Search Menu


Issues list


On the Role of Consistency of the LegalSystem in a Democratic Republic

ISBN 978-9985-870-26-6

Cover image



The World Bank Doing Business Ranking of Quality of Justice: Critical Analysis

1. Introduction

Economic theory, econometrics, and game theory have recently shed new light on the study of diverse social phenomena, including the law and the causes of social and economic development. Economic tools have been used in attempts to understand the effect of laws and regulations in the economic development of a given society. If ‘good laws’ and ‘good institutions’ somehow ‘cause’ economic development, it would be worthwhile to try to identify such laws and institutional arrangements, understand how they foster development, and replicate them everywhere.

The World Bank and other international organisations have generated rankings and measures based on these ideas, including those featured in the six Doing Business reports published by the IFC since 2003 (hereinafter ‘the Reports’). *2 The Reports focus on business legislation around the world, purporting to objectively measure, compare, and report on the quality of laws and regulations affecting businesses in different countries. *3 Their aim is to identify the world’s ‘best practices’ related to the aspects of business regulation their authors consider relevant for entrepreneurship and to benchmark the regulations of other countries in comparison to such best practices. They measure regulation quality, for example, in the categories ‘starting a business’, ‘hiring and firing workers’, ‘getting credit’, ‘registering property’, and ‘enforcing contracts’. The Baltic States have enjoyed high rankings in the Reports, being among the top 30 in the ‘ease of doing business’ ranking since its inception in 2006 (when among 155 countries Estonia ranked 16th , Latvia 26th , and Lithuania 15th ) through to the latest, 2009 release (where among 181 countries they ranked 22nd , 29th , and 28th , respectively). *4

One of the pillars in the Reports’ consideration of the ease of doing business is the measure of ‘enforcing contracts’, for which Estonia ranks 30th , Latvia fourth, and Lithuania 16th . Its aim is to measure the quality of courts, on the basis of the number of procedures, cost for the plaintiff, and the time it takes to enforce a hypothetical contractual dispute in a given economy. The lower these three figures, the higher the ranking.

But does this mean anything? Does a good place in the rankings mean that the Baltic States have ‘better laws’ or ‘better courts’, fostering development better than those of countries with worse ratings? More importantly, does it make sense to reform in order to have a better ranking? Unfortunately, the answer might not be ‘yes’. Using simple law and economics theory, this paper seeks to scratch the surface of the ‘Enforcing Contracts’ section of the Reports, showing that, of the three measures the Reports now employ in relation to contract enforcement, only ‘days to enforce’ seems theoretically sound and giving some suggestions as to how to complement said measure to more meaningfully reflect efficiency in dispute resolution. Of the three measures — procedures, cost, and time — the first needs more solid theoretical foundations, the second is at odds with economic theory, and the third needs improvement to guide policy.

Section 2 offers a brief overview of the Reports, their background, and the issues addressed in this paper. Section 3 critically analyses the measure of number of procedures, and then Section 4 addresses the measure of costs in the light of economic theory. Section 5 looks at the impact of delay in enforcement and makes suggestions for improvement. Finally, Section 6 offers the author’s conclusions.

2. The Reports and enforcing contracts: An overview

The Reports’ stated goals are “to advance the World Bank Group’s private sector development agenda” by “motivating reforms through country benchmarking”, “informing the design of reforms”, “enriching international initiatives on development effectiveness”, and “informing theory”. *5 As these statements and the Reports’ titles suggest, the authors — a World Bank team led by economist Simeon Djankov — have a particular view of the effect of the regulatory environment on development. For them, ‘law matters’ for development, and better (frequently fewer) regulations lead to growth and job creation. *6 Moreover, they very explicitly try to promote legal reform according to their findings, highlighting “top reformers”, pointing at “who is not reforming”, showing “success stories”, and ranking countries according to the “ease of doing business”. *7

Intellectually, the Reports are inspired both by the Law and Finance movement started by Rafael La Porta, Florencio López-de-Silanes, Andrei Shleifer, and Robert Vishny, who tried to apply econometric analysis to the study of the law and legal traditions *8 ; and by the ideas of Hernando de Soto, a Peruvian economist who argues that the legal framework could push people toward the informal economy and prevent them from owning property, doing business, and raising themselves and their countries out of poverty. *9 These ideas were the origin of the New Comparative Economics movement and of the Reports. *10 They are also related to the New Institutional Economics school of thought, founded by Douglass North *11 , for which the institutional framework within which the agents of a given economy operate is determinant of the capability of that economy to achieve development. *12

Although cutting red tape seems in principle a good idea and the wealth of data the Doing Business project is gathering is remarkable *13 , both the Reports and their theoretical background have been subject to serious criticisms. *14 Furthermore, because of the impact in the media and the endorsement by the World Bank and other international institutions (most notably, the United States Millennium Challenge Corporation) *15 , the shortcomings of the Reports may be potentially harmful. *16 For some critics, the procedure for testing the Reports’ underlying hypotheses has been different from the one normally applied in economic research, which involves the risk of the measurement of institutional performance remaining subject to aprioristic policy recommendations:

The indirect cost of [the Reports] from the adoption of defective policies could therefore be huge, for two reasons. Firstly, the authors of the preliminary research are responsible for the subsequent reports. [Although probably insignificant, there could be a] risk that they will tend to search for or interpret the new information in such a way that it confirms their preconceptions[… Secondly], the fact that an institution as relevant as the World Bank is involved in the project covers up its defects and vouches for its conclusions which, in a normal situation, would be taken as preliminary, having a limited effect on policy. For these two reasons — the incorrect procedure and the participation of the World Bank in the project — there is considerable risk that such preliminary conclusions will be taken as final [determinations] and used for establishing wrong ‘best practice’ standards and for taking mistaken decisions in institutional reform. *17

