p. 270. In later years, Marx was less sanguine on the benefits of free trade: ‘[I]deas on freetrade [h]ave long since lost any and every theoretical interest, even if they may still be of some practical interestto some state or another’ (Capital, vol. III, p. 922).

5 At its Gotha congress in 1876, the German socialists declared themselves ‘étrangers à la lutte entre la protectionet le libre-échange’, but in Stuttgart in 1899, echoing the POB, their view had dramatically changed: ‘La politiqueprotectionniste est inconciliable avec les intérêts du prolétariat, des consommateurs, de l’évolution économique et poli-tique . . . et la démocratie’ (cited in Milhaud, Congrès socialiste, p. 41). On German socialists and trade, see Ashley,Tariff history, p. 61; Steenson, After Marx, pp. 92–4.

Economic History Review, 61, 2 (2008), pp. 326–359

© Economic History Society 2007. Published by Blackwell Publishing, 9600 Garsington Road, Oxford OX4 2DQ, UK and 350 MainStreet, Malden, MA 02148, USA.

and commercial policy and extolled instead the benefits of globalization.6 Thishistory was well known to an older generation of historians, Gerschenkron andBairoch in particular, who attributed the revival of worker support for free trade onthe continent to the import of cheap grains from the New World.7

Recently, historians have challenged this view. Tilly claimed that globalizationgenerated a backlash among European workers.8 The modern nation state, whichis identified with the rise of the welfare state, was partially a response to theunfettered forces of international economic integration. Helpless in the wake ofthese forces, workers demanded safety nets because they contested free trade.Rejecting a class-based analysis,Trentmann situated workers’ attitudes to marketsin their political and cultural contexts. During the mid- to late-Victorian period,the dominant ideology of consumption had conditioned workers, like othergroups, to remain loyal to openness.There were signs of stress, however. Althoughthe British Labour Party officially supported free trade at its 1904 conference, thismasked growing unease in certain political circles and industrial sectors withlaissez-faire.Trentmann brought forth evidence that after 1900, when cheap graincould no longer be assured, many on the left turned toward the regulation ofmarkets to secure standards of living, although they remained hostile to protec-tionism. In conjunction with the Fabians and others who claimed better safeguardsand new social entitlements for workers in an increasingly risky and competitiveenvironment, these dissident voices raised serious concerns about the prevailingorthodoxy on the benefits of openness.9

In this paper, I propose a reinterpretation of workers’ attitudes—through theprism of organized labour—towards free trade. Because of data limitations, Icannot identify individual workers’ preferences towards free trade, but I do haveevidence on how workers’ organizations represented their choices. Certainly, thereis not a simple and direct mapping from individuals, even where they are homo-geneous, to their political representatives, because the latter may have their ownpriorities and preferences. In this regard, I follow the standard approach that seestrade policy as the outcome of supply (political parties) and demand (individualand groups of workers and firms).10

I will argue that the older school of historians was right. In many parts ofEurope, organized labour was an active participant in the free trade movement andfor, literally, bread and butter issues. However, the grain invasion is only part of thestory, because the left’s support for an open commercial policy came 20 years after

6 Lujo Brentano, the well-known German social reformer, was an earlier supporter of free trade. See Hage-mann, ‘Verien fùr Sozialpolitik’, pp. 154–60.The classic statement by Bernstein on trade is ‘Zum Kampf’, p. 693.On German marxism before the War, see Gay, Dilemma, pp. 213–50. Fletcher, Revisionism, p. 145, describedBernstein as a ‘good Cobdenite’.

7 Bairoch, ‘European trade’, pp. 131–5; Gerschenkron, Bread, pp. 33–6.8 Tilly, ‘Globalization’. See also Noiriel, ‘L’historien face aux défis’; van der Linden, ‘National integration’.

This line of reasoning has become the standard view in recent ‘world history’ textbooks. Bayly, Birth, p. 243,wrote: ‘After the 55 years from about 1815 to 1870, when free trade was the order of the day, economicprotectionism on national lines was more and more apparent as the century drew to an end. While flows ofinternational trade, labour, and capital grew exponentially, nation states sought rigorously to control and directthem to their own ends’. Critics of globalization echo this position. Silver, Forces of labor, p. 17, is typical: ‘[T]helate–nineteenth and early twentieth century globalization of markets produced a strong countermovement fromworkers and other social groups’.

9 Trentmann, ‘Fiscal reform’; idem, ‘Wealth’; idem, ‘Political culture’; idem, ‘National identity’.10 See, for example, Mayda and Rodrik, ‘Why are individuals more protectionist?’

BELGIAN LABOUR AND GLOBALIZATION 327

© Economic History Society 2007 Economic History Review, 61, 2 (2008)

the initial decline in prices. An alternative explanation, frequently alluded to, isthat certain leaders on the left held the view that globalization was the harbingerof the socialist international. However, this line of reasoning cannot explaincontinued support for globalization even as the vision of an international wanedbefore 1914, and it is not certain how organized labour would have convincedits rank and file of this idea anyway.11 Instead, my claim rests on the appeal madeby labour to its constituents that increasing economic integration orglobalization—the preferred term in the period was internationalism—was com-patible with more and better labour protection and social entitlements. Global-ization affected the pace, if not the process, of structural change, and this resultedin a new set of risks faced by workers in both import-competing and exportsectors. Unions, and at times paternalistic employers, helped to insulate skilledworkers from these shocks, but the bulk of semi- and unskilled male workers, andwomen and children, fell outside these traditional safety nets. After the mass ofEuropean male workers gained voting rights, their political representativesdemanded more and better labour and social safeguards to cushion them againstthe new hazards at the workplace, including long hours, poor and unsanitaryconditions, greater risks of injury, and lay-offs.

For both ideological and practical reasons, the left tied these demands to freetrade. Protectionism was seen as incompatible with the development of socialreform.12 Continental protectionism tended to redistribute income to landlordsand employers in key sectors, and, in any case, tariffs did not generate sufficientstate revenues to support the innovative package of reforms demanded by labour.Free trade would force states to find alternative sources of revenue based on newand, in principle, more equitable tax measures that would lay the basis for labourprotection and social entitlements at the national level. Unlike tariff protection,which tended to guarantee employment security for targeted industries andregions, the ‘welfare state’ would be universal.

At the practical level, the narrative behind organized labour’s adoption of freetrade is more nuanced and contextualized.While free trade became a fixture in theBelgian left’s programme, like any political group the POB occasionally supportedprotection for selected groups of workers, especially those faced by import com-petition in key electoral districts. In fact, this mixed strategy had the effect ofempowering labour in its demands for social legislation. As ever, the ideologicalmake-up of the Belgian Parliament was divided across and within parties andcoalition brokering and strategizing were the norm. While a majority of liberalsmay have supported an open door commercial policy, they needed the POB’ssupport to strengthen their place in Parliament; similarly, labour needed to findsupport elsewhere to ensure that their demands for social reform were acted upon,and while some conservatives recognized the need for social protection, the POBused their conditional backing for free trade in coalition with liberals to garnerimproved legislation. To be clear, if labour support for free trade was seen to be

11 On the rise and fall of the socialist international before 1914, see Donald, ‘Workers of the world’. On liberalinternationalism, see Angel, Illusion. More successful was governmental internationalism, an approach exploitedby the Belgian state and its offices. See Herren, Hintertüren zur Macht.

12 Of course, after these reforms had been instituted, workers might have demanded tariff protection to preservetheir vested interests. This was not the case with Belgian socialists, even in the interwar years, although certainlabour groups have made this claim during the most recent wave of globalization.

328 MICHAEL HUBERMAN


unambiguous, its bargaining power would have been weaker, because it could nothave exercised pressure on the liberals. In this way, labour used its support of tradeas a lever for social reform. I conclude that globalization was the handmaiden ofthe modern nation state, not its arch-rival.

The Belgian example is particularly relevant because it is illustrative of the majorturning points that European labour traversed in adopting free trade. Belgianeconomic and political development was an amalgam of European models. LikeBritain, its industrialization was built on coal and textiles, and as in Germany andFrance, its agricultural sector remained politically, if not economically, significantinto the late-nineteenth century. However, as a small open economy, its develop-ment resembled that of the Netherlands and the Scandinavian countries. Theworkforce was preponderantly unskilled. The development of the Belgian welfarestate was also representative. In previous work, I have shown that in a sample ofEuropean countries it sits on the regression line that best fits the relation betweenlabour and social legislation and levels of openness.13 This paper goes one stepfurther. It is my claim that the dynamics underlying the relationship betweenopenness and social legislation in Belgium illustrate the wider tendencies acrossthe continent.

As elsewhere, organized labour’s support for free trade in Belgium was notinitially assured, coalescing only in the 1880s; until then, it was difficult for labourto find agreement among its constituents because individual attitudes to tradewere based on sectoral interests. I attribute the sluggish move towards free trade tolimited inter-industry factor mobility that created divisions over commercial policyamong workers.With the expansion of the train network and subsidized transpor-tation in the 1880s, mobility increased and a greater number of workers came totake a unified position on trade—the central tenet of standard trade theory—thatwas reflected in the stance of the POB at its first congress. There were of courseworkers in declining sectors whose mobility was limited, and for other groups,trade was not a salient issue. Labour was compelled to represent their interests aswell, an outcome entirely consistent with the claims of the new school of politicaleconomy that workers’ identity was less uniform than trade theory would presume.However, all workers sought better regulation, and the POB was inclined tosupport an open commercial policy in Parliament to this end. In fact, in theabsence of labour’s pressure, Belgium may have returned to protectionism likesome of its continental neighbours.

The paper proceeds as follows. The first section addresses the question of whyBelgian workers were slow to adopt a pro-free trade stance. Section II puts forwardthe proposition that the train network increased labour mobility and made workersinto a class, at least from the perspective of trade theory. In section III, a new dataset on wages is used to test this proposition. The wage data confirm that themid-1880s marked a transition from a world of immobile to mobile labour. Thetiming meshes with the POB’s acceptance of the virtues of laissez-faire, exactlywhat one would have anticipated from standard theory. Sections IV and V docu-ment how the POB exploited its role as the voice of labour to derive improvedlabour regulation and social entitlements. The conclusion, extending from the

13 Huberman and Lewchuk, ‘Labour compact’, p. 29; Huberman, ‘Working hours’, p. 986.



Belgian experience, describes how labour elsewhere in New and Old Worlds, bothbefore 1914 and after 1945, endorsed free trade to gain expanded safety nets.

I

Workers in Belgium, a small Old World country, rich in labour relative to land,were in good position to benefit from the wave of globalization before 1914. On theexport side, Belgium’s labour-intensive industries like textiles ought to haveexpanded, and on the import ledger the country would have benefited from cheapforeign grains. In an open economy, real wages in principle should have convergedto levels in less labour-abundant and more land-rich countries. Belgium was alsorich in capital relative to land. All together, because of its factor endowments, size,and dependence on trade, Belgium would have escaped the curse of diminishingreturns and growth would have continued unabated.14 With regard to commercialpolicy in the three-factor trade model, it is well known that making inferences ofattitudes to trade from the relative proportions of factor endowments is notstraightforward. Here I follow Rogowski’s simplification that ‘the land-labor ratioinforms us fully about any country’s endowment of those two factors’.15 Wherelabour is relatively abundant, regardless of the capital–labour ratio, it would havemuch to gain from supporting free trade, while landed interests would oppose it.The problem for Belgium is not so much that it had multiple endowments, butthat, even assuming the simplified model, factor rewards did not move in line withtheory. In this section, I concentrate on the evolution of wages and land rents andhow their movements affected the attitudes of key social groups towards trade.16

Belgium was an active player in the shift toward free trade initiated by theCobden-Chevalier treaty of 1860.17 It signed its own trade agreement with Francein 1861, followed rapidly by treaties with Britain (1862), the Netherlands (1863),and later with Prussia (1865). As measured by tariff revenues as a percentage of thevalue of imports (average [1], figure 1), Belgium’s commitment to openness wasstable throughout the period. However, this method of calculation masks signifi-cant sectoral differences.The initial trade treaties lasted normally for 10 years, butwhen agreements with two of its major trading partners, France (1872) andGermany (1878), were not renewed, tariff schedules were altered. For three maintypes of traded goods, wheat, livestock, and cotton goods, figure 1 converts tariffschedules that were based on weight and volume into ad valorem rates. (Average [2]is an unweighted average of the major sectoral tariffs.) After 1873 and for theremainder of the period, Belgium escaped the rise in grain duties that gripped its

14 For a survey of recent models of trade and growth, see Helpman, Economic growth, pp. 14–35.15 Rogowski, Commerce, p. 6. See also P. Dutt and D. Mitra, ‘Political ideology and endogenous trade policy: an

empirical investigation’, NBER working paper no. 9239.16 For various historical applications of the standard trade model, see Gourevitch, Politics; Irwin, ‘Industry or

class’; idem, ‘Political economy’; O’Rourke and Williamson, Globalization. The latter cite George Stigler: ‘IfCobden had spokenYiddish, and with a stammer, and if Peel had been a narrow, stupid man, England would havemoved towards free trade in grain as its agricultural classes declined and its manufacturing and commercialclasses grew’ (p. 78). Trentmann and Daunton, ‘Worlds of political economy’, have exposed the assumptionsunderlying these models, emphasizing the heterogeneity of labour and capital. I return to these debates below.

