|The Policy Consequences of Partisan Polarization in the United States
The following is a very rough sketch of the argument and evidence for a book chapter in preparation on the effects of partisan polarization on a number of facets of the policy process in the United States.
In order prevent the “mischief of faction,” the founders designed a system of institutional politics that preserves the status quo absent broad levels of societal consensus. Indeed, many (if not most) of the great achievements of the American national government have come about during the periods of broad bipartisan consensus. However, over the past twenty-five years, we have witnessed the near collapse of such a consensus, at least among political elites. By almost all measures of partisan polarization, the divide between Democratic and Republican members of Congress have widened deeply over the past twenty-five years, reaching levels of partisan conflict not witnessed since the 1920s. Figure 1 illustrates the growing divergence of House Democrats and Republicans as measured by DW-NOMINATE scores.1
There is a small, but rapidly growing literature on the causes of the recent partisan polarization stressing such factors as the realignment of Southern politics (McCarty, Poole and Rosenthal 1997, 2003; Perlstein 2002; Stonecash, Brewer, and Mariani 2003), increasing economic and social inequality (McCarty, Poole, and Rosenthal 1997, 2003, Stonecash, Brewer, and Mariani 2003, Phillips 2002, Hicks 2003, Rosenthal n.d.), the effects of congressional districting (McDonald and Grofman 1999; Carson, Crespin, Finocchiaro, and Rohde 2003), and party-enhancing legislative reforms (Rohde 1991, Snyder and Groseclose 1999, but see McCarty, Poole, and Rosenthal 2000). A growing number of scholars have also studied the effects of elite polarization on mass behavior finding significant effects on the structure of mass partisanship (Hetherington 2001 and McCarty, Poole, and Rosenthal 2003) and on the level of trust in government (King 1999 and Hetherington nd). However, relatively less attention has been paid to the effects of polarization on public policy outcomes.2 In this essay, I propose to explore four primary hypotheses about the effects of polarization on public policy in the United States.
Polarization makes it harder to build legislative coalitions, leading to policy “gridlock.”
Polarization is not ideologically neutral; it has a conservative effect, especially on economic and social policy.
Polarization has had negative effects on the efficient functioning of the administrative state and the judiciary.
Polarization has fundamentally altered the balance of power among our national institutions at the expense of Congress. It has encouraged executives and judges to act unilaterally without legislative consent.
Many of these propositions are not original to me, and have been developed in supported in diverse research endeavors. However, there has been little attempt to provide a systematic account of these diverse phenomena and link them explicitly to political polarization. Thus, my plans are to synthesize a number of strands of the literature on American political institutions around the questions of the consequences of partisan polarization for contemporary policymaking and institutional performance. In those instances where I identify gaps in the literature on key questions, I will supplement with original data analysis on the effects of polarization on legislative productivity and policy outcomes. I now sketch the rough contours of the argument.
Polarization and Gridlock
Polarization may simply produce greater variation in policy, at times producing new policy initiatives and at other times rolling them back. However, given the dispersion of authority in American policymaking systems, there is strong theoretical reason to believe that the main effect of polarization is to produce less public policy. Systems of separation of powers and bicameralism require an extraordinary level of consensus to pass new legislation so that the enacting coalitions are quite often required to be bipartisan (Krehbiel 1999).
Two broad perspectives have been applied to explain patterns in significant legislation in the post-war period: partisan theories and pivot theories. It is easy to see the crucial role that partisan polarization plays in both. In partisan theories, new policies emerge through inter-branch bargaining with partisan actors. Therefore, when different branches are controlled by different parties (e.g. divided government) significant policy enactments are less likely to emerge. However, such a prediction depends on the extent of policy disagreement are across parties. Clearly, executive-legislative relations during the Eisenhower years were quite different than the inter-branch confrontations of the late 1990s. Thus, predictions about the effect of divided governments should be conditioned on the level of partisan polarization. The fact that increased polarization has occurred coincident with increased frequency of divided government suggests a double-whammy against significant policymaking.3
Pivot models such as those proposed by Krehbiel (1999) and Brady and Volden (1998) also suggest a key role for polarization in policymaking. These models identify necessary conditions for policy change based on the support of pivotal actors. They argue that the two binding constraints on policymaking in the U.S. are the presidential veto and cloture in the Senate. Thus, new policies require a House majority, sixty senators to bring cloture, and presidential support or a 2/3s vote of each chamber to override a presidential veto. These models define the “gridlock interval” as the region of the policy space between the senator pivotal on a presidential veto (the 33rd or 67th most conservative depending on whether or not the president is a liberal or conservative) and the senator pivotal on cloture (currently the 40th or 60th senator depending on the president’s position). Prior to the liberalization of cloture in 1975, the “filibuster” pivot was the 33rd or 67th senator. Any policy whose status quo lies in this interval will not be changed because at least one of the two pivotal actors would veto any move. Thus, when the gridlock interval is large, significant policymaking will be rare.