In other words, because of the involvement of the World Bank and other international development aid institutions in the crafting of the rankings, and the potential (or effective) conditioning of aid and investment on how well a country scores in those rankings, the indicators have the potential of being very influential. Therefore, such measures have to be designed and tested with the utmost care and, further, have to be methodologically and theoretically sound, to avoid generating widespread implementation of inappropriate reforms or wrongly punishing the right policies. *18 This is particularly important in view of the emphasis the Reports’ authors (and, ultimately, the World Bank) place on highlighting the ‘good’ and ‘bad’ economies, those that have reformed according to the Reports’ metrics or not *19 , on suggesting the use of the Reports to guide reform *20 , and on how ‘what gets measured gets done’ *21 , suggesting that reforms that do not get measured or do not help the country improve its rankings may be overlooked despite their necessity or convenience.

The ‘Enforcing Contracts’ chapter focuses on measuring procedural laws and judicial institutions. The authors argue that courts should be “fast, fair and affordable” *22 , and they try to measure these qualities by setting up a hypothetical contractual dispute, examine the procedural laws of the countries surveyed to see how the dispute would be handled, check with local counsel, and use the information to build the measurement indices. *23 The indices were originally four: number of mandatory procedures requiring interaction between the parties and/or the court; cost incurred by the plaintiff; estimated time to resolve the dispute; and a measure that disappeared with the 2005 report, termed procedural complexity. *24 With this information, they rank the surveyed countries, stating that courts in richer countries (which arguably have the most ‘efficient’ courts) have fewer procedures *25 , are less costly *26 , and take less time to resolve the dispute. *27 In the 2004 report, the authors blame poverty, legal tradition, and procedural complexity as the main causes of court ‘inefficiency’, and suggest various reforms. *28 The theoretical and methodological background of the indicators and suggestions consists of a research paper authored by Djankov and others (hereinafter ‘the BP’) that argues that legal tradition and ‘formalism’ of procedure are associated with less desirable courts. *29

At first glance, the assertions might seem plausible, but this is less clear once the level of scrutiny is increased. If upon close scrutiny the measures do not prove methodologically or theoretically sound, their continued usage in the rankings may mislead governments into pushing for unnecessary or even incorrect reforms, hampering rather than fostering development. If that were to prove the case, the World Bank should reform the Reports forthwith, abandoning such measures for ones more meaningfully reflecting a legal system’s efficiency of enforcement of business contract arrangements.

T he problem is that the ‘enforcing contracts’ chapter does not pass such scrutiny. Firstly, the Reports assume a certain judicial dispute resolution procedure to be ideal, one that has very few steps and formalities. This is most apparent in the BP, which explicitly assumes the ‘ideal’ (in the sense of better justice) of the neighbour dispute and constructs the ‘formalism’ index to measure departures from said ideal. Thus was the ‘Number of Procedures’ indicator derived, under which a country scores better the fewer steps are necessary for enforcing a contract. As Section 3 attempts to show, ranking based purely on number of procedures appears baseless, as claiming the neighbour dispute is ‘ideal’ is a mere assertion without sufficient theoretical or authoritative backing. Moreover, a simple reductio ad absurdum shows the ‘ideal’ justice promoted by this chapter of the Reportsis arguably no justice at all. It may even fail the test of econometric regression. The assertion that using fewer procedures leads to ‘better’, or more ‘efficient’ justice, whatever that may mean, is not obviously true and needs to be argued properly.

Secondly, the Reports, elaborating on the neighbours’ ideal, posit that enforcing a contract in court should be cheap, fast, and fair, meaning that the (hypothetical) plaintiff should be able to go to court, get a judgment, and collect the debt in as short a time as possible, without any lawyers or appeals and paying as little as possible in court and attorneys’ fees. For the Reports, a court system that is able to achieve such a state of affairs would be ‘efficient’. Section 4 seeks to show that this idea is completely at odds with the basic literature on the economic analysis of procedural law and rational choice. For these theories (not disproved or argued against by the Reports or the BP), the aim of procedural law is to minimise the total costs of settling disputes for society as a whole: the ‘administrative’ cost of deciding the claim for the parties and the state and the cost of the case being decided ‘wrongly’ (not according to substantive law), and thus giving the wrong behavioural signals to society (taking into account the number and the quality of cases allowed to be filed). These costs are minimised in principle by doing the contrary of what the Reports prescribe: with out-of-court settlements that save on time and money spent on trials, which are more likely to come about the costlier going to court is for the parties; by enhancing the role of professional lawyers and judges; and through the existence of appeals.

However, the time measure does give some hint of judicial economic efficiency given the disruption that judicial delay may have on plaintiff claims. Section 5 thus argues that costs and procedural hurdles may be redeemed as efficiency gauges if the focus is placed on the issues that exacerbate the disruptive effect of time — particularly the enforceability of settlements themselves.

If the rationale for defining and detecting ‘best practices’ is flawed, the policy implications may be correct only by chance. Therefore, if the Reports are not framing the correct picture as regards enforcing contracts, then using their indicators to inform policy and reform may be harmful and should be discouraged.

