Voter Preferences, Institutions, and Economic Freedom

Transcription

1 Voter Preferences, Institutions, and Economic Freedom George R. Crowley, John A. Dove, and Daniel Sutter* Manuel H. Johnson Center for Political Economy, Troy University Author s: Crowley, Dove, Sutter, The authors thank the Free Market Institute at Texas Tech and the Templeton Foundation for financial support. 1

2 1.Introduction Economists since Adam Smith have extolled the virtues of the market economy in generating wealth and prosperity. The last twenty years have witnessed the compilation of two metrics of economic freedom across nations, to measure in a consistent and systematic fashion how closely nations approach the ideal of voluntary transactions. The measures have enabled a new generation of empirical studies providing important new evidence on the beneficial effects of markets. Nations with more economic freedom have higher standards of living, faster economic growth, less poverty and extreme poverty, longer life expectancy, and more health and happiness. In short, almost all of the good things in life are more prevalent in nations with more economic freedom. The enormous benefits of economic freedom make an understanding of the factors or forces affecting its level in a nation become paramount. Across the globe the good news has been an increase in economic freedom as measured by the Fraser Institute, from 5.34 in 1980 to 6.87 in 2011 (Gwartney, Lawson and Hall, 2013, p.15). Yet the 20 th Century witnessed substantial growth in the public sector across the world, and the United States has fallen from 2 nd internationally in 2000 to 17 th place in 2011 (with a score decline from 8.65 to 7.74 over the period). Substantial changes in direction are possible even in short time periods; the U.S. witnessed a pronounced shift toward greater government in the financial crisis and elections of 2008, followed two years later by election of a Republican majority in the House of Representatives led by Tea Party fiscal conservatives. As a consequence, discretionary federal fiscal spending is expected to decline in Fiscal Year 2014 for the second straight year. Economist Tyler Cowen observed in 2000 (p.1) that questions regarding the determinants of freedom do not always hold a central place in mainstream academic [economic] discourse. 2

3 The determinants of economic freedom have begun to receive considerably more attention from researchers over the past 15 years. Nonetheless, regression analyses of economic suggest that much of the variation remains to be explained (Leeson, Ryan and Williamson 2012, Hefner and Witte 2013). We use a newly constructed vote index for the U.S. Congress tied to the components of the Fraser Institute s Economic Freedom of the World index as a control for citizen preferences in an analysis of state-level economic freedom using the Fraser Institute s Economic Freedom of North America index. Controlling for citizen preferences for freedom allows us to shed new light on the role of political or ideological persuasion in social change the old question of whether everyone must become a libertarian for libertarian policies to prevail. Yet our measure of preferences is based on a state s Congressional economic freedom vote score, and not decisions by state government and thus is independent of the policies measuring state economic freedom. Our preliminary results show that Congressional voting is robustly correlated with state economic freedom, and in the expected direction. This paper proceeds as follows. Section 2 reviews some of the relevant literature examining changes in economic freedom. Section 3 provides a rationale for and discussion of the economic freedom vote index for Congress. Section 4 presents an analysis of state level economic freedom using a state s Congressional delegation s economic freedom vote. Section 5 offers a brief conclusion and directions for future research. 2. Literature Review Economic freedom has far more often been used as a right hand side variable in regression analysis than as a dependent variable. Nonetheless, the endogeneity of the institutions of the market has been raised in a number of areas in prior research. Most of the analysis to date has 3

4 focused on economic freedom at the national level, which is perhaps natural given the wide variation in reliance on markets across nations relative to across states in the U.S. A number of papers have considered the potential for both political institutions and economic institutions to diffuse across national borders (Leeson and Sobel 2007, Dean and Leeson 2009). Other papers have explored the impact of particular events on freedom. For instance, Hall, Lawson and Wogsland (2011) find that joining the European Union resulted in a statistically significant but modest in magnitude increase in economic freedom. La Porta et al. (2002) find that constitutional and legal regimes are associated with economic freedom. Several studies have attempted to advance the analysis of the effect of economic freedom on growth by establishing causality between freedom and growth. Causality tests consequently allow the potential endogeneity of economic freedom, and specifically that growth could contribute to freedom. Dawson (2003) finds that the level of economic freedom in a nation and the use of markets and property rights components of the Fraser index, Granger cause economic growth, with no reverse causality. Changes in economic freedom are closely related to growth, but causality here runs in both directions, while both the level of freedom and changes in freedom cause investment, with no evidence of reverse causality. Farr, Lord and Wolfenberger (1998) find that the level of economic freedom Granger causes the level of GDP per capita, while Heckelman (2000) also finds that economic freedom and its components Granger cause growth. Hefner and Witte (201x) find evidence of a unit root in the level of economic freedom in a panel of nations. Because economic freedom scores are bounded, economic freedom may not technically be able to follow a random walk, but their result does suggest that shocks to economic freedom, when they occur, tend to persist. They also find that a wide range of 4

