Experimenting with Conditional Fees in the Netherlands: A Law & Economics Analysis PDF

Experimenting with conditional fees in the Netherlands Velthoven, B.C.J. van; Wijck, P.W. van; Boom, W.H. van Citation Velthoven, B. C. J. van, & Wijck, P. W. van. (2017). Experimenting with conditional fees in the Netherlands. In W. H. van Boom (Ed.), Litigation, Costs, Funding and Behaviour. Impl...

Experimenting with conditional fees in the Netherlands Velthoven, B.C.J. van; Wijck, P.W. van; Boom, W.H. van Citation Velthoven, B. C. J. van, & Wijck, P. W. van. (2017). Experimenting with conditional fees in the Netherlands. In W. H. van Boom (Ed.), Litigation, Costs, Funding and Behaviour. Implications for the Law (pp. 129-154). London and New York: Routledge. doi:10.4324/9781315592701 Version: Not Applicable (or Unknown) License: Leiden University Non-exclusive license Downloaded from: https://hdl.handle.net/1887/47874 Note: To cite this publication please use the final published version (if applicable). EXPERIMENTING WITH CONDITIONAL FEES IN THE NETHERLANDS Ben van Velthoven and Peter van Wijck1 Abstract An experiment allowing lawyers to provide services on a no-win, no-fee basis in personal injury cases started in the Netherlands in 2014. Two conditional fee variants were introduced in addition to the standard hourly fee. This paper investigates the implications of this on access to justice. Based on insights from Law & Economics, this paper presents a simulation model. The model simulation suggests that the experiment will not be successful. 1. Introduction In the Netherlands, until recently, lawyers were not permitted to provide services on a no-win, no-fee basis, nor could the amount of their fee vary in proportion to the outcome of the particular case. The traditional line of argument in support of these restrictions has been one of professional ethics. To be of most value to his client, a lawyer, when representing a client in a case, should be independent, impartial and certainly not influenced by any direct financial interest in the outcome of that case. These restrictions have, however, been under discussion for quite some time now. Conditional and contingent fee arrangements have been introduced and are actual practice in other countries such as the UK and the US, so why not in the Netherlands? Moreover, restrictions apply to lawyers, who are members of the Dutch Bar Association. As such, they have a monopoly over advocacy in court. In personal injury cases, however, legal assistance is frequently provided by claim agents and other non-legal professionals. This is possible because the large majority of personal injury cases is settled out of court. Not bound by the rules of the Bar, claims agents have, over time, obtained a substantial share of the market for legal services in personal injury cases by offering potential clients varying no- win, no-fee arrangements. Due to this loss of business, lawyers have started to complain about this competitive disadvantage. The Dutch Competition Authority agreed with this and stated 1 Both associate professor in Law & Economics at the Leiden Law School. We thank A.K. Abhelakh, R.Th. Bocxe, Th. Kremer, A.M. Mooyman and M. Speelmans for information on the handling of personal injury cases in the Netherlands. That information was an essential input for choosing the key parameters of our simulation model. 1 that the regulation that prohibits no-win, no-fee constitutes a violation of competition law.2 In 2004, the first proposal of the Dutch Bar Association to start a no-win, no-fee experiment was blocked by the Minister of Justice, who called upon the traditional line of argument referred to above. Since then however, government budget problems have led to a sharp rise in court fees and substantial cuts to Legal Aid. A second proposal of the Dutch Bar Association for such an experiment got a more favourable reception because it was thought it might help to improve access to justice for the less well-to-do in society. So, on 1 January 2014, a five-year experiment with conditional fees in personal injury cases commenced. During this, claimants and lawyers may waive the normal hourly fee and instead opt for a no- win, no-fee arrangement. Under this arrangement, the claimant does not have to pay any fee if the case is lost. However in a successful case, the lawyer is entitled to the normal hourly fee plus a pre-specified mark-up to compensate for risk and costs. The mark-up is either 100 or 150 per cent of the hourly fee, depending on whether it is the claimant or the lawyer who bears the risk of additional costs, such as medical expertise and court fees. Further, it has been established that the claimant’s payment to the lawyer cannot exceed a certain percentage: 25 or 35 per cent of the financial result. What consequences can be expected to arise from these systems? More specifically, how does this influence the behaviour of parties involved in a dispute? This type of question has been studied extensively in the Law & Economics literature, where scholars typically develop models providing stylised representations of the relevant process. Such literature, however, does not immediately answer our question. The Dutch experiment with conditional fees takes place in a setting that is characterised by the continental way of cost shifting: the party that loses in court must, in principle, pay for the litigation costs of the successful party. Until now, such a combination of conditional fees and cost shifting has hardly been studied in the Law & Economics literature and if so, only on a theoretical level. In the present paper, we study whether the no-win, no-fee arrangement that is currently made available in the Netherlands really improves access to justice.3 To that end, we built our own model to discover whether the conditional fee arrangement, in comparison to hourly fees, can in fact be beneficial to claimants and their lawyers. Given that, regrettably, hard empirical evidence is not available, a simulation was made based on certain key parameters that were distilled from conversations with stakeholders in Dutch personal injury litigation. Our paper proceeds as follows. Section 2 provides a concise review of the Law & Economics literature on contingent and conditional fee arrangements. Section 3 describes the details of the Dutch experiment for both variants. Section 4 presents the outline of our model, which is then used in Section 5 for comparative simulations of the hourly fee and conditional fee arrangements, as they are currently available in the Netherlands. In Section 6, a reflection is made on the premises of our model simulations. A conclusion is provided in Section 7. 2 For more details see Faure and Philipsen (2010). 3 Van Boom and de Jong (2014) interviewed various players in the personal injury field at the start of the experiment and observed that the expectations were very modest. 