3. Puzzling maths: The fewer the procedures, the better the court

3.1. A thin concept of an ideal court

The Reports have a special concept of what desirable courts administering desirable contract enforcement processes look like. For them, an ideal court uses few procedures and specialises in commercial matters; procedures do not require writing, lawyers, or legal argumentation; the judge is a layperson who does not need to give legal justification for the decision rendered; going to court is free or very cheap, with no court or attorney fees; there are few rules of evidence; and there are no appeals. *30 Such a court would be ‘fast, fair, and affordable’. Moreover, according to the Reports’ number of proceduresindicator, the fewer procedures necessary to enforce the model contractual dispute, the better the court. *31

The description is rather surprising, since it would count as ‘ideal’ procedures that might intuitively not seem fair. In the extreme, the ‘best’ possible court under this scheme would have only one procedure and a (merchant) lay judge. This can hardly be anything else than a plaintiff orally telling her story and the lay judge deciding on the spot without consulting the law, what the defendant has to say, or anything else. Let us imagine, for example, the following dispute settlement process.

D borrows € 100 from P and promises to pay back in three months the € 100 plus € 10 interest. One month later, the legislature validly passes a statute prohibiting and making void charging interest for non-banking loans of less than six months’ time term, but allowing the recovery of the principal. At the end of the third month, D pays P € 100. P claims the interest, D refuses to pay, citing the statute. P goes to the court, which does not charge any fees and therefore has a long queue, and when his turn arrives he explains the situation to the judge orally. She hears P and believes his contention that D owes him € 110 and not € 100. Nobody writes anything down, and she gives no justification for the decision. *32 She then asks P to go with the sheriff to D’s house, notify him of the decision, and seize the € 10 from D’s assets. The decision of the judge, who is a layperson, cannot be appealed. Let us imagine the same case, but now D never paid back the € 100, and P is a person with a bad reputation in the neighbourhood. P comes to the courthouse, the judge hears his case, and when P is finished the judge denies the claim immediately without giving any explanation. Again, there is no appeal, and P cannot question the judge’s impartiality.

Albeit arguably extreme, the process depicted above bears many characteristics the Reports consider desirable: it is free, is fast, and has few procedures, with no writing, no lawyers, no legal justifications, and no appeal. It would be almost instant justice — if it were justice at all. And, of course, it is far from being a ‘fair’ procedure. D’s clear legal right to not pay interest was denied in the first case. He had no opportunity to present a defence or correct the judge’s mistake in applying the law. Some procedures can be superfluous, but others may have a justification in adding fairness. For example, without proper notification of lawsuits against her, a person may not be able to adequately defend her rights; the plaintiff has the whole statute of limitations period to prepare the lawsuit, but the defendant may have just a few days to gather all the evidence and counter the plaintiff’s arguments; the participation of lawyers and professional judges charged with the duty of iura novit curia help to ensure respect for the rights of each party; and so on.

More generally, the different procedures and institutions within a dispute settlement process may be devices to, apart from reducing the ‘cost of error’, maintain what Shapiro calls the basic social logic of courts — the triad decision structure of the impartial third party deciding the dispute — from breaking down into a bullying two against one. *33 He argues that, although the basic social logic of courts is the triad of the parties and an impartial third, the triad is unstable because when the third decides the social logic becomes a defeat by two against one. Thus, a “substantial portion of the total behaviour of courts in all societies can be analyzed in terms of attempts to prevent the triad from breaking down into two against one”. *34 The process’s fairness arguably contributes to convincing the losing party “that he should obey the third man because he has consented in advance to obey” *35 by participating in the process.

On a more legalistic note, the International Covenant on Civil and Political Rights sets forth in its Article 14 what a minimally fair procedure should be. The Reports’ ideal court is a far cry from what this covenant prescribes. *36 The World Bank being an international institution related to the United Nations — in turn, devoted to fostering respect for human rights as embodied in international law — the Reports adopting such a weak concept of a ‘fair’ process is rather contradictory. *37

3.2. In search of a rationale

What is the Reports’ rationale for advocating such a model of dispute settlement? The Reports state that “fewer procedures are associated with reduced time and cost and with perceptions of improved fairness”. *38 Let us focus first on the fairness perception. The Reports mention that fewer procedures and less of what they term procedural complexity are associated with “perceived fairness” and “less corruption”. *39 These two measures, subjective in nature, are built upon figures from the World Business Environment Survey, a structured survey administered at the managerial level in enterprises around the world. *40 That seems odd, because the 2004 report harshly criticises subjective measures a few pages earlier and purports to depart from them *41 , on several grounds: “[a] large body of evidence shows that survey questions on perceptions do not always elicit meaningful responses”; are subject to design biases, varying responses due to the different scales used, uninformed answers, lack of a reference point, and sample selection issues; and “are often driven by general sentiment but do not provide useful indicators of specific features of the business environment”. *42 Therefore, when one applies the Reports’ own logic, the inference is somewhat questionable. Furthermore, the measures have the problem of reverse causality: it may be the case that because of the negative general perception of the courts, some procedures have been introduced to enhance fairness, not the other way around.

