-
1 The Evolution of the Electoral System Change Scholarship
1.1 Left Electoral Threat
The classic explanation of electoral system reform in Western Europe is that Conservative and Liberal parties introduced PR as protection from rising Left electoral threat – the possibility of losing elections to Socialist parties (Boix, 1999; Rokkan, 1970). This hypothesis treats parties as unitary actors and cannot explain the Belgian case.
Subsequent scholarship is more nuanced. Blaise et al. (2004) show that PR was introduced because of democratisation and the disproportional effects of existing majoritarian systems, not Left electoral threat. Penadés (2008) elucidates the electoral system preferences of Socialists parties. Calvo (2009) highlights the emergence of multi-party races as suffrage was extended. Notably, Kreuzer (2010) recommends a forward-looking, historically sensitive approach that examines intra-party divisions and cross-party alliances. In response, Boix (2010) re-analysed the evidence from his 1999 article, demonstrating that electoral threat remains important after these factors are considered.1.2 Left Existential Threat
Ahmed (2010, 2013) embeds electoral system change in Western Europe and the United States in the process of democratisation. She argues that PR in Belgium was introduced in response to high Left existential threat, defined as the electoral viability of the Socialist Party and its ideological radicalism, including its propensity to use extra-constitutional measures such as mass demonstrations or revolutionary actions.1x The concept does not distinguish clearly between short- and long-term Left electoral threat and Left extra-institutional threat. The author examines intra-party divisions and cross-party alliances. She observes that most countries transitioned to PR from majoritarian systems with single-member districts (SMDs) and multi-member districts (MMDs), not from single-member district plurality (SMP).2x A single-member district (SMD) is a constituency that elects one member of parliament. A multi-member district (MMD) is a constituency that elects more than one deputy. Single-member district plurality is an electoral system, also known as ‘the first-past-the-post system’, in which the candidate that wins the most votes wins the single seat in that district. Therefore, SMP and PR represent different means of Left containment.
The Belgian case does not fit well in this account. First, Ahmed cites Marks et al.’s (2009) analysis of Left-Party radicalism when operationalising Left existential threat. While Marks et al. classify the Belgian Workers’ Party (BWP) as moderate, Ahmed considers it radical, without explaining the re-classification. However, after universal male suffrage (with compulsory voting and plural voting) was adopted in 1893, the Socialists moderated their position (Liebman, 1979; Polasky, 1992, pp. 454-455; Vandervelde, 1925). Ahmed (2013, p. 169) cites Polasky (1992, p. 452) to show that the Socialist leader Emile Vandervelde was more radical than his predecessor but Polasky (pp. 454-455) writes that Vandervelde was a moderate who, after 1894, embraced parliamentary channels to implement reform. Second, Ahmed examines the failed electoral system reform proposals of 1893, 1894 and January 1899 but excludes the last phase of negotiations, when pure PR was finally adopted (June 1899 to 24 November 1899).1.3 Interaction between Extra-Institutional Threat and Electoral Threat
Barzachka (2014) argues that incumbents’ preferences for electoral systems are shaped by their perceptions of two distinct factors – electoral threat (the possibility of losing elections) and extra-institutional threat to the regime (the possibility of being ousted from power through mass protests, civil unrest, etc.) The article compares seven cases from nineteenth-century Belgium and two cases from post-communist Bulgaria. Focusing on actors’ perceptions of threats allows cross-regional generalisation because different factors could influence perceptions in different contexts. Threats could come from various parties, coalitions or factions, not necessarily from the Socialists.
In Belgium, Catholic office-holders’ perceptions of electoral and extra-institutional threats varied over time, shaping their motives (seat-maximisation or tactical seat loss) and electoral system preferences (majoritarian system with MMDs; majoritarian system with SMDs; mixed systems; or pure PR) during different periods. Deputies from different Catholic factions perceived electoral threat differently and preferred different solutions. The article underscores that Vandenpeereboom’s 1899 mixed system bill, a flagrant seat-maximising endeavour, would have passed, had it not been for the rapidly rising extra-institutional threat from the unified opposition. The Catholics engaged in tactical seat loss – choosing to lose some seats (without relinquishing legislative majority) to stop the civil unrest. They introduced pure PR because they faced high extra-institutional threat and low electoral threat.
This account correctly identifies the levels of extra-institutional threat in all seven cases and the levels of electoral threat in five cases. It correctly emphasises that, in 1899, incumbents traded a relatively small loss in the electoral arena for a substantial gain in the regime transition arena. It operationalises electoral threat as total seats gained by the Catholic Party and relies on aspects of contemporaneous sources that emphasise its dominant position. Consequently, it considers electoral threat to the Catholic Party in 1899 as low. However, analysis of the results of 1896 and 1898 partial elections at the district level shows that the Catholic Party won the large, urban MMDs with a slim majority. Many Catholic leaders feared that the Socialist-Liberal alliances in those districts could cost them future elections (Emmenegger & Walter, 2019).1.4 Back to Electoral Threat
Emmenegger and Walter (2019) contend that PR was adopted in Belgium because the Catholic Party faced increasing electoral threat from Liberal-Socialist cartels in the large urban MMDs. The majoritarian electoral system awarded all seats in the district to the party that eked out a simple majority. Seven districts, ranging from 18 seats (Brussels) to six seats (Leuven and Mons) were endangered. If the Catholics lost several of these districts, they could lose their majority. The account offers robust empirical support to Boix (2010) and fits the growing electoral reform literature on internal party divisions and political geography (Cox et al., 2019; Leemann & Mares, 2014; Schröder & Manow, 2020; Walter & Emmenegger, 2019).
While Emmenegger and Walter highlight the importance of electoral threat, they do not fully consider the role of extra-institutional threat. The authors acknowledge that in June 1899, the government’s mixed electoral system proposal faced significant obstruction by the Liberals and the Socialists in the Chamber (Emmenegger & Walter, 2019, p. 448). However, they do not mention that the opposition organised mass demonstrations in Brussels and other cities to impede the bill. Immediately before the protests started, the Catholic office-holders had enough votes to introduce Vandenpeereboom’s seat-maximising, mixed electoral system (Goblet d’Alviella, 1900, pp. 114-117). The authors do not examine how the extra-institutional threat contributed to the adoption of pure PR in November 1899. Without the extra-institutional pressure, the electoral system reform outcome would have been very different. -
2 Using Theory-Guided Process-Tracing to Evaluate Competing Explanations
Process-tracing represents an important tool for hypothesis/theory testing3x Büthe (2002) argues that process-tracing allows hypothesis generation, but not testing. Gerring (2007) adopts an intermediary position on the issue. in single case studies (Beach, 2016; Beach & Pedersen, 2013; Bennett, 2010; Bennett & Checkel, 2015; Brady & Collier, 2010; Crasnow, 2017; Falleti, 2016; George & Bennett, 2005; Hall, 2003; Kreuzer, 2016; Mahoney, 2012; Tannenwald, 2015; Waldner 2015a; 2015b). Evaluating rival causal explanations can be resolved by process-tracing that is complete, consistent, coherent and theoretically informed.
Process-tracing is the “method [that] attempts to identify the intervening causal process – the causal chain and causal mechanism – between an independent variable (or variables) and the outcome of the dependent variable” (George & Bennett, 2005, p. 206), “the use of evidence within a case to make inferences about causal explanations of that case” (Bennett & Checkel, 2015, p. 4) or “the temporal and causal analysis of the sequences of events that constitute the process of interest”, where the order of events “is causally consequential” (Falleti, 2016, p. 457).
The present analysis relies on inference-based and narrative-based process-tracing.4x See Bates et al. (1998) for analytic narratives, Hall (2003) for systematic process analysis and Kreuzer (2016) for Bayesian process-tracing. Researchers using inference-based process-tracing identify additional events, facts and counterfactuals from the case. Based on this information, they construct hoop tests, smoking-gun tests, straw-in-the-wind tests and/or doubly decisive tests to evaluate competing hypotheses. Hoop tests are necessary but not sufficient conditions for affirming causal inference. Passing a hoop test increases confidence in the hypothesis but does not confirm it. Failing a hoop test eliminates the hypothesis. Straw-in-the-wind tests are neither necessary nor sufficient for affirming causal inference. Passing them slightly bolsters the hypothesis, while failing slightly undermines it (Beach, 2016; Beach & Pedersen, 2013; Bennett, 2010, p. 210; Collier, 2011, p. 825; Mahoney, 2012; 2015; Van Evera, 1997, pp. 31-32; Waldner, 2015a; 2015b). For researchers who use narrative-based process-tracing, a well-constructed narrative, based on a holistic understanding of the case, is better at evaluating competing arguments than inference-based analysis, which uses discrete pieces of evidence (Crasnow, 2017).