2 2. Literature review In the United States, the principal means of funding litigation in personal injury cases is through no-cure, no-pay, with, in case of success, a direct and proportionate relationship between fees and damages (contingent fees). This fee system of damages based arrangements (DBAs) has been studied extensively in the Law & Economics literature. General conclusions, however, cannot easily be drawn, as Gravelle and Waterson (1993, p. 1218) have already noted: “the move to contingent fees will create a complex web of effects because changes in the contract between the plaintiff and his lawyer influence the decisions taken at all stages of the litigation process.” Nevertheless, the following points can be said to have sufficient theoretical and empirical support.4 Firstly, contingent fees tend to enlarge the access to justice for cases that are relatively strong and where the amount at stake is relatively high. If capital markets are imperfect, claimants may be unable to finance hourly fees even when litigation has good prospects of success. A contingent legal fee arrangement is a device to solve this liquidity constraint. Furthermore, if claimants are more risk-averse than lawyers, a contingent fee arrangement can facilitate an increase in net utility as a result of risk sharing. Secondly, bearing part of the risk means that lawyers filter out cases with poorer prospects of success and lower value claims. Overall therefore, it is not clear whether the total number of claims filed will actually be higher under contingent fees. However, the average quality of the claims definitely will be. Whether or not a lawyer is willing to take a case on a contingent fee basis thus provides a strong signal about the quality of the case. Thirdly, lawyers who charge an hourly rate, may feel tempted to work more hours on a case than might be warranted by the (additional) expected award to his client. Contingent fees align the incentives more effectively by giving the lawyer a shared interest in the claim. Hence, contingent fee lawyers will put in less effort on a case than hourly fee lawyers. Fourthly, contingent fees do not create a perfect alignment of incentives, as both the lawyer and the client only get a share of the award. It will usually be to the lawyer’s advantage to obtain a quick settlement even if that settlement is below the expected gain from court proceedings, because this will allow him to minimise his efforts. This tendency is reinforced as most clients lack the necessary information and knowledge needed to assess the recommendations of and actions taken by their lawyers. In consequence, lawyers have considerable control over their clients in the settlement process. There is, nevertheless, an important check on contingent fee lawyers, as Kritzer (2004, p. 178) points out and which “comes from the incentives created by the long-term need to maintain a profitable portfolio of cases: these incentives dictate both delivering outcomes for clients (to facilitate a flow of clients) and properly preparing cases (to get good offers from opposing parties).” Conditional fee arrangements (CFAs) were made lawful in England & Wales in 1995. Their scope was extended in 2000 by the Access to Justice Act, which abolished Legal Aid in most personal injury cases. Under a conditional fee arrangement, the lawyer receives no fee at all if the case is lost, but is paid a normal hourly fee plus a mark-up (the success fee) if the case is won. The unsuccessful claimant is thus protected from his lawyer’s bill. He would 4 Schwartz and Mitchell (1970), Gravelle and Waterson (1993), Helland and Tabarrok (2003), Kritzer (2004), Moorhead (2010). 3 however, still be subject to the defendant’s legal costs due to the English rule of cost shifting. As a result, CFAs are typically combined with after-the-event (ATE) insurance policies, purchased when the claim is initiated, to cover these costs in return for a premium.5 A comparison of the American DBA and English CFA systems must take account of the difference in cost allocation. Emons and Garoupa (2006) and Melamed (2006) study the two systems under the American rule of cost allocation. The conclusion is that contingent fees are more efficient than conditional fees. The conditional fee may solve the contingent fee’s problem of encouraging lawyers to keep their work to a minimum, but it gives rise to other significant issues. The conditional fee creates an incentive for lawyers to work as many hours as possible and simultaneously, an incentive to settle as opposed to going to court. This is because upon settlement the lawyer can expect an unusually high return per hour, due to the mark-up. Alternatively, if the case proceeds to court, the lawyer not only risks losing any reward for future efforts but also the reward for the hours he has invested in the case until then. The more hours the lawyer has worked, the less incentive he has to raise civil proceedings. Hyde (2006) shows that the cost allocation rule has no impact on the extent of legal effort under contingent and conditional fees. However, the comparison of the two fee structures depends critically on who takes the relevant decisions in the course of the litigation process, the client or his lawyer. Conditional fees result in more legal effort if litigation is determined by the lawyer and not by the client. 3. The Dutch experiment, three options On 1 January 2014, a five-year experiment with conditional fees in personal injury cases began. During the experiment claimants and lawyers may opt for no-win, no-fee instead of the normal hourly fee arrangement. The no-win, no-fee arrangement comes in two variants.  Variant 1: The lawyer is entitled to an hourly rate with a mark-up of (at most) 100 per cent if the case is won. His total invoice cannot exceed 25 per cent of the financial result, which is defined as the sum of the damages and any legal costs that are reimbursed by the defendant as a result of cost shifting. If the case is unsuccessful, the lawyer receives no fee at all. However, the claimant bears all the specific costs such as medical expertise and court fees. Moreover, upon losing the case the claimant is responsible for any legal costs that shift from the defendant’s side.  Variant 2: The lawyer pays for all the specific costs such as medical expertise and court fees and for any cost shifting from the defendant’s side upon losing the case. If the case is unsuccessful, the lawyer receives no fee at all. In a successful case, however, the lawyer receives an hourly fee with a mark-up of (at most) 150 per cent. His total invoice cannot exceed 35 per cent of the financial result. Moreover, he is entitled to any specific costs that are reimbursed by the defendant as a result of cost shifting. Thus, claimants have three options to choose from, provided they can find a lawyer to 5 In 2000, to further safeguard access to justice, ATE premiums and success fees were made recoverable from the defendant if the case was won. This weakened the claimants’ incentives to monitor their lawyers’ efforts. Indeed, litigation costs increased by approximately 25 per cent after this policy change, according to Fenn, Grembi and Rickman (2014). This recoverability rule was reversed in 2012. For more details see Peysner (2014). 4 represent them under each of the three fee arrangements. 4. Set-up of the model To study the implications of the conditional fee experiment on access to justice, a simple theoretical model of the process of handling a personal injury dispute was built. That model is at the heart of our computer simulation in Section 5. To be manageable, the model is an abstract representation of the dispute handling processes that can be observed in the real world, concentrating on the key elements. In Section 6, we return to the assumptions underlying our model and reflect on how they may affect our results. 4.1 General structure The structure of our model is illustrated in Figure 1.6 Person A, the alleged injurer, has caused an accident that brought personal injury to person B, the victim. Victim B considers whether it is worthwhile to sue alleged injurer A for damages. From then on, B is the claimant and A the defendant. As a first step, claimant B needs to consult a lawyer, for information and advice and for representation in court. Therefore he investigates whether he can find a lawyer who is willing to take his case on. Secondly, claimant B, duly informed by his lawyer, verifies whether the expected net benefit of legal action is positive. If so, he can credibly inform the defendant that civil proceedings may be brought. Thirdly, defendant A may make an offer to settle, if he wishes to avoid the case being brought to court directly. Finally, it is a matter for claimant B, or his representative, to decide whether the offer made by the defendant is acceptable. Otherwise, the case will end up in court. Figure 1 General structure of the dispute handling process --------------------- Insert about here --------------------- The model proceeds on the basis that both the claimant and the defendant are risk neutral.7 Further, we begin with the assumption that the claimant has sufficient financial means to cover his legal costs, if any, so there is no liquidity constraint. Initially, it is also assumed that the settlement decision is controlled by the claimant. In Section 5.3, an analysis is provided with respect to what would happen, if the settlement decision were to be controlled by the lawyer. The impact of a liquidity constraint is discussed in Section 5.4. 