Printer Friendly

Coupon settlements: compensation and deterrence.


Defendants--innocent or not--have learned to fear the antitrust class action. By combining the claims of thousands, or even millions, of antitrust plaintiffs into a single suit, the class action poses a serious financial threat for even the wealthiest firms. As with most antitrust disputes, class actions are usually settled following class certification. Alarmed by possible abuses, Congress passed the Class Action Fairness Act (CAFA) of 2005. (1) Among other things, CAFA made it easier for defendants to remove class action suits to federal court. (2) It also addressed possible abuses in coupon settlements. (3) Specifically, prior to CAFA, the face value of coupons issued was used to calculate the attorneys' fee award under a share-of-recovery calculation. To deal with the perception that such settlements often resulted in "cash for the lawyers and (worthless) coupons for the class," CAFA seeks to better align the interests of the class lawyers and the class members by altering the calculation of attorneys' fees when coupons are part of a settlement. Only the value of the coupons actually redeemed may now be used in calculating attorneys' fees under a share-of-recovery calculation. This change may influence settlements in many ways.

In this article, we focus on several economic aspects of coupon settlements and CAFA. In part II, we describe the elements of private antitrust enforcement. Part III examines the economic effects of coupon settlements on the firm. Parts IV and V discuss the provisions of CAFA and the economics of class action litigation. Part VI examines the economic impact of the CAFA coupon rule. Finallly, part VII discusses the effects of coupons on deterrence, and we present concluding remarks in part VIII.


Those who have been injured by an antitrust violation may pursue private remedies under section 4 of the Clayton Act, (4) which provides in relevant part: "any person who shall be injured in his business or property by reason of anything forbidden in the antitrust laws may sue therefor ... and shall recover threefold the damages by him sustained, and the cost of suit, including a reasonable attorney's fee." (5) This, of course, raises many issues of standing and antitrust injury, which we do not address here. (6)

Under some circumstances, the possibly small claims of a large number of antitrust victims can be combined into a class action. This may be necessary for antitrust victims to obtain compensation for their injuries. Suppose, for example, that a price fixing conspiracy in the retail gasoline market resulted in a $0.05 per gallon overcharge. Due to the enormous volume of gasoline sold, the total overcharge would be quite large while the average consumer's injury might be $50-$100, which is too small to pursue individually. The antitrust class action, however, could provide a means of recovery. No matter how just this may seem, before certifying a proposed class, the court must conduct a multipronged inquiry. Rule 23(a) of the Federal Rules of Civil Procedure sets out four prerequisites for class certification that must be satisfied by a designated class representative. These are (1) numerosity, which means that the class is so large that joinder is impractical; (2) commonality, which means that there must be at least one question of law or fact that is common to every member of the class; (3) typicality, which means that the class representative must be a member of the class and that his or her claim is typical of those of the other class members; and (4) adequacy, which requires that the class representative fairly and adequately protect the interests of all class members. (7) For our purposes, we will assume that the requirements of Rule 23(a) have been met. Rule 23(b) sets out four criteria, one of which must be met by the proposed class representative, for class certification. (8) The most common antitrust classes are certified under Rule 23(b)(3), which provides for class certification under a predominance standard. (9) Generally, Rule 23(b)(3) requires two things. First, "questions of law or fact common to the members of the class predominate over any questions affecting only individual members." (10) Second, a class action must be "superior to other available methods for the fair and efficient adjudication of the controversy." (11) When these demanding standards are met, a class may be certified.

From an enforcement perspective, class actions are desirable when an antitrust violation has imposed relatively small damages on a large number of victims. Absent a class action, few, if any, will be able to sue for damages. Not only does this leave the victims uncompensated, but it weakens the deterrent effect of private remedies. There is, however, the possibility of abuse due to principal-agent problems. (12) In most class actions, there are very few class representatives and they have a relatively minor incentive to truly represent the interests of the class members. In contrast, the class lawyers have a large economic interest in the case. (13) Nonetheless, because the class lawyers receive only a portion of the recovery, their interests and those of the class members are not perfectly aligned. (14) This imperfect alignment of interests is exacerbated in coupon settlements. The class lawyers must be paid in cash while the class members may receive only coupons, which are actually worth far less than their face value. This undermines the compensatory role of section 4 of the Clayton Act. As we shall see, coupons undermine the deterrent effect of section 4 as well.


Coupon settlements have some objectionable characteristics that CAFA does not remedy. Often, coupons do not provide much, if any, compensation to the class and, as a result, little deterrent effect. This can be seen in the context of a simple model of the firm.