Both process-tracing variants examine alternative hypotheses and counterfactual scenarios, derived from theory and/or empirical analyses. They take sequencing seriously and share the standards of completeness, consistency and coherence.2.1 The Process-Tracing Standards: Completeness, Consistency and Coherence
2.1.1 Completeness
Excellent knowledge of the case and the inclusion of all relevant facts are pre-conditions for uncovering causal inferences and testing competing hypotheses (Beach, 2016; Beach & Pedersen, 2013; Bennett, 2010; Bennett & Checkel, 2015; Capoccia & Ziblatt 2010, p. 943; Collier, 2011; Hall, 2012; Mahoney, 2012; Ricks & Liu, 2018; Trampusch & Palier, 2016). Completeness is important because if the narrative begins too early, or ends too late, it risks incorporating irrelevant factors. If it starts late, ends early or leaves out important events, it risks omitting potential causal variables (Bennett, 2010), turning points or counterfactuals (Crasnow, 2017). Furthermore, an incomplete analysis could reach the correct conclusion for the wrong reasons.
Complete inference-based process-tracing must include a causal graph, event-history maps, descriptive inferences and causal inferences underlined by strong causal mechanisms (Waldner, 2015a, pp. 249-250). Complete narrative-based process-tracing must: accurately pinpoint the beginning and the end of the process; include all relevant events (Crasnow, 2017); correctly identify all potential causes; and construct a theoretically informed, generalisable account. The two approaches are complimentary.2.1.2 Consistency
Consistency is the extent to which the causal mechanisms in a hypothesis/theory fits the empirical facts. To ensure consistency, researchers conducting narrative-based process-tracing must look for events that contradict the hypothesis/theory under evaluation or events that should have occurred if it were true but did not (Crasnow, 2017, p. 12). Consistency requires an excellent understanding of the case and commitment to considering all relevant evidence (Crasnow, 2017, p. 12). This recommendation also applies to inference-based process-tracing (Beach, 2016; Mahoney, 2012, 2015; Van Evera, 1997; Waldner, 2015a, 2015b).
2.1.3 Coherence
An account that is complete and consistent is also coherent and more credible. It must incorporate as many elements of the case as possible.
Case studies contain so many potential points of fit – so many details that might not be able to be accounted for – that when they do all fit into an account, our confidence this account is a good one should be increased. (Crasnow, 2017, p. 12)
While this recommendation applies directly to narrative-based process-tracing, it resembles passing many difficult tests in inference-based process-tracing.
2.2 Process-Tracing Procedures for Evaluating Competing Arguments
The present article follows a set of eight consecutive procedures established in the process-tracing literature (Crasnow, 2017; Ricks & Liu, 2018; Trampusch & Palier, 2016; Waldner, 2015a). The sequence of steps begins with identifying the ontological and epistemological approaches of each account and concludes with an examination of their causal mechanisms. Table 1 summarises the sequence of procedures and their application.
The article adopts Waldner’s conceptualisation of strong causal mechanisms or “invariant causal principles that generate the links between events” (Waldner, 2015a, p. 242), as opposed to ‘mechanisms-as-events’ (Crasnow, 2017; Mahoney, 2012, 2015). For example, to explain what causes fire, one should emphasise the principle of combustion, not the striking of the match or the pouring of oil. Strong causal mechanisms are more easily generalisable to a large number of cases (Waldner, 2015a, pp. 224-243).2.3 Constructing a Theoretically Informed Narrative
Electoral system reform is a type of institutional change. Following Renwick (2010) and Falleti (2016), the article identifies several propositions from the institutional change literature that guide the narrative of electoral system reform in Belgium.
The article assumes that there are three necessary but not sufficient conditions for change: an opening for change must emerge (the old institution must be discredited or removed); actors must identify an alternative (Blyth, 1997; Campbell, 2004); and they must have the capacity to adopt it. Systematically thinking how electoral or extra-institutional threats influence these conditions helps determine the importance of each threat and explain why and when it mattered. The processes through which these conditions emerge could develop simultaneously or sequentially and could interact. A serious problem with an existing institution could prompt a search for alternatives. The opening could result from an exogenous shock (Pierson 2004) or emerge gradually (Mahoney & Thelen, 2010, 2015).Table 1 Process-Tracing Procedures for Evaluating Competing ArgumentsProcedures Description Application in Present Article 1. Identify
the Ontology- establish whether the ontological orientation of each account is probabilistic or deterministic
- clarify own ontological assumptions
(Hall, 2003; Trampusch & Palier, 2016)- the three accounts share a deterministic ontology and assume that actors have some power to shape the institutions that constrain them
- the present article adopts a deterministic ontology2. Establish the Epistemological
Approach- specify whether each account is deductive or inductive; many studies combine elements of both approaches (Trampusch & Palier, 2016)
- identify epistemological assumptions
(Trampusch & Palier, 2016)- the three accounts follow a deductive format, yet their arguments are based on a detailed understanding of the case
- the same applies for the present article3. Specify
Testable Hypotheses/
Theories- identify testable theories/hypotheses a priori (Beach, 2016; Crasnow, 2017; Hall, 2012; Mahoney, 2015; Ricks & Liu, 2018; Trampusch & Palier, 2016; Waldner, 2015a)
- consider multiple causal factors working together (Zacks, 2017)- three hypotheses are tested here: Left existential threat
interaction b/w electoral and extra-institutional threat
electoral threat
4. Construct Timelines - create a timeline of all relevant events (Waldner, 2015a)
- compare it to the timelines of rival accounts to find omitted events (Waldner, 2015a)- Appendix I cross-references the available primary and secondary sources
- Appendix II creates a timeline based on the sources in Appendix I
- Appendix II reconstructs the timelines of the rival accounts and compares them to the first timeline5. Write the Narrative - a complete and coherent narrative, is more compelling (Tannenwald, 2015) and credible (Crasnow, 2017)
- the narrative should be theoretically informed (Crasnow, 2017)- the narrative re-examines the historical sources of the three accounts and includes new evidence from the historical press
- the narrative of this article is chronological and structured around insights from the institutional change literature6. Construct Causal Graphs and Event-History Maps - create causal graphs and event-history maps for each argument (Ricks & Liu, 2018; Waldner, 2015a)
- compare the graphs to the timelines and narratives of their accounts and to each other
- if timeline is incomplete, examine how the missing evidence fits with the account- Appendix III compares the causal graphs and event-history maps of each account 7. Identify Alternative Choices, Events and Outcomes - evaluate timelines, narratives, causal graphs and event-history maps to identify counterfactuals (Ricks & Liu, 2018) and create causal mechanism tests (Waldner, 2015a)
- excellent understanding of the case and willingness to consider evidence that could undermine one’s hypothesis are necessary (Crasnow, 2017; Ricks & Liu, 2018)- the two inference-based tests below rely on counterfactuals derived from the narrative:
- explaining the timing of PR adoption
explaining why the 1899 Mixed System Bill was withdrawn8. Examine Causal Mechanisms - for each argument, identify strong causal mechanisms that are generalisable (Waldner, 2015a) - mechanisms include:
- concern about existential threat
- offensive and defensive seat- maximisation, tactical seat loss, ‘playing-it-safe’
- seat-maximisationInstitutional change is contested. The winners from the status quo defend it; the losers seek reform. Actors could have conflicting goals or remain indifferent. The outcome could be unintended (Mahoney & Thelen, 2010, 2015). The article examines how actors’ initial preferences emerge and whether they shift and if so, why and when. It also analyses actors’ coalition-building efforts. Stasis occurs when actors do not believe reform is necessary or disagree about the alternatives. Change happens when the factors promoting reform prevail over those protecting the extant institution.