The Reports also mention that “[c]omparing by income quartiles […] the richest jurisdictions […] have the lowest number of procedures” and that “legal tradition is also associated with the efficiency of contract enforcement”, with the countries with Nordic legal origin having on average a lower number of procedures than the ones with a French legal origin. *43 However, it is not at all clear how ‘number of procedures’ relates to development or a country being relatively rich, nor is it obvious how a Nordic or French legal origin may be better or worse for development, or what the direction of causality is, despite the claims of the Law and Finance movement. *44

More importantly, the theory behind the inferences is not so clear either, the only suggestion being that “a higher number of procedures is associated with more opportunities in the judicial system for extracting bribes”. *45 If the judge in a given case is the only official able to take a bribe to speed up the process or favour one party, it may be irrelevant whether she takes, say, 10 bribes of € 100 in an equal number of procedures or one € 1000 bribe just before issuing a judgment. The assertion would make more sense if the procedures that open bribe opportunities have to be performed by different and independent officials, which may lead to a sort of ‘tragedy of the anti‑commons’ bribing scenario. *46 Unfortunately, the Reports do not explain in any depth why procedures are relevant to development. We should look for a rationale elsewhere.

The Reports are an elaboration of previous scholarly research studies published in peer‑reviewed journals. *47 The idea of the desirable process being very simple comes from the BP, which tries to show that courts’ performance and quality are determined by the level of ‘procedural formalism’ (how the law regulates their operation), trying then to measure empirically the determinants of such ‘formalism’ and its consequences for the quality of dispute resolution in courts. *48 The authors loosely apply the concepts of judicial efficiency, effectiveness, performance, and quality. Ultimately, the two measures of outcomes are the estimated duration *49 of dispute resolution and the ‘public’ perception of the quality of the legal system (using data from the World Business Environment Survey). *50 The two measures are subjective and subject to the objections mentioned above.

It is also apparent from the BP the manner in which the particular ‘ideal’ court notion seen here originated:

In a theoretical model of an ideal court, a dispute between two neighbors can be resolved by a third on fairness grounds, with little knowledge or use of law, no lawyers, no written submissions, no procedural constraints on how evidence, witnesses, and arguments are presented, and no appeal [Shapiro 1981]. […]

According to Shapiro [1981], the essence of an idealized universal court is the resolution of a dispute among two neighbors by a third, guided by common sense and custom. Such resolution does not rely on formal law and does not circumscribe the procedures that the neighbors employ to address their differences. Yet courts everywhere deviate from this ideal. *51

Then, the authors generate indices of ‘formalism’ as departures from their ‘neighbour model’, which looks a lot like the hypothetical case at the start of this section. They define seven aspects of it, including professionalism (professional judges and lawyers), whether procedures are oral or written, the need for legal justification, the characteristics of the rules of evidence, the existence of appeal, the characteristics of the engagement formalities (service of process), and the procedure count. *52

As above shown, the authors do not argue why they think the ideal court is the neighbour model; instead, they simply cite Shapiro *53 as authority. The problem is that Shapiro’s book does not seem to provide much authority for the claims the BP makes. He does not seem to propose a normative ideal court procedure to which all courts should approximate but, rather, suggests what the basic logic of the institution of a court is and how that basic logic has given way to modifications and additions in different countries and regimes. In short, Shapiro’s ‘ideal’ looks not like an ideal in the sense of ‘the best court’ but an ideal in the sense of the ‘basic idea of a court’. *54 Moreover, as argued above, procedures could be viewed as efforts to prevent Shapiro’s triad from breaking down; and, in that sense, the prescriptions of the BP and the Report as to the number of procedures and the lack of necessity of active participation of the defendant are directly at odds with what Shapiro describes as the logic of consent. *55 The example at the beginning of this section would not elicit D’s consent to the first decision and thus would fail as a triad.

Additionally, comparing the reasoning of the BP and the Reports reveals a peculiar shift. The BP identifies ‘formalism’ as the cause of ‘bad quality’ courts; with legal origins as the root of formalism. That was still the argument in the 2004 report, with ‘procedural complexity’ aggregating all formalism indicators except number of procedures. But since 2005 that indicator has been absent without trace or explanation. The only measured ‘cause’ of the ‘quality’ of courts in the subsequent reports is (apart from cost) the number of procedures. But the number of procedures is notwhat the BP pointed to as ‘cause’. So even if we considered the BP to be correct, it still would not provide support for the claim that fewer procedures is associated with ‘better justice’.

In sum, the court model the Reports promote has yet to be properly argued and justified.

3.3. Having fewer procedures does not mean having faster trials

Additionally, although the 2004 report claims that “fewer procedures are associated with […] reduced time” *56 , this claim does not seem to be supported by the data in that report or the subsequent ones. The countries ranking in the top and bottom 10 positions as regards number of procedures seldom match the corresponding top and bottom 10 for average length of proceedings. The 2004 report shows only Tunisia among the top 10 for both variables and only Angola among the bottom 10 for both measures. In the 2008 and 2009 reports, only Singapore and Hong Kong are among the 10 countries with fewest procedures and among the 10 with shortest trials, while only East Timor is among the bottom 10 in both counts. Irish processes, with 20 procedures, are 515 days long; Estonian ones are shorter (at 425 days) even though they involve 36 procedures. Sierra Leone has twice the Irish procedure count but the same process length, and in Brunei the process takes a very similar 540 days but has the highest number of procedures: 58. A simple regression between the two variables, using the data from the Doing Business site (calculated with the 2009 report’s methodology) *57 , confirms that the correlation between these variables is weak and procedure number explains less than 10% of the trial length. No correlations ever reach above 0.307, with R2 never above 0.094, weakening any claims that having fewer procedures in general means having quicker trials. *58

In sum, the contention that the use of fewer procedures to enforce the hypothetical contract is associated with fairer and faster courts needs further argumentation and proof if it is to be convincing. If procedure number is not an independent ‘outcome’ measure of court quality *59 , the Reports’ advocacy for trials with few steps seems misplaced. *60