Suppose that a cartel has conspired to fix prices at the joint profit maximizing level, i.e., the cartel acts like a multiplant monopoly. In figure 1, consumer demand is given by D and the competitive supply is the industry marginal cost curve (MC). A profit maximizing cartel will sell [Q.sub.1] units of output where marginal revenue (MR) equals marginal cost. The resulting profits (or aggregate overcharges) are equal to ([P.sub.1] - [P.sub.c])[Q.sub.1], where [P.sub.c] is the competitive price, which is equal to MC. A class action suit is filed and subsequently settled for the full overcharge. Instead of paying the class members in cash, however, the defendants issue coupons that provide for rebates on future purchases. For purposes of this illustration, assume that each coupon has a face value (r) equal to the overcharge, i.e., r = [P.sub.1] - [P.sub.c], and that [Q.sub.1] coupons are issued to the class members. We will assume (perhaps heroically) that the class members actually receive the coupons. Thus, we assume away any burdensome proof of purchase or other administrative hurdle to receiving the coupons. These coupons influence the demand function during the redemption period. Specifically, the demand is discontinuous. Over the output range from zero to [Q.sub.1], the demand shifts up by the amount of the rebate. (15) For Q > [Q.sub.1], the demand coincides with the original demand. This can be seen in figure 2. During the coupon redemption period, the industry marginal cost is also discontinuous: for Q [less than or equal to] [Q.sub.1], industry marginal cost is MC + r; for Q > [Q.sub.1], industry marginal cost is MC. The impact on the colluding firms will depend upon their behavior after the case has been settled and the coupons have been issued. There are several cases to consider.


A. Tacit collusion

First, suppose that the firms in the industry have learned the value of output restriction. Following the settlement, they continue to collude albeit tacitly. In that event, the firms will gravitate to the output where marginal revenue (MR) is equal to marginal cost plus the rebate (MC + r). In this case, the marginal revenue is discontinuous due to the discontinuity in the demand function. As we can see in figure 2, however, marginal revenue equals MC + r at an output of [Q.sub.1]. As a result, the price will actually rise during the redemption period to [P.sub.2] even though output remains at [Q.sub.1]. The coupon redemption rate will be 100 percent. (16) Consumers pay a net price of [P.sub.2] - r or [P.sub.1]. The impact on the cartel is hardly punitive. The excess profits during the redemption period are just as high as they were during the overt conspiracy period:

[PI] = ([P.sub.2] - MC - r) [Q.sub.1],

but [P.sub.2] = [P.sub.1] + r and therefore,

[PI] = ([P.sub.1] - MC)[Q.sub.1]

which was the profit during the conspiracy period. The firms earned illegal monopoly profits during the overt conspiracy period. The coupons took none of that away, and therefore, the colluders disgorged none of their ill-gotten gains. Following the redemption period, industry marginal cost will return to MC. If the firms continue to collude tacitly, price will fall back to [P.sub.1], output will continue to be [Q.sub.1], and profits will remain equal to the joint profit maximizing level. (17) The consumers, of course, continue to be overcharged by the difference between [P.sub.1] and [P.sub.c].


B. Competition

Now suppose that the antitrust class action causes the collapse of the cartel and that the firms behave competitively during and after the redemption period. During the redemption period, industry supply will be MC + r for all Q [less than or equal to] [Q.sub.1] and MC for all Q > [Q.sub.1], As in the preceding case, demand also becomes discontinuous. This is depicted in figure 3. As a result, price will fall to [P.sub.c] and so, of course, output will rise to [Q.sub.c]. The net effect will be that consumers will pay the competitive price. Those with coupons will receive a discount (or rebate) equal to r; those without coupons will pay [P.sub.c]. (18) The firms will earn no excess profits during the redemption period for those consumers who were overcharged. The coupons provide a discount off the competitive price during the collusive period. The coupons force the colluders to disgorge the impermissible monopoly profit earned during the collusive period. In effect, consumers recover the overcharges as discounts from the coupon settlement. (19)


C. Alternative face value

In some cases, the face value of the coupons is less than the per unit overcharge: r < [P.sub.1] - [P.sub.c]. This is a clever negotiating coup for the defendants. Just how clever depends upon the competitive behavior of the firms during and after the redemption period. This can be seen in figure 4. The industry marginal cost is MC + r for all Q [less than or equal to] [Q.sub.1], but MC + r does not equal [P.sub.1] because by assumption r is less than [P.sub.1] - [P.sub.c]. For Q > [Q.sub.1], industry marginal cost is MC.


If the firms collude tacitly during the redemption period, the results are precisely the same as those described in part III.A. The industry profits will be the same:

[PI] = ([P.sub.2] - MC)[Q.sub.1]

and the net price will be the same, i.e., consumers will pay [P.sub.2] in cash and surrender a worthless coupon with a face value of r = [P.sub.2] - [P.sub.1]. Nothing changes for the consumers or for the (now tacitly) colluding firms. If, however, the firms compete, the industry supply, which is discontinuous, will equal demand at an output of [Q.sub.c], which is the competitive output. Price, of course, will be equal to the competitive price of [P.sub.c]. Those consumers with coupons pay a net price of [P.sub.c] - r and, therefore, obtain some actual benefit from the coupons. Since r < [P.sub.1] - [P.sub.c], however, the colluding firms retain some of the collusive overcharge: ([P.sub.2] - [P.sub.c] - r)[Q.sub.1]. Those consumers without coupons were not injured by the overcharge (according to the antitrust laws at least). (20) They pay the competitive price during and after the redemption period.


