Assessing congressional responses to growing presidential powers: the case of recess appointments.

Black, Ryan C. ; Lynch, Michael S. ; Madonna, Anthony J. 等

In 2007, President George W. Bush indicated that he planned to fill vacancies on several key independent boards and commissions by using his presidential recess appointment power. Anyone receiving such an appointment could serve for up to two years and would not be subjected to the "advice and consent" of the Senate. Senate Republicans supported Bush's proposed usage of the recess appointment clause. Senator Judd Gregg (R-NH) described it as "a reasonable management tool" to overcome potential Democratic obstruction (Billings 2007). Democrats, however, were highly critical of Bush's current plan as well as his previous decisions to make recess appointments. Senate Majority Leader Harry Reid (D-NV) criticized the president's penchant for making "controversial recess appointments" (Billings 2007). Senator Chuck Schumer (D-NY) went further, characterizing Bush's past recess appointments as a "punch in the face" (Congressional Record, June 9, 2005). (1) In November, Reid announced plans to keep the Senate in perpetual session, denying Bush the opportunity to make any more recess appointments. (2)

The decision to block the president from making recess appointments was an unprecedented procedural maneuver that had far-reaching policy consequences. After Reid's announcement, large numbers of vacancies at several independent agencies and commissions could not be filled. These understaffed boards could not implement policy or enforce previously made decisions. Vacancies effectively shut down the Federal Election Commission and prevented it from ruling on many important issues such as the decision to approve public financing funds for Senator John McCain's (R-AZ) presidential campaign (Murray 2008). (3) The five-member National Labor Relations Board had only two members for over two years, calling into question the legitimacy of the decisions issued by the board. (4) The bipartisan Securities and Exchange Commission operated without any Democratic nominees for almost six months.

The unfilled vacancies on these boards have greatly altered the ideological makeup of these bodies. If Bush had been allowed to appoint members to boards without interference from the Senate, it seems clear that the boards would have produced rulings and policies more consistent with Bush's own views. Senate Democrats were able to craft independent boards, and therefore policy, more to their liking by preventing Bush from using his recess appointment power.

The Senate's decision to block the president from utilizing his recess appointment power poses an interesting question for scholars of inter-branch politics. The number of presidential recess appointments has increased steadily over the past three administrations with virtually no credible attempt by the Senate to curtail the practice. This behavior is consistent with political science literature that has pointed to the growing unilateral power of the president (Cooper 2002; Howell 2003; Mayer 2001; Moe and Howell 1999a, 1999b). These scholars argue that while the president alone can initiate the use of unilateral powers, members of Congress must act collectively to prevent the use of these powers. Because the costs of collective action by Congress can be prohibitively high, Congress is seldom able to check the growing power of the executive branch. In this article we seek to evaluate why Congress was able to achieve collective action and take aggressive action to counter the power of the executive.

We argue that for Congress to effectively check increasing presidential powers two conditions must exist. First, the president's use of a unilateral power must create high political costs for members of Congress. Second, these costs must be felt by a sufficiently large number of members to reach a threshold necessary to attain collective action. In most interbranch conflicts these conditions are not met and the president is able to garner ever-growing power relative to the Congress (Howell 2003). However, the case of recess appointments is a rare contemporary example of the Congress successfully checking the unilateral powers of the president.

We argue that the Senate was able to successfully check Bush's unilateral power to make recess appointments for two reasons. First, Bush's frequent use of such appointments increased the political costs of allowing recess appointments for many members of the Senate. If Bush had been allowed to make recess appointments unabated in his final year in office, his appointees would have altered policies generated by independent agencies in a way that held high political costs for Democratic members of Congress. Second, we argue that the key to understanding this episode is the recognition that the number of members sufficient to achieve collective action was relatively low. The rules and practices of the Senate allowed Reid, in his role as majority leader, to abolish recesses--thus ending the possibility of recess appointments. While Republicans generally supported Bush's use of recess appointments, Reid only needed the support of his Democratic Senate colleagues to successfully counter the president. This low threshold for collective action provided the Senate with a unique opportunity to successfully check a unilateral power of the president.

To make this case, we take a two-pronged approach. First, we review Senate responses to controversial recess appointments and compare them to Reid's recent actions. Second, using data from all civilian nominations and recess appointments from 1987 to 2006, we analyze the partisan incentives for Reid and the Democrats to block the president. We find that President George W. Bush's usage of the recess appointment clause had far exceeded that of his predecessors. Thus, Reid's refusal to allow the Senate to go into recess has had a major effect on the composition of the bureaucracy. Our results suggest that had Reid allowed the Senate to go into recess, recess appointments would have been used to fill approximately 54% of all vacancies on major independent boards or agencies. Such a result would have represented an unparalleled expansion of unilateral presidential powers that would have generated high political costs for Democratic members.

The lessons learned from this episode have major implications for the study of separations of powers. Presidents can expand their use of powers, but there is a limit to how far they can push. Excessive presidential actions face retaliation by Congress (Corley 2006). How Congress can retaliate depends on the rules of the game. For a unilateral power like presidential orders, supermajorities in the House and Senate must come together to pass legislation and defend it from a presidential veto. For recess appointments, a simple change in Senate procedure allows a bare majority in the Senate to prevent presidential use of a unilateral power. Just as the president can creatively use ambiguities in the Constitution to expand his powers, Congress has the option of using ambiguities in their own rules of procedure to successfully check presidential powers.

The article proceeds as follows: (1) we reevaluate the constitutional foundations for recess appointments; (2) we review the limited political science literature on recess appointments and how it ties into the broader debate on unilateral presidential powers; (3) we examine past attempts by the Senate to curtail presidential use of recess appointments; (4) we consider the various factors that could influence a president's decision to utilize recess appointments; and (5) we examine recess appointment use in both a descriptive and a multivariate setting. Finally, we conclude with a discussion of how recess appointments may be used in future administrations.

The Nominating Process

The U.S. Constitution provides that nominations to the executive and judicial branches be made by the president and subjected to "the advice and consent of the Senate" (U.S. Constitution, Article II [section] 2). This provision all but ensures conflict between the executive and legislative branches during the appointment process. Both branches of government would like to exert a disproportionate level of influence over the bureaucracy and judiciary to ensure its preferences are represented.