4. Forgetting economics: What appears to be cheap justice might be actually costly

Even if one were to concede that a fast and simple judicial process is somehow ‘ideal’ or desirable from the individual plaintiff’s standpoint, the Reports proposed process still may not be the most efficient from the point of view of society as a whole. Another element of the contract enforcement ranking is the enforcement cost for plaintiffs, computed such that a jurisdiction scores better the cheaper it is to go to court and enforce the hypothetical contract. *61 The underlying assumption is that justice should be cheap and within the reach of everybody, rich or poor, regardless of the amount the dispute concerns, to promote entrepreneurs resolving their disputes in court and avoid informal justice. For the Reports,

Courts have four important functions. They encourage new business relationships, because partners do not fear being cheated. They generate confidence in more complex business transactions by clarifying threat points in the contract and enforcing such threats in the event of default. They enable more sophisticated goods and services to be rendered by encouraging asset-specific investments in their production. And they serve a social objective by limiting injustice and securing social peace. Without courts, commercial disputes often end up in feuds, to the detriment of everyone involved.

Companies that have little or no access to courts must rely on other mechanisms, both formal and informal — such as trade associations, social networks, credit bureaus, and private information channels — to decide with whom to do business. Companies may also adopt conservative business practices and deal only with repeat customers. Transactions are then structured to forestall disputes. Whatever alternative is chosen, economic and social value may be lost. *62

In other words, if it is cheap to go to court for an entrepreneur when a new customer defaults, she does not need to be too careful in choosing customers and making client‑specific investments, expanding her potential market, and improving her capacity to compete. The wide availability of low-cost contract enforcement would promote entrepreneurship and thus be development-enhancing.

This ‘entrepreneur-empowering’ story looks plausible. However, little thought has been given to the efficiency of a cheap judicial contract enforcement scheme. It may well be the case that cheap enforcement for a person as an entrepreneur may mean costly government spending and heavy taxes for her as a taxpayer (or reallocation of funds from other programmes, such as those for hospitals and schools), perhaps an unreasonable high price to pay.

4.1. The bill that the plaintiff picks up is not the whole story

The data notes reveal, interestingly, that the enforcement cost indicator only comprises the cost incurred by the plaintiff as court, attorneys’ and enforcement fees. *63 However, the cost of the court system may be much greater, and such costs are not captured by the index. Perhaps in some cases the judicial branch is exclusively funded by explicit fees to the users, but that may not be the case always: the fixed costs of the judicial branch (such as salaries and rental costs) may be borne by the government budget from taxes on businesses and consumers. Moreover, following the Reports’ logic, an ideal commercial court could be one that enforces all contracts for free, coupled with a 100% government subsidy of attorney’s fees, in which case the whole cost would have to be borne by the taxpayers. But if two systems had identical procedures and equal total cost, only differing in that one is 100% funded by fees and the other 100% funded by taxes, it would be odd to say that the second is ‘less costly’ and therefore preferable from society’s point of view. Furthermore, there could be specific taxes paid at the time of signing a contract that make subsequent enforcement in court less expensive, faster, or even possible. In our example, the second system may be funded by a stamp tax paid upon signing a contract as a condition for its validity, making all contracts for which the tax was not paid effectively unenforceable. Such a system would be ‘free’ for the Reports yet clearly increases the cost for the plaintiff. *64

That the state bears part of the judicial burden might help to explain why richer countries (with presumably bigger budgets) have faster and ‘cheaper’ (in the Reports’ sense) courts, assuming, for example, that the judiciary’s main costs are fixed, such as salaries and costs of facilities. On the other hand, it could be argued that computing the cost as a percentage of the per capita income *65 already incorporates the increased resources available in rich countries. Again, countries with similar per capita income might have different judicial expenditure; countries with a similar judicial budget could use such resources with different degrees of efficiency. In any case, the Reports fail to take into account both a part of the cost and how efficiently it is incurred. *66

More importantly, focusing on how affordable courts are for plaintiffs and simplistically stating that the more affordable the better overlooks the simple economic logic of demand and supply of goods and efficient allocation of resources. The lower the price of an item, the greater the quantity demanded. But because the production of that item is not cost-free and resources are not infinite, producing too many of those items would lead to waste. Therefore, there is a point at which the marginal cost of producing an item is equal to the price the demand side is willing to pay for the last unit offered, such that the quantity offered is equal to the quantity demanded, and the price paid is equal to the marginal cost. Producing more or less than this would be wasteful. *67

This is no different for contract enforcement services. If going to court is too cheap, there will be just too many lawsuits. *68 The service is not costless; it needs judges, books, computers, courtrooms, clerks, lawyers, sheriffs, expert witnesses, juror compensation, and so forth. As in any enterprise, some costs will be fixed and some will depend on how many lawsuits are brought. There will then be a marginal cost of the enforcement service: the cost of having one more lawsuit. In the end, somebody will have to pay the bill. If the only one paying is the government and the cost for the plaintiff is zero, then she will file a suit even if the expected benefit for her is very small, regardless of the litigation cost for society. In a world of limited resources, subsidised demand may lead to queues, inefficient allocation, delay, and waste. A free court would be ideal for the Reports’ cost indicator but would be a nightmare of a court system, with a huge budget that is never enough and full of unresolved cases that will take years to conclude.