Now, let us turn to a legislative effort to address some congressional concerns with class actions, including coupon settlements. On February 18, 2005, the Class Action Fairness Act was signed by President George W. Bush. (21) CAFA modified various aspects of class action litigation in an effort to protect defendants from runaway state court juries, protect class members by ensuring fair settlement awards, and promote efficiency by consolidating indirect purchaser suits in federal court. In short, the authors of CAFA addressed several apparent problems with the former system. Substantial jury awards in state court may have given the appearance that the defendant was being "hometowned." In other words, local juries favored local plaintiffs at the expense of out of town defendants. At the same time, coupon settlements in antitrust class actions were often perceived as undercompensating class members and overcompensating counsel. Class members frequently received coupons of little or no value rather than cash. (22) With this in mind, CAFA provides two main areas of reform: expansion of federal jurisdiction over certain cases and new provisions for coupon settlements and calculation of attorneys' fees. CAFA also has implications for indirect purchaser class actions.

A. Expanded federal jurisdiction

One of the central features of CAFA is an expansion of federal jurisdiction over cases that meet certain criteria. Specifically,

"[t]he district courts shall have original jurisdiction of any civil action in which the matter in controversy exceeds the sum or value of $5,000,000, exclusive of interest and costs, and is a class action in which (A) any member of a class of plaintiffs is a citizen of a State different from any defendant; (B) any member of a class of plaintiffs is a foreign state or a citizen or subject of a foreign state and any defendant is a citizen of a State; or (C) any member of a class of plaintiffs is a citizen of a State and any defendant is a foreign state or a citizen or subject of a foreign state." (23)

Prior to the passage of CAFA, an action was subject to federal jurisdiction only if it met the requirement of complete diversity. This meant that federal jurisdiction was not applicable to cases in which any of the named class representatives and any of the defendants were from the same state. For example, if one class representative and one defendant were both from the state of Florida, the action could not be moved to federal court on the basis of diversity. This often prevented multistate class actions from being reviewed at a national level. CAFA ensures that if any class member's state differs from any defendant's state, and the monetary criterion is met, then the action is subject to federal jurisdiction. (24)

Applicability of federal jurisdiction also changed with respect to the total amount of damages sought by the class. Previously, federal jurisdiction was only applicable if the value of at least one class representative's claim exceeded $75,000. At such a high threshold, this requirement was rarely met. (25) Under CAFA, the standard for federal jurisdiction is based upon the total amount sought by the class rather than on one individual's claim. These changes ensure that more class actions are heard in federal court, rather than in state courts, where enormous damages have been awarded. State court and federal court juries are drawn locally, however; therefore, the practical significance of this change is an empirical matter. Time will tell whether the damage awards more closely reflect the actual damages suffered.

B. Coupon settlements

With respect to coupon settlements, CAFA has two main provisions. First, the settlement must be reviewed and approved by the court and found to be "fair, adequate, and reasonable." (26) The second component, however, changes the method for calculating attorneys' fees in class action settlements. Prior to CAFA, attorneys' fees were based on the value of the coupons issued to the class. Now, CAFA requires that in a coupon settlement, attorneys' fees must be based on the value of coupons actually redeemed. (27) This change may have a substantial impact on fees depending upon the redemption rate. Suppose that the attorneys' fee based on coupons issued would have been F and that the redemption rate is the fraction r. Under CAFA, the attorneys' fee awarded (A) would be reduced: A = rF. If r were, say, 0.10, A would be only 10 percent of F.

There are two ways that a coupon settlement offer can be presented to the court. (28) First, some combination of cash and coupons may be offered to the class with attorneys' fees to be determined by the court and awarded separately, i.e., in addition to the cash and coupons offered to the class. In this case, the cash component only needs to be large enough to cover the administrative costs associated with issuing the coupons because attorneys' fees will not be deducted from it. A second type of offer involves a combination of cash and coupons from which the attorneys' fees must be paid. In this event, the cash component must be large enough to cover the attorneys' fees in addition to the administrative costs. For ease of exposition, the analysis that follows focuses on the latter type of coupon settlement, but the principles apply to the first type as well.

Settlements may also include components other than coupons, such as injunctive relief. When this is the case, a share-of-recovery calculation is not available since injunctive relief is difficult to monetize. If the settlement does not incorporate coupons (or at least a portion of the settlement has noncoupon components), then the calculation of attorneys' fees not based on the value of coupons is calculated "based upon the amount of time class counsel reasonably expended working on the action." (29) This is referred to as a lodestar calculation and is defined as the product of a reasonable number of hours spent by class counsel during the course of the litigation times a reasonable hourly attorneys' fee, often adjusted for various factors such as risk. Of course, a lodestar calculation could be used in lieu of a share-of-recovery calculation with a coupon settlement. Presumably, class counsel will opt for the calculation which yields the greatest fee award.

Recognizing that many coupons resulting from a settlement are typically unredeemed, CAFA allows the court to approve a settlement that provides for the distribution of some portion of unclaimed coupons to charitable organizations. (30) The value of unredeemed coupons allocated to such an organization cannot be used as a basis for calculating attorneys' fees. This provision makes some sense because CAFA is designed to reward the class lawyers for the benefits that they obtain for the class members rather than for third parties.

