Necessary costs and expenditure incentives under the English rule

https://doi.org/10.1016/S0144-8188(02)00087-XGet rights and content

Abstract

In jurisdictions employing the English rule, the loser is normally ordered to bear the winner’s costs, but only to the extent that these costs were ‘necessary’ for the attainment of justice. The usual approach to modeling the English rule does not take into account this limitation on fee-shifting, instead assuming that any and all costs can be shifted. We identify the conditions under which legal expenditures always decrease as the definition of necessary costs becomes more restrictive. It is also shown that increased uncertainty over the definition of necessary costs can increase or decrease the expenditure of risk neutral litigants. If litigants are uncertain about the amount of damages, rather than liability, then limiting the award of costs has no effect on expenditure.

Introduction

Much has been written comparing the incentives of litigants to file suit, settle and litigate under the ‘American’ and ‘English’ cost allocation rules (Cooter & Rubinfeld, 1989, Gould, 1973, Landes, 1971, Posner, 1972, Reinganum & Wilde, 1986, Shavell, 1982, Spier, 1994). While most attention has been paid to whether out-of-court settlements are more likely under one rule than the other, our contribution is to the literature examining the effects on litigation expenditure (Braeutigam, Owen, & Panzar, 1984; Hause, 1989, Katz, 1987). A clear understanding of how trial expenditures are affected by fee-shifting is not only important because it is a key determinant of whether an out-of-court settlement is likely to occur, but it is also important in its own right. This is reflected in the widespread concern in many countries that gaining access to the legal system is too costly.

The standard approach to modeling the English rule has been to assume that the winner at trial is able to shift their entire legal costs onto the loser. In contrast, the American rule requires each party to pay their own legal costs. Gong and McAfee (2000) explain the standard intuition regarding the effects of fee-shifting on expenditure as

Since fee-shifting works as a subsidy to winning and a tax on losing, one might reasonably expect fee-shifting to increase the stakes of the trial and thus to encourage greater legal expenditures than under the U.S. system.

By positing a more accurate model of the English rule, we seek to gain a better understanding of the effects of limits on fee-shifting under the English rule.

The above characterization of the English rule is misleading as it is normal for courts in jurisdictions employing the English rule to give judges discretion as to how they ‘award costs’ in civil litigation. The courts can order the loser to pay the full costs incurred by the winner, but such an award of ‘indemnity costs’ is very rare. More commonly, costs are awarded against the unsuccessful party on a partyparty basis, meaning that only those costs that were necessary in the sense of being ‘fair and reasonable having regard for all the circumstances’ can be shifted onto the loser (Cook, 1991). This limiting of awards of costs is usually justified on the grounds that the loser should not be forced to bear excessive expenditure incurred by an imprudent litigant. Rather, they should only bear those costs incurred by a careful litigant. This suggests that the practice of awarding party–party costs makes the English rule more ‘American’ than previously recognized. While this shortcoming of the stylized model of the English rule has been previously noted, we offer a more thorough treatment of the topic than has occurred to date (Bowles, 1987; Gong & McAfee, 2000; Pfennigstorf, 1984).

In the event of billing disputes, the court can request taxation (i.e., scrutiny) of the bill to determine which costs were necessary. This procedure can be used to resolve disputes between a lawyer and his client and also for disputes arising between the winner’s lawyer and the loser in awards of party–party costs. Taxation is essentially a mechanism to control lawyer agency problems (e.g., artificial inflation of costs). While it is an integral part of the English rule, taxation is not the focus here. Rather, we seek understanding of a more fundamental question: how does limiting the extent of costs that can be shifted onto the loser affect litigation expenditures in the absence of lawyer agency problems? This is a legitimate question since, as mentioned above, limits on fee-shifting appear to be motivated purely by concerns of fairness, not agency problems. We assume that lawyers act in accordance with their client’s wishes and that only necessary costs are in fact shifted.

