The effect of judicial selection processes on judicial quality: the role of partisan politics.

Sobel, Russell S. ; Hall, Joshua C.

The quality of a state's judicial system is an important determinant of economic growth and vitality. The decisions made within state judicial systems affect the degree to which private property rights are well-defined and enforced, which is an essential building block for entrepreneurial activity and economic growth. The key link between free-market institutions, such as secure property rights, and entrepreneurial activity has been demonstrated by Kreft and Sobel (2005) and Ovaska and Sobel (2005). Bad court decisions often infringe on the individual liberties and freedoms that are essential underpinnings for civil society and a well-functioning market economy (see Dorn 1985 and others in that special issue of the Cato Journal on this general topic).

In addition, decisions made within state judicial systems also have important effects on the cost of doing business in a state. Poor liability rules reflected in state judicial decisions have been blamed for high medical malpractice rates, high workers" compensation rates, and high automobile insurance rates in many states. Judicial decisions also impact the costliness of mandates and other regulations faced by businesses. Thus, it is clear that the judicial system is important for economic activity, and thus so is the selection mechanism that is used to determine the membership of state courts.

States use different methods for the selection of judges. In 29 states, the governor or legislature appoints judges, while in 21 states popular elections are used to select judges. (1) Among those states that elect judges, 8 rely on partisan elections, in which the judicial candidates run as a member of a particular political party, while the 13 other states use nonpartisan elections. (2)

The extremely politicized nature of partisan elections casts some doubt on their ability to efficiently select unbiased judges. The 2004 state supreme court election in West Virginia provides a good example. This was the most expensive judicial election in the nation in 2004, with the Democratic incumbent spending $376,000 to the opponents $540,000. In addition, this spending was dwarfed by the spending on political advertisements and publicity by other groups in the state. One pro-incumbent group spent more than $1 million by itself, while the largest anti-incumbent group spent $2.5 million. The organized labor union interests in the state took the incumbent's side while the business associations supported the challenger. The candidates themselves ran on political platforms reflecting their party affiliation. The incumbent, well known for his reputation of deciding cases for labor interest and against business interests, stressed this reputation in his media ads, while the challenger outwardly vowed to decide cases in a more business friendly manner if he were to be elected.

In partisan election states, citizens are used to this political bias in judicial elections, and there is no veil of judicial decisions being unbiased. Public opinion polls from these partisan election states confirm this claim. Geyh (2003) cites polls from Texas and Ohio that find more than 80 percent of adults believe that campaign contributions to judges running for election have a significant influence on subsequent judicial decisions they make.

If an unbiased and fair judicial system is important, it is clear that having judicial candidates run for office on political platforms, with heavy support from labor or business groups, works contrary to this goal. In this article, we focus on two issues related to the determinants of judicial quality. Do states using partisan elections have lower judicial quality? If so, do states with partisan judicial elections have judicial outcomes that clearly differ depending on which political party controls the court?

The article proceeds by first reviewing the previous literature on judicial appointment versus judicial election. We then continue by presenting data on the quality of state judicial systems and on other measures of judicial outcomes, and attempt to uncover the impact of partisan elections, and of specific party control.

Literature Review

An enduring question in the existing literature is whether judges should be appointed or elected. This topic has engendered considerable debate among scholars and according to Dubois (1986) is the single most discussed topic in legal scholarship. Among the many arguments put forth by those in favor of appointing judges instead of electing them is the hypothesis that judicial elections undermine public confidence in the judiciary and thus are incompatible with a strong and independent judiciary (see, e.g., Link 2004). The research on this topic is so large and extensive that we refer the reader to Hanssen (2004) for a more comprehensive review. However, for our purposes, we focus here specifically on several recent studies that examine specific measurable outcomes of the judicial process, and how those outcomes differ between the two groups of states.

