ECONOMIC ANALYSIS OF THE NEW CIVIL FEES SYSTEMS IN SPAIN: IMPACTS ON CONGESTION AND ACCESSIBILITY OF ENFORCEMENT INSTITUTIONS 1

Transcription

1 ECONOMIC ANALYSIS OF THE NEW CIVIL FEES SYSTEMS IN SPAIN: IMPACTS ON CONGESTION AND ACCESSIBILITY OF ENFORCEMENT INSTITUTIONS 1 Juan S. Mora-Sanguinetti & Marta Martínez-Matute 2 Banco de España-Eurosystem Version: June 15, Simposio-Spanish Economic Association (SAEe) [Bilbao, December 2016]. Abstract Compared to other European countries, Spain has traditionally lacked a general system of court fees. At the same time, Spain has one of the highest rates of litigation of the OECD, with impacts on the efficacy of the enforcement institutions and general economic efficiency. In 2002, the Congress passed a system of court fees forcing the legal entities and enterprises to pay a fee in the case they used the judicial system. In 2012, the fees were extended to individuals. As a result of a strong opposition by the legal profession, the system of 2012 was amended in 2013 and abrogated in These systems, while having a bounded period of enforcement, allows us to make a precise experiment on the impact of court fees in a judicial system. Thus, this article examines empirically whether court fees significantly reduced the litigation rates and the congestion of the civil jurisdiction in Spain. In order to do this, we collected a broad database of quarterly data on the real workload of Spanish courts. This study concludes, somehow surprisingly, that the effects of the fees were not homogeneous. They depended on the type of civil judicial procedure used by citizens and businesses. The adoption of the system of 2002 reduced the litigation rate under the form of verbal or exchange judgments but did not reduce the congestion rates in general except in the case of the ordinary procedures. The extension of the fees in 2012 reduced the litigation under the form of exchange judgments and payment procedures but had reverse effects on the ordinary procedures. Consequently, the congestion of the judicial system was only reduced in a limited set of procedures. This research also finds that the observed effects were stronger in the provinces where the courts showed higher amounts of unresolved cases. JEL Classification: K41, E51, G2 Keywords: enforcement institutions, litigation rates, courts fees, judicial efficacy. 1 We are grateful to Ildefonso Villan Criado (CGPJ) for his advice in the use of the judicial performance data. We also thank Gabriel Doménech and the seminar at the Bar Association of Valencia ( Los costes de la justicia: costas procesales, tasas judiciales y justicia gratuita, Valencia, 2015) for their comments and suggestions. The views expressed are those of the authors and should not be attributed to the Banco de España or the Eurosystem. 2 All authors serve at the Research Department of the Banco de España-Eurosystem. Contact information: Juan S. Mora-Sanguinetti, Senior Economist and Lawyer. juans.mora@bde.es. Tel. (+34) Fax: (+34) Banco de España. Alcala 48, 28014, Madrid. 1

2 1 Introduction Nowadays almost all European countries charge a fee for the use of the judicial system. According to the European Commission for the Efficiency of Justice (CEPEJ, 2014) in 2012 there were court fees, with different legal configurations and tax burdens, in all countries covered by the Commission (42) except for Luxembourg. Spain entered the group of countries with an active system of court fees in With the Law 53/ Spain adopted a system of fees paid by legal entities and enterprises (which we will call in this article: "old" fee system). Later, in 2012, after the approval of the "new" court fees Law (Law 10/2012), the payment was extended to individuals. The extension of court fees to individuals generated a strong opposition by the legal profession and it was recurred several times before the tribunals (for an extensive analysis of the case law on court fees, see DOMÉNECH, 2014). The "new" fees survived in their original configuration, only for a short time while the Royal Decree Law (RDL) 3/ amended them and, in , they were abrogated for individuals (thus, since then, they were only forcing legal entities and businesses). As it will be discussed below, the court fees in their old configuration were applicable in the civil and administrative jurisdictions and were extended to the social (labour) jurisdiction later. It should be noted that none of the reforms (the old or the new ) affected the rules about who should bear the payment of the fees (the fee-shifting rules). In any of their configurations, the fees were paid by those who initiated the conflict before the courts, regardless of which person or business turned out to be winning part of the conflict. Spain seems to be an interesting country to study empirically whether or not court fees meet their legal and political objectives. Spain has one of the highest levels of litigation at the international level. Specifically, it would have the third higher litigation rate of the OECD, whether measured in per capita terms or in relation to the country's wealth 7 (PALUMBO et al. 2013a and 2013b) 8. These results are represented in Figure 1. Litigation rates have a direct (and negative) impact on the levels of efficacy of the judicial system (PALUMBO et al., 2013a) and this has several negative impacts on economic efficiency 9. Along with this, another interesting fact is that in the long term, the revenue achieved with the fees has been historically low. 3 Spain had no fees since 1986, when the system of 1959 was suppressed. The abrogation took place through Law 25/1986 of 24 of December de supresión de las tasas judiciales. 4 Law 53/2002, of 30 of diciembre, de Medidas Fiscales, Administrativas y del Orden Social. 5 RDL 3/2013, of 22 of February, por el que se modifica el régimen de las tasas en el ámbito de la Administración de Justicia y el sistema de asistencia jurídica gratuita. 6 RDL 1/2015, of 27 of February, de mecanismo de segunda oportunidad, reducción de carga financiera y otras medidas de orden social. 7 Spain is the country with the third highest per capita rate of litigation (after Russia and the Czech Republic, followed by Greece and Italy). The litigation rate is measured as the ratio between the number of new conflicts of civil nature brought to the courts in a given year in relation to the population. In addition, Spain is the country (after Russia, the Czech Republic and Greece and followed by Italy) with the fourth highest rate of litigation brought to the courts in a given year in relation to its GDP (measured in PPP -purchasing power parity- in US dollars). 8 The countries (or legal systems) covered by the research were: Australia, Austria, Belgium, the Czech Republic, Denmark, England and Wales (United Kingdom), Estonia, Finland, France, Germany, Greece, Hungary, Iceland, Ireland, Israel, Italy, Japan, Korea, Luxembourg, Mexico, Netherlands, New Zealand, Switzerland, Northern Ireland (United Kingdom), Norway, Poland, Portugal, Russia, Scotland (United Kingdom), the Slovak Republic, Slovenia, South Africa, Spain, Sweden, and Turkey. 9 Through effects on entrepreneurship and the size of companies (GARCÍA-POSADA, M. and J. S. MORA- SANGUINETTI (2014 and 2015) or even the efficiency of the housing markets (MORA-SANGUINETTI, 2012). 2