17 Even before Cobden-Chevalier, under the leadership of Walthère Frère-Orban, Belgium had negotiatedbilateral trade agreements with Britain and the Netherlands. See Kossmann, Low Countries, p. 232; Degrève,Commerce extérieur, pp. 34–6. Because of the large number of treaties signed in the 1860s and the extension ofmost favoured nation clauses, the special tariff was widely applied and the general tariff rarely invoked. Bairoch,‘European trade’, p. 43.



large continental neighbours. There was free trade in cereals, but the livestocksector garnered increasing protection over the period. The commercial history oftextiles is more complex. In 1882, while the weaving sector saw its level ofprotection decline, the yarn sector guarded itself from a decrease in duties.18 Thiswas to the disadvantage of weavers, who depended on supplies of yarn at worldprices. It was only after a lengthy debate on the tariff structure in 1895, an episodeto which I will return to later, that duties on yarn were cut back.19 As for thecapital-intensive sectors, the level of protectionism remained low in most sectorsthroughout the period.20 Certain industries, like glass manufacturing, thatdepended on skilled labour, were able to push through higher rates of protection,but these duties seem to have been an exception and had little effect on overallaverages.

Trade flows and price changes mirrored commercial policy. Belgium’s level ofopenness (values of its exports and imports divided by GDP) was about 60 percent in 1870, rising to about 80 per cent by the First World War. In Europe before1914, only Switzerland had a comparable degree of external exposure.21 Not all

18 On tariffs in cotton textiles, see Scholliers, ‘Industrie cotonnière’, pp. 134–7. The overall tariff in manufac-turing varied between 5% and 8%. On some items, like linen and hemp, duties were abolished in the early roundof tariff negotiations. See Bairoch, ‘European trade,’ pp. 42, 76; Estevadeordal, ‘Measuring protection’, pp. 106–7;O’Rourke and Williamson, Globalization, pp. 98–9; Belgium, Annales parlementaires, 1894–5, passim.

19 Irwin, ‘Interpreting’, proposed that tariff structures before 1914 were uneven (and non-protectionist)because they were intended to insure the coordination of excise and custom taxes, the most important source ofgovernment revenues in the period. For Belgium, Bairoch, ‘European trade’, p. 59, estimated that the share ofcustom duties in total central government revenue was relatively small, varying between 7.1% and 9.3%. Pollard,‘Free trade’, provides an excellent summary of effective protection in the period. He concluded that Belgium, likemost European countries, had a ‘system of moderate tariffs’ (p. 45).

20 In 1913, Belgium (after Switzerland) had the lowest tariff rates on sheet iron. Grunzel, Protectionism, p. 158.21 Exports per capita (in $1990) in 1913 were: Belgium, $954; the Netherlands, $702; Denmark, $500.

Population and export values are from Maddison, World economy, pp. 183, 359.

0

5

10

15

20

25

1870

Tar

iif r

ate

(%)

average(1) average(2) grains

livestock cotton yarn cotton textiles

1880

1890

1900

1910

Figure 1. Tariff rates by sectorNotes and sources: Average (1) is from Foreman-Peck and Lains, ‘European economic development’, pp. 103–5, and Degrève,Commerce extérieur, p. 39; average (2) is from Bairoch, ‘European trade’, pp. 42, 76; yarn (plain, 30–40,000 metres) and textiles(plain, 1st class, no. 35) from Scholliers, ‘Industrie cotonnière’, p. 137; all other figures from Belgium, Annales parlementaires,1894–5.



sectors participated in this trend, however. Table 1 gives export and import ratiosfor some important sectors. Wheat and other grains flowed in at impressivevolumes and coal shipments remained strong. The trade balance in livestockproducts improved somewhat. The fate of the textile sector is evidence of thelobbying power of certain groups of employers and workers. Realizing thedemands of spinners, the increased duties on yarn reduced imports to the detri-ment of weavers. Trade performance in the capital-intensive sector was stronger.Exports of processed iron and steel increased ten-fold from 1870 to 1900, andthen doubled from 1900 until 1913.22 The contraction in the export ratio in thedecade before the FirstWorldWar represented increased national demand and nota decline in competitiveness.

The overall picture is that the grain invasion in Belgium had mixed results. Inthe standard trade model, the import of foreign grains would have engendered aprice shock, lowering agricultural rents and promoting a transfer of resources tothe labour-abundant sector. Arable prices did contract by about 30 per cent from1870 until 1914 (figure 2) and production fell by about 25 per cent.23 However,exports of labour-intensive goods, here represented by cotton textiles, scarcelyprogressed before the tariff revision of the mid-1890s. Some resources weretransferred to dairy production, whose output doubled, but, sheltered behind a

22 Trade figures in this paragraph are from Degrève, Commerce extérieur, pp. 841–5. By 1913, the export ratioof processed iron and steel was 10 to 1.

23 All values from Blomme, Economic development, pp. 396, 418–9.

Table 1. Sectoral performance in international trade

Sector 1875 1880 1890 1900 1910

GrainsImport ratio 0.39 0.49 0.66 0.77 0.83Trade balance (%) -64.8 -84.3 -197.5 -271.4 -431.5

LivestockTrade balance (%) -20.0 -7.8 -8.1

MiningExport ratio 0.32 0.30 0.30 0.30 0.26Import ratio 0.07 0.07 0.11 0.18 0.28Trade balance (%) 27.1 25.3 20.7 14.3 -26.6

Iron and steelExport ratio 0.24 0.24 0.18 0.10Import ratio 0.05 0.03 0.02 0.02Trade balance (%) 20.5 21.7 12.9 5.7

Cotton spinningExport ratio 0.03 0.05 0.06 0.08 0.10Import ratio 0.03 0.02 0.04 0.06 0.06Trade balance (%) 14.2 20.1 20.3 25.0 28.7

Cotton textilesExport ratio 0.14 0.22 0.20 0.25 0.52Import ratio 0.05 0.08 0.10 0.12 0.10Trade balance (%) 13.5 25.6 25.0 32.8 47.1

Notes and sources: Export ratio = exports/production; import ratio = imports/consumption = imports/(production –exports + imports); trade balance = (exports - imports)/production. Grains are wheat, rye, and oats. Livestock is livestockfarming. Agricultural values from Blomme, Economic development, pp. 358, 366; mining and iron and steel from Mitchell, Historicalstatistics, pp. 430, 499; textiles from van Houtte, Industrie textile, pp. 257–9; all trade figures from Degrève, Commerce extérieur,pp. 180–6; 291–301; 361–7; 636–7, 844–5, except the livestock trade balance, which is from Blomme, Economic development,p. 280.



tariff wall, this sector produced mainly for the domestic market.24 Clearly, the graininvasion did not precipitate significant changes in the distribution of employment,as in theory it ought to have (table 2). In the 50-year period from 1846 to 1896,the share of workers in agriculture fell by only 0.5 per cent per annum; thereafterand until the First World War by 1.07 per cent per annum, a transition that will beexplained below. As for textiles, its numbers hardly changed; in the 50 years before1896, its growth rate was 0.03 per cent and its share of employment actuallydeclined.25 Again, it was only after the tariff revisions of the mid-1890s that textileemployment grew rapidly, at a rate of 2.70 per cent per annum. In contrast, theevolution of the relatively unprotected capital-goods industries, engineering andiron and steel, was steady, and this was also the case for the handful of otherindustries dependent on skilled workers.

Although the vision of a socialist international was honoured and the fall in grainprices was welcomed favourably, Émile Vandervelde, the leading Belgian socialistof the period and a specialist in agrarian conditions, documented the overcrowdingof the Belgian countryside. He reported that the number of tenants doubledbetween 1846 and 1880, amounting in the later year to 68 per cent of the ruralpopulation.26 In a more recent assessment, Blomme tracked the decline of

24 Livestock prices increased by 4%. The growth of the livestock sector is represented by the small change infeed grain prices in fig. 2. See Blomme, Economic development, p. 415.

25 The figures on employment in cotton spinning from the censuses are: 1846—6,980; 1880—7,153;1896—8,073; 1910—14,713. In the same years, weaving employed: 15,270; 18,935; 11,929; 25,012.The sourceof this information is Gadisseur, ‘Output per worker’, pp. 142–4.

26 Vandervelde, Socialisme agricole, p. 12. Rowntree reported similar results in Land & labour, pp. 70–89.

0

5

10

15

20

25

30

35

40

45

Whe

at

Pri

ce d

ecli

ne (

%)

Denmark

Belgium

Rye Barley

Oats

Figure 2. Real grain price declines: Belgium and Denmark, 1870–4 to 1909–1913Sources: Belgium: product prices for 1877–1913 from Blomme, Economic development, pp. 418–9; cost of living from Scholliers,‘Century’, pp. 111–3; Denmark: 1870–1913 from O’Rourke, ‘Grain invasion’, p. 785.



self-sufficient farmers and the rise in the number of day labourers.27 In the 1880s,pull factors that would have attracted labourers into industry remained weak.‘[U]rbanization of the countryside’, one authority has written, ‘was more typical ofindustrialization than the exponential growth of existing cities’.28 Internal migra-tion to urban centres had slowed down after 1850 and, while Belgians had a longtradition of short-distance relocation to France, emigration to the New World wascomparatively rare.29 The result was ‘a substantial degree of hidden unemployment[in the countryside] which, certainly in the middle of the 19th century, must haveassumed grotesque proportions in certain regions’.30

The contrast with Denmark, whose transformation is cited by O’Rourke andWilliamson as the classic open-economy response to the grain invasion, could nothave been more stark.31 While the volume of imports from the New Worldimpacted on domestic grain prices and production in the same order of magnitudein the two countries (figure 2), the response was hardly similar.The key variable inthe O’Rourke and Williamson model of late-nineteenth-century globalizationis the wage–rental ratio. In both countries, land rents barely changed duringthe period of the grain invasion. In Belgium, overcrowding in the countrysidesustained high rents, but in Denmark, the unchanging price of land reflected the

27 Blomme, Economic development, pp. 223–4.28 Van den Eeckhout, ‘Belgium’, p. 190. The percentage of the population living in towns with over 25,000

inhabitants increased by about 50% from 1846 to 1880, but by only 35% from 1880 to 1910 (calculated fromFlora, State, economy and society, p. 253). Belgium had a comparatively higher proportion of its population insmaller centres. In 1900, 25.4% of its population lived in cities of 5,000 to 20,000; the equivalent figure forFrance is 11.3%; and for Germany, 13.5% (Mahaim, Abonnements d’ouvriers, p. 149). More generally, see deBrabander, Regional specialization, pp. 46–95.

29 Even cross-border movement receded after 1890, the implications of which are discussed in the next section.See Lentacker, Frontière franco-belge, pp. 266–75.

30 Blomme, Economic development, p. 223.31 O’Rourke and Williamson, Globalization, pp. 60–2; O’Rourke, ‘Grain invasion’, pp. 782–5.

Table 2. Employment distribution by industry, 1846–1910

Branch

Employment (‘000)(% of total / % of total excluding services)

1846 1896 1910

Agriculture 778.7 (0.55/0.63) 773.0 (0.31/0.45) 550.9 (0.22/0.32)Mining and quarrying 60.6 (0.04/0.05) 155.0 (0.06/0.09) 178.0 (0.07/0.10)Food, beverages, and tobacco 52.2 (0.04/0.04) 104.1 (0.04/0.06) 118.3 (0.05/0.07)Textiles 94.5 (0.07/0.08) 96.0 (0.04/0.06) 139.3 (0.06/0.08)Clothing and footwear 86.8 (0.06/0.07) 157.1 (0.06/0.09) 184.7 (0.07/0.11)Wood and furniture 48.8 (0.03/0.04) 89.8 (0.04/0.05) 106.9 (0.04/0.06)Paper, printing, and publishing 6.2 (0.00/0.01) 23.8 (0.01/0.01) 29.7 (0.01/0.02)Chemicals 3.8 (0.00/0.00) 22.3 (0.01/0.01) 40.3 (0.02/0.02)Bricks, pottery, cement, and glass 20.0 (0.01/0.02) 39.9 (0.02/0.02) 63.7 (0.03/0.04)Iron, steel, and non-ferrous 14.2 (0.01/0.01) 56.2 (0.02/0.03) 75.0 (0.03/0.04)Engineering 34.9 (0.02/0.03) 69.2 (0.03/0.04) 104.7 (0.04/0.06)Other manufacturing 8.7 (0.01/0.01) 26.1 (0.01/0.02) 31.7 (0.01/0.02)Construction 21.0 (0.01/0.02) 93.9 (0.04/0.06) 106.0 (0.04/0.06)Service 185.7 (0.13) 751.4 (0.31) 783.5 (0.31)

Total excluding services 1,230.5 1,706.2 1,729.0Total 1,416.2 2,457.6 2,512.5

Source: De Brabander, Regional specialization, p. 82.



profitability of the cooperative movement, a strategic player in the upsurge inlivestock production that was already apparent before the grain invasion. Theseorganizations were weak in Belgium, whose dairy exports lagged behind those ofDenmark.32 Moreover, relative to Belgium, Danish migration overseas was sub-stantial.33 Table 3 summarizes the combined influences of population movementsand changes in agricultural production in Belgium and Denmark. While thewage–rental ratio hardly changed in the former, it nearly doubled in the latter.