In a study of state employment discrimination charges filed during the 1973 to 2000 period, Besley and Payne (2003) find that states appointing their judges have significantly lower levels of filings when compared with states electing their judges. They hypothesize that there are two potential reasons why judicial outcomes may differ between elected and appointed systems. First, the choice of the judicial selection method might cause a difference in what types of individuals actually become judges. This can happen not only because the selection criteria is different, but also because different types of potential judges may self-select into the potential pool of judicial candidates. The second source of potential difference stems from the incentive effects created by the need for elected judges to face popular reelection. Facing a reelection constraint may alter the incentives judges face when deciding cases. Besley and Payne conclude that it is the incentive effect of facing reelection that causes the difference they document in their data.

Tabarrok and Helland (1999) and Helland and Tabarrok (2002) examine state tort awards, and hypothesize that when state judges are elected they will have a stronger incentive to redistribute wealth from out-of-state businesses to in-state plaintiffs (who are, of course, voters). Their prediction is that states with judicial elections should have larger trial awards. In addition, they predict this effect to be even larger in states using partisan elections. Using large samples of cases from across the United States, they find strong support for their hypothesis that elections, and their partisan nature, matter in both of their studies. Thus, like Besley and Payne (2003), Tabarrok and Helland (1999) and Helland and Tabarrok (2002) find that judicial outcomes differ significantly between states electing judges and those appointing judges.

In addition to examining differences between states that elect judges and those that appoint judges, the literature has also examined whether there exist differences between states electing their judges on a partisan basis and those electing judges on a nonpartisan basis. Hanssen (2000) finds, for example, that partisan elected judges side less frequently with challengers to a regulatory status quo, and Hall and Brace (1996) find that partisan elected judges dissent less frequently on politically controversial issues. Even within the literature on the importance of judicial independence, such as Hanssen (9.004), judges elected on a partisan basis are considered the category least independent, with nonpartisan elections being the second least independent.

Absent from this previous literature, however, is an analysis of whether the outcomes from judicial systems using partisan elections differ depending on which political party holds a majority in the court. That is, is there evidence that judicial outcomes differ between Republican majority state courts and Democratic majority state courts? The absence of an inquiry into whether party control matters in state judicial systems is made more intriguing by the relative abundance of literature on how party control affects legislative and executive decisionmaking and outcomes. Even when party control is not a primary focus of the study, political party affiliation is nearly always included as a control variable to capture any party control effect. Some recent representative studies of this literature are Pjesky and Sutter (2002) and Lopez and Sutter (2004).

The issue of whether party control matters for judicial outcomes is uniquely important for several reasons. First, if there are differences, previous studies that simply group them together to find the average outcome for elected judicial systems may be averaging away the true magnitude of this election effect. Second, to the extent that there are differences depending on party control, this can create problems in the long term with stability of judicial decisions in a state, creating uncertainty and a larger amount of litigation attempting to overturn previous decisions. (3) Finally, if we can find no difference in judicial outcomes depending on party control this would suggest that the substantial resources devoted by parties to win these elections represents a waste of resources from a societal perspective, equivalent to the social losses associated with rent seeking and political lobbying in the legislative branch of government.

Data and Empirical Findings

To measure the quality of state judicial systems we exploit a new survey-based ranking of state legal liability systems conducted by the U.S. Chamber of Commerce. This index scores state judicial quality on a scale of zero to 100 based on a nationwide survey of lawyers. (4) While there are potential problems with any survey-based measure, we can see whether we can replicate the results found by other studies comparing appointive and elective systems before proceeding to examine the impact of party control within elected systems. If our findings are similar, this would suggest that this measure is consistent with other, more specific and limited measures of judicial outcomes. However, as we proceed, we will also supplement our judicial quality data with additional data that potentially can reflect the quality of state judicial systems. These additional measures include state use of eminent domain for the takings of private property, state medical malpractice cases filed, and state unemployment compensation premiums.

Table 1 presents data showing the average state ranking and the underlying average score by method of judicial selection. As can be seen in the table, states selecting judges through appointment have an average ranking in the survey of 21.1 (out of 50), with an average underlying judicial quality score of 61.0, compared with an average ranking of 31.5 (out of 50), and an average underlying score of 53.5, for states selecting judges by elections. This finding is consistent with the studies cited earlier, namely that state judicial quality declines when elections are used to select judges. In Table 1 we also show the average ranking and underlying score for the two subsets of elections, partisan and nonpartisan. The difference between these two groups is even more striking than the difference between elected and appointed systems in general. States using partisan elections to select judges have a judicial quality score substantially lower than states using nonpartisan elections; however, both subsets remain below the average for states that appoint judges. Thus, the selection of judges by election lowers judicial quality, and this impact is even larger when those elections are of a partisan nature. Both of these results are consistent with the findings of previous literature. We now turn to regression methodology to see whether these differences remain once other variables are taken into account.