The narrative considers how the structural, cultural and institutional contexts influence actors (Greener, 2005). Institutions may influence cultural norms or strengthen status quo supporters (Capoccia, 2015). Interests and norms influence actors’ perceptions of how existing institutions function and whether reform is needed. They also influence institutional innovation and selection (Blyth, 1997; Campbell, 2004; Hall 1993, 2010). The article also examines agency – how political entrepreneurs identify problems with extant institutions and propose solutions (Campbell 2004); how they use discourse to develop and disseminate their ideas and build coalitions to remove or alter status quo constraints (Schmidt 2010); and how their opponents do the same. -
3 The Narrative: Electoral System Reform in Belgium
3.1 The Institutional Status Quo
Originally, Belgium’s two-ballot majoritarian system (MR) was used in 32 MMDs and 9 SMDs. The 1831 Constitution contained a provision that one deputy must represent every 40,000 inhabitants. Consequently, the more populous districts elected more than one representative. Elections for half of the Chamber of Representatives took place every two years. The franchise was highly restricted and inscribed in Article 47 of the Constitution. This was a major obstacle to suffrage reform because the Constitution could only be revised by a 2/3 majority in a Constituent Assembly (Barthélemy, 1912; Delfosse, 2004, p. 184; Gilissen, 1958, pp. 91, 123; Mahaim, 1900, p. 82; Pilet, 2007, p. 23; Stengers, 2004, pp. 256-257).
3.2 Actors
From the 1830s to the 1880s, Belgium had a two-party system. The Liberal Party, which espoused secularism and represented wealthy industrialists, was strong in the urban centres of Wallonia, while the Catholic Party was strong in rural Flanders. Nevertheless, the Catholics also obtained seats in some urban Walloon districts and Brussels. The Liberals had strongholds in the Flemish cities of Ghent and Antwerp. Until the 1880s, the electoral system allowed both parties access to office. The Liberals were in power from 1847 to 1854; 1857 to 1870; 1878 to 1884. The Catholics governed between 1856 and 57 and 1870 and 1878 (Devresse, 1990, p. 27; Gilissen, 1958, pp. 102-117).
Both parties were internally divided. The Liberal Party included doctrinaires, who opposed franchise extension, and progressives/radicals, who supported it. The Catholic Party had a conservative/reactionary faction, which supported the Church’s active involvement in politics and society and a moderate bloc. The BWP (the Socialists) emerged in the 1880s, entering parliament in 1894 (Collier, 1999, p. 91; Devresse, 1990, p. 39-45; Gilissen, 1958, pp. 102-117; Linden, 1920, p. 269).3.3 No Opening for Change, No Clear Alternative (1860s-1870s)
Initially, PR was a novel, elitist concept endorsed by a few forward-thinking Liberal and Catholic politicians, including Count Goblet d’Alviella, a Liberal. Early attempts to introduce PR (in 1866, 1871 and 1878) failed because there was no opening for institutional change. Most politicians expected to benefit from the vagaries of the existing electoral system and opposed reform (Goblet d’Alviella, 1900). Furthermore, they had a poor understanding of the alternative – PR (Hare). Some impetus for change came when, in 1878, Victor d’Hondt developed a new method for transferring votes into seats that was easier to understand than Hare’s (Barthélemy, 1912, pp. 533-535). However, it was not until problems with the current electoral institutions became obvious that PR was seriously considered.
3.4 No Opening for Change, Alternative Clarified (1880s)
During the 1880s, the constraints of existing electoral institutions gradually began to erode but did not disappear. The forces of institutional reproduction (seat-maximisation under the status quo system and limited understanding of alternatives) were stronger than the forces of institutional transformation (increasingly disproportional results, the emergence of a Socialist Party and electoral system innovation).
At the time, Belgium was the fastest industrialising country in the world. Due to population shifts, the district magnitude of the industrial centres in Wallonia increased, according to the aforementioned constitutional provision (Barthélemy, 1912, p. 513). The overall effect was the overrepresentation of the small rural Flemish districts and the underrepresentation of the large urban constituencies. The system was beginning to favour the Catholics, though both parties hoped to use it to their advantage (Barthélemy, 1912, p. 534; Cameau, 1901 pp. 51-53; Dupriez, 1901, p. 157; Goblet d’Alviella, 1900, pp. 43-60; Van der Smissen, 1920, p. 342; Woeste, 1927, p. 369).
Progressive Liberals and moderate Catholics who suffered losses under the status quo wanted PR but did not have majority to implement reform (Goblet d’Alviella, 1900, pp. 43-60). Instead, they focused on simplifying and disseminating the idea. The Reformist Belgian Association for Proportional Representation (RBAPR), established in 1881, issued brochures and organised lectures and mock elections. In 1888, RBAPR unsuccessfully proposed a PR bill based on a refined version of D’Hondt’s method (Barthélemy, 1912, pp. 533-535; Goblet d’Alviella, 1900, pp. 49-60). A clear alternative that would facilitate future change emerged.
Industrialisation also undermined the status quo by creating a large working class and a Socialist Party without legislative representation. The Socialists adopted an extra-institutional strategy for the improvement of working conditions, wage increases and universal male suffrage. The 1886 mass workers’ protests precipitated the 1887 unsuccessful constitutional reform proposal of the progressive Liberals. Yet, by the time PR (d’Hondt) was discussed in the Chamber in 1888, the extra-institutional threat had abated. The bill failed because the majority of both parties and even Prime Minister (PM) Beernaert, a moderate Catholic and a founding member of RBAPR, did not think PR was necessary5x Annales Parlementaires: Séance du 24 janvier, 1888. Bruxelles, Belgique: Chambre de Représentants, p. 423. (Liebman, 1979 p. 83; Vandervelde, 1925, pp. 37-41). Most actors did not find problems with the status quo and had limited interest in reform. It was not until the Socialists started to demonstrate again for universal male suffrage that an opportunity for change appeared.3.5 Extra-Institutional Threat Destabilises the Status Quo, No Capacity to Implement PR (1890-1893)
Between August 1890 and February1892, the workers mobilised three times to demand the right to vote. The strikes and demonstrations garnered between 80,000 and 200,000 participants each time (Devresse, 1990, p. 44; Gilissen, 1958, p. 21; Liebman, 1979, p. 86; Polasky, 1992, pp. 454-455; Strikwerda, 1997, pp. 91-95). The progressive faction of the Liberal Party, led by Paul Janson, a long-time universal male suffrage advocate, called for constitutional reform. While the reactionary Catholics and doctrinaire Liberals opposed universal male suffrage, some moderate Catholics had started to believe that it might be necessary to appease the Left. Beernaert managed to persuade the king and moderate Catholics to reform the Constitution. After the February 1892 miner’s strike, the Chamber voted to take up the constitutional reform bill sponsored by Janson (Goblet d’Alviella, 1900, pp. 63-66; Van der Smissen, 1898, p. 559; Vandervelde, 1925, p. 45).
The moderate Catholics needed to placate the Socialists and defend their seats in the large urban districts. Consequently, Beernaert proposed the following measures: universal male suffrage; plural voting (which gave up to two additional votes to wealthier and more educated men); compulsory voting; PR (d’Hondt). The last three measures were electoral threat safeguards (Frère-Orban, 1895; Goblet d’Alviella, 1900, pp. 64; Mahaim, 1900, p. 387; Vandervelde, 1925, p. 43).
The moderate Catholics were concerned about high electoral threat to the party and their faction/themselves (consistent with all three accounts). They also worried about the immediate high extra-institutional threat from the Left (consistent with the Left threat and the interaction accounts). The fact that the Constituent Assembly passed the reform only after the Socialist held another mass strike in the spring of 1893 is further evidence of that (Cameau, 1901, pp. 107-8; Goblet d’Alviella, 1900, pp. 64-69; Mommaert, 1900, pp. 88-91; Van der Smissen, 1898, pp. 563).
