An important question in the fields of political science and political economy is why some demographic and professional groups have more presence and influence than others in politics. Whether viewed from the perspective of ‘political selection’ (e.g., Besley, Reference Besley2005) or ‘descriptive representation,’ (e.g., Pitkin, Reference Pitkin1967), the occupational and educational backgrounds of legislators have significant consequences for public policy-making due to the role that professional socialization and educational training plays in shaping law makers' experiences, perceptions, interests, and behaviors.
In response, scholars have sought to better understand how institutions impact the degree to which occupational groups enter politics. For instance, research on Germany and Austria finds public servants more likely to become members of parliament (MPs) when they are guaranteed re-employment, receive compensation for holding their office in abeyance, and when ineligibility and incompatibility rules are not present (Braendle and Stutzer, Reference Braendle and Stutzer2010). In Russia, studies have found more businessman candidates in regions where democratic institutions (i.e., political parties, free media, and government transparency) are weak (e.g., Gehlbach et al., Reference Gehlbach, Sonin and Zhuravskaya2010), and in China, entrepreneurs seem to participate more in politics in regions where markets are less developed (Li et al., Reference Li, Meng and Zhang2006). Meanwhile, the ratio between salaries for the US Congress and those offered in the private sector appear to impact both the average ‘quality of entering politicians’ and turnover in office (Mattozzi and Merlo, Reference Mattozzi and Merlo2008: 599).
The high share of MPs with law backgrounds has also drawn considerable attention (e.g., Hain and Pierson, Reference Hain and Pierson1975; Besley, Reference Besley2005: 54; Moliterno, Reference Moliterno2009; Bonica, Reference Bonica2017; Robinson, Reference Robinson2017). Explanations for the prevalence of MPs with educational or occupational training in law tend to focus on the socialization experience of law school (e.g., Miller, Reference Miller1993, Reference Miller1995), lawyers' professional networks and fundraising abilities (e.g., Bonica, Reference Bonica2017), the structure of the legal profession (e.g., Robinson, Reference Robinson2017), and restrictions in legislatures concerning legal representation of clients before state entities (Rosenson, Reference Rosenson2006). A related factor is whether MPs are permitted to simultaneously work elsewhere while holding public office since attractive alternatives in the private sector may dis-incentivize lawyers from running for a political position (e.g., Gagliarducci et al., Reference Gagliarducci, Nannicini and Naticchioni2010).
Despite these advances, previous attempts to explain the proliferation of lawyers and law graduates in legislatures have overlooked one of the most basic structural factors of a democratic polity – its electoral system. At stake here is the common usage of a ‘first-pass-the-post’ (FPTP) single-member district (SMD) electoral system in the institutional ensemble of many former British colonies including the USA where lawyers are particularly prominent in legislatures, a feature which stands in contrast to many other industrialized democracies where elections rely on some variant of proportional representation (PR) (e.g., Norris, Reference Norris2004; Farrell, Reference Farrell2011).
As this article explains, a parliament's electoral system can play an important role in influencing the degree to which its membership is comprised of citizens with a background in law defined here as possessing either a university degree in law (‘law graduates’) or occupational experience working in the field of law (‘lawyers’). More specifically, this article proposes that SMD elections are more conducive to the election of MPs with a background in law than PR systems. This presumed causal relationship stems from differences between whether electoral systems aim to achieve ‘principal-agent’ or ‘microcosmic’ models of representation (e.g., McLean, Reference McLean and Held1991; Farrell, Reference Farrell2011: 10–11). On the one hand, SMD elections presume a principal-agent model of representation which incentivizes political parties to select MPs with training in the field of law. On the other hand, PR elections presume a microcosmic model of representation which incentivizes parties to select MPs from diverse demographic and occupational backgrounds.