4.2 Hourly fees Firstly, we address hourly fees. Suppose defendant A has caused harm H, H > 0, to claimant B. Claimant B wants to sue defendant A to obtain damages. To clarify the relationships of our 6 The structure of the model builds on classic papers in the field, especially Landes (1971), Gould (1973), Posner (1973) and Shavell (1995). 7 An individual is said to be risk neutral if he is indifferent to a certain amount of money and an uncertain amount with an equal expected value. A risk neutral individual effectively makes decisions comparing the expected values associated with choices. 5 model we proceed by backwards induction. That is, we start with the last step of the dispute handling process and then follow the parties’ decision-making back to the first step. Step 4: Settlement or court proceedings? In the final step, claimant B must choose between going to court and accepting defendant A’s settlement offer, if any. Let us first determine B’s expected net benefit of litigation. Under hourly fees, the lawyer will send an invoice for all the hours, h, spent on the case at the current remuneration rate, w. Hence, the lawyer’s total fee is equal to wh. Moreover, B’s legal costs also encompass any specific costs for medical expertise, court fees and the like, denoted by s. Let B’s subjective probability of success in court be equal to PB, with 0 ≤ PB ≤ 1. If B prevails in court, A has to fully compensate harm H. Under the continental rule of cost shifting, A must also reimburse the legal costs of the claimant.8 If B loses in court, however, he has to pay for the legal costs incurred by defendant A. These are denoted by t.9 Claimant B’s expected net benefit of litigation, LB, is then equal to: LB = –(wh + s) + PB(H + wh + s) − (1−PB)t = PBH − (1−PB)(wh + s + t). (1) As regards the alternative option of a settlement: claimant B will only be interested in accepting a settlement offer from defendant A, if it makes him, at least, as well off as the result he can expect from court proceedings. Let us assume that the costs of bringing about a settlement are a fraction α (0 < α < 1) of the costs of a court action. Then, the claimant’s legal costs in case of a settlement are α(wh + s). Hence, to be acceptable to claimant B, the settlement offer must at least amount to Z, where: Z = LB + α(wh + s) = PBH − (1−PB)(wh + s + t) + α(wh + s). (2) Otherwise, the claimant will decline the offer and wait for the judge to give his final ruling in court. Step 3: Court proceedings or settlement offer? In this step it is up to defendant A to decide whether making an acceptable settlement offer is a good alternative to waiting for the outcome of civil proceedings. Let A's subjective probability of success in court be equal to PA, 0  PA  1. If A prevails, his legal costs, t, will be shifted to B. If A loses, he has to compensate claimant B for his harm, H, and his legal costs, wh + s. Consequently, defendant A’s expected total costs of litigation are given by: LA = t − PAt + (1−PA)(H + wh + s) = (1−PA)(H + wh + s + t). (3) 8 We acknowledge that the Dutch rule of cost shifting is somewhat mixed in character. Court fees and expert witness costs are fully shifted, however lawyer’s costs are subject to quite a detailed tariff schedule. We pass over the specifics of that schedule. 9 There is no need to differentiate these costs in lawyer’s fees and specific costs. Note that the defendants in personal injury cases are almost always represented by liability insurers. 6  A can also try to settle the dispute out of court. In order to be acceptable to the claimant, the settlement offer should at least amount to Z, as derived in (2). Clearly, it is not in A’s interest to pay more than strictly necessary, so the offer will be equal to Z.10 Moreover, A will incur legal costs, αt, in bringing about a settlement. Hence, defendant A’s total costs in case of a settlement are given by: Z + αt = PBH − (1−PB)(wh + s + t) + α(wh + s + t). (4) Comparing A’s expected total costs of litigation (3) and A’s total costs in case of a settlement (4), it follows that A will aim for a settlement if and only if: (PA + PB – 1)(H + wh + s + t) ≤ (1 – α)(wh + s + t). (5) Step 2: Credible threat? In the second step, claimant B must decide whether it is credible for him to bring civil proceedings. This is the case if his expected net benefit upon doing so is positive. From the previous discussion we know that the claimant’s expected net benefit of litigation equals LB; see (1). If the case ends in a settlement, the claimant can expect to receive an amount Z, which after deduction of legal costs also yields a net benefit of LB, see (2). Hence, the claimant will be interested in proceeding to court if LB is positive, that is if: PB > (wh + s + t)/(H + wh + s + t). (6) If so, the claimant has a credible case against the defendant. 11 Step 1: Finding an lawyer That brings us to the first step where the claimant is looking for a legal representative. In the context of hourly fees, finding a lawyer will not be a problem given that a lawyer can send an invoice for all the hours he will spend on the case, in and outside of court, with or without success, at his current remuneration rate, w. Throughout our analysis, we assume that the remuneration rate w (at least) covers the opportunity cost of his effort, v.12 Summary Table 1 presents an overview of the expected net benefits of the parties under an hourly fee arrangement. As the claimant can readily find a lawyer to take on the case, three possible 10 Assuming that defendant A either knows the amount Z from his interaction with claimant B or at least has obtained sufficient information for a reliable guess. 11 Note that Z > 0 whenever B has a credible threat of proceeding against A. For if LB > 0, then a fortiori LB + α(wh + s) > 0. 12 The opportunity cost of a lawyer’s hour is the net return he can earn by putting the hour to its best alternative use. The opportunity cost is zero, if the lawyer, upon declining the present case, should spend his time idly due to lack of assignments. The opportunity cost is w, if the lawyer has sufficient other cases to choose from to earn him an average fee of w per hour. 7 outcomes remain: no threat, settlement and court proceedings. Table 1 Hourly fee arrangement: overview of expected net benefits No threat Settlement Court proceedings Defendant A 0 −Z – αt −(1−PA)(H + wh + s + t) Claimant B 0 Z − α(wh + s) PBH − (1−PB)(wh + s + t) B’s lawyer 0 α(w – v)h (w – v)h where: Z = PBH − (1−PB)(wh + s + t) + α(wh + s) 4.3 Conditional fees, variant 1 We now discuss the first variant of the conditional fee experiment. Step 4: Settlement or court proceedings? In the last step of the dispute handling process, the claimant has to choose to go to court or rather, accept the defendant’s settlement offer, provided there is one. What can be said about B’s expected net benefit of litigation? First, B has to pay for his own specific costs, s. Should the court rule in his favour, B is awarded an amount of damages H and legal cost reimbursements wh + s which the defendant requires to pay. At the same time, he must pay his lawyer’s