The New Separation‐of‐Powers Approach to American Politics
Abstract and Keywords
This article aims to elaborate on the logic of the new separation-of-powers approach and draw its implications for American politics. The first three sections in the article discuss the new separation of powers as it applies to the bureaucracy, the courts, and the presidency. A survey of a series of works that emphasize the new separation-of-powers approach to American politics is provided in the article as well.
American politics is undergoing a quiet revolution. The study of American politics has long been organized by sub‐fields—Congress, the presidency, the courts, and the bureaucracy. The advantage of this approach is that it allowed political scientists to gain expertise on a relevant portion of American politics and to create the scientific study of each institution. The liabilities of this approach are equally (p. 200) clear: it has created somewhat of a silo approach to American politics in which discoveries and knowledge in one sub‐field or silo tended to have little impact on the others.
Since 1990, a new approach to studying American politics has emerged based on rational choice methods that we call the “new separation‐of‐powers approach.” Often more implicitly than explicitly, the traditional approach to American politics emphasized separation of powers as the independence of the branches. In contrast to the traditional approach, the new one explicitly studies the interaction of the institutions, incorporating a broader range of actors and embedding them in models of strategic interaction. The results are dramatic. Scholars demonstrate that we cannot fully understand the behavior of one institution without understanding it in the context of the others.
Although traditional scholars always recognized the effect of one institution on another, they had no way to provide a systematic characterization of these interactions, many of which were instead characterized as caveats on the more general operation of the primary institution. In contrast, the new approach sees the constraints posed by other institutions as the foundational issue that defines the context, and thus the nature, of an institution's power to act.
The purpose of this chapter is to exposit the logic of the new approach and then to draw its implications for American politics.1 Although still in its early stages, the new separation‐of‐powers approach has enhanced both theoretical and empirical analyses. Essentially, it allows scholars at once to study interlinked phenomena occurring in multiple institutions, which had previously not been systematically considered.
The new approach embeds institutional actors in the political environments that impose systematic constraints on them. For example, rather than seeing the president as largely unconstrained, due to her broad powers, the new approach formalizes the effect of the president's need to account for congressional authority and preferences. This move is important because it implies that the characteristics and behavior of outside actors are as important as those of actors inside an institution in explaining variations in the behavior of the institution under study.
Moreover, the new approach demonstrates that these constraints can best be understood in a strategic context: to further their own goals, actors in each branch must anticipate the reactions of actors in the other branches. Although students of American politics have always known this, the new models provide the game‐theoretic technology for studying how this occurs. For example, the judiciary is not seen in isolation, even for judges with lifetime tenure. In the case of judicial determination of the meaning of a statute, judges will take the likely congressional responses into account. Because being overturned by Congress is costly, judges have an incentive to act strategically in their decisions so as not to be overturned; and this requires making decisions more favorable to Congress.
Another potential for the approach involves its implications for empirical analysis. The two biggest problems that hinder students of American political institutions are: too few degrees of freedom and confounds to causality—including selection bias, spuriousness, and simultaneous causality. As we show below, the new approach has the potential to address both of these problems.
Lack of sufficient variation is a fundamental problem in the study of many American political institutions. Consider the case of the presidency: the United States has had only forty‐four presidents. This creates a “degrees of freedom problem:” the number of relevant variables affecting presidential choice vastly exceeds the number of cases, making it difficult to make inferences about presidential behavior. The approach provides an elegant solution to this problem in two ways. First, it increases the number of cases by reconceptualizing the independent variable; many presidents face variations in their environment, such as the partisan composition of the Congress. Thus, the unit of analysis becomes not just a presidency, but a president–Congress combination. Second, by incorporating other institutions, it allows scholars to study interactions that happen over and over—in the case of the president these include appointments, budgets, and vetoes.
More generally, the new separation‐of‐powers approach has helped authors unpack causality. This is particularly difficult for students of political institutions, since causality tends to flow both to and from the institution and other variables. In most cases, to identify the unique causal effects of a particular variable under study, the solution is to look for other information, outside the phenomenon being studied. The embrace of separation‐of‐powers models allows students to do this by introducing structure—the constraints that institutions put on public officials in making policy—and strategy—the way in which the actors involved condition their behavior on that structure as well as other players' actions or, importantly, potential actions. The introduction of structure imposes constraints, which in turn help identify causal effects.
The ability to understand causality has resulted in repeated demonstrations that appearances in American institutional politics can be deceiving. Studying an institution in isolation tends to make it appear more important and central to decision‐making than studying it within the separation‐of‐powers system. The reason is that, when studying one institution in isolation, it is often easy to miss how behavior in that institution anticipates reactions from another (Weingast and Moran 1983).
For example, many observe the bureaucracy make choices without any systematic political input from elected officials, and infer that the bureaucracy therefore chooses the policy it wanted without much influence by elected officials. These scholars rely on an internalist bureaucratic logic that focuses on the organization itself to explain bureaucratic choices (e.g. Derthick and Quirk 1985; Wilson 1980). The separation‐of‐powers approach, in contrast, embeds the bureaucracy in a political context and argues that bureaucracies face a complex incentive system. Bureaucratic officials understand these incentives and, anticipating the likely reaction of interest groups and political officials, make policy choices.
This chapter proceeds as follows. The next three sections discuss the new separation of powers as it applies to the presidency, the courts, and the bureaucracy. Our conclusions follow after a penultimate, synthetic section.
2 The Presidency
As noted above, the traditional approach to the presidency tends to study the president in isolation. In a literal sense, this claim is false. Every case study of the president contains major actors and institutions outside the president: the president constantly struggles not only with other institutions—Congress, the courts, and the bureaucracy—but also others—interest groups, the media, and foreign countries. Per the degrees of freedom problem, studying cases implies too many relevant independent variables, far larger than the number of cases. Similarly, studying divergent cases makes it harder to infer the systematic effect of particular institutions on the presidency. Indeed, because the characteristics of the other institutions and the environment exhibit so much variation, the lack of a method of accounting for these differences has obscured systematic influences.
The new separation‐of‐powers approach to the presidency resolves this problem by explicitly embedding the study of the president in a spatial model that allows us to study systematically her interaction with other institutions, typically in the context of something the president does again and again. This approach mitigates the degrees of freedom problem by studying multiple instances of the same category of behavior.
Consider the literature on budgets. Historically, the literature on budgeting differs from other literatures on the presidency in that the interaction of Congress and the president has always been central (e.g. Fisher 1975; Wildavsky 1964).
The advantage of the new separation‐of‐powers approach is that spatial models and other techniques provide a method of accounting for differences in the other institutions and the environment so that more systematic inferences can be made. For example, consider Kiewiet and McCubbins's asymmetric veto model (1988). They demonstrate a remarkable asymmetry in the presidential veto threat that had previously not been understood as a major principle of presidential power in budgeting: the veto threat works only when the president wants a budget that is less than that wanted by the median in Congress.
To see the logic of this conclusion, consider the following spatial model for a particular budget problem (Figure 11.1). The line represents the range of budgets, from 0 on the left to very large budget on the right.