The first three elements of Beernaert’s proposal were adopted, but PR was so controversial that the PM withdrew it. The reactionary Catholics from the small rural Flemish districts were not directly threatened by Liberal-Socialist alliances. Both reactionary Catholics and doctrinaire Liberals thought PR was too complex and unnecessary and disapproved of Beernaert for allowing the adoption of universal male suffrage. Beernaert and de Smet de Naeyer, another moderate Catholic and RBAPR member, prevented the inclusion of the electoral system in the constitution, which would have become a major obstacle to future reform (Cameau, 1901, p. 108; Dupriez, 1901, p. 160; Frère-Orban, 1895; Goblet d’Alviella, 1900, pp. 66-69; Van der Smissen, 1898, p. 562; Van der Smissen, 1920, p. 377).3.6 Disagreements over Alternatives, No Capacity to Implement Preferences (1894)
Uncertainties about the effects of the new electoral rules prompted Beernaert to attempt reform again before the 1894 legislative elections. His bill for a mixed system (PR in the 32 MMDs and the extant majoritarian system in the 9 SMDs) was opposed by Charles Woeste, a prominent reactionary Catholic, who disliked Beernaert and proposed a different solution – dividing the MMDs into SMDs. Woeste (from Aalst, a Flemish four-member district6x Session de 1894-1895: Liste des membres de la Chambre, 1894. Bruxelles, Belgique: Chambre de Représentants.) and other representatives from the small, rural, Flemish constituencies preferred redistricting into SMDs because they did not face electoral threat from Liberal-Socialist cartels (Woeste, 1898, pp. 18-19) and because SMDs seemed more straightforward (Van der Smissen, 1920, pp. 354-355). Elected in Tielt, a two-member constituency in Western Flanders, Beernaert, nevertheless, was concerned about Liberal-Socialist cartels in the large urban MMDs (Cameau, 1901, pp. 110-111; Goblet d’Alviella, 1900, pp. 69-74). Finally, many representatives who expected to continue winning under the existing system did not want change (Cameau, 1901, pp. 110-111; Dupriez, 1901, p. 162; Goblet d’Alviella, 1900, p. 74). Beernaert resigned after his bill failed (Barthélemy, 1912, p. 358; Goblet d’Alviella, 1900, p. 77). Two other PR initiatives were rejected for similar reasons (Goblet d’Alviella, 1900, pp. 78-83).
The historical record shows that office-holders were not concerned about Left extra-institutional threat at this time. Once in the legislature, the Socialists were satisfied with the constitutional revision, moderated their position and, having expected success under universal male suffrage, vowed to continue working for the abolition of plural voting through parliamentary channels. The extra-institutional threat remained low until June 1899 (Collier, 1999; Liebman, 1979; Polasky, 1992, pp. 454-455; Vandervelde, 1925).
The analysis is consistent with the explanations that underscore electoral threat and the interaction between electoral and extra-institutional threats. The 1894 reform is inconsistent with Ahmed’s (2013) argument. The author does not mention the mixed electoral system and argues that Left existential threat was high because the Belgian Socialists were electorally viable and ideologically radical. While the evidence supports the electoral viability claim, it contradicts the claim about their ideological radicalism (see Marks et al., 2009).
In 1895, a mixed system (including PR) was adopted for local elections (Goblet d’Alviella, 1900, pp. 89-92). It took another four years before electoral system reform at the national level was introduced in 1899. This hiatus provides the first plausible counterfactual – that if Left existential threat had remained high, PR would have been adopted sooner (see Test 1).3.7 The Status Quo Becomes Unsustainable, No Agreement on the Alternative (1894-1899)
The combined effects of plural voting, compulsory voting and a majoritarian electoral system with districts of various sizes protected the Catholic Party from the Left electoral threat in the first elections under universal male suffrage in 1894. The Catholics gained 12 seats (from 92 in 1892 to 104 in 1894). Their seat share increased to 111 in 1896, and to 112 in 1898. The Socialists won 28 seats in 1894 and 1896 and 27 in 1898. The Liberal Party went to the brink of extinction from 60 seats in 1892 to 20 in 1894 and 13 in 1896/8 (Stengers, 2000, pp. 135-136, 2004, pp. 259-260). The aggregate results seem to indicate that the electoral threat to the Catholic Party was low.
Examining district-level results, however, reveals a more complex picture. The total Catholic vote share was declining. Catholics in the small MMDs and the SMDs in rural Flanders, indeed, faced low electoral threat from the Socialist-Liberal cartels. Nevertheless, electoral threat was rising for Catholics in the large, urban, MMDs, who competed against Socialists-Liberal cartels (Goblet d’Alviella, 1900, pp. 84-85, pp. 92-93, pp. 96-99; Mahaim, 1900; Stengers, 2000, pp. 136-137, 2004, pp. 259-260). They were especially vulnerable in Brussels (18 seats), Antwerp (11), Liège (11), Ghent (9), Charleroi (8), Leuven (6) and Mons (6). When discussing the motives behind the 1899 mixed system proposal, Barthélemy explains that, “in five of these districts, the anti-clericals were sure, by their union, to take the total number of seats and the same result was probable in the other two” (Barthélemy, 1912, p. 540).
In 1894, the Socialists won all of the seats in Charleroi and Mons. A cartel won Liège, where six seats went to the Socialist and five to the Liberals (Stengers, 2000, p. 135). The 1896 election in Brussels was very close.[O]ne could have calculated that without the support at the second ballot of some thousand or so conservative liberals, having together fifteen thousand votes for the Catholic list, and the blanc vote of many thousands of others, the Catholics would have been defeated. (Stengers, 2000, p. 137)
An upcoming census was expected to increase the size of the MMDs. Furthermore, many Catholic politicians were concerned about the future of the Liberal Party and did not want to see the Socialists as their only opposition. While in 1896, the moderate Catholics won Brussels and other key MMDs, their future control of these constituencies was uncertain (Stengers, 2000, pp. 136-137).
The highly disproportional results outraged the Liberals and the Socialists, worried moderate Catholics and delegitimised the electoral system. An opening for change appeared.
Several solutions were discussed before pure PR. During these deliberations, actors’ preferences aligned with their varying perceptions of electoral threat and their beliefs about the applicability of each system to Belgium (Cameau, 1901, pp. 118-119; Dupriez, 1901, pp. 163-164; Goblet d’Alviella, 1900, pp. 94-106; Stengers, 2000, p. 137; Woeste, 1933, pp. 153-160).3.8 Alternative I: Redistricting into SMDs (January 1899)
In January 1899, Charles Woeste, an influential reactionary Catholic and outspoken supporter of redistricting into SMDs convinced the king to support this measure. The king issued a statement that SMDs and PR are equivalent alternatives to the status quo. PM de Smet de Naeyer, who preferred PR at the run-off elections and agreed to divide the Brussels constituency in two, but strongly opposed SMDs, resigned (Dupriez, 1901, p. 164; Goblet d’Alviella, 1900, pp. 104-106; Woeste, 1933, pp. 153-160). Vandenpeereboom, a reactionary Catholic, became PM in January 1899. Woeste continued to unsuccessfully advocate for SMDs, antagonising much of the Chamber and his party (Goblet d’Alviella, 1900, pp. 106-108; Woeste, 1933, pp. 161-163).
Interestingly, between 1896 and 1899, Woeste did not seem worried that the cartels would threaten the Catholic majority. He thought that the Socialists and the Liberals have conflicting economic interests (Woeste, 1933, p. 138, 1898, p. 8) and believed that the two-ballot majoritarian system protected the Catholics (Woeste, 1933, pp. 87-89, pp. 141). Woeste wanted to preserve it, while introducing SMDs. According to him,The king did not appear reassured on the subject of the 1900 elections. This last fear seemed exaggerated; the status quo was not as menaced as one said. It did not seem plausible that we would lose at the same time Brussels, Nivelles, Antwerp and Philippeville. (Woeste, 1933, p. 155)
Woeste thought that PR would benefit the Christian-Democratic wing, led by Daens who challenged him in Aalst (Woeste, pp. 16, 40, 68, 84-88).
Neither group could implement its preference. However, each succeeded in impeding the adoption of the reform their opponents preferred. The stalemate deepened, the status quo persisted, but the problem with the extant system was not resolved.3.9 Alternative II: Vandenpeereboom’s Mixed System Bill (April to June 1899)
Vandenpeereboom consulted the Catholic parliamentary group. Beernaert and his allies wanted PR in the urban MMDs; Woeste and his clique demanded SMDs; many were undecided and some did not want change. Vandenpeereboom met separately with each faction before unveiling, in April 1899, a mixed system bill. The bill introduced PR in only seven large MMDs, divided the two-member districts into SMDs, and made no change in medium-sized MMDs (Goblet d’Alviella, 1900, p. 110; Woeste, 1933, p. 162-163). The Socialists and both progressive and doctrinaire Liberals understood that the reform would strengthen Catholic dominance and vehemently opposed it (Barthélemy, 1912, pp. 539-540; Goblet d’Alviella, 1900, pp. 111-114; Mahaim, 1900, p. 83; Vandervelde, 1925, p. 51).