3 Figure 1: Litigation rates per capita and as a percentage of GDP (civil jurisdiction) Note: In dark blue: ligation rates per capita. In light blue: litigation rates divided by the GDP (corrected by PPP). Source : MORA-SANGUINETTI (2013) and PALUMBO et al. (2013a and 2013b). The literature identifies different objectives raised by governments to justify the adoption of a system of court fees: from a policy of "deterrence" of the dispute (with the ultimate goal of reducing the "congestion" of the judicial system) to an instrument of provision of funding for the administration of justice. This last objective, as it was discussed before, seems to have been not achieved in Spain (at least relative to other countries) 10. For a review of this literature see, among others, SHAVELL (1997) or MERY NIETO (2015). The literature makes it clear that neither extreme is beneficial: low (or inexistent) court fees may lead to high levels of litigation (which lead to judicial collapse and thereby reduce the effective exercise of the right of citizens to access the judicial system) but a high court fees may prevent some "meritorious" cases to arrive to the courts. This article analyses the impacts of the systems of court fees adopted in Spain in 2002 and 2012 on the litigation rates by civil procedure and the efficacy of the judicial system. To do this, we collected a novel database of the workload of Spanish courts at the local level by quarters between 2001 and The article concludes that the introduction of the court fees of 2002 limited the number of conflicts that took the form of exchange or verbal judgments but, by contrast, did not reduce the congestion rate of the civil jurisdiction except in the case of solving ordinary procedures. The extension in 2012 of fees to individuals reduced the number of conflicts brought by citizens to the courts in the form of exchange or payment (monitorio) procedures. Although this effect could be expected, the new fees had the reverse effect on the ordinary procedures. The congestion of the judicial system was only reduced in a limited set of procedures and not in all of them. Moreover, the effect of the fees were not homogenous among the Spanish provinces (although the design of the judicial system is common to all of them). The rest of the article is organized as follows: section 2 analyses the adoption of the system of new court fees of 2012 and its evolution in the long term, paying attention to its gradual abrogation in 2013 and Section 3 provides the details on how we have built the database and the judicial indicators used in the econometric exercise. Section 4 presents the empirical strategy of this article and the controls included in the estimations. Section 5 shows the results of the models and discusses the main findings. Finally, Section 6 provides some brief conclusions. 10 According to the CEPEJ, in 2010 the court fees covered 4.1% of the budget of justice in Spain, compared with the European average of 22.3%. 3

4 2 The systems of court fees in Spain ( ) Since the abrogation of the court fees system of 1959 (which took place in 1986) 11, there has been two recent schemes of court fees: the fees of the Law 53/2002 (named in this article as "old" fee system) and the system of "new" court fees passed with the Law 10/2012, which was amended afterwards. Figure 2 provides an outline of the court fees enforced in Spain through time and the jurisdictions affected by each system. The "old" court fees were in force between 1st January 2003 to 21 November The payer of these fees were only the legal entities and enterprises and were payable in the case of a conflict in the civil or the administrative jurisdictions. Fees were made up of a fixed part (which depended on the type of procedure used) 12 and a variable part, consisting of 0.5% of the amount disputed (if the amount was below 1 million euros) and 0.25% if the amount was over 1 million. This variable part of the fee had a maximum of The new fee system entered into force on 22 November 2012, abrogated the old scheme and extended the payment to individuals, keeping them also for legal entities and enterprises. The "new" fees were also payable in the social jurisdiction for first time (as well as in civil and administrative jurisdictions). As in the previous system, the fees included a fixed part, by type of procedure 13 and another part which was variable (consisting of 0.5% of the amount of the procedure up to 1 million euros and 0.25% if the amount was over 1 million). The variable part had a maximum of 10,000 ). As it was already discussed, the extension of the court fees to individuals generated much controversy and several judicial appeals (DOMÉNECH, 2014). In a short time (around 3 months), the government decided to amend the new system through the Royal Decree-Law (RDL) 3/2013 ("Reform I" in Figure 2) (which entered into force on February 24, 2013). This reform diminished the variable part for individuals to 0.1% of the amount of the dispute and established its maximum in Finally, in a second reform ("Reform II"), which entered into force on 1 February 2015 (RDL 1/2015), the court fees for individuals were abrogated. As it was introduced, this article investigates the impacts of court fees (both the "old" and the "new" system) on the litigation rates and the efficacy of the courts of the civil jurisdiction. A definition of both measures (litigation and efficacy) will be provided in section 3. This paper also studies the impact of the first reform of the new fees ("reform I", i.e. the reform of the variable part of the fee for individuals). There is no sufficient statistical information as for today to fully cover the analysis of "reform II", which completely abrogated the fees for individuals. It seems relevant to note that the adoption of both the old and the new fees in Spain has not coincided with other major legislative initiatives in the civil procedural regulations. In fact, the most important change in those regulations took place in 2000 with the new Civil 11 Op. cit. footnote The fixed part differs among the different civil procedures (verbal, ordinary, payment or exchange procedures or in the case of an extrajudicial execution or an insolvency proceeding). For instance, if the procedure used is the verbal one, the fixed part of the fee was 90 and if it was ordinary, 150. Also different types of appeals (recurso de apelación, recurso de casación o recurso por infracción procesal) had specific fixed payments. Different fixed parts of the fee were also applicable if the procedure was initiated in the other jurisdiction affected by the fees (the administrative jurisdiction). 13 In the civil jurisdiction, the fixed part differed among different procedures (verbal/exchange, ordinary or payment procedures) or in the case of an extrajudicial execution, a specific insolvency proceeding (concurso necesario) or different types of appeals (recurso de apelación, recurso de casación or recurso por infracción procesal). If the procedure used was the verbal one, the fixed part of the fee was 150 and it was 300 if it was an ordinary procedure. Different fees were payable in the other jurisdiction affected by the court fees system of 2012 (the administrative and the social/labour jurisdiction). 4

