FROM AGRICULTURAL PROTECTION TO FREE TRADE IN BELGIUM, 1834–1873

Figure 1 depicts the evolution of the yearly price averages of wheat, rye, and potatoes for the period 1830–1890, as well as the dates of the votes on the Corn Laws under consideration (dotted vertical lines).Footnote ¹⁸ From this figure, it appears that the Belgian legislators did not simply react to price changes. The law of 22 February 1850 is particularly intriguing because it buried the protectionist sliding scale law at a time when prices were low. The price dip around 1865, however, did not automatically produce any legislation. Figure 1 also includes the ad valorem tariff equivalent of the specific tariffs that were in force.Footnote ¹⁹ While the tariffs of the Belgian sliding scale law, expressed as ad valorem equivalents, were lower than the British ones, the evolution is comparable.Footnote ²⁰

FIGURE 1

PRICE EVOLUTION, AD VALOREM TARIFF EQUIVALENT, AND YEARS OF IMPORTANT VOTES

Sources: Gadisseur, Le produit physique, pp. 756–61; and authors' own calculations.

After Belgian independence in 1830 the prices of wheat and rye started to fall because of good harvests, sparking demands for agricultural protection.Footnote ²¹ The protectionist law of 31 July 1834 introduced a sliding scale inspired by the British Corn Law of 1828, with an increasing import tariff as the prices of wheat and rye declined. When the price reached 19 Belgian francs (BEF) per 100 kg, the highest tariff of 7.5 BEF was imposed (39 percent ad valorem), and if the prices dropped below 15 BEF, imports were banned. At the other extreme, imports were free when prices exceeded a certain threshold, and in the case of extremely high prices, exports were prohibited. The idea behind the sliding scale was to guarantee agricultural producers a minimum price that was considered remunerative. In 1844 the prices of wheat and rye once more descended to an alarmingly low level, sparking calls for a more restrictive protectionist law. In February 1845, 21 members of the House of Commons introduced a bill that became known as the “Law of Famine.” It had the intention of guaranteeing corn producers even higher prices than the 1834 law by introducing two new scales.

However, this pinnacle of agricultural protectionism was never promulgated because of the subsistence crisis of 1845–1848. The failure of the potato crop in 1845 and of the rye harvest in 1846 produced extremely high prices until 1848 (see Figure 1), causing hardship for a large majority of the population. The suffering was compounded by an outbreak of cholera in 1848.Footnote ²² The government suspended the sliding scale law in September 1845, and temporary measures left the import of foodstuffs free until 1850. The yearly renewable measures also prohibited the export of foodstuffs between 1845 and 1848.

After the end of this severe subsistence crisis, the temporary laws were replaced by the law of 22 February 1850, which fixed the new agricultural tariffs for an indefinite period. The law was an important step towards free trade, but also showed many signs of a compromise. The genesis of this law was long and complex. When the prices began to descend at the end of 1848, the Liberal government of Charles Rogier wanted to extend the temporary measures until the consequences of British repeal (which took effect in February 1849) became clear.Footnote ²³ At the end of 1849, Rogier proposed a new bill that prolonged the temporary measures for two years, but by that time the prices of wheat and rye had once more descended to such alarmingly levels that the agricultural protectionists were goaded into action. The parliamentary commission studying the bill argued that agriculture should receive protection as long as industry remained guarded from competition.Footnote ²⁴ Lengthy discussions in the House of Representatives then led to the law of 22 February 1850. It brought the tariff for wheat to 1 BEF per 100 kg, which amounted to a duty of 4.8 percent relative to the average price of 1850.

The law of 22 February 1850 was a mixed triumph. It ended the sliding scale of 1834 and is therefore the equivalent of British repeal. Yet many proponents of free trade voted against the law because they thought that the tariff of 1 BEF was too high. Many protectionists also opposed it because they hoped for the return of a protectionist sliding scale.

