Spain has the one of the world's lowest business bankruptcy rates (number of ....
The extremely low Spanish business bankruptcy rates have been noted before.
THE SPANISH BUSINESS BANKRUPTCY PUZZLE
Marco Celentani Universidad Carlos III and FEDEA Miguel García-Posada Universidad Carlos III and FEDEA Fernando Gómez Universitat Pompeu Fabra February 2012
Abstract Spain has the one of the world’s lowest business bankruptcy rates (number of business bankruptcies divided by number of firms). We propose a hypothesis to explain this fact, which hinges on the idea that the unattractiveness of bankruptcy procedures leads Spanish firms to reduce the risk of bankruptcy. We argue that the Spanish corporate bankruptcy law grants low creditor protection relative to an alternative insolvency institution, the mortgage system, without achieving sizeable ex-post efficiency gains, and it is also very unattractive for company managers in terms of expected liabilities. The Spanish personal insolvency law, widely used for small companies in other countries, is too severe towards the individual debtor. This institutional framework may have negative repercussions on innovation, business returns and productivity. We show that this hypothesis is compatible with stylized features of firms’ capital and asset structures and profitability. We conclude with a brief discussion of potential policy implications.
1
1. Introduction Spain has one of the world’s lowest business bankruptcy rates, i.e. number of business bankruptcies divided by the total number of firms in the economy1. This suggests that Spanish firms rarely enter a formal bankruptcy procedure and instead try to deal with financial distress in alternative ways2. According to a report of Euler Hermes (2007), with data on 30 countries -both developed and emerging economies- for the year 2006, Spain had the second lowest bankruptcy rate, just after Poland, as shown in Table 13. The differences in the orders of magnitude with other Western European countries are also remarkable. For instance, while in Spain there were less than 3 bankruptcy filings per 10,000 firms in 2006, in Italy, Germany, U.K. and France there were 26, 96, 115 and 179, respectively. [TABLE 1 HERE] When a firm defaults on its debt, filing for bankruptcy is just one of the available options. Alternatively, a firm can engage in a private debtrestructuring process with its major creditors (a private workout) or it can sell major assets with high liquidation value to satisfy the outstanding debts before closing the business. Other procedures are based on the law of secured transactions, which allows creditors to seize the assets that serve as collateral for the loan, e.g. a foreclosure proceeding. Another option is a “friendly foreclosure”, in which the secured lender repossesses the property with the consent of the borrower in exchange for cancelling the outstanding debt; after that, the lender can sell the property to a third party to recover its credit. Foreclosures –either “friendly” or “unfriendly”- may play a major role as an alternative mechanism to bankruptcy in the case of Spain, given the wide use of mortgage debt by Spanish firms. Figure 1 shows the weight of mortgage debt on total bank debt of non-financial firms of Spain, France, Germany and U.K. for the period 1999-2011, highlighting the greater importance of this type of debt in Spain. [FIGURE 1 HERE]
1
By "Bankruptcy" we mean a legal procedure that imposes court supervision over the financial affairs of a firm or individual that has broken its promises to creditors or honors them with difficulty. Henceforth we will use the terms "bankruptcy" and "insolvency" as synonyms. 2 By “financial distress” we mean a situation in which a firm is close to default and it needs to take corrective action, such a selling major assets, merging with another firm or filing for bankruptcy. For a discussion see Ross, Westerfield and Jaffe (2005, pp. 830-846). 3 The current crisis in Spain has made the number of bankruptcy filings soar in absolute terms but not in relative terms; i.e., Spain still has one of the lowest business bankruptcy rates. See Appendix A for more details.
2
Since low business bankruptcy rates in Spain are accompanied by low business exit rates4, one could argue that the former is just a consequence of the latter. This possibility is considered in Table (2) where we list conditional business bankruptcy rates, i.e., the number of business bankruptcies divided by the total firms that exit the economy. In 2006 Spain has the lowest rate, three times less than the country with the second lowest (Czech Republic)5. [TABLE 2 HERE] There are two competing hypotheses on the extremely low bankruptcy rates in Spain. The one supported in this paper is that Spanish firms reduce the risk of bankruptcy because of the unattractiveness of the bankruptcy procedure, i.e., the distribution of Spanish firms in terms of their bankruptcy risk differs from that of other countries with higher bankruptcy rates. The alternative hypothesis is that the distribution of Spanish firms in terms of their bankruptcy risk is not different from that of other countries, but what differs is the legal threshold that separates insolvent from non-insolvent firms. In other words, the Spanish bankruptcy legislation is “softer”, in the sense that makes filing for bankruptcy a legal requirement only when firms are in a situation of extreme financial distress and it leaves more room for private workouts. A firm would enter a bankruptcy procedure only in rare events, when private workouts fail, or when its financial condition is so desperate that a private workout is not even attempted. This view would imply that the low bankruptcy rates are nothing but a statistical reflection of a legislation with an unusually low “implied insolvency test” that assigns more firms in financial distress to informal workouts (that are difficult to document) and fewer to formal bankruptcies (that are recorded in official statistics). Apart from some theoretical objections, we reject this second hypothesis on empirical grounds. Specifically, we find that Spanish firms involved in bankruptcy procedures are not in worse financial conditions than their foreign counterparts. Therefore, we will focus on the first hypothesis and we will analyze the distribution of all Spanish firms –both solvent and insolvent- in terms of their bankruptcy risk, to assess whether they deliberately reduce their probability of filing for bankruptcy. The extremely low Spanish business bankruptcy rates have been noted before. But given that there is no accepted explanation for these low rates, we refer to them as the Spanish business bankruptcy puzzle (SBBP). This paper is a first step to attempt to explain the SBBP. The main objective of the paper is to propose an explanation that is compatible with a number of related aggregate stylized facts 4
Both Núñez (2004) and López-García and Puente (2006) find that in Spain exit rates are lower than in other OECD countries. 5 To enhance comparability across countries we do not take into account exits from industries in which the public sector may play a major role, such as education, health, social and personal service activities, etc.
3
that we document. Because the data we use are aggregate, we cannot test our view. We can simply use the data as a guide to propose a coherent explanation and as an indication of how useful it may be to pursue this line of research. Figure 2 shows a positive relationship between (the log of) bankruptcy rates and economic development, as measured by per capita GDP at PPP, for the countries included in the study of Euler Hermes (2007), where the red dot represents Spain. Its correlation coefficient is 0.48. This finding is consistent with Claessens and Klapper (2005)6. This paper does not attempt to provide a complete explanation for the positive relation exhibited in Figure 2. Figure 2 constitutes instead a graphical representation of the observation that motivates this work, and of the reason why we believe it is important to explain this observation. The observation is that Spain has extremely low business bankruptcy rates. The reason why we think it is important to explain this is that low bankruptcy rates may indicate that the legal environment discourages business risk-taking, with obvious repercussions on growth and on the riskier components of its drivers, such as innovation.
[FIGURE 2] There is ample literature, both within and outside economics, arguing that the institutional framework dealing with the creditor-debtor relationships and the insolvency of firms is very relevant for economic outcomes. The ways in which legal systems distribute claims against assets of insolvent firms to debtors and creditors and assign the rights to control these assets to creditors, debtors or third parties (such as judges or insolvency practitioners) obviously influence expost outcomes, i.e., the allocation of resources after the insolvency procedure has been initiated. But legal provisions for insolvent firms also have important ex-ante effects, because they affect productive activities and their financing and also because they ultimately influence the probability that firms become insolvent. In this paper we propose the idea that, in order to explain the Spanish business bankruptcy puzzle, it is necessary to keep into account both the ex-ante and the ex-post repercussions of the Spanish bankruptcy law. The view that we propose in this paper can be summarized as follows: (1) Creditor/debtor orientation of insolvency law, (2) efficiency of bankruptcy procedures and (3) mortgage efficiency have an impact on (4) the choice of capital and asset structures; (4)
In a regression of bankruptcy rates on per capita GDP, they find that the latter has a positive and very significant coefficient. See page 268 for further details. 6
4
has an impact on (5) the probability of bankruptcy filings and on (6) firms’ performance, measured in terms of profitability. The paper is organized as follows. In section 2 we discuss the related literature. In section 3 we describe the Spanish bankruptcy legal procedure and we compare it with the ones of France, Germany, Italy and the UK. In section 4 we reject the alternative hypothesis on the SBBP, namely that is a consequence of an unusually low “insolvency test” implied by the Spanish bankruptcy legislation. In section 5 we expose our view of the Spanish experience. In section 6 we contrast our view with aggregate data on capital and asset structure of nonfinancial firms in Spain, France, Germany, and Italy. In section 7 we conclude and provide a brief discussion of our findings and of the policy implications of our view. The Appendices contain details on data sources and on the evolution of business bankruptcy rates during the current economic crisis. 2. Related literature Our work relates to existing literature in economics, finance, law and economics and law and finance. Many authors have already stressed the importance of how and how efficiently the bankruptcy code splits claims and control rights on the assets of financially distressed firms or, in other words, of (1) the creditor/debtor orientation and (2) the efficiency of bankruptcy procedures. For instance, La Porta, Lopez de Silanes, Shleifer and Vishny (1997 and 1998) have made international comparisons of insolvency procedures constructing indices of their propensities to allocate these claims and control rights to creditors.7 La Porta, Lopez de Silanes, Shleifer and Vishny (1997) also use indices of the “efficiency of the judicial system,” the “rule of law” and “corruption” elaborated by the Business International Corporation and the International Country Risk Guide. Djankov, Hart, McLiesh, Shleifer (2008) have designed a survey to provide a quantitative measure of the losses in debt enforcement around the world and of the causes of these losses, e.g., what part is due to legal costs, duration or inefficient decisions regarding the liquidation or continuation of an insolvent firm. There is a vast literature on the ex-ante implications of bankruptcy codes and, in particular, on (4) firms’ capital and asset structures. La Porta, Lopez de Silanes, Shleifer and Vishny (1997) find that shareholder protection encourages the development of equity markets, and to a lesser extent that creditor protection encourages the development of credit markets. Qian and Strahan (2007) study how creditor protection rules affect price and non-price terms (such as debt maturity) in bank loans in a sample of 60 different countries. In their international comparisons of capital structures, Rajan and Zingales (1995) 7
See also the independent classification of Lopez, Garcia and Torre (2009).
