Friday, August 07, 2015

Reducing delays in litigation by reshaping the incentives of litigants

by Shubho Roy.

Judicial delays are a major problem in India. There have been a number of attempts to solve these through introducing new legislation or tweaking existing laws. The tweaks usually involve putting ad-hoc numerical limits on the number of proceedings or delays. This approach has failed. A different approach is to create incentives for parties to not delay legal proceedings. This approach has been used worldwide with success. One example we show here, from the US, is Rule 68 of the Federal Rules of Procedure which sets up an interesting game to speed up litigation.

The problem

In enforcing contracts, India ranks 186 out of 189 countries. Judicial delays in criminal cases probably cause even more harm. To solve this problem the government has tried quite a few things. Amongst them are:

  1. The Arbitration Act and Conciliation Act, 1996 was made with the objective of providing litigating parties (mostly in commercial disputes) a system outside the court system through arbitrators, but within a legal system of the Arbitration Act. This law succeeded an older law of 1940 and was supposed to make India's law aligned with international law of arbitration.
  2. The Code of Civil Procedure which governs court proceedings in civil disputes was amended in 1999 (effective from 2002) requiring courts provide a maximum of three adjournments to a party in a case (times a party can delay a court proceeding for the day). This rule appears to be followed more in its breach. Similarly the costs imposed on parties for adjournments are puny compared to the actual costs in an adjournment. E.g.  Bombay caps costs for each days proceeding at Rs.100.

These attempts have not resulted in improved arbitration or reduced court delays. As a recent arbitration order against India in an arbitration under a bilateral investment treaty notes: An international investor could not enforce the arbitration award (i.e. legally collect the award) after winning the arbitration for a period of 8 years.

The government proposed an ordinance to amend the Arbitration Act, 1996 to speed up the process of arbitration. Amongst other changes, two key proposals are:

  1. A time of limit of 9 months for arbitrators to finish proceedings. If the time for proceedings exceeds the limit, the arbitrator will have to apply to the High Court to get an extension. The High Court may prevent arbitrators with long delays from taking up new proceedings.
  2. The government will cap total fees payable to an arbitrator.

These quantitative restrictions and price controls have three features:

  1. They are not new, and have been tried multiple number of times before.
  2. They have been a resounding failure in the past.
  3. They have many unintended consequences.

While speeding up arbitration is a step in the right direction, at the end if the losing party does not cooperate, the coercive power of the state has to be exercised. The Ease of Doing Business report notes that a contract enforcement in India involves 46 steps. Arbitration proceedings constitute only a fraction of those steps.

This award is symptomatic of what is wrong with squeezing the balloon in one place. We just create incentives for parties who want to litigate and delay to move their delaying tactics to other areas including:

  1. Appointment of arbitrators: When parties disagree whether an arbitration is required or who should be an arbitrator, the courts have to step in to start arbitration proceedings or appoint arbitrators. Parties unwilling to cooperate, will just use the same old delaying tactics in Indian courts to delay the appointment of arbitrators.
  2. Execution of arbitration awards: After winning an arbitration, the winning party still has to go to the court to force an unwilling losing party to pay up. Only a court order can block and transfer money out of a bank account or hold auction for a property of the loser. Again, this requires the winning party to go file an application before the court to get court official to assist in forcible takeover of properties, or get bank account records changed (usually called execution proceedings). The losing party can again use time tested delaying tactics in execution proceedings to lengthen out the suffering of the winning party.

Emphasising a single bad metric may have many bad unintended consequences. Arbitration proceedings should not be judged solely on the basis of time taken for the award. The quality of the award is also an important desirable feature of an arbitration. Arbitrary limits on time and fees work against the quality of the awards.

With the nine month deadline in the mind of arbitrators and a probable reduction of fees, the arbitrator will have the incentive to:

  1. Hurriedly finish arbitrations and push out a low quality award. Which will then be challenged in appeal before courts, thereby burdening the judiciary again.
  2. Take up more number of arbitrations to have the same level of income as before. This will have the same effect of pushing down the time and effort an arbitrator allocates to each case.