5 variables describing an economy, like the shares of manufacturing, mining, and agriculture in GDP, and the consumption of goods which might portend social change (e.g., appliances and media) overall have little explanatory power over changes in economic freedom. Hefner and Witte s interesting analysis leaves the question of what does indeed drive changes in economic freedom unclear, but does allow possible scope for political and cultural entrepreneurs to change a nation s course. One previous attempt to use voter preferences as a determinant of economic freedom is Crampton (2002). He examines a panel of 25 OECD nations and uses a measure of the median voter derived from the Manifesto Research Group analysis of party platforms. Changes in freedom preferences measured by manifesto positions on policy matters like regulation, free enterprise, and the welfare state are found to be significant determinants of changes in economic freedom, and in the expected direction. The result is robust to inclusion of variables controlling for legal or political institutions. Crampton s method of measuring voter preferences though draws on the same policies included in measures of economic freedom. For instance, Margaret Thatcher s election would shift the calculated position of the median voter in Britain and result in policy changes reflected in an increase in economic freedom. In our analysis we use voting on federal issues as a measure of voter preferences in a state, so that the policy acts measuring preferences are not part of state level economic freedom. Most of the research to date has examined economic freedom at the national level. One exception is Leeson, Ryan and Williamson (2012), who examine the policy components of state economic freedom. They focus on spending by state based free market think tanks, along with spending by lobbying groups, although they also include gross state product per capital and several measures of partisan control of state government in robustness analysis. Think tank and 5

6 political lobbying spending has little systematic effect on the policies related to economic freedom, and regression analysis of several components of freedom exhibit little explanatory power, despite the inclusion of state and year fixed effects. 1 The determinants of state level economic freedom remain murky. 3. Congressional Economic Freedom Vote Index Methodology Many Congressional vote indices exist, raising the question of the marginal value of any new index. While indices have been compiled by groups generally supportive of limited government and freer markets, no index to date attempts to track votes on economic freedom directly and comprehensively. The index which comes the closest is compiled by FreedomWorks. Other indices either contain votes dealing with ideological or social issues unrelated to economic freedom, like the American Conservative Union (ACU), or are focused only on one aspect of economic freedom, like the National Taxpayers Union. An example from the U.S. Chamber of Commerce s How They Voted index illustrates a problem of inconsistency with economic freedom. The Chamber s 2012 vote index for the U.S. Senate includes roll call vote #95, to reauthorize the Export-Import Bank, which has been identified by fiscal conservatives as a protectionist agency that provides political privileges to well-connected firms at the expense of all other citizens (de Rugy and Castillo 2014, p.4). 2 We construct an index scoring each member of each house of the 112 th U.S. Congress ( ) based on votes affecting economic freedom in the United States. Our index is based on selected (as discussed below) roll call vote data for years 2011 and Roll call votes are scored as either improving or worsening economic freedom, with a 1 indicating a vote in favor of 1 They do however find that state based think tanks appear to significantly affect attitudes toward markets and limited government as reflected in the General Social Survey. 2 The FreedomWorks vote score counted a vote for the U.S. Export-Import Bank as a vote against freedom in