C. Indirect purchasers

While the expansion of federal jurisdiction applies to all class actions, class actions involving indirect purchasers in particular will be affected by CAFA. Indirect purchaser class actions were previously confined to state court. Indirect purchasers are precluded from filing a claim for antitrust damages in federal court. Under Illinois Brick. (31) Indirect purchaser claims can only be filed in state court in those states that have Illinois Brick repealer laws or that otherwise recognize indirect purchasers as having standing. Under CAFA, indirect purchaser class actions can be removed to federal court for all pretrial proceedings if they meet the previously specified criteria for federal jurisdiction. (32) These cases can then be settled collectively under the supervision of the federal court. This substantially changes the litigation costs associated with indirect purchaser class actions. Under CAFA, all pretrial proceedings need take place only once, rather than be duplicated in each jurisdiction where a case has been filed. For multistate indirect purchaser class actions, CAFA will result in substantial cost savings for both the defendants and the plaintiffs.


The basic model of litigation reveals why most legal disputes (including antitrust class actions) are resolved through settlement rather than trial. (33) We begin with the basic model and then introduce coupons. (34) We start with the heroic assumption that the class and the defendant agree that monopoly overcharge damages are equal to a sum that we denote as D. (35) In the event that the class wins at trial, the damage award will be trebled to 3D. In addition, the plaintiff class will receive its costs ([C.sub.p]). (36) Victory, however, is never certain. Let p represent the subjective probability (or relative likelihood) that the class attaches to its winning. To the class, the net expected value of the suit can be represented as

E[award] = p(3D + [C.sub.p]) - [C.sub.p],

where E denotes the expected value. (37) In a very real sense, the suit is simply an asset with an expected dollar value equal to the expected net award of p(3D + [C.sub.p]) - [C.sub.p]. The class owns this asset and can "sell" it to the defendant through a settlement. The class presumably will not settle (i.e., sell) for less than the net expected value of the suit. (38) In economic terms, this sum represents its reservation price.

The defendant's exposure is equal to the trebled damages (3D) plus the cost of the suit ([C.sub.p]) plus its own litigation costs ([C.sub.D]). But these sums must be adjusted to reflect the fact that a loss is not certain. In determining whether to settle, the defendant will look at the expected loss of going to trial, which will be less than the exposure. Accordingly, the defendant will not be interested in "buying" the suit from the class unless the defendant values the suit more than the class values it. For the defendant, there is no upside: the value of the suit is the expected loss associated with going to trial, which is what the defendant avoids in a settlement. The defendant's expected loss of going to trial can be represented as

E[loss] = q(3D + [C.sub.p]) + [C.sub.D],

where q is the defendant's subjective assessment of the probability of losing and [C.sub.D] denotes the defendant's legal costs, which he bears whether he wins or loses.

For a settlement to occur, the value of the suit to the defendant must exceed the value to the class, otherwise the defendant will be priced out of the market, i.e., it will not "buy" out the class's ownership interest in the suit. In other words, a settlement will occur if the expected loss to the defendant exceeds the expected award to the class. Mathematically, settlement requires that

E[loss] = > E[award]


q(3D + [C.sub.p]) + [C.sub.D] > p(3D + [C.sub.p]) - [C.sub.p].

Collecting these terms on the left-hand side, we see that a settlement will occur when

(q - p)(3D + [C.sub.p]) + [C.sub.D] + [C.sub.p] > 0.

This will always hold if discovery and legal precedent cause the convergence of the probabilities, i.e., if q = p. When the class's assessment of its chances of winning equals the defendant's assessment of the class's chances of winning, the first term on the left-hand side is zero. Since the sum of the litigation costs must necessarily be positive, the condition for a settlement is satisfied.

To illustrate this point, suppose that the damages (D) are $15 million and the anticipated costs of litigation ([C.sub.p]) are $5 million. (39) If the class believes that its probability of winning is 0.8, then we can find the expected value of the suit by substitution:

E[award] = (0.8)($45 + $5) - $5,

which equals $35 million. Since the class expects to receive $35 million by going to trial, it will not take less than that to avoid a trial.

Now assume that the defendant agrees that the damages are $15 million. Also assume that the defendant believes that the probability of its losing is 0.8 (i.e., q = 0.8) and that its costs of litigation will also be $5 million. Now, the expected loss associated with a trial will be

E[loss] = (0.8)($45 + $5) + $5,

which is $45 million. This sum represents the maximum price that the defendant will pay to buy the class's right to go to trial. Since the class will take anything above $35 million and the defendant will pay anything less than $45 million, we should expect a settlement, which is what will happen.

In this example, we have assumed that p = q = 0.8. As a result, we have

(0.8 - 0.8)($45 + $5) + $5 + $5 > 0.

Since $10 million in litigation costs can be avoided by settling, it is economically efficient to settle. This is the economic explanation for why the vast majority of cases settle.

Settlement talks can break down if the parties are too optimistic about their chances of winning at trial. In the model, optimism is captured by a high value of p for the class and a low value of q for the defendant. There will be no settlement if the class's assessment of p exceeds the defendant's assessment of q by enough to swamp the litigation costs. In other words, if p is sufficiently larger than q, the expected loss to the defendant will be less than the expected award to the class. As a result, the value of the asset (i.e., the suit) is greater for the class than for the defendant and, therefore, the defendant will not be willing to buy it.