Belgium's tariff history became even more complicated thereafter, because the “permanent” law of 1850 did not last long. A new subsistence crisis struck between 1853 and 1857, because of bad harvests from 1853 to 1855 and difficulties with the imports of foodstuffs during the Crimean War (1854–1856). Prices rose even higher than in 1845–1848.Footnote ²⁵ The government reverted to the typical temporary crisis measures of free imports and export prohibition in yearly renewable laws. The Catholic Prime Minister Pieter De Decker, a known free trade proponent,Footnote ²⁶ moved for a new permanent law at the end of 1856. According to contemporary economists, the resulting legislation—the law of 5 February 1857—inaugurated an era of free trade in Belgian corn tariffs.Footnote ²⁷ But the law did leave a tariff of 0.5 BEF for 100 kg of wheat (1.7 percent ad valorem at the high prices of 1857).

The law of 5 February 1857 functioned until a new period of high prices in 1871–1873, when a Catholic government removed all remaining duties on foodstuffs. The law of 2 January 1873 was accepted unanimously in the House of Representatives.

AN ECONOMETRIC ANALYSIS OF REPRESENTATIVES' VOTING BEHAVIOR

Our empirical analysis seeks to disentangle the relative importance of representatives' personal economic interests, their party affiliation, and the economic interests of their constituents in explaining their votes on the Belgian Corn Laws in the period 1834–1873. We study voting on five laws: the sliding scale law of 1834, the law of famine of 1845, the Repeal law of 1850, the law that lowered the tariff to 0.5 BEF in 1857, and the removal of all remaining tariffs in 1873.Footnote ²⁸ The first four roll-call votes will be analyzed quantitatively. Because the removal of tariffs in 1873 was unanimous, it is analyzed qualitatively in the next section.

Data

The votes of the MPs in the roll-call votes are collected in the published parliamentary procedures.Footnote ²⁹ Only votes actually cast are considered, while absentees are disregarded. Table 1 shows the percentages of votes, party affiliation, and personal interests in the roll-call votes under consideration.

TABLE 1

DISTRIBUTION OF VOTES, PARTY AFFILIATION, AND PERSONAL INTERESTS

^aThe constituencies of Diekirch, Roermond, Maastricht, and Luxembourg were omitted as they were transferred to the Netherlands in 1839. Further, three representatives were dropped because of missing personal data.

^bOne representative was also dropped because of missing data.

Sources:

See the text.

The party affiliations of the MPs were collected from the main published biographical dictionary of the members of Belgian Parliament.Footnote ³⁰ Party formation in Belgium revolved about differences over the role of the church and religion that pitted Catholics against Liberals.Footnote ³¹ Party affiliation is represented by three dummy variables: CATH, LIB, and UNION. The first two dummies stand for the respective attachment to the Catholic and Liberal party. UNION stands for a political current of Unionists, which strived for cooperation among the central political parties in the first years of independence. The Unionists are of limited importance for our analysis and they disappeared as tensions between Catholics and Liberals mounted in the 1840s. In line with British literature, we expect Catholic representatives (Conservatives) to favor agricultural protection more than the Liberals.

Representatives' personal economic interests are derived from biographical dictionaries of Belgian Parliament and similar biographical sources for other political institutions.Footnote ³² Large landowners are found in the list of persons eligible for the Senate. Eligibility was conditional on paying a high level of taxes, which required owning a large amount of land.Footnote ³³ Laureyssens' list of board members was employed to find representatives involved in the industry.Footnote ³⁴ The personal economic background of representatives is captured by four mutually exclusive dummy variables: PERLAND indicates eligibility for the Senate because of landownership, PERIND indicates membership on the board of industrial and financial companies, PERBOTH indicates a combination of landownership and interests in industrial and financial companies. As in Britain, portfolio diversification was a reality in Belgium, with industrial classes investing in property and landowners investing in the financial and industrial sectors.Footnote ³⁵ Finally, PERFREE indicates activity as lawyer, military officer, civil servant (before 1848), notary, doctor, journalist, university professor, or engineer (with no large landownership or interests in industrial or financial companies). We expect landowners to favor agricultural protection and industrialists to prefer free trade because this would allow for lower wages.

The economic and geographic characteristics of the constituencies come from data for the year 1846, but the economic and demographic structure did not change much over the period of the analyzed votes (1834–1857).Footnote ³⁶ The population and industrial censuses of 1846 allow us to capture the urbanization (the share of the population living in cities) and industrialization of the constituencies. The general demographic structure of the constituency is captured by its area (area) and population (population), and their ratio, the population density (popdensity). The variables popcitrel and needyrel are the percentage of the population living in cities and the percentage officially recognized as needy. The number of needy people is an indicator for poverty and potential political instability (food riots). Industrialization is measured by the percentage of the population working in sectors in which (private) firms employed on average more than 100 people (indrel100).

