The expertise effect: the impact of legal specialists’ intervention on the timely delivery of laymen's judgments

France relies exclusively on lay assessors to solve conflicts between entrepreneurs and employees for the longest time compared to any other civil jurisdiction. Given the lack of legal training of labor judges, the paper tests whether an intensification in the offer for legal services, whether coming from professional judges or specialized labor lawyers, induces delays in the procedure. Using a two-stage least squared estimating method and an innovative dataset on French labor courts between 2013 and 2017, the manuscript estimates the effect of an increase in the intervention of legal experts in labor litigation. The analysis showed that referrals to a professional judge delay labor cases’ solutions but refuted the long-standing economic theory that lawyers induce delay to acquire higher fees. This study indicates that the current structure of employment courts, entirely left in the hand of laymen with little knowledge of the ever-changing labor legislation, should be reconsidered to foster collaborative solutions accounting for industrial experience and, at the same time, legal competency.


Introduction
France is one of the few countries to rely solely on legally untrained lay assessors (prud'hommes) for the judicial resolution of labor disputes in the first instance. The court in charge of dealing with employment disputes is called Conseil de prud'hommes (CPH), and it is composed of lay judges that are occupationally close to the custom and practices of the parties (Ray & Rojot, 2013). The court was established to solve disputes in a peer-to-peer forum (Lefebvre, 2002) and 1 3 fulfill a significant social function (Machura, 2001) by imposing a constant dialogue between judges representing bourgeois and proletarians (Mouriaux, 1982). Today, all verdicts still require a cooperative solution that reconciles the views of distant social partners-unionists and employers associations' representatives-into a single co-written judgment. The balance between opposite views is the essence of the mandate of prud'hommes (Boulmier, 2002), and it is crucial for the conservative social order (Michel & Willemez, 2007) and industrial peace (Cottereau, 2002). Nevertheless, while in the past the norm was to settle litigation by pacifying social partners (Barthélémy & Cette, 2015) with a peaceful solution (Kieffer, 1987), the current situation does not allow real negotiations among litigants (Chappe & Doriat-Duban, 2003). A shared verdict would foster the law's protective function and economic efficiency (Barthelemy & Cette, 2015). However, the representation of both employers and employees simultaneously (Orif, 2015), complemented by a politicized judiciary, results in long delays and antagonistic arguments between social partners' views (Moriaux, 2006). The litigation market before CPHs is, hence, full of unfinished cases (Larivière, 1997).
The labor jurisdiction of prud'hommes is the most affected by litigation delays. Despite the smallest number of cases per year, the average processing time of employment cases is approximately triple compared to District Courts (Tribunals d'Instance) and more than double vis-à-vis High Courts (Tribunals de Grande Instance) (see Fig. 1). Eventually, all legal actors -plaintiffs, judges, clerks, and administrative staff -complain about the excessive duration of lawsuits among CPHs (Breda et al., 2017). Even though the primary goal of justice is not to convey cases to fast conclusions (Bufford, 1996), excessive legal cases chasing too few judges (Church, 1982) frustrates plaintiffs seeking compensation (Heise, 2000). It further involves significant social costs (Chappe, 2012) in terms of deterrence, accuracy (Fenn & Richmann, 1999), and a higher financial burden on litigants (Connolly & Smith, 1983). The theoretical reasons for litigation delays can be multiple: backlogs (Boyum, 1979), inconsistent jurisprudential case law (Berlemann & Christmann, 2017), complexity and uncertainty of the law (Di Vita, 2007), high procedural and bureaucratic formalism (Djankov et al., 2002), or inadequate calendaring systems (Somerlot et al., 1989) without third-party supervision (Ransom, 1923). 1 The court burdens taxpayers 2 more than any other jurisdiction (Serverine, 2012), with delays undermining the "very essence" of justice (Vereeck & Mühl, 2000) in diverting judicial resources solely to gain time (Di Vita, 2012).