invoice, which is denoted by JC. If the court finds against B, his lawyer is not entitled to any fee. However, B still has to bear the legal costs, t, that are shifted from the defendant’s side. Hence, the claimant’s expected net benefit of litigation is: LB = –s + PB(H + wh + s – JC) − (1−PB)t = PB(H + wh – JC) − (1−PB)(s + t), (7) Let us consider the lawyer’s invoice in more detail. Under the first variant of the conditional fee experiment, the lawyer receives no fee at all if the case is lost. However, in the case of success, the lawyer can send his client an invoice for the normal hourly fees, wh, plus a mark- up of 100 per cent. The total amount of that invoice cannot exceed 25 per cent of the claimant’s financial result defined as the sum of damages, H, and legal cost reimbursements, wh + s. The lawyer’s invoice after successful court proceedings is thus given by: JC = min{2wh, 0.25(H + wh + s)}. (8) Next, we turn to the settlement. Claimant B will only be interested in accepting a settlement offer if it makes him at least as well off as a he would be due to the judgment in court proceedings, by yielding a net amount of LB; see (7). To reach that goal, the settlement offer should also compensate the claimant for the legal costs in settling the case. These consist of the claimant’s specific costs, αs, and the invoice from his lawyer, JS. Upon a settlement, the lawyer is entitled to the normal hourly fee, w, for the settlement hours, αh, plus a mark-up of 100 per cent. The invoice is however, subject to the condition that the total invoice should not 8 exceed 25 per cent of the claimant’s financial result, as given by the settlement amount Z. Thus, the settlement offer will only be successful if it is at least equal to: Z = LB + αs + JS, where JS = min{2αwh, 0.25Z}. (9)13 Step 3: Settlement or court proceedings? In the third step, defendant A considers the option of settling the case instead of going to court. As before, A’s expected total costs of litigation are equal to: LA = (1−PA)(H + wh + s + t). (10) Alternatively, A might aim for a settlement. He can proceed in that direction by offering the claimant the minimum acceptable amount Z, given by (9). As defendant A also incurs legal costs αt in bringing about a settlement, he will aim for a settlement if and only if Z + αt ≤ LA. Step 2: Credible threat? It follows from the previous discussion that the claimant’s expected net benefit is equal to LB, no matter whether the dispute handling process ends in litigation or a settlement. Hence, the claimant has a credible threat if and only if, LB is positive, that is if: PB > (s + t)/(H + wh + s + t – JC). (11)14 Step 1: Finding an lawyer In order to be able to take his case to court, claimant B has to find a lawyer who is willing to represent him. Thereto, the lawyer’s expected return must be sufficient to cover the opportunity costs of his efforts. If the case ends up in court, the lawyer can send his client an invoice JC if he is successful; see (8). On the other hand, the lawyer cannot charge any fee if the case is lost. Ex ante, the lawyer will only be interested in the case if his expected net return is positive, that is if: PBJC − vh ≥ 0, (12) which boils down to PB ≥ 0.5v/w, as long as 2wh ≤ 0.25(H + wh + s). Note that the lawyer can make such a calculation on the back of an envelope, based on his subjective estimate of the claimant’s probability of success in combination with the amount of harm and the expected legal costs in a case of that particular size and strength. Deciding in this way, is well suited to the lawyer who has no experience with the new remuneration rule, as it has only recently been introduced as an experiment. 13 Solving the interdependency of Z and JS yields: JS = 2αwh, Z = LB + αs +2αwh, if αwh ≤ 1/6 (LB + αs); and Z = 4/3 (LB + αs), JS = 1/3 (LB + αs), if αwh > 1/6 (LB + αs). 14 Note that Z > 0 whenever B has a credible threat. For if LB> 0, then a fortiori Z = LB + αs + JS > 0. 9  However, if the experiment runs longer and lawyers become better acquainted with its impact, they will take into account that cases can also end in a settlement. Over time, they will develop good instinct, if not expertise, to predict which cases will end up in court and which cases will be settled. Decision rule (12) still gives good guidance to determine whether it pays to accept a case that predictably ends up in court. For a case that will end in a settlement, the decision rule will be adjusted. Such a case only pays if the fee JS, as given by (9), is sufficient to cover the opportunity costs αvh of the lawyer’s efforts. Summary Table 2 provides an overview of the expected net benefits to the parties under variant 1 of the conditional fee arrangement. There are four possible outcomes: no lawyer, no threat, settlement and court proceedings. Table 2 Conditional fees, variant 1: Overview of expected net benefits No lawyer / no threat Settlement Court proceedings Defendant A 0 −Z – αt −(1−PA)(H + wh + s + t) Claimant B 0 Z – αs − JS PB(H + wh – JC) − (1−PB)(s + t) B’s lawyer 0 JS – αvh PBJC – vh where: JC = min{2wh, 0.25(H+wh+s)} JS = min{2αwh, 0.25Z} Z = PB(H + wh – JC) − (1−PB)(s + t) + αs + JS 4.4 Conditional fees, variant 2 Next, we address the second variant of the conditional fee experiment. As the general line of the dispute handling process does not change in comparison to the first variant, we need not repeat the four steps in full. However, the expected net returns of both the claimant and his lawyer at court and of settlement need adjustment because of the differences in the remuneration rule. To begin, take the lawyer’s expected net return in the case of court proceedings. Under the second variant of the conditional fee experiment, the lawyer has to bear the risk of the specific costs on the claimant’s side, s. If the case is lost, he gets no fee, however he has to reimburse the defendant’s legal costs, t, as a result of cost shifting. In a successful case, on the other hand, he can send his client an invoice for the normal hourly fees, wh, plus a mark-up of 150 per cent. The total amount of that invoice, JC, cannot surpass 35 per cent of the claimant’s financial result defined as the sum of damages payments, H, and legal cost reimbursements, wh + s. Upon winning the case, moreover, he is entitled to a reimbursement of the specific costs, s, he has paid in advance. In the rules of the experiment, it was further laid down that the claimant’s net benefit upon winning his case should never be negative. A successful claimant should therefore never pay a larger sum to his lawyer, JC + s, than his financial result, H + wh + s, allows. This results in an additional upper limit on the lawyer’s invoice: JC 10 ≤ H + wh. The lawyer’s invoice after successful court proceedings is thus given by: JC = min{2.5wh, 0.35(H + wh + s), H + wh}. (13) Altogether, the lawyer’s expected net return in the context of a court case is given by: −s + PB(JC + s) – (1 – PB)t – vh, (14) It can be concluded that the case only pays if: PB ≥ (vh + s + t)/(JC + s + t). (15) Thereafter, the claimant’s expected net benefit of litigation can be assessed. If the case is won, B obtains a financial result H + wh + s, while his lawyer is entitled to the fee JC, discussed above, plus a reimbursement of the specific costs, s. In the event that the case is lost, the claimant has no legal costs at all as they are all taken care of by his lawyer. The claimant’s expected net benefit of litigation is thus given by: LB = PB(H + wh – JC). (16) Instead of waiting for the outcome of the proceedings in court, A can also try to settle the case by offering a settlement amount Z that would bring claimant B – net of settlement costs – into the same position as the decision of the court. B’s settlement costs consist of the lawyer’s fee and the specific costs, αs, he has to reimburse to his lawyer. In principle, the lawyer can claim a mark-up of 150 per cent on the normal hourly remuneration, w, for the settlement hours, αh. His invoice JS, is however subject to the condition that it should not exceed 35 per cent of the claimant’s financial result, Z. A’s settlement offer is thus equal to: Z = LB + αs + JS, where JS = min{2.5αwh, 0.35Z}. (17)15 As a consequence, B’s expected net benefit of a settlement equals LB. His lawyer’s expected net return for a settlement is given by: JS + αs – α(vh + s). (18) Finally, note that claimant B’s expected net benefit equals LB, no matter whether the dispute handling process ends in litigation or settlement. From (16) and upper limit (13) on the lawyer’s invoice, JC ≤ H + wh, it immediately follows that B always has a credible threat, if PB > 0. 