Given the discrepancy between theory and reality, how valid are previous conclusions about the effect of fee-shifting on expenditure? A useful starting point is the observation that if party–party costs are limited to non-discretionary costs, then expenditure incentives are in fact identical under the English and American rules since non-discretionary costs must, by definition, be incurred under both rules. Discretionary costs will be incurred to the same extent under both rules simply because they are borne by the party that incurs them in each case. In practice, however, the distinction between discretionary and non-discretionary costs is difficult to draw, as an award of party–party costs typically allows the shifting of some, but not all, discretionary costs.

Thus, the ‘true’ English rule lies somewhere between the stylized model, wherein unlimited costs can be shifted from winner to loser, and the American rule. The objective of this paper is to determine precisely what it means, in terms of litigation expenditures, to be intermediate between these two rules. If litigant expenditure can be shown to vary monotonically with the extent of fee-shifting, then, using the earlier intuition from Gong and McAfee (2000), the answer is straightforward: expenditure under the English rule will be higher than that induced by the American rule but lower than that predicted by the stylized English rule. Moreover, increasing the stringency of the limit on fee-shifting will always result in lower expenditure levels. One of the main goals of this paper is to identify conditions under which expenditure is indeed monotonic in the extent of fee-shifting.

In addition to the effect discussed by Gong and McAfee (2000), there is an additional way in which limiting the awards of costs to party–party costs can affect incentives. This pertains to the uncertainty that stems from limiting awards of costs to party–party costs—we outline three sources. First, the award of party–party costs takes on two different formulations and it is typically unclear as to which will be applied to a given case. One formulation is that the costs that can be shifted are only those that were necessary for the conduct of the litigation. A classic statement of the principle is by Mallins VC in Smith v Buller:

Costs chargeable under a taxation as between party and party are all that are necessary to enable the adverse party to conduct the litigation, and no more. Any charges merely for conducting litigation more conveniently may be called luxuries, and must be paid by the parties incurring them.1

This formulation draws a distinction between necessary and unnecessary costs, being cited with approval by Develin LJ in Berry v British Transport Commission.2

Another formulation allows costs to be shifted that were necessary and/or proper, where the terms ‘necessary’ and ‘proper’ clearly have distinct meanings. A proper cost is one which was appropriate or suitable in the circumstances, and need not be necessary.3 Uncertainty as to whether proper costs will in fact be allowed will affect litigants’ incentives to engage in certain kinds of expenditures.

A second source of uncertainty arises from the fact that every case is unique, making it hard to infer from other cases what costs the court will consider to be necessary or proper. Thus, litigants will be uncertain as to the outcome of taxation of party–party costs, even if they know which of the above two formulations of the English rule will be used.

The third source of uncertainty is that judges retain discretion to award costs that are greater or less than those that are necessary and/or proper. That is, sometimes formulations other than the two discussed above are applied. For example, courts have awarded costs that are greater than those that are necessary and/or proper in circumstances such as the making of allegations of fraud, knowing them to be false or irrelevant; evidence of particular misconduct that causes loss of time to the court or other parties; the fact that the proceedings were commenced or continued for some ulterior motive, or in willful disregard for known facts or clearly established law; and imprudent refusal of an offer to compromise.4 Courts have awarded lower costs than those which are necessary and/or proper when they have found in favor of one party but they find that this party was also at fault in some significant way.5

Our model assumes that litigation expenditures (probabilistically) determine the outcome of trial, this relationship being described by the legal technology. Risk neutral litigants choose their expenditures in such a way as to maximize their expected payoff from going to trial. We allow litigants’ beliefs about the trial outcome to differ, adopting the traditional ‘divergent beliefs’ approach rather than the ‘asymmetric information’ approach.6 The amount of damages awarded to the plaintiff if they prevail is common knowledge and independent of litigation expenditures.