R is the reversion policy that goes into effect if Congress and the president cannot agree on a budget.2 P is the president's ideal policy and C is the ideal policy of Congress (presumably that of the median member of Congress). P(R) represents that policy to which the president is exactly indifferent with the reversion point, implying that the president prefers all policies in between P and P(R) to either P or P(R).
If Congress proposes a budget at its ideal policy of C, the president will veto it. C is outside of the set of budgets (R, P(R)) that the president prefers to R, so she is better off with a veto that yields R = 0 than with C. Because Congress wants a significantly higher budget than the president, the best it can do is propose P(R), which the president will accept. As seen from the figure, the president's veto threat is effective and forces Congress to adjust what it proposes. Moreover, because Congress can anticipate the policies that the president will accept or veto, no veto occurs. Despite the absence of an observable veto or even an explicit threat by the president, the veto constrains congressional decision‐making: they are forced to pass the budget P(R) instead of their preferred C.3
Now suppose that the relative preferences of Congress and the president are reversed, where the president wants a significantly higher budget than the Congress, as illustrated in Figure 11.2.
In this situation, the president's threat to veto a budget that she feels is too small is not credible. Suppose Congress passes its ideal budget, C. The president may threaten to veto the budget if Congress fails to pass a budget that is higher than C, but this threat is vacuous. If the president vetoes C, she winds up with R, which is significantly worse than C. Hence the president will accept C.
McCubbins (1990) provides a variant on this logic to explain an odd feature of the Ronald Reagan presidency. Reagan sought to transform the relationship of the government and the economy by “getting the government off the back of the American people.” As part of this effort, he sought a dramatic reduction in the budget for domestic programs. In collaboration with his director of the Office of Management and Budget, David Stockman, Reagan's proposed budget cuts in 1981 were legendary and helped solidify his reputation as cutting the government. Despite appearances, domestic spending increased dramatically under Reagan after 1981. Surely Reagan did not intend this result, so how did it happen?
Table 11.1 Party preferences
McCubbins provides a persuasive interpretation that rests on a variant of the above and the observation that Reagan faced divided government: although Reagan was a strong president and had the support of a Republican majority in the Senate, the Democrats retained control of the House. Both parties therefore held a veto over policy‐making.
With respect to policy content, Democrats wanted higher domestic spending and lower defense spending, while Reagan and the Republicans wanted the reverse: lower domestic and higher defense spending. When faced with a trade‐off, both parties would settle for increasing both policies rather than retaining the status quo.
Table 11.1 represents the preferences for both Republicans and Democrats. Domestic programs are represented by the letter d; defense programs by r. A capital letter indicates an increase in the program while a lower‐case letter indicates a decrease. Q is the status quo. Democrats prefer first (D, r); next (D, R); then the status quo, Q; next (d, r), and last (d, R). Republicans' preferences are similar, mutatis mutandis.
Because the two houses of Congress were divided, Congress could not act as one and present the president with a fait accompli budget. Instead, Democrats in the House had to compromise with Republicans in the Senate to produce legislation for the president to sign.
McCubbins argues that this implies the only equilibrium outcome from Congress is that both programs are increased. To see this, consider the preferences in Table 11.1. As is readily apparent, the only outcome that both Democrats and Republicans prefer to the status quo is (D, R), increasing both programs. This is the only budget that can pass both houses of Congress. As to President Reagan, he could threaten to veto such a budget unless Congress passed one that lowers domestic spending while increasing defense spending, but the veto threat is not credible. Because the Democrats in the House prefer the status quo (Q) to lowering domestic and increasing defense spending (d, R), they were better off passing (D, R) and letting Reagan veto it than giving in to Reagan. Despite threats to veto increases in both, Reagan would not do so because of a variant of the above logic. Because he prefers increases in both (D, R) to the status quo (Q), Reagan's veto threat is not credible. Reagan's rhetoric about cutting domestic spending aside, he regularly signed budgets increasing domestic spending.
As a second illustration of the separation‐of‐powers approach, we consider recent models of appointments (Gely and Spiller 1992; Binder and Maltzman 2002; Cameron, Cover, and Segal et al. 1990; Chang 2003; Jacobi 2005; Moraski and Shipan 1999; Nokken and Sala 2000; Snyder and Weingast 2000). Traditional models of the (p. 205) presidency tended to consider appointments the purview of the president despite the fact that formal rules which give the power to nominate to the president also explicitly require the “advice and consent” of the Senate. The reason is that, in general, few nominations are rejected, seeming to imply a norm of congressional deference so that the president typically gets her way. Indeed, many have claimed that because senatorial consideration of appointments is often perfunctory, the Senate exerts little influence or power over them. This view implies that the president is unconstrained in her appointments and will therefore choose the nominee that best furthers her goals.
The new separation‐of‐powers models of the appointment process yields a different interpretation, focusing on both the presidential nomination and the Senate approval of appointments. The most obvious implication is that, because the Senate has a veto, the president is forced to take senatorial preferences into account. Per the observational equivalence, an alternative explanation for the observation that the Senate rarely rejects a nominee is that the president has taken the possibility of rejection into account and never puts forth a nomination that will be rejected. Because rejection is highly embarrassing and can make the president seem weak, the president will seek to avoid rejection, and this forces her to take into account senatorial preferences. Indeed, presidents often vet candidates with the committee prior to announcing a nomination.
Consider a simple model of appointments to a regulatory body or multimember court, such as the Supreme Court (Snyder and Weingast 2000). The president and the Senate have differing preferences over policy. Appealing to a standard bargaining framework, we assume that policy arises in a two‐step process. First, elected officials bargain to produce a target policy; second, they implement the policy through a regulatory board who choose policy by voting. In the first stage, elected officials produce some form of compromise over policy; if not split the difference, at least something for both sides. For simplicity, assume that the relevant policy concerns a regulatory agency whose policy choice we designate R.
Figure 11.3 illustrates this compromise by assuming a president on the left and a Senate on the right. Their compromise regulatory policy is R, perhaps biased toward the president, but a compromise nonetheless.
To implement their target policy, elected officials must appoint the regulatory board so that the median board member's preferences correspond to R. We illustrate this in Figure 11.4 for a five‐member board with members whose ideal policies are located at 1, 2, M, 4, and 5. When the median's ideal is located at the elected officials' compromise policy of R, the board implements the elected officials' compromise policy.
The importance of this model is that it affords comparative static results predicting the effects on regulatory appointments and policy after an election that changes the (p. 206) identity and policy preferences of the elected officials. Suppose, for example, that the new president is on the right, replacing the old one on the left. The new president and Senate bargain to form a new compromise between their ideal points, say in between board members 4 and 5 at policy C.
Most independent regulatory agency board members serve for a fixed term rather than at the pleasure of the president, and can only be removed under circumstances of gross misconduct. Thus, a president–Senate combination that wants to move policy to C cannot do so by firing members from the current board. Instead, they must wait until current members resign, die, or complete their term.