7 freedom and a -1 a vote against. The average across all included votes is a senator or representative overall score, and also ranges from -1 to 1. We define economic freedom based on the Fraser Institute s Economic Freedom of the World and Economic Freedom of North America indices. The indices endeavor to measure the degree to which economic decision-making is left to markets as opposed to subject to government control. We consider legislation to expand the role of government as adversely affecting economic freedom, while legislation to either expand market organization or curtail the role of the state increases economic freedom. Examples of economic-freedom-enhancing legislation include tax code simplification or deregulation, while legislation to expand regulatory power or limit international trade would decrease economic freedom. We examined each individual roll call vote and determined those which most directly impacted economic freedom in the United States. First, purely procedural votes (such as those which approve the Journal, determine a quorum, or express a sense of Congress) were excluded. Of the remaining non-procedural votes, those not pertaining to actual legislation (such as confirmations of government officials or the ceremonial naming of government buildings) were dropped. Similarly, we excluded votes on legislation not clearly affecting economic freedom (such as a spending bill including a multitude of provisions or the Violence Against Women Act). 3 Finally, we excluded any remaining relevant votes which did not themselves affect economic freedom (for instance to recommit legislation, or to invoke cloture and end 3 Votes on legislation which affected only spending (such as appropriations bills) were excluded. Most of these bills include so many specific spending programs that it is quite difficult to assess the overall impact on economic freedom. While it may seem that all spending should adversely affect economic freedom, this is naïve for at least two reasons. First, spending on services such as law enforcement likely improves the legal environment. Secondly, while economic freedom indices include a size of government component, it is unclear to what extent a given vote (which likely increases spending on some programs and cuts spending on others) will change the relative size of government; for example a hypothetical vote which simply renewed all spending programs at current levels would not affect a change in the size of government and so would have no impact on economic freedom. 7

8 debate) or did so in an insignificant way (such as requiring the EPA conduct a study). The included votes were therefore either on passage of final legislation or on approval of amendments to pending legislation directly and clearly affecting economic freedom. As summarized in Table 1, our index includes over 160 House votes and nearly 90 Senate votes, roughly 12% of all roll call votes in the 112 th Congress. 4 [Insert Table 1 about here] For each included vote, we gathered the roll call data from the official House and Senate websites. Votes for legislation (or amendments to pending legislation) increasing economic freedom and votes against legislation adversely affecting economic freedom were scored as 1. A score of -1 was assigned to yea votes adversely affect economic freedom or votes against legislation improving economic freedom. 5 Votes of present and abstentions were coded as 0. Each member of Congress was then assigned a score calculated as an average of their vote scores, with all votes weighted equally. A hypothetical Senator or Representative who voted for legislation that improved economic freedom 100% of the time would score a 1, while the hypothetical member of Congress consistently voting in favor of decreasing economic freedom would score a -1. Our index identifies Mike Lee (R-UT) and Rand Paul (R-KY), with scores of 0.67 and 0.66 respectively, as the Senators with the most economically free voting records in the 112 th Congress. The Senators voting most consistently against economic freedom were Bernie Sanders (a self-described socialist of VT) and a tie between Jack Reed and Sheldon Whitehouse 4 While care was taken to include all votes affecting economic freedom, the assignment is necessarily subjective to some extent. We chose to err in classification by excluding ambiguous votes. A complete list of all votes included, as well as how we chose to score the votes is available upon request. 5 We chose to equally weight yea votes on freedom-improving legislation and nay votes against freedomreducing legislation; likewise for yea votes on freedom-reducing legislation and nay votes for freedom-improving legislation. 8

9 (both D-RI), with scores of and respectively. In the House, Jeff Reed (R-AZ) and Tom Graves (R-GA) had the freest voting records (0.75 and respectively), while Ed Markey (D-MA) and Brad Miller (D-NC) voted most consistently against economic freedom (with scores of and -0.66). The overall average scores are for the Senate and 0.19 for the House. In addition to this overall measure, we also categorized the roll call votes by the area of the Economic Freedom of the World most directly affected. Some components of the index had only a small number of votes, so we focus on sub-indices for the Size of Government, Takings and Discriminatory Taxation, and Labor Market Freedom components, and constructed following the method described above. 4. Data and Model Specification 4.1 Empirical Specification and Description of the Data In order to evaluate the impact, if any, of the Congressional economic freedom scores on statelevel economic freedom, we estimate a number of regression models. The baseline model takes the following form: StateEFS i = α i + CongressionalEFS i + Z i + ε i (1) Where StateEFS i is the economic freedom score in state i for CongressionalEFS i is the economic freedom score for each state s congressional delegation to the U.S. Congress, while Z i is a vector of socioeconomic control variables detailed below. The state economic freedom scores come from the Economic Freedom of North American Index (EFNA), compiled annually by the Frasier Institute. 7 This index provides an was chosen as the baseline year of analysis as it is the most recent year in which state economic freedom scores are available. 9