15 Note that Z > 0 whenever B has a credible threat. For if LB> 0, then a fortiori Z = LB + αs + JS > 0. Solving the interdependency of Z and JS yields: JS = 2.5αwh, Z = LB + αs + 2.5αwh, if αwh ≤ 14/65 (LB + αs); and Z = 20/13 (LB + αs), JS = 7/13 (LB + αs), if αwh > 14/65 (LB + αs). 11 Summary Table 3 presents an overview of the expected net benefits of the parties under variant 2 of the conditional fee arrangement. There are four possible outcomes: no lawyer, no threat, settlement and court proceedings. Table 3 Conditional fees, variant 2: Overview of expected net benefits No lawyer / no threat Settlement Court proceedings Defendant A 0 −Z – αt −(1−PA)(H + wh + s + t) Claimant B 0 Z – αs – JS PB(H + wh – JC) B’s lawyer 0 JS + αs – α(vh + s) − s + PB(JC + s) – (1 – PB)t – vh where: JC = min{2.5wh, 0.35(H + wh + s), H+ wh} JS = min{2.5αwh, 0.35Z} Z = PB(H + wh – JC) + αs + JS 5. Model simulation 5.1 Set up Our model covers the key aspects of the litigation process but is, in essence, still very simple. For one thing, it does not cover the choice of the level of effort by the claimant’s lawyer in response to the remuneration rule. The preventative working of the legal system with respect to defendant A’s choice of level of care and amount of activities are also not explicitly modelled. Yet, the model as it now stands, simple as its essence may be, is difficult to fathom by analytical means. There are many parameters that play a role. Additionally, the upper limits in the conditional fee arrangements thwart the linear structure of the model. Instead, we opt here for a computer simulation of the model. That is, the computer is used to calculate the outcomes of the model outlined in the preceding section, inserting numerical values for the key parameters. Based on various pieces of information that we obtained from Dutch publications and from conversations with Dutch lawyers, claims agents and insurers, we take the following figures and ratios for granted:  The lawyer’s costs in the case of a settlement are related to the level of harm by an increasing but concave function: αwh = 1.21×(46.9 + 31.10 H0.5).  The specific costs in the case of a settlement are 10 per cent of the lawyer’s costs: αs = 0.1αwh.  The legal costs of reaching a settlement are 25 per cent of the total legal costs of bringing a case to court and proceeding all the way to a judgment by the court: α = 0.25.  The legal costs on the defendant’s side are 50 per cent of the legal costs on the claimant’s side: t = 0.5(wh + s).  The opportunity rate of a lawyer’s effort equals his remuneration rate: v = w. That is, if the lawyer does not agree to take on the case offered to him, he has a sufficient number of 12  other cases to choose from to earn an average fee of w per hour.16 Other important parameters of our model are the degree of harm, H, and the subjective probabilities of success of the parties involved, PA and PB. In the literature, we found some figures with respect to payments for damages for settled cases.17 The frequency distribution is reported to be: ≤ € 5,000 38.5% € 5,000 0.25(H + wh + s). This results in H < € 1,082,000 (in somewhat rounded figures) after substituting 0.25wh = 1.21×(46.9 + 31.10 H0.5) and s = 0.1wh from the initial settings of our simulation. 25 Note that JC = 0.25(H+wh+s) if the restriction on the lawyer’s invoice is active. From (12) we know that the lawyer is only interested in the case if PBJC ≥ vh. Substituting v = w, s = 0.1wh and 0,25wh = 1.21×(46.9 + 31.10 H0.5) tells us that the lawyer will only take the case if H > € 192,000, even if PB = 1. 15  Secondly, it turns out that the claimant’s expected net benefit of litigation is always positive once he has been able to find a lawyer to represent his case. The probability of success is large enough then.26 The lawyer’s readiness to represent the claimant acts as an important signal about the strength of the case. Thirdly, whether a credible threat will end in a settlement or court proceedings depends on the subjective probabilities of success of the two parties. If PA + PB ≤ 1, there is always a settlement offer that is acceptable to both the claimant and the defendant, as they can save on legal costs. Conditional fees, variant 2 The general line of the results for variant 2 of the conditional fee arrangement is very similar to variant 1 discussed above. See the right column of Figure 2. However, some key numerical findings need adjustment. Most notably: a lawyer will in general be able to cover his opportunity costs if PB ≥ 0.52.27 There is a complicating element, however, as the total amount of his invoice cannot exceed 35 per cent of the claimant’s financial result. This upper limit on his invoice appears to be active for H < € 831,000.28 If so, the probability of success must be correspondingly higher to cover his opportunity costs. As a result, until H = € 71,000, the lawyer will not be interested in the case even if he is certain to win.29 Compensation rates Even more interesting than the dispute handling process as such, at least from the claimant’s perspective, is the (expected) compensation rate. What is the claimant’s net expected return as a percentage of his original harm? Figure 3 presents the results. Figure 3 Compensation rate: claimant-control, myopic lawyers --------------------- Insert about here --------------------- Under each of the three fee arrangements, the compensation rate goes up as the degree of harm increases and as the claimant’s probability of success increases. The reason is straightforward: both of these lead to a growth in the claimant’s expected net benefit of litigation. The stronger the claimant’s initial position is vis-à-vis the defendant, the better his end result will be. The (expected) compensation rate, however, does not vary according to the defendant’s 26 Cf. (11). 27 Neglecting for a moment the upper limit on the invoice, the lawyer will be interested if PB ≥ (vh + s + t)/(2.5wh + s + t). Cf. (15). After substituting v = w, s = 0.1wh and t = 0.5(wh + s) this boils down to PB ≥ 1.65/3.15 = 0.52. 28 Cf. (13). The upper limit is active if 2.5wh > 0.35(H + wh + s). This results in H < € 831,000 after substituting 0.25wh = 1.21×(46.9 + 31.10 H0.5) and s = 0.1wh. 29 Note that JC = 0.35(H+wh+s) if the restriction on the lawyer’s invoice is active. The lawyer is only interested in the case if his expected net return (14) is positive. Substituting v = w, s = 0.1wh and 0,25wh = 1.21×(46.9 + 31.10 H0.5) tells us that the lawyer will only take the case if H > € 71,000, even if PB = 1. 