Introduction
On July 27, 2001 the House of Representatives adopted an amendment to H.R. 2620, the
Fiscal 2002 Veterans Administration/Housing and Urban Development appropriations bill, to
ban the expenditure of funds for activities that might hinder nationwide reductions in drinking
water arsenic levels. The amendment was adopted despite strong opposition from the majority
party leadership, including both the majority leader and the majority whip.
earlier, the House had adopted three amendments to the Fiscal 2002 Interior Appropriations bill
prohibiting oil drilling in certain locales; these too were passed over strong Republican
objections. Majority party losses or “rolls”
between the 102d and 107th Congresses (1991-2000), the House majority party was rolled more
than 600 times—a number that represents over nine percent of all recorded votes held in the
chamber during that period.
Two conditions make the examples above noteworthy, however. First, since achieving
majority party status at the outset of the 104th Congress (1995-1996), the Republican leadership
has worked to promote cohesiveness and to discourage defections from the party’s preferred
policy positions (Evans and Oleszek 1997; Fleisher and Bond 2004). Second, each of the four
majority party rolls we describe was included in the 2001 iteration of the League of Conservation
Voters’ (LCV) annual congressional scorecard. LCV is a coalition of pro-environment leaders
1
Congressional Quarterly’s Congressional Roll Call 2001, H288.
2
For our purposes a “roll” is a vote in which a majority of the majority party votes on the losing side. One could use
the positions of the majority leader and majority whip to determine the party's position, however we use the majority
of the majority to conform with the extant literature (Cox and McCubbins 2005) and because there are instances in
which the two party leaders do not vote in the same direction. It is worth noting the position of the two majority
party leaders and the majority of the majority coincides on more than 80 percent of votes for Democratic majorities
and more than 90 percent of votes for Republican majorities.
3
Note that we use all non-unanimous votes in making this determination, while some scholars use only votes on
final passage. See Krehbiel and Woon (2005) for a discussion of vote types and inferential error.
1