Suppose, as above, that 3D + [C.sub.p] equals $50 million and that [C.sub.p] = [C.sub.D] = $5 million, but that p = 0.85 while q = 0.6. Thus, the class believes that the relative likelihood of its winning is 0.85 while the defendant believes that its chances of losing are only 0.6. In this event, the expected value to the class is

E[award] = (0.8)($50) - $5,

which is $37.5 million, while the expected loss to the defendant is

E[loss] = (0.6)($50) + $5,

which is $35 million. Since the maximum amount that the defendant is willing to pay ($35 million) is less than the minimum amount that the class is willing to accept ($37.5 million), no settlement will occur. (40) The beliefs of one of the parties will be vindicated at trial.

As seen in the economic model of litigation, a decision to settle depends on the relative probabilities of success as well as beliefs about damages and costs of litigation. A defendant may choose to settle if he can avoid additional costs associated with a trial or because the probability of losing is high. A decision to settle, therefore, does not mean that the defendant is necessarily guilty. Rather, a defendant may find it in his financial interest to avoid a trial because the cost of settling is less than the expected loss of a trial. Additionally, the size of a settlement depends on the same factors. If both p and q are small, the expected value of litigation for both the plaintiff and the defendant will be low relative to the trebled damages undiscounted by the probability of success. If the suit is pursued, the range of settlement values available will also be low. While class action settlements are often criticized for undercompensating class members, a small settlement does not necessarily signal abuse of the class. For example, suppose that the damages (arguably) are $100 million and litigation costs are $10 million for each side. If the liability case is weak, p and q will be small, say, 10 percent. Then E[award] = (0.10)($300 + $10) - $10, which is $21 million. The E[loss] = (0.10)($300 + $10) + $10, which is $41 million. Thus, a settlement of, say, $31 million is a good deal for the plaintiffs even though it is a far cry from $310 million.


A. Coupon settlements

When settlement talks fail because the parties are too optimistic (i.e., when p is sufficiently greater than q), coupons may come to the rescue, but that may not be a blessing. First, we should recognize that most antitrust class members are largely passive participants in the litigation process. (41) The active participants are the class attorneys, who really make the decisions on behalf of the class. If the class lawyers are satisfied, then a settlement may be reached. The trick is to figure out how to do that when the condition for a settlement does not appear to be satisfied. One way is to appeal to the economic interest of class counsel, which flows from the fees that the court will award. There are two bases for the fee award: (1) the lodestar, which depends upon the hours expended, the risks undertaken by the attorneys, the tasks performed, and so on, and (2) a share of the recovery.

We focus here on the share-of-recovery method, which means that the case is subject to a contingent attorney's fee. If the defendant prevails, the class counsel will get nothing. If the class prevails, class counsel will receive a share of the recovery ([alpha]). For simplicity, we shall assume that a is applied to the trebled damages (3D) and awarded separately. (42) In addition, we ignore other litigation costs (which can be substantial) to reduce notational clutter. As a result, the expected award will be p(3D) + [alpha]p (3D) or

E[A] = (1 + [alpha])p(3D).

The sum that is expected to go to the class members is p(3D) and the sum expected to go to the class counsel is [alpha]p(3D). A proposed settlement will look good to risk neutral class counsel provided that there is enough cash to cover the expected attorneys' fees: [alpha]p(3D). Thus, the defendant must put together a settlement package that will generate a fee award of [alpha]p(3D) in order to settle the case.

Suppose a creative offer were made by the defendant: cash for the lawyers of [alpha]p(3D) and coupons with a nominal value of p(3D) for the class members. If the redemption rate (r) on the coupons is less than 100 percent, this offer may lead to a settlement (assuming that the court approves it). (43) The class lawyers are no worse off financially than they would be by going to trial. As long as the defendant's expected loss is greater than the actual value of the cash-coupon offer, the defendant will want to settle. In other words, if

(1 + [alpha])q(3D) > [alpha](p(3D)) + r(p(3D)),

then the defendant will want to settle for anything less than that amount. The defendant will then be better off than it would be by going to trial. The class members, of course, do not receive p(3D) because not all coupons are redeemed, i.e., r < 1. In the extreme, the coupons could be essentially worthless, the redemption rate would then be zero, and the class members would get nothing. In that event, the class lawyers would get the cash and the defendant would have settled for much less than the class's expected net recovery at trial. In effect, the class lawyers would get 100 percent of the actual recovery.

In our previous numerical example, we assumed that 3D = $45 million, p = 0.85, and q = 0.6. Continuing to ignore the other litigation costs, the expected damage award to the class at trial, therefore, is $38.25 million. If the class lawyers anticipate receiving a fee award equal to 20 percent of the recovery ([alpha] = 0.20), then the class lawyers expect a fee award of (0.20)($38.25 million), which equals $7.65 million. If the defendant is prepared to pay $7.65 million in cash plus coupons worth $38.25 million in nominal terms, the class lawyers' fee will exhaust the cash. The defendant's maximum settlement offer is (1 + [alpha])(q3D) which is equal to (1.2)(0.6)($45) or $32.4 million, because it is less convinced that the plaintiff will win. The real value of the settlement offer will be no more than $32.4 million if the coupon redemption rate is less than 64.7 percent. (44) In that event, the class members will actually receive $24.75 million in benefits rather than the nominal value of $38.25 million. The class counsel then receive a fee equal to 30.9 percent of the real value of the coupons rather than 20 percent. But things could be much worse for the class and much better for the defendant. Suppose that the redemption rate were 20 percent (r = 0.20). In that event, the actual value of the class compensation is only $7.65 million. The class lawyers still get $7.65 million in fees. Thus, the lawyers get 50 percent of the actual value of the coupons. The defendant is happy because the actual cost to settle is only $15.3 million.