In the general nomination process, the president first submits a formal nomination to the Senate. The Senate's presiding officer refers that nomination to the relevant committee, where the committee chair schedules a confirmation hearing. For the nomination to proceed, a majority of the committee must then report the nomination to the floor. If a majority of senators vote to confirm, the nominee then simply waits for a formal commission from the president. Historically, a negative floor vote is the rarest way for a nomination to fail. More often, nominations are defeated before they reach the Senate floor. For either ideological reasons or time constraints, committee chairs will often not schedule a confirmation hearing. When the formal congressional session terminates, it takes any unconfirmed nominations with it. (5) If scheduled, the nominee could also fail to receive a majority vote in the committee. This blocks the nomination from reaching the full Senate, effectively killing the nomination. Finally, opponents of the nomination can filibuster the nomination. To overcome the filibuster, supporters of the nominee would need 60 votes to invoke cloture. Without cloture, the nomination would fail when the session ended. (6)

The president has institutional advantages that may allow him to circumvent the Senate. Article II, Section 2 of the Constitution states, "The President shall have Power to fill up all Vacancies that may happen during the Recess of the Senate, by granting Commissions which shall expire at the End of their next Session." That is, the Constitution empowers presidents to avoid the Senate and install individuals in vacancies without senatorial consent. When the president makes a recess appointment, he can fill a vacancy with his preferred candidate. Once there, the recess appointee enjoys the same formal authority as congressionally confirmed nominees. So long as the recess appointment follows the law, the appointee enjoys all the powers of a confirmed nominee. Even the pay is the same.

The difference between recess appointees and confirmed nominees stem from the length of time they may serve (Hogue 2007). A confirmed nominee serves the full time allowed in the position (which may be a statutorily limited time period or at the pleasure of the president). The recess appointee, however, may only serve until the end of the next Senate session.

Before a president can make a recess appointment, there must be a vacancy (this formal existence of a vacancy may be difficult to determine). After a vacancy exists, the president may install a recess appointee only while the Senate is in recess. (7) Once appointed, the recess appointee may serve only until either (1) the Senate confirms a nominee to the position (the recess appointee or someone else) or (2) the end of the Senate's next session, whichever comes sooner. (8)

Because presidents use recess appointments to influence policy making (Black et al. 2007), Congress has established a number of ways to prevent them from being made. (9) For instance, Congress can often legislate when a vacancy exists and thereby preclude a recess appointment. More specifically, Congress can legislate that individuals serving in term-limited positions must continue serving after the expiration of the terms in a holdover capacity until a successor is confirmed to ensure the smooth functioning of government. When a person serves in a mandatory holdover capacity (i.e., where the statute suggests that the person shall continue to serve after the expiration of his term) or where the holdover capacity is for a specified length of time (e.g., for one year), the holdover--even though the appointee has served longer than the specified term--does not create a vacancy (Hogue 2007). On the other hand, when federal statutes allow--but do not require--the officeholder to serve in a holdover capacity, a vacancy exists that the president can fill with a recess appointment. (10)

As we argue more fully below, the Senate may adjust its calendar so as to block the president from making recess appointments. Indeed, Reid's decision to keep the Senate in pro forma session with the understanding that no real business would be conducted appears to have been one low-cost solution to blocking recess appointments. (11)

Recess Appointments in Political Science

Despite the potential negotiating leverage provided by the recess appointment clause, such appointments have received scant attention from scholars of interbranch relations. This dearth of attention is due to a combination of factors. First, throughout most of American history, the recess appointment clause has been relatively noncontroversial. In the early Senate, the time between formal sessions of Congress was quite long. Thus, the impetus behind placing the clause in the Constitution was to ensure the good workings of the bureaucracy and judiciary during these periods. As several prominent legal scholars have pointed out, recess appointments were made almost exclusively during these long intersession recesses (Carrier 1994; Halstead 2005; Rappaport 2005). It was not until recently that presidents have expanded their use of recess appointments to include intrasession recesses--or recesses occurring within a formal Senate session (Carrier 1994).

Second, most of the literature on the nominating process focuses on explaining the relative lack of manifest conflict between the president and Senate. These scholars start with the observation that nearly all presidential nominees are confirmed by the Senate (Krutz, Fleisher, and Bond 1998). Despite this common observation, scholars generally conclude that the president does not unilaterally dominate the appointment process.

Scholars adopting a formal theory approach suggest the president rationally anticipates legislative action and makes his nominations accordingly (Calvert, McCubbins, and Weingast 1989; Hammond and Hill 1993; Moraski and Shipan 1999; Nokken and Sala 2000; Snyder and Weingast 2000). This literature stresses the importance of the two branches' ideological preferences but also the confirmation and reversion forecasts for the nominees. Senators compare how a nominee would shift the likely policy outcome of the board, agency, or court once approved and compare that to the current policy output. Presumably, if the former is a certain degree worse than the latter, the senator would vote against confirmation and preserve the status quo. Presidents, seeking to get their nominees approved, would take into account the Senate's calculation and nominate individuals based not only on ideology but also the likelihood of confirmation.

Recent empirical work has moved past the high success rate of executive nominations and focused on the increasingly long delays between nomination and confirmation. The longer a nominee is unconfirmed, the less time he or she is spending passing policies favored by the executive branch. McCarty and Razaghian (1999) find that the presence of large numbers of senators not in the president's party increases the amount of delay in the nomination. Binder and Maltzman (2002) reinforce this finding, demonstrating that institutional constraints like divided government and the ideological distance between the president and opposition party median play a pivotal role in delaying the confirmation of a nominee.

Scholarship on executive-legislative relations has focused on presidential unilateral powers or powers where the president has the advantage of moving first and forcing Congress or the courts to react to his move (Moe and Howell 1999a, 1999b). Presidents can use this first-mover advantage to significantly impact policy outcomes. Scholars have examined unilateral powers such as executive orders (Mayer 2001), presidential signing statements (Cooper 2002), and executive agreements (Moe and Howell 1999a). Black et al. (2007) argue that because presidents move first when making recess appointments, and because recess appointments have historically been used to fill positions that have the greatest policy impact, recess appointments constitute a unilateral power. The recess appointment clause allows the executive branch to forgo the need to anticipate the Senate's preferences and bypass increasingly long nomination delays. Corley (2006) finds historical evidence for this, demonstrating that presidents were more likely to use recess appointments if they lacked partisan support in the Senate.

