September 16, 2005
Conglomerate Junior Scholars Workshop: Gamage & Kedem on the Consideration Paradox
Posted by Christine Hurt

Welcome back to the final installment of Conglomerate Junior Scholars Workshop!  Our last (but not least) featured paper is Resolving the Paradox of the Consideration Doctrine by David Gamage and Allon Kedem.  Both authors are recent law school graduates.  Allon is currently in a federal clerkship, and David is a visiting assistant professor at the University of Texas School of Law in the Emerging Scholars program.

Our expert reviewer for this paper is Larry Garvin, of the Moritz College of Law at Ohio State University.  Prof. Garvin writes in the areas of contract, sales, and behavioral law and economics.  He also has a fortchoming textbook from Aspen with Gordon on entrepreneurial finance.  Prof. Garvin's comments appear above and below the fold:

In Resolving the Paradox of the Consideration Doctrine, David Gamage and Allon Kedem set themselves one of the persistent problems in the common law of contract: why the law of consideration insists on bargain, on the one hand, but does not police the adequacy of consideration, on the other. How, then, can we justify consideration doctrine in light of our assumption that contract law seeks to give effect to the intentions of the promisor at the time of promise? Surely parties intending to be bound should then be bound, whether the promise is based on bargain or gift.

Gamage and Kedem start by showing the weaknesses in the traditional explanations and the significant incoherence of current doctrine. They then suggest that we should abandon our premise that effecting the intent of the parties to be bound is a laudable object. Rather, they propose that we ask whether the parties wanted the option to be legally bound at all. They base this less-than-intuitive shift on option theory, particularly the work of Aghion and Hermalin. In a world of welfare maximizers with asymmetric information about the likelihood of breach, and contracts consisting only of two terms – the size of the promise and the level of sanctions for breaching – we will see a “signaling spiral.” Parties with above-average reliability will offer proof of their reliability. This lowers the perceived (and actual) value of promises from the remaining promisors. Those above the new average will likewise seek to signal their relative reliability, and so forth. Of course, those with below-average reliability will try to mimic their more reliable colleagues. As the authors demonstrate formally, we would see either a pooled equilibrium at an inefficient level of sanctions or a separating equilibrium, with unreliable promisors offering excessively large promises and reliable ones offering excessively large sanctions for breach. Manipulating the damages measure and the other constants in the model can affect whether allowing the parties to be bound enhances welfare or not, from which the authors conclude that the models “can only disprove the dominant wisdom that parties who take advantage of a legally binding option necessarily desire the existence of that option.” But the data allowing the efficient result to be calculated do not, and probably cannot, exist.

How, then, to separate binding from non-binding promises? Gamage and Kedem propose that loosening the assumption that contracts contain only two terms may let us do so. Suppose contracts also may define the scope of a promise and its level of return payments. Then we may see a way out. Some types of promises resist commodification, making bargains over these terms “socially awkward.” Looking at a wide range of scholarship in anthropology, sociology, philosophy, and economics, they conclude that social norms barring this sort of unseemly behavior are strongest in many gift relations, particularly those centered in trust-building and status-enhancing. In contrast, exchange relations allow for this sort of bargaining. So, less intuitively, do altruistic promises, though resort to interdependent utility comparisons yields the same result. Where social norms inhibit the use of consideration, the authors propose a default rule of unenforceability. They recognize, though, that they have no way to determine whether the parties to these promises would prefer the option of enforceability, so they also propose some tie-breaking factors – whether the welfare of society would rise were the law involved, whether the obligations were explicitly promises, whether the welfare gains of enforcement to the parties are clear