This is where CAFA intervenes. When a settlement involves coupons, CAFA requires the court to base its fee award on the actual value of the recovery. Thus, the fee depends on the value of the coupons actually redeemed rather than the nominal value of the coupons issued. For the class lawyers to get a fee award of [alpha]p (3D), the value of the nominal coupons must be increased enough so that the coupons actually redeemed are worth p(3D). Otherwise, the court will not award a fee of [alpha](p (3D)) to the class lawyers and they will not agree to settle because their expected fee is higher if they go to trial and get cash for the class. Due to the coupon rule, therefore, some settlements will not be reached that might have been reached before CAFA was passed.

Suppose that the coupon redemption rate turned out to be 50 percent in our numerical example. As a result, the actual value of the coupons would be half of the nominal value or $19.125 million (redeemed coupons). If the court awarded 20 percent of the true value of the coupons to the class counsel, they would receive only $3.825 million rather than $7.65 million. This, of course, is not inappropriate since the lawyers are being compensated on the basis of the total actual recovery rather than some fanciful figure. When the court uses the redeemed value of the coupons, it heightens the lawyers' sensitivity to these values. This better aligns the interests of the lawyers and the class members. In our view, this is a good thing.

B. Attorney compensation

Depending on the redemption rate, the value of the case to the class attorneys may be lower under CAFA. Redemption rates are typically very low for a variety of reasons. (45) Assume, for example, that $50 million in coupons are issued to the class as part of a settlement. If the expected redemption rate is ten percent, then the coupons are worth $5 million. The attorneys' fees now must be calculated as a portion of the $5 million redeemed, not the $50 million issued to the class. This will have an obvious effect on the willingness of the class lawyers to file suits or agree to settle cases once they are filed.

CAFA also directly impacts the timing and structure of attorneys' fees. As a result of the coupon provision, the collection of attorneys' fees is delayed until the close of the redemption period, when the actual redemption rate can be ascertained. The delay in receiving the fee can have a substantial effect on the value to the attorneys. For example, suppose that the fee awarded is $10 million. If the appropriate interest rate is, say, ten percent, the present value of the fee is $9.09 million if payment is delayed one year. If it is delayed for two years, the present value drops even further to $8.26 million. (46) At the time the settlement is proposed and until the redemption period ends, the attorneys' portion is subject to a great deal of uncertainty. Since the fee award is conditional upon the value of the coupons actually redeemed, the attorneys' fees at the time of settlement are expected values, dependent upon the realized redemption rate.

Restricting attorneys' fees to a portion of the coupons actually redeemed could shift counsel toward using the lodestar method. (47) Because of the uncertainty of the redemption rate as well as the delay in payment, counsel could prefer the lodestar calculation. In addition, it is also possible that class counsel may delay settlements in order to increase the amount of time and resources dedicated to the case, in order to increase their potential fee. (48)

C. Settlement approval

In addition to altering the calculation of attorneys' fees, CAFA changes the court's obligations with respect to settlement approval. CAFA requires the court to approve a proposed settlement only when it has determined that it is "fair, adequate, and reasonable for the class members." (49) Judicial review of proposed settlements should have been conducted on that basis prior to CAFA's passage. To the extent that appropriate review was already being conducted, this provision may not have much effect on class actions or their settlements. Many past coupon settlements, however, ultimately resulted in extremely low redemption rates, little, if any, benefit to class members, and windfall fees to class attorneys. In these circumstances, CAFA may heighten the sensitivity of some courts to this approval obligation. If so, class members may realize some benefit from this provision.


As a general matter, private enforcement will deter antitrust violations if it makes them unprofitable. The use of coupons instead of cash in antitrust class action settlements may impair the deterrent effect of such settlements by reducing the expected cost of committing an antitrust violation. This is easy to see for a risk-neutral defendant. (50)

Suppose that a prospective antitrust violator will earn profits of [[PI].sub.1] without committing a violation. An antitrust violation will increase profits to [[PI].sub.2] if the violation is not successfully challenged. Suppose that the damages (D) are equal to the difference between [[PI].sub.1] and [[PI].sub.2]:

D = [[PI].sub.1] - [[PI].sub.2].

Now, the expected value of the violation is (51)

E [[[PI]] = p[[PI].sub.2] + (1 - p)(-3D),

where p is the probability of getting away with the violation and (1 - p) is the probability of a successful suit. If

[[PI].sub.1] > E [[PI]],

i.e., if the legal profit is larger than the expected profit of an antitrust violation, the violation will not pay on average and, therefore, it will be deterred.

The effect of coupon settlements is to reduce the downside risk by reducing the actual value of D. Analytically, we can see that

[partial derivative]E[[PI]]/[partial derivative]D = (1-p)3<0

since (1 - p) and 3 are both positive. Thus, a decrease in D will increase E [[PI]]. In that event, some violations that had been deterred will not be deterred.