Since the definition of the gridlock interval is non-partisan, there need not be any one-to-one connection between partisan polarization the width of the gridlock interval. However, as we can see from Figure 2, there is a very strong correlation.
More than ¾’s of the variance in the width of the gridlock interval in the post-War period is accounted for by polarization and the 1975 reforms. However, the concepts are statistically distinct enough to incorporate both into econometric models.
Beyond its direct consequences for the difficultly of legislative coalition-building, partisan polarization may have other effects that retard policymaking. For example, King (1997) and Hetherington (nd) argue that elite polarization has contributed to lower levels of trust in government which in turn weakens the public’s demand for redistributive programs. Greater polarization may also increase the incentives to engage in strategic disagreements or the “blame game” which lead to bargaining failure and gridlock (Gilmour 1995, Groseclose and McCarty 2000).
My hypotheses can be tested on existing data sets such as that collected by David Mayhew in his landmark study of major legislative enactments following World War II. Below are results from a simple regression where a measure of party polarization (from the U.S. House) has been added to Divided Government (standard errors in parentheses).
Major Laws = 27.12 + .168*Divided Government –25.6*polarization
(7.32) (1.99) (12.29)
The effect of polarization is very substantial, accounting for a difference of seven major laws between its trough in 1968 and its peak in 1999 (the average number of major laws is only 11). As would be suggested by the partisan theory, I included an interaction between polarization and divided government but it was never significant and generally had the wrong sign. I also tried to use the measure of the gridlock interval in addition to polarization, but its coefficient had the wrong sign.
I have also extended this analysis to the more fine-grained coding of legislation developed by Howell, Adler, Cameron, and Riemann (2000). They classify new public laws as corresponding to four levels of significance - level A corresponding roughly to Mayhew’s data though trivial enactments at level D. If I run the same model as above (with or without a linear time trend), polarization has a statistically and substantively significant negative effect on legislation at all levels, save D. It turns out that polarization actually increases the output of triviality.
Thus, my preliminary and admittedly crude analysis is consistent with polarization having a large negative effect on the legislative output of Congress.
The Conservative Bias of Polarization
The increased propensity for gridlock and stalemate can have profound consequences for policy outcomes. First, a politically polarized Congress will have difficulty in responding to economic shocks.4 Second, gridlock erodes the value of non-inflation indexed social benefits by preventing the legislative actions necessary to prevent the erosion of real benefit levels. Over the long term the level of non-indexed benefits will converge to the levels preferred by the most fiscally conservative pivotal decision maker. This dynamic suggests that, especially in the case of social spending and non-indexed programs, political polarization biases policy in a fiscally conservative direction.
As an example of this phenomenon, consider the Federal minimum wage. While there have been numerous unsuccessful attempts to index it to the cost of living, absent new legislation, inflation erodes its real value. Absent deflation, the status quo is always moving in a conservative direction. Thus, if the conservative pivot moves to the right, there should be fewer and smaller adjustments to the nominal rate, depressing the real value of the wage. Below I report the estimates of a simple model of the annual real minimum wage since its inception in 1936. The primary independent variables are polarization, a dummy variable for unified Democratic governments, an interaction between polarization and unified government, and a time trend.
Real Min. Wage = 11.18 -14.01*Polar +3.55 *UniDem – 5.75*Polar*UDG + .092*year R2 =.743
(.078) (1.37) (1.51) (2.59) (.007)
We can see that polarization has a large, depressing effect on the level of the real minimum wage. In fact, the difference between the predicted wage at minimum level of polarization and the maximum level is about $5. Polarization has also wiped out the Democratic premium on the wage. At the minimum level of polarization, the real wage is about $.60 higher during unified Democratic governments. At its maximum level, Democratic governments actually lead to lower real minimum wages.5
This analysis can be extended with a variety of other data on specific policy outcomes. In particular, I can supplement it with preliminary results of my work on the relationship between polarization and TANF policies in the U.S. states.