As this lengthy summary suggests, Gamage and Kedem make an intricate and layered argument. Their article seems to me especially innovative in its use of option theory to study consideration (though other topics in contract have benefitted from its use, as recent articles by Avery Katz in Virginia and Bob Scott and George Triantis in Columbia illustrate). And one does not expect an article filled with game theory to bring the insights of anthropology and sociology to bear on its central issues. This eclecticism is, I think, a strength of the article, and adds richness and context to its analysis. All that said, there are a few spots where the analysis seems less than complete or less than persuasive. The incompleteness may have been a strategy, to avoid adding bulk to an already substantial paper, but I think in places it leaves important questions underanswered. (This was especially noticeable as the article proceeded, which may suggest that the later sections are more nearly works in progress than the former.) As I read through the game theory, for example, I waited in vain for the references to Akerlof’s work on lemons markets, which seems to me quite closely related to the authors’ analysis. (Ditto some of the work of Michael Spence.) I could also add important writers on altruism and the nature of promising – Jon Elster and Diego Gambetta come to mind, but there are many others – whose work might add texture to that section of the paper. The discussion of trust-building promises might make a bow in the direction of transaction-cost economics. One might also take a closer look at some of the classics of the consideration literature, such as Edwin Patterson’s defense of consideration. And the authors state that they believe they are the first scholars to observe that “social norms sometimes prevent parties from voicing even nominal consideration.” Surely not?

More substantially, I’m not sure all of the arguments cohere. The long discussion of game theory ends by suggesting that because the authors’ model can find the option to be bound either welfare-enhancing or welfare-diminishing, that we cannot assume that the parties to an agreement would wish the option to be bound. But, as they point out, this depends entirely on the values they set for their constants. Are their values realistic, either for promises overall or for defined classes of promises? It may be that for a realistic range of values (particularly those for damages and for the costs associated with enforcement), the option to be bound is either clearly welfare-enhancing or clearly welfare-diminishing. (I note as well that they model the costs of fulfilling a promise by squaring the size of the promise, a not-trivial step which does not seem empirically supported.) This leads to a query about one of the assumptions of the Aghion-Hermalin model that Gamage and Kedem adapt here. Aghion and Hermalin point out that their model assumes that the parties have access to only one signal, and they show that restricting access to that signal may improve efficiency. But if the parties have access to many signals, restricting access to only a subset may actually reduce efficiency. Gamage and Kedem do much with the literature on extra-legal aspects to promise enforcement here, and with great effect, but I was left a little uncertain whether their game-theoretic analysis fully takes this into account. The parties to an actual agreement will draw conclusions about the likelihood of performance from more than the signals set out in the course of negotiation, and changing consideration doctrine will affect only some of them.

The authors’ use of the commodification literature also could be stronger. It might, for example, take into account the conflicting threads of contractarian and non-contractarian analyses of marriage and related promises (for a handy reference, see Bob Hillman’s excellent treatment in The Richness of Contract Law). And I wonder whether the authors should put quite as much emphasis on the signaling rationale for charitable promises as they do. Certainly some donations give status, in a relatively benign version of Veblenesque conspicuous consumption. Naming opportunities on professorships, buildings, colleges, or whole universities come to mind. But many donations, especially more modest ones, seem more clearly explained by altruism, particularly if made without the expectation of public acknowledgment. In addition, some rather broad statements in the conclusion may need a bit of qualification. I’m not sure that their tie-breaker rule about welfare-enhancement yields a line drawn just where they want it. Any small promise will, if enforceable and enforced, bring about enforcement costs well in excess of the value of the claim, whether the promise is basically donative or bargain-based. And their statement that “[t]he default rule should always be against involving the law in private disputes” may not capture their point here. If they distinguish private disputes from those involving social welfare-enhancement, then the line drawn cuts jaggedly across traditional bargain and traditional gift alike. (I should note that Gamage and Kedem point out that many of the problems of traditional analysis arise because the law tries to determine which promises are socially valuable. But don’t they take this into account as one of their tie-breakers?)

Finally, the authors give two examples of promises that their theory leaves unenforceable, but that might properly be enforced should empirical work show that the harm from potential enforcement is low or the costs of barring proof of reliability high. Their examples are charitable subscriptions and merchant firm offers. I suppose I can accept this for charitable subscriptions, but merchant firm offers? As Learned Hand pointed out in Baird v. Gimbel, there is always the possibility of making an option contract that will have the effect of a firm offer. But is this remotely practicable? Put otherwise, what is to be gained by requiring the empty recitals of an option contract when merchants already have means of making offers that are not firm? Perhaps another example would be more apt. Most of these comments are peripheral, and most of the others probably reflect more a need for more complete exposition than a need for rethinking. The core of the article – a broadly-based, fresh, and nuanced analysis of a thorny problem – is important and worthwhile, and I think it will make a real contribution to the already distinguished literature on consideration.