A third crucial issue affecting prud'hommes is that despite the recent legislator's intention to professionalize the judges, they are not compulsorily required to hold a law degree (Ouaissi, 2017). The lay assessors' unfamiliarity with the law translates into a perpetual linguistic and legal insecurity in legal argumentation (Willemez, 2012). Eventually, it exposes judgments to legal flaws (Peskine & Wolmark, 2021), with judgments based on "humanity and common sense" (Gillet, 2015). In effect, the familiarity of prud'hommes with the industrial subculture is not per se sufficient to impose the structures and rationality that shall inform the legal thoughts of professional jurists (Newman, 1987). Traditional labor law is already challenged by the emergence of new technologies (Todolí-Signes, 2021), new recruiting methods, and atypical labor relations (Goldin, 2020). In recent years, all these difficulties have aggravated the already heated and nervous debates among prud'hommes, creating a real discontent among prud'hommes in doing their work (Béroud & Meynaud, 2021). The net result is that judges have interminable discussions (Masson, 2019), and delays in labor proceedings are notorious (Desrieux & Espinoza, 2019).
In this context, in which coherent legal reasoning is required by increasingly complex legal issues (Masclet et al., 1995), it is plausible that legal expertise might impact the final decisions of lay assessors (Martire & Montgomery-Farrer, 2020). Indeed, when courts follow legal cognitive errors, resources are wasted with 1 There are also quantitative indicators of judicial performances (Marciano et al., 2019), identified in external factors (Cashman, 2020), such as low financial endowment , and internal factors related to justice crucial players (Economides et al., 2013), such as lawyers, judges, rule reformers, scholars (Menkel-Meadow & Garth, 2010), and litigants (Miller, 1997), sometimes unwilling for a changeover (Gillespie, 1977). 2 French taxpayers have to compensate litigants for violations of the so-called "reasonable time principle" (Rainey et al., 2017) before the European Court of Human Rights (ECHR). The ECHR sanctions "unreasonable", "excessive", and "inordinate" (Edel, 2007) delays in all judicial matters (Schabas, 2015) with a "spectrum of reparations" (Ichim, 2014). Labor litigation is the one that imposes the highest number of convictions (Métin & Doudet, 2012) for psychological tensions (Poirier, 2013), "faulty functioning" of justice, frustration of litigants (Connolly & Smith, 1983) and credibility of the judiciary (Salado Osuna, 2012). unnecessary long decisions (Wistrich & Rachlinski, 2013). The prud'hommal ignorance of the law obliges them to recourse frequently to professional legal actors' advice (Bordieu, 1987). In a typical situation, legal technicians (Iweins, 2016) and juridical talents (Kelly, 2009) would seek logical legal argumentation to foster consistency (Schlanger, 2003) and enforceable and theoretically sensible litigation (Cotterrell, 2013). Nevertheless, in the context of prud'hommes, only lawyers and professional judges are qualified individuals with appropriate legal training (Perrin & Gaune, 2010) and informational advantage (Corby et al., 2019). On the one hand, professional judges reduce the number of disputes that cannot be solved (Cavanagh, 1998). On the other hand, lawyers contribute with their experience (Audier, 2016) and skills (Spurr, 1997) in giving valuable advice (Gordon, 2010).