5 Procedural Law (CPL) (Law 1/2000) which entered into force on January 8, 2001, 14 therefore, before the entry into force of the "old" fees analysed here. Figure 2: The system of court fees in Spain after 2002 New CPL (Law 1/2000) Old Law on court fees (Law 53/2002) New Law on court fees (Law 10/2012) Reform I (RDLaw 3/2013) Reform II (RDLaw 1/2015) Time 8/1/2001 1/1/ /11/ /2/2013 1/2/2015 Fixed amount per civil procedure, fixed amount per administrative procedure + VARIABLE amount (0.5% to 1m) (0.25% for the rest). Maximum Fixed amount per civil procedure, fixed amount per administrative procedure, fixed amount in the social jurisdiction (casación and suplicación) + VARIABLE amount (0.5% to 1m) (0.25% for the rest). Maximum Fixed amount per civil procedure, fixed amount per administrative procedure, fixed amount in the social jurisdiction (casación and suplicación) + VARIABLE amount (0.5% to 1m) (0.25% for the rest). Maximum Fixed amount per civil procedure, fixed amount per administrative procedure, fixed amount in the social jurisdiction (casación and suplicación) + VARIABLE amount (0.5% to 1m) (0.25% for the rest). Maximum Court fees paid by legal entities and enterprises Fixed amount per civil procedure, fixed amount per administrative procedure, fixed amount in the social jurisdiction (casación and suplicación) + VARIABLE amount (0.5% to 1m) (0.25% for the rest). Maximum Fixed amount per civil procedure, fixed amount per administrative procedure, fixed amount in the social jurisdiction (casación and suplicación)) + VARIABLE (0.1%). Maximum Court fees paid by individuals Source : Own elaboration 14 This Law abrogated the old CPL of Royal Decree of February 3rd (1881), de promulgación de la Ley de Enjuiciamiento Civil. 5

6 3 Measuring litigation rates and efficacy rates in the Spanish civil jurisdiction We have compiled data from the General Council of the Judiciary (CGPJ) at the court level, by type of procedure and on a quarterly basis (from the first quarter of 2001 to the second quarter of 2015). The database of the CGPJ provides statistical information on the actual workloads of the courts and tribunals. Therefore it makes available information on flows of new conflicts arrived to the court system (channelled to the judicial system by the lawyers), solved conflicts and pending conflicts (waiting to be solved in the pile of the specific court) by quarter. In this paper we analyze the impact of court fees in the performance (and litigation rates) of the civil jurisdiction. Figure 2 provides a scheme of the Spanish judicial system. Figure 2: Simplified scheme of the Spanish judicial system Conflicts of an individual or a legal entity Conflicts of private nature Conflicts with the administration Threats to the public interest Conflicts with an enterprise related to the labour law (dismissal) Civil jurisdiction Jurisdiction for suits under administrative law Criminal jurisdiction Out-of-court conciliation Mercantile courts First instance (and instruction) courts Labour (social) jurisdiction concursales Declaratory judgment Conciliation hearing Ordinary judgment (ordinario) Verbal judgment Payment procedure (monitorio) Exchange and cheques (cambiario) Trial Execution Execution Execution Execution Source: Own elaboration. 6

7 Figure 3 graphs the entry of new cases in the civil (and commercial) jurisdiction by quarter and by type of procedure. Figure 4 shows the annual growth rates in the number of new cases in the civil jurisdiction by procedure. There seems to be an increase in those numbers even after the entry into force of the various systems of court fees. This may have to do with, on the one hand, the increase in the GDP and the complexity of the Spanish economy in the long run and, on the other hand, with the increase in workloads brought by the crisis environment in the last years. These observations seem to coincidence with previous findings by the literature. Following Palumbo et al. (2013a), the number of cases which are brought to the court system of a country is a function, among others, of the frequency of disputes in a society, which is influenced by the volume and complexity of economic transactions. In addition, the crisis environment (since 2008) 15 may have brought some more conflicts to the judicial system because during a recession there may be more breaches of contracts, resulting in increased litigation 16. In addition to the cycle considerations there may be other factors may affect the rate of litigation, such as the incentives affecting the market for lawyering or the complexity of the local economy (MORA- SANGUINETTI and GAROUPA, 2015). All this justifies the need to construct an econometric model taking into account all the mentioned factors in order to find out if there was an effect of the court fees on the litigation rates. 15 The crisis of the Spanish economy began in the fourth quarter of 2008 and lasted until Although this is what it seems to be observed in the Spanish case, the interaction between litigation and the economic cycle is still being discussed in the literature. For instance, under some assumptions, GINSBURG and HOETKER (2006) argue that litigation should increase in economic booming instead. MORA-SANGUINETTI et al. (2016) analyze the effects of judicial performance on the credit market, and suggest that those effects, which show differences among the different phases of the economic cycle, may partially depend on the specific judicial procedure analyzed. In our sample, beginning in 2001, we cover several years of expansion and recession. 7