16 probability of success. The latter parameter does affect the defendant’s choice between court proceedings and settlement. Having said that, to try and reach a settlement, the defendant has no need to offer the claimant a more generous net benefit than he can expect to obtain in court. Comparison The stage is now set to make a mutual comparison of the three fee arrangements. It is apparent from Figure 2 that any of the remuneration rules may be acceptable to both the claimant and his lawyer, rather than taking no action at all, provided the degree of harm and the claimant’s probability of success are high enough. However, it is also clear from Figure 2 that the hourly fee arrangement is acceptable to both the claimant and his lawyer for a far greater set of parameter configurations than any of the conditional fee variants. Cases that end up being settled or proceeding to court in column 1 (hourly fees) may end up as no lawyer in column 2 (variant 1) or column 3 (variant 2). In these cases, the claimant would clearly be better off under hourly fees. More specifically, no lawyer will be interested in representing the claimant under conditional fees for a level of harm below € 192,000 in the case of variant 1 and € 71,000 in the case of variant 2. Moreover, Figure 3 shows that the compensation rate under hourly fees is, in most cases, higher and never lower than under conditional fees. Hence, from the point of view of the claimant, the hourly fee arrangement is definitely the better option. As to conditional fees, neither of the two variants leads to a consistently higher compensation rate vis-à-vis the other. For some parameter configurations variant 1 is the better option, for other parameter configurations variant 2 will be preferred by the claimant. 5.3 Simulation results: lawyer-control, myopic lawyers In our literature review we noted that it is critical in the context of conditional fees whether litigation is claimant- or lawyer-controlled. The simulations in the preceding section started from the model as it was laid out before, where the settlement decision was assumed to be claimant-controlled. Let us now address lawyer-controlled decision-making. The major change is in Step 3 of the model and only for conditional fees. Under hourly fees, the lawyer is indifferent as to whether the case settles or goes to court, provided his remuneration rate per hour equals his opportunity cost, as has been assumed. The lawyer’s income does not change if he is writing hours on a new case instead of going to court and writing additional hours for the current case: an hour is an hour. The lawyer is not indifferent, however, when it comes to conditional fees. There may be cases where a settlement might be mutually acceptable to the claimant and the defendant, but where the lawyer might prefer to go to court. Court proceedings provide the scope for obtaining a 100 or 150 per cent mark-up on the additional hours in court, in return for sacrificing his opportunity cost and the certainty of obtaining a mark-up upon settling. Going to court may be worthwhile if the probability of success, PB, is high enough. If the upper limit on the invoice does not apply, the lawyer prefers court proceedings to a settlement if PB > 0.62 under variant 1, and PB > 0.64 under 17 variant 2.30 The defendant might try to raise his settlement offer to avert court proceedings. At the end of the day however, that will not help to change the lawyer’s preference as his conditional fee does not depend on the financial result, but on the number of hours worked. Conditional fees give the lawyer a strong incentive to go to court in respect of cases with relatively high probabilities of success. This can be seen from our simulation results. See the graphs on the right hand side of the bottom row of Figure 4 (lawyer-control) in comparison to the corresponding graphs in Figure 2 (claimant-control). Figures 4 and 2 also indicate that the lawyer’s incentive to raise a court action based on conditional fees decreases with the degree of harm. The reason is that the upper limit on the lawyer’s invoice comes into effect for lower levels of harm,31 making court proceedings less attractive.32 Figure 4 Outcome of dispute handling: lawyer-control, myopic lawyers --------------------- Insert about here --------------------- 5.4 Liquidity constraint While presenting the model in Section 4, we noted in passing that the claimant might face a liquidity constraint upon taking judicial steps. This liquidity constraint is more stringent under an hourly fee arrangement than under each of the two conditional fee variants. Under hourly fees, the claimant faces the worst case scenario (losing his case in court) which means he would have to reimburse his lawyer’s invoice, the specific costs and the defendant’s legal costs as a result of cost shifting, a total sum of wh + s + t. Under conditional fees, on the other hand, the lawyer receives no fee at all if the case is lost in court. Hence, in the worst case scenario, the claimant has only to pay for the specific costs and the defendant’s costs that result from cost shifting under variant 1, a total sum of s + t. As far as variant 2 is concerned, the claim on the claimant’s resources is nil as all legal costs are covered by his lawyer. In principle therefore, conditional fees may provide relief to those claimants with insufficient financial means such that they cannot credibly contemplate raising court proceedings where hourly fees apply. This relief is, however, only very partial as Figures 2 and 4 point out. Lawyers only take cases on a conditional fee basis that have relatively high probabilities of success and very high levels of harm. The selection of cases with high probabilities of success may have a positive connotation, in that claimants, as well as lawyers and courts, are 30 Cf. Tables 2 and 3. Leaving aside the possibility of an upper limit on the invoice, the lawyer will prefer court proceedings over a settlement under variant 1 if PB2wh – vh > 2αwh – αvh, which boils down to PB > (v + 2αw – αv)/2w = 0.62, if w = v and α = 0.25. Similarly, the lawyer will prefer court proceedings over a settlement under variant 2 if PB(2.5wh + s + t) – (vh + s + t) > 2.5αwh – αvh, which boils down to PB > 2.025/3.15 = 0.64, if w = v, α = 0.25, t = 0.5(wh + s), and s = 0.1wh. 31 More specifically: for H below € 1,082,000 under variant 1 and below € 831,000 under variant 2. Cf. Section 5.2. 32 Note that lawyer-control does not affect the claimant’s (expected) compensation rate. For if the lawyer were indeed to be tempted to go to court more often, the claimant can still count on his expected net benefit of litigation, given by (7) for variant 1 and (16) for variant 2. In case of a settlement, the claimant’s net result would not be better, as it is in the defendant’s direct interest not to offer more than strictly necessary. If the lawyer fares better at court proceedings, it is the defendant who bears the (additional) costs. 18 prevented from wasting time and financial means on (too) weak cases. On the other hand, conditional fees do not improve access to justice for the less well-to-do by very much at all if lawyers only take the few cases where a very high degree of harm is present. 