Junior Scholars | Bookmark

TrackBacks (0)

TrackBack URL for this entry:

Links to weblogs that reference Conglomerate Junior Scholars Workshop: Gamage & Kedem on the Consideration Paradox:

Comments (12)

1. Posted by Ethan Leib on September 16, 2005 @ 17:57 | Permalink

I understood close to none of the game theory. Yet I was virtually sure that it didn't need to be there and that it was getting in the way of a more comprehesive effort to elaborate upon your central ideas that the consideration doctrine could be used to help cabin certain agreements from being commodified. My instinct is that you guys should develop two separate articles. One 25-pager for a peer-reviewed law & econ journal containing all your models and a second one that more carefully develops you view of the consideration doctrine.

Another real problem is that you purport to resolve the paradox which you define as courts saying they care about consideration but refusing to look into the adequacy of consideration. I'm not sure anything you say about commodification actually addresses that "paradox." More basically, I'm not sure you actually have sufficient doctrine or empirical evidence that consideration serves this function to show that your account is anything other than normative.

Finally, I wonder whether we couldn't police the commodification problem through remedies rather than through enforceability simpliciter. That is, just because parties don't want to commodify their relationship doesn't mean that they mean to operate outside of law altogether. Indeed, perhaps we vindicate these non-commodified agreements by recognizing them not as unenforceable but as good candidates for specific performance.

Admittedly, I may have missed a whole dimension because I couldn't quite get how the game theory elucidated anything.

Those are my two cents. Good luck; I enjoyed reading it.

2. Posted by David Gamage on September 17, 2005 @ 12:54 | Permalink

First of all, I’d like to thank Christine and the other Conglomerate folks for hosting this event, and for allowing us to participate. Thanks also to Larry Garvin for insightful comments, and a better summary of our piece than we were able to come up with. We will likely draw on his description when rewriting the abstract. I’ll respond to some of the comments below.

3. Posted by David Gamage on September 17, 2005 @ 13:37 | Permalink

Larry, thanks again for reviewing our piece. We will definitely draw on your comments as we revise the paper.

As you point out, this is a work in progress. In particular, the last couple sections were very rough in the draft we posted last week. We’ve already revised these sections considerably since that draft.

One of the main dilemmas we are facing as we finish the paper is how to cover our topic while remaining within the new length requirements of the major law reviews. We’ve already cut substantially, and yet the paper is still too long. Unfortunately, this limits our ability to engage with the existing literature. If anyone here who has read the piece has suggestions for cuts we should make, we would be most appreciative. (I’ll respond to Ethan’s comments below, as to why I don’t believe it would make sense to split the piece into two articles). Aghion & Hermalin’s work follows from a substantial body of literature (including Akerlof’s work), and we consequently rely on this literature as well. To the extent we can fit in more citations, we will try to discuss some of the other major pieces in this thread of scholarship.

Similarly, our discussion of the commodification literature is highly truncated. There is simply no way we can do justice to this literature in our paper, doing so would require a book. We did not mean to imply that all charitable promises are of the status enhancing type. We meant to argue that the consideration doctrine can distinguish between charitable donations made out of altruism and ones made for signaling purposes, and that consideration should generally be available only for the former. We will try to make this more clear.

As for the more substantial points, Larry is absolutely correct in pointing out that the models’ results entirely depend on the settings for the constants. Might promises be universally welfare enhancing or diminishing within realistic settings for the models? Possibly. But how can we know what settings are realistic? In the absence of a comprehensive empirical analysis, we must reach some preliminary conclusions. And we doubt whether a comprehensive analysis of this sort would even be possible.

In regard to our squaring of the size of the promise in the cost term of our model, some exponent is needed, although our choice of exponent is arbitrary. An economically rational promisor should only promise if the benefits of doing so exceed the costs at some initial level. Yet the costs must eventually exceed the benefits or else the promise would be infinitely sized. We originally included another constant as the exponent term, but replaced that constant with the squaring exponent in order to make the equation more readable. We will add a better footnote explaining this.