10 overall score for each state s level of economic freedom on a scale of 10, with 0 being the least economically free and 10 being the most economically free. This overall score is derived from a number of sub-scores in various categories, which include measures on the Size of Government, Takings and Discriminatory Taxation, Labor Market Freedom, Regulation of Credit Markets, Business Regulations, Legal System and Property Rights, and Sound Money. The first three sub-scores are specific to each state, while the latter scores are taken from the Economic Freedom of the World Index. Therefore, along with the overall state economic freedom score, we also employ the Size of Government, Takings and Discriminatory Taxation, and Labor Market Freedom sub-scores as separate dependent variables. We use three different tallies of Congressional votes, the average scores for the state s House and Senate delegations, and the average for the state s entire Congressional delegations, with scores for the three corresponding sub-components listed above calculated similarly. Thus, for each of the EFNA scores, we estimate six regressions using each of the main independent variables of interest both with and without controls. Finally, along with looking at just the EFNA scores for 2011, we also consider 5 and 10-year averages of the EFNA scores as dependent variables. Overall, this provides a total of 18 separate regressions for each category under analysis. We also include a number of socioeconomic control variables identified in the literature as possible determinants of state economic freedom. These include the state population (in 10,000s), population density, the median age in each state, the percent of the population that is male, the percent of the population that is white, the unemployment rate in each state, per capita gross state product, and the percent of each state s population with a bachelor s degree or 7 Data with detailed descriptions on the construction of the index are available at 10

11 higher. 8 Finally, we also include a set of dummy variables based on U.S. Census regions. The regional dummies were included to mitigate any omitted variable bias resulting from the inability of cross-sectional regression analysis to detect unobserved differences across states. Table 2 provides the summary statistics for each of the variables discussed above. [Insert Table 2 About Here] Section 4.2 Results Overall some interesting results emerge from the regression analysis. Table 3 considers the overall economic freedom scores for each state with the overall Congressional economic freedom score included as the main independent variable of interest. [Insert Table 3 About Here] All of our tables of regression results are organized similarly. The first six columns present regressions against the state s House, Senate, and overall Congressional delegation economic freedom scores, both with and without controls. Columns 7 through 12 display the results with a 5-year average EFNA score as the dependent variable, while columns 13 through 18 employ a 10-year average of the overall EFNA score. The state Congressional economic freedom score increases state economic freedom in each specification in table 3, and the effects are statistically and economically significant. The overall state delegation exhibits greater explanatory power than either chamber score, suggesting that averaging more vote scores provides a better measure of state preferences regarding economic freedom. Next, table 4 regresses the Size of Government EFNA sub-score against each of the overall Congressional economic freedom scores. 8 State population, population density, median age, percent male, percent white, and educational attainment were all taken from U.S. Census data. These data are freely available at The unemployment rate and gross state product data were taken from the Bureau of Labor Statistics, with data freely available at Each of these variables is from 2011, except educational attainment which is from 2009, the most recent year available. 11

12 [Insert Table 4 About Here] The results would appear far less robust than those found in table 3. Of note, the Congressional economic freedom vote score has the expected positive sign in each case, but attains statistical significance in only 9 of the 18 specifications, and only once at better than the 10% level. Further, 6 of the 9 statistically significant specifications do not contain the controls. Table 5 displays regression of the EFNA Size of Government sub-category against the Size of Government sub-category from the Congressional economic freedom scores. [Insert Table 5 About Here] As can be seen, the Congressional vote score has a positive and economically significant impact in each case, but only 2 of the 18 results attain statistical significance. Tables 6 and 7 show regressions of the EFNA Taxation sub-category against the overall Congressional economic freedom score and Taxation Congressional economic freedom sub-score respectively. [Insert Table 6 About Here] [Insert Table 7 About Here] Table 6 shows that 10 of the 18 results are statistically significant, with all results economically significant and positive. However, 9 of the 10 significant results occur only without the control variables. Table 6 presents the results for the Taxation Congressional economic freedom score sub-index. Here again statistical significance is rarely attained (only 4 of the 18 results). Yet the congressional vote score generally has a positive and economically significant coefficient, with only 1 specification (in column 14) netting a negative result. 12