Recess Appointments and Senate Responses

Recent presidents have drastically increased their use of recess appointments. (12) Figure 1 reports the percentage of all appointments made via recess appointment from 1987 to 2006. While Bush has made the most recess appointments of any previous president (Hogue and Bearden 2008), other presidents have made frequent use of recess appointments. Ronald Reagan and William Clinton had years where more than 10% of their appointments were recess appointments. Also, Bush's use of recess appointments has been steady and increasing since the beginning of his presidency (Black et al. 2007).

Further, we suggest that Figure 1 underestimates the policy consequences of the Bush administration's large number of recess appointments. Specifically, the raw count data does not account for recess appointments Bush would have made his last two years in office. These two years are likely to see a sharp increase in the number of recess appointments due to the presence of divided government in the final two years of his administration. We review previous responses to recess appointments, arguing that these proved inadequate in the past.

[FIGURE 1 OMITTED]

While President George W. Bush has used recess appointments more frequently than his counterparts, he was certainly not the first president to do so. He is, however, the first to have his ability to utilize the recess appointment clause blocked by a Senate staying in permanent session. So why has the Senate just now started to push back against the president on this issue?

The Senate has previously tried several different methods to stop, or at least slow, presidential use of recess appointments. As with all unilateral powers, congressional attempts to combat recess appointments have required the cooperation of the courts or the support of a large coalition of senators. Because of the need for high levels of collective action to stop recess appointments, previous attempts to limit presidential use of recess appointments have largely been unsuccessful. In this section we document these previous attempts and examine why they were ineffective.

Relying on the Courts

During the presidency of George W. Bush, Democrats actively filibustered the confirmation of federal appeals court nominee William Pryor, largely because of the conservative reputation he gained while serving as the attorney general of Alabama. A Republican-led Senate coalition fell seven votes shy of invoking cloture to end this filibuster. Following this defeat, Bush circumvented the Senate by recess appointing Pryor, during a ten-day Senate recess in February 2004.

Outraged Democrats, led by Senator Ted Kennedy (D-MA), filed an amicus curiae brief in support of a lawsuit that challenged the legality of Pryor's appointment. Kennedy asserted that it was unconstitutional to make recess appointments during the short intrasession recesses that occur during a congressional session. Eight months later, the Eleventh Circuit Court of Appeals rejected Kennedy's challenge, ruling that the Constitution "does not establish a minimum time that an authorized break in the Senate must last to give legal force to the President's appointment power under the Recess Appointments Clause" (quoted in Jansen 2007). Tom Daschle (D-SD) announced that the Democrats would place holds on all pending judicial nominees unless President Bush promised to refrain from recess appointing other judges. The president ultimately acceded to that request. (13)

Scholars of the president's unilateral powers have argued that presidents have long made use of the ambiguous language in the Constitution to incrementally increase their power relative to the Congress (Moe and Howell 1999b). In theory, the courts could play an important role in clarifying the ambiguous nature of constitutional provisions and limiting presidents' use of vague provisions.

However, this case illustrates why lawsuits tend to be an unpopular way to respond to the growing use of recess appointments or other unilateral powers. First, they are time consuming. Cases must slowly make their way through the courts and the appeals process. Second, relying on the courts can be a risky proposition for Congress. If the courts agree with the views of members of Congress, then the president's use of a unilateral power can be constrained. But if the courts side with the president, then the president, who was relying on his own interpretation of constitutional gray areas, will now be able to act with legal authority in an area now made black and white thanks to the court's ruling. Kennedy's attempt to bar intrasession recess appointments backfired, providing presidents with a clear legal precedent allowing them to make recess appointments during intrasession recesses. Accordingly, relying on the courts is a generally ineffective way to deter presidential use of recess appointments or other unilateral powers.

Threats of Senate Delay

In 1996, President Clinton used his recess appointment power to appoint Sara M. Fox to the National Labor Relations Board. Fox, a former staffer to Senator Ted Kennedy, had developed a reputation as a staunch supporter of labor. Republicans expressed outrage at Clinton's recess appointment. Senator James Inhofe (R-OK) claimed the recess appointment violated the current practice, which required presidents to inform the majority and minority leaders of all recess appointments prior to making them. This practice was outlined in a letter by then Senate Majority Leader Robert C. Byrd (D-WV), responding to recess appointments made by President Reagan. (14) Inhofe lambasted Clinton's decision as "unconstitutional" and "egregious" (Congressional Record, September 11, 2000).

Inhofe's response was not limited to speeches. The senator pledged that he would "place a 'hold' on every judicial nominee which reaches the Senate calendar for the remainder of the Clinton presidency" (Boyer 1999). He concluded that the only way Clinton's nominees would gain confirmation was "over [his] dead body" (Preston 2000). Inhofe secured pledges from 16 additional Republicans to support such a maneuver. However, in February of 2000, President Clinton still maintained high approval ratings and Majority Leader Trent Lott (R-MS) refused to honor Inhofe's holds (Lewis 2000). A test vote revealed the obstructionist Republicans controlled much less than the 41 votes necessary to sustain a filibuster.

Not all threats of delay are as ineffective as in the Fox case. In 1989, George H. W. Bush chose not to recess appoint William Lucas, in part because of threats of retaliation. Lucas had been nominated to head the civil rights division of the Justice Department. Civil rights groups opposed his candidacy, arguing he lacked the legal experience to competently execute the duties of the post (Marcus 1989). The Democratic-controlled Judiciary Committee rejected Lucas' nomination on a 7-7 vote that broke down near party lines. (15) The Republican minority leader, Robert Dole (R-KS), publicly requested that Bush grant Lucas a recess appointment. However, threats of delaying future nominations by majority leader George Mitchell (D-ME) were apparently sufficient in persuading the president to back away from recess appointing Lucas. Instead, Bush nominated him to head the Office of Liaison Services, a post that did not require Senate confirmation (Marcus 1989; Devroy and Dewar 1989.)

In the case of Lucas, the appointment would have occurred relatively early in Bush's presidential term. (16) This made the threat of retaliatory delays by Senate Democrats all the more dangerous. The fact that the threat came from the majority leader, who holds the most important position in terms of legislative scheduling, further served to dissuade Bush from making the recess appointment.