8 Figure 3: Entry of new cases in the civil (and commercial) jurisdiction by quarter and by type of procedure Source : Own elaboration using CGPJ data (2015). Source: Own elaboration using CGPJ data (2015). 8

9 Figure 4: Annual growth rate in the number of cases admitted in the civil jurisdiction by procedure Source: Own elaboration using CGPJ data (2015). We will estimate the impacts of the court fees in the litigation rate per capita and by type of procedure. The litigation rate is obtained as the quotient of new cases brought to the courts divided by the population. The subscript "j" denotes the province (we add all the new cases of all the courts of the same province), and "t" the quarter. Litigation rate j, t Newcases Population j, t j, t 9

10 Then, we will estimate the impact of court fees on a classic measure of "efficacy" of the judicial system: the "congestion" rate, which can be computed for each type of procedure [see, among others, GARCÍA-POSADA and MORA-SANGUINETTI (2015) or PONTICELLI and ALENCAR (2015)]. Congestion rate j, t Pending cases j, t 1 New cases Cases resolved j, t j, t The congestion rate is defined as the ratio of the sum of pending cases (measured at the beginning of the quarter), plus the new cases measured in a specific quarter divided by the resolved cases in the same quarter. A lower congestion rate is related to a greater efficacy of the courts of the province. 4 Identification strategy and empirical results 4.1 Capturing the effects of the adoption of court fees in the civil jurisdiction We have built three dummy variables which capture the entry into force of the "old" court fees, the "new" court fees and, finally, the "reform I" (as it was named in the previous sections). Given the current availability of data, we do not have enough observations to analyse the effect of "reform II". In order to avoid the risk of capturing the impacts of various systems at a time in each model, we restrict our period of analysis in each model to a window which contains only the time period before and after the date of entry into force of the previous or the next reform. Thus, to capture the effect of the introduction of the first ("old") court fees system we included a dummy variable that takes value 1 during the period in which that system was enforced in Spain (from 1 st January 2003). Our period of analysis in this case is a window which contains only the years 2001 to As before, the subscript "j" denotes the province, and "t" the quarter. Reform 52/2002 t 0 if t 2003T1 1if t 2003T1 For the analysis of the effects of the introduction of the new court fees in 2012, we analyzed the observations about the rate of litigation and court congestion in the period (window) between 2003 and the first quarter of The dummy variable takes the following values: Reform10/2012 t 0 if t 2013T1 1if t 2013T1 Finally, for the analysis of the "reform I we analyze the observations for both the litigation rate and the congestion rates from the first quarter of 2013 until the end of The dummy variable in this case takes the following values: RDLaw 3/2013 t 0 if t 2013T 2 1if t 2013T Control variables As it was discussed in section 3, the volume of litigation and the rate of congestion of the courts may be the result of multiple factors not directly related to the introduction of court fees. On the one hand, we include controls for the macroeconomic environment of each province in each quarter: the income per capita, whereas the higher the income, the higher the litigation, the sectoral composition of the economy of the province i.e. the weight of manufacturing, 10

11 construction, agriculture and services, as the composition may affect litigation (Palumbo et al. 2013, find evidence that points to lower litigation in the industrial sector), and the default rate (measured as the NPL ratio). The NPL ratio is the proportion of non-performing loans claims on the total available credit. To construct this variable we obtained data from the Central Credit Register of the Bank of Spain (CIR), which contains information on all 17 Spanish loans to non-financial companies granted by credit institutions operating in Spain above a reporting threshold of 6,000. As loans to businesses are often higher than the threshold for registration, CIR provides in practice the entire population of loans to companies in Spain. We refer as "non-performing loan" a loan that is in default or close to being in default. The regulation of the Bank of Spain states that a loan is in this situation if the payment of principal or interest has been delayed for more than 90 days or there are any circumstances that makes full repayment of the loan is unlikely. In addition to those macroeconomic controls, we include in the models the number of companies, in per capita terms, to account for the "social complexity" that can influence the rate of local litigation (CARMIGNANI and GIACOMELLI, 2010, MORA-SANGUINETTI and GAROUPA, 2015), and finally we control the model by the presence of lawyers per capita in the province considered. There is evidence to suggest that lawyers and litigation rate may be correlated (although the relationship is not necessarily endogenous) 18 To illustrate this issue, Figure 5 shows the very different size (on average) of the markets for lawyers at the provincial level in Spain We are including any instrument through which banks can provide credit to businesses: financial loans, commercial loans, letters of credit, leasing, factoring, repos, securities lending and loans or credits transferred to third parties. 18 In addition to the above references, see HANSSEN (1999) or BUONANNO and GALIZZI (2012). 19 The graph shows the average number of lawyers in the period between the entry into force of the new Civil Procedural Law and the entry into force of the Law on Access to the Legal Profession, i.e

12 Figure 4: Average number of lawyers (* 1000) per capita by province. Fuente : Own elaboration using data from the census of the Consejo Consejo General de la Abogacía Española. 4.3 Identification model for the civil jurisdiction We set up a regression model which relates our measures of litigation or judicial efficacy (the rate of court congestion) with one of the dummy variables specified in Section 4.1 (the two systems of court fees and the first reform), the set of controls specified in section 4.2 and provincial "fixed effects", which capture the characteristics of the Spanish provinces that do not vary over time or do so very slowly. We also add a group of quarterly variables (Q2, Q3 and Q4) as a seasonal adjustment, as Figure 3 clearly showed a seasonal pattern in the entry of new conflicts in the civil jurisdiction. Thus, with these variables we aim to have seasonally adjusted series of litigation. Q2, Q3 and Q4, take the following values: 12