13 Finally, tables 8 and 9 present regression of the EFNA Labor Market Freedom score against overall Congressional economic freedom score and Congressional Labor Market Freedom sub-scores respectively. [Insert Table 8 About Here] [Insert Table 9 About Here] The results for the Labor Market Freedom component are quite strong. The Congressional vote score variable attains statistical significance in each specification in table 8 and 12 of 18 instances in table 9. Senate vote scores, both overall and for the sub-index, have more explanatory power over state labor market policy, observed by both the goodness-of-fit of the specifications and the insignificance of the House delegation sub-index in table 9. By contrast, the overall congressional delegation score correlated more closely with overall state economic freedom. In particular, the Senate labor market freedom vote sub-index correlates extremely highly with state labor market freedom in table Conclusion and Directions for Future Research Economic freedom offers a cornucopia of benefits, making an understanding of factors affecting its expansion and recession an important task for economists. We have employed a newly constructed index of economic freedom for members of the U.S. Congress as a means to control for voter preferences in an analysis of state level economic freedom. The results, which should be considered highly preliminary, are encouraging. Voting in Congress on matters related to economic freedom seems to correlate with state policies affecting the balance between markets and government. The results are robust to inclusion of socioeconomic control variables as well as regional dummy variables. 13

14 The alignment of state economic freedom and Congressional voting suggests the relative unimportance of sectional economic interests. Public choice theory suggests that voters will vote their economic interests, which might differ in federal versus state policy decisions. Economic self-interest, for instance, might lead representatives from agricultural states to vote for federal crop price supports, despite an underlying preference for economic freedom. Our results, although clearly preliminary, suggest that at least on average preferences for markets are consistent across jurisdictional levels. 9 The index of Congressional voting employed here has been constructed for just one session of Congress, necessitating a cross-sectional analysis. The robustness of the results obtained here suggests that extending the congressional economic freedom vote index to prior years could yield a valuable tool for research on the determinants of economic freedom. 9 Conceivably Representatives and Senators all vote their district s or state s economic interests against freedom when issues arise, but the opportunities to benefit from government interventions are equally distributed across states. 14

15 References Cowen, Tyler Why Does Freedom Wax and Wane? Some Research Questions in Social Change and Big Government. Arlington VA: Mercatus Center. Available at: 20Wane.pdf Crampton, Eric You Get What You Vote For: Electoral Determinants of Economic Freedom. Journal of Private Enterprise, 18: Dawson, John W Causality in the Freedom-Growth Relationship. European Journal of Political Economy, 19: Dean, Andrea, and Peter T. Leeson The Democratic Domino Theory: An Empirical Investigation. American Journal of Political Science, 53(3): De Rugy, Veronique, and Andrea Castillo The US Export-Import Bank: A Review of the Debate over Reauthorization. Mercatus Center, available at: Farr, W. Kenn, Richard A. Lord, and J. Larry Wolfenbarger Economic Freedom, Political Freedom, and Well-Being: A Causality Analysis, Cato Journal, 18(2): Gwartney, James, Robert Lawson, and Joshua Hall Economic Freedom of the World: 2013 Annual Report. Vancouver, BC: Fraser Institute. Hall, Joshua C., Robert A. Lawson, and Rachael Wogsland The European Union and Economic Freedom. Global Economy Journal, 11(3), Article 7. Heckelman, Jac C Economic Freedom and Economic Growth: A Short-Run Causal Investigation. Journal of Applied Economics, 3: Hefner, Frank, and Mark David Witte A Random Walk to Economic Freedom? European Journal of Comparative Economics, 10(1): La Porta, Rafael, Florencio Lopez-de-Silanes, Cristian Pop-Eleches, and Andrei Shleifer The Guarantees of Freedom. Harvard Institute of Economic Research Discussion Paper #1943. Leeson, Peter T., Matt E. Ryan, and Claudia R. Williamson Think Tanks. Journal of Comparative Economics, 40: Leeson, Peter T., and Russell S. Sobel The Spread of Global Economic Freedom. In Economic Freedom of the World: 2007 Annual Report. Vancouver, BC: Fraser Institute. 15

16 Table 1: Summary of Roll Call Votes Included Year Chamber House 92/949, 9.7% 72/659, 10.9% Senate 32/235, 13.6% 57/251, 22.7% 16