5
suggest that bankruptcy laws may be one of the important reasons why firms in different countries have different capital structures and conjecture that the creditor friendliness of the bankruptcy code may be a determinant. Davydenko and Franks (2008) study the impact of bankruptcy codes on firms’ asset structures. In particular, they compare the asset structure of firms that defaulted on their bank debt in France (which has a debtor-friendly code), Germany and UK (whose bankruptcy laws are much more creditor-friendly). They conclude that lenders respond to poor creditor protection under bankruptcy by requiring more collateral and by relying on types of collateral that minimize the dilution of their claims. Giannetti (2003) finds that institutions that favor creditor rights are associated with higher leverage and greater availability of long-term debt. Acharya, Sundaram, and John (2011) propose and test a model in which the optimal leverage depends on the creditor friendliness of the bankruptcy code but also on the anticipated liquidation value of the firm. With respect to the literature on (5) the probability of bankruptcy filings, Claessens and Klapper (2005) study bankruptcy rates around the world. They find that a proxy for a country's overall institutional quality -”rule of law”- and some features of bankruptcy systems are associated with more bankruptcies. In particular, they find that bankruptcy rates increase with (creditor oriented) restrictive reorganization rules (e.g., creditors' consent for reorganization) and with (debtor oriented) provisions of automatic stay of creditors' claims during the insolvency process. Unfortunately these results do not fit into the Spanish case. Since Spain has high rule of law and its bankruptcy system has both restrictive reorganization rules and an automatic stay provision, Spain should exhibit, according to Claessens and Klapper, high bankruptcy rates. A different approach to study (5) is to find the determinants of bankruptcy filings conditional on financial distress8. A pioneer work in the area is the empirical study of Gilson et al. (1990), who analyze the factors that make financially distressed firms restructure their debt privately rather than through formal bankruptcy. They find that firms more likely to restructure their debt privately have more intangible assets, owe more of their debt to banks and owe to fewer lenders. Morrison (2008) studies why US small firms rarely file for bankruptcy when dealing with financial distress. He argues that there are cheaper and speedier insolvency procedures for these firms, such as foreclosures, out-of-court liquidations and private workouts. However, some of these procedures, unlike the bankruptcy system, face major coordination and asymmetric information problems that may hamper its use. Thus he identifies the conditions under which these problems are not so important so those procedures can be implemented. Several studies have analyzed the impact of insolvency institutions on (6) firms’ 8
In mathematical terms, we can decompose the probability of filing for bankruptcy into 2 terms:
6
performance. Acharya and Subramanian (2009) study the link between bankruptcy codes and innovation. They find that, when a bankruptcy code is “creditor-friendly”, inefficient liquidations cause leveraged firms to shun risks, and hence innovation, whereas by promoting continuation upon failure, a “debtor-friendly” code induces greater innovation. While this type of ex-post inefficiency occurs in firms from any industry, the greater risk inherent in innovative industries accentuates its deadweight loss. Thus, a “creditorfriendly” code discourages innovation relative to a relatively more “debtorfriendly” one. Another strand of literature studies the impact of personal bankruptcy laws on the performance of small firms and, in general, on entrepreneurship. Personal bankruptcy laws may be used by non-corporate businesses and by small corporate firms (Berkowitz and White, 2004). When a business is non-corporate, its debts are personal liabilities of the firm’s owner. When a firm is a small corporation, lenders often require personal guarantees or security in the form of a second mortgage on the owner’s home, which wipes out the owner’s limited liability. The impact of personal insolvency laws on entrepreneurship depends on two opposite effects. On the one hand, very severe personal bankruptcy laws minimize the scope for moral hazard and they may in turn reduce the risk premium charged to entrepreneurs (the “credit supply effect”). On the other hand, more lenient ones (for instance, allowing some debt discharge or high exemption levels) provide partial insurance against business failure (the “insurance effect”). Since entrepreneurship is very risky, this partial insurance provides risk-averse agents with incentives to undertake entrepreneurial activities. The available evidence seems to suggest that the second effect dominates the first one. Fan & White (2003), making use of the state variation in bankruptcy exemption levels in the US, find a positive relationship between exemptions and the number of start-up companies. They argue that higher exemption levels benefit potential entrepreneurs who are risk averse by providing partial wealth insurance and therefore that the probability of owning a business increases as the exemption level does. Armour and Cumming (2008), using panel data for fifteen countries in Europe and North America, find a negative relationship between severity of personal bankruptcy laws –measured in terms of years after bankruptcy required for a discharge to be available, exemptions and restrictions on the debtor’s civil and economic rights, amongst others- and entrepreneurship, as measured by selfemployment rates. Armour (2004), by means of cross-country analyses, finds a negative correlation between the severity of personal insolvency laws –measured by the years until a discharge is allowed- and venture capital. He argues that such a correlation is the consequence of the lower demand for venture capital by entrepreneurs in countries where personal bankruptcy is very severe. Since venture capital is used to finance high-risk start-up companies with new business ideas but limited access to other sources of funding, too severe personal insolvency laws may deter innovation.
7
3. Spanish Insolvency Law in context The current Spanish insolvency law dates from 2003, though its entry into force was delayed until September 1 20049. Prior to that, the Spanish regime was notoriously chaotic and inefficient (Cerdá and Sancho, 2000). It was sharply divided in two different procedures. One of them (quiebra) could be initiated by both the insolvent debtor and the creditors, and implied that the firm’s management was taken over by creditor-appointed representatives who were essentially in charge of liquidating assets and paying creditors in due order, although an agreed restructuring was also possible. The rules were very rough on the debtor and archaic –mostly in the 1885 Commercial Code, but also in an earlier Commercial Code from 1829. The procedure was complex and lengthy 25 years was not unheard of- and the ex-post efficiency in terms of asset realization was meagre to say the best. The other was essentially –because it could also end up in liquidation eventually- a restructuring procedure (suspensión de pagos) which could be initiated solely by the debtor, who typically would retain control of the firm. The procedure, though substantially simplified from that of the quiebra, was still cumbersome, and the occasions for debtor and creditor opportunism were plenty. Various attempts to radically reshape Spanish insolvency Law failed, until the 2003 Bankruptcy Act (Ley concursal, LC in what follows). The old arrangements probably were an important explanatory factor behind the observation that Spain had, at the turn of the millennium, an astonishingly low number of bankruptcy proceedings relative to the number of firms. In Claessens and Klapper (2005) out of 35 countries in Europe, America and Asia, Spain had the lowest formal bankruptcy rate, of 0.02%, only close to Peru (roughly double the Spanish rate) and Portugal (about four times the Spanish rate). The new insolvency regime applied by Spanish courts since 2004 may be summarily described by the following features: 1. It is a unified procedure, eliminating the previous two avenues to channel firm insolvency. It is also unified in the sense that it serves both firms and individuals, though there is a simplified procedure that the Court may follow using several criteria (less than 50 creditors; liabilities not above € 5,000,000; assets not above € 5,000,000; anticipated restructuring agreement; foreseen sale of the firm as a going concern, etc)10.
9
The current Act has been extensively modified twice, one in March 2009 and the other at the end of 2011, both trying to cope with various dysfunctional features in the initial design. For instance, formal workout negotiations on the brink of bankruptcy filing have been facilitated. 10 The latest reform has introduced a much more open text. The initial LC set much more rigid criteria, namely that the liabilities did not exceed € 1,000,000 and that the firm had simplified accounts and no audited books. That figure was raised to € 10,000,000 in the reform of March 2009 (Real Decreto Ley 3/2009), as a result of the greatly increased workload for the Mercantile courts due to the crisis.
8
2.
Both the debtor and the creditors may initiate the proceedings. In practice, since its entry into force in September 2004, around 87% have been so-called voluntary filings, on the debtor’s initiative12. The debtor and the firm managers are encouraged to file early through different means: first, they may file even when no actual insolvency exists, as long as the inability to face payments is imminent; second, the debtor and firm management are under a legal duty to file in two months from actual insolvency, and this will be presumed after 3 months of default in tax and social security contributions, or salaries. If they do not file in the prescribed time, there will be a presumption that insolvency is not without fault, which may imply serious personal liabilities for management; third, the filing is simplified, because only the accounts, a list of creditors and assets, and a brief explanation of the situation of the firm is required for filing being complete; fourth, if they do not file, but the creditors do (the so-called necessary filings), the rule is that firm management will be taken over by court-appointed representatives. In 2009 the LC has been modified to coordinate early filing with incentives for debt renegotiation: if the debtor is negotiating a proposal for restructuring agreement (convenio anticipado), it may avoid the obligation to file for insolvency by notifying the court that it is in such negotiation process. Then, a window of 3 months opens for such a process, with an additional 1 month to file if renegotiation fails and insolvency has not been avoided, in which no procedures may be initiated by creditors.