Conversely, if the fees are too high, there will be too few lawsuits. *69 Lawsuits have ‘positive externalities’, benefits that are perceived not by the parties but by society. For instance, the decision in a case between two persons may clarify a point of law that prevents (or promotes) the filing of similar cases between other people. Likewise, the likelihood that a lawsuit will be successful against a person if she breaches a contract may affect her decision to breach it; that makes publicising the outcome of breach of contract lawsuits important to deter (or positively sanction) contract breaching. *70 The Reports’ account of the courts’ development ‘functions’ quoted above could be construed as a story of positive externalities. Therefore, it may be the case that judicial enforcement of contract should be subsidised to promote development. Such considerations, however, should not be translated into a simplistic grading scale in which ‘the cheaper the better’; on the contrary, all cost and benefits should be taken into account if we want to accurately measure court ‘efficiency’.

4.2. The peculiar relationship between cost and chances of settlement

Besides overlooking the cost borne by the state, the Reports fail to consider the courts’ efficiency in generating the usual instruments by which disputes are solved: settlements. As the famous Hollywood quote puts it, “the whole idea of lawsuits is to settle” *71 , and some argue that fewer than 10% of the civil disputes filed in the US actually require a trial to be concluded. *72 Consequently, the Reports fail to consider more than 30 years of scholarly research in the economics of procedural law. Ever since the seminal work of Landes *73 , economic analysis has shown that, for an equal cost of error or quality of judicial decision, a more expensive court proceeding is preferable to a less expensive one because it promotes settlements and discourages frivolous suits, which in turn saves on costs, prevents queues, and promotes efficiency. *74

How is that so? Procedural rules have the function of regulating the application of substantive law. *75 Their use generates costs, not only ‘administration’ costs (such as courts, lawyers, and time) but also the cost of the court deciding ‘wrongly’ (i.e., contrary to what the substantive law prescribes) and consequently affecting incentives and behaviour in society, what Cooter & Ulen call the ‘error’ cost. *76 If we are concerned about court efficiency, we should assess whether the particular procedural law and court system concerned minimises the sum of administrative and error cost. If for a dispute an out-of-court settlement could imitate the outcome of the court decision (i.e., the trial cost of error is the same as the settlement cost of error) and the cost of reaching a settlement is lower than the administrative cost of trial, then it is socially preferable that the dispute be settled out of court to minimise administrative costs. In such a case, the only costs remaining would be the error costs. Arguably, these costs could be minimised by requiring more information from the parties (assuming that the more and better information judges have, the less the chance of error), providing for closer scrutiny of the evidence and its handling, maximising the chances for the parties to interact under the court’s auspices (providing opportunities to negotiate a settlement in an ‘impartial’ environment), requiring extensive and sophisticated legal argumentation from the parties’ counsel (that may better guide the judge in the decision), having professional judges available who understand such arguments, and allowing appeals (which are especially designed to correct mistakes). *77 That is the opposite of what the Reports prescribe.

Furthermore, from a rational choice standpoint, an efficient judicial procedure system is one that gives incentives to the parties to minimise costs and thus promotes settlements, discourages frivolous suits *78 , and minimises the risk of obtaining wrong judgments. What kind of procedure delivers these outcomes?

A settlement is possible when the expected value of the trial for each party (the judgment’s expected present value net of its expected cost) is lower than the expected value of the settlement net of its cost. In other words, if the plaintiff is better off going to trial than accepting the defendant’s offer for settlement, then she will not co-operate; she will go to trial to get more. If the defendant would lose less by letting the trial continue than the price the plaintiff is demanding to settle, then he will not co-operate and would refuse to settle. Since a settlement could achieve the same outcome as the trial, the only thing the parties would be bargaining for would be the trial’s expected cost — that is, what the parties could save and divide between them. To put it differently, the savings in transaction costs create a co-operative surplus, which is equal to the joint costs of litigating net of the costs of settlement. Therefore, the higherthe expected cost of trial, the lower the non-co-operative reserve price, the greater the co-operative surplus and the wider the room for bargaining to achieve a settlement. Conversely, if the parties have different expectations as to the outcome of the trial (value times probability) and they are both too optimistic, the non-co-operative reserve price would rise, reducing the surplus and with it the chances of settlement. Then, the higher the accuracy with which the parties could predict the judgment, the higher the chances of reaching an agreement. *79

Similarly, if the expected cost of suing is too low for the plaintiff but too high for the defendant, the former could file a suit even knowing that the expected net value of the trial judgment is low or negative (a frivolous ‘nuisance’ suit) and ask the defendant to buy her off under the threat of continuing the suit, thus causing the defendant to incur great losses. In such cases, again, the costlier the trial is for the party who would eventually lose (in this case, the plaintiff), via either a high bill or a ‘loser pays all’rule for allocating trial costs, the less the incentive to file nuisance suits. *80

4.3. The efficiency of having appeals

Finally, appeals are an instrument to reduce the error cost at low administrative cost, thus promoting social efficiency. *81 To consider a hypothetical situation similar to the one Cooter & Ulen use in their book *82 , imagine that in a dispute a judge, at an administrative cost of € 1000, has a 40% probability of deciding wrongly, generating € 25,000 of error cost. Then, the expected social cost of the judgment without appeal is € 11,000. Imagine now that a court of appeals is created, that parties file an appeal when the trial judge makes a mistake, and that the appeal court has the same 40% error probability as the trial judge and an administrative cost of € 1000. In such a case, the expected social cost of the dispute settlement would be reduced to € 5400 because the reduction in the cost of error would more than compensate for the higher administrative costs. *83 Of course, this is just a numerical example and different numbers yield different results. However, it shows that appeals are not generally ‘inefficient’, as the Reports suggest they are. Quite the contrary, under conservative assumptions, potential efficiency gains are apparent once the error cost is considered. Moreover, the appeal stage is also subject to the strategic interaction for a settlement, which will discourage filing appeals and encourage settlements if appealing is costly and the parties know the expected value of the appellate decision. In fact, few cases would reach appeal, and the plaintiff would consider the appeal’s expected value at the beginning, when deciding whether to sue or not. *84 Therefore, appeals may be helpful tools for preventing dispute settlement error cost, rendering the court system more efficientand not less. This is the opposite of what the Reports suggest.