The rest of this paper is structured as follows. First, it reviews the scholarly literature on the causes and consequences of having a large number of legislators with an educational or occupational background in law. It then considers how FPTP-SMD and closed-list PR electoral systems may differentially impact the proliferation of lawyers and law graduates in parliaments with the former more conducive to MPs with a background in law than the latter. This conjecture is then tested and prima facie evidence is found in support of this relationship through an analysis of elections conducted in three large parliaments with mixed-member electoral systems combining SMD and PR tiers: Japan, South Korea, and Germany. The article concludes by discussing the implications of these findings and suggestions for future research.
1. Literature review
1.1. The proliferation of lawyer-legislators
Comparative research has repeatedly found that educated professionals comprise a large share of the world's legislators, and among white-collar occupations present in parliaments, lawyers are particularly prominent (e.g., Miller, Reference Miller1995; Norris and Lovenduski, Reference Norris and Lovenduski1995; Norris and Franklin, Reference Norris and Franklin1997; Best and Cotta, Reference Best and Cotta2000; Best and Vogel, Reference Best, Vogel, Martin, Saalfeld and Strøm2014; Bonica, Reference Bonica2017). A glaring example of this phenomenon is the USA where the disproportionate influence of lawyers in politics was famously observed nearly two centuries ago by de Tocqueville (Reference de Tocqueville1838/2003). Since gaining independence in the late eighteenth century, lawyers have comprised 71% of US Senators and 62% of its House of Representatives (Bonica, Reference Bonica2017: 4) as well as 100% of Supreme Court judges, 59% of US presidents, 68% of vice presidents, and 63% of cabinet members (Robinson, Reference Robinson2017: 667). Lawyers have also comprised a high proportion of US state-level legislators estimated at 40% within state senates and 25% within state houses (Cohen, Reference Cohen1969: 569). As Robinson (Reference Robinson2017) notes, lawyers currently constitute only 0.4% of the US population, but in 2015–16, lawyers (37%) were the largest occupational group in the US Congress followed by ‘business and banking’ (26%) and ‘public service and politics’ (23%).
Aside from the US case, lawyers are numerous in many other countries' legislatures including the UK and former British colonies (Podmore, Reference Podmore1977). In the 622-member German Bundestag (GBT), Braendle and Stutzer (Reference Braendle and Stutzer2016: 713) found more legal professionals (102) than any other category including educational professions and researchers (86) or public administrators (62). Lawyers have also been prominent in Canadian politics comprising roughly a fifth of legislators and almost three-fifths of Canadian Prime Ministers (22 of 30) (Chatoor, Reference Chatoor2013: 4). Similarly, Braendle (Reference Braendle2015: 9) found a strong presence of European Parliament members with a degree in law (20%), and a large-scale study conducted by The Economist (2009) has estimated that lawyers comprise about 20% of the world's politicians – a greater share than any other professional background including business (16%), diplomacy (12%), or the military (11%).Footnote 1
A long-standing question among scholars has been whether legislators with a law background think and act differently than their counterparts without legal training (e.g., Eulau and Sprague, Reference Eulau and Sprague1964). In the US case, due to American Bar Association (ABA) accreditation requirements, the high costs of obtaining a law degree, and the need to pass a bar exam, lawyers are more likely to be white, male, and affluent compared to non-lawyers (Shepherd, Reference Shepherd2003; Michelson, Reference Michelson2013). Moreover, lawyers often ‘come from more “elite” family backgrounds’ and ‘earn more than typical Americans’ (Robinson, Reference Robinson2017: 694) resulting in a US Congress dominated by representatives from affluent households (see Carnes, Reference Carnes2013, Reference Carnes2018). Miller's (Reference Miller1993: 1) interview research found ‘lawyers have more positive attitudes toward the courts’ than non-lawyers and Eakins (Reference Eakins2006) found lawyers in politics to be more ambitious than non-lawyers. Polsby (Reference Polsby, Thaysen, Davidson and Livingston1990: 114) also identifies a critical mass effect whereby the ‘occupational culture of Congress’ in the USA is ‘dominated by lawyers' ways and lawyers' jargon.’ As Davidson (Reference Davidson and Katzmann1988: 92) notes, ‘oral debates and written reports display lawmakers' intense fascination with legal procedures, concepts, and terminology. And members often reveal a knack for turning substantive issues into matters of legal or procedural detail – perhaps a belated manifestation of legal training.’Footnote 2
One alarming possibility is that lawyer-legislators may be systematically biased toward advancing their own industry and personal gains at the expense of general public interests (Kagan, Reference Kagan2001). While, in some countries like Russia, businesspeople commonly run as legislative candidates in order to ‘enjoy unparalleled access to policy decisions’ that help their businesses profit (Szakonyi, Reference Szakonyi2018: 325; see also Braendle, Reference Braendle2016), a study of roll call votes in the US Congress and state legislatures regarding the $265 billion a year tort business found lawyers ‘systematically less likely to vote in favour of tort reforms that restrict tort legislation but more likely to support bills that extend tort law compared to legislators with different professional backgrounds’ (Matter and Stutzer, Reference Matter and Stutzer2015: 357).