3. Once the petition for bankruptcy has been formalized, the courts – Mercantile courts, specialized in commercial matters- will start the procedure (sección primera) and decide whether to accept or reject the petition. Courts should do so summarily – the LC, very optimistically, foresees that this will happen in one day in case of petition by the debtor. Roughly 19% of petitions are rejected, to a higher degree when the petition is not voluntary on the debtor’s side13. In case of creditor’s initiative, opposed by the debtor, typically because insolvency is not convincingly shown. In the case of debtor’s initiative, it is essentially because some of the formal requirements are missing or defective, possibly on purpose, when the debtor may try to hastily fulfill the duty to file, or get a head start in petition over creditors. 4. The second section of the procedure deals with the court-appointed administrators. The general rule is that there will be one designated by the court from different pools14. He or she may be a practicing lawyer selected by the court from the list provided by the Bar Association, or an auditor, economist or commercial expert (titulado mercantil) also from the 12
Source: Consejo General del Poder Judicial. Specifically,17.6 % for voluntary filings, 30.4% for necessary filings. Source: Consejo General del Poder Judicial. 14 The number of administrators has gone down to one from the initial number of three (one in the simplified procedure) only in the late 2011, with entry into force in January 1st 2012. 13
9
lists provided by the relevant professional bodies15. When the proceeding is especially important due to the number of creditors, annual turnover, employees, etc, the court may appoint a creditor from the upper third part of credits in terms of amount. Administrators will be compensated over the debtor’s assets, on a variable basis depending essentially on the value of the assets and the volume of credit. As many criticisms were raised against excessive compensation, the LC was changed in 2009 to fix a cap on the level of compensation. No performance incentives (neither in restructuring nor in liquidation) are built into this compensation scheme. The insolvency administrators take over management when the court so decides –more commonly in creditors’ initiated procedures- and in the remaining cases they oversee current management, and have to authorize all transactions outside day to day business of the firm. They also draw the list of assets and creditors, have to give an opinion on all restructuring plans that may be presented, and are in charge of drafting the liquidation plan of the firm’s assets, unless the debtor himself has presented an early liquidation plan that has obtained court approval. 5. The court declaration of insolvency that starts the formal procedure determines an automatic stay in all unsecured credits until the end of the procedure, and interest cease to accrue, with very limited exceptions. Secured creditors over assets that are integrated in the debtor’s production process are also affected by the stay, for the minimum of 1 year or the date in which a restructuring plan that does not affect their rights is approved. 6. Insolvency administrators produce a list of all the debtor’s assets (sección tercera) and credits against the insolvent debtor (sección cuarta). Credits are subject to the following ordering: (i) preferential credit (créditos contra la masa) will be the first to be paid, and comprise salaries for the last month of activity, the costs of the procedure itself, including compensation for the insolvency administrators, plus the new debt incurred by the firm in its activities after the insolvency declaration. This means that new funds for the on-going operation of the firm, even new financial debt, will enjoy priority over old secured debt; (ii) secured credit of all kinds over specific assets of the firm (créditos con privilegio especial); (iii) privileged credit (créditos con privilegio general) such as other labor credits, and public and tort creditors up to a certain amount; (iv) ordinary credits (créditos ordinarios) as the residual category: all credits that do not belong to any other type; (v) subordinated credit (créditos subordinados) which includes those of closely related parties (managers,
15
The administrators may belong to a specialized firm having both lawyers and economists, auditors or commercial experts
10
shareholders, etc.) and some other kinds, such as interest, sanctions and fines, etc. 7. The LC provides for some claw-back actions and procedures aimed at redressing the assets of the debtor –now, presumably, with the creditors as residual claimants- of the harmful consequences of actions that took place prior to the insolvency declaration. Thus, advanced payments and transactions with related parties –managers, shareholders- may be clawed-back at the initiative of the administrators or the creditors, and the proceeds will increase the debtor’s assets. The possibility of clawback also affects the grant of secured status to credits replacing existing credit. In some –but very few indeed- cases Spanish courts ruled to eliminate the securities granted to banks having re-financed the existing debt. This created huge alarm in the financial sector, and led to a change in the LC in order to exempt re-finance transactions from claw-back, if some conditions are met: the re-finance plan is agreed by 3/5 of existing credit, and the plan is ok-ed by an independent expert appointed by the Commercial Registry (Registro Mercantil)16. 8. After the list of assets and creditors is approved by the court, the common phase of the procedure ends, and we may have reorganization or liquidation. A restructuring or reorganization plan may be proposed both by the debtor and by the creditors. Data shows that in virtually all cases –nearly 97%- it is the debtor who has the initiative of the plan17. The debtor may also make use of the opportunity to present an anticipated plan, together with the petition for bankruptcy (with the support of at least 10% of outstanding credit) or within one month of the court declaration of insolvency (with the support of at least 20% of outstanding credit). Anticipated plans have been popular in terms of presentation, but they have been approved less frequently than ordinary plans18. The plan has to be informed by the insolvency administrator or administrators, approved by a majority of ordinary creditors, and finally authorized by the court. Secured and privileged creditors will not be affected by the plan and will fully enjoy their legal rights against the insolvent firm unless they vote in favor of the plan. The plan may not – except in extraordinary cases that need to be justified by the court- imply a loss for ordinary creditors beyond 50% of their nominal value, nor a delay in payment of more than 5 years. For anticipated plans, if there is a
16
In the latest reform the stay agreed in refinancing plans approved by 75% of bank creditors may be imposed upon the remaining bank creditors who do not have secured credit. 17 Source: Consejo General del Poder Judicial. 18 For instance, until 2008 anticipated plans accounted for a 23% of all the proposed reorganization plans, but they only accounted for a 14% of the approved plans, while the remaining 86% were ordinary plans.
11
viability plan with external financing, it is easier to overcome these limits. 9. If the debtor so requests, liquidation starts. Also, if no plan is presented or reaches approval19, or if the approved plan fails, the insolvency administrators submit a liquidation plan to the court, in order to sell the assets and pay the creditors in the order summarized above.20 10. The LC mandates that the court should examine the potential liabilities of the debtor –the firm’s management in the case of a company- when there is liquidation of the firm or when a reorganization plan is agreed, except when the hair-cut is of less than 33% or the delay in payments is less than 3 years. The LC contains some presumptions of fault on the part of the debtor or its managers, and taking them into account, and considering the evidence presented, the court may declare the bankruptcy to be fortuitous (concurso fortuito) or guilty (concurso culpable). A finding of guilt may imply a judgment against the individual manager involving incapacitation to run a company from 2 to 15 years, payment of damages to the firm or to creditors, and even, in case of liquidation, the obligation to face the unpaid sums in favor of the creditors. These liabilities are independent of the criminal liabilities that may apply if a criminal behavior –fraud, embezzlement- is found and proven in front of a Criminal Court. As will be shown in Section 5 below, an important explanatory dimension lies in the abilities of courts, and court-appointed administrators or trustees to separate firms that should survive as a going concern from those that are not viable and should be liquidated and the assets, tangible and intangible (if at all possible) distributed to the creditors. In the Spanish case, there are several factors that lead us to estimate that the degree of specific training and experience both of mercantile courts and of the administrators these courts appoint are relatively poor. First, the bankruptcy history of Spain had traditionally made bankruptcy law a marginal and not too visible area of legal practice and expertise, reducing both the incentives and the opportunities to gain practical experience "on the ground". Moreover, traditionally the number of people of experience in the area, given the extremely low number of bankruptcies, was very limited. Moreover, until very recently Spain did not have specialized courts on these matters, and even today the degree of specialization is far from complete, as mercantile courts have to deal with other matters (corporate, transport, intellectual property) that lie far apart from bankruptcy. Mercantile judges, Only in 11% of the total bankruptcy filings a reorganization plan was presented, and only in 5% of these filings a plan was approved. 20 The debtor may also present a liquidation plan to the court, with the petition for bankruptcy or at any time 15 days later than the list of assets and creditors has been produced. 19
12
despite their high legal proficiency, have not, and still do not receive any training about firm valuation or financial matters more generally. As to administrators, the appointment system, by relying on the lists21 provided to courts by the official bodies or associations of the relevant professions does not seem to encourage professionals to develop above-average expertise nor, coupled with the compensation system –lacking any performance-related component- to provide incentives for maximizing recovery value. Only very recently, after the latest reform, two improvements have been introduced: first, that courts are able, even only when required by the circumstances, to select particular administrators, out of the general appointment system; second, that firms -comprising both lawyers and economists or accountants- may be nominated as trustees, and not only individual professionals. In sum, apparently the most important decision-makers inside the bankruptcy proceedings -court and administrators- are generally not too well equipped to make the decisions based on the viability or non-viability of the firm in bankruptcy. How similar or different is the current Spanish insolvency law to those of other developed economies? Most commentators, both in economics and in law, tend to ascribe an insolvency regime to one or the other of two ideal types of regime. There are creditor-friendly regimes, essentially driven by creditors and focused upon maximizing the net recovery of their credit. Debtor-friendly regimes, on the other hand, are mostly concerned about keeping the firm as a running enterprise, and allow space for the debtor to reorganize and be back in business, while at the same time keeping an eye on safeguarding the interests of other stakeholders, particularly employees. The UK22 is typically considered to possess a clear creditor-oriented system. This orientation may have probably weakened following legislative changes earlier in this decade, but the full effects of these changes may not yet be entirely perceptible. Although various insolvency procedures co-exist in the UK, the most important regimes for business insolvency are summarized here. Prior to the Enterprise Act (2002), secured creditors were almost entirely in charge of bankruptcy under an administrative receivership scheme. The holder of a floating charge on the business23, commonly one bank providing the bulk of finance to the company, could appoint, with almost no other constraints, as soon as there was a default in the loan, a receiver who would take over the entire company, and would try to maximize recovery for the holder of the floating charge. Of course, this normally did not imply piecemeal liquidation of the assets, but the sale of the business to a new entrepreneur. Floating charges 21
The lists were open to all professionals with at least five years of general experience as lawyers, accountants, auditors, etc. 22 The main source for insolvency Laws in Europe is McBryde, Flessner, and Kortmann (2003). 23 Davydenko and Franks (2008) present evidence that financing of each individual firm is more concentrated in UK firms than in other European countries.