The trend in Belgian real wages requires precise analysis. As consumers, workersin all sectors, protected and unprotected, would have benefited from the fall in theprice of bread, an important item in their diets. Urban industrial workers, whowere at least partly isolated from the pools of unemployed and immobile workersin the countryside, ought to have seen their real wages rise.This was the case in theskilled labour, capital-intensive sector, but the story in the semi- and unskilledsectors was different. With the exception of mining, demand for workers in thelabour-intensive manufacturing sector, like textiles, was not forthcoming. Thedomestic market may have been too small or too poor and, in any case, commercialpolicy cut off key inputs like yarn.34 Cotton textiles were not competitive at worldprices and the industry stagnated until the mid-1890s. As a result, the growth ofaverage industrial wages was held back, certainly compared to other open econo-mies of the period. According to Williamson’s wage indexes for unskilled daylabourers, between 1870 and 1913 wages rose by 2.63 per cent per annum inDenmark; 2.73 per cent in Sweden; and a modest 0.92 per cent in Belgium.35

Belgium’s unbalanced economic development was manifested in the divergentattitudes of its leading social actors towards free trade. Despite the growth ininternational exposure far exceeding that of other small countries, its ruralpopulation—landlords and tenants—carried an important economic and politicalweight similar to that of France and Germany.36 Conservative and Catholic parties

32 In 1866, Denmark and Belgium exported the same quantity of butter to the UK. By 1894, Danish exportshad increased more then tenfold, while those of Belgium had stagnated. Suetens, Politique commerciale, p. 134.

33 Hatton and Williamson, Age of migration, p. 10, reported that Belgian emigration in its peak decade(1881–90) was only 25% that of Denmark. Stengers, Émigration, pp. 60–1.

34 The price of cotton yarn began to fall and output per person began to rise in the 1890s, about the time oftariff reform (van Houtte, Industrie textile, pp. 231, 269). For contemporary statements on the internationalcompetitiveness of the Belgian textile industry, see Belgium, Annales parlementaires, 1894–5, p. 1531. Again, it wasonly after the tariff reform of 1895 that Belgian cotton textile exports rose in line with trade theory. Between 1900and 1913, exports to Argentina increased fourfold, and those to Brazil and the UK threefold. Van Houtte,Industrie textile, p. 173.

35 O’Rourke and Williamson, Globalization, p. 19; Williamson, ‘Evolution of global labor’, pp. 178–80.36 Van Molle, Katholieken en Landbouw, pp. 361–8.

Table 3. Wage–rental ratios: Belgium and Denmark

Year

Belgium Denmark

Land rents Wages Wage–rental Wage–rental

1880 100 100 100 1001885 81 121 149 1751910 101 122 121 223

Sources: Belgium: Blomme, Economic development, p. 171, 398–9; Denmark:O’Rourke, ‘Grain invasion’, p. 787.



had a stronghold in the countryside. Lacking comparative advantage in agricul-tural production, rural interests were favourable to tariff protection. Urban andliberal Belgium was less than unanimous about the merits of free trade. Antwerp,whose fortunes were based on its external commerce, had a strong predispositionto laissez-faire, but Ghent cotton-textile manufacturers demanded protection fromforeign competition.37 The left was also divided on the merits of openness.Although the vision of a socialist international was honoured and the fall in grainprices was not lost on contemporaries, the left was sympathetic to rural demandsand the protectionist arguments of selected groups of workers. Belgian socialistswere more inclined than their German comrades to uphold rural values and tosupport tenants and small landholders who wished to remain on the land.38

Certain groups of urban, skilled workers may have had a strong interest inmaintaining the status quo. The labour aristocracy benefited from cheap, foreigngrain and certainly did not object to the fact that hordes of cheap labourersremained confined to the countryside, not least because it made city housingaccessible. Workers in import-competing sectors, such as weaving, campaignedactively with their employers for greater protection. As a result of these competinginterests, the left was split on the benefits of free trade and many of these fissureswere present in the POB even after it proclaimed its support for free trade in 1885.

The divergence in attitudes among social actors towards commercial policy doesnot sit well with Stolper-Samuelson.39 The basic model depends critically on theassumption that factors of production, while immobile internationally, are per-fectly mobile within the domestic economy. Factor mobility assures that tradeaffects labour, capital, and natural resources in the same way, no matter where theyare employed in the economy. The implication is that owners of the same factorshare the same preferences with respect to trade policy. However, in the 1880s,Belgian labour was hardly mobile from agriculture to industry, or within industry.The country seems to have been closer to the fixed-factor world of Ricardo-Viner,in which the gains of openness are not distributed evenly to the abundant factor.40

In these models, where factors are immobile, attitudes to trade policy will cleavealong industry as opposed to class lines, exactly as they did in Belgium until 1885.

These models, because they assume either identical preferences or a represen-tative worker, need be interpreted as rough guides to identifying attitudes to trade.As there was a mix of capital and uses for land, there was range of worker types,from mobile to immobile, and from skilled to unskilled. Moreover, labour was splitalong deep-seated ‘fault’ lines of language and religion, alongside trade and eco-nomic issues.41 The point here is not to dismiss these distinctions and give exclu-sive weight to an ideal worker, but to situate the move to free trade by organizedlabour as part of a political process that accommodated the multiple interests,economic and non-economic, of its supporters.

37 Scholliers, ‘Industrie cotonnière’, 134–5.38 On the agrarian question and European socialists, see Gerschenkron, Bread, pp. 33–6; Mitrany, Marx against

the peasant, pp. 7–41. For Vandervelde, see his ‘Lois sociales’; Influence des villes; and Exode rural.39 On trade models, see Hiscox, International trade, pp. 7–10.40 The standard text is Magee, ‘Three simple tests’. See Hiscox, International trade, pp. 12–34, and idem,

‘Interindustry factor mobility’, for a recent discussion of these models and their applications.41 The ‘fault line model’ is from Witte, Craeybeckx, and Meynen, Political history, pp. 12–13. I return to this

model below.



II

Some time before the turn of the century, workers tended to forsake sectoralinterests and began to share common ground on economic issues. The turningpoint, Belgian historians assert, was the extension of the rail network and thewidespread use of the subsidized transport that allowed workers to commutenon-trivial distances on a daily basis.42 The standard account is as follows. In 1870,the state began to buy up private lines, leading to the establishment in 1884 of theSociété nationale des chemins de fer vincinaux (SNCV). Public ownership assuredthat control of the rail network would remain in Belgian hands. Initially it wasgroups of employers, especially in the iron and steel and mining industries in thesouth, that demanded state intervention to assure a plentiful and steady supply oflabour.They received backing from Catholic groups and even liberals who believedthat the rail system would stem the rise in urban support for socialism in thecities.43 A dense, secondary train network linking rural villages with larger conur-bations was built rapidly over a small land mass. Workers could stay in thecountryside but have jobs in the city, thereby breaking down regional and sectoraldivisions.

Provoked by employers’ demands and support from the main parties, the stateactively encouraged train travel. In the 1870s, the state put in place ticket plansbased on distance that subsidized workers’ travel.44 In 1885, about 4 per cent of thelabour force commuted on a daily basis to their place of work; in 1895, 6 per cent;and by the eve of the FirstWorldWar, the number had jumped to 15 per cent.Thesefigures are an underestimate because they do not include the number of workerswho travelled less than six times per week and they do not adjust for the fact that thevast majority of the commuters were male.The average commuting distance (in onedirection) also increased, from 12 kilometres in 1870 to 19 kilometres in 1910.45

The greatest number of workers bought tickets for travel within the fast growingindustrial region of Brabant and to its immediate south, although workers camefrom as far as western Flanders and Limbourg. Luxembourg was an importantstarting point for travel to Liège and Namur.

Push and pull factors contributed to the rapid rise in the number of passengersbefore the turn of the century. On the demand side, annual growth of industrialproduction after 1895 and until the FirstWorldWar was almost three times greaterthan in the 20 preceding years.46 On the supply side, cross-border movements ofworkers had fallen off as Belgians met resistance from French unions.47 This earlierflow had repercussions for the success of the train network and workers’ identifi-cation with it. Workers had previous experience of commuting, and family andsocial institutions had adapted to it. On another level, the network was resolutely

42 Cassiers, ‘Rôle de l’État’; idem, ‘Fonction importante’; Polasky, Democratic socialism, pp. 18–22; idem,‘Transplanting workers’.

43 Van den Eeckhout, ‘Belgium’, p. 190.44 With the discount, the average worker spent approximately 5% of his weekly salary on transportation. All

statistics in this paragraph are from Mahaim, Abonnements d’ouvriers. For a treatment of this source, see Pasleau,‘Migrations’.

45 Commuting distances varied by region: 25.6% of commuters travelled 16–25km to Brussels; the corre-sponding figure for Liège was 20.0%; for Ghent, 6.2%.

46 Vandermotten, ‘Tendances longues’, p. 267.47 Lentacker, Frontière franco-belge, pp. 266–75; cf. Strikwerda, ‘France and Belgian immigration’, pp. 127–9.



national. Turned back by the French, even as certain socialist leaders spokepassionately about internationalism, workers came to see that their welfare wastied more closely to national policy.

Figure 3 gives the breakdown of ticket holders by industry for 1910. Workersmoved to industry and not vice versa.48 Skilled workers in engineering and met-alwork commuted into Antwerp and Brabant, but their numbers were compara-tively small.49 Many workers in these sectors were already established in urbancentres.The train network had a greater impact on the livelihoods of unskilled andsemi-skilled workers. The network encouraged the match between the mass ofunderemployed that lived in the countryside and in smaller towns in the west andthe extreme south, and the demand for factory workers in the Brabant region andelsewhere in the north.

Belgian labour and social historians have long claimed that the train systembroke down divisions within the working class.The local buyers’ market for labourin the countryside had contracted by the early 1890s and a new national labourmarket took root, in which ‘commuting workers regulated supply and demand’.50

Socialists, who initially were hesitant in their support for the expandednetwork—they preferred that workers congregate in urban centres for electoralreasons—came to see that the train system was instrumental to the fortunes of thePOB, whose formation coincided with the extension of the network. Although thenetwork did not put an end to the various divisions among its supporters, it didresolve the agrarian question. More optimistically, Vandervelde, the leader of thenew POB, saw its evolving role in the making of the Belgian working class. The

48 I discuss regional specialization in the next section.49 Pasleau, ‘Migrations’, pp. 180–1.50 The phrase is from Mahaim, Abonnements d’ouvriers, p. 175: ‘Les abonnnements d’ouvriers sont devenus le

régulateur de l’offre et de la demande’.

0

50000

100000

150000

200000

250000

300000

350000

Min

ing

Facto

ry

Constr

uctio

n

Daywor

kers

Admin

istra

tion

Other

Num

ber 6-7 tickets/wk

1 ticket/wk

10.0%

30.1%

14.9% 15.6%

10.4%

19.1%

Figure 3. Occupations of ticket holders, 1906Source: Mahaim, Abonnements d’ouvriers, p. 103.



origins of the rail system, he wrote in 1888, could be attributed to employers’demands, but the scheme had unintended consequences:

They [employers] had hoped to severely weaken unions, to flood the market with cheaplabour, and to thereby lower workers’ ‘standard of life’. But, contrary to expectations,the opposite occurred. Flemish coalmen—who travel as long as two hours a day to workin the mines of the Borinage and in central Belgium—and masons, carpenters, andplasterers who come to Brussels every morning, are in daily contact with other groupsof workers.They have acquired the same needs, make the same demands, and they arguethe same opinions. Every night, we hear socialist songs in the trains bringing them backhome, and the electoral progress of our Party can be attributed to them.51

Overall, the train network accelerated the structural change of the Belgianeconomy in the context of globalization that elsewhere in northern Europe wasachieved by emigration.The new transport system is posited to have had the effectof transforming an immobile labour force into a mobile one, the crucial assump-tion underlying Stolper-Samuelson. The question remains as to how the newlyformed POB was able to channel this transformation in the labour market to itsadvantage, and what this meant for commercial and social policy.

III

In this section, exploiting new wage data, I test Vandervelde’s proposition aboutthe link between the rail system and class formation. The logic is straightforward:a class consensus on economic issues, represented here by worker support for freetrade, is more likely where factors are mobile. I use rate of return differentials asan indicator of factor movement.Where mobility is high, wage differentials tend tobe arbitraged away and all workers share in the benefits of openness (Stolper-Samuelson); where mobility is low, wage dispersion persists and workers’ welfaredepends on their sector of employment (Ricardo-Viner).