If lawyer-legislators share such biases, then why do so many lawyers enter politics in the first place? In the US case, lawyers hold many intermediate feeder positions for US state and national legislatures such as the positions of the state attorney general and public prosecutor (Shepherd, Reference Shepherd2003: 654; Francis, Reference Francis2014: 80; Robinson, Reference Robinson2017: 691). Another reason is their ability to ‘enter and leave politics without damaging their legal careers’ (Miller, Reference Miller1993: 2). Lawyers can use legislative positions as stepping stones to advance their careers (Hain and Pierson, Reference Hain and Pierson1975), and even losing an election campaign can help increase their name recognition and network of contacts.Footnote 3
Some scholarship also suggests the socialization experience of obtaining a law degree is a strong factor in shaping law graduates' desire to enter politics. As Miller (Reference Miller1993: 4) notes, regardless of whether law graduates ever become practicing attorneys, law school socialization has a long-term influence and ‘professional training produces changes in individuals which stay with them throughout their subsequent careers.’ For instance, legal training may enrich a future politician's logic, reasoning, debating skills, and ability to persuade (e.g., Podmore, Reference Podmore1977).Footnote 4 Yet, it might also make them ‘accustomed to representing clients’ [private] interests against the interests of all others, including the public interest’ (Moliterno, Reference Moliterno2009: 1262). As Miller (Reference Miller1995) has argued, legal training also leads lawyers to approach problems differently by preferring a ‘case-by-case approach in attacking a problem’ and placing legal analysis over ‘consideration of broader values’ (Miller, Reference Miller1993: 5). Law graduates might also be more likely to ‘think of themselves as qualified to run’ or have ‘an elected official as a family member or as a friend’ (Robinson, Reference Robinson2017: 687) thus providing them with more insider knowledge and self-confidence than other potential candidates. Relatedly, Eulau and Sprague (Reference Eulau and Sprague1964) found many who chose to enter law school already had an above average interest in politics. Nevertheless, a generally higher level of interest among law graduates for joining politics still does not explain the variation in MPs with a law background within and across countries – perhaps electoral systems do?
1.2. Electoral systems and occupational representation
As Besley (Reference Besley2005: 52) has argued, compared to inheritance, lotteries, and force, ‘elections are now the most commonly used institution for selecting a political class, but enormous diversity exists across electoral institutions.’ While scholars have not yet considered how electoral system differences impact the election of lawyers, the possibility of a systematic link between electoral systems and candidate selection methods has been identified by Hazan and Voerman (Reference Hazan and Voerman2006: 148) who propose that candidate selection methods will be more decentralized under SMD plurality electoral systems, but more centralized in multi-member districts. Partially confirming these conjectures, a study comparing the candidate selection methods of 512 parties in 46 countries concluded that on average ‘parties competing in unitary systems will tend to adopt more centralized and exclusive selection processes, whereas parties under federal countries will adopt more inclusive and decentralized selection mechanisms’ (Shomer, Reference Shomer2014: 543).