13
agreed after the enactment of the Enterprise Act do not give rise to such power. The bank may, however, under some conditions, appoint an administrator who takes over the management, although he owes duties also to other creditors and to the company itself. The procedure may be initiated by the secured creditor even without court order and proof of insolvency. The company management is entirely replaced, and the administrator is supervised by the court, and by a committee of creditors. The administration implies a stay for non-secured creditors, but the individual enforcement of secured credit needs to be authorized by the administrator or the court. As for liability of management following an insolvency procedure, it is mainly criminal, and based on a finding of fraudulent trading that may be attributed to the actions or decisions of the managers. France is commonly placed at the opposite end of the spectrum, essentially due to the ample powers of the insolvency courts, at the service of the goal of preserving the company and employment. Among the different procedures, the redressement judiciaire seems to be by far the most important insolvency regime for companies. To initiate the procedure, the inability of the firm to meet current liabilities with liquid assets needs to be ascertained. The request may come from the debtor himself –who is under the duty to file in a period of 15 days after having ceased payments, the breach of which may imply severe sanctions-, the creditors, the public prosecutor and the court itself. After the procedure is formally opened, typically the debtor remains in possession and control of assets, and the management remains in place, although they will be subject to authorization by a special judge (juge commissaire). In rare cases, the management will be replaced by an administrator, who also needs to report to the special judge. The procedure determines a stay both for unsecured and secured creditors, with the exception of employees, for certain amounts. Reorganization plans play a major role in the proceeding, although it may also end up in liquidation of firm assets, and orderly payment to creditors. Reorganization plans are divided in continuation and transfer plans, depending on whether the current debtor will still be in control of the business. All players may have the initiative to present a plan, though the court is solely responsible, after hearing all affected parties, to decide about its approval, and the other players hold no veto power. In fact, the court may use certain sticks –a long moratorium upon creditors, imposition of liabilities- to guide the goal and content of the plan in the desired direction. Finally, managers of the insolvent firm may be found liable if they have, with fault, caused the insolvency situation. A finding of liability leads to monetary sanctions –up to full payment of unpaid debt- and incapacitation for managing firms. Germany has also experienced a process to some extent similar to the UK. It was considered a very creditor-friendly country, but in 1999 the new regime (Insolvenzordnung) softened that character. The procedure may be initiated by
14
the debtor himself or by the creditors, when inability to pay current debt can be shown. The debtor has to file for insolvency within 3 weeks of finding its own inability to pay. Also potential insolvency, as in the Spanish LC, may allow filing. The court does not only verify the factual insolvency, but also runs a level of assets test. If the assets are unlikely to be able to cover the costs of the procedure, the court will reject the filing. As soon as the procedure is formally started, the court appoints an administrator who would replace current management. Creditors may change the administrator in the first creditors’ meeting, though this happens only in exceptional cases, mostly of very large insolvent debtors in which substantial amounts are at stake for the financial creditors. The administrator is overseen both by the court and by the creditors, who could replace him. Less commonly, the court may authorize the debtor to remain in control in which case the court will appoint a supervisor to oversee management. Non-secured credits are stayed after formal insolvency is declared. Secured creditors over movable assets are stayed up to the first creditors’ meeting. Secured creditors over real estate are not automatically stayed, but the court may stay individual enforcement at the request of the administrator, if it is deemed to prevent adequate reorganization. If the creditors do not decide in favor of preserving the company, and no plan has been presented, the administrator must proceed to liquidate assets and pay creditors in an orderly fashion. Reorganization plans may be presented by the debtor or by the administrator, and they have to be approved by a majority – number and value- of affected creditors. Those who are not affected by the plan are not entitled to vote. The plan has then to be accepted by the debtor, and confirmed by the court, which can refuse confirmation only under specific grounds. Liability imposed on management of the insolvent firm seems not to play a large role, although the violation of the duty to file may trigger liabilities vis-à-vis creditors under general principles of tort Law. Italy has traditionally known a wide array of insolvency procedures24 (fallimento, concordato preventivo, amministrazione controllata, liquidazione coatta amministrativa, amministrazione straordinaria delle grandi imprese insolventi), but one of them –fallimento- stands out, by far, as the most widely used scheme. The triggering event is again the inability to regularly meet current liabilities. The debtor, creditors, the public prosecutor and the court itself may initiate the procedures. Data show that many filings are brought by creditors. In such cases, the debtor has to be heard and may oppose the insolvency declaration. This explains the relatively high rate of rejected filings in the Italian system. Once the procedure is formally opened, the debtor is dispossessed of the company, and a court appointed administrator (curatore fallimentare) takes over, under the supervision of the court. The procedure implies an automatic stay of 24
For Italy see Stanghellini (2008).
15
non-secured claims, but secured claims may also be affected to some extent. Liquidation is the most common result. In 2006 a major reform was introduced in the regulation of “fallimento” Creditors’ rights are strengthened, especially by the creation of a creditors’ committee, which may request the substitution of the administrator, supervises his activities, may request liquidation at any time and has to approve any reorganization scheme proposed by the debtor, the administrator or a third party. Additionally, in 2005 an important reform was also introduced concerning reorganization schemes before bankruptcy (“concordato preventivo”). Here, the initiative and room for maneuver for the debtor have been expanded, the plan proposed by the firm in trouble has to be approved by a majority of creditors, although exceptionally it may be imposed if the court considers that is beneficial to them. This reform has increased manyfold the rate of use of this alternative vis-à-vis the full bankruptcy procedure (Rodano, Serrano-Velarde and Tarantino, 2011). So far our institutional analysis has been exclusively focused on corporate bankruptcy law. However, as explained in section 2, for small businesses the personal bankruptcy law is also relevant. The analysis of personal insolvency laws is especially important in order to understand the SBBP, since Spanish small firms account for a large proportion of the total stock of firms and their bankruptcy rates are the lowest. Figure 3 shows the evolution of bankruptcy rates by firm size since the onset of the LC, classifying businesses into 3 categories: micro (less than 10 employees), small (between 10 and 49) and medium & large (more than 49). The bankruptcy rates –bankruptcy filings per 10,000 firms- of micro firms averaged 1.5 before the start of the crisis (2004-2007) and even during the crisis they have never exceeded 12, much lower than any value for small, medium and large firms26. Furthermore, micro firms account for around a 95% of the total firms in Spain27, which implies that the low business bankruptcy rates mainly reflect the extremely low bankruptcy rates of those firms. [FIGURE 3] As also explained in section 2, the literature suggests that personal insolvency laws are an important determinant of entrepreneurship. Figure 4 shows the evolution of bankruptcy rates, distinguishing between self-employed and companies28. While the bankruptcy rates of companies have almost reached 40 during the crisis, the bankruptcy rates of self-employed have never reached 2, i.e., around 20 times lower. Since these data are quarterly, the bankruptcy rates have been multiplied by 4 to make them comparable with the annual bankruptcy rates displayed in the other tables and figures. 27 Source: Instituto Nacional de Estadística, Directorio Central de Empresas. 28 Since these data are quarterly, the bankruptcy rates have been multiplied by 4 to make them comparable with the annual bankruptcy rates displayed in the other tables and figures. 26
16
[FIGURE 4] Therefore, the available evidence seems to suggest that filing for personal bankruptcy in Spain is a very unattractive option. The analysis of the law corroborates this idea, since in Spain there is no debt discharge. In other words, all the present and future income of the debtor -except for some exemptions defined in the law- must be used to pay back her pre-bankruptcy debts. Since the fixed costs of filing for bankruptcy are high (compensation of the insolvency administrators, lawyers’ fees, etc) while the benefits are almost none in the absence of a discharge (the slim chance of reaching a debt-restructuring agreement with creditors), personal insolvency is very unappealing. Thus, entrepreneurs, unincorporated firms and small firms that pledged personal guarantees to obtain credit have strong incentives to avoid filing for bankruptcy. By contrast, in France the discharge is immediate (the so-called “fresh start”), while in UK it is permitted after one year. This might help explain why, for instance, in 2006 the bankruptcy rate for French micro firms was 208 per 10,000 firms, while in Spain was 1.5, and the bankruptcy rate for French self-employed was 139, while in Spain was 0.129. In Germany the discharge is not immediate, and only takes place after 6 years, a period in which the debtor must use part of her income to repay the pre-bankruptcy debts and prove his “honest behavior” and “good faith”. Cross-country comparisons of other features of personal insolvency laws that also determine the degree of severity of these laws yield quite similar conclusions. Armour & Cumming (2008) analyze other three: “exemptions”, “disabilities” and “composition”. “Exemptions” relate to the level of personal wealth that cannot be seized by the court to pay unsecured creditors. “Disabilities” measure restrictions on the civil and economic rights of the bankrupt debtor, such as the incapacitation to run a company or to vote. “Composition” measures the majority threshold required for creditors to make an agreement on the debtor’s discharge which can bind a dissenting minority. Among our five countries of interest, the Spanish personal insolvency law is: a) quite severe in terms of “exemption”; b) the most severe in terms of “disabilities”; c) quite lenient in terms of “composition”30. However, even if de jure creditors’ agreement on a debt discharge is not so difficult to achieve, de facto they do not have incentives to accept such a discharge. Finally, it is worth mentioning that limited liability can also be wiped out by the Law under corporate bankruptcy if the firm’s managers are found guilty, i.e., if they have caused the insolvency situation with fault. In several countries, such 29 30
The figures for France were computed using Altares (2010) and Eurostat’s business demography. For more details see Armour & Cumming (2008, pp. 310-315).