Our empirical specification takes the following form:

[Y.sub.i] = [[alpha].sub.i] + [[[beta].sub.i] + [X.sub.i] + [[delta].sub.i][Z.sub.i] +_ [[epsilon].sub.i]

where [Y.sub.i] = judicial quality within a state as measured by a state's score on the state liability system ranking, [X.sub.i] = a vector of control variables, [Z.sub.i] = 1 if a state elects its judges and 0 otherwise, and [[epsilon].sub.i] is the random error term. The coefficient of interest is [[delta].sub.i]. A negative and statistically significant estimate would support the hypothesis that the election of judges decreases the quality of a state's judicial system. A positive and statistically significant value of [[delta].sub.i] would imply that judicial elections lead to an increase in judicial quality. If [[delta].sub.i] is not significantly different from zero, it would suggest that the method of judicial selection does not influence judicial quality.

The vector of control variables, [X.sub.i], consists of state-level data that might influence [Y.sub.i], the measure of judicial quality in a state. Appendix Table 1 gives the sources and definitions of all of the variables used in the regression. The control variables we include reflect education levels within the state, number of lawyers per capita, percent voting Democratic in the 2000 presidential election, and judicial salary level. States with higher education levels are generally found to have better judicial quality. Although judicial salaries are generally included in an analysis such as ours, their impact is theoretically ambiguous. The number of lawyers per capita should be negatively related to a state's judicial quality because more lawyers lower the cost of litigation (Hanssen 1999). However, we run our regression both including and excluding this variable because of potential problems with endogeneity. It is, of course, likely that more lawyers result from a system with more litigation and redistribution.

The results of our regression analysis are presented in Table 2. As expected, the sign on ELECTIVE is negative and statistically significant at the 10 percent level. The coefficient on the elective variable can be interpreted as showing that a state moving from an elected system to an appointive system, other things equal, could expect an increase in their ranking on the U.S. Chamber of Commerce's index by around 4.5 points. Depending on where a state was in the rankings that could be quite a significant jump. For example, fewer than 4.5 points separate the 9th ranked state, Indiana, from the 26th ranked state, New Jersey.

Table 3 breaks down the elective variable between those states in which judges run in partisan elections and those in which elections are nonpartisan. Here we wish to see whether the real difference in judicial quality found in the regression results presented in Table 1 is not simply a difference between elective and appointive states, but rather between partisan elective states and appointive states. The regression results presented in Table 3 give some credence to the idea that it is the partisan nature of judicial elections, not just the elections themselves, which really influence state judicial quality. The sign on PARTISAN is negative and statistically significant at the 1 percent level. Although the sign on NONPARTISAN is negative, it is statistically insignificant at the 10 percent level. Thus, the results in Table 3 would suggest that there is not a significant difference in judicial quality between nonpartisan elective states and appointed states, but that there is a significant difference for states using partisan elections. In other words, the divergence seen between elective states and appointive states is really due to the subset of elective states using partisan elections.

We now turn to an examination of whether party control of state supreme courts matters in states with partisan elections. Unfortunately, given the relatively small number of states in these subsamples, regression methodology cannot be used. However, Table 4 presents the mean values of our key measures of judicial quality, split by party control of the judicial system, as well as showing the overall means for elected versus appointed system states.