13 1if quarter t is a second quarter within a specific year Q2t 0 otherwise 1if quarter t is a third quarter within a specific year Q3t 0 otherwise 1if quarter t is a fourth quarter within a specific year Q4t 0 otherwise Our estimates are obtained using the following econometric model: Litigation jt j K k Reform 52/2002 jt kcontrol jt 1Q2 t 2Q3t 3 k 1 Q4 t jt Litigation represents the rate of litigation for each civil procedure (ordinary, verbal, payment, jt or exchange procedures or the procedure involved in a bankruptcy incident). Represents the provincial fixed effects, Control is a matrix that includes the controls explained in section 4.2 and jt is the error term of the model. k jt We cluster the errors so that standard errors are robust to heteroskedasticity and serial correlation. Finally, it should be recalled that the inclusion of fixed effect by province and our set of dummy variables per quarter help to partially mitigate the omitted variable bias. j 5 Estimation results 5.1 Impacts of the old court fees (Law 52/2002) Tables 1 and 2 show the results of the models which estimate the effects of the "old" system of court fees (Law 52/2002). The adoption of the "old" court fees implied a significant decrease in per capita litigation if we observe the verbal and the exchange procedures. However, the fees did not reduce the litigation taking the form of ordinary procedures. Furthermore, the old system was significantly associated with an increase in the litigation taking the form of payment procedures (monitorios). Conceivably, litigation taking the form of a payment procedures may react differently when confronted to an increase in the costs to access the judicial system for several reasons: firstly, the monitorios were new procedures (born with the new CPL of 2000) with some attractive features: they were simpler and faster to resolve. As new procedures, they absorbed an important part of the new conflicts arriving to the courts (MORA- SANGUINETTI, 2010). Also, as they were "simpler" and potentially faster than the rest of old procedures, they were cheaper to use for citizens and enterprises. Therefore they may absorb litigation when the prices (part of the price is the court fee) to use the judicial system increase. In summary, we may be observing a safe haven effect by which the new payment procedures concentrate more demand as a result of the increase in the price to litigate. The adoption of the "old" court fees seems also to be related to a decrease in the growth rate of all procedures. That is, they had a significant moderating effect on the increase in litigation in Spain. The effect is statistically significant in the case of all procedures except for the payment one. Finally, the old court fees do not seem to have a significant effect in reducing judicial congestion rates except for the case of the ordinary procedures. 13

14 If we focus solely on the analysis of the most congested provinces in all procedures (i.e., we restrict the sample to the 5 most congested provinces out of 50) or we focus the analysis on the provinces which are most congested in only one of the procedures (38 provinces out of 50), we observe that the effects observed above get heightened. That is, almost all the impact of the adoption of the old court fees was observed in the most congested provinces. We can provide two explanations for this effect: on the one hand, the congestion is a cost for the litigants (as we can expect a conflict lasting longer in a congested court). In the most congested provinces, the direct cost of congestion was joined by the cost implied by the adoption of the fees. On the other hand, we should think about the "quality" or merit (composition) of the conflicts which reach the courts in each province: it is conceivable that in the most congested provinces there is a larger presence of non-meritorious cases which may be more affected by the fees. Unfortunately, there is no specific statistical information on the composition of cases which arrive to the courts, so we cannot control for that composition in the econometric analysis. 14

15 Table 1 PERIOD: Tasas de congestión de los asuntos iniciados (en términos per cápita) Tasas de crecimiento interanual de los asuntos iniciados (en términos per cápita) Procedimientos iniciados (en términos per cápita) Verbales Monitorios Ordinarios Cambiarios Concursales Verbales Monitorios Ordinarios Cambiarios Ejecuciones Concursales Verbales Monitorios Ordinarios Cambiarios Ejecuciones Concursales Ejecuciones VARIABLES Reform 52/ ** *** 0.303*** *** *** ** *** *** *** (0.0529) (0.0129) (0.0803) (0.0103) (0.0103) (0.0378) (0.0145) (0.0408) (0.0213) (0.0561) (0.370) (0.0532) (0.0762) (0.0880) (0.277) Non-performing loan ratio 4.385*** 0.129* 1.270*** 3.709*** * *** ** *** *** *** 16.49*** *** * 5.689* (1.022) (0.0769) (0.235) (1.292) (0.145) (0.168) (0.253) (4.247) (0.0949) (0.357) (0.164) (0.594) (5.583) (6.295) (0.808) (1.108) (1.435) (3.093) Lawyers per capita *** *** *** *** *** ** ** 0.277*** *** ** ( ) ( ) ( ) (0.0147) ( ) ( ) ( ) (0.0286) ( ) ( ) ( ) ( ) (0.0743) (0.0451) ( ) (0.0182) (0.0182) (0.0380) Number corporations per capita 20.77* * 16.75*** 16.48*** 17.09** 301.5*** 1.989* 28.02*** 13.30*** 62.60*** ** 28.37* 59.08** (11.64) (1.702) (3.113) (24.74) (4.542) (5.494) (7.031) (109.8) (1.175) (8.605) (3.158) (19.77) (62.85) (83.86) (10.84) (16.76) (23.79) (31.45) GDP per cápita * *** *** ** *** * (0.0206) ( ) ( ) (0.0399) ( ) ( ) ( ) (0.274) ( ) (0.0145) ( ) (0.0274) (0.107) (0.236) (0.0182) (0.0295) (0.0353) (0.0712) Manufacturing ** * *** (0.0144) ( ) ( ) (0.0205) ( ) ( ) ( ) (0.171) ( ) ( ) ( ) (0.0139) (0.105) (0.162) (0.0197) (0.0199) (0.0284) (0.0452) Construction *** *** *** *** * (0.0143) ( ) ( ) (0.0280) ( ) ( ) ( ) (0.166) ( ) ( ) ( ) (0.0131) (0.156) (0.226) (0.0218) (0.0234) (0.0323) (0.0430) 15 Services *** *** *** *** ** *** ** ** * * ( ) ( ) ( ) (0.0139) ( ) ( ) ( ) (0.109) ( ) ( ) ( ) ( ) (0.0854) (0.122) (0.0137) (0.0151) (0.0149) (0.0406) Second quarter ** *** *** *** *** *** ** (0.0172) ( ) ( ) (0.0271) ( ) ( ) (0.232) (1.833) (0.0217) (0.0378) (0.0330) (0.0870) Third quarter *** *** *** *** *** *** 2.648*** *** 0.815*** 1.730*** 1.979*** (0.0315) (0.0102) (0.0131) (0.0419) ( ) ( ) (0.359) (2.324) (0.0463) (0.0628) (0.0712) (0.191) Fourth quarter * 0.982*** ** 0.135*** *** (0.0237) ( ) ( ) (0.0314) ( ) ( ) (0.222) (1.976) (0.0314) (0.0434) (0.0411) (0.139) Constant ** *** *** *** *** *** ** ** *** ** 15.66* *** ** (0.763) (0.132) (0.247) (1.182) (0.207) (0.294) (0.336) (11.87) (0.149) (0.475) (0.250) (0.944) (7.133) (8.946) (1.334) (1.428) (1.670) (3.018) Observations 2,000 1,277 2,000 2,000 2,000 2,000 1,850 1,028 1,850 1,850 1,850 1,850 2,000 1,195 2,000 2,000 2,000 2,000 R Number of provinces