17
as Spain and France, a finding of guilt leads to monetary sanctions up to full payment of unpaid debt31. However, the criteria under which fault is presumed are more vague in the Spanish code, leaving more room for the judge’s discretionality. For instance, the French legislation includes sanctions for debtors who do not file for bankruptcy in time. But the possibility of liabilities conditional on having filed in time is remote. This means that French company administrators know that a voluntary filing for bankruptcy, provided it is timely, reduces the probability of a sanction very substantially. By contrast, the Spanish legislation includes sanctions for late filing (in this case guilt is presumed) but also for gross negligence (culpa grave) in the running of the business. This means that even a timely filing does not drive to 0 the probability of a finding of guilt32. As a consequence, filing for bankruptcy is a very unattractive option for Spanish managers.
4. Testing differences in “implied insolvency tests”. As explained in the introduction, an alternative explanation of the SBBP could be that the Spanish bankruptcy legislation is very “soft”, in the sense that makes formal bankruptcy a legal requirement only when firms are in a situation of extreme financial distress, leaving more room for private workouts. In other words, a firm would enter a formal bankruptcy procedure only in very rare cases, when private workouts fail or when its financial condition is so desperate that a private workout is not even attempted. This Coasian view would mean that the extremely low bankruptcy rates in Spain are just a statistical reflection of a legislation with a very low “implied insolvency test” that assigns more firms in financial distress to private workouts and fewer to bankruptcies, but that has no implications in terms of efficiency. There are theoretical objections to this explanation. First of all, the Spanish bankruptcy law, like the bankruptcy laws of France, Germany and UK, gives incentives for early filing through several mechanisms, so that the firm enters the bankruptcy procedure as soon as possible, to avoid further deterioration of its financials: this contradicts the idea of an abnormally low “implied insolvency test”. Furthermore, private workouts are often unfeasible due to high bargaining costs: they may fail due to coordination and asymmetric information problems33. Regardless of the strength of these theoretical objections, the hypothesis can also be tested empirically by studying whether Spanish firms involved in bankruptcy procedures are in worse financial conditions than their foreign 31
Another civil sanction that guilty managers may face is the incapacitation to run a company for some period of time. 32 Since the onset of the LC until 2010, out of a total of 17,333 bankruptcy procedures 835 were found guilty, i.e., almost 5 %. 33 See Gilson et al. (1990) for a discussion.
18
counterparts or not. This is done by constructing several indicators of financial distress, liquidity, profitability, leverage and solvency for a sample of bankrupt firms in Spain, Germany, France and U.K for the period 2007-2008, and comparing their differences in means and medians across countries. The sample comes from the database Amadeus, which provides firm-level data for a number of European countries. Details about the database and the sample selection criteria can be found in the Appendix. To measure financial distress we use the Altman's Z-Score (Altman, 2000). The higher the Z, the lower the probability of financial distress34. To capture profitability we use return on assets (ROA), computed as Ebitda over total assets. Liquidity is evaluated with the CURRENT RATIO, which is the ratio of current assets to current liabilities. To assess leverage we calculate 3 variables: total debt over total assets (LEVERAGE 1); financial debt over total assets (LEVERAGE 2)35; and the INTEREST COVERAGE RATIO, which is the ratio of ebitda to interest expense. To capture solvency we use the proportion of “balance-sheet insolvent” firms in each country. A firm is said to be “balancesheet insolvent” when it has negative net assets, i.e., the book value of its assets is lower than the book value of its liabilities36. The means and standard deviations (in parentheses) of those indicators for each country are shown in Table 3. Each panel of Table 3 is organized so that the country with the highest mean appears in the second column, the country with the second highest mean in the third column and so on. The presence of an asterisk (*), two asterisks (**) or three asterisks (***) next to a particular figure indicates that the difference between the mean of that particular variable for that particular country and the mean of the same variable for the country that has the closest but lower mean is statistically different from zero at a 10% (*), 5% (**) or 1% (***) confidence level. This is found through one-sided p-values of 34
The Altman's Z-Score is a weighted sum of four variables that represent liquidity, solvency, profitability and leverage. Both the weights and the variables are chosen by using discriminant analysis to find the linear combination that best predicts bankruptcy. The Z-Score has several versions depending on the type of firms to be analysed. The version used in this paper is the one for non-listed firms that do not necessarily belong to the manufacturing sector. The exact formula is:
Z = 6.56⋅ X1 + 3.26⋅ X2 + 6.72⋅ X3 +1.05⋅ X4
where X(1)=(Current Assets-Current Liabilities) / Total Assets; X(2)= Retained Earnings / Total Assets; X(3)= Earnings Before Interest and Taxes / Total Assets; X(4)= Book Value of Equity / Total Liabilities. 35 By total debt we mean both financial debt and accounts payable. The difference between the two measures is that trade credit is excluded from the computation of Leverage 2. This follows Rajan & Zingales (1995), which argue that trade credit may be used in some industries for transactions purposes rather than for financing, consequently overstating the amount of leverage. Given that the level of accounts payable and accounts receivable may jointly be influenced by industry considerations, and the industry distribution of bankrupt firms differs across countries, it seems appropriate to use an alternative measure of leverage unaffected by the gross level of trade credit for robustness. 36
Although there are firms that can be balance-sheet insolvent without being in financial distress because the book value of their assets is much lower than their market value (for instance, some companies have very valuable brands, but these resources only show up in the balance sheet as intangible assets when the company is purchased by another one, in the form of goodwill). We thank María Gutiérrez for pointing out this idea.
19
two-sample t-tests37. Table 4 is structured in the same way, but showing medians instead. The statistical differences are found through one-sided pvalues of Fischer exact tests. Table 3: Financial Distress Indicators: means and standard deviations (s.d.) Each panel is organized so that the country with the highest mean appears in the second column, the country with the second highest mean in the third column and so on. The presence of an asterisk (*), two asterisks (**) or three asterisks (***) next to a particular figure indicates that the difference between the mean of that particular variable for that particular country and the mean of the same variable for the country that has the closest but lower mean is statistically different from zero at a 10% (*), 5% (**) or 1% (***) confidence level. This is found through onesided p-values of two-sample t-tests.
Panel A: Altman’s Z-Score country mean (s.d.) N
Germany 1.29** (2.90) 372
Spain 0.83*** (2.42) 629
U.K. 0.09* (3.40) 571
France -0.41 (4.23) 194
Panel B: Return on assets (ROA) country mean (s.d.) N
Germany 3.37 (13.94) 384
U.K. 2.74 (13.06) 504
France 1.82 (12.94) 148
Spain 1.45 (7.70) 614
Panel C: CURRENT RATIO country mean (s.d.) N
Germany 1.27** (0.73) 428
France 1.16 (0.59) 193
Spain 1.12*** (0.54) 602
U.K. 0.96 (0.50) 600
Panel D: INTEREST COVERAGE RATIO country mean (s.d.) N
U.K. 1.66 (6.14) 419
Germany 1.35 (4.34) 348
France 0.75 (7.91) 136
Spain 0.51 (2.75) 590
Panel E: LEVERAGE 1 37
Two-sample t-tests for the equality of means can be implemented with and without the assumption of equal population variances. In order to ascertain whether this assumption is plausible, a couple of tests for the equality of variances have been implemented in each case. The selected tests have been those of Brown and Forsythe (1974), since they are robust to non-normality -unlike the traditional F-test- and the variables of this study have been found to be non-normal.
20
country mean (s.d.) N
Spain 67.50* (27.20) 651
U.K. 64.33***(34.22) 491
France 53.50 (27.58) 197
Germany 51.80 (34.33) 432
Panel F: LEVERAGE 2 country mean (s.d.) N
Spain 46.74 (25.26) 650
U.K. Germany 46.63***(41.13) 37.86***(32.62) 497 435
France 16.88 (17.77) 194
Panel G: % Balance-sheet insolvency firms country mean (s.d.) N
U.K. 29.21 (45.51) 647
France 28.93* (45.46) 197
Germany 23.49***(42.43) 562
Spain 14.16 (34.89) 657
Table 4: Financial Distress Indicators: medians Each panel is organized so that the country with the highest median appears in the second column, the country with the second highest median in the third column and so on. The presence of an asterisk (*), two asterisks (**) or three asterisks (***) next to a particular figure indicates that the difference between the median of that particular variable for that particular country and the median of the same variable for the country that has the closest but lower median is statistically different from zero at a 10% (*), 5% (**) or 1% (***) confidence level. This is found through one-sided p-values of Fischer exact tests.