In sum, the economic analysis of procedural law yields findings in direct contradiction with what the Reports depict as desirable courts. The costlier the trial, the greater the chance of a settlement, and thus the lower the social cost of resolving the dispute. Moreover, if trial fees are too low they may generate a non-optimal number of lawsuits. Additionally, the involvement of professionals and the existence of appeals may prevent or reduce the generation of error cost, making a system with professionals and appeals, ceteris paribus,more efficient than a system without lawyers or appeals. In sum, the Reports’ ‘cheap’ justice for the plaintiff is ‘costly’ for society.

Nevertheless, room might remain for the Reports to do some good. Since settlements’ likelihood depends on the parties’ perceptions, the information the Reports provide on the actual costs and length of trials could help parties with cases identical to the hypothetical to calculate better the cost of trial and the value of the surplus to be divided.

5. Hamlet’s dread: The disruptive effects of the law’s delay

Hamlet’s first lines in Act III of that Shakespeare play are among the most famous in world literature. A few words after “To be or not to be”, Hamlet mentions “the law’s delay” as one of life’s tragedies. *85 Centuries later, the delay in enforcing contracts around the world is still significant, as the Reports, for instance, point out. *86 Although court delay does not obviously relate to development directly *87 , it has been argued that strong protection of property rights is a necessary condition for development *88 , and having an effective enforcement system addressing property rights violations thus might be development-enhancing.

In the context of a legal conflict, the delay to resolve the dispute may disrupt the underlying property rights — and economic efficiency in general — in several ways. Firstly, and fundamentally, property rights that are in dispute cannot be bought or sold on the market at their usual market value (if they can be sold at all); there is a probability of the seller losing the dispute totally or in part and the right then losing its value or even becoming worthless, determining a risk premium. Every day that passes without the dispute being resolved is one more day in which the affected resources cannot be routed to the most efficient use, one more day of opportunity cost and economic loss.

Secondly, the passage of time allows the evidence backing a legal claim to decay, making it more difficult to gather witnesses (who may forget, become incapacitated, or even die before appearing in court) or necessary documents (which may get lost or damaged), etc. Moreover, delay in resolving a dispute may trigger statutes of limitation of related rights, affecting their value. As the probability of a lengthy process grows, the probability of having a worse case to put to trial also grows, and the value of the legal claim and the underlying right diminishes.

Thirdly, the very interest of the parties in the dispute may shift or disappear with the passage of time. Plaintiffs and defendants may go out of business; corporate parties may be sold, taken over, dissolved or their management be changed or retire; and parties that are natural persons may die or grow tired of going to court. Furthermore, some disputes may be very time-sensitive and the value of pursuing them may change dramatically, depending on whether they could be resolved rapidly or not. For instance, a dispute over a corporate merger that is not resolved quickly may affect the share price of the companies involved, reduce or destroy the profit the parties expect from the deal, and remove the incentive to make it, eliminating the efficiency gains that may result from the merger.

Fourthly, and related to the first issue, time is money and a euro today is worth more than a euro tomorrow. If the successful plaintiff is not granted interest and inflation adjustment on the amount claimed for the time of the enforcement process at the end of it, or the interest rate is lower than the (inflation-adjusted) market rate, every day of delay in enforcement implies economic loss for the plaintiff. Conversely, the defendant may have a financial benefit from the delay, as she does not have to pay unless and until the court’s decision has been issued, upheld (if appealed), and executed.

Finally, parties in court systems subject to the above-mentioned factors may be well aware of them and adjust their behaviour strategically to manipulate the delay to their advantage. For a defendant, for instance, it would often be financially and economically efficient to let the process continue as long as possible because the likelihood of an adverse court decision and its expected amount diminish the later the decision is issued. Conversely, known court delay may also strategically benefit a plaintiff, who may file a frivolous claim against a corporate merger proposal and take advantage of the delay to ‘kill’ the deal. Likewise, a person suspecting she will be sued may engage in perverse forum shopping, precluding the claim against her by filing a counterclaim in a court that lacks jurisdiction but will take too long to issue a decision declining jurisdiction, forcing the appropriate court to wait and improving her negotiating position vis à vis the suing party.

In these contexts, settlements that prevent trials and save on total social cost are less likely. If the delay in the process asymmetrically benefits defendants, if the cost of postponing the dispute resolution (including the present value of the eventual judgment) is lower for her than the cost of settlement, she will have an incentive to try to make the litigation as slow and difficult as possible — filing obstructive appeals, introducing extraneous witnesses, contesting every procedural issue, contesting the court’s jurisdiction, etc. — instead of trying to settle, investing effort, time, and money that may have better uses. Therefore, efficiency would require mechanisms to prevent this kind of abuse. Note that the abuse comes not from the mere existence of appeals or other procedural hurdles but from the economic and financial effect the delay has on the plaintiff’s claim and the underlying property right. In other words, because of glitches in the system, time favours one party over the other, and those problems should be solved just enough for the incentives to be again in favour of settling.