Threats of delay are a potential way for the Senate to convince presidents to reduce their use of the recess appointment clause. If a president knows his use of recess appointments will result in all future nominees facing delay, he may avoid them. However, for the Senate to make good on its threats of delay, a sufficient coalition of at least 41 senators must remain dedicated to use the filibuster to maintain delay. Such collective action is hard to achieve. While the president's partisan foes may be interested in attempting delay, presidential allies are very unlikely to actively work to stop recess appointments. The short-term policy goals of senators in the president's party are likely to be well served by recess appointments, and these short-term considerations are likely to trump any concerns presidential allies have about the long-term growth of the powers of the president. Additionally, broad efforts to delay all presidential nominees are costly, in terms of time and energy consumed. Any time spent delaying presidential nominations means other potentially important issues and legislation are being delayed as well. The fact that retaliation is costly to maintain means that even opponents of a president may be unwilling to sustain delay if it means sacrificing other policy and electoral goals. These factors were certainly in play when Inhofe was unable to garner 41 Republican votes to retaliate against Clinton's recess appointment of Fox, despite the fact that Republicans had 55 seats in the chamber at the time.

Reid and Permanent Session

On November 16, 2007, Reid announced for the first time that the Senate would hold pro forma sessions to "prevent recess appointments" (Hogue and Bearden 2008). These sessions are short meetings in which no business is conducted. If a pro forma session is conducted once every three days, then the Senate is not considered to be in recess (Hulse 2007). By running these sessions during traditional congressional breaks, Reid was able to keep the Senate from technically going into recess and President Bush was prevented from using of his recess appointment power.

Compared to lawsuits and threats of delay, pro forma sessions are an effective and low cost way to prevent recess appointments. Reid, acting as Democratic majority leader, only needs the members of his party to allow him to call these sessions. Unlike threats of delay, other important legislation is not risked by the use of pro forma sessions, and senators do not need to actively take part in filibusters or other dilatory activities, keeping the costs to maintaining an anti-recess appointment coalition low. In fact, the only real cost is that one senator must physically come to the Senate to preside over pro forma sessions. During the 2007 Thanksgiving break, the Senate conducted four pro forma sessions to prevent a formal recess from occurring (Hogue and Bearden 2008). Freshmen Senator Jim Webb (D-VA) presided over the first of these sessions. After calling the chamber to order, Webb, the only senator in the chamber, immediately adjourned the Senate until the next pro forma session. The entire session lasted less than 30 seconds (Hulse 2007).

Projecting Bush's Usage of the Recess Appointment Clause

We argue that Reid's decision to keep the Senate in permanent session was motivated by Bush's increasing use of recess appointments. Recess appointments can have a dramatic effect on the composition of the federal bureaucracy, and in turn, a dramatic effect on the policies generated by the bureaucracy. This potential change in bureaucratic policy making was politically costly to Senate Democrats, giving them a strong incentive to collectively support Reid's elimination of recesses. If Bush's frequent recess appointments drove Democrats to end recesses, then we would expect data on recess appointments to confirm that Bush's usage of the recess appointments exceeded that of his predecessors, even when controlling for all other possible explanations for recess appointments. In this section we demonstrate that the Bush administration's use of recess appointments greatly differed from that of previous presidents.

To test our hypothesis, we present a multivariate analysis of all civilian nominations and recess appointments made between 1987 and 2006. (17) Our dependent variable considers whether a president chooses to fill a vacancy by formally nominating an appointee and seeking confirmation by the Senate (coded as 0) or whether a recess appointment was used to bypass the Senate (coded as 1). Our data contains 414 recess appointments and 9,593 traditional appointments.

To consider how frequently Bush used recess appointments relative to other presidents we include dummy variables in our model for each president (Reagan, Bush, Clinton, and Bush). We also consider if the party controlling the Senate helps explain recess appointments. We include dummy variables for the divided and unified periods of the Clinton and Bush presidencies. (18)

We control for other factors that may influence a president's decision to issue a recess appointment. First, and perhaps most importantly, presidents want to use recess appointments to influence policy (Corley 2006; Black et al. 2007). Thus, major independent boards and commissions are a likely target of recess appointment use. These boards are generally smaller than other agencies, and one or two nominees can greatly sway the boards' ideology. Second, members are usually appointed to fairly short terms (especially compared to judges who enjoy lifetime tenure). Because of this, the one- to two-year tenure of most recess appointees is not a major limitation for presidents seeking to fill those positions. Consistent with Nixon (2009), we code major independent agency as 1 if the agency to which the person was nominated or recess appointed is one of the following: the Consumer Product Safety Commission, Equal Employment Opportunity Commission, Federal Communications Commission, Federal Election Commission, Federal Energy Regulatory Commission, Federal Reserve Board of Governors, Federal Trade Commission, Interstate Commerce Commission, National Labor Relations Board, National Transportation Safety Board, Nuclear Regulatory Commission, and Securities and Exchange Commission. Our data include 390 board vacancies, of which 47 were filled via recess appointment. We also control for other high-profile nominated positions that may be targeted for recess appointment. Specifically, we control for cabinet-level positions, ambassadors, and judges. (19)

We also anticipate that presidents will be more likely to make more recess appointments during certain times in a four year term. Following Black et al. (2007), we anticipate that presidents will make many more recess appointments at the end of their presidencies, especially when they are going to be replaced by a new president from the other party. We control for this lame-duck period, which we define as time period between the November presidential election and the presidential inauguration on January 20, with the use of two dummy variables-one for lame-duck periods preceding a party change and another for lame-duck periods preceding no change in the party of the president. Since opposition in the Senate may try to delay nominations made in an election year, we control for the run up to a presidential election from January of an election year up to the date of the election in November.

Additionally, as the case studies have suggested, dilatory tactics have become commonplace in the Senate. Supporters of President Bush's decisions to make recess appointments, like Senator John Cornyn (R-TX), go so far as to assert the increase in recess appointments is primarily due to increasing use of holds and filibusters in the Senate. Cornyn has argued that in several instances, these holds or filibusters were detrimental to national security, and thus, the recess appointment was necessary (Congressional Record, September 19, 2006). Accordingly, we control for increasing dilatory tactics by measuring the yearly average of days between a nomination's submission and a terminal action by the Senate for the year in which the nomination is made (Corley 2006, Black et al. 2007).