Panel A: Altman’s Z-Score country median N
Germany 1.15*** 372
Spain 0.74 629
France 0.62 194
U.K. 0.53 571
Panel B: Return on assets (ROA) country median N
Germany 5.04* 384
U.K. 3.83 504
France 3.66 148
Spain 2.87 614
Panel C: CURRENT RATIO country median N
France 1.09 193
Germany 1.08*** 428
Spain 1.03*** 602
Panel D: INTEREST COVERAGE RATIO
21
U.K. 0.99 600
country median N
Germany 1.50 348
U.K. 1.38 419
France 1.35 136
Spain 1.04 590
Germany 55.42 432
France 50.97 197
Germany 33.25*** 435
France 10.91 194
Panel E: LEVERAGE 1 country median N
Spain 73.42*** 651
U.K. 63.42*** 491
Panel F: LEVERAGE 2 country median N
Spain 46.61*** 650
U.K. 36.65 497
From the analysis of tables 3 and 4 the following findings on the bankrupt firms of the sample can be summarized: a) In terms of the Altman's Z-score, German firms have the best financials, followed by the Spanish ones; b) German companies may be slightly more profitable than the rest (ROA); c) German and French firms have the highest levels of liquidity, followed by Spanish and British firms, in that order (CURRENT RATIO); d) Spanish and British firms are the most leveraged (LEVERAGE 1 and LEVERAGE 2)38; e) Spain has, by far, the lowest proportion of balance-sheet insolvent firms. These findings lead to the conclusion that Spanish firms are not in worse financial conditions than their European counterparts, with the exception of Germany, whose bankrupt firms seem to be in better health than the rest39. Thus we reject the hypothesis that the Spanish bankruptcy code implies a lower “insolvency test”40. 38
The fact that the most leveraged firms come from Spain and UK can partially be attributed to the housing bubble / burst that those countries have experimented during the period of the sample (2007-2008), since the proportion of bankrupt firms that belong to the construction sector is 34% in Spain and 27% in UK, while only a 6% in Germany and France. 39 This may be a consequence of a provision of the German bankruptcy code. Following a bankruptcy filing, the court runs a level of assets test. If the assets are unlikely to be able to cover the costs of the procedure, the court will reject the filing. This prevents firms in a extremely bad financial situation to enter into the bankruptcy procedure, which biases the distribution of bankrupt firms towards healthier ones. 40 A robustness analysis was also carried out. Since the distribution of bankrupt firms across countries differ in two aspects which can potentially influence the results, the proportion of very large firms and the proportion of firms with consolidated accounts, those firms were removed from the sample and the same analysis was implemented. The results of that analysis -
22
5. A thought model for the Spanish insolvency law One way of describing bankruptcy laws is by characterizing their ability to promote ex-ante and ex-post efficiency. One would normally think that ex-post efficiency simply requires the maximization of the expected value of the assets of the firm, conditional on the firm having entered a bankruptcy procedure. The reason is that control rights should be assigned in such a way as to maximize the value of the assets and financial claims on the assets could then be assigned in any way that promotes ex-ante efficiency.41 In a similar way, it seems relatively natural to identify ex-ante efficiency with the maximization of the expected value of a firm.42 If it were really possible to assign control rights and financial claims on the assets independently of one another, it would then follow that ex-post efficiency would be a necessary condition for ex-ante efficiency. But there are a number of reasons to think that this is not true in practice and that a sizable trade-off between ex-ante and ex-post efficiency exists. First of all, bankruptcy procedures have substantial legal and administrative costs. This means that a bankruptcy law with lower ex-post efficiency but that induces a lower probability of reaching an insolvency state may be preferable. A probably more important factor is that several individuals involved in the bankruptcy procedure are likely to be cash-constrained (first and foremost the debtor). This implies that not all transfers are possible and that the assignment of control rights and of financial claims are not independent of each other. To give an example, think about a situation in which the maximization of the value of assets requires assigning control rights to the debtor who would keep the firm as a going concern. Imagine, however, that ensuring the availability of credit that promotes ex-ante efficiency required assigning substantial financial claims to the creditor. If the debtor were not cash constrained, all that would be needed would be for the debtor to make a transfer to the creditor. But, given available upon request- lead to the same general conclusion: Spanish bankrupt firms are not in worse financial conditions than their European counterparts. 41 For this to be the appropriate notion of efficiency, one needs to assume that preferences are quasi-linear in wealth, that no party is cash constrained, and that the only assets that are involved in the bankruptcy procedure are the assets of the firm. This may not be the case, for instance, when the manager/entrepreneur’s reputation also depends on the final outcome of bankruptcy, when workers have firm-specific human capital, or when the losses to creditors may precipitate them in insolvency. 42 It may be important to determine whether one should think about the maximization of the value of a potential firm that has not been set up yet, or the maximization of the value of the existing firms. But because, we will not discuss this issue we prefer to ignore it.
23
that the debtor is likely to be cash constrained, this is not possible. In this situation ex-ante efficiency may therefore require assigning control rights to the creditors. But because creditors are inherently biased towards liquidation (as the nature of their claims implies that they do not fully reap the gains of upside potential, but they suffer from the downside risk), an ex-post efficiency loss would arise. A paper that examines in detail the trade-off between ex-ante and ex-post efficiency is Ayotte and Yun (2007) and because we think that it is especially suited to analyze the Spanish case we want to summarize some of its arguments and conclusions. Ayotte and Yun (2007) start from the observation that bankruptcy laws either allocate significant control rights to third parties, such as judges or insolvency practitioners (IP), or allow them to mediate in the allocation of these rights to debtors and creditors. The reason why such an arrangement may be superior for debtors and creditors is that third parties can act on “soft” information (e.g., recent evolution of cash flows) that is difficult to describe and that is therefore not contractible. In other words, the discretion of judges or IP’s can enhance the efficiency of ex-ante contracts between debtors and creditors43. One of the main tasks to be performed by judges and IP’s in a bankruptcy procedure is to determine whether the value of the firm as a going concern (either under current or under new management) is higher or lower than what can be realized through a piecemeal liquidation of assets. Depending on their ability to make sound business judgments, this task can be executed with lower or higher probabilities of type I and type II errors44 (respectively, maintaining as a going concern a firm that is worth more if liquidated piecemeal, liquidating piecemeal a firm that is worth more as a going concern). This implies that the ex-post efficiency gains of judicial discretion are increasing in the judges and IP’s ability to make business judgments. As an alternative the bankruptcy law could assign the control rights in a more mechanical way that foregoes the potential gains of judicial and IP’s discretion. But in this case, the bankruptcy law should focus on the ex-ante perspective and should assign control rights to creditors to promote the ex-ante availability of credit by maximizing the recovery of credit upon bankruptcy and reducing the scope for debtor’s moral hazard. Based on the previous premises, Ayotte and Yun’s (2007) point is that to a certain extent laws should be best responses to the abilities of its enforcers, more than the other way around. The reason is that it is simpler and more 43
It is important to notice that the assumption is not that judges or IP’s have superior information. What is important is that judges or IP’s provide a technology to include soft information in contractual agreements. 44 We follow the terminology used by White (1994).
24
economical to adjust a bankruptcy law to the existing distribution of professional skills than to wait for a decade or so for an appropriate professional elite –in terms of skills and experience- to emerge in the judiciary and the professional pools of insolvency administrators, possessing the appropriate competencies to provide the tailored enforcement envisioned by the law.45 Ayotte and Yun’s (2007) conclusion is that the optimal trade off between ex-post efficiency and ex-ante efficiency depends on the ability of the third parties that may be assigned control rights in bankruptcy. If judges and IP’s have high abilities in separating out viable from nonviable firms, the bankruptcy law should make use of these abilities to promote ex-post efficiency. But if judges and IP’s have low abilities in discerning viable from nonviable firms, the expost gains of discretion are lower and the optimal insolvency law should be more creditor oriented, in the sense that it should assign ample control rights and financial claims to creditors. What is more important is that Ayotte and Yun (2007) also ask the following question: What would happen if, in the face of low ability of bankruptcy judges and IP’s, the legal system failed to optimally respond with a creditor oriented bankruptcy code? This question is interesting for the Spanish case, because, as the discussion of the previous section should have clarified, the Spanish bankruptcy code is relatively debtor oriented and yet makes use of judges and IP’s who probably know the law well, but who have no special skills, nor clear incentives to make appropriate business decisions regarding the continuation of the firm. The answer to the question provided by Ayotte and Yun (2007) is that “[w]here the bankruptcy code does not provide enough creditor protection to make lending feasible, our model predicts that credit contracts will be written so that distress is resolved outside of bankruptcy, thus reducing bankruptcy usage rates.”46 47 In the next section we will show that a view similar to the one described above is compatible with the stylized features of Spanish businesses balance sheets and profit and loss accounts when they are compared to those of businesses subject to different bankruptcy laws in other European countries. The view that we propose can be summarized as follows. The relatively debtor oriented 2003 Spanish Bankruptcy act, and the low efficiency of bankruptcy procedures in Spain -in particular the low ability of judges and administrators to determine which firms should be liquidated piecemeal and which should be maintained as a going concern and reorganized- imply that, relative to their foreign Ayotte and Yun (2007) mention that the Bankruptcy and Composition Act of the Slovak Republic has been amended 14 times in 10 years. 46 Ayotte and Yun (2007), page 5. 47 However, this prediction implicitly assumes that there are alternative institutions where credit contracts can be efficiently enforced (e.g. the mortgage system). This may not happen in the case of France, whose bankruptcy code grants little creditor protection and whose judges are biased towards firms’ continuation (hence achieving low levels of ex-post efficiency) but it exhibits very high bankruptcy rates. 45
25