My source is the international data set on wages and hours collected by the USDepartment of Labor in 1900 in its Fifteenth annual report, under the supervision ofCarroll Wright.52 The Report was restricted to official government sources only,including reports on foreign countries (in fact, about 65 per cent of observations forBelgium actually came from the latter sources), mainly US and British consularreports.53 For all countries, the Report recorded wages (dollars per day) by occupa-tion, of which there were nearly 4,500 types (for example, the average wage of a

51 This is my own translation. Vandervelde and Destrée, Socialisme, p. 308. I have located one of these songs(Belgium, Annales parlementaires, 1894–5, p. 1576):

Braves fermiers de modeste culture,Vous vous plaignez tout haut et justement;Mais l’argent pris à notre nourriture,On pourrait bien vous le rendre autrement;Vous payez tout sans mesure et sans trêve;Le sol, l’outil; l’habit, le mobilier;Demandez donc plutôt qu’on vous dégrèveSans renchérir le pain d’ouvrier.

52 Huberman, ‘Working hours’, pp. 968–9, reviews the biases contained in the Report.53 The Report listed these Belgian sources: Ministère de l’Intérieur et de l’Instruction Publique, Annuaire

statistique, 1887, 1898; Office du Travail, Revue du travail, 4 vols., 1896–9; Commission du Travail, Rapport,3 vols., 1887; Ministère de l’Agriculture, de l’Industrie et des Travaux Publiques, Salaries et budgets ouvriers enBelgique au mois d’avril 1891; Ville de Bruxelles, Minimum de salaire, Enquête du mai 1896.



cotton spinner in a selected firm in 1872). I have grouped the Belgian observationsinto 16 major industries.54 This breakdown is similar to that used by Hiscox, whostudied inter-industry mobility for the same period for a sample of countries,including the United States and France.55 Table 4 gives the descriptive statistics ofthe sample by industry. Although not exhaustive, the data certainly cover workersin all regions and industries who exploited subsidized train travel (figure 3).Conforming to Belgium’s industrial structure, observations are concentratedin mining (16.7 per cent), iron and steel (14.5 per cent), and construction(10.8 per cent).Women comprised 10 per cent of the sample; female labour forceparticipation, which was declining over the period, was concentrated in cotton(33.0 per cent), manufacturing (14.8 per cent), and services (9.1 per cent). TheReport also recorded hours of work for slightly more than 25 per cent of the sample.As for periodization, the first observation is for 1830, but the bulk of the data isrestricted to the period 1870–1900. The last three columns of table 4 give thebreakdown of the sample for these decades. With the exception of cotton andwoollen industries, the observations are generally spread out evenly over theseyears.

After adjusting for cost of living changes, I have assembled the data by industryin five-year intervals beginning in 1870. In other work, I used regression analysisto correct for the unbalanced nature of the sample (both over time and acrossindustries), but for this paper there were a sufficient number of data points to allow

54 Huberman, ‘Working hours’, p. 970, used a smaller sample (N = 172) of observations for Belgium. In thispaper, I use all Belgian observations contained in the Report.

55 See Hiscox, International trade, pp. 12–34; idem, ‘Interindustry factor mobility’.

Table 4. Wage observations by industry ($ per day)

Industry ObservationsMean (median) wage

($ per day) s.d.

Percentage of observations by decade

1870–9 1880–9 1890–9

Mines 438 0.57 (0.55) 0.22 0.24 0.32 0.18Iron and steel 378 0.71 (0.67) 0.24 0.22 0.52 0.18Construction 284 0.59 (0.58) 0.19 0.24 0.35 0.36Machinery 220 0.66 (0.65) 0.17 0.29 0.42 0.26Cotton 219 0.60 (0.53) 0.39 0.09 0.53 0.21Clothing 152 0.64 (0.64) 0.28 0.13 0.34 0.38Services 149 0.62 (0.62) 0.29 0.13 0.50 0.35Furniture 146 0.66 (0.70) 0.20 0.25 0.31 0.35Glass 127 0.99 (0.85) 0.68 0.18 0.28 0.50Brick 97 0.73 (0.73) 0.24 0.16 0.26 0.55Wool 80 0.60 (0.58) 0.24 0.06 0.51 0.26Stove 78 0.69 (0.66) 0.21 0.28 0.33 0.37Paper 77 0.69 (0.69) 0.28 0.21 0.29 0.47Rail 74 0.75 (0.57) 0.53 0.20 0.47 0.31Farm labour 49 0.28 (0.24) 0.18 0.18 0.27 0.29Food 42 0.74 (0.75) 0.29 0.14 0.38 0.45Men 2,368 0.68 (0.65) 0.31 0.22 0.39 0.29Women 242 0.36 (0.33) 0.18 0.06 0.41 0.39(w/h)*100 691 1.04 (0.97) 0.55 0.01 0.52 0.43All 2,610 0.65 (0.63) 0.32 0.20 0.39 0.30

Source: US Department of Labor, Report.



use of the more simple averaging procedure.56 Figure 4 provides a check on myprocedures.The top panel compares my series for mining with that of Puissant forthe Borinage region. Puissant’s series is generally lower because it is restricted topit workers, while my series includes workers above and below ground.That said,the two series move together. The second panel compares my figures for maleworkers in cotton textiles with series produced by Scholliers for a single firmlocated in Ghent, one for male spinners and the other for all workers, male andfemale. From the outset, Scholliers’s series for all workers seems to track wagesfrom the Report, and after 1885, his series for male spinners and my estimatescoincide as well.

In the last panel of figure 4, I have calculated a series of average wages for non-agricultural workers.57 Employers’ demand for an abundant and cheap laboursupply was satisfied. Over the period, the annual growth rate of wages was 0.5 percent; Scholliers, who constructed an average real wage series for seven industries,calculated a growth rate of 0.8 per cent per annum. Panel c also traces the fates ofworkers in two sectors. Until 1880, farm labourers saw their wages contract; ironand steel workers, many of whom were skilled, were more fortunate.This is exactlythe result anticipated in a world of immobile labour. After 1880, farm labourers’wages began to converge to the average wage, a finding that is consistent withincreased mobility.

Figure 5 presents a more systematic analysis of trends in inter-industry mobility,using information from all occupations. I use the coefficient of variation (c.v.) ofreal wages to test for increased labour mobility. Using all observations and withoutany controls, the c.v. fell by slightly more than 20 per cent from 1850 to 1900, andby about 35 per cent from peak to trough.58 The long-run trend repeats whenobservations are grouped into occupations, but note that between 1875 and 1880,wage dispersion actually widens, a result that is consistent with Ricardo-Viner.After 1880, Stolper-Samuelson took hold. The c.v. fell because the expansion ofsubsidized train transport narrowed wage differentials.

Undoubtedly, occupational grouping affords only a basic control of the data.Wage convergence may have been the result of changes in the geographicaldistribution and composition of industry, without any change in the mobility ofworkers. However, the geographical makeup of Belgian industry was stable in theperiod.59 Consider also the likelihood that wage narrowing was the outcome ofchanges in the wage premium paid to skilled versus unskilled workers.60 I havecontrolled partially for heterogeneity by separating adult male workers fromwomen and adolescents, using the occupational breakdown. As expected, the wage

56 More precisely, I use a five-year moving average centred on 1870, 1875, etc. I have experimented withdifferent specifications following the technique used in ‘Working hours’, but regression analysis performs poorly.All wages in figs. 4–6 have been deflated by Scholliers’s ‘Century of wages’ price index (1852–4 = 100). Othersources cited in this paragraph are Puissant, Évolution, pp. 630–1; Scholliers, Wages, manufacturers and workers,pp. 227–8.

57 The non-agricultural wage series in fig. 4, panel 3, is weighted using employment figures from the censusesof 1846 and 1896. Source: De Brabander, Regional specialization, p. 82.

58 Scholliers, ‘Industrial wage differentials’, exploiting census wage data, estimated a c.v. of 0.31 in 1846 and0.18 in 1896, about the same order of magnitude that I report for the two closest years in fig. 5. The averagedecline in wage dispersion for Belgium is similar to that reported by Hiscox, International trade, for the UnitedStates and Britain. For France, the c.v. was relatively flat. I discuss the French case below.

59 De Brabander, Regional specialization, pp. 67–9; Vandermotten, Barsten in België.60 Krueger and Summers, ‘Reflections’.



0.40

0.45

0.50

0.55

0.60

0.65

0.70

0.75

0.80A

B

C

1870

1875

1880

1885

1890

1895

1900

Puissant

US Report

0.35

0.40

0.45

0.50

0.55

0.60

0.65

0.70

0.75

1870

1875

1880

1885

1890

1895

1900

Ghent – male spinners

Ghent – all workers

US Report

0.00

0.10

0.20

0.30

0.40

0.50

0.60

0.70

0.80

1870

1875

1880

1885

1890

1895

1900

Non-agriculture

Iron and steel

Agriculture

Figure 4 a. Wages in mining; b. Wages in cotton manufacture; c. Wages in industryand agricultureNotes and sources: All figures are in dollars per day. Coal miners from Puissant, Évolution, pp. 630–1; cotton spinners fromScholliers, Wages, manufacturers and workers, pp. 227–8; all other series from US Department of Labor, Report. The non-agricultural wage series in panel 4c is weighted using employment figures from the censuses of 1846 and 1896; de Brabander,Regional specialization, p. 82. All series were constructed using five-year averages. Cost of living index: Scholliers, ‘Century’,pp. 111–3, 1852–4 = 100.



dispersion for men is smaller than for all workers, but again the direction of changeindicates increased factor mobility, especially in the 1880s (figure 5). These wageseries have not been corrected for hours of work and it may well be the case thatwages adjusted for changes in the relative attractiveness of working in differentindustries over time. If this is the case, then the observed shifts in wage differentialscould represent changes in compensating differentials and not changes in the sizeof the labour market. I control for different conditions of work using the subset ofobservations that also report work hours. The results reported in the figure 5 aremixed. The c.v. of wages per hour did increase in the period until 1880, but thenit declined rapidly. It would appear that conditions of work might have mattered inthe early years, an unsurprising result because of the absence of factory legislation.That said, the downward trend in adjusted wages with the advent of the trainsystem in the 1880s is unmistakable and is in line with the other series.

Figure 6 examines the question of skill in more detail. It may well be the casethat the narrowing of wage dispersion was the result of changes in skill levels in onedirection or another, regardless of changes in mobility. I examine this possibility bycomparing the 10th and 90th percentiles of the wage distributions of three indus-tries that, it is claimed, experienced technical and organizational changes before1900.61 The deskilling of high-wage workers or the accumulation of new skills bylow-paid workers would have led to a rise in the pay ratio.The ratios fluctuate overthe period, especially in mining, but any trend in skill within industries was minor,if not insignificant.62

61 Scholliers, ‘Industrial wage differentials’, p. 107.62 Industrial specialization was pretty much stable in the last half of the century, despite some increased

concentration in engineering and textiles. De Brabander, Regional specialization, pp. 67–9; Vandermotten, Barstenin België, pp. 280–1.

0,00

0,05

0,10

0,15

0,20

0,25

0,30

0,35

0,40

0,45

0,50

1850

Coe

ffic

ient

of

vari

atio

n

All

Occupation

Men

Wages per hour

1860

1870

1875

1880

1885

1890

1895

1900

Figure 5. Interindustry variation in wagesSource: US Department of Labor, Report.



Vandervelde was right. The train system permitted employers to draw on anabundant supply of cheap labour in the countryside and keep wages low. However,increased mobility had unanticipated consequences. By the late-nineteenthcentury, labour had coalesced into a homogeneous entity, at least from an eco-nomic point of view. Subsidized train transport was indeed the smoking gun in theformation of the Belgian working class. There is a flip side to this argument. Thecreation of integrated ‘capital’ went along with the formation of unified‘labour’—and for the same reasons.63 Evidence on sectoral investments and port-folio de-diversification is consistent with this.

In the absence of direct evidence or survey information of individual prefer-ences, the wage data can be used with caution to infer workers’ attitudes towardtrade policy.64 It does not follow, of course, that because workers socialized witheach other, they had put aside their religious, linguistic, or regional identities, orthat trade issues were their primary concern. The expansion of the rail networkwould have had a similar salutary effect on membership in Catholic unions thatwere tied to the Conservatives. But from the left’s viewpoint, that is where thegrassroots was located. It was the POB’s challenge to find the mix of commercial

63 The proposition here is that ‘capital’ would have concentrated their investments in a single sector of activity,because if capital were mobile, the convergence of profit differentials would have made them indifferent betweendifferent sectors of activity. There is evidence of this for the mining sector (see Wauthelet, ‘Accumulation’). Inprinciple, when capital is less mobile, we should expect more equity financing as opposed to borrowing, becauselenders are more reluctant to invest in specific assets and they charge premia for the added risks. Also, whencapital is less mobile, returns vary between sectors and investors have a greater incentive to diversify theirportfolios (Hiscox, ‘Interindustry factor mobility’, p. 403). Profit rates for a sample of firms are not available totest this proposition, but historians of finance (Duriviaux, Banque mixte) report that Belgian banks acted as themain lenders in capital markets, while the equity market was of secondary importance.

64 O’Rourke and Sinnott, ‘Determinants of attitudes’, use survey evidence of individual preferences.

0,00

0,10

0,20

0,30

0,40

0,50

0,60

0,70

1870

10-9

0 w

age

rati

o

Mines Cotton Iron and steel

1875

1880

1885

1890

1895

1900

Figure 6. Wage distributions: selected industries (10–90 centiles)Source: US Department of Labor, Report.



and social policy that accommodated differences among its supporters and that atthe same time provided higher standards of living to all.