The referral to professional judges (départage), acting as a tiebreaker (Desrieux & Espinoza, 2015), should be intended to find a compromise between employees' protection and entrepreneurial needs, end lay assessors' impasses (Hunter-Falck, 2015), and overcome particularly contentious principles of law (Tyl, 2008). The referral to professional judges, however, is costly. From a certain point of view, it increases the marginalization of non-legally qualified judges due to the complexity of the law (Burges et al., 2014) but reduces the historical and cultural context of judgment, elements not per se sufficient to shape judicial opinions (Garofalo, 2011) but essential in the labor domain. At the same time, the natural inclination of lay assessors, lacking legal training, to seek help from jurists to achieve rigorous, logical, and formalistic opinions (Posner, 2001) is costly in terms of time (Desrieux & Espinosa, 2020). In this framework, the role of lawyers in extending lawsuits' timing is unclear. 3 Empirical evidence suggests that attorneys overburden their law firms with excessive workloads and maintain an unsustainable volume of cases that prolong proceedings (Tahura, 2022). Nonetheless, in a setting in which judgments are delivered in a "non-canonical" way, legal technicians might endorse coherent legal reasoning to overcome legal complexity (Masclet et al., 1995) and create predictable behavioral patterns in the delineation of rights (Buchanan, 1974). Specialized lawyers represent professionals with a long-term perspective (Israël, 2019) and experts in the law's most complex legal and social aspects (Ramseyer & Rasmusen, 2010). Therefore, instead of inducing delay to wear opponent down (De Vos, 2004) and expand disputes (Mather, 2013), specialized lawyers might straightforwardly suggest the logical deductions to be written consistently in the judgments (Steiner, 2018). The net result could ultimately be reduced delays and conflicts in the litigation process among CPHs.
To answer these questions, the manuscript tests whether an intensification in referrals to professional judges (départage) and an increase in the concentration of specialized labor attorneys impact the duration of employment litigation, using an innovative dataset consisting of 201 labor courts between 2013 and 2017, and a twostage least squared estimating method. To instrument the endogenous variables of interest -referrals to professional judges and labor-specialized lawyers -the paper relies on three instrumental variables: the number of notaries per 10.000 inhabitants in 1865, the presence of a law school in the territorial jurisdiction of the court, and the ubiquity of the professional judge and prud'hommes. The results indicate that cases referred from prud'hommes to professional judges extend the litigation timing. At the same time, an increase in the market for specialized attorneys' services has no impact on the litigation pace. The final argument is that experienced jurists' intervention is dependent on many historical factors, and much of the troubles of employment litigation can be identified in structural flaws and unresolved issues.
The paper is organized as follows. Section 2 describes the data and the endogeneity issues to be tackled in the empirical strategy. Section 3 presents the empirical results discussed in the following Sect. 4, whereas conclusions are drawn in Sect. 5.

Data and methodology
The paper relies on a balanced panel dataset of 1005 observations, equivalent to 201 labor courts' annual observations, between 2013 and 2017. The data is publicly available from the Ministry of Justice (MoJ), the National Institute of Statistics and Economic Studies (INSEE), the Statistical Office of the European Union (EURO-STAT), the National attorneys' bar council (CNB), the Ministry of Education (MoE) and the National Library of Paris (NLP). The dataset is limited to Metropolitan France, excluding overseas departments and regions (DROM), overseas territories (TOM), and overseas collectivities (COM), in light of their special legal status, departmental autonomy, and different availability of the data. One limitation of the dataset is that the variance within courts is not visible. Public and manageable data is available only at the aggregate level to prevent forum shopping, and the French criminal law prohibits any kind of empirical analysis at the judge level. A second important aspect to consider is that MoJ official data lack any information regarding the lawsuits' duration in two courts: Carcassone and Créteil. Consequently, recognizing the impossibility of replacing missing values, the two courts were suppressed from the dataset, reducing the sample from 1,015 observations to 1,005.
