One of the most distinguishing features of the American national government is the way separated institutions, especially Congress and president, share the power to make public policy. For example, the executive branch implements the laws the legislative branch creates; the Senate approves or rejects the president's appointments to the Supreme Court. Most striking, the president can block a bill passed by Congress by exercising the veto power and Congress may then attempt to override the veto. Although the separated powers system is a critical element of public policy making in the United States, many parts of it are only poorly understood by political scientists. For example, consider the presidential veto: what happens to a bill after it is vetoed? Under what circumstances does it die? When Congress repasses a vetoed bill, does Congress make concessions to the president? Under what circumstances are the concessions large, and when are they small? More broadly, what causes a bill to be vetoed? Although vetoes are rare events, many involve the great legislative issues of the day. So, answers to these questions would help us understand an important aspect of the American system of separated powers. Unfortunately, although political scientists know quite a lot about the frequency of vetoes, they do not know the answers to the above questions. Thre are two reasons for this. First, these questions approach vetoes from an unusual perspective. They assume a single veto is not an isolated event but just one stage in a bargaining process between Congress and president. They assume a sequential bargaining approach to the veto. Most political scientists have tended to approach each veto as a separate event, obscuring the bargaining dynamic between the two branches of government. Second, answering these questions requires data on individual vetoed bills, or even a random sample of all bills, rather than simply the aggregate number of vetoes per time period. These types of data have never been collected, and the process of collecting them is tedious and time consuming. This research project targets both of these problems. First, the investigators examine vetoes from a sequential bargaining perspective. Second, they collect data on individual bills. The researchers will collect data on about one hundred fifty bills in veto "chains" (bills vetoed and repassed), a random sample of four hundred vetoed bills (about fifty five per cent of all vetoes since 1946), and a random sample of about four hundred general bills; a total of about nine hundred bills (allowing for some overlap among the samples). By combining theory and data, the investigators will substantially enhance our understanding of the veto process and the role it plays in policy formulation.