IV

Drawing on examples from Old and New Worlds, Engerman and Sokoloff foundthat polities with labour scarcity and greater income equality generally fostered abroadening of the franchise.65The Belgian case sits well with their model. For mostof the century, local elites were hard-pressed to extend voting privileges, and it wasonly after the social upheavals of 1886 that the ruling Conservative Party wascompelled to lay the groundwork for a larger electorate.66 Nevertheless, whenuniversal male suffrage was introduced in 1893, it granted multiple votes to certainproperty holders.The new electoral system actually increased the representation ofrural landlords.67 Compared to France and Germany, Belgium’s experience ispuzzling because polities with more international exposure have been found tohave greater democratic participation.68 The underlying idea here is that in lessdeveloped and labour-abundant countries, openness reduces income gaps betweenworkers and elites. Along these lines, Acemoglu and Robinson asserted that eliteswould have the incentive to extend the franchise, as voters who saw their wagesrising would be less likely to support redistributive policies because they fearedrising tax burdens.69 The problem with this argument is that economic forcesthemselves did not ensure reduced income inequality, and workers were compelledto turn to their political organizations to seek higher wages and better workingconditions.

Globalization before 1914 generated only modest wage gains in Belgium, evenin the period of strong inter-industry factor mobility. Productivity change from1870 to 1900 was twice as great as the increase in the average wage.70 Workers didnot and could not capture an increasing share of the surplus. Despite mobilizingworkers in a series of national strikes, unions were less successful in negotiating

65 Engerman and Sokoloff, ‘Evolution of suffrage’.66 Boix, Democracy, pp. 12–13, offers an alternative explanation. In his model, the move to democracy is

associated with the rise in capital mobility that allowed elites to protect their wealth from redistributive policy. Iwill return to the role of capital flight in the next section.

67 Men over 34 years of age with a family and living in a ratable house were given a second vote, while thosewith some property or professional qualifications obtained yet another. On the development of le suffrage universelplural, see Kossmann, Low Countries, p. 373. On rural voting patterns, see van Molle, Katholieken en Landbouw,p. 362.

68 J. E. Lopez-Cordova and C. M. Meissner, ‘Globalization and democracy, 1870–2000’, unpublished manu-script, University of Cambridge (2004); K. H. O’Rourke and A. M. Taylor, ‘Democracy and protectionism’,NBER working paper no. 12250 (2006).

69 Acemoglu and Robinson, Economic origins, p. 41. The Belgian experience is closer to an earlier model ofAcemoglu and Robinson, ‘Democracy, inequality, and growth’, pp. 40–1, in which elites grant electoral reform toprevent widespread social unrest and revolution.

70 Using new information on worktimes, Huberman, ‘Working hours’, p. 984, calculated that output perworking hour in Belgium grew by 1.73% per annum between 1870 and 1913. Vandermotten, ‘Tendenceslongues’, p. 283, estimated productivity gains of 2.6% in the period 1840–86; 1.2% in 1880–96; and 1.5% in1896–1910. Sectoral evidence confirms that productivity gains outstripped wage increases. In cotton spinning,productivity increased by 3.2% p.a. from 1850 to 1910; van Houtte, Industrie textile, p. 230. In mining, theincrease of output per worker was about 1% p.a. from 1860 to 1900 in the Borinage; Puissant, Évolution, p. 630.



wage increases and union density remained low.71 Employers did not hesitate toexploit divisions along religious and linguistic fault-lines.72 Ever the keen socialobserver, Vandervelde wrote that even in the heartland of industrial conflict, themining regions, workers had become distrustful of collective action.73

The POB was not automatically guaranteed of workers’ allegiance because ofcompeting loyalties. Electoral politics was competitive. Figure 7 traces the evolu-tion of support for the POB in the key elections before and after the introductionof universal suffrage in 1893. Like the other major parties, the POB vote wascomposed of various elements. It had to compete with Catholic parties that hadestablished strong ties to workers, and, initially, support for the POB was limitedto the mining and iron and steel districts in the south and in Brabant. By the early1900s, however, the party had made inroads into rural and conservative Flanders.Altogether, the POB found support among a quarter of eligible voters by 1914.The problem faced by the POB was therefore twofold. It had to look beyond itsnatural base of support to get its agenda acted upon in Parliament, and at the sametime its programme had to satisfy the diversity of its membership.

In 1894, during the first Parliament session after the extension of the franchise,the Conservatives introduced a bill intended to make comprehensive changes to

71 The Netherlands had a smaller population than Belgium before 1914, but nearly double the number ofstrikes per year between 1896 and 1913; Flora, State, economy and society, pp. 696, 753. Belgian unions lost morethan 60% of their strikes between 1896 and 1913; Neuville, Naissance, p. 217.

72 Catholic and socialist unions had approximately equal membership in 1914; Leboutte, Puissant, and Scuto,Siècle d’histoire, p. 94. On political and ethnic divisions at the local level, see Strikwerda, House divided, pp. 181–96,261.

73 ‘Il faut dire, les Borains [residents of the Borinage mining district] ont été souvent victims de leur bonne foi: il s’esttrouvé des individus qui, ayant la confiance de leurs camarades, en ont profité pour les voler’ (Vandervelde, cited inPuissant, Évolution, p. 325).

Election results, 1892-1914

51.141.4

48.543.9 42.8

45.4

28

18.9

24.325.6 24.5

21

22.5 26 30.3

53.6

13.2

0%

10%

20%

30%

40%

50%

60%

70%

80%

90%

100%

1892

other

Workers

Liberal

Catholic

1894

1898

1900

1904

1914

Figure 7. Election results, 1892–1914Source: Mackie and Rose, International almanac, pp. 39–45.



Belgium’s tariff structure. Faced by the renewed protectionism of its two maintrading partners, France and Germany, the Belgian government was under strongpressure from employers in key sectors and from landlords to retaliate. TheBelgian government was, no doubt, in a weaker position than its continentalneighbours to push for protectionism because the various pressure groups behindit were probably less well concentrated and organized. They were split acrossreligious and linguistic affiliations and the collective action problem constrainedtheir influence, as it did the free traders. In this context, government leaders putforward the argument that tariff changes were necessary because of declining staterevenues. The government proposed to harmonize the tariff system, which meantrates of around 5 to 10 per cent across a broad range of manufacturing andagricultural products.

In an earlier round of tariff renegotiations before universal suffrage, in 1884,Conservatives and certain Liberals had claimed to speak on behalf of workers,invoking their support for protectionist measures in cotton spinning and otherimport-sensitive sectors.74 And at the outset of the 1894/5 tariff debate, citing theexamples of large continental neighbours, the government anticipated that extend-ing the voting list would lead to greater support for protection.75 However, con-trary to the government’s calculation, more democratic representation was notnecessarily antithetical to increased support for free trade among voters. Althoughthe parliamentary exchanges did not reveal clear-cut cleavages along class lines, asRogowski and others using standard trade theory would have predicted, by the endof the debate the parties had moved in this direction. The Liberal Party wascomposed of ardent defenders of laissez-faire whose views dominated the protec-tionist sentiments of certain employer groups among its partisans.The Conserva-tive Party saw itself as the spokesperson of protectionist landlords and rural values,at the expense of some urban support for lower tariffs. Finally, the POB could nottake for granted its constituents support for free trade, as it had to reckon withpockets of workers who retained strong regional attachments, skilled workers inthe capital-intensive industries who may have wanted to separate themselves fromthe mass of unskilled labourers, and still others who may have been tempted by theappeal of Catholic unions and were less inclined to speak out against the govern-ment’s planned changes.76

Certainly, the leadership of the POB may have had a different outlook oninternationalism than the rank and file.Vandervelde was a member of the executiveof the Second International, of which he was president from 1900, and he andothers participated actively and hosted many of the meetings organized by Euro-pean social reformers on working conditions. These conferences were an impor-tant venue for the exchange and dissemination of information. That said, if thepamphlets published by the POB and distributed to their supporters are any

74 Belgium, Annales parlementaires, 1884–5, p. 1490.75 The debate was prolonged, extending from March to June 1895 and covering more than 600 pages in

Belgium, Annales parlementaires, 1894–5.76 The fault-line model of Witte, Craeybeckx, and Meynen, Political history, p. 13, is consistent with the

emergence of a cleavage along class lines in the period before 1914: ‘[R]ecent Belgian political history isdominated by three intertwining problems—the socio-economic, the religious philosophical and the languagedispute . . . The labor vs. capital conflict came to a head around the mid-nineteenth century’.



indication, the concept of a socialist international was not discussed extensively inthe parliamentary debates, a reasonable strategy given the closing of the Frenchborder to Belgian migrant workers.

The POB’s arguments in support of free trade sought to balance the competinginterests of its constituents. In Parliament and elsewhere, the POB sought toconvince workers of the benefits of free trade, based on their immediate needs asconsumers and producers. These arguments were well rehearsed by earlier pro-ponents in Belgium and elsewhere.The import of cheap grain from the NewWorldhad a direct effect on workers’ pocket books. Across the country, the labourmovement had established Maisons du Peuple, cooperatives whose objective was toensure that workers had a cheap and steady supply of bread, the main staple intheir diets.77 These cooperatives, which were key to mobilizing support for the left,tied workers’ fate to the state of international trade. Echoing Manchester liberals,Belgian socialists argued that low grain prices would translate into reduced wagedemands on employers.78 An open-door policy would also set competitive pricesfor meat products and, according to the respected social reformer Hector Denis,provide a supplementary source of nutrients to sustain Belgians over their longworkdays.79 Providing detailed documentation of family budgets, Denis also cal-culated the effect on workers’ clothing expenses of higher tariffs on cotton tex-tiles.80 On the production side, the POB’s representatives observed that theproposed tariff rates on key inputs like wood and raw materials would raisebuilding and housing costs;81 to remain competitive, workers in the export sectorwould face downward pressure on their wages. Finally, taking a page from basictheory, the POB saw that free trade would act as a ‘wake-up call’, forcing employ-ers in export-challenged industries to make the necessary investments in technol-ogy and organizational changes.82

While these arguments were standard fare, the novelty of the POB’s positionon openness, and the manner in which it distinguished itself from the Liberalswho extolled the virtues of the market, was the link it proposed between freetrade and social programmes.83 Globalization had accelerated the structuralchange of the Belgian economy. Grain prices fluctuated considerably and therewas no mechanism in place to smooth consumption flows.84 Workers whoremained in the countryside had traditional sources of income security to fall

77 Belgium, Annales parlementaires, 1894–5, p. 1575. On the cooperative movement, see Leboutte, Puissant, andScuto, Siècle d’histoire, pp. 91–2.

78 During the debate, all political parties referred to free trade as ‘le manchestérianisme’. It can be noted inpassing that Stella Cobden Sanderson, daughter of Richard Cobden, was Vandervelde’s sister-in-law.

79 Belgium, Annales parlementaires, 1894–5, p. 1856.80 Ibid., p. 1912.81 Ibid., p. 1571.82 Ibid., pp. 1571, 1879, 1895.83 The aim, as one parliamentary representative put it, was to: ‘. . . [S]’aventurer à concilier l’interventionnisme en

matière de travail et de salaire avec la proscription de toute protection douanière’ (Belgium, Annales parlementaires,1894–5, p. 1685). The British Liberal party made a similar argument a decade later during the great budgetdebate. See Murray, People’s Budget, pp. 49–50. On the influence of the Belgian social insurance model on the UKvariant, see Harris, ‘Poor law’, pp. 428–9.