Consumer class actions have led to large verdicts in state courts. To protect defendants from being "hometowned" by state court juries, CAFA makes it easier to remove class action suits to federal court. The actual effect of this change is clearly an empirical matter and the evidence is not in yet. Time will tell whether this CAFA provision actually reduces the frequency of suits or the severity of judgments.

CAFA also apparently protects class members from their own lawyers. It is a common perception that class action lawyers profit from class action suits while class members do not. This is most visible when the class action lawyers receive handsome fees in cash while the class members receive coupons of dubious value. CAFA addresses that concern by requiring the court to base the attorneys' fees on the value of the coupons redeemed rather than the nominal value of the coupons issued. We have argued that this change may have several effects: (1) it may better align the interests of the class members and their lawyers, (2) it may reduce the number of suits filed, which is not necessarily a good thing, and (3) it may delay settlements so the class lawyers can build up time for a lodestar fee award. The net effect, again, is an empirical matter that must await the empirical evidence.

AUTHORS' NOTE: We thank the Limberopoulos Family Trust and the Warrington College of Business Administration at the University of Florida. We also thank our colleague, William Page, for sparking our interest in this subject. We received valuable advice and suggestions from Christopher Leslie, John Lopatka, and Richard Romano, who cannot be blamed for what follows. Some of what follows depends on Roger D. Blair & Christine A. Piette, Coupons and Settlements in Antitrust Class Actions, 20 ANTITRUST 32 (2005).

(1) Class Action Fairness Act, Pub. L. 109-2, 119 Stat. 4 (codified in scattered sections of 28 U.S.C.).

(2) 28 U.S.C. [section] 1453.

(3) 28 U.S.C. [section] 1712.

(4) 15 U.S.C. [section] 15.

(5) 15U.S.C.[section]15.

(6) These issues are examined extensively in PHILLIP E. AREEDA, HERBERT HOVENKAMP & ROGER D. BLAIR, ANTITRUST LAW, [paragraphs] 335-44 (2nd ed. 2000).

(7) Id. at [paragraph] 331. For a compact survey with ample citations, see ABA SECTION OF ANTITRUST LAW, ANTITRUST LAW DEVELOPMENTS 928-41 (5th ed. 2002).

(8) Id. at 941.

(9) Id. at 943.

(10) Id.

(11) Id.


(13) This can become excessive. Recently, for example, a major law firm was indicted for allegedly paying its "clients" to file suits in securities class actions. See Julie Creswell, U.S. Indictment For Big Law Firms in Class Actions, N. Y. TIMES, May 19, 2006, at A1.

(14) For a compact treatment of incentive contracts, see JEFFREY M. PERLOFF, MICROECONOMICS 665-81 (4th ed. 2007).

(15) Suppose a consumer was willing to pay $10 for a product. If she is given a coupon with a $2 face value, then she will be willing to pay $10 plus the coupon for the item. The coupon cannot be used for anything other than to purchase the product in question.

(16) We are assuming that any consumer entitled to a coupon has one and will use it on future purchases. In reality, many coupons go unclaimed or unused due to imperfect information, inconvenience, onerous conditions, and soon.

(17) Even if the collusive profits were returned in cash, the conspiratorial effort would be profitable if tacit collusion, which is lawful, replaces overt collusion, which is unlawful, in the post-litigation period.

(18) For ease of exposition, we have modeled the free entry and exit case, which is why MC is horizontal. In this case, incumbents that have issued coupons worth r = [P.sub.1] - [P.sub.c] could simply exit and not redeem them. We assume for purposes of exposition that this does not occur. In actual markets, the redemption period is a short-run period and many coupons are not redeemed.

(19) In addition, they benefit from the suit because the collusive pricing stops.

(20) Those consumers who are priced out of the market by the colluding firms clearly suffer economic injury, which is the usual deadweight welfare loss. These losses are not cognizable antitrust damages because proof is impractical (if not impossible).

(21) Class Action Fairness Act of 2005, Pub. L. No. 109-2, 119 Stat. 4 (codified in scattered sections of 28 U.S.C.).

(22) Class Action Fairness Act, [section] 2(a)(3)(A), 28 U.S.C. [section] 1711 note.

(23) 28 U.S.C. [section] 1332(d)(2).

(24) There are several exceptions to this rule. See 28 U.S.C. [section] 1332(d)(3) & (4).

(25) Suppose, for example, that there had been a price fixing conspiracy in the Florida gasoline market. If the conspiracy raised the price by $0.05 per gallon, total overcharges would be about $400 million since some eight billion gallons of gasoline are sold in Florida each year. A consumer class action, however, would not have met the monetary threshold since no consumer would have purchased 1.5 million gallons of gasoline.

(26 28 U.S.C. [section] 1712(e).

(27) 28 U.S.C. [section] 1712(a).

(28) No doubt there are myriad variations on these two approaches, but for present purposes, we will focus on these two simple cases.

(29) 28 U.S.C. [section] 1712(b)(1).

(30) 28 U.S.C. [section] 1712(e).

(31) Ill. Brick Co. v. Illinois, 431 U.S. 720 (1977), confined standing to direct purchasers and denied standing to those who purchased from resellers. The only exception is for cost-plus contracts with predetermined quantities. For an analysis, see Herbert Hovenkamp, The Indirect Purchaser Rule and Cost-Plus Sales, 103 HARV. L. REV. 1717 (1990).