17 Table 1: Summary Statistics Variables Observations Mean Std. Dev. Min Max EFNA Overall Score EFNA Overall Score 5-year average EFNA Overall Score 10-year average EFNA "Size of Government" Score EFNA "Size of Government" Score 5-year average EFNA (Size of Government" Score 10-year average EFNA "Taxation" Score EFNA "Taxation" Score 5-year average EFNA "Taxation" Score 10-year average EFNA "Labor Market Freedom" Score EFNA "Labor Market Freedom" Score 5-year average EFNA "Labor Market Freedom" Score 10-year average House Overall Economic Freedom Score Senate Overall Economic Freedom Score Average Delegation Overall Economic Freedom Score House "Size of Government" Economic Freedom Score Senate "Size of Government" Economic Freedom Score Average Delegation "Size of Government" Economic Freedom Score House "Taxation" Economic Freedom Score Senate "Taxation" Economic Freedom Score Average Delegation "Taxation" Economic Freedom Score House "Labor Market Freedom" Economic Freedom Score Senate "Labor Market Freedom" Economic Freedom Score Average Delegation "Labor Market Freedom" Economic Freedom Score Population (In 10,000s) Population Density Median Age % Male % White Unemployment Per Capita GDP % Bachelor's Degree Northeast Midwest West South

18 Table 2: Overall State EFNA Economic Freedom Score Dependent Variable = One, Five, and Ten Year Average Score House, Senate, and State Average Scores = Overall Congressional Economic Freedom (1) (2) (3) (4) (5) (6) (7) (8) (9) (10) (11) (12) (13) (14) (15) (16) (17) (18) VARIABLES One Year One Year One Year One Year One Year One Year Five Year Five Year Five Year Five Year Five Year Five Year Ten Year Ten Year Ten Year Ten Year Ten Year Ten Year Constant 6.449*** 5.628** 6.623*** *** *** 4.892* 6.729*** *** *** 5.559* 6.842*** *** (0.0917) (2.700) (0.0853) (2.703) (0.0834) (2.629) (0.0865) (2.541) (0.0801) (2.540) (0.0786) (2.483) (0.0881) (2.798) (0.0788) (2.719) (0.0788) (2.672) House 0.758*** 0.581* 0.696*** 0.497* 0.622*** 0.612* (0.209) (0.299) (0.197) (0.282) (0.201) (0.310) Senate 0.746*** 0.789*** 0.699*** 0.713** 0.726*** 0.927*** (0.226) (0.280) (0.213) (0.263) (0.209) (0.281) State Average 0.903*** 0.986*** 0.836*** 0.870*** 0.804*** 1.108*** (0.232) (0.336) (0.218) (0.317) (0.219) (0.341) Population (in 10,000s) ** * ** ** * ** * * ( ) ( ) ( ) ( ) ( ) ( ) ( ) ( ) ( ) Population Density * * ( ) ( ) ( ) ( ) ( ) ( ) ( ) ( ) ( ) Median Age * (0.0440) (0.0479) (0.0464) (0.0414) (0.0450) (0.0438) (0.0456) (0.0482) (0.0472) % Male (0.0385) (0.0357) (0.0364) (0.0363) (0.0335) (0.0344) (0.0399) (0.0359) (0.0370) % White e ( ) ( ) ( ) ( ) ( ) ( ) ( ) ( ) ( ) Unemployment (0.0545) (0.0503) (0.0499) (0.0513) (0.0472) (0.0472) (0.0565) (0.0506) (0.0508) Per Capita GDP 2.80e-05** 3.53e-05*** 3.40e-05*** 2.57e-05** 3.25e-05*** 3.11e-05*** 2.18e-05* 3.08e-05*** 2.88e-05*** (1.06e-05) (1.06e-05) (1.04e-05) (9.97e-06) (9.98e-06) (9.79e-06) (1.10e-05) (1.07e-05) (1.05e-05) % Bachelor's Degree * ** ** ** ** ** ** ** *** (0.0222) (0.0210) (0.0212) (0.0209) (0.0197) (0.0200) (0.0231) (0.0211) (0.0215) Northeast ** * ** * * (0.337) (0.318) (0.310) (0.317) (0.298) (0.293) (0.349) (0.319) (0.316) Midwest ** ** ** ** ** ** * (0.219) (0.211) (0.211) (0.206) (0.198) (0.199) (0.227) (0.212) (0.214) West *** *** *** *** *** *** *** ** ** (0.235) (0.237) (0.235) (0.222) (0.223) (0.222) (0.244) (0.238) (0.239) Observations R-squared Standard errors in parentheses *** p<0.01, ** p<0.05, * p<0.1 18

