19
supportive of Krehbiel’s (1986) model of Senate leadership and followership. After controlling
for a variety of other factors, we find that senators on the threshold of retirement do, in fact, tend
to initiate more filibusters than their colleagues, but that this effect is statistically marginal.
Those willing to entertain more liberal thresholds of statistical significance, though, will interpret
these results as being consonant with Krehbiel’s model of Senate leadership, which posits that
the chamber’s leaders use their scheduling prerogatives to capitalize on senators’ desires to
maximize their personal/political utility in something resembling an iterated prisoners’ dilemma
game and in a manner that allows cooperation to develop on the Senate floor.
While our new data indicate that retiring senators may be somewhat more likely than
other senators to lead filibusters, that conclusion should be tempered by the substantive
observation that cooperation certainly does not completely break down as senators approach
retirement, as a rational choice model might strictly predict. Of the 99 senators who retired
during the course of our 27-year time series, only 19 (19.19 percent) opted to filibuster during
the final Congress of their tenure; similarly, of the 45 “complete career” senators in our dataset,
only 12 (26.67 percent) led a final Congress filibuster. The fact that only roughly one senator in
four or five engages in the most obstructive form of dilatory behavior as they prepare to retire
from the chamber indicates to us that while purely rational considerations certainly enter into
senators’ behavioral calculations, they do not necessarily predominate. Rather, our findings are
consistent with a model of senators’ behavior in which internalized norms of cooperation usually
continue to mediate the most obstreperous kinds of conduct even when purely rational
considerations would entice utility maximizers to defect in large numbers.
Convention