counterparts, Spanish firms and their creditors choose capital and asset structures that imply a lower risk of reaching a bankruptcy procedure. 6. Reducing bankruptcy risk? The aggregate evidence In the introduction we have already documented that the risk of bankruptcy is substantially lower for Spanish firms and the comparative analysis of section 2 is our way of claiming that premises (1) and (2) of the view summarized at the end of the previous section hold for Spain. To convince the reader that this view is not incompatible with available evidence, in this section we turn to international comparisons of stylized facts that are related to the activities in which Spanish firms and their lenders can engage to reduce the risk of bankruptcy. Those activities are: (a) Holding a high proportion of mortgage debt, since it facilitates the use of foreclosures as an alternative to bankruptcy; (b) Choosing low financial leverage, because it reduces the likelihood of financial distress; (c) High levels of lenders’ screening and monitoring, since they reduce moral hazard and adverse selection problems. Following these ideas, in each of the subsections of this section we will investigate whether aggregate data indicate that for Spanish firms, compared to their foreign counterparts: (a) The weight of tangible fixed assets (the assets that can be used as mortgage collateral) is higher; (b) Leverage is lower; (c) The weight of debt held by banks is higher, since banks are more likely to screen and monitor firms than other creditors that have smaller stakes, such as trade creditors. In this section we compare aggregate and median ratios of tangible assets, profitability, leverage and bank debt of Spanish firms with those of firms in France, Germany and Italy –three countries with higher bankruptcy rates than Spain- using data on balance sheets and profit and loss accounts of nonfinancial firms available from BACH (Bank for the Accounts of Companies Harmonized) and BACH-ESD (European Sectoral references Database). The Appendix contains a description of the firms contained in BACH for each country, as well as some distinctive features of BACH-ESD. In Table 5 we summarize the balance sheets of the firms included in BACH for the year 2006. [TABLE 5 HERE] BACH does not provide individual information of companies, but the aggregate value of each financial account for some industry-size bins in each country. Therefore, a ratio between two of these accounts yields the aggregate ratio, which is just a size-weighted mean. By contrast, BACH-ESD provides the median -among other characteristics of the
26
distribution- of 28 ratios. Since the values of aggregate ratios may be distorted by those of the largest firms of the bin, we will instead report the median ratios when available48. Given that the countries that we consider have different sectoral compositions and different size distributions, to make the comparisons as meaningful as possible we will look separately at the most important sectors in the 1-digit NACE Rev. 1.1 classifications (D-K) and we will also distinguish among small firms (with a turnover lower than 10 million €), medium firms (with a turnover between 10 and 50 million €), and large firms (with a turnover over 50 million €). 6.a Tangible fixed assets and profitability. If the bankruptcy code does not grant enough creditor protection, a possible response for a firm that attempts to obtain credit is to choose an asset structure that makes it easy to obtain mortgage debt since, in the event of financial distress, secured creditors could foreclose on the assets that were pledged as a collateral. This implies that the mortgage system would become a de facto insolvency institution under which some firms are liquidated piecemeal49. However, this alternative mechanism to deal with financial distress will only be undertaken if the mortgage institution is efficient enough, so that creditors enjoy high (discounted) recovery rates. According to a survey of the European Mortgage Federation (2007), for instance, the usual interval between mortgage foreclosure and the actual distribution of the proceeds of the sale is 7 to 9 months in Spain, 12 months in Germany, 15-25 months in France and 5-7 years in Italy. This means that the mortgage system is very efficient in Spain and Germany, but not in France and Italy. In Spain bankruptcy procedures take much longer than foreclosures. According to Van Hemmen (2008), the median length of a bankruptcy process in 2007 ranged between 20 and 23 months50. Furthermore, Spanish secured creditors will not suffer any of the dilution of claims related to the bankruptcy process if they avoid it and foreclose on the collateral instead. This makes the mortgage system a very attractive alternative insolvency institution.
48
In the latter case, aggregate ratios have also been computed as a robustness analysis. The results are available upon request. 49 The use of foreclosures –either “friendly” or “unfriendly” ones- as an alternative to bankruptcy requires that the majority of the firm’s debt is mortgage debt. In such a case the firm, when entering into financial distress, can pay back its debts with unsecured creditors and other stakeholders (trade creditors, workers, Social Security) with the remaining liquidity and default on its mortgage debt, triggering the foreclosure process. 50 Those figures mainly correspond to a pre-crisis situation. The economic crisis made the number of bankruptcy filings soar since 2008, implying a congestion of the courts and a dramatic increase in the median length of the bankruptcy process: between 27 and 35 months in 2008, between 31 and 36 in 2009 and between 28 and 36 in 2010 (Van Hemmen, 2009, 2010, 2011).
27
Table 6 shows the ratio of tangible fixed assets to total assets. Higher ratios are expected to be associated with greater use of mortgage debt, since tangible fixed assets (land and building, plant and machinery) are the assets that can be pledged as mortgage collateral. The Table corroborates our hypothesis by showing that Spanish and German firms generally have substantially higher ratios than their French or Italian counterparts. In particular, of the total of 21 size/industry classes, Spain has the highest ratio in 14 cases and the second highest in 6 cases. [TABLE 6 HERE] Nevertheless, the fact that Spanish firms hold a higher proportion of tangible fixed assets may imply that they distort their asset structures, with the ensuing productive inefficiencies. In other words, Spanish firms may over-invest in tangible fixed assets (e.g. purchasing machinery instead of renting it, since in the first way it can be included in the mortgage contract) to increase the available credit and reduce funding costs, at the expense of not choosing the amounts of inputs that minimize production costs. This should be reflected in low profitability. The available evidence, displayed in Tables 7-9, is consistent with this hypothesis. Table 7 shows that the return on assets or ROA –measured as EBITDA over total assets- is substantially lower for Spain, especially in the case of SMEs51. [TABLE 7 HERE]
Furthermore, ROA can be decomposed into 2 factors, Profit Margin and Asset Turnover52: ROA= EBITDA*100/(Total Assets)=Profit Margin*Asset Turnover where Profit Margin=[EBITDA*100/Turnover] and Asset Turnover=[Turnover/Total Assets] While the values for profit margins are nearly always higher for Spanish firms (Table 8), the values for asset turnover are always lower for them (Table 9). This means that the low ROA of Spanish firms is driven by a low asset turnover, which indicates low levels of operating efficiency. [TABLE 8 HERE]
51
Large firms are often believed to be less dependent on the local institutions for creditor protection because they have easier access to equity financing and to credit from foreign sources. Therefore, we do not expect the proposed effect to be so strong in their case. 52 Asset Turnover measures the capacity of the firm’s assets to generate revenue, known as operating efficiency, while Profit Margin is influenced by the firm’s market power.
28
[TABLE 9 HERE]
Moreover, the asset turnover can be decomposed in the following way: Asset Turnover=[Turnover/Total Assets]= Tangible Fixed Asset Turnover * [Tangible fixed assets/Total Assets] where Tangible Fixed Asset Turnover=Turnover/Tangible fixed assets We have already explained that the second term, the proportion of tangible fixed assets over total assets, is higher for Spain than for the other countries (Table 6). Therefore, the first term, the Tangible Fixed Asset Turnover, must be much lower, as it is in fact the case (Table 10). Therefore, we can conclude that the key driver of the lower ROA of Spanish firms is the low operating efficiency of the investments in productive capacity, as reflected by the Tangible Fixed Asset Turnover53. This suggests that Spanish firms may over-invest in tangible fixed assets to relax their credit constraints, at the expense of some productive inefficiencies. [TABLE 10 HERE] The high efficiency and high creditor protection of the mortgage system relative to bankruptcy may bias the choice of business projects by Spanish firms. Specifically, companies may choose projects that require a high proportion of tangible fixed assets over projects with potentially higher productivity but a lower proportion of these assets. This would deter innovative activities, which mainly rely on intangible assets, while it would favor other lower-productivity activities such as construction. For instance, according to Arce et al. (2008), in Spain the less productive sectors (such as construction) would have benefited more from the strong credit growth between 1995 and 2007, one of the reasons being that those sectors produce assets that can be used as collateral on loans. 6.b Leverage. To document the extent of leverage, we consider two different measures. The first one, capital and reserves to total assets, measures which % of assets are funded without making use of debt. Therefore, the higher this ratio, the lower the leverage54. This measure shows that, in median terms (see Table 11) Spanish firms exhibit substantially higher ratios (i.e. lower leverage) than those of the remaining countries. Averaging over sectors, small, medium and large firms have a It is worth noticing that the tangible fixed asset turnover, as well as the asset turnover, is computed using the assets’ net value, i.e., net of accumulated depreciation. However, we have also computed them using the assets’ gross value in order to rule out that differences in accounting rules across countries, in interaction with tax systems, are the factors behind the differences in these ratios. 54 Since assets=debt + capital & reserves, that ratio amounts to 1-(Debt/Assets). 53
29
difference of 4, 13 and 10 points, respectively. Spanish firms have the lowest leverage for 15 of the 21 sizes/industry bins and the second lowest in other 3. [TABLE 11 HERE] The second measure, interest expenses to EBITDA, assesses the capacity of the firm to honor its debt payments: the higher this ratio, the higher the leverage55. In terms of this measure (see Table 12) Spanish and French firms exhibit substantially lower leverages than German and Italian firms. Among small firms the differences are of the order of 5 points (with Germany) and 10 points (with Italy). Among medium firms the differences are of the order of 3-4 points. By contrast, among large firms the differences are of the order of only 1.5 with Italy but of 14 points with Germany. [TABLE 12 HERE] Therefore, according to Tables 11 and 12, Spanish firms can be singled out for their low leverage. Because high levels of leverage are normally associated with higher probabilities of bankruptcy, the evidence contributes to explain the low rates of Spanish business bankruptcies. However, in order to claim that the lower leverage of Spanish firms is a consequence of intended actions to reduce bankruptcy risk it is necessary to discuss the possible effects of other known determinants of leverage. First, notice that by controlling for size and industry in Tables 11 and 12 we factor out the effects of two well-documented determinants of financial leverage. Second, while the evidence is sometimes mixed, low leverages are normally expected when (i) capital and personal tax rates are low (because the tax savings of issuing debt are lower); (ii) the weight of tangible fixed assets is lower (because tangible fixed assets can be used as collateral); (iii) returns are higher (because retained earnings are an alternative source of financing).57 But because Spanish firms, as compared to the other three countries, do not face lower taxes58, they have a higher weight of tangible fixed assets and their projects yield lower returns, our finding of lower leverages is all the more remarkable and makes our conjecture of a relationship with the risk of bankruptcy all the more sensible. 55
EBITDA stands for Earnings Before Interests, Taxes, Depreciation and Amortisation, a very gross measure of profit whose main virtue is its comparability across countries. 57 See, for instance, Titman and Wessels (1988) and Rajan and Zingales (1995). 58 According to the OECD Tax Database, in 2006 Spain had the second highest Corporate Income Tax Rate of the four countries (35%), a higher Small Business Corporate Tax Rate than France (the only other country with such a tax) and the second highest top marginal rate in its Personal Income Tax (45%, also applicable for sole proprietorships), although it had the second lowest Effective Tax Rate on Dividends. However some of these rates were lowered in subsequent years. For more information see www.oecd.org/ctp/taxdatabase.