Yes, Aghion and Hermalin’s caveat that their model’s results may not hold if promisors have many signals also applies to our analysis. We note that signaling spirals only occur within groups of promisors with similar observable characteristics – such that other signals are assumed to not be available. We could probably do a better job explaining this and discussing its implications and will try to do so as we revise the piece.

Your points about our tie breaker arguments and our example in the conclusion about merchant firm offers is well taken. We have already revised these sections considerably since the draft you read.

Thanks again for your comments. They have been very helpful. (I plan on responding to Ethan’s points in a separate post, which I hope to have up later this afternoon).

4. Posted by Paul Gowder on September 17, 2005 @ 18:31 | Permalink


Why not just make the costs (1-Pi)XY where Y is anything above one? That would make the costs increase faster than the benefits but not quite so fast as squaring.

5. Posted by David Gamage on September 17, 2005 @ 20:59 | Permalink


I found your comments to be very discouraging. Our game theory arguments were intended to be accessible to readers lacking a sophisticated background in economics. That an intelligent reader like you was unable to make sense of the game theory places doubt on our entire project.

As for your suggestion of dividing the piece, I fail to see how this could work. Our analysis that anti-commodification norms make consideration socially unavailable under certain circumstances doesn’t do much on its own. An entire paper dedicated to this proposition wouldn’t have any implications for policy. Only by combining this analysis with our models can we claim that providing a legally binding option is likely to be welfare enhancing in circumstances where consideration is socially available, but not in circumstances where norms block the use of consideration. Without understanding the game theory section, did you find our paper to be of any value?

Moreover, our economic analysis doesn’t do much without the section on commodification. We do not claim to have significantly advanced the economics literature. Our extensions of Aghion & Hermalin’s work are trivial from an economics perspective, our contribution lies in applying the work to a new area – namely the consideration doctrine. Our target audience is scholars who think about consideration, not economists.

Were you not able to follow even our attempt at an intuitive description of our model in the introduction to Part II and section A of Part II? We tried to develop our discussion in stages, limiting our use of mathematics to the very end, and describing our argument with text before even using graphs. Allon spent considerable time and effort rewriting my language in the hopes of making it comprehensible to a larger audience. I’d appreciate any suggestions on how we could make that portion of our paper more accessible.

I find this discussion to be disturbing on a more general level. If those of us who use economic analysis cannot find a way to convey the essential points of this analysis to a wider audience, we risk either having the economic insights ignored or else segregating off certain areas of policy so that only those who understand economics can fully engage in the policy debate. To some extent, I believe the latter has occurred in my primary field of taxation. I view one of my scholarly roles as attempting to translate economic arguments so that non-economic minded scholars can appreciate the arguments. It is not enough to merely say economic models argue for X. Economic analysis purposefully abstracts away from values that many consider to be important, myself included. I believe some of the pathologies inherent in our tax policy have arisen because non-economists fail to fully understand the implications of economic theory, while economists ignore other values that nevertheless affect policy outcomes. Our collective failure to create an effective dialogue prevents us from thinking about tradeoffs between these values in a systematic fashion.

Back to our paper: We never meant to imply we offered a descriptive picture of how the consideration doctrine actually arose. I actually have a theory on how the doctrine developed, which has nothing to do with our arguments here. Instead, we meant we “resolve the paradox” in the sense that we describe a functional role for the doctrine that has implications for how the doctrine should be applied. We will try to make this clearer as we revise.

6. Posted by David Gamage on September 17, 2005 @ 21:00 | Permalink


Adding a coefficient in front of the cost term as you suggest does not solve the problem by itself. If both parts of the equation – the cost and the benefit terms – start with zero cost and zero benefit for promises of size zero, then adding a coefficient makes the cost term either larger than the benefit term for all promises or smaller than the benefit term for all promises. In other words, either no promises would be made or all promises would be of infinite size. In order to solve this, we would need to add an additional constant to the benefit term so that making a promise of size zero would result in benefit for the promisor. There is nothing inherently wrong with this approach, but using an exponent instead of a coefficient seemed a little more intuitive to me. At some level, we should expect to see increasing marginal costs to increasing the size of a promise. Modeling increasing marginal costs requires the use of an exponent.