Finally, the president's popularity may play an important role in determining whether or not he will utilize a recess appointment. Corley (2006) finds that presidents utilize recess appointments when they have high levels of political capital. (20) One way to account for available political capital is to control for the president's popularity. We do so by including a president's average monthly Gallup approval rating during the month in which the appointment or nomination was made.

Results

Since recess appointments are rare relative to traditional appointments (about 4% of all appointments), we employ a rare events logistic regression model to analyze our data (King and Zeng 2001). Table 1 reports the results of our model.

As expected, presidents are more likely to make recess appointments to major boards and agencies. There is no statistical evidence that presidents are more likely to make recess appointments to highly visible cabinet positions, and presidents are less likely to make recess appointments to ambassadorships and judgeships than to other positions. Presidents are more likely to make recess appointments during the year of a presidential election, and presidents are more likely to make recess appointments after they had become lame ducks, regardless of whether or not they were being replaced by someone of their political party. Neither presidential approval rates nor delay in the Senate have a significant effect on presidents' decisions to use recess appointments. (21)

To explore the differences in how often presidents use recess appointments, we calculate predicted probabilities for each president filling a vacancy with a recess appointment, accounting for both unified and divided party control of the Senate. These are graphically displayed in Figure 2.

While Bush's use of recess appointments was fairly consistent with that of other presidents when the Democrats controlled the Senate (0.08% for non-board positions and 2.30% for board positions), Bush's use of recess appointments far exceeded his predecessors when Republicans controlled the Senate (2.42% for non-board positions and 6.61% for board positions). This is certainly counterintuitive. If the recess appointment provision is a way for the president to overcome obstruction is the Senate, why would he use it more often when he had greater support in the chamber?

Comparing Bush's use under unified control to Clinton's--the only other president to enjoy a period of unified party control of the presidency and the Senate--is somewhat shocking. Under unified party control, Bush made recess appointments at a rate 12 times greater than Clinton (2.42% versus 0.20%) for nonboard positions and 10 times greater than Clinton (6.61% versus 0.62%) for board positions. This indicates that Bush was far more aggressive in his use of recess appointments than presidents operating in a similar political context. This helps to explain Reid's motivations for blocking Bush's heavy use of recess appointments. Indeed, Bush's use of recess appointments under unified party control exceeded that of all other time periods examined. (22)

Anticipating Bush's Usage for the 110th Congress

To provide context to Reid's decision to move the Senate into permanent session, we think it is important to consider Bush's anticipated rates of recess appointments had the Senate maintained regular recesses throughout the 110th Congress. To do this we simulate predicted levels of recess appointments for three time periods during the 110th Congress: 2007; the run up to the 2008 presidential election from January 2008 to November 2008; and the lame-duck period from November of 2008 to January of 2009. We evaluate the rate of recess appointments Bush would have likely made given divided control of the Senate. The results are reported in Figure 3.

[FIGURE 2 OMITTED]

[FIGURE 3 OMITTED]

The trends are clear. Had Reid not moved the Senate into permanent session, he could have anticipated Bush using his recess appointment power at ever increasing rates. The simulated rates for 2007 (0.85% for non-board positions and 2.35% for board positions) are fairly low relative to Bush's use under unified control of the Senate, but in 2008 Reid could have anticipated a sharp increase in the use of recess appointments. During the run up to the 2008 presidential election, Bush's simulated rates of recess appointments climb to 2.72% for non-board positions and 7.52% for board positions. After the November election, when it became clear that Democratic president-elect Barack Obama would replace Bush, Bush's simulated rates reach 33.80% for non-board positions and 53.89% for board positions. These simulated rates dramatically demonstrate the potential political costs of allowing Bush to continue to make recess appointments in the final two years of his administration. Given the large number of vacancies on independent boards and agencies during this period and the important policy making role exerted by these positions, Bush could have had an enormous impact on the policy made by boards and agencies had he been allowed to make recess appointments to these positions. Democrats had a clear incentive to support Reid's decision to block all future recess appointments by not allowing to Senate to recess.

Discussion and Conclusion

Unilateral powers give presidents a clear advantage in their ongoing struggles with Congress to control the federal bureaucracy, but unilateral powers have limits. Powers such as those granted by the recess appointment clause can only be used as long as Congress allows it. Once Congress is driven to act, unilateral powers can be revoked.

In this article we establish that George W. Bush's use of recess appointments was vastly different than that of previous presidents. Bush used recess appointments far more than other presidents and used them extensively even when his party had control of the Senate. Coming into the last two years of the Bush presidency, Reid considered the pattern of Bush's past recess appointments and foresaw an election and post-election period with a potential party change that would predict even higher rates of usage, as high as 54% of board vacancies. This, coupled with a vast number of vacancies on policy making, independent boards and agencies, gave Senate Democrats all the incentives they needed to reach collective action and lend their support to Reid's plan to move the Senate into permanent session.

Moreover, as Senate majority leader, Reid had the power to change how the Senate conducted recesses and he only needed the support of members of the Democratic caucus to accomplish this. This previously unrecognized institutional feature of the Senate allowed Democrats to directly confront Bush in a way they could not do when they were in the minority. When Republicans controlled the Senate, they had little policy incentive to prevent Bush from using recess appointments, and Democrats were forced to try less effective methods of blocking recess appointments such as Kennedy's unsuccessful lawsuit. Permanent session was much less costly to maintain than were previous attempts to use delay to prevent presidents from making recess appointments.

The results of majority party control of a Senate process, like the scheduling of recesses, had an enormous impact on policy. This practice prevented Bush from filling hundreds of vacancies, including key policy-making positions on the Federal Election Commission, National Labor Relations Board, or Securities and Exchange Commission. While we will never know what the difference in the rulings issued by these boards would have been, had Bush been allowed to use recess appointments to fill these boards, there is little question that the results would have benefited Bush and hurt Senate Democrats.

What does this episode tell us more broadly about interbranch politics? While recent literature has argued the presidents' use of unilateral power has grown and grown over time, it is possible for Congress to check presidential unilateral powers. But this is likely to happen only in rare circumstances. The rules of the Senate allow a single party to decide how often, if at all, recesses will occur. Thus, in the case of recess appointments, party leadership can check the president's power to make recess appointments as long as they are supported by a majority of their party. There is no need to gain the support of senators in the minority party or anyone in the House.

