In their recent paper “Anti-Herding Regulation,” forthcoming in the Harvard Business Review, Ian Ayres and Joshua Mitts argue that many well-intentioned public policy regulations potentially harm rather than help situations. That is, because they seek to pool — or herd — groups of people, treating them as equal, they miss or mask important differences among the regulated, thus magnifying systematic risk. Anti-herding regulation, on the other hand, can produce socially beneficial information, in their words steering “both private and public actors toward better evidence-based outcomes.” Left to their own, or with various carrot-and-stick incentives, some groups, anyway, would instead fare better if allowed to separate or diverge.Typically and for separation of powers considerations, courts and the court process conduct a structurally different role from administrative agencies. However, when the legislature or a higher court devises a presumption that regulates conduct, judges are not as free to use their discretion in interpreting the law as they usually are, particularly as they would be with an opened-ended goal like deciding custody “in the best interests of the child.” What child custody statutes with presumptions do is to assume that the road to the “best interests” of each child is the same: that is, that a single solution will prove to be best for all children. In the law and economics framework of Ayres and Mitts, the statutes promote a “pooling” rather than a “separating” equilibrium.In purely financial or commercial settings, this sort of forcing may not be inappropriate, and may in fact be constitutionally necessary. However, with very few exceptions, whole groups of children cannot be forced into similar parenting situations as long as parents are acting within the quite broad variance given to “fitness.” It is this type of pooling that may in fact threaten systematic risk for those children least likely to have resources to cope with them — those whose parents cannot get along even to the extent of co-parenting well, or where there are the barriers to trust posed by such conditions as substance abuse, mental illness, or coercive-control intimate partner violence, or even where the parents struggled to maintain a viable financial life when living in a single household, now divided into two.Presumptions in child custody are naturally disfavored by three groups — the judges who lose their discretion and seemingly a part of their “raison d’etre” (though deciding contested custody cases is difficult and uncomfortable), the helping professionals who otherwise would assist in making determinations based on “best interests,” and whatever group is disfavored by the particular way the presumption is set. Custody rules, including the ones examined here, are generally set with a great deal of advocacy but without a great deal of empirical research behind them, and are rarely tested carefully to see whether justice to the individual affected children is being served. This paper presents an initial foray into such a test, looking at the effect of a strong child shared custody presumption on the behavior of parents and judges during and within five years of the original custody proceeding. There are limitations to this empirical approach: most of the cases do not provide close glimpses into what the parents were thinking at any given point in time, what arrangements judges would have ordered absent the presumption, or what professionals such as custody evaluators would have done differently. Even with these limitations, the empirical picture is a mixed one: some couples do fine with shared custody, and are able to adjust over time without result to acrimonious processes. Their children are presumably better for the frequent and meaningful contact with both the parents. The successes must be weighed against another group of cases where, at best, shared parenting seems to take place at the price of considerable parental conflict and continued litigation. The emerging factual pattern seems to fit within Ayres and Mitts’ prescription for a separating equilibrium, while the presumption, bolstered by strong community support for shared parenting, pushes for a pooling equilibrium highlighted by equally shared custody. Particularly troublesome (and unstable) are cases involving indications of domestic violence and/or substance abuse as well as those from the lower half of family incomes and the increasing number of unmarried couples affected by custody and child support orders.