The paper uses a two-stage least squared estimating method to analyze the data, including least trimmed quantile regressions as a robustness check, removing a proportion of largest and smallest observations to reduce outliers' influence on the explanatory variable (Neykov et al., 2012). The literature suggests an optimal α-trimming proportion between 5 and 10 percent (Konker & Bassett, 1978) to produce a reliable assessment of precision (Koenker & Hallock, 2001). Therefore, the dataset is trimmed at 2.5, 5, 7.5, and 10 percent of the highest and lowest observations of litigation duration to assess the true mean of our dependent variable conditional on the levels of all independent controls. The standard econometric model, upon which coefficients are obtained, is the following: time c,t = ß 1 ln(depart c,t ) + ß 2 as(law c,t ) + X c,t + α t + ε c,t

3
The dependent variable time c,t represents the monthly duration of lawsuits registered yearly for each CPH, and it is transformed according to the natural logarithm. The coefficients of interest are derived from two principal explanatory variables: ln(depart s,t ) and as(law c,t ). The first is the number of cases referred to a professional judge and labeled as "judged" by the Ministry of Justice (to exclude mere referral culminated in settlement or abandoned before the judgment). The variable is transformed according to the natural logarithm to interpret coefficients in percentage terms. The second explanatory variable is the number of specialized labor lawyers. 4 Specialized lawyers usually operate in the court in which they are resident because of a familiarity with the personalities and propensities of judges competent for their specific discipline (Kuhne, 2019). The variable cannot be transformed according to the natural log to estimate elasticities because of some zero values. Therefore, the variable is multiplied by ten and transformed according to the inverse hyperbolic sine (or arcsinh). The multiplication does not affect slope estimates (Wooldridge, 2013), but the multiplication reduces approximation error in interpreting arcsinh results as elasticities to less than half of a percent (Bellemare & Wichman, 2020).
It has to be acknowledged that the relationship between litigation and its crucial actors gives a standard endogeneity problem (Ginsburg & Hoetker, 2006), with lawyers inducing unnecessary litigation  and vice-versa (Buonanno & Galizzi, 2014). In our case, it is conceivable that the proportion of cases sent to départage is related to the low legal expertise of prud'hommes, which might have lain down in resignation over difficulties in favor of their professional judicial colleague. Correspondingly, an increase in the number of specialized labor attorneys might induce swifter solutions for judges, who might rely directly on the legal conclusions given by professional experts in the field. The literature struggles to use institutional insight (Angrist & Pischle, 2008) to find appropriate instrumental variables (IV) for legal actors (Mora-Sanguinetti & Garoupa, 2015). Educational costs suggest that proximity to law schools might be a good instrument for lawyers (Carmignani & Giacomelli, 2010). However, French students' mobility is high (Lemistre & Moreau, 2009) because public benefits (e.g., social security and housing allowance) encourage enrollment in universities far from the parental nest (Ryan, 2001). Therefore, in light of the difficulty in addressing the exact direction of mobility, but keeping in mind that legal professionals enjoy, in France, a close relationship with law faculties (Dubury, 2001), we instrument our endogenous variables with a dummy equal to 1 in the presence of a law faculty in the jurisdiction of the CPH and zero otherwise.
A second IV proposed to tackle endogeneity issues is the rate of notaries per 10,000 inhabitants, registered by the Ministry of Justice in 1865, when France had already acquired Nice and Savoy but had not yet lost Alsace and Lorraine to the German Empire. The variable is sufficiently far in time from our period of analysis to be reasonably considered exogenous. The figure of the notary intervenes in complex legal issues (Tissot, 2018) to sort out complexity (Delmas, 2020) and produce acts with warranted public trust and high juridical accuracy (Deschenes, 1983). Nevertheless, while they share with judges the best legal knowledge, paid consulting practice, and even teaching activity (Débax, 2017), they endorse a different vision from lawyers who are intrinsically associated with the dispute and litigation in favor of their client (Monahan, 2015). Therefore, we should expect a positive correlation between notaries and professional judges because both are complementary in using the public trust to interpret the complexity of the law at its finest (Jamin & Xifaras, 2015). On the contrary, lawyers' role is not to pursue the law but to challenge its principles in favor of their clients (De Vareilles-Sommières, 2016), according to a private entrepreneurial logic far from the defense of the public interest (Delmas, 2015).