A single cost cap for arbitrator fees also ignores the complexity of modern arbitrations. Arbitrations today are not just limited to legal questions, complicated contracts in engineering, construction, high end services require specialist arbitrators with technical knowledge. Capping costs has a high risk of driving out competent and therefore expensive aribtrators outside India.

Reshaping incentives

In order to make progress, we should look deeper. We should understand the incentives of the parties to a litigation and then use policy interventions to modify these incentives. One useful example is from the US: Rule 68 of the Federal Rules of Civil Procedure. This is a more nuanced approach which discourages parties to litigate.

Rule 68: Winner beware

Rule 68 involves civil cases where the plaintiff (the suing party) is seeking monetary damages against the defendant (the party being sued). The rule has the following proposition:

At any time before the trial starts, the defendant can make an offer to the plaintiff to settle the case. Two copies of the offer terms are made. The plaintiff can accept or reject the offer. If the plaintiff accepts the offer, the cost of the trial is eliminated.

If the plaintiff rejects the offer, the judge is informed about the rejection but not the terms of the offer that was rejected (this is kept in a sealed copy with the court). If the plaintiff wins, there can be two scenarios at this point:

  1. The sum awarded in the judgment is higher than the sum offered by the defendant before the trial started.
  2. The sum awarded in the judgment is lower than the sum offered by the defendant before the trial started.

In the second case, the plaintiff has to bear the entire litigation costs incurred by the defendant from the date the offer was made by the defendant. The offer is not seen by the judge before the trial to prevent the judge's final determination from getting coloured by the offer of the defendant. The judge comes to the determination of judgment amount through the independent judicial process.

This rule is an elegant way to reduce litigation. At the beginning of a case, the judge has very little information about the merits of the case: In contrast, the parties know much more, having lived through the dispute. They are also in a better position to understand the true value of their economic loss. However, every plaintiff (who believes she will win) has an incentive to ask for more damages than actually suffered. Conversely, every defendant who knows that he has a weak case still has some incentive in drawing out a litigation, thereby delaying the eventual payout she has to make.

When an offer is made to settle, every plaintiff takes it as a signal about what the defendant thinks about the merits of her case. A high offer is interpreted by the plaintiff as that the defendant considers that the plaintiff is on strong legal grounds to win. This may push the plaintiff to continue with the trial, with the hope of getting a higher award in judgment rather than the settlement. However, by transferring the trial costs in case of a lower judgment value, a good counter incentive is created for the plaintiff. The plaintiff has to think hard about the offer and cannot reject it summarily.

Similarly, defendants have an incentive to offer lower settlement amounts because it may be used as a signal that the defendant has a good case. However, this rule gives an incentive to the defendant to make a fair and generous offer, knowing that if the court gives a lower amount the defendant will make significant savings in litigation costs.

The rule thus sets up an economic game where there is a strong incentive for both parties to avoid judicial systems without doing injustice and reducing the burden on the state.


Few problems are as important to India's emergence as a mature market economy and successful liberal democracy, as the problem of making courts work better. One element of this is a fresh approach to the administrative aspects of how courts work. The second element is to rethink rules in a way that is grounded in thinking about incentives. Compare and contrast the sophistication of Rule 68 with the 9 month and price capping rules that we are proposing in our arbitration law.


  1. Rule 68 only deals with incentives of parties to litigate. It does not deal with incentives of intermediaries ( In this case Lawyers). In the current system where lawyers charge per hour or per appearance has economic incentive to ensure that the case does not conclude.

    There shall be law which allows lawyers to charge fees in combination of lump sum and hourly basis but there shall be cap on no of hours which can be billed. Further the fees shall be deposited by the parties in the court and the court will disburse 50 % of the fees in various stages and balance 50% on judgement.

    This will ensures that lawyers from both sides are incentivised for early conclusion of the case.

    Rajan Mehta

    1. There are multiple problems with the Indian legal system and payments to lawyers. However, when analysing a policy problem, one tries to do it keeping everything else the same (cetrus paribus). Even with the same set rules governing the payment for lawyers this rule can create the right nudges.

      I do admit that the legal profession has captured the regulation and fees structure. However, I find large sophisticated parties, have substantial pricing power due to the oversupply of lawyers. If you are a large bank you rarely pay per hour rates and pay lump sump per case, where there is incentive for lawyers to finish cases.