While there may be a link between SMDs and decentralized candidate selection procedures, SMD electoral systems exhibit other characteristics conducive to the selection of lawyer candidates. First, the literature on electoral systems has identified the differing character of SMD and PR in terms of competition structure and nomination patterns as the former is uninominal (district magnitude = 1), while the latter employs multi-member districts with correspondingly higher district and party magnitudes (e.g., Duverger, Reference Duverger1954; Rae, Reference Rae1967; Taagepera and Shugart, Reference Taagepera and Shugart1989; Norris, Reference Norris2004; Farrell, Reference Farrell2011). Moreover, a general finding is that PR (especially when using a closed-list) tends to be more inclusive of diversity than SMD when it comes to electing otherwise under-represented population groups such as women and young people (e.g., Matland and Studlar, Reference Matland and Studlar1996; Norris, Reference Norris2004; Joshi, Reference Joshi2013; Joshi and Kingma, Reference Joshi and Kingma2013; Reiser, Reference Reiser2014).
Second, SMD creates an adversarial winner-take-all contest. This incentivizes candidates in competitive districts to excel in matters such as persuasion, fundraising, and perseverance to out-compete adversaries. Of particular importance is a candidate's personal appeal linked to the ‘personal vote,’ whereas party reputation and ideology are relatively more important in closed-list PR systems (Carey and Shugart, Reference Carey and Shugart1995). As Herron et al. (Reference Herron, Nemoto, Nishikawa, Herron, Pekannen and Shugart2018: 451) note, candidates elected under SMD systems are often expected or motivated to ‘work for local interests, while candidates under PR work for broader interests.’ For this reason, political parties often see law graduates and lawyers who have been trained to vigorously advocate on behalf of a narrow client base as attractive candidates for SMDs (Lawless, Reference Lawless2012; Francis, Reference Francis2014).
Third, as Galasso and Nannicini (Reference Galasso and Nannicini2017) point out, political parties' top preference is often to select ‘loyal candidates’ (especially party elites and long-term party members) as evident in candidate nominations for the top of their PR-lists and for safe SMD seats. However, parties may experience a backlash from electorates in SMD contests if they do not choose ‘high-quality’ candidates (those with more education and higher incomes) such as lawyers, which were found more often to be representing competitive SMDs under Italy's previously mixed electoral system (ibid).
2. Causal logic and hypotheses
The discussion above highlights the propensity of lawyers and law graduates to enter parliaments when an FPTP-SMD electoral system is in place, and it illustrates that different electoral systems might lead to different results for political selection. Regarding the latter, electoral systems differ in character based on the degree to which they are linked to different logics or models of representation. A crucial distinction here is between ‘principal-agent’ and ‘microcosm’ models (McLean, Reference McLean and Held1991; Farrell, Reference Farrell2011: 10–11).
First, a principal-agent model of representation lies behind SMD elections because the district magnitude is only one. Under these circumstances, legal training will likely facilitate agentic representation because legal advocacy of the type embedded in the lawyer–client relationship studied in law school and practiced in the legal profession corresponds to a parliamentary representative who acts on behalf of a specific geographic constituency based on the incentive of re-election. For this reason, given the skills of those with a law background and the fact that entrance to a law degree program is often available only to those with a high level of intellectual acumen, political parties are likely to choose a sizable number of lawyers to serve as representatives under SMD systems.
By contrast, PR elections imply a microcosmic model of representation in which a legislature is supposed to resemble the population at large. As articulated by John Adams, this vision expects parliament to function like ‘an exact portrait, in miniature, of the people at large, as it should think, feel, reason, and act like them’ (quoted in McLean, Reference McLean and Held1991:173). As Farrell (Reference Farrell2011: 10) points out, this is very ‘similar to the governing principle behind public opinion polls: that is, the notion of a representative sample.’ Hence under PR elections (especially of the closed-list variety), diversity of representatives across occupational and demographic groups is usually preferred by party elites in comprising candidate lists over concentration from a single profession in order to better inform overall decision-making and to demonstrate to different segments of voters that people like them will be included as representatives.
