How Lame Are Lame Ducks?

2016 Update

Under normal circumstances, members of Congress face a number of perverse incentives. Ironically, a lame duck session of Congress mitigates some of these perverse incentives at the same time that it diminishes the discipline of electoral accountability. During a lame duck session, members are less likely to vote at all, and when they do vote, they are less inclined to follow the wishes of their party or their constituents.

A lame duck session of Congress occurs when legislators meet after an election has been held but before the next Congress has taken office. Lame duck sessions are often criticized by the victorious party in the election, and a common critique is that the lame duck members—undisciplined by electoral constraints—vote irresponsibly. There are subtle but statistically significant differences between voting patterns in regular and lame duck sessions, as revealed by analysis of more than 52,000 House and Senate roll call votes. 

During a lame duck session, members are slightly less likely to side with their own parties and less likely to vote at all. These patterns persist in very lame duck sessions—those that take place following the loss of majority status within a single house. In these sessions, however, a new pattern emerges: Senators become less likely to cast bipartisan votes. Beyond these voting patterns, it is difficult to say whether members vote more or less “responsibly” during lame duck sessions of Congress. Our analysis supports the primary findings of the existing literature on lame ducks. Past studies have found lame duck legislators to be less likely to indulge most special interests, but others suggest some legislators may be more likely to indulge one particular special interest: their next employers. In this paper, we explain how incentives change for lame duck legislators, briefly review past research on lame ducks, and present our statistical findings that support and add to the existing literature. 

WHAT MAKES A LAME DUCK SESSION SPECIAL? 

Lame duck sessions differ from regular legislative sessions in that some members will have lost their reelection bids and others will be on the brink of voluntary retirement. 

For many, the mere presence of members who will not be returning is prima facie evidence that lame duck sessions are undesirable. As Yale law professor Bruce Ackerman says, “It is utterly undemocratic for repudiated representatives to legislate in the name of the American people.” Indeed, John Nagle of Notre Dame has demonstrated that these sorts of concerns animated the push for the 20th Amendment, which, after 1933, eliminated the long and mandatory lame duck session that had followed each election since the founding of the Republic.

The force of this critique is blunted by the fact that the vast majority of members win their reelections: Since 1962, just 6 percent of representatives and 13 percent of senators have lost reelection. And in the average session, only about 5 percent of members retire.

In the 2016 election, the departing member rate was especially low. Just 3 percent of representatives and 6 percent of senators who sought reelection lost their bids, while 6 percent of representatives and 5 percent of senators retired.

So most legislators who sit during a lame duck session will have just passed an electoral test. A large body of literature examines the reasons that incumbents are unlikely to lose their seats. These factors include name recognition, a fundraising advantage, and uncompetitive, gerrymandered districts. Current estimates suggest the incumbent advantage is between 8 and 10 percentage points.

Among those members who do retire or lose their reelection bids, lame duck sessions might be expected to alter incentives for a number of reasons:

  1. Nonreturning members may feel less of an obligation to serve the interests of their constituents since they no longer need their approval for reelection. Depending on the point of view, this may result in more “ideological” votes or more “principled” votes. It is also likely to result in diminished work effort.
  2. Nonreturning legislators are also free to ignore whatever political bargains (commonly called “logrolls”) they have struck with their fellow members, especially party leaders. This has an ambiguous effect on public welfare because, on the one hand, logrolls tend to facilitate policies that favor parochial interests, but on the other hand, logrolls can also help minority interests register intense preferences.
  3. These members may feel less pressure to cater to special interests whose financial and organizational support is often needed for political success. This may make lame ducks more likely to serve the general population. 
  4. There is an important caveat to the previous factor: Nonreturning members may be more likely to cater to the interests of one particular special interest—their next employers. And for many ex-members, this next employer is a lobbyist or a client of a lobbyist.

On balance, it is not clear whether these altered incentives are likely to produce better or worse public policy.

Empirical evaluation of lame duck sessions is complicated by the fact these sessions occur at the end of the year, when more complex and controversial bills are likely to come up for a vote. This is because large, complex pieces of legislation take longer to assemble and controversial legislation takes longer to whip. It is possible, therefore, that votes cast near the end of the year will tend to be more momentous and more controversial, whether they take place during an election year or not. For this reason, we believe it is important to control for any confounding “time of year” effect when assessing the influence of lame duck sessions on voting patterns. 

WHAT DO WE KNOW ABOUT LAME DUCK VOTING PATTERNS? 

In recent years, a number of researchers have documented differences in voting patterns between regular and lame duck sessions of the House of Representatives. These authors have found House members less likely to participate in roll call votes and more likely to cast votes that are inconsistent with the ideological wishes of their party leaders. There is also some evidence that members who do not face electoral constraints are more likely to ignore the ideological wishes of their constituents.

We build on this research, extending it to the Senate and introducing controls to account for different voting patterns that might emerge at the end of the year, whether members are voting in a lame duck session or not. We also assess the effect of a “very lame duck” session, that is, one that takes place following the loss of majority status in the chamber. 

Our analysis is based on more than 29,000 House and 22,000 Senate roll call votes from 1939 through 2014. A roll call vote is a recorded, chamber-wide vote on any measure. It could be a vote on the final passage of a bill, an amendment, or a procedural vote. 