84 Bloome’s index (Economic development, p. 415) reports a decline in grain prices in 1893–4 of 10% and thena rise of a similar proportion in 1894–5, the year of the parliamentary debate. That said, prices continued theirlong-term downward movement, until the decade before 1914. In Britain, a similar break permitted protectioniststo argue against free trade; Trentmann, ‘Fiscal reform’; idem, ‘National identity’. In Belgium, the cooperativemovement, without raising doubts about the advantages of free trade, sought intervention from local authorities.



back upon, like their home gardens, but the new jobs to which they now com-muted had created an additional set of risks and uncertainties that arose in partfrom their long hours and dangerous conditions. There were also selected groupsof skilled workers, for example in glass manufacturing, and less skilled workersin import-competing industries like spinning, that lobbied hard to maintainlevels of protection. Like any political party, the POB was sensitive to thesedemands and it favoured some modest degree of protection. However, protec-tionism was not the long-term objective. In response to the government’s plan touse tariff reform to generate revenues, the POB insisted that as a small country,Belgium’s external exposure could not be compromised. Moreover, the pro-posed tariffs would ultimately benefit landlords and selected employers at theexpense of workers, widening the already large income gaps. Belgian socialistsmaintained that tariff policy was inherently inequitable and the protectionagainst risk it provided was uneven across regions and industries, depending asit did on the lobbying power of the key actors.85 To drive home this point tosceptical workers, the POB argued that protectionism sustained the ancienrégime. Here the critique of German reformists, most notably Bernstein, of Bis-marckian commercial policy was influential.86

The POB saw the reduction in general tariffs as a first step to fiscal reform.Borrowing from the writings of the fervent free-trader Henry George, the POBproposed a more equitable system of income and property taxation alongsidereduced tariff barriers.87 Vandervelde argued in Parliament that removing tradebarriers would pressure the state to find alternative sources of revenues; under thenew system, since all segments of the population would share the burden, all wouldhave the right to benefit.88 The promise of fiscal reform was key to rallying a broadcoalition of workers behind free trade. Skilled workers may have preferred to havetheir unions administer social insurance policies, but the appeal of subsidizeduniversal programmes was greater. Recognizing that globalization generates losersas well as winners, the POB proposed that state revenues would be channelled toexpand existing social entitlements like unemployment insurance, and to fund newinitiatives such as job exchanges. Workers in sectors that faced strong foreigncompetition would rely on state programmes to smooth income over periods of jobloss and shorten unemployment spells. In agriculture, the POB sought an insur-ance scheme to stabilize incomes from dramatic and unforeseen changes inprices.89 Thus when his political opponents chided Vandervelde for declaringunabashedly that ‘we [POB] demand zero protection for all industries’—when in

85 Vandervelde’s words were direct: ‘Il est inutile de rappeler que, dans un pays comme la Belgique, qui ne produit pasla moitié de ce qui est nécessaire à sa subsistance, les droits protecteurs constituent un impôt direct, payé au profit dequelques-uns, par la grande masse des consommateurs’ (Belgium, Annales parlementaires, 1894–5, p. 879); ‘[Leprotectionnisme est] un système dans lequel le grand capitaliste de Gand, et les agrariens de Flandres se mettent d’accordpour savoir ce que les ouvriers individuels devront payer non pas en faveur des ouvriers agricoles, mais en faveur descapitalistes et des propriétaires’ (p. 1891).

86 Fletcher, Revisionism, pp. 125–44.87 Belgium, Annales parlementaires, 1894–5, p. 878. George’s classic work, Protection or free trade, was published

in French in 1888. Numerous editions of this text can be found in the Vandervelde Archives in Brussels. I thankRobert C. Allen for pointing out the influence of George on European socialists.

88 In the case of many small firms, employer resistance to redistributive tax policy was less well organized thanin Germany. On issues of risk and levels of skills and firm size, see the important study by Mares, Politics of socialrisk, pp. 12–63.

89 Belgium, Annales parlementaires, 1894–5, pp. 879–80.



fact labour had supported selected protectionist measures—he responded thathis support of free trade was conditional upon a new redistributive policy, aprogramme he referred to as ‘l’alternative’.90 For the left, openness was seen ascomplementary to redistributive policy, a line of reasoning pre-empting the argu-ments of Rodrik and others by 100 years.91

In the end, the ruling party was only partially successful in getting its tariffchanges through, and they were limited for the most part to dairy products.Livestock tariffs, like those in glass manufacturing, remained high, but their overallimpact was small. In 1913, manufacturing tariffs were half that in France andGermany.92 The coalition of anti-protectionists had won a large victory with thereduction of duties on imported cotton yarns. It was worker pressure that helpedopen up the textile sector to world markets. More fundamentally, the tariff debatehad enduring effects on the direction of social and commercial policy in Belgiumin the period until 1914, if not thereafter.93 The POB had demonstrated to theother political parties a commitment to free trade, albeit one that was tied to socialand fiscal reform. Given the POB’s electoral share, future governments would havebeen ill-advised to raise the tariff question again, and, in fact, after the 1894/5debates, protectionism seems to have lost favour even among Conservatives.Labour had turned the debate about free trade into a debate about national socialpolicy.We should bear in mind that Acemoglu and Robinson asserted that, duringperiods of globalization, elites would have an incentive to extend the franchise,because new voters whose wages were rising would not support redistributivepolicies. In Belgium, in contrast, workers saw that free trade and redistributionwent hand-in-hand.

V

So far, I have argued that putting aside the enduring fault-lines that characterizeBelgium, the change in inter-industry mobility in the 1880s had the effect ofcoalescing attitudes towards free trade, as projected by traditional trade theory.These attitudes became apparent in the 1894/5 tariff debate. But if the debatemade clear organized labour’s link between commercial policy and social reform,it did not ensure that the programme would be realized. How, then, was the POBprogramme achieved?

A plausible alternative story is that the ruling party pre-empted workers’demands and instituted a rudimentary welfare state, because capital in the firstwave of globalization was not as mobile as it was in more recent rounds.94 Fearinghigher taxes on their immobile capital that was invested in mines and iron andsteel, the ruling elites would have preferred to introduce some reforms to mollifylabour. The problem with this line of reasoning is that by the 1890s, owners of

90 Belgium, Annales parlementaires, 1894–5, p. 879.Vandervelde’s exact words were: ‘Nous ne demandons pas deprotection pour aucune industrie’ (ibid., p. 1675).

91 Rodrik, ‘Open economies’; for a critical assessment, see Iversen and Cusack, ‘Welfare state’.92 O’Rourke and Williamson, Globalization, p. 98.93 On the interwar period, see Berman, Primacy of politics.94 On capital mobility, see Bordo, Eichengreen, and Irwin, ‘Globalization today’. On capital and the welfare

state, see Boix, Democracy, pp. 171–203.



Belgian mines and steel mills had investments in France and Germany, and thatoverall Belgium’s capital exports on a per capita basis were comparatively signifi-cant.95 Paradoxically, labour itself offered support for foreign investment, observ-ing that their jobs were dependent on the association between capital outflows andexports.96 The story then was not of weak capital but of growing labour confidencein using its influence.

In the tripartite politics of the period, coalition building was frequentlyresorted to both by the ruling and opposition parties. All parties had experienceof forming alliances at the local level, and with the move to proportionalrepresentation (PR) in 1899, coalition politics became a fact of life in nationalpolitics.97 The moment was propitious for the POB. The tariff debate gave theparty credibility, a necessity in forming cross-party alliances, if only because itdemonstrated steadfastness in its critique of the government’s commercialpolicy.98 But it was not evident, a priori, who was labour’s natural partner. Onsocial issues, except for the religious question, the POB may have been closer tocertain groups of Conservatives that advocated reforms, albeit to curtail social-ism; as for the Liberals, on ideological grounds they were less inclined tosupport labour and social legislation that impinged on who could work andwhen.99 But because the labour party showed that its support for free trade wasconditional upon social reform, it had created a viable bargaining strategy. Sincethe Liberal Party needed worker support if it ever hoped to regain power, itmoved closer to the POB.100 By the late 1890s, a Lib–Lab coalition was in placewhose combined influence effectively pressured the ruling Conservatives toforsake protectionism and to introduce new social legislation.101

The change in the electoral system strengthened Belgium’s attachment to freetrade. Rogowski observed that PR polities are most likely to support opennessbecause representatives of large electoral districts are more insulated from regionaland sectoral pressures.102 After reviewing a number of key episodes—and contraryto the received wisdom—Rogowski found that in PR systems, political parties tendto be ‘more disciplined, more coherent, and more ideological’ and more secureagainst sudden shifts in voter sentiments. Large constituencies in small countrieslike that found in Belgium make PR more exact, further encouraging workers,

95 On Belgian foreign ownership and investments, see van der Wee, ‘Investment strategy’, pp. 82–7; Pollard,Peaceful conquest, p. 156.

96 Polasky, Democratic socialism, pp. 70–1; on French socialists and capital exports, see Berger, Premièremondialisation, pp. 53–62.

97 Strikwerda, House divided, pp. 277, 317, gives some examples.98 PR demands trust among political parties of diverse stripes. Excluding extremist factions, various combi-

nations of political parties can come together at any time to set the policy agenda, but to do so, the parties needto demonstrate their willingness and ability to be reliable, if not trustworthy, partners.

99 Puissant, ‘1886’, pp. 86–91; Strikwerda, House divided, p. 262.100 The liberal strategy paid off. The liberals won more seats than the socialists between 1904 and 1914;

Kossmann, Low Countries, p. 477.101 The POB did forsake its traditional demands regarding women’s right to vote; Hilden, Women, p. 301.102 Rogowski, ‘Trade’, p. 208, summarized the main idea: ‘When automakers or dairy farmers entirely dominate

twenty small constituencies and are a powerful minority in fifty more, their voice will be heard in a nation’scouncils.When they constitute but one or two per cent of an enormous district’s electorate, representatives maydefy them more freely.’ For a similar argument, see Persson and Tabellini, ‘Constitutions’, pp. 86–90.



employers, and landlords to vote along ‘class’ as opposed to regional or sectorallines.103

The adoption of PR was also crucial to the timing and components of the reformpackage. In Persson and Tabellini’s model of social spending, government expen-ditures are greater in countries with PR than those with majoritarian systems.104

Before 1890, Belgium was notorious for its poor working conditions, and socialentitlements were unknown. The country was, in Marx’s phrase, a ‘capitalist’sparadise’, and in Vandervelde’s eyes, a ‘workers’ purgatory’.105 Factory inspectionbegan in Britain in 1833, in Germany in 1853, and in France in 1874, but wasintroduced in Belgium only in 1889. The first factory acts were also approved inthat year, but in 1900, the country still ranked 10th out of 16 European countriesin the tightness of its labour laws and in the provision of social spending. Evenlatecomers like Spain and Sweden had superior safety nets at this date.106 Themove to PR in 1899 opened the door to state funding.107 Under pressure from theLib–Lab coalition, which exploited their control of the key committees wheresocial and commercial policy was designed, Belgium’s body of social protectionrapidly improved. Old age pensions were introduced in 1900, accident compen-sation in 1903, and unemployment insurance in 1907. Under the latter legislation,the state subsidized municipal authorities that in turn funded members of union-run voluntary insurance schemes.108 The number of such programmes expandedrapidly, increasing from 310 in 1909 to 634 by 1913.109 Similar labour laws andsocial insurance programmes were in place elsewhere, but the Belgian case standsout because of the speed of their introduction. By 1914, the overall protectionoffered to workers was only slightly inferior to that in France and Germany, and itfar surpassed Denmark’s at this date. It should be noted that throughout this

103 Rogowski, ‘Trade’, p. 215, reported that, in the 1980s, the ratio of voters to districts in Belgium was aboutthe same as in other small countries like Sweden and Denmark.The Netherlands is an extreme case with only oneconstituency (perfect PR). It is difficult to find comparable figures for 1914, because PR was introduced after theFirst World War in the Netherlands, but comparing Belgium in 1911 and the Netherlands in 1919, the ratio ofvoters to districts was about the same in the two countries; Mackie and Rose, International almanac, pp. 39–97.The Belgian method of seat allocation was devised by a national, the lawyer and mathematician Victor d’Hondt,in 1878.

104 Persson and Tabellini, ‘Constitutions’, p. 87. The idea is that PR systems often result in coalitions that usestate budgets to maintain the support of its various members.

105 Vandervelde, Assurance ouvrière, p. 4; Marx cited in Polasky, Democratic socialism, p. 5.106 For an international comparison, see Huberman and Lewchuk, ‘Labour compact’, p. 21.107 Progressive income taxation was introduced in Belgium immediately at the end of hostilities, but the

direction of the government’s fiscal position was set in the decade prior to the war; Kossmann, Low Countries,p. 651. According to Lindert, ‘Social spending’, p. 10, government social spending rose from 0.26% of nationalproduct in 1900 to 0.43% in 1910.This increase was greater than in any other country Lindert studied, includingDenmark and the UK. Lindert’s figures for Belgian government expenditure may be an underestimate, becausesome of the programmes, like unemployment insurance, were partially financed by sub-national authorities;Vanthemsche, ‘Unemployment insurance’, p. 356. Lindert, Growing public, pp. 165–6, reported that, as a result ofpressure from the trade unions and the POB, these years saw an increase in spending on education in additionto social programmes.

108 The intention was not to strengthen the union movement, but to exploit the institutions in place to ensuredelivery; Vanthemsche, ‘Unemployment insurance’, p. 353.

109 Benefits were distributed in two ways. The Ghent system gave the benefits directly to individual membersof the union sponsored funds; the Liège system provided subsidies to unions. On this, see Goossens, Peeters, andPepermans, ‘Interwar unemployment’, pp. 292–4; Harrington, ‘Trade and social insurance’, pp. 221–8. Thefigures in the text refer to the more popular Ghent programmes; Vanthemsche, ‘Unemployment insurance’,p. 356.



period, there was no movement to raise trade barriers and levels of opennessactually increased.Vandervelde’s ‘alternative’ policy mix was realized, with labourchoosing social over tariff protection.

Size may have mattered here. Economists and political scientists have claimedthat the provision of public goods is greater where populations are more homo-geneous and in small polities.110 The Belgian experience points to the importanceof the latter. Strong inter-regional mobility, made possible by the dense trainnetwork, broke down local and sectoral interests. Smallness had clearly abetted thetransformation of the labour market and acted as a countervailing force in oppo-sition to long-standing historical divisions that militated against the provision ofpublic goods on a national scale.