To sum up, SMD elections presume a principal-agent model of representation which I argue incentivizes political parties to select more MPs with training in the field of law than would be the case under PR elections. This leads us to the following hypotheses. Although the literature tends to conflate the two, I first distinguish between ‘law graduates’ (those with a university degree in law) and ‘lawyers’ (those whose occupation is in the field of law). My null hypothesis (H0) is that politics simply attracts law graduates and lawyers in general, so there should be no significant difference in their share elected under SMD or PR systems. By contrast, my first hypothesis (H1) is that despite differences across nations in their share of law-graduate MPs, in each individual parliament, law graduates will hold a greater share of SMD seats than PR seats.Footnote 5 My second hypothesis (H2) is that despite differences across nations in their share of MPs in law-related occupations (‘lawyers’), in each individual parliament, lawyers will hold a greater share of SMD seats than PR seats.
3. Methodology
To test these hypotheses, I examined three large national parliaments with mixed electoral systems combining both SMD and PR elections. This approach enables isolating the impact of distinct electoral systems in a context where two different elections are held side-by-side. It also serves as a convenient means to control for factors such as country-specific conditions, election timing, type and level of legislature, and partisan competition. In other words, mixed electoral systems allow one to compare election rates of lawyers and law graduates under different electoral systems while holding most other relevant factors constant.
My analysis begins with South Korea's unicameral parliament, the Korean National Assembly (KNA) (대한민국 국회Daehanminguk Gukhoe), which employs a parallel ‘majoritarian’ mixed-member (MMM) electoral systems involving strict ‘separation’ between electoral tiers. This is a case where the two electoral tiers disallow dual candidacy. In other words, a candidate must either contest an SMD seat or be placed on a PR-list. They cannot do both simultaneously. I then replicate this analysis on the mixed-member electoral systems of the German Bundestag (GBT) and Japanese House of Representatives (JHR) (衆議院Shugiin) which comprise the lower houses of their respective bicameral national parliaments. The GBT and JHR, respectively, have mixed-member proportional (MMP) and MMM systems involving contamination between their two electoral tiers by allowing dual candidacy.Footnote 6
These three national assemblies share certain similarities. All have a large number of seats and are located in advanced industrialized democratic countries that are members of the Organization for Economic Cooperation and Development (OECD). In each of these parliamentary chambers, voters have two votes. The first vote is for an SMD seat based on a plurality FPTP formula, while the second vote is for the PR tier involving closed-lists compiled by political parties over which voters have no say in determining who the candidates are or in what order on the list.Footnote 7 In all three cases (Germany, Japan, and South Korea), the parliaments are genuinely influential institutions in national politics, and MP salaries are substantial and identical for members elected via PR-lists and SMD seats. Reported monthly salaries (and their US$ equivalents) are 1,294,000 yen ($11,570) for the Japanese National Diet (国会Kokkai), 9,780.28 euros ($11,145) plus an allowance of 4,339.97 euros ($4,946) for the GBT, and 11,496,667 won ($10,096) for the KNA.Footnote 8 In these three assemblies, law is a prominent background (occupation or university major) among current representatives as shown in Table 1. In the JHR and KNA, law is the most common university degree held by MPs. In the GBT, law is the most common previous occupation among its current MPs.
Table 1. Most common occupations/university majors among members of parliament
Data Source: Compiled from member bios on parliamentary websites.
Note: ‘Political Sci.’ includes International Relations and Diplomacy.