This model makes predictions about how every appointment will affect the median—move to the right, stationary, or to the left. Suppose that board member 4 is the first to leave the board. No matter what a nominee's preferences are, the president cannot move the median to the right: the three leftmost board members are fixed, implying that so too is the median. If, instead, member 2's term is up, then the elected officials can move policy from the current median to the ideal of member 4 by appointing a new board member to the right of 4. This is illustrated in Figure 11.5
If elected officials appoint a new member at A to the right of 4, then 4 becomes the new median board member, and regulatory policy will move from the old policy, R, to the new policy 4. Elected officials can move policy to their target policy of C when they have the opportunity to replace one of the now three leftmost members—members 1, old M, or 4—with someone to the right of C. In general, after three appointments on a five‐member board, a president–Senate combination can move the median anywhere. This implies that after three appointments, the model predicts that the president and Senate have reached their target policy so that further appointments need not be used to move the median.
Snyder and Weingast apply this model to the National Labor Relations Board (NLRB), a major agency making labor regulatory policy. They test this model's predictions by examining every regulatory appointment from 1950 through 1988, a total of forty‐three appointments. Given some assumptions about presidential–Senate preferences, the model yields a prediction about how a new appointment will move the median—to the left, none, or to the right.
Table 11.2 Regulatory appointments to the NLRB
Predicted sign of change
Actual sign of change
Source: Snyder and Weingast 2000, table 2.
Table 11.2 summarizes the results. As can be seen, most appointees are as predicted: when the predicted sign change of the median is negative (a move to the right), 60 per cent of the actual appointees move the median to the left; when the predicted sign is 0, two‐thirds are as predicted; and when the predicted sign is positive, 60 per cent are as predicted. Moreover, using a nested‐regression framework, they test this model against both the null model and the presidential dominance hypothesis (predicting that only the president's preferences matter). The separation‐of‐powers model outperforms both the presidential dominance and null models.
In short, the model shows the importance of the interaction of the president and Senate with respect to appointments. Both branches matter, and the president must take the Senate's preferences into account when making nominations, lest she risk failure.
The new separation‐of‐powers approach provides powerful new insights into presidential behavior. The use of spatial models and other techniques affords an accounting system that allows us to keep track of different parameters in the political environment, in turn, allowing us to say how presidential behavior varies with the political environment. The models of veto behavior, budgets, and appointments all show the systematic influence of Congress on the president. To further her goals, the president must anticipate the interaction with Congress.
3 The Courts
As with the other branches, the courts operate within, and constitute part of, the political system: the policy positions of Congress, the president, and the bureaucracy constrain judges in their policy‐making role. The new separation‐of‐powers analysis shows how the formal constraints that operate on the judiciary force it to consider (p. 208) the likely responses of the other institutional players to its decisions. For instance, being overturned by Congress is institutionally costly to the courts, as overrides make the courts appear weak, lower their legitimacy, and waste judicial resources. As such, we expect courts to make their decisions in a way that avoids congressional override. Understanding the powers and preferences of the elected branches provides central information in predicting judicial action.
This is not a unidirectional effect: the courts also constrain the other players in separation‐of‐powers games. Because judicial action shapes policy outcomes, Congress, the president, and agencies will anticipate court decisions, and the potential for judicial review will be taken into account during the law‐making process. Just as courts prefer not to be overruled by Congress, Congress generally prefers not to have its legislation struck down or altered by the courts. If Congress cannot force the judiciary to adhere to its own policy preferences, then its members must take judicial preferences into account when they write legislation. Consequently, the position of the judiciary shapes the behavior of Congress, the president, and bureaucratic agencies. These two effects together show how judicial interactions with the other branches of government shape the application of law and limit the power of Congress and the executive.
The new separation‐of‐powers analysis treats judicial decisions not as one‐shot cases determined by their idiosyncratic characteristics, but as repeated iterations of interactions between the branches. Whether a court reviews the constitutionality of legislation, the interpretation of a statute, or an administrative decision, judicial action is always subject to the responses of other political bodies. This interaction offers a means of predicting judicial decision‐making behavior, and accounting for its variation. Central to this analysis is the recognition that judicial action takes place within the context of a political environment that will react to, and anticipate, judicial action.
Judicial literature has long recognized that the judiciary is subject to responses from the elected branches (e.g. Dahl 1957; McCloskey 1960; Rosenburg 1991). In particular, traditional legal literature emphasized judicial vulnerability to the elected branches of government, through their control of Federal Court jurisdiction, the threat of impeachment, and control over budget and appointments. However, prior to the new separation‐of‐powers literature, these formal constraints were never modeled in terms of their effect on judicial decision‐making. Rather, judicial scholarship on how judges make decisions was framed by a debate between traditional legal scholars, who emphasized the role of judicial character in the voluntary exercise of self‐restraint (e.g. Bickel 1986; Fuller 1978), and political science's attitudinalists, who empirically established that judicial decisions were strongly correlated with individual characteristics of judges, such as the party of the appointing president (e.g. Segal and Spaeth 1993).
This debate was informative about what determines judicial preferences, but it told us little of what constraints operated on the judiciary, given those preferences. In particular, it gave little consideration to the position of the other institutional players in determining the outcome of cases. Figure 11.6 illustrates this point, in stylized form. (p. 209) The line represents a range of possible policy choices and judicial outcomes. C, P, and J represent the locations of the ideal policies for Congress, the president, and the court, respectively.
According to the attitudinalists, cases will be decided at J, the point that represents the preferences of the judiciary. According to traditional legal scholarship, judges will do their best to make determinations at point L, the exogenously determined “best” legal outcome, regardless of its relationship to J. For the attitudinalists, the other players, Congress (C) and the president (P), are relevant only in terms of the correlation that can be anticipated between P (and to some extent the Senate) and J, due to the appointment power; whereas for the traditional legal scholars, P and C may be somewhat informative as to the limits of public acceptability, which informs judicial legitimacy, but not in a formally predictable way.
Although new separation‐of‐powers models provide an approach to the appointment process that determines the position of J (as we discussed in the previous section), most such models of the courts begin with a variant of the attitudinalist model: that judges have a consistent set of goals based on ideology, political preferences, and broader values. The new separation‐of‐powers approach poses new questions, particularly: taking the position of J as given, how will judges decide issues? And how will those decisions be affected by the positions of the other institutional players?
The first step in answering these questions is to set aside the idea that the judiciary is the unconstrained last mover, or the last mover who is constrained only by internal norms. Although legislative overrides occur infrequently, the ubiquitous possibility of congressional override of judicial statutory interpretation shapes judicial behavior. This point was first formally analyzed by Marks, who ascertained when Congress will be unable to change judicial alterations of legislative policy, even if the majority of legislators do not support the judicial alteration. Marks (1998) showed that both committees in Congress and bicameralism expand the range of stable equilibria the judiciary can institute.4
Figures 11.7a and 11.7b illustrate Marks's insight. HM is the median House voter, HC is the median House committee member. J1 in Figure 11.7a and J2 in Figure 11.7b are two examples of possible court positions. Marks showed how a court's ability to determine policy hinges on the positions of HM and HC. The distance between the preferences of the House median and the House committee median prevents any agreement being able to be reached to overturn any judicial ruling within that (p. 210) range. For example, in Figure 11.7a, a court at J1 could effectively entrench a policy X1; because the Committee and the House have opposing preferences, they cannot agree how to change this ruling. The addition of bicameralism expands the range in which courts can impose their policy preferences: Marks showed that the judiciary can act free from fear of being overturned in the range of the maximum distance between not only HM and HC, but also the Senate equivalents, SM and SC.