The decline in hours of work illustrates the success of the new legislation.Belgians had the longest working day in Europe into the late-nineteenth century,600 hours per year more than the British and 200 hours more than the French in1880. The first piece of legislation curtailing the length of the working day forwomen and children was passed in 1889, and by 1913 the country had convergedto the European average, labouring about 100 hours per year less than inFrance.111 The comparison with France is indicative of some of the broader forcesunderlying the role of legislation. France had well-entrenched regional labourmarkets, with their own distinctive work patterns and rhythms. In 1892, it passedcomparable legislation to that of its smaller neighbour, but the move to greaterstandardization of work-times was more apparent in Belgium.112 Union power wasabout the same in the two countries. I surmise that legislation had more effect inBelgium because labour markets had become more integrated.113

The new legislation encountered opposition from employers who claimed thatshorter hours would impair their international competitiveness. But globalizationseems to have had the opposite effect, exactly as labour leaders predicted during the1894/5 tariff debate.114 Vandervelde argued with great sophistication that, in con-trast to German enterprises that passed the costs of protection on to consumers,increased external exposure would force Belgian firms to modernize; and because ofthe resulting productivity increases, firms would be able to absorb the added costsof regulation and social programmes, which would in turn have a positive effect onworker output.Vandervelde observed this chain of events after the passage in 1889of work-time legislation in the cotton textile industry. New machinery was installed,the average firm size increased, and cloth exports to Britain rose. On the same basis,he argued for further limitations to the working day in textiles, and in 1909 headvocated a nine-hour working day in coal mining. In the end, globalization didimprove Belgian workers’ wellbeing; the vehicle, however, was not the classicaleffects of rising wages, but labour and social legislation.

110 Alesina, Baqir, and Easterly, ‘Public goods’.111 Huberman, ‘Working hours’, p. 982.112 Excluding services and agriculture, the standard deviation of weekly work hours in France was 5.5 before

and after 1892. For Belgium, it had declined from 6.2 to 5.2 after the introduction of hours legislation in 1889;data set described in Huberman, ‘Working hours’, pp. 990–1.

113 For Belgium, see Vandervelde, ‘Inspection’. For France, see Rist, Réglementation.114 Vandervelde, ‘Inspection’; idem, ‘Lois sociales’; idem, ‘Journée de neuf heures’; idem, ‘Labour reforms’. Van

Houtte, Industrie textile, corroborated many of Vandervelde’s claims about cotton textiles. Between 1896 and1910, the average size of spinning firms increased from 190 to 216 workers (p. 102). Over the same period, yarnoutput increased by more than 80% (p. 257) and exports doubled (p. 265).



VI

In the 1860s, when Belgium made its initial move to reduce tariff barriers in theaftermath of Cobden-Chevalier, the government assured sceptics that the newliberal trade order enveloping Europe would have a positive effect on wages anddampen workers’ militancy. 115 However, for much of the second half of thenineteenth century, wages did not move as expected, and pay rises were unevenlydistributed across industries. In the absence of inter-industry mobility, workers’fate was determined by their sector of employment. The extension of the trainnetwork and subsidized transport in the mid-1880s changed this dynamic. Wagedispersion across regions and industries narrowed and, despite deep fault-linesalong social, religious, and language issues, the mass of unskilled workers began toshare common interests, at least on economic questions. Organized labour came toback free trade as standard trade theory would have predicted, but its support forlower tariffs was conditional upon more and better social legislation. The left sawthe tariff system as inherently inequitable and, in its place, sought to introduce aprogressive income tax to pay for the new programmes. This position was clearlyenunciated by the POB in the parliamentary debates on commercial policy in1894/5 and was achieved in the following decade in coalition with the liberals.Workers took the benefits generated by free trade in the form of better workconditions and social entitlements, as opposed to increased monetary rewards.Politics had trumped markets.

The link between globalization and the extension of the franchise and changesto the electoral system was two-sided. On the one hand, globalization was thetrigger for workers to become active in electoral politics, because they could notrely on the market to raise their wages and they were compelled to turn to theirelected representatives to get the promised gains of openness and raise their livingstandards, broadly defined to include their conditions and entitlements of work.On the other hand, the move to a more representative electoral system helped todeepen workers’ attachments to globalization. Regardless of the causality, the basisof the Belgian welfare state was not created in opposition to or as a backlashagainst the pressures of economic integration, but was complementary to it.Indeed, if workers had exercised their voices to moderate the forces of globaliza-tion, their claims for more labour regulation would have been weaker.

The logic of the POB was evident in the approaches of the left to commercialpolicy in many European countries. In Germany, a new generation of socialistschose openness and universal social programmes over protectionism and Bis-marck’s targeted welfare programmes; in France, groups of workers welcomedglobalization on similar grounds; and in liberal Britain, contemporaries referred toLloyd George’s fiscal reforms of 1909 as the ‘Free Trade Budget’—althoughhistorians have since called it the People’s Budget—because of its link betweenredistributive policy and traditional trade policy, and in contrast to the attempt bythe Conservatives earlier in the decade to initiate tariff reform to protect jobs.116

115 Witte, Craeybeckx, and Meynen, Political history.116 France: Berger, Première mondialisation, pp. 74–80; Germany: Bernstein, ‘Zum Kampf’; UK: Dangerfield,

Strange death, p. 30; Murray, People’s Budget, pp. 49–50; Hellwig, ‘Origins’, pp. 121–3. Harrington, ‘Trade andsocial insurance’, pp. 141–333, compared the association between openness and unemployment insurance inBelgium, Great Britain, Switzerland, and the United States.



Across the Atlantic, the story was different because labour was relatively scarceand, consequently, support for free trade was weaker, especially among unions.Nevertheless, Democrats in the US Congress in 1913 insisted on and succeededin reducing tariffs and introducing an income tax simultaneously. Woodrow Wil-son’s Secretary of Agriculture echoed the terminology of Vandervelde: ‘Think ofit—a tariff revision downwards . . . not dictated by the manufacturers . . . A pro-gressive income tax! I did not think we would live to see these things’.117

It was in Europe’s small countries that the connection between social policy andglobalization took hold most firmly and persisted into the interwar years andafter.118 On the basis of the Belgian experience, it seems that the virtues of beingsmall in the wave of globalization before 1914 were threefold. First, smallnesspermitted the rapid transition from a world of immobile to mobile labour; second,small European countries tended to institute electoral reform based on PR thatconsolidated organized labour’s role as the voice of workers; and third, smallnessfacilitated the provision of public goods because it acted as a countervailing forceagainst historical divisions, linguistic, ethnic, or religious.

Labour’s support for globalization in the period before 1914 foreshadowed itsacceptance of increased levels of openness in the golden age of European growthafter 1945.119 In both periods, the welfare state was extended as trade expanded.However, there was an important difference between the two. During the secondwave of globalization, workers moderated wage demands to obtain increased levelsof social entitlements; in the period before 1914, labourers’ political demandsarose, in the case of Belgium, in the absence of upwards wage pressure.That said,both periods attest to the view that workers support globalization when politicalinstitutions are in place that allow them to share in its benefits.

Université de Montréal, CIRANO, CIREQ

Date submitted 10 March 2005Revised version submitted 9 February 2006Accepted 4 May 2006

DOI: 10.1111/j.1468-0289.2007.00396.x

Footnote referencesAbs, R., ‘Les statuts du parti ouvrier belge de 1885–1894’, Socialisme, 94 (1969), pp. 466–72.Abs, R., Histoire du parti socialiste belge, 1885–1900: synthèse historique (Brussels, 1974).Acemoglu, D. and Robinson, J. A., ‘Why did the West extend the franchise? Democracy, inequality, and growth

in historical perspective’, Quarterly Journal of Economics, 115 (2000), pp. 1167–99.Acemoglu, D. and Robinson, J. A., Economic origins of dictatorship and democracy (Cambridge, Mass., 2005).

117 The first post-Civil War income tax sponsored by William Jennings Bryan was voted in 1893 (lateroverturned by the Supreme Court) as part of a tariff reform bill. The income tax enacted in 1914 was part of apackage of reform measures including tariff reductions. See Sanders, Roots of reform, pp. 226–7. I thank GeraldFreidman for this reference.

118 The relation between tariff reductions and direct taxation was more evident in Denmark than in Sweden.The issue separating the two was the attitude of the agrarian sector to protectionism. Swedish farmers were moreprotectionist and they could absorb the burden of paying their share of contributions for social insurance; Daneswere in favour of free trade and, because of international competition, insisted that the state finance the new socialprogrammes. See Baldwin, Politics, pp. 65–94.

119 Eichengreen, ‘Institutions and economic growth’.



Alesina, A., Baqir, R., and Easterly, W., ‘Public goods and ethnic divisions’, Quarterly Journal of Economics, 114(1999), pp. 1243–84.

Angel, N., The great illusion: a study of the relation of military power to national advantage (Toronto, 1913).Ashley, P., Modern tariff history (New York, 1905).Bairoch, P., ‘European trade policy, 1815–1914’, in P. Mathias and S. Pollard, eds., Cambridge economic history of

Europe,VIII.The industrial economies: the development of economic and social policies (Cambridge, 1989), pp. 1–160.Baldwin, P., The politics of social solidarity: class bases of the European welfare state, 1875–1975 (Cambridge, 1990).Bayly, C. A., The birth of the modern world, 1780–1914 (Oxford, 2004).Berger, S., Notre première mondialisation: leçons d’un échec oublié (Paris, 2003).Berman, S., The primacy of politics: social democracy and the making of Europe’s twentieth century (Cambridge, 2006).Bernstein, E., ‘Zum Kampf gegen die Zollschraube’, Sozialistische Monatshefte, 2 (1901), pp. 690–711.Blomme, J., The economic development of Belgian agriculture, 1880–1980: a quantitative and qualitative analysis

(Leuven, 1993).Boix, C., Democracy and redistribution (New York, 2003).Bordo, M., Eichengreen, B., and Irwin, D., ‘Is globalization today really different from globalization a hundred

years ago?’ in S. Collins and R. Lawrence, eds., Brookings Trade Forum (Washington, D.C., 1999), pp. 1–50.De Brabander, G. L., Regional specialization, employment and economic growth in Belgium between 1846 and 1970

(New York, 1981).Cassiers, I., ‘Le rôle de l’État à l’apogée du libéralisme, 1850–1886’, Contradictions, 23–24 (1980), pp. 121–43.Cassiers, I., ‘Une fonction importante de l’État libéral: la création et la gestion des chemins de fer 1850–1914’,

Contradictions, 23–24 (1980), pp. 153–64.Dangerfield, G., The strange death of liberal England (New York, 1935).Degrève, D., Le commerce extérieur de la Belgique, 1830–1913–1939: présentation critique des données statistiques

(Brussels, 1982).Delsinne, L., Le Parti ouvrier belge des origines à 1894 (Brussels, 1955).Donald, M., ‘Workers of the world unite? Exploring the enigma of the Second International,’ in M. H. Geyer and

J. Paulmann, eds., The mechanics of internationalism: culture, society, and politics from the 1840s to the FirstWorldWar(2001), pp. 177–204.

Duriviaux. R., La Banque mixte: origine et soutien de l’expansion économique de la Belgique (Brussels, 1947).Van den Eeckhout, P., ‘Belgium’, in C. G. Pooley, ed., Housing strategies in Europe, 1880–1930 (Leicester, 1992),

pp. 190–220.Eichengreen, B., ‘Institutions and economic growth: Europe afterWorldWar II’, in N. Crafts and G. Toniolo, eds.,

Economic growth in Europe since 1945 (Cambridge, 1996), pp. 38–72.Engerman, S. and Sokoloff, K., ‘The evolution of suffrage institutions in the New World’, Journal of Economic

History, 65 (2005), pp. 891–922.Estevadeordal, A., ‘Measuring protection in the early twentieth century’, European Review of Economic History,

1 (1997), pp. 89–125.Fletcher, R., Revisionism and empire: socialist imperialism in Germany, 1897–1914 (1984).Flora, P., State, economy and society in western Europe, 1815–1975, vol. 2 (Frankfurt, 1983).Foreman-Peck, J. and Lains, P., ‘European economic development: the core and the southern periphery,

1870–1910’, in S. Pamuk and J. G. Williamson, eds., The Mediterranean response to globalization before 1950(2000), pp. 76–106.

Gadisseur, J., ‘Output per worker and its evolution in Belgian industry, 1846–1910’, in R. Fremdling andP. K. O’Brien, eds., Productivity in the economies of Europe (Bamberg, 1983), pp. 141–51.

