Risk Management in Commercial Litigation: IT Solutions?

April 30, 1998

Richard Harrison

Richard Harrison is a partner at Laytons, the national law firm.He works out of both the London and Manchester office and deals with commercialand corporate litigation with an increasing IT bias. Notwithstanding some of thesceptical content of this article, he is an enthusiastic supporter of theincreased role of computers in legal practice. He can be contacted on 0171 8428000 or by e-mail to rmh@laytons.com.

Risk management is a phrase on everyone’s lips at the moment.In order to make money out of conditional fee arrangements, it is accepted thatlawyers and insurers should be in a position to use sophisticated risk appraisalsystems. Computer technology and expert systems, as will be well known to everyreader of this magazine, may provide a way forward.

But as with all computer applications, we must sketch out the issues to bedealt with and what we want to achieve on paper. How therefore can we assessrisk in civil litigation cases? The first task is to identify the type, natureand extent of the risks facing those who engage in civil litigation. We can thendecide how to deal with those risks.

Identifying Risk

I identify and discuss five areas below. They are: legal risk, factual risk,procedural risk, financial risk and counterparty risk. They sometimes overlap intheir source and effect.

Legal risk

The first type of risk is legal risk. Assuming that a certain set of factscan be proved to exist, and that no gloss on those facts is going to affect theresult, the task of the lawyer is to predict how a judge will determine theoutcome by applying relevant legal principles.

At first sight this is an easy task. The starting point for the training ofmost lawyers is of course the university legal problem in which A, B and C getinvolved in a particular concatenation of facts and the question posed is:‘advise A’. The student’s assumed knowledge of relevant areas of the lawwill enable him or her to extract the principles necessary to predict theoutcome. However, it soon becomes clear to the student that even for the mostapparently straightforward legal problem there is no ‘right’ answer becauseeither the law is unclear and awaits determination by the House of Lords or, asis most likely, the facts under consideration are subtly but sufficientlydistinguishable from the relevant precedents.

One of the skills of the academic lawyer is the extraction of relevantprinciple and relevant fact and the application of one to the other. Thefascination of law is in predicting the extent to which variation in the factsunder consideration will give rise to a different legal result. The craft liesin assessing and identifying those particular facts which will affect the resultachieved. So the answer to any legal problem, will always contain an element of‘it depends’.

Another reason why the likely legal outcome of a particular dispute isinherently unpredictable is that the judicial decision-making process involvesat least two subjective processes: ‘reasonableness’ and ‘discretion’.Judges can effectively manipulate their findings of facts to achieve the desiredoutcome but they will characterise that outcome as a decision of law. Further,of course, even the best judges have been overturned on appeal. They did notnegligently get the law wrong at first instance; other judges simply took adifferent view or gave different weight to particular considerations.Ultimately, on any one point, three House of Lords judges can prevail over atotal of seven other highly qualified and conscientious lawyers who haveconsidered the point with great skill and care. What hope then for the poorlawyer deciding how best to advise a client and whether to offer a conditionalfee?.

To put those assumptions bluntly, even the very best legal advice must beexpressed or understood to be subject to the following heavy qualifications:

  • I may be wrong
  • the judge may give particular facts greater or less weight than I would
  • the judge may have his own view about what is reasonable
  • the judge may be overturned on appeal.

Factual risk

But, even so, as every practising lawyer knows, the law itself is easy. Itsapplication is not. As shown above, the content of the law is based on the factsdealt with. The process of identifying and presenting those primary facts formanipulation is itself fraught with uncertainty and therefore risk.

The lawyer advising on the future outcome of a case must assume facts whichmay or may not emerge in Court. He is not in control of those facts and can onlyadvise on the basis of caveats, which are not always appreciated by clients. Toput the assumptions bluntly once again, any sensible legal advice will besubject to the following equally heavy qualifications:

  • the judge may not believe what you say at trial, in the witness box and under cross-examination
  • the judge may prefer what the opposition’s witnesses have to say
  • all potential witnesses for your side may not (a) measure up to what you tell me they will say and (b) remain co-operative and controllable
  • the documents presently available may not tell the full story or further documents may emerge which cast a different light on events.

In summary, the facts which a client tells a lawyer and the version of eventswhich a supportive lawyer wants to believe and put forward to the court tosupport his legal case is not definitive. The legally relevant facts will onlyemerge after a long and apparently artificial process which bears no relation tohow they are discussed in the lawyer’s office. The conference room is not thecourtroom and a client may not be wise to be impressed by a bullish lawyer.

Procedural risk

Any case may eventually get to court. However, in the absence of an effectivefast-track procedure or a judicial case management system which really works(and reforms we are told are imminent), the nature of litigation means thatprocedural frustrations are a fact of life. A determined opposition can set upobstacles and cause delay; the case can get bogged down in drawn out wranglingsover procedural and ‘interlocutory’ issues and the courts’ listing systems(and the availability of the necessary ‘expert’ witnesses) do not serve tomake life easier.

As mentioned, this does not really affect the ultimate ability of a case toget tried at some stage, but it does affect budgeting. All litigation to somedegree or other is an exercise in project management but the increased costs ofdealing with these procedural uncertainties at any particular stage must bebuilt into the risk management programme. A case which is going to be decided insix months may well make the risks identified above bearable; a case which willlast three years will magnify those risks and increase the financial exposure.

An unexpected obstacle may make all the difference to the fundamental issuein all litigation which is cost-effectiveness. The constant question must be: isparticipation in the process going to lead to a net benefit? If it is, carry on;if it isn’t, then stop.

Financial risk

This leads to the next element, which depends on balancing what needs to beinvested with what can be achieved.