Finally, considering that professional judges are said to marginalize non-legally qualified colleagues (see above, Burges et al., op.cit.), it is conceivable that départage is affected by the ubiquitous presence of both judiciaries in the same building. A ubiquity might induce prud'hommes to fulfill their mandate without bothering their colleagues, in light of the well-known "conformity" psychological phenomenon that induces, for fear of ostracism and "normative influence" (Kassin et al., 2020), behavioral changes to match the actions of others in face-to-face interactions (Ciccarelli & White, 2018). Indeed, there is evidence that lay judges conform to the expectancies of others (Kaplan & Miller, 1987) in the hope of fitting in and appearing suitable for work (Hewlin, 2009). Therefore, we introduce a dummy IV that takes a value of 1 if the referring tribunal and prud'hommes operate in the same courthouse and zero otherwise. It is reasonable to expect a negative correlation only with lawsuits ended in départage. Lawyers should be indifferent whether courts' offices are located in the same building.
The model includes a vector of control variables, represented by X c,t to account for legal and socio-economic determinants of litigation delay. All controls are constructed by aggregating the yearly data available at the urban agglomeration level, reshaped according to the area of jurisdiction of all CPHs. Control variables are lagged back to 1 year, a method that is considered the most effective strategy (Reed, 2015) and capable of wiping out potential simultaneity with the dependent variable (Bellemare et al., 2017). Legal procedural controls include the rate of conciliated and abandoned lawsuits -representative of settlement and premature discontinuation of the case -the share of backlogged and cleared cases -capturing the congestion of the court and the proportion of judgments in comparison to incoming cases -as well as the claimants' losing rate -indicative of a court that is less prone to give reason to claimants. The socio-economic controls include the rate of the "intermediaries" -regrouping professions between the manager and the employee like a liberal nurse or a sports instructor -and the rate of workmen, employees, executives, artisans and traders among the active population. Further controls include the unemployment rate -crucial for courts' engorgement since almost 95 percent of lawsuits are brought for employment termination -and "halo" individuals -not working and not actively looking for a job or not quickly available to occupy one. Both variables are important: an increase in unemployment jeopardies the possibilities of labor reintegration and might justify a more protracted and exhausting legal action. At the 1 3 same time, halo individuals are synonyms for barriers to the availability of employment. The model also includes yearly indicators (α t ) to restrict the correlation of unobserved outcomes components within annual observations. Finally, ε c,t represents the usual error term. Errors are two-way clustered at years and courts' level.
Descriptive statistics of all previously described variables are provided in Table 1.

Empirical results
A reduced form of the first-stage coefficients and cluster-robust test statistics for weak and under identifications are provided in Table 2. Each column represents a different alpha-trim quantile, according to the sample size suggested in Koenker & Bassett (op. cit.). The complete estimations and all first-stage tests of the robustness of the instruments are reported in the Appendix (see Tables 4 and 5).   The instruments are robust to heteroscedasticity and autocorrelation (Baum et al., 2007) but must pass several tests to be accepted. To deem each instrument valid, the rule of thumb is a first-stage F-statistic above 10 (Staiger & Stock, 1997). Nevertheless, this value is a conservative lower bound for multiple instruments (Bun & De Haan, 2010). The most recent literature (Stock & Yogo, 2005) suggests more precise critical values for the Sanderson-Windmeijer tests on under-identification and weakly identified instruments, to be set at 13.91 (5 percent significant level) and 12.83 (15 percent maximal IV size). The first stage F statistics and weakness and under-identification tests are above the mentioned thresholds.
The second stage of the two-stage least squared regressions are reported in Table 3. Again, each column represents a different proportion of the alpha proportion smallest and largest observations removed from the dataset. The clustered standard errors imposed an alternative to the Durbin-Wu-Hausman test, which was no longer valid. The difference-in-Sargan statistic test suggests that the instrumented variables are not exogenous. Therefore, it was corrected to consider both instrumented variables as endogenous.