16 Table 2 PROVINCIAS MÁS CONGESTIONADAS EN ALGUNOS DE LOS ASUNTOS PROVINCIAS MÁS CONGESTIONADAS EN TODOS LOS ASUNTOS Tasas de congestión de los asuntos iniciados (en términos per cápita) Tasas de congestión de los asuntos iniciados (en términos per cápita) Verbales Monitorios Ordinarios Cambiarios Concursales Verbales Monitorios Ordinarios Cambiarios Ejecuciones Concursales Ejecuciones VARIABLES Reform 52/ *** (1.444) (0.205) (0.407) (0.429) (0.739) (0.458) (0.0651) (0.0993) (0.107) (0.369) Non-performing loan ratio 28.89* ** * ** *** * (11.27) (22.82) (1.368) (3.071) (2.617) (2.498) (6.698) (7.629) (0.969) (1.264) (1.659) (3.695) Lawyers per cápita * * 3.895* *** 0.271*** *** ** (4.365) (5.004) (0.648) (1.123) (1.183) (1.634) (0.0679) (0.0517) ( ) (0.0163) (0.0166) (0.0342) Number corporations per cáp ** 192.7*** 335.9* ** ** (221.3) (543.5) (45.14) (42.01) (33.69) (123.8) (61.02) (104.0) (11.99) (21.80) (27.61) (37.59) GDP per cápita *** ** * 0.707*** * (0.460) (0.635) (0.0751) (0.0800) (0.0833) (0.219) (0.145) (0.351) (0.0254) (0.0446) (0.0503) (0.110) Manufacturing 0.882* * ** (0.339) (0.490) (0.0498) (0.0583) (0.0887) (0.203) (0.139) (0.199) (0.0247) (0.0255) (0.0352) (0.0571) Construction * (0.304) (0.728) (0.0452) (0.0403) (0.0578) (0.245) (0.182) (0.284) (0.0251) (0.0257) (0.0359) (0.0514) Services ** ** * ** * (0.290) (0.656) (0.0397) (0.0773) (0.0724) (0.217) (0.106) (0.159) (0.0154) (0.0175) (0.0155) (0.0497) Second quarter ** * ** *** *** *** * (0.802) (0.797) (0.0672) (0.119) (0.0424) (0.394) (0.298) (2.488) (0.0271) (0.0468) (0.0398) (0.108) Third quarter 3.183** *** 0.948** 2.202*** 2.279*** 3.055*** *** 0.886*** 1.845*** 2.264*** (0.693) (0.662) (0.0971) (0.220) (0.196) (0.391) (0.435) (3.161) (0.0480) (0.0703) (0.0797) (0.231) Fourth quarter * ** *** *** 0.164*** * 0.592*** (1.069) (0.367) (0.0972) (0.153) (0.0840) (0.445) (0.271) (2.654) (0.0364) (0.0520) (0.0481) (0.170) Constant * ** *** ** (25.35) (51.85) (3.653) (6.895) (3.417) (20.57) (8.649) (10.28) (1.526) (1.696) (1.746) (3.676) Observations , ,520 1,520 1,520 1,520 R-squared Number of province_id (1) Provincias con más congestión en todos los procedimientos: Ciudad Real, Las Palmas, Madrid, Murcia y S.C. Tenerife (2) Provincias con más congestión en algunos de los procedimientos: Alicante, Almería, Baleares, Barcelona, Burgos, Cádiz, Castellón, Ciudad Real, Córdoba, Cuenca, Cáceres, Gerona, Granada, Guadalajara, Huelva, Huesca, Jaén, Las Palmas, León, Lérida, Madrid, Málaga, Murcia, Navarra, Orense, Palencia, Pontevedra, S.C.Tenerife, Segovia, Sevilla, Soria, Tarragona, Teruel, Toledo, Valencia, Zaragoza y Ávila. 16

