Professor Eric Posner (and co-authors) on a classic strategy - and law's responses to it
What do labor organizers, Germans, and criminals have in common? Each of their respective bêtes noire - union-busting employers, the Emperor Tiberius, and clever prosecutors - are masters of a classic strategy: divide and conquer. Employers may try to buy off some employees and punish others to prevent unionization or gain an advantage in negotiations. Tiberius famously played German tribes against each other to ensure Roman dominance. And, in the well-known Prisoner's Dilemma, a nameless prosecutor separates conspirators and offers each a deal that makes squealing attractive. These are only a few examples of what surely must be one of the oldest moves in human interaction. But is "divide and conquer" just a loosely-used label, or do these different examples have something in common? If they do, does that tell us anything interesting about what we should think or do about uses of the strategy?
Chicago's Professor Eric Posner, along with co-authors Kathryn Spier and Adrian Vermeule (both at Harvard Law School) tries to answer these questions in his draft paper titled, not surprisingly, "Divide and Conquer", which was presented at last week's Works in Progress (WiP) talk. Prof. Posner and his co-authors come to the conclusion that there is a common thread running through many incidents labeled as "divide and conquer," and that there are interesting implications of this that might not be apparent when looking at each type of incident in isolation. Posner and his co-authors first classify different methods of divide-and-conquer behavior, then investigate the use and limits of legal rules for countering divide-and-conquer behavior in situations where it is harmful.
"Divide and conquer" is such a well-known and intuitive concept that a formal definition might seem unnecessary. Given the frequency with which the term is used, however, some precision is helpful. In particular, it was important for Prof. Posner to distinguish divide and conquer from the related 'strategy' of tertius gaudens ("the third rejoices") in which someone simply profits from others' internal discord. Divide and conquer requires that the actor profiting from this discord also be responsible for causing it. More technically, divide and conquer requires a unitary actor (either an individual or a group that has overcome collective action problems) who is in some way competing with a set of multiple actors. The unitary actor must create, exacerbate, or exploit collective action problems among his adversaries, and then profit from this discord. In each of the examples above, employers, emperors, and prosecutors all create such discord and profit from it - they are, therefore, practicing "divide and conquer".
Armed with this definition, Prof. Posner looks at a wide variety of other divide-and-conquer situations and asks whether there are any lessons to be drawn from their similarities. A key observation is that, from a utilitarian point of view, divide and conquer is not necessarily a good or bad thing. One party (the unitary actor) clearly profits at the relative expense of others (the multiple actors being divided and conquered), but this doesn't say anything about the net impact of the move, or its effects on outsiders. Sometimes the move seems clearly problematic, as in the case of targeted bribery of legislators or a tyrant who grants just enough favors to a few elites to stay in power. In other cases, the move is clearly beneficial - with the best example being the prosecutor in the Prisoner's Dilemma. His divide-and-conquer move hurts the criminals, surely, but benefits society by making crime much more costly. In many situations, however, it is not so clear whether the benefits of a divide-and-conquer move outweigh the costs. Whether Tiberius' or employers' divide-and-conquer moves are beneficial or not depends on a lot of specific facts and, ultimately, how one feels about unions and barbarian tribes. Put more formally, it is often unclear ex ante whether a given divide-and-conquer move will have externalities, and whether those externalities will be positive or negative.
Prof. Posner argues that a single rule praising or condemning divide and conquer, therefore, seems like a bad idea. A more targeted approach that discourages divide and conquer when we think the effects are negative, and possibly encourages it in cases where we think effects are positive would seem to be a better option. This matches up with what we see - the divide-and-conquer move is made frequently in many spheres without any legal restriction. In some cases, however, the law specifically blocks some types of divide-and-conquer moves. For example, labor laws prevent many types of union-busting divide and conquer tactics, making it illegal for employers to buy off groups of workers during union elections or other activities. Similarly, bankruptcy law blocks many types of threats or side payments that might be used to force through a Chapter 11 plan. The Constitution also bars some types of divide and conquer strategy that the federal government might use against the states: "[n]o preference shall be given by any regulation of commerce or revenue to the ports of one state over those of another" (Article 1 Sec. 9).
In each of these cases, Prof. Posner observes, the legal solution to divide-and-conquer moves that are perceived as problematic is to introduce a non-discrimination rule. Such a rule forces the unitary actor to treat each of the opposed multiple actors equally, or at least (as in bankruptcy law) to treat each class of actor equally. Like all rules, these non-disclosure rules risk being either underinclusive ("bad" divide-and-conquer moves might still be legal) or overinclusive ("good" or "neutral" moves might be unjustifiably banned). This is unavoidable - it is simply really hard to know in advance whether a set of proscribed moves will be big enough to capture everything harmful and small enough not to impose excessive costs of its own.
This problem, of course, is not limited to divide and conquer - Prof. Posner observed that it is simply another manifestation of a common legal problem - that of rules vs. standards. Here, we have a large category of similar behaviors - divide and conquer - and a wide variety of very specific rule-based approaches. Further, each of the legal disciplines in which these rules are crafted and applied are relatively divorced from one another - lessons learned about good rules (or standards) governing divide and conquer in, say, the bankruptcy context are unlikely to be communicated to other contexts in which the problems turn out to be quite similar.
While Prof. Posner does not offer a "unified theory" for dealing with divide and conquer that would cut across these divides, or even suggest that such an approach would be useful, the insight that similar lessons might be learned from what have previously been considered to be distinct problems is, I think, a valuable contribution. Put differently, the paper does not suggest either a new problem or a new answer to an old one, but a new place to look for both.
It's not immediately clear, however, whether this is an important revelation or not. As one commenter asked, if Tiberius himself read this paper, would he act or consider his actions any differently? In the same vein, Prof. Posner commented in his talk that he remains unsure about whether this paper is profound or obvious. Having attended a year's worth of WiP talks, I now understand that this distinction is much finer than it would first appear to both writers and readers. I am optimistic, however, that Prof. Posner and his co-authors are on to something useful here. Cutting across the insularity of disciplines to find simple, common threads is, I think, helpful and productive. This paper seems like a good example of that.