To test whether lame duck and very lame duck sessions are associated with altered legislative behavior, we focus on three main voting patterns: share of missing votes, party unity, and bipartisan votes. The share of missing votes is simply the number of members who fail to vote on a roll call vote divided by the total number of members in the chamber. Party unity is the share of members who vote with their party on each roll call vote, calculated in terms of Republican unity, Democratic unity, majority party unity, and minority party unity. Finally, a bipartisan vote is a roll call vote that garners majority support from both major parties. We hypothesize that during lame duck sessions, members will be less likely to vote, and when they do, they will be less likely to vote in line with their party leadership. 

Figure 1 shows how lame duck voting patterns differ from regular session patterns. These estimates control for the idiosyncratic patterns of particular congresses and for the time of the year. Except where indicated by hatched coloring, the estimated effects are statistically significant. In most cases, the effect is relatively modest and larger in the House than in the Senate.

Viewed in light of typical voting patterns, the largest effect of a lame duck session is to reduce member participation. Typically, about 9 percent of representatives and 10 percent of senators skip roll call votes, but during a lame duck session, an additional 4.2 percent of representatives and 2.75 percent of senators miss votes. In other words, representatives are 46 percent more likely and senators are 27 percent more likely to miss a vote if it happens during a lame duck session.

Lame duck sessions persistently reduce party unity, but not by much. Typically, about 83 percent of House members and 80 percent of senators vote with their party; but during a lame duck session, party unity is reduced by about 3 percentage points. The pattern persists across all party categories except Senate Republicans and the Senate minority party (which are typically one and the same). Lame duck sessions have no statistically significant relationship to bipartisan roll call votes—those that are supported by majorities of both parties.

Figure 2 shows how voting patterns during very lame duck sessions differ from those of regular sessions. As with lame duck sessions, the effects of a very lame duck session are typically stronger in the House than in the Senate. As with lame duck sessions, very lame duck sessions are associated with a greater tendency to skip votes. Very lame duck sessions are also associated with less party unity, though in both chambers the effect seems to be driven by Democrats and by the majority party. (In fact, very lame duck sessions are associated with slightly more unity among the Senate minority party.) Interestingly, a very lame duck session reduces the odds of a bipartisan Senate vote by about 14 percent but has no statistically significant effect on a bipartisan House vote.

WHAT DO WE KNOW ABOUT HOW LAME DUCK SESSIONS AFFECT LEGISLATION? 

Beyond the voting patterns, the effect of lame duck sessions on policy is a more difficult question to tackle and has accordingly received less attention. Table 1 lists major pieces of legislation passed during lame duck sessions. No discernable ideological pattern is obvious. The diverse list includes the impeachment of President Clinton, the adoption of major trade deals such as the landmark 1994 expansion of the General Agreement on Tariffs and Trade (GATT), and the creation of the Department of Homeland Security. 

As we note above, theory predicts that nonreturning members will be more likely to ignore the ideological preferences of their constituents and their party. Indeed, empirical tests suggest that members are more likely to depart from their typical ideological voting patterns during lame duck sessions.

Theory also predicts that vacating members will be less responsive to (most) special interests. And there is some evidence for this as well. One study examined individual votes on a protectionist measure adopted during the 1982 lame duck session of the House of Representatives. The researchers found that 66 percent of nonreturning members opposed the measure while 57 percent of returnees supported it. A more recent study examined an auto manufacturers bailout bill that passed the House during the 2008 lame duck session and later died in the Senate. The author found that—unlike returning members—retiring representatives were not responsive to industry campaign contributions.

This does not, however, mean that lame duck sessions cause members to ignore the entreaties of all special interests. President Harding’s controversial proposal to subsidize the construction of US merchant ships is illustrative. The bill was ultimately taken up in a lame duck session of the 67th Congress, where defeated Republicans overwhelmingly supported the measure. When it came time to seek post-congressional work, a “grateful president” saw to it that these members were “disproportionately rewarded [with administration positions] . . . for their unpopular stand.”

CONCLUSION 

Under normal circumstances, members of Congress face a number of perverse incentives. They are under constant pressure to indulge the desires of special interests because these groups are well organized, well informed, and often well-heeled. They are also encouraged to favor the priorities of party leadership, for leadership controls both the agenda and committee assignments, which affect legislative and fundraising success. And, of course, members are incentivized to favor the often-parochial priorities of other colleagues because the reciprocal support of other members is necessary for legislative success of their own priorities. 

The political process as a whole permits small, concentrated interests to wield disproportionate power. It allows agenda setters to manipulate the outcome of votes. And it permits winning coalitions to adopt policies that benefit themselves or push costs onto others outside the coalition. Even though voters themselves sometimes reward bad policy choices, regular elections are typically thought to be the best bulwark against these pressures.

Ironically, a lame duck session of Congress mitigates some of these perverse incentives at the same time that it diminishes the discipline of electoral accountability. Lame duck sessions seem to make members more independent. During a lame duck session, members are less likely to vote at all, and when they do vote, they are less inclined to follow the wishes of their party or their constituents. Lame duck members may be less inclined to indulge special interests, but they may make an important exception for their next employers. Given the fact that 80 to 90 percent of members win their reelection bids, these effects are unsurprisingly small.

Publication Note: This report was originally released in 2014. We have updated it to account for the most recent lame duck session of Congress, allowing us to analyze an additional 2,875 votes. 

Mercatus AI Assistant
Ask questions about this research.
GPT Logo
Mercatus AI Research Assistant
Ask questions about this research. Mercatus Chatbot AI More Details
Suggested Prompts:
Ask us anything. We use OpenAI's ChatGPT 4o base model to answer any question about Mercatus research.