Compare this to the president's power to issue executive orders. If Congress opposes an executive order, a bill overturning it must be written and introduced. It must be passed by both chambers of Congress, requiring at minimum the support of party leadership, key committee chairs, and a bare majority of both houses. If there is a filibuster threat, 60 members of the Senate will have to support the measure. If the president is willing to use his veto power to protect his executive order, two-thirds of the House and Senate will be required to block action by the president. This threshold for collective action is simply too high, and Congress is unlikely to effectively counter executive orders, unless the orders are extremely unpopular. In short, Congress is likely to stop presidential use of unilateral powers only in cases where presidential action is extremely unpopular or in cases, such as that of recess appointments, where the threshold of collective action is relatively low.

What does Bush's use of recess appointments tell us about how future presidents will use recess appointments? Bush's increased use of appointments while his party controlled the Senate seems to give us a clue. We think that future presidents will be more likely to use recess appointment power when their party controls the Senate. This practice will give Obama and other presidents an effective tool to overcome minority party threats of delay. If Obama uses recess appointments, it will likely be with the consent of Senate Democrats. For this next Congress, Republicans will be forced to rely on relatively ineffective tools, such as holds, threats of filibusters, and possible court cases to convince Obama to not make appointments. In the unlikely event that Republicans win control of the Senate majority in 2010, they will likely revert to Reid's policy of permanent session, again preventing presidential use of recess appointments.

AUTHORS' NOTE: We thank Jamie L. Carson, Andrew D. Martin, Gary J. Miller, Steven S. Smith, and James F. Spriggs for comments on earlier drafts of this manuscript. Lynch thanks the Kansas University Institute for Policy and Social Research and Owens thanks the George H. W. Bush Library Foundation for generous financial support of this research. All errors remain our own.

References

Angrist, Joshua A., and Jorn-Steffen Pischke. 2009. Most Harmless Econometrics: A Empiricist's Companion. Princeton, NJ: Princeton University Press.

Berger, Joseph. 2010. "Democrats Defend Appointments." New York Times, March 29.

Billings, Erin P. 2007. "Reid Mulls Pro Forma Sessions." Roll Call, November 15.

Binder, Sarah A., and Forrest Maltzman. 2002. "Senatorial Delay in Confirming Federal Judges, 1947-98." American Journal of Political Science 46: 190-99.

Binder, Sarah A., and Steven S. Smith. 1997. Politics or Principle? Filibustering in the United States Senate. Washington, DC: Brookings Institution Press.

Black, Ryan C., Anthony J. Madonna, Ryan J. Owens, and Michael S. Lynch. 2007. "Adding Recess Appointments to the President's 'Tool Chest' of Unilateral Powers." Political Research Quarterly 60 (4): 645-54.

Boyer, Dave. 1999. "GOP Senator Vows to Block Clinton Picks." Washington Times, December 21.

Calvert, Randall L., Mathew D. McCubbins, and Barry R. Weingast. 1989. "A Theory of Political Control and Agency Discretion." American Journal of Political Science 33: 588-611.

Carrier, Michael A. 1994. "When Is the Senate in Recess for Purposes of the Recess Appointments Clause?" Michigan Law Review 92: 2204-47.

Cooper, Phillip. 2002. By Order of the President: The Use and Abuse of Executive Direct Action. Lawrence: University Press of Kansas.

Corley, Pamela. 2006. "Avoiding Advice and Consent: Recess Appointments and Presidential Power." Presidential Studies Quarterly 36: 670-80.

Curtis, Thomas A. 1984. "Recess Appointments to Article III Courts: The Use of Historical Practice in Constitutional Interpretation." Columbia Law Review 84 (7): 1758-91.

Devroy, Ann, and Helen Dewar. 1989. "Mitchell Opposes Recess Appointment." Washington Post, August 5.

Garcia, Rodelio. 2001. "Recess Appointments Made by President Clinton." Congressional Research Service Report RL30821.

Hammond, Thomas H., and Jeffrey S. Hill. 1993. "Deference or Preference? Explaining Senate Confirmation of Presidential Nominees to Administrative Agencies." Journal of Theoretical Politics 5: 23-59.

Halstead, T. J. 2005. "Recess Appointments: A Legal Overview." Congressional Research Service Report RL33009.

Hogue, Henry B. 2007. "Recess Appointments: Frequently Asked Questions." Congressional Research Service Report RS21308.

Hogue, Henry B., and Maureen Bearden. 2008. "Recess Appointments Made by President George W. Bush, January 20, 2001-October 31, 2008" Congressional Research Service Report RL33310.

Howell, William G. 2003. Power without Persuasion: The Politics of Direct Presidential Action. Princeton, NJ: Princeton University Press.

Hulse, Carl. 2007. "Democrats Move to Block Bush Appointments." New York Times, November 21. Jacoby, Mary. 1993a. "Is It Political or Legal? That's What Court Must Decide in the Fight over Recess Appointments." Roll Call, July 22.

--. 1993b. "President Limited in Ability to Bypass Hill on Appointees." Roll Call, July 29.

Jansen, Bart. 2007. "Wary of Recess Appointments, Senate Democrats Weigh August 'Pro Forma' Sessions." CQ Today, August 2.

King, Gary, and Langche Zeng. 2001. "Logistic Regression in Rare Events Data." Political Analysis 9: 37-63.

Kosuke, Imai, Gary King, and Oliva Lau. 2007. "relogit: Rare Events Logistic Regression for Dichotomous Dependent Variables" in Kosuke Imai, Gary King, and Olivia Lau, "Zelig: Everyone's Statistical Software," http://gking.harvard.edu/zelig.

Krutz, Glen L., Richard Fleisher, and Jon R. Bond. 1998. "From Abe Fortas to Zoe Baird: Why Some Presidential Nominations Fail in the Senate." American Political Science Review 92: 871-81.

Lewis, Neil A. 2000. "2 Clinton Judicial Choices Are Approved After Lott Forces Votes." New York Times, February 11.

Marcus, Ruth. 1989. "Panel Kills Nomination of Lucas." Washington Post, August 2.

Mayer, Kenneth. 2001. With the Stroke of a Pen: Executive Orders and Presidential Power. Princeton, NJ: Princeton University Press.

McCarty, Nolan, and Rose Razaghian. 1999. "Advice and Consent: Senate Responses to Executive Branch Nominations 1885-1996." American Journal of Political Science 43: 1122-43.