We also have two variables that measure changes in the franchise that may have affected support for free trade. One of them measures the percentage of voters living in cities before the law of 12 March 1848 (cityvoterel7) and the other the percentage of voters living in cities after the law of 12 March 1848 (cityvoterel8).Footnote ³⁷ The law of 12 March 1848 lowered the amount of taxes payable to obtain voting rights.Footnote ³⁸ It increased the number of voters living mainly in more urban constituencies and can thus be part of the explanation of the shift to free trade. For measures of constituents' stake in agriculture, we used the agricultural census, which provided the percentage of farmland sown in wheat and rye (wheatryerel) and in potatoes (potatoerel); the number of livestock (dairy cows of more than two years old plus pigs older than two months), the percentage of agricultural land worked by owner operators (propratio) as opposed to tenants; and the percentage of farms over ten hectares large (farm10ha). We expect representatives whose constituencies specialize in wheat and rye to favor agricultural protection. We also expect more support for protectionist measures if more land is leased rather than owned, because large landowners would expect protectionism to raise their rental incomes and they would be more effective lobbyists. The preference of constituents in districts that specialize in livestock production is harder to predict: as consumers of cereals they gain from lower cereal prices, but the laws under consideration in 1850 and 1857 also incorporated tariffs on livestock.Footnote ³⁹

We also measure the influence of the subsistence crisis, because we expect that representatives from constituencies where the subsistence crisis struck hard would be more likely to support agricultural trade liberalization. The impact of the subsistence crisis is captured by a dummy variable (POTATOCRIS), which is one if the fraction of the needy exceeded 25 percent of population and if at least 4 percent of the agricultural land was used to grow potatoes.Footnote ⁴⁰ This variable is one for 14 (out of a total of 41) constituencies, which are grouped in two geographical clusters: East and West Flanders and a smaller region east of Brussels. The first cluster was particularly severely hit. Most farms in this region were very small, intensively worked, and leased for a high rent. The typical household here combined cottage industry with farming.Footnote ⁴¹ When the potato crop failed, these agricultural producers became dependent on the market—or charitable institutions—for food.

TABLE 2

CORRELATION MATRIX (all districts weighted equally)

Sources:

All results are from the authors' calculations.

Table 2 displays the correlations of these variables. Note how population density correlates positively with the percentage of the population living in cities and with agricultural specialization in cereal and potato production. More densely populated constituencies also tend to have less land owned by the farmers and fewer farms over ten hectares. Because the percentage of the population in cities and the fraction of voters in cities are almost perfectly correlated, we use only the fraction of the voters in cities in our analysis. Because the share of the population that is needy was measured in 1846, at the height of the subsistence crisis, it is positively correlated with cereal and potato production.

General Analysis

After eliminating nine abstentions in the 1850 vote and seventeen in the 1857 vote, we can analyze the roll-call votes of 1834, 1845, 1850, and 1857 as a binary variable.Footnote ⁴² We apply probit analysis to each vote separately to allow for maximal flexibility and account for differences in the content of the laws. General differences in historical circumstances are, as far as they are orthogonal to our variables, absorbed into the constants of the regressions. Table 3 presents the marginal effects of each explanatory variable in a probit regression.Footnote ⁴³

These marginal effects represent the change in the probability of a vote for trade liberalization due to a unit increase in the relevant variable, while all other explanatory variables are set equal to their mean values. For dummy variables, the marginal effect is calculated by changing membership from the reference group to the group indicated by the dummy. For logarithmic variables, it concerns a one percent increase.

The representatives' personal economic background only affects voting in a statistically significant way in 1857. Representatives with industrial interests are, other things being equal, 41 percent more likely to vote for liberalization in 1857 than great landowners and 39 percent more likely to vote yes if they also own land in addition to their industrial holdings. But in the other votes, the representatives' personal economic interests had no significant effect.Footnote ⁴⁴ Party affiliation, by contrast, is the most important variable in terms of its magnitude and its statistical significance, except for the 1857 vote, when both the Catholic and Liberal parties were divided. All other things being equal, a Liberal representative was in 1834, 1845, and 1850 some 60, 52, and 77 percent more likely to support agricultural trade liberalization than a Catholic representative.