Some cases require more expensive experts to provide investigation andreports than others. Some cases have more documents to be examined and morewitnesses to be interviewed than others. It is not often brought out but thereal reason why litigation is so expensive is because it is not a linearprocess. Time must constantly be spent revisiting issues and re-examiningevidence in the light of the emergence of other evidence and in the light ofprocedural developments. The more complex the facts, the more time needs to bespent on their evaluation and re-evaluation.

So there are upfront costs (disbursements) and the costs of lawyers’ time.Lawyers can either charge clients for their time or gamble it on a conditionalfee arrangement. They may even fund the disbursements or the insurance premiumsfor costs (their own or the opposition’s in the event of losing).

Counterparty risk

As a very broad rule, and the exceptions will not be discussed here,litigation is only worthwhile if the opposition will be in a position to meet anaward of damages or costs. The solvency of a counterparty and the accessibilityof its assets must therefore be a source of continuous concern.

Evaluating risk

You can evaluate risk on a qualitative case-by-case basis, judging prospectsof success on the best assessment possible in the light of known facts andtaking into account the matters highlighted above, or it can be done on aquantitative basis. In fact the two may be mixed.

There is no mystery to the techniques of quantitative risk assessment. Theunderwriting decision is based on assuming that, given a sufficiently widesample, what has happened before will happen again and variances will balancethemselves out over time. Trends and changes can be spotted, tracked andincorporated. Statistical analysis and probability theory may well come into itsomewhere.

To apply the process to litigation, all we need is a sufficient number ofsimilar cases (which would be nice!). The more similar cases are to each other,the smaller the sample needed to produce a realistic model for risk assessment.

The lawyer who sits in his office overlooking the High Street and sees threepedestrians trip over a loose paving stone will be able to advise the first onewho knocks on the door with some confidence. By the time he has achieved twosettlements and advertised the fact, the third will be very profitable. If thecircumstances involved and the techniques used can be applied to otherincidents, a conditional fee practice in such cases can emerge. If this lawyercan take on five cases in the same town or 50 round the country, the numberinvolved will cancel out any lack of similarity and generate a profitable bookof such cases.

Given the factual dependence of most commercial cases, the issue ofsimilarity is not so easy.

Similarities can be found in the type of incident (the tripping accident asreferred to above, the car crash, the cargo claim) and the common elements andvariables can be identified and evaluated. Similarities can also be found in thetype of client: if one has a client whose modus operandi is generallypredictable and whose activities give rise to the same kind of claims, at leastone of the areas of factual risk identified above can then be minimised.

There are two methods to consider: the statistical sampling methodology(which is used when one has the benefit of a database of similar cases and themore impressionistic ‘standalone analysis’ methodology. Both are reviewedhere and, for the latter, an example of one particular quantitative tool isgiven.

The Statistical Sampling Methodology

One should first build up an appropriate database of historical information(the ‘institutional memory’) which should use relevant and pertinent datafields. One can then:

  • invoke an appropriate query to generate a sample of relevant similar cases
  • formulate a definition of success
  • apply that to the sample to ascertain whether they were successful or not or whether they reached a particular point on a scale of success
  • identify the characteristics of the facts and conduct of those cases which caused the extent of success or the lack of it
  • evaluate each of those factors and apply a numerical weighting based on your assessment of how it contributed to the outcome
  • establish a scale of numbers and a correlative table of ‘conditional fee suitability’
  • identify the factors and score them for the particular case under consideration
  • review the outcome of the scoring exercise and consider where the new case comes on the conditional fee scale.

Stand-alone Case Analysis Methodology: Sample

Without a statistical sample of similar cases (this will be the most commonsituation in commercial cases and it may well be that in other circumstances thestatistical sampling method is flawed), one must revert to the individual caseassessment technique (or the ‘litigators’ gut feeling’ methodology). Thisinvolves a quantitative assessment technique based on a number of factorsrelevant to any particular matter, without necessarily having the benefit of thesampling exercise referred to above.

This involves bringing in the stand-alone case analysis methodology, a sampleof which is given below.

The technique involves two distinct groups of questions which are applied toany particular fact situation. The first set (‘group A’) deals with the riskareas of fact and law as outlined above. The second set (‘group B’)comprises an objective means of analysing some of the issues arising from theneed for witness and documentary evidence as discussed above. From this, aco-efficient on a scale of 1-10 can be derived and applied to the scale ofresponses to a conditional fee proposal: most cases, it is expected, will liesomewhere between the extremes of ‘don’t touch with a bargepole’ to‘bite their hands’ off’. It is not insignificant that positive reactionscan only be seen to commence at 8, a figure and an outcome which, it seems tome, will occur but rarely.

This may appear relatively simplistic and even, given some of the obviousvariables involved, fairly useless. However, it is possible to envisage havingsuch a methodology, producing exactly the same results, with exactly the samesort of input translated from the the sketchy algorithm given above into anattractive and whizzy 32-bit Windows 98 application, marketed to forwardthinking lawyers as a solution to the Woolf/Irvine revolution and winning awardsin the legal press.

This would be far more impressive and credible. It would be called ‘LitiG8™’for Windows 1.1.1 and it would be marketed to progressive law firms as a sign oftheir overall IT awareness and commitment to change, excellence and quality. Itwould appear in the IT kiosks which Lord Woolf foresees outside courtrooms roundthe country. And very few people would dare to point out that the Emperor, infact, had no clothes.


The above article may convey an impression of cycnicism about the use ofcomputerised systems as well as about the effectiveness of applying riskmanagement techniques to the litigation process. The author recognises that suchan impression may not be usual for the content of this magazine. The author is akeen member of the Society and an enthusiast for the digital future of law. Itis however salutary even for enthusiasts to realise that the future of law isnot entirely digital. To convert a bad or unworkable process into an attractivecomputer application does not produce a good or workable process, it produces adangerous computer application.