As Legal Action readers will be well aware, the prospects of success test is a key element of merits assessment for civil legal aid in England and Wales. The highly detailed test (at Civil Legal Aid (Merits Criteria) Regulations 2013 SI No 104 reg 5, as amended) defines the chances of success in percentage bands, from ‘very good’ with a likelihood of success of 80 per cent or more, via ‘good’ (60–79 per cent), ‘moderate’ (50–59 per cent) and ‘marginal’ (45–49 per cent), to ‘poor’ (under 45 per cent). ‘Poor’ cases will not be funded; nor, unless certain restrictive conditions apply, will marginal ones. Prospects of success features not only as a free-standing test, but also as an element of the proportionality test (reg 8) and the reasonable private paying individual test (reg 7).

The test is ubiquitous in the UK: while in Scotland and Northern Ireland, no such test appears in statute or regulation, it forms an important part of the selection of cases for legal aid. Binding guidance on the application of the Northern Irish statutory ‘reasonableness’ test mandates the application of a prospects of success test (Guidance to the Northern Ireland Legal Services Agency: prospects of success and cost-benefit, Department of Justice Northern Ireland, February 2016), with percentage bands comparable to those in England and Wales. Similarly, the Scottish Legal Aid Board will consider whether a case is likely to succeed when assessing legal aid applications, and non-family applications where the prospects are assessed as ‘fair’ or ‘poor’ (between 1 and 6 on a ten-point scale) are likely to be refused.

From a UK viewpoint, we might be tempted to think of the prospects of success test for civil legal aid as inevitable. This is not, however, the case, and major weaknesses in the theory and practical application of the test should lead to it being seriously questioned and even, arguably, abandoned.

Looking outside the UK, the decisive prospects of success test is far from universal. Sweden, for example, has a similar legislative scenario to that in Northern Ireland and Scotland: there is no statutory prospects of success test, but the issue is considered relevant in interpreting the requirement that it must be reasonable for the state to pay for legal representation (Domstolsverkets handböcker, Rättshjälp, chapter 8, para 8.21Not available to the public.). However, the guidance cautions that if the court that is dealing with the main matter is also assessing legal aid (as is often the case in Sweden), it must be careful not to form a substantive opinion before hearing all the evidence and arguments. Therefore, only where the case is evidently hopeless or of meaningless value should legal aid be refused on grounds of prospects of success (para 8.3).

In Finland and Norway, the chance of success can never be decisive in the assessment of civil legal aid. Norway, like Scotland, Northern Ireland and Sweden, has a statutory merits test that legal aid will not be granted if it is unreasonable for assistance to be paid for from public funds (Rettshjelploven, 1981, § 16(5)). However, the guidance there explicitly states that applications must not be decided purely on the question of whether the party is likely to succeed in the action (SRF-1/2017, para 7.6). Finnish legal aid eligibility criteria do not include any measure of likelihood of success at all, and thus legal aid is, in theory, available for very weak or even hopeless cases (Oikeusapulaki 5.4.2002/257, chapter 1).

There are practical reasons in all three countries that partially explain the different approach. In Sweden, as seen above, the fact that courts assess legal aid makes a significant difference. Meanwhile, a grant of Finnish or Norwegian legal aid does not protect against inter partes costs, and clients may well self-impose a prospects of success test to reduce the risk of having to pay the other side’s costs. Certainly, lawyers in Finland find this to be the case, although, anecdotally, Norwegians are not deterred to the same degree. The effect is nonetheless that litigants with mediocre or even poor prospects of success can, in principle, receive legal aid in these jurisdictions.

The willingness of some jurisdictions to fund weak cases reveals a difference of opinion as to the fundamental purpose of legal aid. Is it primarily a tool for achieving a fair trial or is it an instrument for providing access to court for people without means on the same terms as those who have resources? While these aims are not, of course, mutually exclusive, putting a fair trial at the centre of decisions on legal aid makes prospects of success only marginally relevant, if at all – a fair trial is a right irrespective of the strength of a case. Indeed, it may be that representation (and therefore legal aid) is even more important if a case is borderline or weak, as it may make the difference between failure and success. The application of a prospects of success criterion tends towards a situation where the party with the stronger case will be represented and the one with the weaker position will not – surely not a fair trial. A fair trial-focused legal aid system will therefore be cautious of using prospects of success as a decisive element of merits testing for civil legal aid.