Moe, Terry, and William Howell. 1999a. "Unilateral Action and Presidential Power: A Theory." Presidential Studies Quarterly 29 (4): 850-72.

--. 1999b. "The Presidential Power of Unilateral Action." Journal of Law, Economics, and Organization 15 (1): 132-79.

Moraski, Bryon J., and Charles R. Shipan. "The Politics of Supreme Court Nominations: A Theory of Institutional Constraints and Choices." American Journal of Political Science 43 (4): 1069-95.

Mosk, Matthew. 2008. "Vacancies on FEC Filled as 5 Win Senate Approval." Washington Post, June 25.

Murray, Matthew. 2008. "Nominations Fight Freezes McCain's Cash Haul." Roll Call, May 27.

Nixon, David C. 2009. The Independent Regulatory Commissioner Database. http:// www2.hawaii.edu/dnixon/IRC.

Nokken, Timothy P., and Brian R. Sala. 2000. "Confirmation Dynamics: A Model of Presidential Appointments to Independent Agencies." Journal of Theoretical Politics 12: 91-112.

Preston, Mark. 2000. "Inhofe Will Block All Clinton Judicial Nominations." Roll Call, January 10.

Rappaport, Michael B. 2005. "The Original Meaning of the Recess Appointments Clause." UCLA Law Review. 52: 1487-578.

Snyder, Susan K., and Barry Weingast. 2000. "The American System of Shared Powers: the President, Congress, and the NLRB." Journal of Law, Economics, and Organization 16: 269-305.

Tillman, Seth Barrett. 2009. "Terminating Presidential Recess Appointments: A Reply to Professor Brian C. Kalt." Northwestern University Law Review Colloquy 103: 268-91.

York, Michael, and Stephan Barr. 1993. "Bush Appoints Supporter In Postal Board Dispute." Washington Post, January 9.

RYAN C. BLACK

Michigan State University

MICHAEL S. LYNCH

University of Kansas

ANTHONY J. MADONNA

University of Georgia

RYAN J. OWENS

University of Wisconsin-Madison

(1.) Schumer was commenting on Bush's 2004 recess appointment of William Pryor to the Eleventh U.S. Circuit Court of Appeals.

(2.) With only the support of Senate Democrats, Reid was able to change the Senate's recess practices. Reid replaced traditional recesses with pro forma sessions, such that the Senate was technically in session but did not conduct any business.

(3.) This impasse lasted for six months, from January to June 2008 (Mosk 2008).

(4.) The National Labor Relations Board had only two members from January 1, 2008, until President Obama's recess appointment of Craig Becker in March 2010 (Berger 2010). On June 17, 2010, the Supreme Court ruled that three members were required to render decisions (see New Process Steel v. National Labor Relations Board, 130 S. Ct. 2635 (2010)).

(5.) Senate rules state that "Nominations neither confirmed nor rejected during the session at which they are made shall not be acted upon at any succeeding session without being again made to the Senate by the President; and if the Senate shall adjourn or take a recess for more than thirty days, all nominations pending and not finally acted upon at the time of taking such adjournment or recess shall be returned by the Secretary to the President, and shall not again be considered unless they shall again be made to the Senate by the President" (The Rules of the Senate, Rule XXXI, clause 6).

(6.) Obstruction of executive branch nominations are more commonly referred to as "holds." The hold is essentially a threatened filibuster, where a senator notifies the majority leader he or she intends to obstruct the nomination. Overcoming a hold requires the majority leader to schedule the final Senate vote on the nomination, forcing the filibuster. Supporters of the majority leader would then need to invoke cloture (Binder and Smith 1997).

(7.) Perhaps not surprisingly, defining when the Senate is in recess for purposes of making an appointment has been the subject of considerable discussion (Carrier 1994; Curtis 1984). While it seems clear that the general intent behind the recess appointment power was to allow presidents to install nominees to important government positions while the Senate took its lengthy (six to nine month) intersession recesses (Carrier 1994; Rappaport 2005), presidents in recent times have made extensive use of the recess appointment power during intrasession recesses. There is disagreement as to how long an intrasession recess must be before the president credibly can make a recess appointment. Some have argued that a duration of 5-10 days is not long enough (Halstead 2005), and they seem to have the support of the courts. For example, in Kennedy v. Sampson, 511 F.2d 430 (D.C. Cir. 1974), the Circuit Court of Appeals for the District of Columbia prohibited the president from using his pocket veto for an intrasession recess appointment that lasted only six days. In response, President Carter's Department of Justice advised him that intrasession recesses must last at least 30 days to allow a valid recess appointment. Recent presidents made recess appointments during short intrasession recesses, however. George W. Bush made an appointment after only seven days of recess, which was later upheld by the Eleventh Circuit Court of Appeals (see Evans v. Stephens, 387 F.3d 1220 (11th Cir. 2004)). It should also be noted that while the constitutional language appears to suggest that the vacancy must happen (i.e., arise) when the Senate is in recess, federal courts and attorneys general have interpreted this language to mean simply that a vacancy must happen to exist while the Senate is in recess. See, for example, United States v. Woodley, 751 F.2d 1008 (9th Cir. 1985); United States v. Allocco, 305 F.2d 704 (2d Cir. 1962); 10p. Att'y Gen. 631 (1823); 12 Op. Att'y Gen. 449 (1868); 41 Op. Att'y Gen. 80 (1960).

(8.) Recess appointments made during an intrasession recess can serve longer than appointments made during an intersession recess, as the intrasession recess appointee may serve the duration of the principle session, plus the rest of the next. The intersession recess appointee, however, only serves the duration of the principle session.

(9.) One alternative available to presidents would be to name an acting official to a particular job. If the president has reason to believe that acting official would serve his goals, this might be a reliable strategy. Yet, because presidents arguably know with more certainty the preferences of their recess appointees, they are more likely to avoid shirking by such agents.

(10.) Under federal law, if the vacancy occurred while the Senate was in session, the recess appointee will not receive compensation until confirmed by the Senate, unless certain specified conditions are met. For more on these salary restrictions, see Hogue (2007, 4). Similarly, if the Senate rejects the formal nomination of a recess appointee, the appointee can continue to serve until the completion of the term but will not be paid. Federal law also prevents the payment to an individual appointed to the same job consecutively (Hogue 2007).

(11.) One author goes so far as to say that if the Senate wants to terminate recess appointees, it could simply convene, terminate its session, and then immediately start up a new one (Tillman 2009).