Yet, the warring Catholic factions had reached a compromise. Even moderates like Beernaert and de Smet de Naeyer supported the proposal. On June 27, a procedural vote to set the discussion date (88 for; 16 against; and 11 abstentions) indicated that the bill would pass7x Annales Parlementaires: Séance du 27 juin, 1899. Bruxelles, Belgique: Chambre de Représentants. (Goblet d’Alviella, 1900, pp. 114-117). Incumbents were finally capable of implementing a solution.
The parliamentary transcripts of June 27-29 describe vivid scenes of opposition representatives hurling colourful insults, playing musical instruments, throwing paper balls at the Catholics and even fighting8x Annales Parlementaires: Séance du 27 juin, 1899. Bruxelles, Belgique: Chambre de Représentants; Annales Parlementaires: Séance du 28 juin, 1899. Bruxelles, Belgique: Chambre de Représentants; Annales Parlementaires: Séance du 29 juin 1899. Bruxelles, Belgique: Chambre de Représentants. (Barthélemy, 1912, p. 541; Goblet d’Alviella, 1900, p. 114-117; Mahaim, 1900, p. 84; Vandervelde, 1925, p. 52). Emmenegger and Walter (2019) argue that this is why the bill failed, without mentioning the strong extra-institutional pressure outside the Chamber.
Historical press accounts9x They provide a contemporaneous view from a different perspective and new evidence, against which the well-known historical sources and the recent explanations could be compared. show that the obstruction inside parliament would not have stopped the passage of the bill, without coordinated extra-institutional pressure from the streets – mass demonstrations, demanding equal universal male suffrage. The extra-constitutional threat escalated dramatically between June 28 and July 1. The government was prepared for this, having called the national guard and the gendarmerie to Brussels and established a security perimeter around the legislature in late June.10x Annales Parlementaires: Séance du 27 juin, 1899. Le Courrier de Bruxelles, 27 June, 1899, p. 1; Le Peuple, 27 June, 1899, La zone néutre en état de siege, p. 1; Le Peuple, 29 June, 1899, Interpellations sur les brutalites de la Gendarmerie: Situation révolutionnaire, p. 1.
Parliamentary transcripts reveal that the Socialists deliberately used the extra-institutional threat to block the reform. For instance, on 29 June, Defnet, a Socialist, proclaimed:Either you withdraw your bill or you are going to create riots in Brussels and all of our big cities…It is time: withdraw your bill! If not, the duel between you and the people will intensify further. The demonstrations will continue and become more numerous.11x Annales Parlementaires: Séance du 29 juin 1899. Bruxelles, Belgique: Chambre de Représentants.
After the legislative debates ended, the Socialists and Liberals joined the crowds outside. Mounted police were dispatched and several people were wounded, as the country came close to a civil war (Barthélemy, 1912, p. 540-541; Goblet d’Alviella, 1900, pp. 117-118). “[I]n Brussels, it came to a real riot, with, in the center of the city, the beginning of barricades, and gardarmes charging with sabers drawn” (Stengers, 2000, p. 137).
On 30 June 1899, Le Peuple (the daily Socialist newspaper) described the previous night in Brussels as “bloody” and “revolutionary”.12x Le Peuple, 30 June, 1899. Interpellations sur les brutalites de la Gendarmerie, p. 1. Le Courrier de Bruxelles (a major Catholic Party newspaper) suggests that office-holders felt seriously threatened:The real nature of the movement led by the Socialists became obvious last night. It cannot be denied anymore: it is not a political, legal, constitutional movement, but a revolutionary push that would blow away all of our institutions.13x Le Courrier de Bruxelles, June 30, 1899. p. 1.
Le Peuple’s third special edition for July 1 shows that incumbents began to moderate their positions in direct response to the revolutionary threat:
At 3:15 am, it was announced to us that the Chamber had suspended its public meeting to deliberate in secret on the situation, Mr. Vandenpeereboom has made a declaration, allowing us to believe [in] the motives of appeasement. At this moment, the Left deliberates on the conditions of peace.14x Le Peuple, 1 July, 1899. Du sang! p. 1.
While politicians and party newspapers are likely to exaggerate the blame of their opponents and minimise their own, triangulating primary and secondary sources reveals that the spiking extra-institutional threat began shifting actors’ preferences.
Vandenpeereboom agreed to suspend the bill and consider amendments. The Chamber approved. While some Socialists wanted to continue protesting, Vandervelde and six other Socialists asked their supporters to stop (Barthélemy, 1912, p. 541; Goblet d’Alviella, 1900, pp. 118-120). Yet, an alternative was not found and the opening for change remained. This episode provides the second counterfactual, the possibility that without extra-institutional threat, the mixed electoral system bill would have been adopted (Test 2).3.10 Alternative III: Pure PR Adopted (July to November 1899)
The road to pure PR in Belgium was convoluted. Proponents of different electoral systems fought each other and status quo defenders, and sought support (or at least abstention) from their undecided colleagues.
In July, the government created a 16-member commission (including Beernaert, de Smet de Naeyer, Woeste and Vandervelde) to examine possible solutions. The commission debated and rejected four alternatives: redistricting into two-member districts (13 against, Woeste abstained); the Socialist proposal for an electoral reform referendum (11 against, 5 for); pure PR (8 against, 7 for, 1 abstention) and Vandenpeereboom’s mixed system bill (8 against, 8 abstentions). No agreement was possible and Vandenpeereboom stepped down (Goblet d’Alviella, 1900, pp. 122-126; Woeste 1933, pp. 169-197).
De Smet de Naeyer became PM again and introduced a bill for pure PR (d’Hondt) (Goblet d’Alviella, 1900, p. 126; Woeste, 1933, p. 171). The Liberals supported PR, but the Socialists demanded an electoral system referendum and – if not – the abolition of plural voting (equal universal male suffrage). Vandervelde unsuccessfully sponsored a bill to revise Art. 47 of the Constitution. Some Socialists (Vandervelde, Bertrand, Heupgen) supported PR, if accompanied by equal universal male suffrage. These demands and incumbents’ memories of past protests heightened their views of extra-institutional threat. Many other Socialists (e.g. Defuisseaux) opposed PR for seat-maximisation reasons (Barthélemy, 1912, p. 544; Goblet d’Alviella, 1900, pp. 129-134; Woeste 1933, pp. 172-173).
The Chamber debated the government’s bill between 12 September and 24 November. Catholic deputies from the small MMDs, protected their constituencies from redistricting. SMD supporters in the Senate exerted pressure. The Socialist proposed another referendum. Several redistricting proposals were rejected. Albert Nyssens, a Catholic PR proponent, sponsored a quorum amendment to the government’s bill that garnered support from some radical Socialists and reactionary Catholics, including Woeste. The radical Socialists wanted to eliminate the need for Liberal-Socialist cartels in the urban MMDs, the conservatives – to derail PR. The adoption of Nyssens’ proposal would have undermined the government bill and deepened the stalemate. The debates in the Chamber were heated. Schollaert, himself a Catholic PR opponent, asked Nyssens “to withdraw his proposition, so as to not help the opposition open a crisis from which no one saw an exit” (Goblet d’Alviella, 1900, pp. 138). Nyssens consented (Goblet d’Alviella, 1900, pp. 138).
The Chamber passed the bill for pure PR for national legislative elections in Belgium on 24 November by a slim majority (70 for, 63 against and 8 abstentions). Its supporters were 5 Liberals and 65 Catholics, including previous PR opponents who voted in favour because they thought it was the only feasible option, preferable to the abolition of plural voting, or a further escalation of the crisis. This preference change highlights the importance of the elevated extra-institutional threat. And while pure PR meant that the Catholic Party would relinquish some seats to the opposition, the party would not lose power. The Senate approved the bill on 22 December (61 for, 26 against and 6 abstentions). The law was promulgated on 30 December (Goblet d’Alviella, 1900, pp. 138-140).15x Annales Parlementaires: Séance du 24 novembre, 1899. Bruxelles, Belgique: Chambre de Représentants, pp. 51-67. -
4 Using the Narrative to Assess the Relative Importance of Threats
The theoretically informed narrative shows how and when perceptions of extra-institutional threat and electoral threat influenced the adoption of PR in Belgium.
The high extra-institutional threat created an opening for change in 1893. Yet, because that threat subsided by 1894 and because incumbents were not able to implement their preference, the 1894 reform failed. Increasing extra-institutional threat blocked Vandenpeereboom’s mixed bill in the summer of 1899. In the fall of 1899, many Catholic politicians were concerned about the possibility of another escalation. The potential crisis caused many Catholics to change their preferences to full PR in November, according to the logic of tactical seat loss.