Conversely, a legal aid scheme that attempts primarily to level the playing field between those with means and those without will see a prospects of success test as central. The chance of winning a case is, after all, one of the factors that a reasonable litigant will consider carefully before deciding whether to invest their money in legal representation. Caution should be exercised here, however; prospects of success is not the only measure to which a privately paying litigant will pay attention, and therefore a decisive test does not correctly mirror individual risk-taking behaviour. Rather, as the reasonable private paying individual rule suggests, there are three major elements in such a decision: cost; gain sought; and risk of losing. Particularly in the kinds of cases for which legal aid is available, some potential gains will be so considerable that a reasonable person will proceed with a poor chance of success, even if the cost is high. A person might well invest a significant sum of money in trying to keep their home, even if the chance of success is less than half. A purely commercial approach is insufficient.

The mathematical language of statistics lends an aura of objectivity to the prospects of success test, but what does it really say? ‘Per cent’, of course, literally means ‘in every hundred’. Thus, if a case has a 60 per cent chance of success, this means that if it were possible to find 100 identical cases and follow their progress, 60 of them would win.2Statistics work best on large numbers and we would not expect the figures to work out exactly with only 100 cases, and much less so with a sample of ten, but if we could amass 10,000 cases then we would expect to find 6,000 of them winning.

It seems unlikely that this mathematical meaning is in the minds of practitioners, yet if the ranges of percentage prospects of success are small, only a genuinely mathematical meaning can make sense. It is very hard to explain the difference between 50 per cent and 60 per cent likely to succeed unless the statistical definition is used; both are roughly equal chances of success or failure and both might be thought reasonably likely to succeed. Given the discussion below about accuracy of predictions, it seems highly unlikely that there is in fact a ten per cent difference in how many cases actually succeed, depending on whether they are predicted to have a 50 per cent or 60 per cent chance of success, but this is what is being asserted.

Bearing in mind that the prospects of success test in England and Wales defines one category, ‘marginal’, as indicating only a five per cent range of likelihoods of success, it is clearly anticipated that applicants’ lawyers are able to make very accurate predictions of the outcomes of cases. However, there is very little evidence that, in fact, they can. One rare study of lawyers’ predictive ability in the context of legal aid decisions in England suggested that lawyers are actually very bad at forecasting outcomes (Tamara Goriely, Pieta Das Gupta and Roger Bowles, Breaking the code: the impact of legal aid reforms on general civil litigation, Institute of Advanced Legal Studies, 2001). Predictions were approximately double the actual success rates across the range: of cases predicted as having above an 80 per cent chance of success, only 47 per cent succeeded; of those predicted as having a 60–80 per cent chance of success, only 34 per cent were successful; and of those predicted as 50–60 per cent likely to succeed, 30 per cent were, in the event, successful.

In non-legal aid cases, predictability of outcomes is important in conditional fee arrangements and, in that context, a study (David Higham, ‘Does justice play dice? Can lawyers predict the chances of success in litigation?’, Nottingham Law Journal, vol 12(1), 2003, page 20) concluded that ‘lawyers can predict the prospects of success in litigation in a way that is useful and valid. What they cannot do is a sum in objective probability’ (page 30). An objective probability prediction is precisely what is being asked for in the prospects of success test, so this is a worrying conclusion.

Northern Irish guidance makes a point that will ring true to practitioners: ‘the prospects of success could be said to be “unclear” in almost every case as only limited information is available at the outset of a case and the prospects of success may only become clear shortly before trial’ (Guidance to the Northern Ireland Legal Services Agency: prospects of success and cost-benefit, para 6.2, page 7). Nonetheless, a prediction that differentiates within narrow bands must still be given by the lawyer, at the time of the legal aid application.