In Gilligan and Krehbiel's models of procedural choice in legislatures, a committee exerts costly effort to acquire private information about an unknown state of the world. Subsequent work on expertise, delegation, and lobbying has largely followed this approach. In contrast, we develop a model of information as policy valence. We use our model to analyze a procedural choice game, focusing on the effect of transferability, i.e., the extent to which information acquired to implement one policy option can be used to implement a different policy option. We find that when information is transferable, as in Gilligan and Krehbiel's models, closed rules can induce committee specialization. However, when information is policy‐specific, open rules are actually superior for inducing specialization. The reason for this surprising result is that a committee lacking formal agenda power has a greater incentive to exercise informal agenda power by exerting costly effort to generate high‐valence legislation.

During the early years of the New Deal, President Franklin Delano Roosevelt faced an enormous challenge. He believed that dramatic policy innovation was urgently needed on a wide range of issues, including agriculture, trade, banking, employment, and social insurance. However, although Roosevelt was well versed in policy, crafting workable proposals for such a far-reaching agenda was well beyond his capacity, or that of his immediate advisors. Instead, the president had to rely on a broader array of politicians and bureaucrats to develop new policies. Although these individuals shared some common goals (e.g., ending the Great Depression), they also had deep-seated ideological disagreements among themselves, and with the president, about what policies the federal government should pursue.Classic and contemporary theories of bureaucratic politics would suggest that this situation was ripe for exploitation and inefficiency. Weber (1991) famously argued that the typical ruler of a modern state is reduced to being a "mere dilettante" when dealing with institutionalized bureaucratic actors. More recently, scholars building on the work of Crawford and Sobel (1982) have argued that policymaking

One way that principals can overcome the problem of informational asymmetries in hierarchical organizations is to enable whistleblowing. We evaluate how whistleblowing influences compliance in the judicial hierarchy. We present a formal model in which a potential whistleblower may, at some cost, signal noncompliance by a lower court to a higher court. A key insight of the model is that whistleblowing is most informative when it is rare. While the presence of a whistleblower can increase compliance by lower courts, beyond a certain point blowing the whistle is counterproductive and actually reduces compliance. Moreover, a whistleblower who is a "perfect ally" of the higher court (in terms of preferences) blows the whistle too often. Our model shows an important connection between the frequency of whistleblowing and the effectiveness of whistleblowing as a threat to induce compliance in hierarchical organizations.

This online appendix contains all formal proofs, as well as analysis of an extension of a model with alternative assumptions about bargaining. We begin by establishing a general result on optimal policy development in our baseline model and our model of internal capacity.

One way that principals can overcome the problem of informational asymmetries in hierarchical organizations is to enable whistleblowing. We evaluate how whistleblowing influences compliance in the judicial hierarchy. We present a formal model in which a potential whistleblower may, at some cost, signal noncompliance by a lower court to a higher court. A key insight of the model is that whistleblowing is most informative when it is rare. While the presence of a whistleblower can increase compliance by lower courts, beyond a certain point blowing the whistle is counterproductive and actually reduces compliance. Moreover, a whistleblower who is a "perfect ally" of the higher court (in terms of preferences) blows the whistle too often. Our model shows an important connection between the frequency of whistleblowing and the effectiveness of whistleblowing as a threat to induce compliance in hierarchical organizations.