Polarization and the Executive and Judicial Branches
One of the clearest manifestations of partisan polarization is increasing level of contention associated with presidential nominations to executive and judicial positions. The “confirmation mess” not only has the potential to drive otherwise qualified individuals from pubic life, but contributes to large numbers of vacancies in the American bureaucracy and judiciary which impairs the effectiveness of these institutions.
In a recent study of executive appointments, McCarty and Razaghian (1999) find very strong evidence that party polarization in the Senate has contributed to the dramatic increase in the length of time required to confirm appointees that is demonstrated in Figure 3. In their model specification for the post-WWII era, they estimate the following effects for divided government, partisan polarization, and their interaction (standard errors in parentheses):
ln(confirmation time) = -0.988*divided + 1.459*polar + 2.155*polar*divided + other controls
(0.350) (0.469) (0.610)
Given this specification, the polarization since the 1970s has increased confirmation times 35% during unified governments and about 90% during divided governments. While the effects on individual nominees are not large (perhaps a month or more on average), accumulating them over hundreds of nominees has had a large impact on the staffing of the executive branch.
While they do not consider polarization directly, Binder and Maltzman (2002) find that the distance between the president and the median of the opposite party has a large effect on the duration of the confirmation process for federal judges. They find that a two standard variation in this distance lowers the daily probability of confirmation by 60%.6 Since polarization accounts for much of the variation in this measure, this is strong evidence that polarization is consequential for the large vacancy rates in the federal judiciary. 7 In my essay, I’ll reexamine Binder and Maltzman’s data to look specifically for a polarization effect.
Perhaps more important than polarization’s effects on the confirmation process is the question of to what extent the courts themselves are polarized. Using data on the ideal points of Supreme Court justices developed by Martin and Quinn (2002), I have plotted the difference between the 67th percentile conservative and the 33rd percentile conservative on the court for each year. While the pattern does not match the patterns of for the House and Senate, it is worth noting that this measure reached its all-time high shortly after the confirmation of Breyer and continues to increase even absent turnover on the court. There is strong reason to believe that the lower federal courts are even more polarized than the Supreme Court, especially across circuits. In the larger essay, I will pursue such a claim.
Polarization and the Separation of Powers
The power of a legislative body lies in its ability to make laws. As I argued above, a direct effect of partisan polarization is that it impedes the legislative process. However, there is little reason to believe that it similarly impedes other actors in the policy process. A large number of positive theories of policymaking have stressed that an inability to pass corrective legislation allows much greater autonomy for executive and judicial actors in the policy process.8 Several recent empirical studies back-up such claims. Howell (2003) finds evidence that presidents are more likely to act unilaterally during periods of executive-legislative conflict. Lewis (2002) finds that presidents are more likely to create agencies through reorganization authority rather than seek a new statute when inter-branch preferences diverge. Such agencies, he finds, are much less durable and more politicized than statutory agencies.
With respect to the courts, Farhang (2003) finds that Congress is much more likely to empower private litigants and courts to enforce federal statutes rather that executive branch bureaucrats when the preferences of the president and congressional majorities diverge. Thus, he argues that a significant source of the litigation explosion is the result of partisan polarization.
Alt, James E. and Robert C. Lowry. 1994. “Divided Government, Fiscal Institutions, and Budget Deficits: Evidence from the States,” American Political Science Review, 88(4):811-828.
Alesina, Alberto and Howard Rosenthal. 1995. Partisan Politics, Divided Government, and the Economy. New York: Cambridge University Press.
Binder, Sarah and Forrest Maltzman. 2002. “Senatorial Delay in Confirming Federal Judges, 1947-1998.” American Journal of Political Science 46(1):190-199.
Brady, David and Craig Volden. 1997. Revolving Gridlock. Westview Press.
Carson, Jamie L., Michael H. Crespin, Charles J. Finocchiaro, and David Rohde. 2003. “Linking Congressional Districts Across Time: Redistricting and Party Polarization in Congress.” Paper presented at the 2003 Midwest Political Science Association Meetings, Chicago: Ill.
Fiorina, Morris. 1996. Divided Government. Second Edition. Boston: Allyn and Bacon.