Discussion
The empirical coefficients of the first stage of the two-stage least squared regressions, reported in Table 2, suggest a positive correlation between notaries and lawsuits' referrals to professional judges. In the first place, judges and notaries are complementary figures, being both the reflection of today's society (Dauptain, 2017), invested with the function of serving the law to improve the life of our complex societies (Darrois, 2010). In the second place, history suggests that the notarial practice gave birth to modern French law, mastering ever-increasing complex legal questions (Hilaire, 2004). In this sense, the notarial profession presupposes the existence of complex juridical institutions that require a precise, exact, and straightforward legal language (Cardinal, 1956(Cardinal, , 1967. Therefore, it is reasonable to assume that a higher notarial concentration of notaries is attached to an apparent societal complexity in its legal implications, thereby increasing referrals to professional judges from the lay assessors. On the contrary, the negative correlation between labor-specialized attorneys and notaries can be explained by the different legal "tone" and "distinct professional culture" (Monahan, op. cit) between the two figures of legal practitioners. The notarial profession is a category of jurists that prevents lawsuits (Bernard, 1997), while attorneys act in the opposite direction. The labor domain is one of the matters excluded from the scope of application of notarial activity, where lawyers feed off controversy and defend their clients to the hilt. Therefore, the negative correlation between lawyers and notaries can be explained by their different roles in society: on the one hand, they pursue the finest and most detailed juridical solution to public trust issues (notaries) and, on the other hand, ensure a fair trial in the name and interest of a private client's interests (lawyers).
The first-stage empirical results regarding law schools show two positive correlations with the "unions de formation et recherché en Droit" (law faculties). That is explained by the fact that both judges and attorneys share the same educational institution. The French law schools impose lengthy courses in substantial law and civil procedure, creating a local culture saturated with the consciousness of law and legal rights (Schechter, 1996). Despite the high general flexibility of students in France, the peculiarity of legal specializations is that they reduce professional mobility (Bessy, 2010). Highly skilled jurists tend to concentrate where an alma mater is present because academia continues to find the audience for doctrinal scholarship in judges and legal practitioners (van Gestel et al., 2017).
Always sticking to first-stage empirical results, we uncover that the connivance between prud'hommes and professional judges in the same courthouse has important effects on départage. In particular, the ubiquitous presence of both judicial figures in the same building reduces the involvement of the professional judge in the decision-making process. The explanation for this result can be found in a reciprocal monitoring mechanism (Palumbo, 2006) and a "need to belong" effect, which is a form of psychological "conformity" constraint that fulfills individuality and uniqueness with interpersonal relations and membership in a group (Hornsey & Jetten, 2004). This latter concept is particularly suitable for the prud'homale environment since lay assessors follow the typical behavior of those who reinterpret their task and rationalize their behavior to see themselves as independent (Kassin et al., 2017). Conversely, the ubiquity of supervisors and supervisee judges has no statistically significant relationship with labor attorneys. Lawyers are indifferent to the proximity between judges because eventual travel expenses to reach faraway courts are customarily billed to the final client.
The controls reported in Tables 4 and 5 suggest a negative correlation between the claimants' losing rate and both lawsuits sent to professional judges and labor specialized attorneys, even though the statistical significance of this result is not persistent across quantile. The negative correlation with specialized labor lawyers is intuitive: the recourse to experts in employment law should be associated with an increase in the chances of winning the case. That might explain why an increase in the loss rate is associated negatively with specialized attorneys. Simultaneously, a higher number of rejections of claimants' arguments is associated with fewer referrals to professional judges, suggesting that a more attentive legal eye might be good for plaintiffs seeking legal accuracy and recourse to more experienced jurists. It is also interesting to notice that a higher proportion of workmen, artisans, and employees in the active population is associated with fewer referrals to professional judges and fewer labor-specialized attorneys (except for the category of artisans, which is not statistically significant for lawyers). All categories are simple and well-delineated employment classes, existing since immemorial time, with well-defined rights and extensive jurisprudence. The opposite is true for higher intermediary professionals, which are a fluid category of workers (Deauvieau & Dumoulin, 2010), requiring legal expertise (both from the judges and the lawyers) in light of the particular legal complexity in the definition of their rights and duties (Buisson & Gadille, 2016). Finally, while the rate of halo individuals negatively affects départage -because an increase in the population neither in employment nor in education nor training is less of interest to labor judges and more for social services (Frigul & Depoorter, 2010) -the unemployment rate, on the contrary, involves complex security law texts in need for judicial clarification (Willman, 2009).