5.5 When lawyers become sophisticated In the simulation so far, lawyers were assumed to be myopic. The experiment with conditional fees has only recently started. Moreover, in that rather short period of time, only a few cases have been eligible for a conditional fee arrangement, if our simulation results are correct. Therefore, as yet, lawyers have not had much time and opportunity to gain experience in the new remuneration rules. This can of course change with the lapse of time. If so, lawyers may become well-informed and far-sighted with respect to the impact of the remuneration rule. As a consequence, they may obtain sufficient expertise to predict ex ante, whether a new case will end up in court or will settle. Sophisticated lawyers may then become able to understand whether a case that would not be worth the effort if it were to go to court, might nevertheless be interesting enough to accept if there is a prospect that it could be settled. Bringing a case to a successful settlement generally requires considerably less effort than proceeding in court. So, there is a possibility that a settlement may bring the lawyer a positive expected net return in a case where the net return of litigation would be negative. A sophisticated lawyer would accept such a case, whereas a myopic lawyer would recoil and would be reluctant to take on the risk. Figure 5 Outcome of dispute handling: claimant-control, sophisticated lawyers --------------------- Insert about here --------------------- What consequences can be expected if lawyers become sophisticated rather than myopic? Comparing Figures 2 and 5 can provide an answer to this question. From Figure 2, we know that the level of harm should be at least €192,000 under variant 1 and €71,000 under variant 2 before a myopic lawyer might even accept a case that he is certain to win in court. A sophisticated lawyer, on the contrary, is prepared to handle much smaller claims provided the prediction is that the case will end in a settlement. Our simulation therefore suggests that access to justice for claimants will improve over time under the conditional fee experiment. That still leaves the question as to whether the introduction of conditional fees will eventually improve access to justice when compared to the traditional hourly fees. To answer this, we focus on Figure 5 and compare the results in the middle and right-hand columns with those in the left-hand column. It becomes clear that hourly fees are no longer unequivocally the better option. There are cases where the combination of the degree of harm and the claimant’s subjective probability of success are such that a conditional fee lawyer might accept the case and settle it, while the claimant would not have a credible threat or reasonable prospect of bringing a court case under hourly fees. There are indeed, also cases where an hourly fee lawyer might take the case on and reach a settlement or proceed to a final ruling in 19 court, while conditional fee lawyers would not be interested in representing the claimant in this case. Overall however, the overlap is considerable.33 Most parameter configurations that would give access to justice under conditional fees would also be handled under hourly fees and vice-versa. Our simulation therefore suggests that the conditional fee experiment does not really improve access to justice, not even if lawyers become sophisticated in dealing with conditional fees over the course of time. For a final verdict, however, we need more details on the relative frequency of the occurrence of the various (non-overlapping) parameter configurations. 6. Reflection on the assumptions In our model simulations, we have made a number of assumptions with respect to the cost structure of the dispute handling process. These figures and ratios were based on information we obtained from Dutch publications and conversations with Dutch lawyers, claims agents and insurers. Hence, our simulation results are not founded on hard evidence. However, by changing the assumptions, it is possible to investigate the robustness of our results. We re-ran the model simulations under the assumptions that:  The specific costs are, respectively, 5 and 15 per cent of the lawyer’s costs, instead of the original 10 per cent.  The legal costs on the defendant’s side are, respectively, 25 and 75 per cent of the legal costs on the claimant’s side, instead of the original 50 per cent.  The legal costs of reaching a settlement are, respectively, 12.5 and 37.5 per cent of the total legal costs of bringing a case to a final ruling in court, instead of the original 25 per cent. These changes in the legal costs do, of course, affect the details of the outcomes in our figures above.34 The general line of the comparison of hourly fees vis-à-vis conditional fees, however, remains unaltered. To be sure, these were not the only assumptions made. It was also assumed that all those involved are risk-neutral and that the number of hours to be spent on a case varies with the degree of harm but not with the remuneration rule. Under a conditional fee arrangement a lawyer may have an incentive to spend more hours on a case than under an hourly fee arrangement. The reason is as follows. In case of hourly fees, each extra hour generates an extra hourly fee. Since the hourly fee was assumed to be equal to the lawyer’s opportunity costs, the lawyer’s marginal profit of spending an extra hour on a case is zero. In the case of conditional fees, spending extra hours on a case may yield a positive expected profit. This is caused by the mark-up introduced by conditional fees. If it is 33 A similar remark holds for the compensation rate (not depicted here for reasons of space). For some parameter configurations conditional fees give a better result, for other parameter configurations hourly fees would be the better option. Overall, the differences are rather confined. 34 For instance, if α = 0.375 instead of 0.25 we observe a decrease in “no threat” and an increase in “court proceedings” in the case of hourly fees, and a decrease in “no lawyer” for variant 1 and 2. The reason is straightforward. An increase in α implies, ceteris paribus, a decrease in the overall costs of court proceedings. Hence, the claimant’s expected net benefit of court proceedings goes up, giving him more often a credible threat. 20 true that lawyers are inclined to charge more hours that would strengthen our conclusion that an hourly fee arrangement is the better option for claimants. Risk aversion, on the other hand, will make conditional fees relatively more attractive for claimants vis-à-vis hourly fees, as the risk of having to pay for legal costs upon losing the case is transferred, partly or completely, to the lawyer. In principle, lawyers are better equipped to carry that risk, as their portfolio may contain multiple cases where the mark-up in successful cases can compensate for the income loss in the cases with a negative outcome. Much depends on how lawyers actually handle this risk in practice. We have no information on this issue as yet. 7. Conclusion In this paper, we studied whether the current experiment with conditional fees can be expected to improve access to justice in the Netherlands. To that end, we built a model to describe the interaction between the claimant, the claimant’s lawyer and the defendant in the course of the dispute handling process. As hard empirical evidence is, as yet, unavailable, we put the model through a computer simulation, inserting numerical estimates for the key parameters. Testing the robustness of our results by varying the key parameters within a considerable band did not affect the general line of the conclusions. Firstly, our paper shows that simulation can be a useful method of gauging the working of new judicial policy initiatives when rigorous testing is impossible because of insufficient data. Secondly, it is not that easy to fathom the elements of the two variants of the conditional fee experiment. Although the mark-up of 100 per cent or 150 per cent is clear enough in itself, the details with respect to the specific costs, the cost shifting and the upper limits on the lawyer’s invoice complicate the analysis. If our model has difficulty generating an analytical solution to compare the conditional fee arrangements to hourly fees, how on earth will claimants and lawyers in actual practice be able to do this? The problem is even greater as our model by its very nature is only an abstract representation of reality, disregarding all kinds of details and exceptions. This lack of overview alone may be a good reason for the aforementioned parties to stick to the traditional and well-known remuneration rule of hourly fees. Thirdly, when it comes to comparing conditional and hourly fees, it proves to be important to distinguish between myopic and sophisticated lawyers. At the start of the experiment lawyers will be myopic, in the sense that they lack the information and expertise to give an almost accurate prediction of the eventual