Marks assumed that judicial decision‐making was exogenous to legislative structure, and so assumed judges did not act in a sophisticated fashion. Thus he considered that only judicial rulings that occurred within the legislative core, such as X1 in Figure 11.7a, could have a lasting effect.5 However, it follows from his analysis that a sophisticated court positioned outside the equilibrium range, such as J2, would shift its ruling from its ideal point, J2, which would be overruled, to the closest stable equilibrium, in this case X2.
Not only is this simple model informative about the constraints that operate on judicial decision‐making, but it also explains one aspect of congressional behavior, in particular why we should not expect to see frequent congressional override of judicial decisions. While many have concluded that the infrequency of congressional override suggests that this institutional mechanism is ineffective, new separation‐of‐powers models suggest the opposite. Because the threat of congressional override is powerful and credible, it does not need to be exercised (Spiller and Tiller 1996). As with the observational equivalence of the presidential veto, the lack of congressional overrides is consistent with a very powerful threat of these overrides and a very weak one. We do not expect to see these threats exercised: they are operative by their mere potential.
3.1 Further Insights
When the new approach was first applied to the judiciary, some scholars saw it as conflicting with the attitudinalist school. Attitudinalists believe that judges achieve their desired outcomes through sincerely voting their unconstrained policy preferences, whereas new separation‐of‐powers scholars argue that those preferences are exercised in the context of institutional constraints, and so judges strategically incorporate the preferences of other relevant political actors (Segal 1997; Maltzman, Spriggs, (p. 211) and Wahlbeck 1999). However, the new approach provides the logical conclusion of the attitudinalist insight: if there are any costs to the institutional repercussions of pursuing unsustainable outcomes, such as being overturned by Congress, it would be irrational for judges to pursue their preferences without accounting for the limits of their policy‐making capacity. Judges are unlikely to be so shortsighted in pursuing their preferences.
Capturing the limits on the judical pursuit of their policy preferences allows the new approach to improve on the explanatory power of the attitudinal model, as well as to provide new theoretical insights in a number of different areas affecting the judiciary's relations with the other branches. Scholars have shown how the president and the Senate can each shape the ideological make‐up of courts through strategic use of the appointment power. Moraski and Shipan examined when the Senate's advice and consent role will be determinative: its power depends on the Senate's position relative to the president and the existing court median. Only when the Senate is the moderate player does it have a direct influence on the confirmation process (1999, 1077). Jacobi modeled the effect of senatorial courtesy on nominations, and how confirmation outcomes depend on the configuration of preferences among the president, the Senate median, and the home state senator. When the objecting home state senator is the moderate player, the president can strategically draw the equilibrium outcome closer to her preferences (2005, 209). Segal, Cameron, and Cover showed that senatorial support for judicial nominees depends on the relative position of their constituency and the nominee. When both a senator and the nominee are to the left or right of the median constituent, the senator is more likely to vote for the nominee (1992, 111). New separation‐of‐powers analysis has illustrated how and when the ideology of judicial nominees will vary with the ideology of other institutional players.
Similarly, judicial agenda‐setting literature has benefited greatly from the new separation‐of‐powers analysis. This literature documents how judges partially circumvent their institutional inability to initiate their own agendas. For example, judges exercise certiorari voting strategically (e.g. Caldelra, Wright, and Zorn 1999), considering both probable outcomes in deciding whether to grant certiorari (Boucher and Segal 1995) and which cases will most influence lower courts, so as to maximize the proportion of total decisions favorable to their policy preferences (Schubert 1962). Epstein, Segal, and Victor (2002) showed that Supreme Court judges consider both the internal level of heterogeneity of the court, as well as the position of the court median relative to Congress, when deciding between accepting constitutional or statutory interpretation cases.
Appointments and agenda‐setting are just two areas that have benefited from the new approach to traditional questions about judicial behavior, by looking beyond the judiciary itself. Judges anticipate potential opposition to their decisions, and if that opposition can be credibly exercised, they can be expected to adjust their actions to avoid negative consequences. Consequently, the positions of other institutional actors constitute constraints on judicial decision‐making. Incorporating these factors allows for comparative statics, such that we can predict changes in judicial behavior, (p. 212) for example movements resulting from a change in the political administration. At the same time, potential judicial action can constrain other actors: as long as courts can impose costs on Congress, the president, or administrative agencies, those actors can be expected to adjust their decisions to avoid such costs. The new separation‐of‐powers analysis provides a variety of mechanisms for formalizing and predicting when these constraints will be active. As such, it both incorporates an understanding of the judiciary into broader political analysis, and provides important tools and insights to judicial specialists.
4 The Bureaucracy
As with the study of Congress, the presidency, and the courts, the study of bureaucracy also began with relatively limited consideration given to other institutions. Beginning in the late 1970s, however, this started to change, as rational choice theorists (and, to a limited extent, empiricists) began to recognize that there was much greater ability to study their phenomena of interest by incorporating the other branches into their analyses (Huber and Shipan, this volume, review this literature in detail).
4.1 Setting the Table
Prior to the 1970s, the study of bureaucracy was largely confined to historical, normative, and behavioral analyses. With respect to the first, the dominant theory focused on the independence of bureaucrats from other actors in the separation‐of‐powers system, effectively legitimizing and systemizing the “silo”‐based approach (e.g. Bernstein 1955; Mills 1956).6
As Moe (1984, 1990) points out, the normative questions were the domain of another stream of literature: public administration. This literature was primarily interested in how agencies could and should be organized. This focus on internal organization meant that external organization—how agencies fit in a complex system of separation of powers—was largely subordinated.
Finally, the behavioral theory of organization, exemplified by the psycho‐sociological approach of March and Simon (1958; see also Simon 1947; Selznick 1948) also focused primarily on internal organization—how tasks are divided within an organization, what such specialization means for organizational effectiveness in the context of the human cognitive limitations, and what types of pathologies might result.
Following on the pioneering work of Niskanen (1971), by the 1980s scholars began to model the interaction between bureau and other institutional actors in a more fully (p. 213) strategic setting. One of the most important of these was the work that responded to the widely held claim that agencies were unresponsive to Congress and the president. Employing the tools and principles of the new economics of organizations, such as agency theory and transaction cost economics, these scholars began the process of both broadening the explanatory power of models of bureaucratic behavior and using models to help identify causality and sort out apparently observationally equivalent hypotheses.
4.2 Congressional and Presidential Dominance
For those working on more explicitly modeling the interaction between Congress and the bureaucracy, the initial focus was on the variety of instruments that legislators could avail themselves of to control the bureaucracy; and relatedly, the patterns of use of these instruments we would expect to see under various hypotheses of bureaucratic decision‐making.