In the analysis that follows, I focus on statistically significant differences between MPs with law backgrounds in SMD and PR seats across individual parliamentary sessions. To assess legislator backgrounds, information was obtained from MP profiles on parliamentary websites regarding their occupational and educational backgrounds for MPs serving during the last four parliamentary sessions in each country.Footnote 9 Data were also collected on the current membership of the Japanese House of Councilors (JHC) (参議院 Sangiin) which forms the upper house of the Japanese Diet.Footnote 10 In all cases, MPs were coded as having a ‘law background’ based on their either having a university degree in law or being listed on their respective parliament website as having a law-related occupation such as ‘lawyer,’ ‘attorney,’ ‘judge,’ ‘jurist,’ or ‘legal advisor.’ As MP background information is presented differently on parliament websites in each country, the coding of law backgrounds for South Korea and Japan are based on university degrees. For Germany, it is based on occupation.
4. Analysis
My empirical analysis begins with the KNA where over the past four elections, 23.3–31.7% of KNA members have been law graduates. The KNA is an ideal case to test my first hypothesis, because it has a ‘strictly’ parallel electoral system that does not allow dual candidacy while combining FPTP single-member districts with a relatively high-magnitude PR tier comprising one nationwide constituency. In line with my expectations, as shown in Table 2, over the past four KNA elections, there has consistently been a much lower share of law graduates in its PR tier (6.6–23.3%) than in SMD seats (27.0–33.7%).
Table 2. MPs with law backgrounds (LBMPs) in parliaments with mixed electoral systems
Data Source: Author's dataset.
Note: Table includes the Korean National Assembly (KNA), German Bundestag (GBT), Japanese House of Representatives (JHR), and Japanese House of Councilors (JHC).
KNA members also include mid-term replacements. Asterisks mark statistically significant differences as measured by χ 2 tests, P-values: *<0.10, **<0.05, ***<0.01.s
I then analyzed the lower houses of the parliaments of Japan and Germany which allow dual candidacy but nevertheless have a large number of both PR and SMD seats thereby offering a further opportunity to explore the plausibility of my hypotheses. In Germany, MPs with work experience in law-related occupations comprised 19.8–20.8% of the GBT over the past four elections, but resembling the pattern observed for law graduates in South Korea, the share of lawyers in the Bundestag's SMD-tier (21.4–31.0%) has consistently been higher than in its PR-tier (12.3–18.1%).
By contrast, there has been a greater share of law graduates ranging from 34.2 to 37.7% in the JHR over its past four elections. However, even in this law-heavy context,Footnote 11 where laws governing campaign periods (only 12 days) and permissible activities are fairly restrictive (Nemoto, Reference Nemoto, Herron, Pekannen and Shugart2018), and ‘having a legal background increases the likelihood of an MP serving in a high-policy post’ (Pekkanen et al., Reference Pekkanen, Nyblade and Krauss2006: 188), the electoral system is correlated with the election of citizens with educational training in law. In the JHR, the share of law graduates elected in SMDs (38.0–43.4%) has consistently been much higher than in the PR tier (27.2–31.3%).
Lastly, I examined the current composition of the JHC which (like the KNA) bans dual candidacy but practices staggered elections with half of its members elected every 3 years. Law graduates comprise a fairly high share (29.9%) of its 241 current members (as of February 2019) who are much more prevalent in SMD seats (32.3% of 65 SMD seat holders) compared to the PR tier where their share is only 20.8% (out of 96 MPs).Footnote 12
5. Discussion
Across all three countries and every election examined, we observed a regular pattern. There has consistently been a greater share of law graduates and lawyers in SMD seats than in PR seats in Germany, South Korea, and Japan. This gap exceeded 9 percentage points in 12 out of 13 assembly sessions and was statistically significant as measured by χ 2 tests at the P < 0.05 level in nine cases and at the P < 0.12 level in 12 cases.Footnote 13 Overall these findings support my two hypotheses.