      While champerty (conditional fees or no success no fees) is prohibited in law, it is a common practice in many cities. Large businesses and serial litigators (landlords) are able to get this done. In such situations the incentives are aligned.

      While we see lawyers as the biggest players in the delay, I think it is still the parties interested in delay as the main drivers of the delay. Lawyers usually delay on the instruction of parties. Lawyers are in a repeated game of reputation, people who are unable to "deliver" usually lose market share. Lawyers "deliver" delay usually on client instructions.

      Many individuals are usually single time litigators in India and get taken for a ride. But that is the same in almost all markets (with varying levels). Experienced players are better negotiators. Though this is not an apology for the lack of any minimum standards in the legal services industry.

      What I would like you to think about the suggestion keeping everything same. Glad to write about fees.

  2. Given the level of faith and trust prevailing in judiciary (borne out by many many trials and awards e.g. Salman Khan case, and other bribery scandals) the basic premise of the idea that judge will not see the award is not enforceable. In the absence of this, the losing party will always argue malice towards him/her (and again given the faith in judiciary it will be hard to disregard the insinuation). This idea, with all due respect, is not enforceable where one hears cases of rampant disregard for law. Laws and governance are not created in isolation, they are but an outcome of the level of trust and the machinations in the country. This sounds too naive to work in the context of the complex caste/bribe/"source"/power driven system in India.

    1. There are many ways to solve this part of the problem. In most cases the envelope is not literally handed over to the judge (I wrote that to simplify the problem). It is handed to the registrar/court administration not directly dealing with the judge.
      While you could argue that the judge could influence that, there are physical systems of sealing (three envelopes with signed seals) which worked in the pre-computer era.

      Nowadays, it is a simple task of computer encryption with three keys. One for the judge, one for each of the parties. While I agree court corruption is a problem. That problem exists even in the present system. These rules work irrespective of the court corruption problems in the sense they can improve systems along with present levels of corruption. Of course it will not work as well as in a corruption free system. But I look at corruption as a transmission loss, improvement in generation does not take care of transmission loss but still improves the overall system.

      One must also consider the incentives of the judge in opening the folder. It does not help him dispose of the case faster. It has no information about evidence and the complex procedural systems for establishing it. If one of the parties has exercised "source" to influence a judgment, the envelope is irrelevant. The judge will not bother opening it. The judge may merely pass judgment in favour of the party exercising "source".

      So my proposal does not either contribute or reduce the corruption in the judicial system. While I admit corruption plays some role in judicial delays, I think, it is not as much as one hears in the media. A bigger role is played by poor systems which include anti-corruption measures.

      Let us not use corruption as the excuse for not improving other parts of the system. Even in corruption: our present systems to prevent them are "crude" and there are quite a few elegant systems we could discuss.

  3. Thanks for the detailed reply, Shubho!
    I do appreciate the attempt to solve the problem and agree that it is a problem worth solving. My lack of trust in something like this working is that, in general, mechanistic solutions of locks and keys are going to be inadequate to solve structural problems of trust. Another thing to consider is that when I am assessing my probability of winning the case, I will also take into account my ability to win a judgement in my favor (i.e. the usual "source"/caste equations), the potential loss by admitting defeat, in addition to the merit of the case. If I assign a higher probability of winning by transferring judges (think of political influence), if admitting to a loss results in huge negative externalities (think of political cases of land grab....the case might drag on for 20 years but if I admit defeat today, I have to resign my seat today) those cases are not going to be settled by this. Honor killings, harassment by state etc. are difficult to assign value to so they are suspect too. I do agree that this will solve petty cases where it is easy to figure out the award or the loss. So I think it will work in many cases. So may be we deal with this mechanically till we build trust. Yours is not an enviable job though.... Best of Luck!


Please note: Comments are moderated. Only civilised conversation is permitted on this blog. Criticising me is perfectly okay; uncivilised language is not. I delete any comment which is spam, has personal attacks against anyone, or uses foul language. I delete any comment which does not contribute to the intellectual discussion about the blog article in question.

Please note: LaTeX mathematics works. This means that if you want to say $10 you have to say \$10.