Thomas L. Brunell and William Koetzle, "A Divided-Government-Based Explanation for the Decline in Resignations from the US Senate, 1834-1996," Party Politics, 5 (October 1999), 497-505.

First Paragraph:

The US Senate experienced the resignation of two of its most senior members during the 104th Congress. Senators Robert Dole (Republican, Kansas) and Robert Packwood (Republican, Oregon) both left office before the expiration of their terms, Dole to run for president and Packwood to run for cover. These highly publicized resignations raise an interesting question in modern American politics: why are fewer and fewer senators resigning their seats before the end of their term? In the 19th century, for example, it was quite common for senators to resign from office and not serve out their term. Between 1834 and 1850 no less than 47 senators resigned from office, contrasted with only 13 Senate resignation in the period 1980-96. There are undoubtedly numerous reasons for this decline. Two related causal factors are the increase in careerism in Congress and the concomitant institutionalization of the body (Polsby, 1968; Rohde et al., 1985).

Figures and Tables:
Table 1: Breakdown of US Senate resignations, 1834-1996
Figure 1: Total Senate resignations and propostion of senators who could resign in a way favorable to their party

Last Paragraph:
Our starting point was the puzzling decline in the number of resignations from the US Senate. While the rise of institutionalization of Congress is undoubtedly linked to this phenomenon, we have shown that this decline is also linked to a change in the opportunity structure. Specifically, the data presented here demonstrate that senators are significantly more likely to resign when their party will continue to control that seat. Given the natural position of senators in their party -- they are generally high-ranking and well-respected members of their respective political parties -- this is not surprising. The incentive for senators to resign only when their party will retain control of that seat has remained constant over time. Yet the increased incidence of divided party control between senators and the body with the power to appoint a replacement (i.e. the state legislature or the governor) has resulted in a marked decrease in the number of senators who are able to resign with the assurance that a replacement will be from the same party. Occasionally resignations from the Senate still do occur and when they do, they remain overwhelmingly of this type, despite the fact that it is now more difficult to resign with the guarantee that the same party will retain control of the seat.