The electoral threat and the growing disproportionality16x Leemann and Mares (2014). helped open a window for change in 1893, and in 1899. They also shaped the initial preferences of actors and blocked Woeste’s 1899 SMD bill when de Smet de Naeyer resigned. -
5 Inference-Based Testing
5.1 Test 1 (Straw-in-the-Wind): Does the Account Explain the Timing of PR Adoption?
5.1.1 Left Existential Threat
If left existential threat had been constantly high between 1893 and 1899 because the Socialists were ideologically radical, PR for national elections could have been adopted earlier during this period. The empirical evidence weakens the thesis (Fig. 3, Appendix III).
5.1.2 Interaction between Threats
In this explanation, between 1894 and 1899, incumbents viewed both the extra-institutional and electoral threats as low. The extra-institutional threat was indeed low until June 1899, consistent with the timing of PR adoption (Fig. 11 and Fig. 13, Appendix III).
The account correctly identifies electoral threat perceptions in the case of five of the seven reform proposals. Yet, the 1896 and 1898 elections became a source of concern, especially for the moderate Catholics. The 1899 mixed system proposal and the pure PR bill were motivated by perceptions of increasing electoral threat, which partly weakens the explanation (Fig. 11 and Fig. 13, Appendix III).5.1.3 Electoral Threat
The electoral threat from Socialist-Liberal alliances increased during the 1894-1899 period but electoral reform was not proposed sooner because the Catholics preferred different alternatives and could not agree on a solution. The 1896 and 1898 partial elections impacted the reactionary faction (from small rural Flemish constituencies) and the moderate faction (from large urban MMDs) of the Catholic Party differently (Emmenegger and Walter, 2019, pp. 443-444). The test supports the narrative of the electoral threat account (Fig. 18, Appendix III).
5.2 Test 2 (Hoop): Does the Account Explain Why Vandenpeereboom’s 1899 Mixed System Bill, Which Had Enough Legislative Support, Was Withdrawn?
5.2.1 Left Existential Threat
The bill was blocked by mass protests. Had the episode been included, it would have bolstered the argument but is missing from the account (Fig. 4, Appendix III).
5.2.2 Interaction between Electoral and Extra-Institutional Threats
The account examines the episode. It shows that the bill would have passed when there was no extra-institutional threat, but was withdrawn after the extra-institutional threat escalated. The protests added a new dimension to the negotiations, creating a two-level game and encouraging tactical seat loss. The account passes the test (Fig. 13, Appendix III).
5.2.3 Electoral Threat
Evidence that elevated extra-institutional threat in June 1899 stopped the passage of Vandenpeereboom’s mixed system bill is inconsistent with this account. Elevated electoral threat did motivate many Catholic deputies who agreed to vote for the proposal. If electoral considerations had been the only cause, the mixed system would have been adopted because the bill had garnered enough votes. The authors argue that the bill was rejected because “both of the Catholics’ wings still believed their preferred option to be possible” (Emmenegger & Walter, 2019, p. 448) and omit the extra-institutional threat (Fig. 22 and Fig. 24, Appendix III).
-
6 Conclusion
This article makes several contributions to the literature on electoral system reform. Methodologically, it uses theoretically informed process-tracing to evaluate rival explanations of the introduction of PR in Belgium. It demonstrates that the narrative-based and the inference-based approaches to process-tracing are important and complementary tools for hypothesis/theory testing. When applied systematically, both approaches can help evaluate competing explanations. All three accounts examined here (Ahmed, 2013; Barzachka 2014; Emmenegger & Walter, 2019) consider intra-party divisions and Liberal-Socialist cooperation. The Left existential threat and the interaction account underscore the context of democratisation and the extra-constitutional dimension. The most recent account highlights the importance of electoral threat and political geography. This article acknowledges and builds on to the strengths and contributions of this literature.
Yet, the article also finds that all three accounts have weaknesses. The Left existential threat explanation of the Belgian case is incomplete (it excludes the period of June 1899 to November 1899) and inconsistent (the Left existential threat decreased between 1894 and June 1899, which contradicts the theory). These problems undermine the coherence of the account. The interaction account is complete and consistent in five of the seven cases. It is inconsistent regarding electoral threat in the two 1899 episodes, but consistent regarding extra-institutional threat in all episodes. The electoral threat account is correct and consistent when it comes to electoral threat from Socialist-Liberal cartels (which explains office-holders’ preferences). However, it is incomplete because it excludes the extra-institutional dimension in 1899 (and cannot accurately explain why the 1899 mixed system bill was rejected). The account’s incompleteness undermines its coherence.
Empirically, this article finds that both electoral threat and extra-institutional threat mattered for the introduction of PR in Belgium. It accomplishes that by using theoretically informed process-tracing, focusing on event-sequencing and examining how events unfolded in the summer and fall of 1899, both inside and outside of the parliamentary arena. In the late 1890s, many (but not all) Catholic politicians thought that the electoral threat to their party was increasing because of the high uncertainty surrounding elections in the large urban MMDs. This explains why Vandenpeereboom’s bill introduced PR only in those contested districts. Office-holders’ perceptions of electoral threat made them consider electoral system reform, thereby creating an opening for change. The article confirms the importance of extra-institutional threat in blocking the controversial proposal and in pushing incumbents to change preferences towards pure PR in the fall of 1899. It shows that tactical seat loss could also occur if the perceived electoral threat is rising, uncertainty about the outcome of elections is high and extra-institutional pressure endangers regime stability and incumbent status. Additional tests, derived from theory, new interpretations of historical sources and/or new evidence could confirm or undermine these findings. The bibliography and Appendix I aim to increase data transparency (Moravcsik, 2014) and facilitate such research.
The third contribution of the article is theoretical – integrating the literatures on institutional change and electoral system reform allows a systematic analysis of the complete process (the series of attempted and implemented reforms) that led to PR adoption in Belgium. Future research should continue to explore the connections between the two literatures as this could provide new insights about institutional change and electoral reform in other contexts, including electoral system change during democratisation in post-communist Central and Eastern Europe, Latin America, Africa and Asia. References Ahmed, A. (2010). Reading History Forward: The Origins of Electoral Systems in European Democracies. Comparative Political Studies, 43(8/9), 1059-1088. https://doi.org/10.1177/00104140103704.
Ahmed, A. (2013). Democracy and the Politics of Electoral System Choice: Engineering Electoral Dominance. Cambridge University Press.
Barthélemy, J. (1912). L’organisation du suffrage et l’experience belge. M. Giard et E. Brière.
Barzachka, N. S. (2014). When Winning Seats Is Not Everything: Tactical Seat-Loss During Democratization. Comparative Politics, 46(2), 209-229. https://doi.org/10.5129/001041514809387351.
Bates, R. H., Greif, A., Levi, M., Rosenthal, J. -L. & Weingast, B. R. (1998). Analytic Narratives. Princeton University Press.
Beach, D. (2016). It’s All About Causal Mechanisms—What Process-Tracing Case Studies Should Be Tracing. New Political Economy, 21(5), 463-472. https://doi.org/10.1080/13563467.2015.1134466
Beach, D. & Pedersen, R. B. (2013). Process-Tracing Methods. University of Michigan Press.
Bennett, A. (2010). Process Tracing and Causal Inference. In H. Brady & D. Collier (Eds.), Rethinking Social Inquiry: Diverse Tools, Shared Standards. Rowen & Littlefield Publishers, Inc.
Bennett, A. & Checkel, J. T. (2015). Process-Tracing: From Philosophical Roots to Best Practices. In A. Bennett & J. T. Checkel (Eds.), Process-Tracing: From Method to Analytic Tool (pp. 3-37). Cambridge University Press.
Blaise, A., Dobrzynska, A. & Indridason, I. (2004). To Adopt or Not to Adopt Proportional Representation: The Politics of Institutional Choice. British Journal of Political Science, 35(1), 182-190. https://doi.org/10.1017/S0007123405000098.
Blyth, M. (1997). ‘Any More Bright Ideas?’: The Ideational Turn of Comparative Political Economy. Comparative Politics, 30, 229-250. https://doi.org/10.2307/422082.
Boix, C. (1999). Setting the Rules of the Game: The Choice of Electoral Systems in Advanced Democracies. American Political Science Review, 93(3), 609-624. https://doi.org/10.2307/2585577.