Our first main result is to show that while the expenditure of the higher spending party increases steadily with the amount of costs that can be shifted (up to the point where the limit is not binding on either litigant) this is not necessarily true for the other party. Monotonicity of both litigant’s expenditures requires sufficient curvature of the legal technology (with respect to expenditure). Only for symmetric legal technologies, in which both litigant’s expenditures are equal, is this curvature requirement not binding. This clarifies the precise conditions under which the standard intuition—that both litigant’s expenditures are monotonic in the extent of fee-shifting—holds. We then use this result to identify conditions under which a more restrictive definition of necessary costs increases welfare, thus providing a normative basis for the limiting of costs. Limiting costs unambiguously increases welfare if the legal technology is symmetric and exhibits homogeneity of degree zero, suggesting that the American rule is welfare maximizing for suits in which litigation expenditures serve primarily to sway an impressionable jury rather than to allow discovery of the underlying facts of the case.

This first result also indicates the conditions under which further limiting of costs necessarily decreases the scope for pre-trial settlement. Ceteris paribus, parties have more incentive to settle when they anticipate the costs of going to trial to be higher, implying that actual expenditure may be lowest when potential trial expenditures are highest. This point has previously been made by Plott (1987) and Hause (1989), although the former suggests that the higher tendency to settle under the English rule may be welfare decreasing since it may interfere with the attainment of justice. Alternatively, it may be welfare increasing if sufficiently large cost savings result. We then briefly consider suits wherein litigant uncertainty centers on the magnitude of the damages awarded, not the liability of the defendant—such a situation arises in bifurcated trials. Limits on fee-shifting are shown to have no impact on expenditure in these suits. This neatly circumscribes the set of suits for which there is a link between fee-shifting and litigation expenditure.

Our second main result is that increased litigant uncertainty about the definition of necessary costs may either increase or decrease the expenditure of risk neutral litigants. It can decrease expenditure if both litigants’ expenditures are already low (relative to expected necessary costs), while increasing expenditure if they are initially high. This suggests that courts have two policy instruments available to control litigation expenditures: the stringency (i.e., inclusiveness) of the definition of necessary costs and the degree of clarity and precision in defining what an award of party–party costs entails (i.e., which costs are necessary).

Our analysis relates most closely to the work of Braeutigam et al. (1984), Hause (1989), and Gong and McAfee (2000). As discussed in Section 3.2, an advantage of our model over Braeutigam et al. (1984) is that we explicitly model the notion of necessary costs while they do not. This means that it is not possible to determine the general effect of fee-shifting on expenditure from their model. While Gong and McAfee (2000) do explicitly model limits on fee-shifting, the complexity of their analysis prevents them from deriving any general results regarding the effect of fee-shifting on expenditure.7 Rather, their emphasis is on assessing the fairness of settlements.

Hause (1989) is the closest in spirit and approach to our paper as it also focuses primarily on expenditure incentives in a model of the divergent beliefs type. There are two important differences, however, both of which make it easier to discern the nature of the relationship between fee-shifting and expenditure in our paper. First, like Gong and McAfee (2000), he models limits on fee-shifting as a cap on the proportion of a litigant’s costs that can be shifted, while we model it as a well-defined quantity of expenditure. Our approach is more natural in the sense that courts are unlikely to increase the amount of costs that they deem necessary simply because litigants engage in higher expenditures—this is contrary to the spirit of the exercise. Second, although his results are essentially consistent with those that we derive in Section 3, it is not clear precisely what primitive conditions are required to support these results. He shows that monotonicity does indeed hold if the ratio of the litigant’s expenditures does not change with the extent of fee-shifting.8 However, this is an endogenous condition since the levels of expenditures are determined in equilibrium. Our approach is to identify primitive conditions on the legal technology that give rise to these same results.

We next describe the model. In Section 3, we analyze the case where litigants are certain of the definition of necessary costs. We then turn to the case of uncertainty in Section 4. An example is presented in Section 5 and concluding comments are offered in Section 6.

Section snippets

General properties

A plaintiff has filed suit and (for exogenous reasons) is unable to reach a pre-trial settlement with the defendant.

Certain necessary costs

Suppose that both litigants are perfectly informed about z, the court’s definition of necessary costs, and this is common knowledge. We use the labels limited and unlimited to distinguish between our model and the traditional formulation of the English rule.

Limited English rule: necessary costs are finite (i.e., z<+∞).