As for the constituency background, a higher share of agricultural land leased does increase—as expected—the probability of a protectionist vote in 1834 and 1857. Livestock and wheat and rye production boost the likelihood of a protectionist vote. The effects of industrialization and the potato crisis, however, are not significantly different from zero Surprisingly, the probability of a protectionist vote is higher if a greater share of voters live in the cities, which suggests that the increased prominence of city voters due to the law of 12 March 1848 was not a factor in Belgium's move towards agricultural trade liberalization. Perhaps some representatives of urbanized constituencies believed that the protectionism of the sliding scale was in the best interests of their constituents, because it would also restrict exports in times of scarcity. Furthermore, the scale of urbanization was relatively limited in Belgium, even though it was increasing rapidly. Brussels, for instance, only counted 123.874 inhabitants in 1846 and even in the most urbanized constituencies a majority of the population still lived in the countryside.

TABLE 3

MARGINAL EFFECTS PROBIT ESTIMATION

* = Significant at the 10 percent level.

** = Significant at the 5 percent level.

*** = Significant at the 1 percent level.

Notes:

Marginal effects computed at the mean. The standard errors are in parentheses. For the dummy variables (PERIND, PERFREE, PERBOTH, LIB, UNION, and POTATOCRIS), the marginal effects are computed for a discrete change of the dummy variable from 0 to 1. Dependent variable: Vote pro trade liberalization.

Sources:

All results are from the authors' calculations.

Additional Analysis of the Votes in 1850 and 1857

The vote on the law of 22 February 1850 was, as we have noted, a compromise that was not completely acceptable for all parties. A number of radical free trade supporters and protectionists rejected the compromise, which obviously influences the analysis of the 1850 vote. We can get a better sense of what was happening by examining the votes on three amendments to the bill. On 2 February 1850 all present members of the House of Representatives voted for one of the three tariff options proposed by amendments representing the free trade position (a tariff of 0.5 BEF on wheat), the compromise position (a tariff of 1 BEF), and the protectionist position (a tariff of 1.5 BEF). The tariffs amounted to respectively 2.4 percent, 4.8 percent, and 7.2 percent ad valorem at the average prices prevailing in 1850.Footnote ⁴⁵ The marginal effects at the mean of an ordered probit analysis of the votes on the amendments to the 1850 bill are presented in Table 4.

In contrast to the vote on the final compromise, the representatives' personal economic background did affect their votes on the amendments. All other things equal, representatives with industrial interests are 32 percent less likely to vote for the protectionist amendment and 27 percent more likely to prefer the 0.5 BEF amendment than landowners. Representatives with a professional background show a similar, but less pronounced preference for liberalization. Representatives with a mixed background in land and industry are significantly more likely than those whose interests are in land alone to vote for the protectionist option.

Livestock and wheat and rye production in the constituencies are still an important factor in producing support for the protectionist amendment. But as before, party affiliation has the most significant effect: Liberal representatives were 56 percent less likely to vote for the protectionist option, and respectively 37 percent and 19 percent more likely to vote for the middle and liberalization options than Catholic representatives. The ordered probit on the amendments allows one to conclude that the underlying variables still work in the same direction in 1850 as they did in 1834 and 1845. However, during the voting on the amendments in 1850, the representatives had more room to take their personal interests into account.

TABLE 4

MARGINAL EFFECTS OF ORDERED PROBIT ANALYSIS 1850 AMENDMENT VOTE

* = Significant at the 10 percent level.

** = Significant at the 5 percent level.

*** = Significant at the 1 percent level.

Notes:

Marginal effects computed at the mean. The standard errors are in parentheses. For the dummy variables (PERIND, PERFREE, PERBOTH, LIB, and POTATOCRIS), the marginal effects are computed for a discrete change of the dummy variable from 0 to 1.

Source:

All results are from the authors' calculations.