Boix, C. (2010). Electoral Markets, Party Strategies, and Proportional Representation. American Political Science Review, 104(2), 404-413. https://doi.org/10.1017/S0003055410000146.
Brady, H. E. & Collier, D. (Eds.). (2010). Rethinking Social Inquiry: Diverse Tools, Shared Standards. Rowman & Littlefield Publishers.
Büthe, T. (2002). Taking Temporality Seriously: Modeling History and the Use of Narratives as Evidence. American Political Science Review, 96(3), 481-493.
Calvo, E. (2009). The Competitive Road to Proportional Representation: Partisan Biases and Electoral Regime Change Under Increasing Party Competition. World Politics, 61(2), 254-295. https://doi.org/10.1017/S0043887109000100.
Cameau, P. (1901). La représentation proportionelle en Belgique. Librairie nouvelle de droit et de jurisprudence.
Campbell, J. (2004). Institutional Change and Globalization. Princeton University Press.
Capoccia, G. (2015). Critical Junctures and Institutional Change. In J. Mahoney & K. Thelen (Eds.), Advances in Comparative-Historical Analysis (pp. 147-179). Cambridge University Press.
Capoccia, G. & Ziblatt, D. (2010). The Historical Turn in Democratization Studies: A New Research Agenda for Europe and Beyond. Comparative Political Studies, 43, 931-968. https://doi.org/10.1177/0010414010370.
Collier, D. (2011). Understanding Process Tracing. PS: Political Science and Politics, 44(4), 823-830. https://doi.org/10.1017/S1049096511001429.
Collier, R. B. (1999). Paths Toward Democracy. Cambridge University Press.
Cox, G., Fiva, J. H. & Smith, D. M. (2019). Parties, Legislators, and the Origins of Proportional Representation. Comparative Political Studies, 52(1), 102-132. https://doi.org/10.1177/0010414018762369.
Crasnow, S. (2017). Process Tracing in Political Science: What’s the Story? Studies in History and Philosophy of Science, 62, 6-13. https://doi.org/10.1016/j.shpsa.2017.03.002.
Delfosse, P. (2004). Les changements de mode de scrutin en Belgique. In A. Lauren, P. Delfosse & A.-P. Frognier (Eds.), Les systems électoraux: permancence et innovations (pp. 151-184). L’Harmattan.
Devresse, D. (1990). Belgium. In M. Van Der Linden & J. Rojahn (Eds.), The Formation of Labour Movements 1870-1914. E. J. Brill.
Dupriez, L. (1901). L’organisation de suffrage en Belgique. Librarie de la société du recueil general des lois et des arrèts.
Emmenegger P. & Walter, A. (2019). When Dominant Parties Adopt Proportional Representation: The Mysterious Case of Belgium. European Political Science Review, 11, 433-450. https://doi.org/10.1017/S1755773919000225.
Falleti, T. G. (2016). Process Tracing of Extensive and Intensive Processes. New Political Economy, 21(5), 455-462. https://doi.org/10.1080/13563467.2015.1135550.
Frère-Orban, W. (1895). La situation présente, ses origins et ses perils. Vanbuggenhoudt.
George, A. L. & Bennett, A. (2005). Case Studies and Theory Development in the Social Sciences. MIT Press.
Gerring, J. (2007). Case Study Research: Principles and Practices. Cambridge University Press.
Gilissen, J. (1958). L’régime représentatif en Belgique depuis 1890. Le renaissance du livre.
Goblet D’Alviella, E. (1900). La représentation proportionelle en Belgique. Histoire d’une réforme. Weissenbruch.
Greener, I. (2005). The Potential of Path Dependence in Political Studies. Politics, 25(1), 62-72. https://doi.org/10.1111/j.1467-9256.2005.00230.x.
Hall, P. A. (1993). Policy Paradigm, Social Learning and the State. Comparative Politics, 25, 275-296.
Hall, P. A. (2003). Ailing Ontology and Methodology in Comparative Politics. In J. Mahoney & D. Rueschemeyer (Eds.), Comparative Historical Analysis in the Social Sciences (pp. 373-404). Cambridge University Press.
Hall, P. A. (2012). Tracing the Progress of Process Tracing. European Political Science, 12, 20-30. https://doi.org/10.1057/eps.2012.6.
Kreuzer, M. (2010). Historical Knowledge and Quantitative Analysis: The Case of the Origins of Proportional Representation. American Political Science Review, 104(2), 369-392.
Kreuzer, M. (2016). Assessing Causal Inference Problems with Bayesian Process Tracing: The Economic Effects of Proportional Representation and the Problem of Endogeneity. New Political Economy, 21(5), 473-483. https://doi.org/10.1080/13563467.2015.1134467
Leemann, L. & Mares, I. (2014). The Adoption of Proportional Representation. Journal of Politics, 76(2), 461-478. https://doi.org/10.1017/S0022381613001394.
Liebman, M. (1979). Histoire du mouvement ouvrier en Belgique: Les socialists belges (1885-1914). Vie Ouvrière.
Linden, H. V. (1920). Belgium: The Making of a Nation. Oxford University Press.
Mahaim, E. (1900). Proportional Representation and the Debates Upon the Electoral Question in Belgium. Annals of the American Academy of Political and Social Sciences, 15, 381-404.
Mahoney, J. (2012). The Logic of Process Tracing Tests in the Social Sciences. Sociological Methods & Research, 20(10), 1-28.
Mahoney, J. (2015). Process Tracing and Historical Explanation. Security Studies, 24(2), 200-218. https://doi.org/10.1080/09636412.2015.1036610.
Mahoney, J. & Thelen, K. (2010). A Theory of Gradual Institutional Change. In J. Mahoney & K. Thelen (Eds.), Explaining Institutional Change: Ambiguity, Agency and Power (pp. 1-37). Cambridge University Press.
Mahoney, J. & Thelen, K. (2015). Comparative-Historical Analysis in Contemporary Political Science. In J. Mahoney & K. Thelen (Eds.), Advances in Comparative-Historical Analysis (pp. 3-38). Cambridge University Press.
Marks, G., Mbaye, H. A. D. & Kim, H. M. (2009). Radicalism or Reformism? Socialist Parties before World War I. American Sociological Review, 74(4), 615-635. https://doi.org/10.1177/000312240907400406
Mommaert, J. (1900). Le mechanism de la representation proportionnelle d’après la loi du 29 décembre 1899. Société Belge de Librairie.
Moravcsik, A. (2014). Transparency: The Revolution in Quantitative Research. PS: Political Science and Politics, 47(1), 48-53. https://doi.org/10.1017/S1049096513001789.
Penadés, A. (2008). Choosing Rules for Government: The Institutional Preferences of Early Socialist Parties. In J. M. Maravall & I. Sanchez-Cuenca (Eds.), Controlling Governments (pp. 202-246). Cambridge University Press.
Pierson, P. (2004). Politics in Time. Princeton University Press.
Pilet, J.-B. (2007). Changer pour gagner? Les réformes des lois électorales en Belgique. Université de Bruxelles.
Polasky, J. (1992). A Revolution for Socialist Reforms: The Belgian General Strike for Universal Male Suffrage. Journal of Contemporary History, 27(3), 449-466.
Renwick, A. (2010). The Politics of Electoral Reform: Changing the Rules of Democracy. Cambridge University Press.
Ricks, J. I. & Liu, V. (2018). Process-Tracing Research Designs: A Practical Guide. PS: Political Science and Politics, 51(4), 842-846.
Rokkan, S. (1970). Citizens, Elections, Parties: Approaches to the Comparative Study of the Processes of Development. Universitetsforlaget.
Schmidt, V. A. (2010). Taking-Ideas and Discourse Seriously: Explaining Change Trough Discursive Institutionalism as the Forth ‘New Institutionalism.’ European Political Science Review, 2(1), 1-25. https://doi.org/10.1017/S175577390999021X.
Schröder, V. & Manow, P. (2020). An Intra-Party Account of Electoral System Choice. Political Science Research and Methods, 8(1), 1-17. https://doi.org/10.1017/psrm.2018.58.
Stengers, J. (2000). L’établissement de la representation proportionnelle en Belgique en 1899. In P. Delwit & J-M. De Waele (Eds.), Le mode de scrutin fait-il l’élection? (pp. 129-143). Editions de l’Université de Bruxelles.