Unlimited English rule: necessary costs are infinite (i.e., z=+∞).

Although there are infinitely many different limited English rules (i.e., choices of z<+∞), many of these rules

Uncertain necessary costs

In reality, the costs a court deems necessary depend importantly upon the details of the particular case at hand. Thus, prior to the conclusion of the trial, litigants are uncertain as to which costs they will be able to shift to the other party if they win. To capture this idea we now generalize the model by assuming that the litigants are symmetrically uninformed of the court’s definition of necessary costs. Both believe z to be distributed according to F(z) on the interval [0,z̄], where the

An example

In order to gain insight into the restrictiveness of Conditions 1 and 2, we now examine a commonly studied legal technology:15μ(xp,xd)=αxpxp+xd,α<1Let μd=μ and μp=μ+β, where β∈(0,1−α). It is useful to think of α as an objective measure of the strength of the plaintiff’s case. In the absence of any divergence in beliefs (i.e., β=0), both litigants

Conclusions

By modeling the notion of necessary costs, we have obtained a more accurate picture of the expenditure incentives of the English rule of cost allocation. A possibly useful extension to the analysis would be to allow for more than one legal input, as this would allow us to determine the conditions under which such limits on fee-shifting can result in inefficiencies due to input mix distortions. Another direction would be to adopt a principal-agent perspective by modeling lawyers, rather than

Acknowledgements

We wish to thank the Federal Attorney-General for funding the Report. The comments of Hugh Neary and participants at seminars at the University of British Columbia and University of Melbourne are also appreciated.

References (22)

  • A. Katz

    Judicial decision making and litigation expenditure

    International Review of Law and Economics

    (1988)
  • L. Bebchuk

    Litigation and settlement under imperfect information

    Rand Journal of Economics

    (1984)
  • R. Bowles

    Settlement range and cost allocation rules: A comment on Avery Katz’s measuring the demand for litigation: Is the English rule really cheaper?

    Journal of Law, Economics, and Organization

    (1987)
  • R. Braeutigam et al.

    An economic analysis of alternative fee-shifting systems

    Law and Contemporary Problems

    (1984)
  • J. Bulow et al.

    Multimarket oligopoly: Strategic substitutes and complements

    Journal of Political Economy

    (1985)
  • Cook, M. J. (1991). Cook on costs: A guide to legal remuneration in civil contentious and non-contentious business....
  • R. Cooter et al.

    Economic analysis of legal disputes and their resolution

    Journal of Economic Literature

    (1989)
  • J. Gong et al.

    Pre-trial negotiation, litigation, and procedural rules

    Law and Contemporary Problems

    (2000)
  • J. Gould

    The economics of legal conflicts

    Journal of Legal Studies

    (1973)
  • J. Hause

    Indemnity, settlement, and litigation, or, I’ll be suing you

    Journal of Legal Studies

    (1989)
  • A. Katz

    Measuring the demand for litigation: Is the English rule really cheaper?

    Journal of Law, Economics, and Organization

    (1987)
  • Cited by (13)

    • Do fee-shifting rules affect plaintiffs’ win rates? A theoretical and empirical analysis

      2021, International Review of Law and Economics
      Citation Excerpt :

      Rent-seeking models of litigation, where the probability of winning the case depends on both parties’ legal expenses, have shown that the English rule might increase those expenses, make litigation costlier, and therefore reduce the number of litigated cases. See, for instance, Brauetigam et al. (1984); Hause (1989); Farmer and Pecorino (1999); Hyde and Williams (2002), and Baye et al. (2005). Additionally, some papers find that this rule can also affect the rate of plaintiff success at trial.

    • Fee shifting and accuracy in adjudication

      2020, International Review of Law and Economics
    • Conditional versus contingent fees: Litigation expenditure incentives

      2006, International Review of Law and Economics
    View all citing articles on Scopus

    The ideas in this paper derived from work that was undertaken for the Australian Federal Attorney-General which resulted in the Report of the Review of Scales of Legal Professional Fees in Federal Jurisdictions, 31 March, 1998.

    View full text