outcome of each case. Our model simulation shows that, in the interaction between a claimant and a myopic lawyer, the hourly fee arrangement is acceptable for a much more extensive set of parameter configurations than any of the conditional fee variants. More specifically, the degree of harm should be at least €192,000 under variant 1 and €71,000 under variant 2 before a myopic lawyer might even take on a case that he is certain to win in court. Moreover, the claimant’s expected compensation rate is, in most cases, higher under hourly fees and never lower than under conditional fees. Hence, the hourly fee arrangement is definitely the better option from the 21 point of view of the claimant, as long as lawyers are myopic. That finding needs adjustment once lawyers become sophisticated; in the sense that they become more aware and knowledgeable of the way the new remuneration options work. With the lapse of time, lawyers will get the opportunity to gain experience with the new remuneration rules. For a full and final assessment, we need more details on the frequency distribution of the degree of harm and (the subjective perceptions of) the relative strength of the various cases at hand. However, our simulation results point out that most parameter configurations that would give access to justice under conditional fees would also be covered by hourly fees. Thus, the access to justice might grow over time under conditional fees as lawyers become sophisticated, but it would not improve on the traditional hourly fee system. Fourthly, that conclusion might only change if the conditional fee arrangements were to solve a serious liquidity constraint or risk aversion on the part of the claimants. Note, however, that the Netherlands has a system of subsidised Legal Aid for the less well-to-do covering some 36 per cent of the population, in addition to a well-developed system of legal expenses insurance, covering some 42 per cent (partly overlapping) of households.35 So it remains to be seen whether the liquidity constraint problem occurs frequently enough to create a non-negligible demand for conditional fees. The impact of risk aversion will depend on whether lawyers in actual practice are ready to take over the inherent risk of litigation from their clients. This presupposes that they develop a good business model to handle the increased risk on their side.36 There is not much of a stimulus in that direction, however, as long as the market for conditional fee cases remains very limited, as our other arguments suggest. All in all, we do not see any good reason why the experiment can be expected to be a success. References Berntsen ARM and K Nijman, ‘Normering is de normaalste zaak van de wereld...’ (2014) PIV-Bulletin December 12-13. Combrink-Kuiters L, M van Gammeren-Zoeteweij and SL Peters, Monitor Gesubsidieerde Rechtsbijstand 2014 (WLP, Oisterwijk, 2015). Emons W and N Garoupa, ‘US-style contingent fees and UK-style conditional fees: Agency problems and the supply of legal services’ (2006) 27 Managerial and Decision Economics 379-385. Faure MG and NJ Philipsen ‘Fees for claim settlement in the field of personal injury: Empirical evidence from the Netherlands’ (2010) 1 Journal of European Tort Law 75- 101. Fenn P, V Grembi and N Rickman, “No win, no fee”, cost-shifting and the costs of civil 35 Combrink-Kuiters et al. (2015). 36 Cf. Kritzer (2004) on the modus operandi of contingent fee lawyers in the US, arguing that to be successful, contingent fee lawyers must generate and manage a portfolio of cases, similar to an investment portfolio with its associated risk. In personal injury cases in the Netherlands lawyers generally operate from relatively small offices. Hence, the investment that is needed to create a portfolio may well be too high. 22  litigation: A natural experiment (2014) (BAFFI Center Research Paper Series No. 2014- 152, Università Commerciale Luigi Bocconi). Gould JP, ‘The economics of legal conflicts’ (1973) 2 Journal of Legal Studies 279-300. Gravelle H and M Waterson, ‘No win, no fee: Some economics of contingent legal fees’ (1993) 103 Economic Journal 1205-1220. Helland E and A Tabarok, ‘Contingent fees, settlement delay, and low-quality litigation: Empirical evidence from two datasets’ (2003) 19(2) Journal of Law, Economics and Organization 517-542. Hyde ChE, ‘Conditional versus contingent fees: Litigation expenditure incentives’ (2006) 26 International Review of Law and Economics 180-194. Kritzer HM, ‘Lawyer fees and lawyer behaviour in litigation: What does the empirical literature really say’ (2002) 80 Texas Law Review 1943-1983. Kritzer HM, Risks, Reputations and Rewards, Contingency fee legal practice in the United States (Stanford Law and Politics, Stanford CA 2004). Landes WM, ‘An economic analysis of the courts’ (1971) 14 Journal of Law and Economics 61-107. Melamed P, ‘An alternative to the contingent fee? An assessment of the incentive effects of the English conditional fee arrangement’ (2006) 27(5) Cardozo Law Review 2433-2464. Moorhead R, ‘An American future? Contingency fees, claims explosions and evidence from employment tribunals’ (2010) 73(5) The Modern Law Review 752-784. Peysner J, Access to Justice - A critical analysis of recoverable conditional fees and no-win no-fee funding (Palgrave Macmillan 2014). Posner RA, ‘An economic approach to legal procedure and judicial administration’ (1973) 2 Journal of Legal Studies 399-458. Schwartz ML and DJB Mitchell, ‘An economic analysis of the contingent fee in personal- injury litigation’ (1970) 22 Stanford Law Review 1125-1162. Shavell S, ‘Alternative dispute resolution: an economic analysis’ (1995) 24 Journal of Legal Studies 1-28. van Boom WH and M de Jong, ‘Het experiment resultaatgerelateerde beloning – verwachtingen over werking en doelbereiking’ (2014) Tijdschrift Vergoeding Personenschade 69-76. van Wijck PW and BCJ van Velthoven, ‘An economic analysis of the American and the continental rule for allocating legal costs’ (2000) 9 European Journal of Law and Economics 115-125. 23 Figure 1 General structure of the dispute handling process 24 Figure 2 Outcome of dispute handling: claimant-control, myopic lawyers 25 Figure 3 Compensation rate: claimant-control, myopic lawyers Hourly fees Conditional fees, variant 1 Conditional fees, variant 2 H=2,000 % % % 100 100 100 90 90 90 80 80 80 70 70 70 60 60 60 50 50 50 40 40 40 30 30 30 20 20 20 10 10 10 0 0 0 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 PB 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 PB 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 PB H=10,000 % % % 100 100 100 90 90 90 80 80 80 70 70 70 60 60 60 50 50 50 40 40 40 30 30 30 20 20 20 10 10 10 0 0 0 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 PB 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 PB 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 PB H=50,000 % % % 100 100 100 90 90 90 80 80 80 70 70 70 60 60 60 50 50 50 40 40 40 30 30 30 20 20 20 10 10 10 0 0 0 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 PB 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 PB 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 PB H=250,000 % % % 100 100 100 90 90 90 80 80 80 70 70 70 60 60 60 50 50 50 40 40 40 30 30 30 20 20 20 10 10 10 0 0 0 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 PB 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 PB 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 PB H=1,250,000 % % % 100 100 100 90 90 90 80 80 80 70 70 70 60 60 60 50 50 50 40 40 40 30 30 30 20 20 20 10 10 10 0 0 0 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 PB 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 PB 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 PB 26 Figure 4 Outcome of dispute handling: lawyer-control, myopic lawyers 27 Figure 5 Outcome of dispute handling: claimant-control, sophisticated lawyers 28