17 5.2 Impacts of the new court fees (Law 10/2012) The results of the models for the "new" court fees are shown in Tables 3 and 4. The adoption of the new fees seems to be related to a significant reduction in the litigation taking the form of exchange and payment procedures and the litigation in conflicts of insolvency (incidentes concursales). We do not observe a significant impact of the fees in the litigation taking the form of verbal procedures. Finally, the new fees were positively related to the litigation rates in the form of ordinary procedures. As it was discussed, the new fees were the first to force individuals to pay. Thus, we should expect to observe higher effects precisely in the most used procedures by individuals: the procedures involving lower amounts, which are all with the exception of the ordinary procedures (which are reserved for conflicts with an amount higher than 6000 euros). This seems to be the case of the impact of the new fees on the payment procedures. The new fees seem to be related to an increase in the rate of growth of various types of procedures: the ordinary procedures, the payment procedures and, more weakly, the exchange procedures. Finally, these fees significantly decreased judicial congestion in the judicial system when solving verbal, exchange and payment procedures. If we restrict the sample to the most congested provinces (as it was done in the analysis of the old fees), we observe very similar impacts compared with the previous analysis: the effects would be even more pronounced when we concentrate on the 37 provinces with the highest congestion in any of procedures. We point to the same explanations as before to explain these differential impacts between more and less congested provinces. 17

18 Table 3 PERIOD: Procedimientos iniciados (en términos per cápita) Tasas de crecimiento interanual de los asuntos iniciados (en términos per cápita Tasas de congestión de los asuntos iniciados (en términos per cápita) Verbales Monitorios Ordinarios Cambiarios Concursales Verbales Monitorios Ordinarios Cambiarios Ejecuciones Concursales Verbales Monitorios Ordinarios Cambiarios Ejecuciones Concursales Ejecuciones VARIABLES Reform 10/ *** *** *** ** *** *** *** 0.267*** * *** *** *** ** (0.0776) ( ) (0.0206) (0.109) (0.0271) (0.0141) (0.0186) (0.203) (0.0181) (0.0253) (0.0377) (0.0546) (0.576) (0.457) (0.0821) (0.0876) (0.142) (0.254) Non-performing loan ratio * *** *** *** *** *** *** *** *** ** *** ** (0.419) (0.0449) (0.115) (0.693) (0.118) (0.110) (0.0930) (2.705) (0.0804) (0.124) (0.125) (0.288) (2.569) (7.724) (0.570) (0.682) (0.778) (1.413) Lawyers per cápita * 5.10e * *** *** ** (0.0132) ( ) ( ) (0.0249) ( ) ( ) ( ) (0.0515) ( ) ( ) ( ) (0.0123) (0.0918) (0.169) (0.0120) (0.0176) (0.0244) (0.0284) Number corporations per cápita 55.07*** *** 27.13*** 24.22*** 25.00*** 220.5*** 10.18*** 36.98*** 21.44*** 77.10*** 410.3*** *** 43.27* 100.9*** 76.08** (13.34) (1.085) (2.701) (25.07) (4.630) (4.110) (4.115) (70.77) (1.364) (6.529) (3.191) (12.60) (107.9) (72.99) (10.08) (24.43) (21.38) (35.46) GDP per cápita 7.162*** 0.446*** 1.281*** 11.31*** 1.113*** 0.955*** 0.941*** *** 0.727** *** *** ** (0.692) (0.137) (0.203) (1.240) (0.186) (0.181) (0.234) (4.666) (0.125) (0.294) (0.186) (0.448) (7.656) (23.15) (0.833) (1.206) (1.450) (2.258) Manufacturing ** ** *** *** * * * (0.0167) ( ) ( ) (0.0229) ( ) ( ) ( ) (0.143) ( ) ( ) ( ) (0.0121) (0.105) (0.165) (0.0160) (0.0200) (0.0231) (0.0478) Construction *** ** *** * ** (0.0156) ( ) ( ) (0.0286) ( ) ( ) ( ) (0.104) ( ) ( ) ( ) (0.0129) (0.168) (0.238) (0.0222) (0.0294) (0.0293) (0.0430) Services *** *** * * * * ( ) ( ) ( ) (0.0165) ( ) ( ) ( ) (0.0634) ( ) ( ) ( ) ( ) (0.0907) (0.106) (0.0117) (0.0167) (0.0158) (0.0448) Second quarter ** *** ** *** *** *** *** (0.0222) ( ) ( ) (0.0311) ( ) ( ) (0.222) (1.245) (0.0211) (0.0380) (0.0401) (0.0988) Third quarter *** *** *** *** *** *** 2.492*** *** 0.804*** 1.912*** 1.801*** (0.0292) ( ) (0.0121) (0.0355) ( ) ( ) (0.344) (1.528) (0.0480) (0.0556) (0.0750) (0.140) Fourth quarter *** 0.206*** *** * 1.271*** *** 0.260*** 0.258*** 0.502*** (0.0223) ( ) ( ) (0.0337) ( ) ( ) (0.254) (1.367) (0.0349) (0.0447) (0.0481) (0.125) Constant ** *** *** *** *** *** *** *** *** *** ** 15.24** * (0.905) (0.112) (0.277) (1.496) (0.262) (0.316) (0.280) (7.895) (0.150) (0.473) (0.258) (0.772) (9.202) (7.554) (1.176) (1.676) (1.445) (3.907) 18 Observations 2,008 1,685 2,008 2,008 2,008 2,008 2,008 1,426 2,008 2,008 2,008 2,008 2,008 1,601 2,008 2,008 2,008 2,008 R Number of provinces