First, looking at the overall judicial quality scores, there appears to be a sizable difference between partisan election states in which the majority of the state's highest court is Democratic and those in which the majority is Republican. The average score for states with Republican court majorities is 52, suggesting a higher level of judicial quality than in states with Democratic court majorities, whose average score is 41. In Table 4 we also present a few additional, more specific measures of outcomes of the judicial process including the usage of eminent domain, workers' compensation premiums, and medical malpractice suits filed. In states with "worse" judicial systems we would expect more use of eminent domain (less secure property rights), higher workers' compensation premiums, and more medical malpractice lawsuits. From Table 4, it is clear that eminent domain filings occur more often in states where Republicans control the judiciary, with one filing for every 55,054 state residents compared with one filing for every 605,411 residents of partisan election states where Democrats control the highest court.

Interestingly, when Democratic and Republican controlled courts are averaged together, under the category of all elected and then compared with all appointed, this huge difference is completely masked away by the result of averaging. Simply comparing elected versus appointed systems might be meaningless without explicitly controlling for party control, a potentially important problem not addressed by the previous literature on the issue. Workers' compensation premiums (largely dependent on court abuse of lenient benefit provision) are twice as high in states with Democratic controlled high courts. Again, this huge difference is masked when they are averaged and compared with appointed states. Finally, in the case of medical malpractice claims, which we would expect to be associated with judicial quality, there does not appear to be any significant difference between partisan election states where Republicans control the courts and states where the high court is controlled by the Democrats.

Conclusion

The results presented here confirm previous research showing that judicial quality is lower in states that utilize elections to select their judges. Utilizing a new dataset measuring judicial quality across the 50 U.S. states we also find it is the partisan nature of judicial elections that is the primary reason for lower judicial quality in elective states that utilize partisan elections. Although regression analysis is not possible due to the limited nature of the data available, we present some evidence that party control of a state's supreme court seems to matter for judicial quality and some outcomes generally considered to be related to judicial quality, such as workers' compensation premiums and eminent domain usage.

As our results seem to suggest, the outcomes under partisan elections, when taken as an overall average, tend to obfuscate the markedly different outcomes between Democratic and Republican controlled courts. This party difference is important, but has been ignored in previous research. Given the vast amount of previous literature suggesting party control of the legislative and executive branch matters, and the commonplace acceptance of at least controlling for party control in empirical models, we believe that this phenomenon is equally important to control for in empirical analyses of judicial outcomes. Thus we view the main contribution of our article as the insight that it is the partisan nature of elections that causes judicial quality to decline, not simply the electoral process as has been the commonly accepted wisdom from previous research.

More important, this study suggests that the large sums of money spent to influence judicial races are being spent because money can have an impact if it affects which party controls a state's supreme court. This has important implications for those analyzing judicial behavior and the impact of judicial selection processes on judicial outcomes. Models of legislative behavior have long included party affiliation. Perhaps it is time models of judicial behavior do the same.

Appendix: Description of Data

Appendix Table 1 provides a description of the variables used in this article and a detailed list of the data sources.

APPENDIX TABLE 1 DESCRIPTION OF DATA

Variable Definition

Judicial Quality Ranking of the quality of a state's
 liability system on a 100 point scale.

Education Percentage of the state's population 25
 and older with at least a college
 degree in 2003.
Population State population in 2003.
Lawyers Number of lawyers in 2003 divided by
 2003 state population.
Salary Salary of judges of state courts of last
 resort in 2003.
Elective Dummy variable equal to 1 for states
 that elected judges in 2003.
Appointive Dummy variable equal to 1 for states
 that appointed judges in 2003.
Partisan Dummy variable equal to 1 for states
 that elected judges in a partisan
 election in 2003.
Nonpartisan Dummy variable equal to 1 for states
 that elected judges in a nonpartisan
 election in 2003.
Democrat Percentage of the total state popular vote
 for the Democratic presidential
 candidate during the 2000 election.
Medical Paid medical malpractice claims per
 Malpractice 1,000 active, nonfederal physicians,
 2003.
Workers' State workers' compensation benefits per
 Compensation $100 of covered wages, 2003.
Eminent Domain Number of state residents per filed
 eminent domain taking, 1998-2002.
Republican Dummy variable equal to 1 if a state has
 Control partisan elections and a maajority of its
 Supreme Court was Republican in
 2003.
Democratic Dummy variable equal to 1 if a state has
 Control partisan elections and a majority of its
 Supreme Court was Democratic in
 2003.