Moving to the results of the second stage of the two-stage least squared regressions, we grasp some confirmations of international concerns (Fiorentino, 2017) about delays in employment procedures, where referrals to professional judges have been considered a particularly harmful incident for stakeholders (Serverine & Vennin, 1995). An increase in labor lawsuits sent to départage compromises the normal state of affairs because of the possibility of resetting the litigation process to zero before the "new" judge. The most common causes of referral to professional judges are related to dismissal for economic reasons. These cases promote ideological tensions among lay assessors and increase difficulties in mastering the legal issues arising with them (Chevillard, 1997). If the issue is legal complexity, the policy-making suggestions would be to provide prud'hommes with compulsory and extended training in civil procedure, judicial drafting, and ethics, as well as guidance on judicial scheduling and follow-up, which are said to reduce procedural delays (Durand & Henriot, 2014). Commentators argue about the impossibility of prud'hommes in delivering fair, credible, and effective justice on a human and economic level (Fricero, 2021). Nevertheless, a solution might exist if the legislator envisages leaving prud'hommes dealing with the initial conciliatory phase -in which social partners exert their reconciling influence on litigating parties -and, in case of failure, the litigation phase and final verdict to a professional judge.
The vexed hypothesis of lawyers-inducing-delay (Edwards, 1986;Partnoy, 2012;Wistrich, 2008) appears to be unfit for France, where judges portray the logical deductions drawn from attorneys' arguments (Steiner, 2018). The American "sweepingly negative view" of lawyers and their dilatory techniques (Clark, 1992) is balanced by their role in helping keep judgments very short (Andenaes & Fairgrieve, 2015) with straightforward arguments to be applied by the judges. In the specific context of the CPH, specialized attorneys are the only professional figure with indepth knowledge of the law's substantive and procedural rules (Blasi, 1995), able to forthrightly diagnose and tackle legal problems (Krieger & Neumann, 2015). However, the negative correlation between specialized lawyers and employment litigation delays is not statistically significant. This potentially surprising result might be explained by a marginal role of lawyers in the scheduling of oral arguments, which are all in the prud'hommes hands. Highly specialized lawyers suggest legal arguments to be used by lay judges, much less proficient in the law (Stimec & Adijès, 2015). Nonetheless, the timeline for the final translation of these into concrete legal documents is incumbent on the final judicial arbiter.
Eventually, the empirical evidence suggests that some institutional glitches jam the market for labor litigation. An increase in settlement before the judge is beneficial for delay reduction because it avoids consuming litigation (Becker, 2011). Nevertheless, since statistics on dispute resolution mechanisms are in constant decline, this hypothesis in a real scenario is very unlikely. Simultaneously, an increase in the number of lawsuits that are abandoned, cleared, and backlogs are a source of delay. The backlogs are known to hinder courts (Miles, 2012), and further congestion entails an increase in judicial caseload and delays. A higher number of abandoned instances is a waste of time and resources spent on a dead track. Interestingly, increasing lawsuits' clearance rate determines the accumulation of new cases with older lawsuits, delaying the case's solution for a longer time. This evidence might suggest a spiral of inefficiency, in which a higher clearance rate is counterbalanced by congestion of additional litigation. Future studies might investigate further aspects. Concerning socio-economic controls, the different categorizations of labor in the active population have different delaying or time-reducing effects. An increase in active individuals in the intermediary category decreases litigation's duration, even though its statistical significance is at the limit and does not persist after trimming the dataset. The artisans, traders, and business leaders categories embrace individuals in the trade and commercial services, a sector in which borders between goods and services have become quite complex and subtle in recent years (Broussolle, 2012). A shining example is the ever-increasing standardization of work through algorithms (Rosenblat, 2018), which generated many confusing conclusions among prud'hommes (Cherry, 2019). In this sense, the empirical evidence might suggest that lay assessors struggle to keep up with the ever-shifting legislation and the legal complexity inherent in specific categories of workers that induce delays because of thorny labor interactions.