Stengers, J. (2004). Histoire de la législation électorale en Belgique. Revue Belge de Philologie et d’Histoire, 82, 247-270. https://doi.org/10.3406/rbph.2004.4826.
Strikwerda, C. (1997). A House Divided: Catholics, Socialists and Flemish Nationalists in Nineteenth Century Belgium. Rowman & Littlefield.
Tannenwald, N. (2015). Process Tracing and Security Studies. Security Studies, 24(2), 219-227. https://doi.org/10.1080/09636412.2015.1036614.
Trampusch, C. & Palier, B. (2016). Between X and Y: How Process Tracing Contributes to Opening the Black Box of Causality. New Political Economy, 21(5), 437-454. https://doi.org/10.1080/13563467.2015.1134465
Van der Smissen, E. (1898). Létat actuel des partis politiques en Belgique. Annales de l’Ecole Libre des Sciences Politiques, 15 September, XIII, 549-566.
Van der Smissen, E. (1920). Léopold II et Beernaert d’après leur correspondence inédite de 1884 à 1894. Goemaere.
Vandervelde, E. (1925). Le parti ouvrier Belge. L’Englantine.
Van Evera, S. (1997). Guide to Methods for Students of Political Science. Cornell University Press.
Waldner, D. (2015a). Process Tracing and Qualitative Causal Inference. Security Studies, 24(2), 239-250. https://doi.org/10.1080/09636412.2015.1036624
Waldner, D. (2015b). What Makes Process Tracing Good? Causal Mechanisms, Causal Inference, and the Completeness Standard in Comparative Politics. In A. Bennett & J. T. Checkel (Eds.), Process Tracing from Metaphor to Analytic Tool. Cambridge University Press.
Walter, A. & Emmenegger, P. (2019). Majority Protection? The Origins of Distorted Proportional Representation. Electoral Studies, 59, 64-77. https://doi.org/10.1016/j.electstud.2019.02.002.
Woeste, C. (1898). Les élections de 1898 et la cause catholique. Société belge de libraire, Oscar Schepens & Cie Successeurs.
Woeste, C. (1927). Mémoires pour servir à l’histoire contemporaire de la Belgique. Tome I. L’édition universelle.
Woeste, C. (1933). Mémoires pour servir à l’histoire contemporaire de la Belgique. Tome II (1894-1914). L’édition universelle.
Zacks, S. (2017). Relationship Among Rivals (RAR): A Framework for Analyzing Contending Hypotheses to Process-Tracing. Political Analysis, 25(3), 344-362. https://doi.org/10.1017/pan.2017.12.
- * The author would like to thank Royden Clark, Margarita Dilkovska, Carol Mershon, Marina Omar, the anonymous reviewers, and the editors and editorial staff at Politics of the Low Countries. All errors remain mine. The archival research was sponsored by the Dumas Malone Graduate Research Fellowship at the University of Virginia.
-
1 The concept does not distinguish clearly between short- and long-term Left electoral threat and Left extra-institutional threat.
-
2 A single-member district (SMD) is a constituency that elects one member of parliament. A multi-member district (MMD) is a constituency that elects more than one deputy. Single-member district plurality is an electoral system, also known as ‘the first-past-the-post system’, in which the candidate that wins the most votes wins the single seat in that district.
-
3 Büthe (2002) argues that process-tracing allows hypothesis generation, but not testing. Gerring (2007) adopts an intermediary position on the issue.
-
4 See Bates et al. (1998) for analytic narratives, Hall (2003) for systematic process analysis and Kreuzer (2016) for Bayesian process-tracing.
-
5 Annales Parlementaires: Séance du 24 janvier, 1888. Bruxelles, Belgique: Chambre de Représentants, p. 423.
-
6 Session de 1894-1895: Liste des membres de la Chambre, 1894. Bruxelles, Belgique: Chambre de Représentants.
-
7 Annales Parlementaires: Séance du 27 juin, 1899. Bruxelles, Belgique: Chambre de Représentants.
-
8 Annales Parlementaires: Séance du 27 juin, 1899. Bruxelles, Belgique: Chambre de Représentants; Annales Parlementaires: Séance du 28 juin, 1899. Bruxelles, Belgique: Chambre de Représentants; Annales Parlementaires: Séance du 29 juin 1899. Bruxelles, Belgique: Chambre de Représentants.
-
9 They provide a contemporaneous view from a different perspective and new evidence, against which the well-known historical sources and the recent explanations could be compared.
-
10 Annales Parlementaires: Séance du 27 juin, 1899. Le Courrier de Bruxelles, 27 June, 1899, p. 1; Le Peuple, 27 June, 1899, La zone néutre en état de siege, p. 1; Le Peuple, 29 June, 1899, Interpellations sur les brutalites de la Gendarmerie: Situation révolutionnaire, p. 1.
-
11 Annales Parlementaires: Séance du 29 juin 1899. Bruxelles, Belgique: Chambre de Représentants.
-
12 Le Peuple, 30 June, 1899. Interpellations sur les brutalites de la Gendarmerie, p. 1.
-
13 Le Courrier de Bruxelles, June 30, 1899. p. 1.
-
14 Le Peuple, 1 July, 1899. Du sang! p. 1.
-
15 Annales Parlementaires: Séance du 24 novembre, 1899. Bruxelles, Belgique: Chambre de Représentants, pp. 51-67.
-
16 Leemann and Mares (2014).
The introduction of proportional representation (PR) in Belgium has a special place in the electoral system reform literature. Not only is Belgium the first country to adopt PR for national legislative elections, but the Belgian case appears as a conspicuous exception to the conventional wisdom that dominant political parties prefer majoritarian electoral systems. PR in Belgium was introduced in 1899 by a coalition of moderate Catholics and progressive Liberals, when the Catholic Party held 112 of the 152 seats the Chamber of Representatives (Barthélemy, 1912; Pilet, 2007). The Belgian case has re-vitalised research into the origins of PR and re-ignited methodological discussions about good historical analysis (Boix 2010; Kreuzer, 2010). Consequently, much progress has been made in understanding electoral system reform in Western Europe.
Nonetheless, there is little consensus as to why exactly Belgium adopted PR. Several explanations exist: Left electoral threat (Boix, 1999, 2010; Rokkan, 1970); Left existential threat (Ahmed, 2013); and the interaction between electoral and extra-institutional threats (Barzachka, 2014). Recently, Emmenegger and Walter (2019) have put electoral threat back in the spotlight. Combining historical research with election data analysis, they argue that the Catholic Party introduced PR because it was vulnerable to high electoral threat from Liberal-Socialist cartels. Which account best explains the Belgian case?
Process-tracing, with its focus on temporal sequences, carefully constructed narratives and counterfactual analysis, can help evaluate rival explanations of single case studies (Beach & Pedersen, 2013; Bennett & Checkel, 2015; Brady & Collier, 2010; George & Bennet, 2005; Mahoney, 2012; Tannewald, 2015). The present article re-examines the recent scholarship on PR in Belgium. It applies the completeness standard of process-tracing (Crasnow, 2017; Waldner, 2015a; 2015b) and uses theory-guided process-tracing (Falleti, 2016) based on insights from the institutional change literature (Campbell, 2004; Mahoney & Thelen, 2010, 2015).
The article is a replication study. It scrutinises the main historical sources used by these accounts – parliamentary transcripts, participant memoirs and contemporaneous academic research, checking for errors, inconsistencies and/or omissions and paying careful attention to the sequencing of events. To corroborate the review of the well-known primary and secondary sources on the subject (Appendix I), it supplements the analysis with new evidence from 1899 newspapers, from the archives of the Royal Library of Belgium. These sources show the significance of the June 1899 demonstrations.
In Belgium, electoral threat (including threat from Liberal-Socialist cartels) and extra-institutional threat fluctuated and mattered at different points of the electoral system reform process. Over time, these variables interacted and influenced office-holders’ preferences for different electoral systems. When in 1899, Catholic politicians finally agreed on a solution – a mixed electoral system designed to reinforce Catholic dominance – high extra-institutional threat from the Socialists and the Liberals stopped the reform. Concerned about civil unrest and equal universal male suffrage and after exhausting all other options, the Catholics compromised, adopting PR in all districts. Both increasing electoral threat and high extra-institutional threat shaped this outcome.
The recent literature on PR in Belgium ‘reads history forward’ (Capoccia & Ziblatt, 2010, p. 943). Nevertheless, it has inconsistencies and/or omissions that need correction. The present article seeks to rectify these problems.