Farhang, Sean. 2003. “The Litigation State: Public Regulation and Private Lawsuits in the American Separation of Powers System.” Typescript: Columbia University.
Ferejohn, John and Charles Shipan. 1990. “Congressional Influence on Bureaucracy.” Journal of Law, Economics, and Organization 6: 1-21.
Hetherington, Marc J. 2001. “Resurgent Mass Partisanship: The Role of Elite Polarization.” American Political Science Review 95(3):619-631.
Hetherington, Marc J. nd. Why Trust Matters: Declining Political Trust and the Demise of American Liberalism. Forthcoming.
Hicks, Alexander. 2003. “Back to the Future? A Review Essay on Income Concentration and Conservatism.” Socio-Economic Review 1:271-288.
Howell, William. 2003. Power without Persuasion: The Politics of Direct Presidential Action, Princeton University Press.
Howell, William, Scott Adler, Charles Cameron, and Charles Riemann. 2000. “Divided Government and the Legislative Productivity of Congress, 1945-1994.” Legislative Studies Quarterly. 25(2): 285-312.
Gilmour, John. 1995. Strategic Disagreement: Stalemate in American Politics University of Pittsburgh Press.
Groseclose, Timothy and Nolan McCarty. 2000.“The Politics of Blame: Bargaining before an Audience,” American Journal of Political Science. 45(1):100-119.
King, David C. 1997. “The Polarization of American Political Parties and Mistrust of Government,” in Joseph S. Nye, Philip Zelikow & David C. King (eds.). Why People Don't Trust Government. Harvard University Press.
Krehbiel, Keith. 1998. Pivotal Politics: A Theory of U.S. Lawmaking. Chicago: University of Chicago Press.
McCarty, Nolan, Keith Poole and Howard Rosenthal. 1997. Income Redistribution and the Realignment of American Politics. Washington D.C.: American Enterprise Institute.
McCarty, Nolan, Keith Poole and Howard Rosenthal. 2001. “The Hunt for Party Discipline in Congress.” American Political Science Review. 95(3):673-687.
McCarty, Nolan, Keith Poole and Howard Rosenthal. 2003. “Political Polarization and Income Inequality.” Typescript, Princeton University.
McCarty, Nolan and Rose Razaghian. 1999. “Advice and Consent: Senate Response to Executive Branch Nominations 1885-1996.” American Journal of Political Science, 43(3):1122-43.
McDonald, Michael P. and Bernard Grofman. 1999. “Redistricting and the Ideological Polarization of the House of Representatives.” Presented at 1999 Midwest Political Science Association Conference.
Martin, Andrew D. and Kevin M. Quinn. 2002. "Dynamic Ideal Point Estimation via Markov Chain Monte Carlo for the U.S. Supreme Court, 1953-1999." Political Analysis. 10: 134-153.
Mebane, Walter R. and Jasjeet S. Sekhon. 2002. “Coordination and Policy Moderation at Midterm.” American Political Science Review, 96(1):141-157.
Moe, Terry and William Howell. 1999. “The Presidential Power of Unilateral Action.” Journal of Law, Economics and Organizations. 29(4): 850-871.
Perlstein, Rick. 2002. Before the Storm: Barry Goldwater and the Unmaking of the American Consensus Hill and Wang.
Poole, Keith T. and Howard Rosenthal. 1984. “The Polarization of American Politics,” Journal of Politics.
Poole, Keith T. and Howard Rosenthal. 1997. Congress: A Political-Economic History of Roll Call Voting. New York: Oxford University Press.
Poterba, James M. 1994. “State Responses to Fiscal Crises: The Effects of Budgetary Institutions and Politics,” Journal of Political Economy. 104(4): 799-821.
Rohde, David. 1991. Party Leaders in the Post-Reform House.
Rosenthal, Howard. nd. “Politics, Public Policy, and Inequality: A Look Back at the Twentieth Century” in Social Inequality ed. Kathryn Neckerman. Russell Sage.
Snyder, James and Tim Groseclose. 2000. “Estimating Party Influence in Congressional Roll-Call Voting.” American Journal of Political Science, 44(2), pp. 193-211.
Stonecash, Jeffrey M. Mark D. Brewer, and Mack D. Mariani. 2003. Diverging Parties: Social Change, Realignment, and Political Polarization. Westview Press.