Weingast and Moran (1983), for example, studied the strategic interaction between an agency (in this case the Federal Trade Commission) and an oversight committee in Congress. They emphasized the varied and effective instruments—budgets, oversight, authority, and structure—which could be used by Congress to turn bureaucratic policy‐making toward their will. One important move they made was to point out that evaluating bureaucratic behavior was very difficult when only looking at an agency by itself. Addressing the question of who controls whom, they pointed out that lack of oversight could either mean that agencies were unfettered or that threats were so effective that the need to exercise such disciplining mechanisms was minimal. In other words, one could not observe the use of punitive measures and draw conclusions about the nature of control.
McCubbins and Schwartz (1984) made a similar contribution by further emphasizing the importance of an equilibrium theory of agency control. They pointed out that in addition to “police patrol” oversight, Congress could also use “fire alarm” oversight. In other words, to monitor agencies, they could rely on “exception reporting,” eschewing active vigilance in favor of particularized, ad hoc intervention in response to complaints by interest groups or constituents. Such an information‐gathering structure, they argued, would be much more efficient than monitoring all agencies, all of the time.
At the same time, other scholars were shifting the focus from the dyad of Congress–bureaucracy to a different dyadic relationship—between president and bureaucracy. Indeed, the primary criticism of the Congress‐centric approach (referred to as “Congressional Dominance”) was that it ignored the president (Moe 1987; Wilson 1989). Perhaps the best example of an attempt to steer scholars away from Congress and toward the president is the work by Moe (1985). Moe argued that the president's power to appoint, combined with senatorial deference, means that when considering “control” of the bureaucracy, the president is the primary principal. In his study of the (p. 214) National Labor Relations Board (NLRB), Moe argued that one could not understand the agency's policy‐making without understanding this institutional feature.7
The insights of the 1980s were a step forward as they shifted from an internal to external, strategic focus. The criticisms that each of the schools made of each other presaged the emergence of a more recent vein of scholarship which explicitly considered the interaction of multiple institutional actors and therefore brought the “new separation‐of‐powers” approach in earnest to bureaucracy studies.
4.3 Agency Discretion in a Separation‐of‐powers Context
In a number of contexts, scholars have continued to focus on how each of the institutional players interacts to explain bureaucratic behavior. To illustrate this development, we focus on perhaps the primary example that integrates all institutional actors—bureaucracy, presidents, Congress, and the courts: the literature on how much discretion agencies are granted to make policy.
This literature focuses on two related issues: first, the degree of ambiguity or specificity of policy that Congress and the president jointly pass, which the courts often review; and second, the degree to which the structure and process created by legislation and judicial decisions constrain or expand agency discretion.
Connecting legal and historical studies of the Administrative Procedure Act of 1946 (Shapiro 1982; Bonfield 1986; Davis 1978), and the insight concerning the usefulness of ex ante control mechanism from McCubbins and Schwartz (1984), McCubbins, Noll, and Weingast (1987, 1989) argued that agency discretion was in part determined by the procedures dictated by Congress that agency policy‐making must follow. These procedures create a set of requirements—including rules of notice, standing, information gathering, and judicial review—for agency rule‐making. By introducing strict limits on discretion, administrative procedures ensure that outcomes will be closer to an elected official's ideal than if the agency had an unlimited range of options. But the mechanism only works if there is ex post enforcement of the rules. Thus, McCubbins, Noll, and Weingast argue that the courts were crucial in making this an effective control mechanism. If the courts ruled consistently with the intent of Congress, then the bureau would have strong incentives to follow their intent. On the other hand, by implication, if the courts were not aligned with the legislature, such mechanisms would provide the bureau with more latitude to implement policy.
More recently, the work on grants of discretion and administrative procedures has been extended in two ways. First, the range of actors included has been expanded. Second, these extended models have allowed scholars to conduct more detailed empirical analysis. For example, Epstein and O'Halloran (1994, 1996, 1999) address the (p. 215) fundamental question that earlier scholars who examined “runaway bureaucracy” were concerned with: “Can the virtues of separate powers intended by the Founders be maintained alongside significant delegation to the executive [by the legislature]?” Epstein and O'Halloran include four important sets of actors in the politics of agency rule‐making: Congress, its committees, the president, and agencies. They observe that, when passing legislation, congresspersons on the floor must make two choices: first, who should they obtain information from—a potentially biased committee or a potentially biased executive agency? And second, how much discretion should they provide to an agency if the agency is allowed to make policy?8 They show that these trade‐offs will depend on political alignment: the key will be who they can more rely on to act as if they had the Congress's interests in mind. Further, the degree of constraint imposed on agencies, when they are delegated policy implementation, will depend on how aligned the president and agencies are with the Congress: a closely aligned agency can be expected to use its information to implement policies which would be close to those that Congress would prefer; otherwise, it will be highly constrained through the use of strict rules, procedures, and very specific legislation.
Prior to Epstein and O'Halloran's work, scholars had primarily considered the question of legislative delegation in a dyadic form: Congress either delegating to a committee or delegating to an agency. Expanding the set of actors to consider all of these players simultaneously allows one to see that committee and agency are partial substitutes. Indeed, Epstein and O'Halloran's model indicates why the study of dyadic delegation relationships is inappropriately specified: because the legislature has choices and will delegate selectively based on its own strategic considerations, the policies and form of delegation observed will necessarily be a selected sample of the full set of possible delegatory choices. This means that what we observe in practice will not be understood without correcting for such sample selection.
By including a broader range of actors, Epstein and O'Halloran's approach allows them to identify the selection mechanism in order more precisely to consider empirical relationships. Indeed, this is precisely how the new separation‐of‐powers analysis informs their empirical work. They employ a data‐set of federal statutes to measure the degree to which administrative procedures constrain grants of authority to agencies. They find, among other things, that, when the executive branch, including the president and bureaucracy, are closer in their policy preferences to the Congress, Congress will at once be more willing to delegate and will provide less constraining rule‐making procedures when they do delegate to a bureau.
The literature on discretion and administrative procedures has been expanded by the work of Huber and Shipan (Huber and Shipan 2002; see also Huber, Shipan, and Pfahler 2001). Like Epstein and O'Halloran, they point out the importance of understanding variation in administrative discretion—either through the specificity (p. 216) of substantive policy instructions or through the degree of procedural limitations—by examining the roles of various actors in a separation‐of‐powers system. Following Epstein and O'Halloran's logic, they argue that “statutory control” of agency authority will be tighter when there is misalignment between institutional actors—divided government. Their contribution is to extend this finding to examine the interaction between divided government and other factors: divided control of the legislature itself; legislative capacity and substitute mechanisms for control.
As with Epstein and O'Halloran, Huber and Shipan's work provides an example of the leverage obtained by incorporating the separation‐of‐powers approach to explain the American bureaucracy. Theoretically, they broaden the range of variables that can be identified to have an effect. Indeed, they demonstrate that two aspects of separation of powers—divided control between legislative and executive functions and bicameral legislatures—are critical to understanding when and how constraints on delegation to agencies will occur.