(12.) This increase is especially pronounced for controversial intrasession appointments. Both Corley (2006) and Black et al. (2007) cite contemporary accounts suggesting that the usage of intrasession recess appointments has increased sharply in recent years. President George W. Bush drastically increased the presidential usage of the recess appointment clause during intrasession recesses. Presidents Clinton and George H. W. Bush also frequently utilized this mechanism. By contrast, presidents made only two intrasession recess appointments prior to 1947 (Carrier 1994).

(13.) President George H. W. Bush's recess appointment of Thomas Ludlow Ashley to the U.S. Postal Service Board also prompted court action by members of the Senate (York and Barr 1993). Ashley was to fill the seat of board member Crocker Nevin, who had been serving an expired term in a holdover capacity. It was unknown at the time whether the president could vacate a position being held in holdover capacity by recess appointment. The issue on intrasession recess appointments was also brought up in the legal challenge. Senate Democrats--then the majority party--supported a lawsuit brought by Nevin and other board members (Jacoby 1993a). The district court's ruling was ambiguous on the issue of intrasession appointments, instead opting to vacate the Ashley appointment on the grounds that the president could not use the recess appointment power to fill a seat occupied by a member serving in holdover capacity (Jacoby 1993b).

(14.) Byrd wrote that "recess appointments should be limited to circumstances when the Senate, by reason of protracted recess, is incapable of confirming a vitally needed public officer. Any other interpretation of the recess appointments clause of the Constitution could be seen as a deliberate effort to circumvent the constitutional responsibility of the Senate to advise and consent to such appointments" (Boyer 1999).

(15.) Only Dennis DeConcini (D-AZ) joined the Republicans.

(16.) In 2000, Clinton used his power to place Roger Gregory on the Fourth Circuit Court of Appeals. Senator Jessie Helms (R-NC), who had blocked Gregory's appointment on the Senate floor previous to the recess appointment, threatened to retaliate. However, unlike the Lucas appointment, the Senate term was nearing an end, making potential retaliation less harmful to the president. Republican support for retaliation was low and Helms' attempts at delay were not effective.

(17.) Our nominations data come from the Presidential Nominations page of the Library of Congress THOMAS web site, http://thomas.loc.gov/home/nomis.html. Using a commercially available program (QuicKeys), we wrote a macro to automate the data-collection process. In particular, the macro iteratively searched the THOMAS database for all civilian nominations across each Congress and then downloaded these results. The outcomes in THOMAS for each nomination are: (1) Confirmed, (2) Rejected by Vote, (3) Expired and Rejected, and (4) Withdrawn. We downloaded all civilian nominations for each Congress using no other search restrictions. We then manually cleaned these downloaded results as follows: incomplete records (e.g., those lacking nomination dates or position information) were purged from our data. Nominations submitted in groups had two records: one for the group nomination and then one for each individual. These duplicate records were deleted. We used two sources to obtain our data for recess appointments. For Presidents Reagan and George H. W. Bush we examined the appendices of various volumes of the Public Papers of the President, which we retrieved from HEINonline, http://heinonline.org. Data for Presidents Clinton and George W. Bush were considerably easier to obtain, as reports published by the Congressional Research Service provided these data in tabular form (Garcia 2001; Hogue and Bearden 2006).

(18.) Previous studies of recess appointments have controlled for the ideological distance between the president and the Senate (Black et al. 2007; Corley 2006). Since controlling for divided and unified government also serves to provide a control for the ideological differences between the Senate and the president, we do not include a measure for ideological distance in these analyses.

(19.) We define cabinet-level positions by using all jobs that fall into Level I of the federal executive pay schedule. These include most cabinet-level positions and a few other positions, such as the chairman of the Federal Reserve and the director of National Intelligence.

(20.) Even popular presidents have difficulty getting their nominees through the Senate. This is due to a combination of factors including, but not limited to, the large number of positions that necessitate chamber approval, supermajoritarian debate rules and procedures that allow individual senators to hold nominations without revealing their identities.

(21.) While this article finds no effect for presidential approval, Corley (2006) finds the expected positive effect, and Black et al. (2007) find a negative effect. The insignificant result for Senate delay runs counter to both Corley's (2006) and Black et al.'s (2007) results.

(22.) The difference between Bush's and Clinton's use under unified control is significant at the 95% confidence level.

Ryan C. Black is an assistant professor of political science at Michigan State University. His recent work has appeared in such journals as the Journal of Politics, the Journal of Law, Economics, & Organization, and Political Research Quarterly.

Michael S. Lynch is an assistant professor of political science at the University of Kansas. His work has appeared in Urban Affairs Review, Political Analysis, Political Research Quarterly, and the Journal of Politics.

Anthony J. Madonna is an assistant professor of political science at the University of Georgia. His work has appeared in such journals as the American Journal of Political Science, Political Research Quarterly, and Perspectives on Politics.

Ryan J. Owens is an assistant professor of political science at the University of Wisconsin-Madison. His research has appeared in such journals as the American Journal of Political Science, the Journal of Politics, Political Research Quarterly, and Judicature.

TABLE 1
Rare Events Logistic Regression of Recess Appointments

                                Covariate   Standard Error

Bush-Unified                     0.658 *        0.300
Bush-Divided                    -0.422          0.348
Clinton-Divided                 -0.185          0.320
Clinton-Unified                 -1.978 *        0.649
H.W. Bush                       -0.210          0.269
Major Board                      1.042 *        0.237
Executive Level I               -0.937          0.634
Ambassador                      -0.932 *        0.248
Judiciary                       -2.524 *        0.947
Lame Duck-New Party              4.137 *        0.294
Lame Duck-Same Party             2.264 *        0.284
Election Run-Up                  1.215 *        0.283
Monthly Presidential Approval    0.001          0.011
Average Yearly Delay             0.004          0.005
Constant                        -4.471 *        1.075
N = 10,007
Null Deviance = 3448.0
Residual Deviance = 2732.7

Note: This rare events logistic regression model was estimated
using Zelig statistical software (Kosuke, King, and Lau 2007).
We estimate both conventional standard errors and robust standard
errors and report the larger robust standard errors (Angrist and
Pischke 2009).

* signifies p :5 .05 (two-tailed test). President and
Senate-control dummies are relative to the baseline of the
Reagan administration under divided control.