7. Posted by Paul Gowder on September 17, 2005 @ 22:49 | Permalink

David: Oh, of course. You can tell how long it's been since I've tried my hand at proper econ!

8. Posted by Ethan Leib on September 17, 2005 @ 23:02 | Permalink

I tried to follow the game theory but just don't have the sophistication for it. That isn't your fault, of course. And I confess to not taking many hours with the paper to really try to figure it out (assuming I even could have with effort).

I appreciate your general concern about the use of economic ideas--but that is just the way it is. Most contracts scholars these days are probably trained in the law & economics mold, so you can very comfortably use the economic language. If you truly want to popularize, I think you do have to dumb it down further for people like me. But that isn't necessarily a choice you need to make right now--and the really good popularizers tend to prove themselves to the experts first before they have the credibility to help the rest of us understand things. I acknowledge that the way the section is drafted does raise the question of whether such analyses have a place in a general interest law review. That was at the root of my suggestion.

More substantively, I did think there was a paper there without the economics section precisely because a normative account of the consideration doctrine that argues that its function to police commodification is interesting and worth developing further. That you also claim "that providing a legally binding option is likely to be welfare enhancing in circumstances where consideration is socially available, but not in circumstances where norms block the use of consideration" is probably most interesting to economists who use welfare-enhancement as a real baseline to evaluate, well, everything. Since most non-economists think normative arguments are perfectly admissible even when they are not especially welfare-enhancing, this connection, though important to you, is not immediately necessary to me. More, the intuitive idea that it is not useful to have a signal in an issue area where signals are generally discouraged could probably could be developed without models.

This leads to one more problem I tend to have with analysis developed from models rather than empirical data: they seem so speculative before tested with data that I always take them with a bit of salt. Anyway. This goes well beyond your paper, which--my criticisms aside--was provocative and enjoyable.

Regards to Allon.

9. Posted by Heidi on September 18, 2005 @ 8:08 | Permalink

Quick comment after short skim: I thought the models were reasonably accessible (speaking from the point of view of an articles editor for a general law review, but one who finds equations fun). If, you want to trim fat from the piece, Part I is screaming for it.

Re: squaring discussion: you can maybe sidestep questions of empirics by testing to see how much the exponent matters. If your results are unstable with respect to the exponent chosen, your model's probably hokey. If you get the same basic shape to your result over a wide range of exponents, you're probably okay. If your results depend wildly on the coefficient, your model is probably crap without empirical support.

10. Posted by David Gamage on September 18, 2005 @ 12:51 | Permalink

Thanks again Ethan,

I now see your point about the value of the commodification section without the game theory. We could write a paper just based on the commodification ideas, but that would take a significant amount of work beyond what we’ve already done. And I continue to think the game theory can’t stand alone without the commodification piece. We are going to continue developing – and trying to publish – the ideas as a joint paper. But we, or at least I, might return to the topic later to write an additional paper based more on the commodification ideas. I suspect the easiest way for us to do this would be to recruit another co-author who already writes about commodification. Where our current paper tries to extract ideas from the commodification literature without taking sides in the debate, the paper you suggest would almost certainly require more substantial analysis about the proper way of thinking about commodification and the normative implications of anti-commodification norms.

To some extent, I agree with your critique of modeling. I don’t really think of myself as a law & econ person; I have a much stronger background in organizational sociology. But conducting good sociological research requires the time and resources to do empirical studies – resources that I’ve lacked until recently. Models and other forms of arm-chair analysis can be spit out far more rapidly. And I do think models have some value and can be used to guide for future empirical research.

Post a comment

If you have a TypeKey or TypePad account, please Sign In

Recent Comments
Popular Threads
Search The Glom
The Glom on Twitter
Archives by Topic
Archives by Date
January 2019
Sun Mon Tue Wed Thu Fri Sat
    1 2 3 4 5
6 7 8 9 10 11 12
13 14 15 16 17 18 19
20 21 22 23 24 25 26
27 28 29 30 31    
Miscellaneous Links