Daniel L. Chen (University of Toulouse 1 - Toulouse School of Economics Institute for Advanced Studies/Harvard Law School LWP; Harvard University - Law School), Jens Frankenreiter & Susan Yeh (George Mason University School of Law) have posted Does the Law Matter and How Would We Know? on SSRN. Here is the abstract:
Public enforcement of law relies on the use of public agents, such as judges, to follow the law. Are judges motivated only by strategic interests and ideology, as many models posit, rather than a duty to follow the law? We use the random assignment of U.S. Federal judges setting geographically-local precedent to document the causal impact of court decisions in a hierarchical legal system. We examine lower court cases filed before and resolved after higher court decisions and find that lower courts are 29-37% points more likely to rule in the manner of the higher court. The results obtain when the higher court case was decided in the same doctrinal area as the pending case and when the higher court case was decided on the merits. Reversals by the higher court have no significant effects. These results provide clean evidence that judges are motivated to follow the law and are not solely motivated by policy preferences.
And from the paper:
We show that District Court judges react to Circuit Court precedent by adjusting their decision standard in the direction of the precedent. By looking only at cases that have been initiated before, but terminated after the decision in the Circuit Court case was rendered, we hold constant the case sample and avoid the potentially blurring effects of the selection of cases for litigation. We exploit random assignment of Circuit Court judges and the predictive power of biographical characteristics to control for reverse causality and omitted variable bias.
Our investigation of piercing the corporate veil cases from 2000 to 2004 shows that District Court judges are more likely to rule in favor of the plaintiff after a pro- piercing Circuit decision. This effect is large — our estimates show a shift in success rates between 29 and 40 percentage points. The results of our instrumental variable regression suggest that this effect warrants a causal interpretation. This effect is highly statistically significant (p-value .006) when restricting the analysis to Circuit Court cases that decided on the merits of the PCV claim.
Very interesting. On a cursory reading, I cannot assess the methodology, but this paper will be of obvious interest to anyone interested in judicial behavior.
For my broad take on the issues implicated by this research, see The Positive Foundations of Formalism: False Necessity and American Legal Realism.