Variable Source

Judicial Quality State Liability Systems Ranking Study
 (2004), U.S. Chamber of Commerce,
 Washington, D.C.
Education Statistical Abstract of the United States:
 2004-2005 (2004), Bureau of the Census,
 Department of Commerce, Washington,
 D.C.
Population Statistical Abstract of the United States:
 2004-2005 (2004), Bureau of the Census,
 Department of Commerce, Washington,
 D.C.
Lawyers The Lawyer Statistical Report (2004),
 American Bar Foundation, Chicago.
Salary The Book of the States (2004), The Council
 of State Governments, Lexington, Ky.
Elective The Book of the States (2004), The Council
 of State Governments, Lexington, Ky.
Appointive The Book of the States (2004), The Council
 of State Governments, Lexington, Ky.
Partisan The Book of the States (2004), The Council
 of State Government, Lexington, Ky.
Nonpartisan The Book of the States (2004), The Council
 of State Governments, Lexington, Ky.
Democrat Statistical Abstract of the United States:
 2004-2005 (2004), Bureau of the Census,
 Department of Commerce, Washington,
 D.C.
Medical "Number of Paid Medical Malpractice
 Malpractice Claims, 2003," Kaiser Family Foundation,
 www.statehealthfacts.org.
Workers' Workers' Compensation: Benefits, Coverage,
 Compensation and Costs, 2003 (2005), National Academy
 of Social Insurance, Washington, D.C.
Eminent Domain Public Power, Private Gain (2003), Institute
 for justice, Washington, D.C.
Republican Author calculations using justice party
 Control affiliation information available at "Judical
 Selection in the States," American
 Judicature Society, www.ajs.org/js/.
Democratic Author calculations using justice party
 Control affiliation information available at "Judical
 Selection in the States," American
 Judicature Society, www.ajs.org/js/.

References

Besley, T., and Payne, A. (2003) "Judicial Accountability and Economic Policy Outcomes: Evidence from Employment Discrimination Charges." Working Paper, London School of Economics.

Dorn, J. A. (1985) "Introduction: Economic Liberties and the Judiciary." Cato Journal 4 (Winter): 661-87.

Dubois, P. (1986) "Accountability, Independence, and the Selection of State Judges." Southwestern Law Journal 40 (May): 31-52.

Geyh, C. G. (2003) "Why Judicial Elections Stink." Ohio State Law Journal 64: 43-79.

Hall, M. G., and Brace, P. R. (1996) "Justices' Response to Case Facts." American Politics Quarterly 24: 237-61.

Hanssen, F. A. (2000) "Independent Courts and Administrative Agencies: An Empirical Analysis of the States." Journal of Law, Economics, and Organization 16: 534-71.

-- (2002) "On the Politics of Judicial Selection: Lawyers and State Campaigns for the Merit Plan." Public Choice 110 (January): 79-97.

-- (2004) "Learning about Judicial Independence: Institutional Change in the State Courts." Journal of Legal Studies 33 (June): 431-73.

Helland, E., and Tabarrok, A. (2002) "The Effect of Electoral Institutions on Tort Awards." American Law and Economics Review 4 (Fall): 341-70.

Kreft, S. F., and Sobel, R. S. (2005) "Public Policy, Entrepreneurship, and Economic Freedom." Cato Journal 25 (Fall): 595-616.

Link, B. (2004) "Had Enough in Ohio? Time to Reform Ohio's Judicial Selection Process." Cleveland State Law Review 51:123-52.

Lopez, E. J., and Sutter, D. (2004) "Ignorance in Congressional Voting? Evidence from Policy Reversal on the Endangered Species Act." Social Science Quarterly 85 (December): 891-912.

Ovaska, T., and Sobel, R. S. (2005) "Entrepreneurship in Post-Socialist Economies." Journal of Private Enterprise 21 (Fall): 8-28.

Pjesky, R., and Sutter, D. (2002) "Searching for Cincinnatus: Representatives' Backgrounds and Voting Behavior." Atlantic Economic Journal 30 (March): 74-86.