Conclusions
This paper investigated the role of attorneys and professional judges on litigation pace in a judicial setting ruled by lay assessors. The empirical results reported in the manuscript suggest that professionalized judges and lawyers are important figures in the litigation process and that they belong to a system forged in the long history of France. In the end, professional judges step into litigation to solve interpretative disputes between lay assessors, and this intervention is costly in terms of time. At the same time, the empirical results reported in this manuscript disprove the debated lawyers-inducing-delay hypothesis because of historical reasons and practical arguments related to the litigation process, which is affected by institutional glitches and excessive under-endowment (Tournaux, 2016).
It is not uncommon to see tribunals relying upon lay assessors to deal with sensitive legal issues. In our case, we have seen that the Conseil de prud'hommes is a unique setting in which lay assessors have entirely to deal with complex employment legal issues without proper legal training. The context of labor litigation is critical because both employers and employees are likely to prefer to foresee pathological situations of professional breakdown and employment disturbances. If lay assessors resolve labor conflict based on instincts and intuitions, emotional pressures might undermine the prompt and foreseeable resolution of employment disturbances. Lawyers and professional judges intervene to support lay assessors at their best but considering that several labor courts sent more than 50 percent of the total cases to a professional judge, there is a significant problem of division of tasks and responsibilities that must be resolved.
In recent years, the French legislator has focused on contractual flexibility and competitiveness, leaving aside flawless solutions to potential contractual failures without touching the too cumbersome and time-consuming labor procedural litigation schemes If the parties are already litigating before a court, either they are incentivized to settle the case -e.g., with the threat of a harsher punishment if parties' arguments reveal to be true -or they will not cooperate to reach a prompt solution. The non-specialization of lay assessors is a critical factor in determining delays because of the growing modernization and complexification of modern labor law. In this regard, lay assessors -in light of their previous syndical professional activity -might test the water for potential points of agreement and then pave the way for professional judges if their conciliatory attempt fails. This policy-making solution could be a trade-off envisaging legal actors' cooperation and interaction between industrial experience and legal proficiency. Another simple solution to reduce delays in employment litigation could be to reduce the backlogs with effortless but effective methods that have proved, in the past, to be highly successful in this respect. 5 It is recommended that further research shall be undertaken to analyze how litigants would react to incentives favoring the amicable settlement of disputes and reducing their options in delaying litigation settlement or the final resolution of employment disturbances. Eventually, it is plausible that some limitations might have influenced the results obtained in this study because measuring the efficiency of legal systems is exceptionally challenging. For example, lagged litigation statistics significantly reduce prospective endogeneity bias, but they do not rule it out entirely. Furthermore, the limitation of the available data at the aggregated level to comply with existing legal limitations of judge-analytics express the evident difficulty of assessing justice quality and its efficiency. Nevertheless, the design and development of multidimensional studies on complex multi-actor decision-making processes is a challenge that we must face to foster justice served equitably but also efficiently.

Table 4
Estimation results of the first-stage regressions and weakness tests  Two-way clustered standard errors, at year and id level, in parentheses. The table shows the results of the first stage of the two-stage least squared and least trimmed quantile regressions, between the number of referrals to professional judges (départage) and a set of instrumental variables (notaries in 1865, law schools, ubiquity professional and lay judges), in a dataset consisting of 201 labor courts, between 2013 and 2017 ***p < 0.01, **p < 0.05, *p < 0.1 Table 5 First-stage regressions on labor specialized attorneys and weakness tests 1st stage (1) (3) Labor att.
Year Two-way clustered standard errors, at year and id level, in parentheses. The table shows the results of the first stage of the two-stage least squared and least trimmed quantile regressions, between the labor specialized attorneys (Labor att.) and a set of instrumental variables (notaries in 1865, law schools, ubiquity professional and lay judges), in a dataset consisting of 201 labor courts, between 2013 and 2017 ***p < 0.01, **p < 0.05, *p < 0.1