19 Table 4 PROVINCIAS MÁS CONGESTIONADAS EN TODOS LOS ASUNTOS PROVINCIAS MÁS CONGESTIONADAS EN ALGUNOS DE LOS ASUNTOS Tasas de congestión de los asuntos iniciados (en términos per cápita) Tasas de congestión de los asuntos iniciados (en términos per cápita) Verbales Monitorios Ordinarios Cambiarios Concursales Verbales Monitorios Ordinarios Cambiarios Ejecuciones Concursales Ejecuciones VARIABLES Reform 10/ * * *** *** *** ** (1.112) (2.269) (0.219) (0.302) (0.219) (0.320) (0.686) (0.560) (0.0905) (0.0947) (0.171) (0.304) Non-performing loan ratio * ** ** (9.823) (10.49) (0.660) (1.449) (1.117) (1.913) (2.824) (8.973) (0.653) (0.763) (0.884) (1.620) Lawyers per cápita * ** 0.396** ** (0.901) (1.334) (0.132) (0.124) (0.178) (0.104) (0.0852) (0.180) (0.0128) (0.0181) (0.0250) (0.0303) Number corporations per cáp 697.1** *** *** *** *** *** 76.85* (242.5) (241.7) (19.00) (61.35) (37.79) (142.8) (138.3) (106.6) (12.77) (30.41) (26.97) (44.40) GDP per cápita * *** ** ** (21.67) (10.26) (2.393) (2.868) (3.719) (4.302) (9.002) (27.32) (0.981) (1.357) (1.702) (2.702) Manufacturing *** (0.267) (0.601) (0.0824) (0.0369) (0.117) (0.124) (0.127) (0.217) (0.0199) (0.0246) (0.0291) (0.0600) Construction ** * *** ** (0.408) (0.323) (0.0648) (0.0248) (0.0817) (0.0635) (0.199) (0.291) (0.0252) (0.0328) (0.0327) (0.0495) Services * ** * (0.110) (0.316) (0.0531) (0.0305) (0.0663) (0.102) (0.104) (0.117) (0.0137) (0.0190) (0.0178) (0.0562) Second quarter ** * *** *** ** (0.829) (0.842) (0.0599) (0.124) (0.119) (0.306) (0.293) (1.731) (0.0265) (0.0511) (0.0511) (0.131) Third quarter 3.513*** *** 0.899*** 2.246*** 2.343*** 2.781*** *** 0.817*** 2.010*** 1.930*** (0.552) (1.059) (0.129) (0.136) (0.182) (0.474) (0.407) (2.178) (0.0560) (0.0695) (0.0886) (0.181) Fourth quarter 1.775* ** 0.364** 0.252** *** *** 0.276*** 0.277*** 0.490*** (0.807) (1.044) (0.0679) (0.123) (0.0797) (0.312) (0.309) (1.889) (0.0429) (0.0559) (0.0573) (0.162) Constant * * 15.68* (31.53) (27.68) (4.534) (4.883) (6.476) (10.46) (11.41) (8.929) (1.428) (2.051) (1.693) (5.065) Observations ,499 1,166 1,499 1,499 1,499 1,499 R Number of provinces (1) Provincias con más congestión en todos los procedimientos: Ciudad Real, Tarragona, Las Palmas, Castellón, Murcia y S.C. Tenerife (2) Provincias con más congestión en algunos de los procedimientos: Albacete, Alicante, Almería, Baleares, Barcelona, Burgos, Cádiz, Castellón, Ciudad Real, Córdoba, Cuenca, Cáceres, Gerona, Granada, Guadalajara, Huelva, Huesca, Jaén, Las Palmas, León, Lérida, Madrid, Málaga, Murcia, Navarra, Orense, Palencia, S.C.Tenerife, Salamanca, Segovia, Sevilla, Soria, Tarragona, Teruel, Toledo, Valencia y Ávila 19

20 6 Conclusions This article analyzes the impact of court fees on both the congestion and the litigation rates of the civil jurisdiction in Spain. In order to do that, we fit several econometric models which exploit variations observed across time and provinces in Spanish real judicial data and relevant other economic variables which may affect the litigation rates at the local level. Spain is an interesting case to study as it has one of the highest rates of litigation of the OECD (Palumbo et al. 2013), a high level of lawyers per capita (Mora-Sanguinetti and Garoupa, 2015) and has traditionally lacked a system of court fees compared with the rest of the European countries. The paper examines both the system of court fees passed in 2002 (which forced just the legal persons) and the fee system of 2012, which extended the fees to individuals. This second system was in force for a short period of time as it was amended (reducing the fee) in 2013 and abrogated in The adoption and abrogation of two different systems of court fees within a limited time period provide a great opportunity to analyze the real impact of court fees in a judicial system. This research leads to the general conclusion that the effects of the fees have not been homogeneous and have not always coincided with the design intended by the legislator. According to the results of the models, the effects of the fees were uneven among the different types of civil judicial procedures and they may have had some indirect effect, transforming some procedures in "safe haven" of litigation against increased costs. Therefore they did not have a clear (or homogenous) impact reducing both litigation rates or mitigating the congestion of the judicial system. Each type of judicial procedure (which has different regulation and amounts), seems to react differently to increased barriers to litigation. More specifically, the adoption of the system of 2002 reduced litigation under the form of verbal or exchange judgments but did not reduce the congestion rates in general except in the case of the ordinary procedures. We may observe, in the case of that system, a safe haven effect, by which we observe an increased litigation in the form of the payment procedures. The extension of the fees in 2012 seem to be related to a reduction in the litigation under the form of exchange judgments and payment procedures but had reverse effects on the ordinary procedures. Consequently, the congestion of the judicial system was only reduced in a limited set of procedures. All these effects, whether those observed for the 2002 or the 2012 fee system, get strengthened in the provinces with more congested courts. That is, litigation rates diminish the most in the provinces where the courts show higher amounts of unresolved cases. 20