Gay, P., The dilemma of democratic socialism: Eduard Bernstein’s challenge to Marx (New York, 1952).George, H., Protection or free trade: an examination of the tariff question with special regard to the interest of labor (New

York, 1886).Gerschenkron, A., Bread and democracy in Germany (Ithaca, 1943).Goossens, M., Peeters, S., and Pepermans, G., ‘Interwar unemployment in Belgium’, in B. Eichengreen and

T. J. Hatton, eds., Interwar unemployment in international perspective (1988), pp. 289–324.Gourevitch, P., Politics in hard times: comparative responses to international economic crises (Ithaca, 1986).Grunzel, J., Economic protectionism (1916).Hagemann, H., ‘The Verein fùr Sozialpolitik from its foundation until World War I’, in M. M. Augello and

M. E. L. Guidi, eds., The spread of political economy and the professionalisation of economists (2001), pp. 152–75.Harrington, M., ‘Trade and social insurance: the development of national unemployment insurance in advanced

industrial democracies’ (unpub. Ph.D. thesis, Univ. of California at Los Angeles, 1998).Harris, J., ‘From poor law to welfare state? A European perspective’, in D. Winch and P. K. O’Brien, eds., The

political economy of British historical experience, 1688–1914 (Oxford, 2002), pp. 409–38.Hatton, T. J. and Williamson, J. G., The age of migration: causes and economic impact (New York, 1998).Hellwig, T. J., ‘The origins of unemployment insurance in Britain: a cross-class alliance approach’, Social Science

History, 29 (2005), pp. 107–36.Helpman, E., The mystery of economic growth (Cambridge, 2004).Herren, M., Hintertüren zur Macht: Internationalismus und modernisierungsorientierte Außenpolitik in Belgien, der

Schweiz und den USA, 1865–1914 (Oldenburg, 2000).



Hilden, P., Women, work and politics: Belgium, 1830–1914 (Oxford, 1993).Hiscox, M., International trade and political conflict: commerce, coalitions, and mobility (Princeton, 2001).Hiscox, M., ‘Interindustry factor mobility and technological change: evidence on wage and profit dispersion

across US industries, 1820–1990’, Journal of Economic History, 62 (2001), pp. 383–417.Hollande, M., La défense ouvrière contre le travail étranger: vers un protectionnisme ouvrier (Paris, 1913).Van Houtte, F.-X., L’évolution de l’industrie textile en Belgique et dans le monde de 1800 à 1939 (Leuven, 1949).Huberman, M., ‘Working hours of the world unite? New international evidence of worktime, 1870–1913’, Journal

of Economic History, 64 (2004), pp. 964–1001.Huberman, M. and Lewchuk, W., ‘European economic integration and the labour compact, 1850–1913’,

European Review of Economic History, 7 (2003), pp. 3–41.Irwin, D. A., ‘The political economy of free trade: voting in the British general election of 1906’, Journal of Law

and Economics, 37 (1994), pp. 75–108.Irwin, D. A., ‘IIndustry or class cleavages over trade policy? Evidence from the British general election of 1923’,

in R. C. Feenstra, G. M. Grossman, and D. A. Irwin, eds., The political economy of trade policy: essays in honorof Jagdish Bhagwati (Cambridge, 1996).

Irwin, D. A., ‘Interpreting the tariff-growth correlation of the late 19th century’, American Economic Review, 92(2002), pp. 165–9.

Iversen, T. and Cusack, T. R., ‘The causes of welfare state expansion’, World Politics, 52 (2000), pp. 313–49.Kossmann, E. H., The Low Countries: 1780–1940 (Oxford, 1978).Krueger, A. and Summers, L., ‘Reflections on the inter-industry wage structure’, in K. Lang and J. Leonard, eds.,

Unemployment and the structure of labor markets (Oxford, 1987), pp. 17–47.Leboutte, R., Puissant, J., and Scuto, D., Un siècle d’histoire industrielle (1873–1973): Belgique, Luxembourg,

Pays-Bas: industrialisations et sociétés (Paris, 1998).Lentacker, F., La frontière franco-belge. Étude géographique des effets d’une frontière internationale sur la vie de relations

(Lille, 1974).Van der Linden, M., ‘The national integration of European working classes, 1871–1914’, International Review of

Social History, 33 (1988), pp. 285–311.Lindert, P., ‘The rise of social spending: 1880–1930’, Explorations in Economic History, 31 (1994),

pp. 1–38.Lindert, P., Growing public: social spending and economic growth since the eighteenth century, vol. 1 (NewYork, 2004).Mackie, T. T. and Rose, R., The international almanac of electoral history (New York, 1974).Maddison, A., The world economy: a millennial perspective (Paris, 2001).Magee, S. P., ‘Three simple tests of the Stolper-Samuelson theorem’, in P. Oppehheimer, ed., Issues in interna-

tional economics (Stockfield, 1980), pp. 139–43.Mahaim, E., Les abonnements d’ouvriers sur les lignes de chemins de fer belges et leurs effets sociaux (Brussels, 1910).Mares, I., The politics of social risk: business and welfare state development (Cambridge, 2003).Marx, K., ‘Speech on free trade (1848)’, reprinted in D. McLellan, ed., Karl Marx: selected writings (Oxford,

1977), pp. 269–70.Marx, K., Capital, vol. III (New York, 1981 [1886]).Mayda, A. M. and Rodrik, D., ‘Why are some individuals (and countries) more protectionist than others?’

European Economic Review, 49 (2005), pp. 1393–1430.Milhaud, E., Le Congrès socialiste de Stuttgart (Paris, 1899).Mitchell, B. R., European historical statistics, 1750–1975 (New York, 2nd edn, 1980).Mitrany, D., Marx against the peasant: a study in social dogmatism (Chapel Hill, 1951).Van Molle, L., Katholieken En Landbouw: Landbouwpolitiek in België (Leuven, 1989).Murray, B. K., The People’s Budget, 1909/10: Lloyd George and liberal politics (Oxford, 1980).Neuville, J., Naissance et croissance du syndicalisme (Brussels, 1979).Noiriel, G., ‘L’historien face aux défis du XXIe siècle: mondialisation des échanges et crise des états-nations’, Les

conférences Gérard-Parizeau, Série Université de Montréal, 1 (Montreal, 2001).O’Rourke, K. H., ‘The European grain invasion, 1870–1913’, Journal of Economic History, 57 (1997),

pp. 775–801.O’Rourke, K. H. and Sinnott, R., ‘The determinants of individual attitudes towards immigration’, European

Journal of Political Economy, 22 (2006), pp. 838–61.O’Rourke, K. H. and Williamson, J. G., Globalization and history: the evolution of a nineteenth-century Atlantic

economy (Cambridge, 1999).Parti Ouvrier Belge, Rapports annuels (Brussels, various years).Pasleau, S., ‘Les migrations de main-d’oeuvre en Belgique’, in Y. Landry, J. A. Dickinson, S. Pasleau, and

C. Desama, eds., Les chemins de la migration en Belgique et au Québec (Beauport, Quebec), pp. 173–94.Persson, T. and Tabellini, G., ‘Constitutions and economic policy’, Journal of Economic Perspectives, 18 (2004),

pp. 75–98.Polasky, J., The democratic socialism of ÉmileVandervelde: between reform and revolution (Oxford, 1995).Polasky, J., ‘Transplanting and rooting workers in London and Brussels: a comparative history’, Journal of Modern

History, 73 (2001), pp. 528–60.Pollard, S., Peaceful conquest: the industrialization of Europe, 1760–1970 (Oxford, 1981).



Pollard, S., ‘Free trade, protectionism, and the world economy’, in M. H. Geyer and J. Paulmann, eds., Themechanics of internationalism: culture, society, and politics from the 1840s to the FirstWorldWar (2001), pp. 27–54.

Prato, G., Le protectionnisme ouvrier (Paris, 1912).Puissant, J., L’évolution du mouvement ouvrier socialiste dans le Borinage (Brussels, 1979).Puissant, J., ‘1886. La contre-réforme sociale?’ in P. van der Host, ed., Cent ans de droit social en Belgique (Brussels,

1986) pp. 67–108.Rist, C., La réglementation légale de la journée de travail de l’ouvrier adulte (Paris, 1898).Rodrik, D., ‘Why do more open economies have bigger governments?’ Journal of Political Economy, 106 (1998),

pp. 997–1033.Rogowski, R., ‘Trade and the variety of democratic institutions’, Industrial Organization, 41 (1987), pp. 203–23.Rogowski, R., Commerce and coalitions: how trade affects domestic political alignments (Princeton, 1989).Rowntree, B. S., Land & labour; lessons from Belgium (1910).Sanders, E. M., Roots of reform: farmers, workers, and the American state, 1877–1917 (New York, 1999).Scholliers, P., ‘Industrial wage differentials in nineteenth-century Belgium’, in Y. S. Brenner, H. Kaelble, and

M. Thomas, eds., Income distribution in historical perspective (Cambridge, 1991), pp. 96–116.Scholliers, P., ‘A century of real industrial wages in Belgium, 1840–1940’, in P. Scholliers and V. Zamagni, eds.,

Labour’s reward, real wages and economic change in 19th and 20th century Europe (Aldershot, 1995), pp. 106–37.Scholliers, P., Wages, manufacturers and workers in the nineteenth century factory: theVoortman Cotton Mill in Ghent

(Oxford, 1996).Scholliers, P., ‘Mots et pratiques. l’industrie cotonnière gantoise, les crises et la perception patronale de la

concurrence internationale, 1790–1914’, Revue d’histoire du XIXe Siècle, 23 (2001), pp. 121–42.Silver, B. J., Forces of labor: workers’ movements and globalization since 1870 (New York, 2003).Steenson, G., After Marx, before Lenin: marxism and socialist working-class parties in Europe, 1889–1914 (Pittsburg,

1991).Stengers, J., Émigration et immigration en Belgique au XIXe et au XXe siècles (Brussels, 1978).Strikwerda, C., A house divided: Catholics, socialists and Flemish nationalists in nineteenth-century Belgium (NewYork,

1997).Strikwerda, C., ‘France and Belgian immigration of the nineteenth century’, in C. Guerin-Gonzales and

C. Strikwerda, eds., The politics of immigrant workers: labor activism and migration in the world economy since1830 (New York, 1998), pp. 111–44.

Suetens, M., Histoire de la politique commerciale de la Belgique depuis 1830 jusqu’à nos jours (Brussels, 1955).Tilly, C., ‘Globalization threatens labour’s rights’, International Labour and Working Class History, 47 (1995),

pp. 1–24.Trentmann, F., ‘The transformation of fiscal reform: reciprocity, modernization, and the fiscal debate within the

business community in early twentieth-century Britain’, Historical Journal, 39 (1996), pp. 1005–48.Trentmann, F., ‘Wealth vs welfare: the British Left between free trade and national political economy before the

First World War’, Historical Research, 70 (1997), pp. 70–96.Trentmann, F., ‘Political culture and political economy: interest, ideology and free trade’, Review of International

Political Economy, 5 (1998), pp. 217–51.Trentmann, F., ‘National identity and consumer politics: free trade and tariff reform’, in D. Winch and

P. K. O’Brien, eds., The political economy of British historical experience, 1688–1914 (Oxford, 2002), pp. 215–42.Trentmann, F. and Daunton, M., ‘Worlds of political economy: knowledge, practices and contestation’, in

M. Daunton and F. Trentmann, eds., Worlds of political economy: knowledge and power in the nineteenth andtwentieth centuries (Basingstoke, 2004), pp. 1–23.

Vandermotten, C., ‘Tendances longues de l’évolution de la production, de l’emploi et de la productivitéindustriels en Belgique: 1840–1978’, Cahiers économiques de Bruxelles, 86 (1980), pp. 261–301.

Vandermotten, C., Barsten in België (Berchem, 1990).Vandervelde, É., L’assurance ouvrière en Belgique (Brussels, 1889).Vandervelde, É., Le socialisme agricole (Brussels, 1895).Vandervelde, É., ‘Inspection du travail en Belgique’, Revue socialiste, 23 (1896), pp. 318–25.Vandervelde, É., Les lois sociales en Belgique (Brussels, 1897).Vandervelde, É., L’influence des villes sur les campagnes (Brussels, 1899).Vandervelde, É., L’exode rural et le retour aux champs (Paris, 1903).Vandervelde, É., ‘La journée de neuf heures dans les mines’, Revue d’économie politique (1911), pp. 185–209.Vandervelde, É., ‘Labour reforms in Belgium’, in J. E. Solano, ed., Labour as an international problem: a series of

essays comprising a short history of the international labour organisation (1920), pp. 105–31.Vandervelde, É. and Destrée, J., Le socialisme en Belgique (Paris, 1898).Vanthemsche, G., ‘Unemployment insurance in interwar Belgium’, International Review of Social History,

35 (1990), pp. 349–76.Van der Wee, H., ‘Investment strategy of Belgian industrial enterprise between 1830 and 1980 and its influence

on the economic development of Europe’, Belgium and Europe: Proceedings of the International Francqui-Colloquium, Brussels, Ghent, Nov. 1981 (Brussels, 1981), pp. 75–93.

Wauthelet, J. M., ‘Accumulation et rentabilité du capital dans les charbonnages belges, 1850–1914’, Rechercheséconomiques de Louvain, 41 (1975), pp. 265–85.



Williamson, J. G., ‘The evolution of global labor markets since 1830: background evidence and hypotheses’,Explorations in Economic History, 32 (1995), pp. 141–96.

Witte, E., Craeybeckx, J., and Meynen, A., Political history of Belgium from 1830 onwards (Brussels, 2000).

Official publicationsBelgium, Annales Parlementaires: Chambres des Représentants (1884–5).Belgium, Annales Parlementaires: Chambres des Représentants (1894–5).US Department of Labor, Fifteenth Annual Report of the Commissioner of Labor:Wages in Commercial Countries,

2 vols. (Washington, 1900).



Ticket to trade: Belgian labour and globalization before 1914

Documents

Transcript of Ticket to trade: Belgian labour and globalization before 1914