Further, as with Epstein and O'Halloran, they are able to employ the constraints in the system to inform empirical analysis. They show that the specificity of legislation is increasing in institutional misalignment and that these effects are even greater in the presence of intra‐institutional divisions. This result is only possible because they expand the normal delegation question from simply a legislature to a bureau, to a more nuanced view including elected executives (presidents or governors) and multiple houses of the legislature.9
Like many judges, civil servants, and appointees that work within government, agencies are not elected. Indeed, they are enabled, broadly, by legislation that is passed by Congress and approved by the president; their budgets are determined by the same institutions; heads of agencies are often appointed by consent between legislators and the president; and their decisions are often subject to judicial review. This means, almost by definition, that understanding government agencies requires incorporation of politics and processes that occur in institutions of government outside the bureaucracy.
5 A Synthetic Application
To this point we have focused on behavior within particular institutions, emphasizing that explanations of most institutional behavior in American politics require students to look outside the institution of interest. More generally, analysts may also be interested not only in the behavior of any single institution, but additionally in the outcomes of the interaction of a range of actors.
This type of analysis, which looks to institutional interaction not to explain one institution but rather to understand the more complete political system, is probably most telling when the dependent variable is policy outcomes. As an illustration of the power of this approach, we discuss Ferejohn and Shipan's (1989, 1990) study of telecommunications policy, which examines the interaction of Congress, the courts, and an administrative agency.
Ferejohn and Shipan's model shows why legislative monitoring of agency actions will often fail to force an agency to change policy. Once Congress has delegated power to an agency, the agency can often exploit its first mover advantage, and make a decision that will not have the support of a majority of the legislature, but nonetheless neither Congress nor the president can alter the agency decision. Because the agency has the power of making the first move, it can choose a position in the gridlock region that elected officials will not be able to overturn, just as in Marks's model of judicial decision‐making.10 But Ferejohn and Shipan also showed that the overall effect of judicial review of administrative decisions is to move policy outcomes back toward the preferences of the congressional median. This effect is illustrated in Figure 11.8a and 11.8b.
A is the position of the Agency, C is the position of the Committee median, and H is the position of the House median. As is common in spatial models, the model assumes that each of these actors prefers policies closer to their position than further away from their position, regardless of the direction of policy. Using this assumption, C(H) is the point to the left of C that is equally distant from C as H is; thus, C(H) is a point at which C is indifferent between it and H.
A's policy discretion allows it to set policy anywhere between C(H) and H, knowing that political officials cannot overturn any decision in this range. When A < C(H), if A sets policy at its own ideal point, new legislation will be enacted, and under an (p. 218) open rule, H will be the outcome. In the absence of judicial review, then, an agency with these preferences will set policy strategically XA at C(H), which it prefers to H.
With judicial review, the court with an ideal policy of J has the choice of accepting policy XA or striking it down. If it chooses to strike down the policy, it reverts to the status quo, Q. If the status quo is outside the legislative equilibrium range, again legislation will be enacted at point H. Thus J chooses between XA and Q or XA and H. A, however, makes its decision knowing both J's position and J's choice; it will therefore set policy in a way that will not be overturned.
In Figure 11.8a, Q lies outside the legislative equilibrium range—i.e. the interval between C and H—in which C and H cannot agree on any change. Since J < C, the equilibrium outcome will remain at XA = C(H), as J prefers the agency outcome, C(H), over H. But when J < C, as in Figure 11.8b, the agency knows that J prefers H to C(H), and so will strike down any legislation to the left of J(H), the point at which J is indifferent to H. So in Figure 11.8b, the equilibrium outcome is for the agency to choose J(H).
Ferejohn and Shipan show that in every possible permutation, judicial review either has no effect or draws the equilibrium outcome away from A and toward C. This study illustrates how the potential for judicial review influences the interactions between agencies and Congress.
This model also showed how judicial review can draw the policy outcome back towards congressional preferences by mitigating the influence of the presidential veto. The president can veto congressional amendments of agency action, so in effect the president can choose between supporting congressional action and allowing the decision to revert to the agency's decision. This constrains Congress's choice of actions, which in turn broadens the agency's discretionary range. The threat of the judiciary overriding an agency's action has the effect of reducing the agency's range of discretion, and so drawing the equilibrium policy outcome back toward the congressional median. Thus Ferejohn and Shipan showed how judicial action shapes congressional–bureaucratic relations and congressional–presidential relations.
This chapter has surveyed a series of works emphasizing the new separation‐of‐powers approach to American politics. American politics specialists have always understood the importance of interaction among the branches of government, but until recently, they have not had the technology to study this interaction systematically. Rational choice methods, including game theory and spatial models, provide a means of studying this interaction directly. In an important sense, it provides a method of accounting for how differences in one branch affect behavior in another.
An important lesson of the approach is that behavior in each branch is deeply strategic—actors understand that their decisions are part of a complex decision‐ (p. 219) making process. To further their goals, decision‐makers must anticipate reactions by the other branches. This applies to Congress anticipating both a potential presidential veto and judicial statutory interpretation for the president with respect to Senate confirmation of her appointments; for agencies anticipating either coherence or conflict between branches and making policy accordingly; and for the courts with respect to congressional overturning of their statutory and administrative decisions.
Another lesson is the importance of the sequence of decision‐making. Actors that move first can often forestall the action of those downstream. Agencies have a range of discretion based on the gridlock region created by the differences between the views in Congress and those of the president. Similarly, Howell (2003) shows how, by carefully choosing the ideological location of the policy in the executive order, the president can assure a more favorable outcome in the subsequent legislative game.
In a real sense, this approach portends a revolution in the study of American politics. It demonstrates that studying the major institutions in isolation is problematic, implying that the study of American politics is moving toward a more synthetic approach. The separation of powers no longer means that each branch acts separately, but rather in a strategic and interactive environment.
Of course, the theory is just developing. Indeed, the inclusion of broader and more sequentially dependent strategies in the new approach has also led, not surprisingly, to a “rollback” in the richness of the interactions between the actors. Models focusing on a single institution explore a wide range of issues and subtleties, including cooperation and conflict with various degrees of incomplete information, hierarchies, and agency relationships. Because the separation‐of‐powers models require by definition a broader set of players, they have tended to remain with a few exceptions primarily in the realm of so‐called complete information spatial models. In other words, increasing richness in the institutional dimension has been perhaps necessarily accompanied by less richness in another. That said, having established a necessary base in more stark theories of institutional interaction, scholars have already fruitfully begun the process of bringing more strategic and informational richness into the study of the American separation‐of‐powers system (see Chang, de Figueiredo, and Weingast 2001 for a discussion of recent applications in the area of bureaucracy). We expect considerable growth over the coming years.