(32) 28 U.S.C. [section] 1332(d)(2).

(33) This model can be traced to William M. Landes, An Economic Analysis of the Courts, 14 J. L. & ECON. 61 (1971) and John Gould, The Economics of Legal Conflicts, 2 J. LEGAL STUD. 279 (1973). For a summary of the model's results, see THOMAS J. MICELI, ECONOMICS OF THE LAW 156-80 (1997).

(34) Several articles specifically consider coupons in class action settlements. See Geoffrey P. Miller & Lori S. Singer, Nonpecuniary Class Action Settlements, 60 L. & CONTEMP. PROBS. 98 (1997); Christopher R. Leslie, A Market-Based Approach to Coupon Settlements in Antitrust and Consumer Class Action Litigation, 49 UCLA L. REV. 991 (2002); and Note, In-Kind Class Action Settlements, 109 HARV. L. REV. 810 (1996).

(35) In most antitrust cases, there is considerable dispute regarding the damage estimate. Such disagreements can be incorporated into the model. Doing so, however, introduces complications that are not relevant to the issues that we examine here.

(36) Section 4 of the Clayton Act, 15 U.S.C. [section] 15, provides for treble damages plus the cost of suit including a reasonable attorney's fee.

(37) Expected values are probability weighted averages: each possible outcome is multiplied by the probability of that outcome and then summed. If the class wins the suit it will receive treble damages (3D) plus its costs, and the probability of that is p. If it loses, with probability (1 - p), it receives nothing. The litigation costs ([C.sub.p]) are actually incurred in either event. Thus, the expected outcome is

E[A] = p(3D + [C.sub.p]) + (1 - p)(0) - [C.sub.p],

which is equivalent to the expected value expressed in the text.

(38) This assumes that the class is risk neutral. If the class members were risk averse, they would be willing to settle for something less than the expected net award. In what follows, we assume that both the class and the defendant are risk neutral. On attitudes toward risk and their implications for decisionmaking, see ROGER D. BLAIR & LAWRENCE W. KENNY, MICROECONOMICS WITH BUSINESS APPLICATIONS 385-408 (1987).

(39) These figures as well as the rest of the numbers used in our numerical example are purely hypothetical and are used for illustrative purposes only.

(40) The reasons for the disparity in the subjective probabilities of success are indicative of a discovery failure or of inherent optimism. We do not explore the reasons for the disparity here.

(41) In principle, the named class representative is there to participate on behalf of the entire class. In practice, however, the class representative is apt to be an inexperienced litigant and, therefore, rely heavily on the class attorneys for advice.

(42) In the event of a small judgment, class counsel may request a lodestar fee from the court. Here, we assume that class counsel will opt for a share of the recovery that the court will determine, i.e., a is dictated by the court taking into account effort and risks.

(43) Even when the redemption rate is 100 percent, a settlement may be possible. See part III above.

(44) Actual redemption rates are often far less than 64.7 percent. This critical value is an artifact of our numerical example.

(45) Settlements often consist of "coupons or other awards of little or no value," or "confusing notices are published that prevent class members from being able to fully understand and effectively exercise their rights." Class Action Fairness Act, [section] 2(a)(3), 28 U.S.C. [section] 1711 note.

(46) The present value of a dollar to be received one year from now at an interest rate of ten percent is $1.00/1.10 = $0.909. The present value at the end of two years is [$1.00/1.10).sup.2] = $0.826. This arithmetic can be applied to the hypothetical $10 million fee in the text.

(47) This, of course, depends upon the alternatives. Presumably, the class counsel will ask the court for a lodestar award when it expects this to be larger than a share of the recovery.

(48) See Charles B. Casper, Settlements Under The Class Action Fairness Act, 2 THE ANTITRUST PRACTITIONER 2 (2005). Whether class counsel have an incentive to delay depends upon their opportunity costs. Hours spent on a class action that does not appear to be very lucrative are hours that cannot be spent on other matters.

(49) 28 U.S.C. [section] 1712(e).

(50) Risk neutrality means that only expected values are relevant to decisions; the variance in outcomes is not relevant. A risk-neutral person will view betting $10 on the flip of a fair coin as equivalent to betting $100 or $1,000. Risk aversion and risk-loving attitudes can be incorporated in the analysis, but do not add much insight.

(51) Here, we calculate without regard to litigation costs, which may be considerable.


* Department of Economics, University of Florida.
COPYRIGHT 2006 Sage Publications, Inc.
No portion of this article can be reproduced without the express written permission from the copyright holder.
Copyright 2006 Gale, Cengage Learning. All rights reserved.

Article Details
Printer friendly Cite/link Email Feedback
Title Annotation:Antitrust Damages Symposium
Author:Blair, Roger D.; Piette, Christine A.
Publication:Antitrust Bulletin
Date:Sep 22, 2006
Previous Article:More can be better: contribution and setoffs in antitrust.
Next Article:The future of aftermarkets in systems competition: an overview of the American Antitrust Institute's invitational symposium.

Terms of use | Privacy policy | Copyright © 2020 Farlex, Inc. | Feedback | For webmasters