What we do not know yet, however, is whether candidates with a law background become MPs in these assemblies because of self-selection, party selection, or election by voters. One limitation of this study is its inability to identify which one of these causal mechanisms has the greatest impact, but theoretically, much of the literature on political recruitment (following Norris and Lovenduski's (Reference Norris and Lovenduski1995) candidate supply-demand model) would suggest it is party selection. This is because lawyers are often present as MP candidates on the supply side, but on the demand side, political parties largely determine who the candidates are that voters can choose from by accepting or rejecting aspirant candidate's wishes to be nominated as candidates (ibid: 107–115). This in turn heavily influences the choices available to voters and whether citizens pursue candidacy in the first place.
The principal-agent logic embedded in SMD electoral systems also suggests that party selection plays a greater role than self-selection or selection through the electorate. As discussed above, due to their training and experience, there are good reasons to believe that candidates with a law background will be more inclined toward becoming legislators than the average person regardless of the electoral system. However, party leaders should be more likely to select candidates with a law background to represent a constituency when facing an SMD election than under PR-list elections since the latter incentivizes them to broaden and diversify the profile/backgrounds of candidates on the ballot. The role of party selection here is especially salient in non-competitive SMD-FPTP districts and in closed PR-lists where voters have no say over the actual candidates offered to them by parties because candidates and their ranking are determined almost exclusively by party elites. Another commonality amongst the cases studied here is that despite differing candidate recruitment methods across countries and political parties, scholars generally find party elites in control of candidate selection procedures in Japan (e.g., Yu et al., Reference Yu, Chen-Hua Yu and Shoji2014), Germany (e.g., Reiser, Reference Reiser2014; Zittel, Reference Zittel, Herron, Pekannen and Shugart2018), and South Korea (e.g., Jun and Hix, Reference Jun and Hix2010).
While the findings of this study will require subsequent validation obtained by investigating more cases both spatially and temporally and applying sophisticated techniques to control for other possible contravening factors, our hypotheses here have nevertheless received noticeable support under somewhat diverse conditions. As noted above, the assemblies we examined diverge in terms of their mixed-electoral system configurations. First, their share of PR-list seats ranges considerably from 50% in Germany and 38% in Japan to 15% in South Korea. Average district magnitude for the PR tier varies from 18 (ranging from 7 to 33) in the JHR to one nationwide district of 46 for the KNA and a ‘de facto’ national district of 656 for the GBT (on the latter, see Shugart and Wattenberg, Reference Shugart, Wattenberg, Shugart and Wattenberg2001; Zittel, Reference Zittel, Herron, Pekannen and Shugart2018). The legal minimum threshold of party votes for PR allocation also varies from 5% or three SMD seats for Germany and 3% for South Korea to none for Japan. As Germany has the most PR-list seats, it arguably has a much weaker personal vote among the three countries.Footnote 14 Germany also features ‘overhang seats’ (Überhangmandate) which ‘actually increase the total number of seats in the chamber in order to allow all other parties to receive their full complement of list-tier seats despite the bonus obtained by some large parties’ (Shugart and Wattenberg, Reference Shugart, Wattenberg, Shugart and Wattenberg2001: 23).
Another source of variation is the presence of dual candidacy in the GBT and JHR where PR-list members of parliament from the largest parties are assigned by their parties to a district constituency to which they provide constituency service.Footnote 15 Hence, the widespread practice of dual candidacies in Germany and Japan (e.g., Klingemann and Wessels, Reference Klingemann, Wessels, Shugart and Wattenberg2001; Reiser, Reference Reiser2014; Nemoto, Reference Nemoto, Herron, Pekannen and Shugart2018)Footnote 16 introduces an element of ‘contamination’ between the two electoral tiers since ‘dual candidacy systems can generate dually nominated list winners who lose in the nominal component but win seats via the list component’ (Herron et al., Reference Herron, Nemoto, Nishikawa, Herron, Pekannen and Shugart2018: 447). In other words, because there is less separation between the electoral tiers in Germany and Japan, the types of candidates chosen to contest SMD seats may be similar to those on PR-lists.Footnote 17
To sum up, there is considerable diversity amongst the assemblies examined here. Arguably, Germany with its high level of contamination between nominal and proportional electoral tiers, and the presence of dual candidacies in the GBT and JHR present a stricter, tougher test of my hypotheses than either the KNA or JHC where there is a clear separation of the two electoral systems allowing for a more ‘controlled comparison’ akin to a natural experiment (Herron et al., Reference Herron, Nemoto, Nishikawa, Herron, Pekannen and Shugart2018: 446). Nevertheless, across all four legislative chambers, the results were in line with my theoretical expectations about the link between electoral systems and MPs with a law background.