L ikewise, it may be the case that settlement as an institution cannot perform the cost- a nd time-saving function. In some jurisdictions, for instance, settlements may be too expensive or difficult to complete, or they may not even be legally possible. As mentioned above, if the cost of settlement for one party is higher than the cost of trial for him, it will be rational for him to go to trial instead of settle.

Furthermore, even if settlements are not too expensive, they may not be helpful in certain contexts. A settlement in a case may save the parties going to trial but may need another lengthy process to be actually enforced and let the plaintiff collect the amount agreed upon. Many settlements feature the defendant paying a lump sum to the plaintiff and the plaintiff relinquishing her rights and abandoning the trial, all at the same time; the court is notified after the fact that the dispute has been settled and the trial is over. In many situations, however, it would be necessary or convenient that the defendant promise to do additional things (such as refrain from doing something), the parties may have reached an agreement about the legal rights but the amount to be paid is still undecided, or the settlement might be void (unless and) until the judge endorses it. In such cases, some elements would remain pending and there will be a chance of a settlement breach, or, put more plainly, the plaintiff will still have to do a lot to collect her money. Therefore, if a settlement is not equivalent to a court judgment and has to undergo a long process to be enforced, the trial time and cost saved may not be substantial. *89 Imagine, for example, a jurisdiction where a court in average trials takes 100 days to issue a decision but it takes a further 100 days for the successful plaintiff to enforce the judgment and finally collect. If the parties reach a settlement that saves half of the first 100 days (for instance, agreeing on the allocation associated with negligence but contesting the damages), or the settlement has the same status as a normal contract subject to breach (in which case, a trial of, let us say, 50 days will ensue), or if the settlement, even if equivalent to the judgment, requires completion of the post-trial execution process, the time and cost saved by settling will account for only a fraction of the total time and cost involved.

Taking into account the considerations stated above, it may be possible to reform the indicators to more accurately measure the effect of court delay and efficiency. Although designing a statistical indicator is beyond the scope of this paper, three areas may be explored to reform the Reports. First, most objections made above to the number of procedures indicator focused on the trial part, wherein the legal rights are determined and fairness has to be attended to particularly. By contrast, the post-judgment part of the process might prompt fewer fairness concerns, as in that stage the legal entitlements have already been defined. Moreover, settlements tend to be better in saving on trials than they are in saving on judgment enforcement (to which they may be subject just as court decisions are). Therefore, the indicator may still play a useful role if its emphasis were to shift to assessing judgment enforcement and enforcement asymmetries between settlements and court decisions. The Reports, although retaining the opacity in how they reach their conclusions, have started to raise these issues by changing their recommendations. *90 However, they have yet to shed the mechanical assumption that a system with fewer procedures is automatically better.

Second, settlements are costly and the cost differential for each of the parties between settling out of court and going to trial shapes the incentives that guide their behaviour. If the enforcement cost indicator were to focus not only on court costs but also on the cost of alternative dispute resolution mechanisms and their relative enforcement value, or on the availability of financing for plaintiffs to go to court, without simplistically assuming that ‘cheaper is better’, it could yield a wealth of useful information.

Third, how different jurisdictions deal with the disruptive effect of court delay may be something useful to measure. Therefore, it might be worth knowing how easy obtaining a preliminary injunction is in each jurisdiction, whether there is a gap between the interest rate granted to the plaintiff in judgments and the market rate (or whether adjustment of the claim to offset inflation is allowed), whether there are adequate measures to protect the evidence and the defendant’s solvency, and so forth.

6. Conclusions

The first report was entitled Understanding Regulation. ‘Understanding’ implies a humble enterprise, one of observing, theorising, testing and validating. With later issues, the statements of the Doing Business project unfortunately became bolder and their sweeping conclusions more assertive. In six short years, the focus shifted from the search for knowledge to a push for reform, and in that process the concern for backing the indicators with solid theory was left behind somewhat. Thus, the Reports continue measuring legal institutions according to a preconceived supposed ideal, the neighbours’ dispute settlement, whose basis in reasoning is not solid. The measures that are utilised on such a basis are presented as bold facts, but important shortcomings lie under the surface. From these shattered measures, the Reports go further then, to issue policy recommendations, which have a great chance of being flawed and harmful.

Even if we were to concede that fewer procedures and less formalism lead to better courts, the Reports would still not be saved. For more than 30 years now, the economic analysis of procedural law has shown that the most efficient procedural rules (in the sense of reducing the costs for society as a whole) are the ones that establish a court system that elicits the optimal number of lawsuits, minimises the probability and cost of issuing a wrong judgment, and also minimises the administrative cost of resolving the disputes — by promoting out-of-court settlements. Such a system requires a procedure completely different from, and practically the antithesis to, the Reports’ ideal; that is, the costlier and the more legally sophisticated, the better.

However, court delay may be a significant problem, as the passage of time affects the value of claims and the underlying property rights and also may hamper a court system’s ability to reduce the administrative and error cost of disputes. If the Reports’ measures were to shift their focus back to ‘understanding’ and explore how court systems around the world cope with the disruptive effect of delay and shape litigant behaviour, they could contribute greatly to development and justice.

The Reports have the aim of motivating reforms through country benchmarking, informing the design of reforms, enriching international initiatives on development effectiveness, and informing theory. The ‘Enforcing Contracts’ chapter has certainly missed the target.