Disclosure in large English litigation: 2 of 2 - Defeat into Victory?

In part 1 of this post, I summarised the history of disclosure reform in England since the 1990s and its failure to make any real impact on the central problem of disproportionate cost.

In this part, I offer some ideas as to how it might be possible to do better this time.

The nature of the problem

Here are some propositions which I think are useful to reflect on in deciding what can or cannot usefully be done. 

  1. Some aspects of the November 2017 proposals – a duty to provide adverse documents which a party knows of and a duty not to provide documents of no relevance – are obviously potential sources of additional cost. Whether these risks will materialise in a significant way and, if so, whether the benefits will outweigh them is something on which reasonable people can have different opinions.
  2. The working group has not sought to establish a norm on the critical question of whether the future should be one of limited disclosure in the arbitration / model C sense or something more extensive in the traditional litigation / model D sense. This is evidently too controversial to be something that could be resolved in a working group comprised essentially of lawyers and, presumably, operating by consensus or something close.
  3. Success or failure on the central problem – cost / value-for-money – will therefore come down to what is done in practice by judges under what is, despite its great length, a very open-ended practice direction.
  4. One possibility is that each judge will follow his or her own path and that the degree of disclosure you will get will depend on that judge’s instincts. 
  5. Another possibility is that judicial opinion will evolve towards a common position. That seems likely to be a slow and uncertain process unless attended with some objective methods and transparency as to their success or failure.
  6. It’s realistic to expect litigators to advance their clients’ interests, not to lie and not to do work which they know to be valueless.
  7. It’s unrealistic to expect them to cooperate beyond a certain point with counterparties in hostile litigation. General acceptance that the system is out of control tends, in the heat of battle, to be outweighed by considerations of how to achieve the best outcome in the case in hand.
  8. Reasonable cooperation in complex cases is in the eye of the beholder to a significant extent.
  9. The strategies of the last twenty years - exhorting cooperation and punishing people for clear non-cooperation - are crude tools. Not much can be expected from them.
  10. Even when everyone’s on the same side, it can be hard to assess costs and value of a large task in a complex legal project.
  11. In disclosure, it’s much tougher than usual because the crucial interests are not aligned. The party seeking disclosure is incentivised to talk down the cost and talk up the value. Vice versa for the party resisting disclosure.
  12. The winner of a disclosure order is further incentivised to criticise the other party’s efficiency in complying with it, in order to reduce the amount of costs which the winner of the disclosure order will have to pay if it loses the overall case.
  13. Lastly, and this needs saying, there are indications that lawyers as a class tend to be more risk-averse and evidence-resistant than business people when it comes to trying new ways, accepting that there will be failures but welcoming that as a chance to learn. On this topic, chapter 2 of Bruce McEwen's book, Tomorrowland (2017) is worth reading.

I state all these points so as to be clear why this problem is so difficult.

Breaking out

In the technology start-up world, a recognised approach for making progress is as follows:

  • Build a minimum viable product first. This has been defined as follows:
"The minimum viable product is that version of a new product a team uses to collect the maximum amount of validated learning about customers with the least effort."
  • "Validated learning" means that you try out ideas against users, measure the impact as objectively as possible, analyse the results, improve and try again. This is known as the build-measure-learn cycle. Doing it effectively and quickly helps. A lot.

Search online for, e.g. "Lean Startup" / "Eric Ries" if you want more detail.

What's this got to do with disclosure reform?

Rather than continuing to argue about what will or won’t work and laying down complicated bodies of the usual lawyer’s abstractions (rules, exceptions, principles, presumptions…), I suggest that the way to make progress is to run imaginative experiments, measure them dispassionately against relevant goals (justice, costs...), learn the lessons, adjust and try again.

What does measurement involve?

Measurement needs data.  In the disclosure context, this might involve the following:

  • Capturing what’s been ordered, and relevant features of the case, as data, in a simple, anonymised, structured way. This is an art in itself.
  • Gather costs information and qualitative evidence from as many people as possible who’ve touched the process. This would include the judge, the lawyers (senior to junior), the client (non-lawyers as well as legal department) and anyone else of relevance.
  • A really light online survey to capture feedback frequently from each participant might have a part to play here. Design is crucial, with a focus on lightness not detail.
  • Over time, the data can be analysed by people with suitable skills. That means involving data scientists and others with product skills, not just lawyers and judges.

What would an experiment involve?

As well as the familiar variants widely seen in litigation and arbitration, I would suggest that some more radical things could be tried.

The following three experiments are just examples of the sorts of the thing that can be tried and assessed with a suitably scientific, evidence-based mindset.

These might not be the right experiments to try, and they and other experiments might well fail, but the important thing is to accept some risk without being reckless, and to assess the results dispassionately.

Experiment 1: sampling

In some cases, a sample approach may be adequate, particularly where large numbers of transactions are involved. Auditors have taken a sampling approach for years.

An example of a recent order to this effect is the one made in Bank Mellat v HM Treasury [2017] EWHC 2409 (Comm), discussed here by the law firm which applied for it.

In that case, the transactions were selected randomly. With good data and an appropriate mindset one might build on whatever the outcome is of this experiment, e.g. using machine learning tactics to build up the sample rather than sampling randomly.

Experiment 2: disclosure facilitator

Rather than leaving parties' lawyers to propose things and seek judicial comment, in a system where the judge is unlikely to have much time to engage deeply, an experiment could be to create a new neutral role which can facilitate discussion of different approaches to disclosure between the parties. A bit like a mediator but procedural. 

Experiment 3: neutral document review

Assume that two parties each have significant disclosure requests to make of each other, and that the case is large and complex enough that it's foreseeable that one thing will lead to another, with multiple disclosure requests being made throughout the case.

Here's one way this could be handled

  1. The parties each put together databases with an agreed set of content, defined in ways which don’t require expensive human effort e.g. named individuals’ emails for a given date range.
  2. These large databases are then made available to an agreed neutral third party with expertise in reviewing documents.
  3. Each party’s requests are then cost-estimated by the third party.
  4. There may then be a bit of to and fro between the party making the request and the third party, exploring ideas for less costly options for doing the relevant searches and reviews.
  5. The party making the request can then pay the third party for such elements of the work as the requester values sufficiently to pay for. If the court, or a party, sets a budget then the requester can prioritise accordingly. Freedom of choice by skilled people with an interest in success, but subject to reasonable resource constraints, is how the best business decisions are made.
  6. With this approach, interests are fundamentally aligned – the party estimating cost is incentivised to make it low enough to get the job and high enough to make a margin, the party making a request has to pay a reasonable price for it up front and can thus be expected to weigh the value.
  7. There are numerous challenges and details to be figured out here, e.g. the payments up front to be made to fund the estimation process itself, the degree of risk which is acceptable on privilege and how to mitigate that risk. I can't immediately think of anything that makes it impracticable to solve these, though.

The point

The point of this post is not to promote these three ideas for experiments but to suggest that a different mindset is needed if we're not just to see the same old pattern - lots of fine words, no real impact on the problems of cost, waste and injustice.

Can litigious-judicial culture cope with something like this?

There's only one way to find out.

Graeme Johnston