That leaves the vote in 1857 as an apparent enigma. The strongest variable of the previous years, party affiliation, stopped functioning in 1857, while personal economic background and the characteristics of their constituencies take on a much greater importance than before (Table 3). To understand this, we need to examine the discussions on corn tariffs in the years immediately preceding 1857, when corn prices were extremely high (Figure 1). The protectionists had based their demands for agricultural protection on the idea of justice, and not so much on any economic theory.Footnote ⁴⁶ The leader of the protectionists during the 1850s, the Catholic Barthélemy Dumortier, was very explicit that he and his partisans always wanted to defend that part of the population that was suffering. During periods of high prices, they wanted to protect consumers with free imports and an export prohibition. But when prices were low, the protectionists wanted protectionist measures for the agricultural sector. By using this rhetoric, Dumortier hoped to reinstate the sliding scale of 1834.Footnote ⁴⁷ But Prime Minister De Decker pushed for free trade in 1856/57 when the corn prices were still very high, which explains why the free trade law made it through Parliament. In these circumstances, many protectionists could not oppose free imports because of their own rhetoric about defending consumers. In the end, De Decker played a role a bit like Peel's. De Decker was a leading member of the Conservative party, but he had opposed previous protectionist legislation. The bill he brought forth completely split the Catholic and the Liberal parties during the vote. This political confusion generated a large number of abstentions (19.8 percent of all votes) and allowed a significant number of representatives to follow their personal economic interests and the interests of their constituencies.

INTERPRETATION OF THE UNANIMOUS 1873 VOTE

In 1873 the remaining agricultural tariffs were removed. All representatives irrespective of their party affiliation or constituency now voted for free trade, except for three abstentions. Opposition to free trade had been on the wane since the 1850 vote. As Table 5 shows, party affiliation, the strongest explanatory variable in the 1834, 1845, and 1850 votes, lost its force in 1857 and 1873. Abstentions also rose in 1850 and 1857, a sign of growing hesitation on the part of both Liberals and Catholics. The Liberals, who had in general favored trade liberalization, were divided in 1857, but in 1873 they all supported liberalization. The Catholics, who were outright protectionist in 1834 and 1845, began to defect to the free trade side in growing numbers, starting in 1850, and in 1873 all Catholics voted for free trade. During the short debate on the 1873 law in the House of Representatives, the Catholic Minister of Finance Jules Malou claimed that free entry of foodstuffs was now generally accepted. That was in itself a major change as Malou was one of the 21 representatives who had proposed the protectionist “Law of Famine” in 1845 and had voted against the free trade bill of De Decker in 1857. Only one Catholic representative, Léon T'Serstevens, pleaded in favor of agricultural protection. But no one else supported this position, a sign that free trade was now seen as in everybody's interest. Malou replied that “the honorable Mr. T'Serstevens seemed to me to be out of touch as far as our economic system is concerned.”Footnote ⁴⁸

TABLE 5

VOTES BY PARTY

Source:

See footnote 29; and De Paepe and Raindorf-Gerard, Le Parlement belge.

How then can we explain this development towards universal support for free trade in agricultural products? It is tempting to invoke the rising share of the industrial sector in GDP, but industrialization cannot explain the universal preference for free trade by 1873 because industry was geographically concentrated in five or six constituencies, leaving many agricultural constituencies untouched by industrialization until the last quarter of the century. In other districts, industrialization did occur, but it remained limited to small industrial islands in an immense agricultural sea.Footnote ⁴⁹ Nor did agriculture change much between 1850 and 1873: it remained largely traditional until 1880, when agricultural imports and the introduction of new technologies started a very slow process of transformation that was only completed after World War II.Footnote ⁵⁰

What then is the explanation? As many authors have argued in the case of the British repeal, other factors besides economic interests have to be taken into account, and the econometric analysis has to be supplemented with a qualitative analysis of the press and the parliamentary debates. Ideas were important and could affect the votes on agricultural trade liberalization. By 1873 almost every representative, regardless of personal interest, party, and constituency, was in favor of free trade. It is certainly not the case that they were all voting against their own or their constituents' interests.Footnote ⁵¹ It was their perception of these interests that had changed. The representatives came to think that free trade was to their advantage and (more importantly) to the advantage of their constituency. In Belgium, ideas were driving a shift in the perception of interests. That is radically different from Britain where research makes a clear distinction between interests and ideas.