Tabarrok, A., and Helland, E. (1999) "Court Politics: The Political Economy of Tort Awards." Journal of Law and Economics 42 (April): 157-88.

U.S. Chamber of Commerce (2004) State Liability Ranking Study: 2004. Washington: U.S. Chamber of Commerce.

(1) In some states the governor can only appoint from a slate of candidates put forth by a nominating commission., however, the governor generally appoints some or all of the members of this nominating committee. For this reason, the literature generally does not make a distinction between these two alternatives (see Helland and Tabarrok 2002).

(2) The states using partisan elections are Alabama, Illinois, Louisiana, Michigan, Pennsylvania, Ohio, Texas, and West Virginia. Note that while party affiliations do not appear on the ballot in Ohio, candidates are nominated for the court by a party and thus are considered to be running in a partisan election.

(3) See Hanssen (2002) for a discussion of why lawyers themselves prefer judicial selection processes that promote uncertainty, such as merit plans for selecting judges.

(4) The ranking uses a random survey of approximately 1,000 lawyers throughout the United States. The lawyers were first asked which states they were familiar with and then were asked to evaluate those states' legal systems on a variety of criteria. These criteria pertained to the perceived quality of a state's legal system, such as judicial impartiality, judicial competence, and overall treatment of tort and contract litigation. Their answers were then used to construct the overall ranking.

Russell S. Sobel is James Clark Coffman Distinguished Chair and Professor of Economies and Joshua C. Hall is Ken and Bandy Kendrick Fellow, respectively, at West Virginia University.

TABLE 1
AVERAGE JUDICIAL QUALITY RANKING BY SELECTION
PROCESS, 2004

Type of System Average Ranking Average Score

Appointed 21.1 61.0
Elected 31.5 53.5
 Partisan 39.9 47.9
 Nonpartisan 26.4 56.9

SOURCE: U.S. Chamber of Commerce (2004).

TABLE 2
DETERMINANTS OF STATE LEGAL LIABILITY RANKINGS

 OLS Estimates

Variables Model 1 Model 2

Constant 43.640 *** 49.19 ***
 (4.43) (5.18)
Elective -4.524 * -4.629 *
 (1.72) (1.73)
Education 1.063 *** 0.811 ***
 (3.28) (2.74)
Salary 0.00003 0.00002
 (0.35) (0.27)
Democrat -0.166 -0.291 *
 (0.90) (1.69)
Lawyers -2.445 * --
 (1.72) --
[R.sup.2] adjusted 0.261 0.229
Observations 50 50

NOTE: * indicates significance at the 10 percent level,
** at the 5 percent level, and *** at the 1 percent level.
Absolute t-statistics are in parentheses.

TABLE 3
COMPARISON OF PARTISAN VS. NONPARTISAN ELECTIONS

 OLS

Variables Model Model

Constant 42.373 *** 39.881 ***
 (4.43) (3.90)
Partisan -9.360 *** -9.526 ***
 (2.64) (2.61)
Nonpartisan -2.091 -1.364
 (0.73) (0.47)
Education 0.935 ** 0.621 **
 (2.91) (2.22)
Salary 0.00005 0.00002
 (0.71) (0.30)
Democrat -0.154 0.010
 (0.86) (1.19)
Lawyers -2.072 --
 (1.49) --
[R.sup.2] adjusted 0.306 0.272
Observations 50 50

NOTE: * indicates significance at the 10 percent level,
** at the 5 percent level, and *** at the 1 percent level.
Absolute t-statistics are in parentheses.

TABLE 4
COMPARISON OF OUTCOMES BY JUDICIAL
CONTROL OF STATE SUPREME COURT

Variable Republican Democratic

Judicial Quality 52.04 40.97
Eminent Domain Filings 55,054 605,411
Workers' Compensation Benefits 1.09 2.18
Medical Malpractice Claims 19.06 19.90
Number of observations 5 3

Variable Elective Appointive

Judicial Quality 53.50 61.00
Eminent Domain Filings 649,719 535,418
Workers' Compensation Benefits 1.26 1.25
Medical Malpractice Claims 17.13 17.36
Number of observations 21 29