6. Conclusion
This paper offers the first academic study to address the question of whether electoral systems contribute to differential proportions of MPs with a law background by conducting an empirical examination of multiple parliamentary sessions in three advanced industrialized democracies with large legislatures that combine SMD-FPTP and closed-list PR electoral tiers: Japan, South Korea, and Germany. Across four chambers and 13 parliamentary sessions, we found more MPs with a law background in SMDs than in PR seats, and in a majority of individual parliamentary sessions, this difference was statistically significant.
Additionally, this study took into consideration mixed electoral system diversity by studying both parallel MMM systems and a compensatory MMP system. Across both systems, a much higher share of lawyer-legislators in FPTP-SMDs compared to closed-list PR seats was confirmed to be an empirical regularity. South Korea with a clear-cut separation of its two electoral tiers offered the most direct test of the paper's main hypothesis, but the fact that Germany and Japan (which practice dual candidacy) provided almost the same exact results provides additional tentative support for the article's theoretical expectation that SMD is more conducive to law graduates and lawyers entering parliaments than PR electoral systems.
Regarding possible mechanisms (i.e., self-selection, party-selection, and voter-selection) that may explain the link between SMD and lawyers and observed differences between SMD and PR elections, it will be important for future research to delve into which of these mechanisms plays a stronger role than the others by studying the occupational backgrounds of the candidate pool.Footnote 18 It will also be useful for future research to extend the scope of this study to other occupational groups, countries, and time periods as well as to sub-national legislatures. A broader multi-country analysis could also probe the selection effects of complementary institutions beyond just electoral systems and it would be worthwhile to examine the effects of electoral system changes from SMD to PR or vice versa and look into diversity amongst law graduates since persons trained in law can function in various roles as a sole practitioner, member of a law firm, as in-house counsel, in work for the government, or work in a field other than law (Chatoor, Reference Chatoor2013: 11). Lastly, another distinction worth considering in future research is the extent to which electoral systems are related to the election of MPs with a law background who formerly served as prosecutors, public defenders, and judges, or who hail from non-profit and activist backgrounds (Robinson, Reference Robinson2017: 680).
In conclusion, this study provisionally found single-member districts bringing citizens with a law background into legislatures at a greater rate than under PR elections even after taking into consideration national-level characteristics and factors such as election timing, partisan competition, and legislature size. Though these initial findings should be studied further in other contexts and with alternative methods, they tentatively suggest that SMD elections may be better suited for societies which aim to have a high share of lawyers or law graduates as legislators. On the other hand, it appears that those societies interested in minimizing or curtailing the share of lawyers and law graduates in their parliaments might be better served with PR electoral systems.
Supplementary material
The supplementary material for this article can be found at https://dataverse.harvard.edu/dataset.xhtml?persistentId=doi:10.7910/DVN/4KRXD3.
Dr. Devin K. Joshi is an Associate Professor of Political Science in the School of Social Sciences at Singapore Management University. He received his B.A. degree from Stanford University and Ph.D. from the University of Washington. His research interests include good governance, political development, and parliamentary representation. Co-author of the book Strengthening Governance Globally (Paradigm/Oxford University Press 2014), he has published over 30 peer-reviewed journal articles in venues such as International Studies Quarterly, Journal of Contemporary Asia, and Journal of Peace Research.