How did Belgian opinion shift? As McLean and Bustani have argued for Britain, we maintain that the subsistence crisis of 1845–1848 played a part in shaking up the old ideas in Belgium. Previously, Belgium, and Flanders in particular, was thought of as an advanced agricultural economy. The American Henry Colman, who visited Europe in the 1840s, was impressed by Flemish agriculture, which he considered to be the best in Europe.Footnote ⁵² Even King Leopold I remarked on the advanced state of agriculture in his 1843 speech before Parliament, two years before the subsistence crisis: “Belgium, so renowned for its agricultural progress.”Footnote ⁵³ But the subsistence crisis changed this perception for many politicians. The censuses of 1846 had the same effect. The year after the census, its architect, Adolphe Quetelet, concluded that Belgian agriculture simply could not feed the growing population.Footnote ⁵⁴

Quetelet's interpretation remained contested for ten more years, but the subsistence crisis sparked a search for solutions.Footnote ⁵⁵ One of the proposed solutions was free trade. But free trade was not a vague idea floating in the air. It was supported by a powerful nation, the United Kingdom, and very active pressure groups. The Repeal of the Corn Laws in the United Kingdom changed the status of political economy. Before repeal, the idea of free trade was considered a theoretical chimera by many Belgian representatives.Footnote ⁵⁶ Once the economic hegemon of the time chose free trade, it seemed the right policy on the way to industrialization for the young Belgian nation. The British repeal inspired a number of Belgian economists backed by industrialists from the Verviers district to start a campaign for free trade.Footnote ⁵⁷ The young engineer and economist Adolphe Le Hardy de Beaulieu formed the Association Belge pour la Liberté Commerciale, which organized meetings that resembled those of the British Anti-Corn Law League. The association was backed by a small, but very vocal, group of representatives to the House. Free trade economists infiltrated the Société Centrale d'Agriculture, the most prominent learned agricultural society, which counted many members of Parliament in its ranks. The economist and journalist Gustave de Molinari commented continuously on Belgian economic policy in his journal L'Économiste Belge (1855–1868).Footnote ⁵⁸ These economists sought to forge coalitions with agriculturalists, industrialists, and the chambers of commerce in order to put pressure on the government.

The opinion that Belgium was confronted with a structural shortage in food production became more difficult to deny for the protectionists when a new subsistence crisis in 1853–1857 again pushed prices to worrying heights and led to food riots.Footnote ⁵⁹ The Société Centrale d'Agriculture, which began supporting free trade in the second half of the 1850s, confirmed the structural character of the shortage in food production in 1854: “A sad truth, a truth full of dangers, is experienced by the people every day. No one contests it, it is officially recognized. What the economists, who were thought to be theorists and dreamers, had dared to proclaim, is now proved by governmental statistics: the growth of food production is not keeping pace with the increasing population.”Footnote ⁶⁰ The most thorough analysis of the situation (with references to the works of British political economists) was written by Edouard Ducpétiaux, a Catholic social scientist and member of the Central Statistical Commission, who estimated the average shortage in wheat production to be about one-tenth of the harvest.Footnote ⁶¹ Faced with this evidence, more and more representatives came to believe that the subsistence crisis had become endemic and that Belgian agriculture did not produce enough food, even in the case of good harvests, to feed the growing population.Footnote ⁶² In 1855 Prime Minister De Decker acknowledged this as a fact.Footnote ⁶³ The defense of protectionist corn tariffs became a difficult task in these circumstances, since free traders could depict tariffs as an unacceptable taxation on an already dear primary product.

The awareness of the structural shortage in corn production, British repeal, the actions of the Belgian economists, and De Decker's strategy all opened the way for free agricultural trade in Belgium. In the 1860s front-rank economists confirmed that free trade was the best policy choice for Belgium. Auguste Orts, a Liberal representative and professor of political economy at the University of Brussels, told his students that “The absolute liberty of trade has brought the results you can observe today.”Footnote ⁶⁴ Emile de Laveleye, professor of political economy at the University of Liège, wrote that economic facts had proved that free trade yielded prosperity. The British example was telling in his mind.Footnote ⁶⁵ The prominence of the British economy was acknowledged in Parliament: “If England, which in the past recognized our superiority, has surpassed us today, this is because for a long time she has rejected the old doctrine of Protection and Subsidies.”Footnote ⁶⁶