30
6.c Bank debt Banks, together with trade creditors, are the main providers of funds for firms with limited access to capital markets, i.e., non-listed firms. While trade credit is often dispersed among a large number of poorly informed small trade creditors, bank credit, in contrast, tends to be concentrated in a smaller number of better informed lenders (Gilson et al., 1990). Therefore, banks may engage in screening and monitoring activities (henceforth, S&M) in order to safeguard their interests. However, S&M are costly activities, so the extent to which banks will use them will also depend on how the institutional framework protects their rights. For instance, Djankov et al. (2007) argue that banks may respond to poor creditor protection under bankruptcy by screening and monitoring borrowers more carefully. Furthermore, according to Gilson et al. (1990), private workouts are more likely to succeed when relatively less debt is owed to trade creditors and more is owed to bank lenders, since coordination and information asymmetry problems are less severe in such a case. This is because bank lenders tend to be fewer in number than other types of creditors -hence reducing potential hold-up problems- and they have stronger incentives to monitor the firm, which reduces the information asymmetries between the borrower and its creditors. Table 13 reports the weight of bank debt over total debt. It shows that Spanish firms have higher shares of bank debt than the average of the other three countries in 18 of the 21 bins. The table is therefore consistent with the idea that Spanish firms and their lenders react to an unattractive bankruptcy legislation by using means of financing that promote S&M and facilitate private workouts, hence reducing the risk of bankruptcy62. [TABLE 13 HERE]
7. Discussion In this paper we have ignored two factors that could be related to the Spanish business bankruptcy rate. The first is that labor regulations may interact with business financial distress and bankruptcies. The very few works that study those interactions point at the effect of labor regulation on business demography. Claessens and Klapper (2005) suggest a negative relationship between bankruptcy rates and labor regulation. Since restrictive labor laws may 62
Furthermore, since the main providers of mortgage loans are banks, we expect a positive correlation between the weight of bank debt over total debt and the proportion of mortgage debt. In other words, firms whose main creditors are banks may be more likely to avoid bankruptcy via foreclosures.
31
reduce entry rates and entry rates are generally positively correlated with exit rates, more restrictive labor laws may imply lower exit rates and consequently lower bankruptcy rates. Garoupa and Morgado (2006) claim that inefficient bankruptcy procedures may be strategically used by interventionist governments to facilitate employment protection by increasing exit costs for firms. However, our findings do not substantially change when we take into account exit rates, as it was shown in Table 2. The second is that there may be reasons to think that there is some degree of hysteresis in bankruptcy rates and that the Spanish business bankruptcy puzzle may be due in part to the fact that the new law entered into force in substitution of a chaotic and archaic procedure for which there was no demand. We have also explored the possibility of analyzing the effect of the introduction of the 2003 Spanish Bankruptcy act.64 To do this we have investigated whether the capital and asset structures of Spanish firms in 2006 differed substantially from the ones of 2002.65 The only difference worth reporting is that large Spanish firms have partially closed their leverage gap with respect to their foreign counterparts. We are reluctant to draw implications from this observation. But we do take some comfort in the fact that the introduction of the new law, that has led to no changes in the usage rates of formal bankruptcy, has also been accompanied by no changes in the balance sheets of Spanish firms, as this may provide some additional indication of a connection between the usage rates and the financing of productive activities. The view that we have proposed of the Spanish bankruptcy code and its interaction with economic and financial decisions can be summarized in the following terms. • The Spanish corporate bankruptcy law endows the debtor with sufficient leeway to attempt the continuity of firm, but does not guarantee that the continuation decision, which to a significant extent depends on the Court and the administrators or trustees, is made efficiently. In other words, it sacrifices the protection of creditor rights without achieving sizable gains in terms of ex-post efficiency. • The Spanish mortgage system is de facto an alternative insolvency institution, since it grants higher creditor protection and its procedures are faster than those of bankruptcy. • The Spanish personal bankruptcy law is particularly severe towards the individual debtor. Thus, entrepreneurs, unincorporated firms and small firms that pledge personal guarantees to obtain credit have strong
64
We thank Tano Santos for suggesting this exercise. The 2003 Spanish Bankruptcy act entered into force only in the last quarter of 2004. According to bankruptcy experts we have interviewed, it can be claimed that in 2002 there was hardly any anticipation of the new law that could have affected the 2002 balance sheets of Spanish firms. 65
32
incentives to avoid filing for bankruptcy. This may have a negative impact on innovation. • The potential sanctions that a company manager faces in case of bankruptcy make it a very unattractive option from the point of view of expected liabilities. • All the above factors lead to a) the purchase of tangible fixed assets that can be used as mortgage collateral; b) a capital structure with low leverage; c) a high proportion of bank debt, since lenders’ screening and monitoring can offset low creditor protection. • The over-investment in tangible fixed assets, while reducing credit constraints, may generate productive inefficiencies that reduce projects’ profitability. It is unclear whether low usage rates of formal bankruptcy procedures should be considered undesirable in general. But in this paper we have made a first attempt to interpret the strikingly low usage rates of formal bankruptcy procedures in Spain. We have shown that our theory is not contradicted by international comparisons of the stylized features of balance sheets and the profit and loss accounts of nonfinancial firms. According to our view, the low usage rates of the bankruptcy procedure can be interpreted as an indication that the Spanish insolvency legislation is implicitly favoring the choice of projects with low bankruptcy risk but also low returns. As we mentioned at the start, because we only make use of aggregate data, we cannot claim that we have shown that our theory is confirmed by data. But we believe that the stylized facts that we have documented are sufficient at least to encourage further research on this issue along the lines pursued here. We do not have enough evidence to comfortably draw policy implications from our work. But at this stage, if we had to venture directions of change for the Spanish bankruptcy legislation, we would propose to eliminate what we have identified as the causes of the Spanish business bankruptcy puzzle. We would support a corporate bankruptcy regime with more protection of creditor rights (e.g., allowing creditors to force liquidation and to propose a liquidation plan; allowing creditors to speed up the procedure and avoid certain phases in it), more possibilities to direct the appointment of professional insolvency practitioners (and probably actively involving creditors in the appointment, along the lines of the German and English systems and, more recently, of the Italian bankruptcy law), and that relaxes the concern of company managers that an external adjudicator, such as the Court running the bankruptcy proceeding, may impose severe sanctions on them, on the basis of a criterion as vague and elusive as gross negligence in the running of a business. With respect to the personal insolvency law, we believe that some debt discharge would be socially desirable.
33
Appendix A: the evolution of business bankruptcy rates during the crisis. The current crisis in Spain has made the number of bankruptcy filings soar in absolute terms but not in relative terms; i.e., Spain still has one of the lowest business bankruptcy rates. Figure 4 shows that, before the crisis, there were less than 1,000 filings per year. By contrast, in 2009 and 2010 there were around 5,000 per year. In Figure 5 we report the evolution of the bankruptcy rates for several European countries, Japan and U.S for the period 2006-2010. Spain still has remarkably low bankruptcy rates, only slightly higher than those in Greece. [FIGURE 4 HERE] [FIGURE 5 HERE]
Appendix B: The data The main data sources this paper uses are the Bank for the Accounts of Companies Harmonized (BACH), BACH-ESD (European Sectoral references Database) and Amadeus. A.1 BACH and BACH-ESD BACH is a database containing annual balance sheets and income statements of non-financial enterprises for 9 European countries (Austria, Belgium, France,
34
Germany, Italy, the Netherlands, Poland, Portugal and Spain), aggregated by major activity sector and by size. It is the result of the cooperation between the European Commission and the European Committee of Central Balance-sheet data offices (ECCB), whose members are several European central banks. Its main goal is the harmonization of the data to make them comparable across countries. However, perfect comparability has not been fully achieved yet, due to the special characteristics of the national accounting methodologies; but several documents, elaborated by the ECCB and the national central banks, help the researcher determine which comparisons can be made. More information on BACH can be found in European Committee of Central Balance Sheet Data Offices and Bank de France (2010), European Committee of Central Balance Sheet Data Offices and European Commission (2006) and in Cano (1997). This paper uses BACH data on 4 European countries -Spain, France, Germany and Italy- and on the following productive sectors (NACE Rev. 1.1): D – Manufacturing E – Electricity, gas and water supply F – Construction G - Wholesale and retail trade H - Hotels and restaurants I - Transport, storage and communication K– Real State, renting and business activities
Firms are also classified by size, according to the following criterion: Size Class Small Medium Large
Turnover