Aldrich, J. H. 1995. Why Parties? The Origin and Transformation of Party Politics in America. Chicago: University of Chicago Press.Find this resource:
Banks, J. 1989. Agency budgets, cost information and auditing. American Journal of Political Science, 33: 670–99.Find this resource:
—— and Weingast, B. R. 1992. The political control of bureaucracies under asymmetric information. American Journal of Political Science, 36: 509–24.Find this resource:
Bawn, K. 1995. Political control versus expertise: congressional choices about administrative procedures. American Political Science Review, 89: 62–73.Find this resource:
(p. 220) Bernstein, M. H. 1955. Regulating Business by Independent Commission. Princeton, NJ: Princeton University Press.Find this resource:
Bickel, A. M. 1986. The Least Dangerous Branch: The Supreme Court at the Bar of Politics, 2nd edn. New Haven, Conn.: Yale University Press.Find this resource:
Binder, S., and Maltzman, F. 2002. Senatorial delay in confirming federal judges, 1947–1998. American Journal of Political Science, 46: 190–9.Find this resource:
Bonfield, A. E. 1986. State Administrative Rule Making. Boston: Little, Brown.Find this resource:
Boucher, R. L., and Segal, J. A. 1995. Supreme Court justices as strategic decision makers: aggressive grants and defensive dsenials on the Vinson Court. Journals of Politics, 47: 824–37.Find this resource:
Caldeira, G. A., Wright J. R., and Zorn, C. J. W. 1999. Sophisticated voting and gate‐keeping in the Supreme Court. Journal of Law, Economics and Organization, 15: 549–72.Find this resource:
Cameron, C. 1996. Veto Bargaining. New York: Cambridge University Press.Find this resource:
Cameron, C., Cover, A., and Segal, J. 1990. Senate voting on Supreme Court nominees: a neoinstitutional model. American Political Science Review, 84: 525–34.Find this resource:
Canes‐Wrone, B. 2003. Bureaucratic decisions and the composition of the lower courts. American Journal of Political Science, 47: 205–14.Find this resource:
Chang, K. 2003. Appointing Central Bankers. New York: Cambridge University Press.Find this resource:
—— de Figueiredo, R. J. P., Jr., and Weingast, B. R. 2001. Rational choice theories of bureaucratic control and performance. In The Elgar Companion on Public Choice ed., W. F. Shughart III and L. Razzolini. New York: Edward Elgar.Find this resource:
Chappell, H. W., Jr., McGregor, R. R., and Vermilyea, T. 2005. Committee Decisions on Monetary Policy: Evidence from Historical Records of the Federal Open Market Committee. Cambridge, Mass.: MIT PressFind this resource:
Cox, G. W., and McCubbins, M. D., 2005. Setting the Agenda: Responsible Party Government in the U.S. House of Representatives. Cambridge: Cambridge University Press.Find this resource:
Dahl, R. A. 1957. Decision‐making in a democracy: the Supreme Court as a national policy‐maker. Journal of Public Law, 6: 279–95.Find this resource:
Davis, K. C. 1978. Administrative Law Treatise, vol. i. San Diego, Calif.: Davis KC.Find this resource:
de Figueiredo, R. J. P., Jr. 2002. Electoral competition, political uncertainty and policy insulation. American Political Science Review, 96: 321–33.Find this resource:
—— 2003. Endogenous budget institutions and political insulation: why states adopt the item veto. Journal of Public Economics, 87: 2677–701.Find this resource:
—— Spiller, P. T., and Urbiztondo, S. 1999. An informational perspective on administrative procedures. Journal of Law, Economics and Organization, 15: 283–305.Find this resource:
—— and Vanden Bergh, R. G. 2004 The political economy of state‐level administrative procedure acts. Journal of Law and Economics, 47: 569–88.Find this resource:
Derthick, M., and Quirk, P. J. 1985. The Politics of Deregulation. Washington, DC: Brookings Institution.Find this resource:
Epstein, D., and O'Halloran, S. 1994. Administrative procedures, information, and agency discretion. American Journal of Political Science, 38: 697–722.Find this resource:
—— —— 1996. Divided government and the design of administrative procedures: a formal model and empirical test. Journal of Politics, 58: 373–97.Find this resource:
—— —— 1999. Delegating Powers: A Transaction Cost Politics Approach to Policy Making under Separate Powers. New York: Cambridge University Press.Find this resource:
Epstein, L., Segal, J., and Victor, J. N. 2002. Dynamic agenda‐setting on the United States Supreme Court: an empirical assessment. Harvard Journal on Legislation, 39: 395–433.Find this resource:
Ferejohn, J. A., and Shipan, C. R. 1989. Congressional influence on administrative agencies: a case study of telecommunications policy. In Congress Reconsidered, ed. L. C. Dodd and B. I. Oppenheimer, 4th edn. Washington, DC: Congressional Quarterly Press.Find this resource:
(p. 221) —— —— 1990. Congressional influence on bureaucracy. Journal of Law, Economics and Organization, 6: 1–21.Find this resource:
Fisher, L. 1975. Presidential Spending Power. Princeton, NJ: Princeton University Press.Find this resource:
Fuller, L. 1978. Forms and limits of adjudication. Harvard Law Review, 92: 353–403.Find this resource:
Gailmard, S. 2002. Agency expertise, subversion and bureaucratic discretion. Journal of Law, Economics and Organization, 18: 536–55.Find this resource:
Gely, R., and Spiller, P. T. 1992. The political economy of Supreme Court constitutional decisions: the case of Ro0sevelt's Court‐packing plan. International Review of Law and Economics, 12: 45–67.Find this resource:
Grier, K. 1987. Presidential elections and Federal Reserve policy: an empirical test. Southern Economic Journal, 54: 475–86.Find this resource:
Havrilesky, T. 1995. The Pressures on American Monetary Policy, 2nd edn. Dordrecht: Kluwer.Find this resource:
Howell, W. 2003. Power without Persuasion: The Politics of Direct Presidential Action. Princeton, NJ: Princeton University Press.Find this resource:
Huber, J. D., and Shipan, C. R. 2002. Deliberate Discretion: The Institutional Origins of Bureaucratic Autonomy in Modern Democracies. New York: Cambridge University Press.Find this resource:
—— —— and Pfahler, M. 2001. Legislatures and statutory control of the bureaucracy. American Journal of Political Science, 45: 330–45.Find this resource:
Jacobi, T. 2005. The senatorial courtesy game: explaining the norm of informal vetoes in “advice and consent” nominations. Legislative Studies Quarterly, Vol. XXX, May: 193, 218.Find this resource:
Keech, W. R., and Morris, I. L. 1997. Appointments, presidential power, and the Federal Reserve. Journal of Macroeconomics, 19: 253–67.Find this resource:
Kiewiet, D. R., and McCubbins, M. 1988. Presidential Influence on congressional appropriation decisions. American Journal of Political Science, 32: 713–36.Find this resource:
—— —— 1991. The Logic of Delegation: Congressional Parties and the Appropriations Process. Chicago: University of Chicago Press.Find this resource:
Krehbiel, K. 1998. Pivotal Politics. Chicago: University of Chicago Press.Find this resource:
Lupia, A. D., and McCubbins, M. D. 1994. Designing bureaucratic accountability. Law and Contemporary Problems, 57: 91–126.Find this resource:
McCarty, N. 2004. The appointments dilemma. American Journal of Political Science, 48: 413–28.Find this resource:
McCloskey, R. G. 1960. The American Supreme Court. Chicago: University of Chicago Press.Find this resource:
McCubbins, M. D. 1990. Party governance and U.S. budget deficits: divided government and fiscal stalemate. In Politics and Economics in the Eighties, ed. A. Alesina and G. Carliner. Chicago: University of Chicago Press.Find this resource:
—— Noll, R., and Weingast, B. R. 1987. Administrative procedures as instruments of political control. Journal of Law, Economics and Organization, 3: 243–77.Find this resource:
—— —— —— 1989. Structure and process, politics and policy: administrative arrangements and the political control of agencies. Virginia Law Review, 75: 431–82.Find this resource:
—— and Schwartz, T. 1984. Congressional oversight overlooked: police patrols vs. fire alarms. American Journal of Political Science Review, 28: 165–79.Find this resource:
Maltzman, F., Spriggs, J. F., II, and Wahlbeck, P. J. 1999. Strategy and judicial choice: new institutionalist approaches to supreme court decision‐making. In Supreme Court Decision‐Making: New Institutionalist Approaches, ed. C. W. Clayton and H. Gillman. Chicago: University of Chicago Press.Find this resource:
March, J. G., and Simon, H. 1958. Organizations. New York: Wiley.Find this resource:
Marks, B. A. 1998. A model of judicial influence on congressional policy‐making: Grove City College v. Bell. Unpublished Hoover working paper manuscript, Stanford, Calf.Find this resource:
Mills, C. W. 1956. The Power Elite. Oxford: Oxford University Press.Find this resource:
(p. 222) Moe, T. M. 1984. The new economics of organizations. American Journal of Political Science, 28: 739–77.Find this resource:
—— 1985. The control and feedback in economic regulation: the case of the NLRB. American Political Science Review, 79: 1094–116.Find this resource:
Moe, T. M. 1987. Interests, institutions, and positive theory: the politics of the NLRB. Studies in American Political Development, 2: 236–99.Find this resource:
—— 1990. The politics of structural choice: towards a theory of public bureaucracy. In Organization Theory: From Chester Barnard to the Present and Beyond, ed. O. E. Williamson. Berkeley: University of California Press.Find this resource:
Moraski, B., and Shipan, C. 1999. The politics of Supreme Court nominations: a theory of institutional constraints and choices. American Journal of Political Science, 43: 1069–95.Find this resource:
Niskanen, W. 1971. Bureaucracy and Representative Government. Chicago: Aldine Press.Find this resource:
Nokken, T. P., and Sala, B. J. 2000. Confirmation dynamics: presidential appointments to independent agencies, 1953–1988. Journal of Theoretical Politics, 12: 91–112.Find this resource:
Peltzman, S. 1976. Toward a more general theory of regulation. Journal of Law and Economics, 19: 211–40.Find this resource:
Romer, T., and Rosenthal, H. 1978. Political resource allocation, controlled agendas, and the status quo. Public Choice, 33: 27–43.Find this resource:
Rosenberg, G. N. 1991. The Hollow Hope: Can Courts Bring about Social Change? Chicago: University of Chicago Press.Find this resource:
Schubert, G. 1962. Policy without law: an extension of the certiorari game. Stanford Law Review, 14: 284–327.Find this resource:
Segal, J. A. 1997. Separation of powers games in the positive theory of Congress and the courts. American Political Science Review, 91: 28–44.Find this resource:
—— and Spaeth, H. J. 1993. The Supreme Court and the Attitudinal Model. New York: Cambridge University Press.Find this resource:
—— Cameron, C., and Cover, A. 1992. A spatial model of roll‐call voting: services, constituents, president, and interest groups and Supreme Court confirmations. American Journal of Political Science, 36: 96–121.Find this resource:
Selznick, P. 1948. Foundations of the theory of organization. American Sociological Review, 13: 25–35.Find this resource:
Shapiro, M. 1982. Who Guards the Guardians? Athens: University of Georgia Press.Find this resource:
Simon, H. A. 1947. Administrative Behavior. New York: Macmillan.Find this resource:
Snyder, S. K., and Weingast, B. R. 2000. The American system of shared powers: the president, Congress, and the NLRB. Journal of Law, Economics and Organization, 16: 269–305.Find this resource:
Spiller, P., and Tiller, E. 1996. Invitations to override: congressional reversals of Supreme Court decisions. International Review of Law and Economics, 16: 503–21.Find this resource:
Stigler, G. 1971. The theory of economic regulation. Bell Journal of Economics, 2: 3–21.Find this resource:
Vanden Bergh, R. G., and de Figueiredo, R. J. P., Jr. 2003. Political uncertainty and administrative procedures. In Uncertainty in American Politics, ed. B. Burden. Cambridge: Cambridge University Press.Find this resource:
Weingast, B. R., and Moran, M. J. 1983. Bureaucratic discretion or congressional control: regulatory policymaking by the FTC. Journal of Political Economy, 91: 765–800.Find this resource:
Wildavsky, A. 1964. Politics of the Budgetary Process. Boston: Little, Brown.Find this resource:
Wilson, J. Q. 1980. The Politics of Regulation. New York: Basic Books.Find this resource:
—— 1989. Bureaucracy: What Government Agencies Do and Why They Do It. New York: Basic Books.Find this resource:
(*) The authors thank Charles Cameron, Gary Cox, William Howell, Kenneth Shepsle, and Donald Wittman for helpful comments.
(1) The other chapters in this section survey the new literature as it applies to the presidency (Cameron), bicameralism (Cutrone and McCarty), the bureaucracy (Huber and Shipan), and the courts (McCubbins).
(3) This formulation assumes that Congress has full information about the president's preferences. Cameron (this volume; 1996) shows how to relax the model to allow uncertainty about the president's preferences. This model derives from that of Romer and Rosenthal 1978.
(4) In keeping with the models of his era, Marks relied on a model of committee agenda power. The same form of results can be obtained using a more recent model of party agenda power (see Aldrich 1995 or Cox and McCubbins 2005).
(6) The so‐called “capture” hypothesis was further developed in the 1970s by the Chicago School of economists who argued that agencies often served economic interests who supported regulation as a barrier to entry (Stigler 1971; Peltzman 1976).
(7) Indeed, developing Moe's argument further, a series of studies (Keech and Morris 1997; Grier 1987; Havrilesky 1995; and Chappell, McGregor, and Vermilyea 2005) develop both theoretical models and empirical evidence to demonstrate the influence of presidential appointments on agency behavior. Notably, Moe's work presaged the later work (above) of Snyder and Weingast 2000 which then extended the analysis of the NLRB to consider agency, president, and Senate.
(8) This emphasis on the tension between the need to control agencies and the need to exploit the expertise held within agencies is the central concern of much of the literature on delegation to bureaucracies. Banks 1989, Banks and Weingast 1992, Kiewiet and McCubbins 1991, Bawn 1995, Gailmard 2002, Lupia and McCubbins 1994, and de Figueiredo, Spiller, and Urbiztondo 1999 are all examples of contributions to this discussion.
(10) For the purposes of this section, it doesn't matter if the gridlock region is generated involving a mix of veto pivot, filibuster pivot, and the president (à la Krehbiel 1998), the majority party median and floor median in Congress (à la Cox and McCubbins 2005), or the floor median and the committee median (à la Ferejohn and Shipan 1989 or Marks 1988).