19 Table 3: "Size of Government" EFNA Subcomponent Score Dependent Variable = One, Five, and Ten Year Average Score House, Senate, and State Average Scores = Overall Congressional Economic Freedom (1) (2) (3) (4) (5) (6) (7) (8) (9) (10) (11) (12) (13) (14) (15) (16) (17) (18) VARIABLES One Year One Year One Year One Year One Year One Year Five Year Five Year Five Year Five Year Five Year Five Year Ten Year Ten Year Ten Year Ten Year Ten Year Ten Year Constant 6.053*** *** *** *** *** *** *** *** *** (0.145) (4.542) (0.134) (4.569) (0.134) (4.552) (0.133) (4.427) (0.122) (4.500) (0.122) (4.471) (0.135) (4.674) (0.122) (4.714) (0.123) (4.690) House 0.683** * (0.330) (0.503) (0.303) (0.490) (0.307) (0.518) Senate 0.651* 0.940* * 0.942* (0.355) (0.473) (0.323) (0.466) (0.324) (0.488) State Average 0.801** * * (0.372) (0.581) (0.340) (0.571) (0.343) (0.599) Population (in 10,000s) ( ) ( ) ( ) ( ) ( ) ( ) ( ) ( ) ( ) Population Density ( ) ( ) ( ) ( ) ( ) ( ) ( ) ( ) ( ) Median Age (0.0741) (0.0810) (0.0803) (0.0722) (0.0797) (0.0789) (0.0762) (0.0835) (0.0827) % Male (0.0648) (0.0603) (0.0631) (0.0632) (0.0594) (0.0620) (0.0667) (0.0622) (0.0650) % White e (0.0117) (0.0109) (0.0113) (0.0114) (0.0107) (0.0111) (0.0121) (0.0112) (0.0117) Unemployment (0.0917) (0.0850) (0.0865) (0.0894) (0.0837) (0.0850) (0.0944) (0.0877) (0.0891) Per Capita GDP 2.72e e-05** 3.39e-05* 2.33e e-05* 2.88e e e e-05 (1.78e-05) (1.80e-05) (1.79e-05) (1.74e-05) (1.77e-05) (1.76e-05) (1.83e-05) (1.85e-05) (1.85e-05) % Bachelor's Degree ** ** ** ** ** ** ** ** ** (0.0374) (0.0354) (0.0367) (0.0365) (0.0349) (0.0360) (0.0385) (0.0365) (0.0378) Northeast (0.567) (0.537) (0.538) (0.552) (0.529) (0.528) (0.583) (0.554) (0.554) Midwest (0.369) (0.357) (0.365) (0.359) (0.351) (0.358) (0.379) (0.368) (0.376) West *** ** ** ** * * ** (0.396) (0.401) (0.407) (0.386) (0.395) (0.400) (0.408) (0.413) (0.420) Observations R-squared Standard errors in parentheses *** p<0.01, ** p<0.05, * p<0.1 19

20 Table 4: "Size of Government" EFNA Subcomponent Score Dependent Variable = One, Five, and Ten Year Average Score House, Senate, and State Average Scores = "Size of Government" Congressional Economic Freedom (1) (2) (3) (4) (5) (6) (7) (8) (9) (10) (11) (12) (13) (14) (15) (16) (17) (18) VARIABLES One Year One Year One Year One Year One Year One Year Five Year Five Year Five Year Five Year Five Year Five Year Ten Year Ten Year Ten Year Ten Year Ten Year Ten Year Constant 6.385*** *** *** *** *** *** *** *** *** (0.167) (4.545) (0.137) (4.561) (0.134) (4.531) (0.154) (4.420) (0.125) (4.471) (0.123) (4.444) (0.157) (4.685) (0.125) (4.705) (0.124) (4.670) House 0.510* (0.257) (0.349) (0.236) (0.339) (0.241) (0.360) Senate (0.352) (0.450) (0.320) (0.442) (0.321) (0.465) State Average 0.762* (0.395) (0.601) (0.361) (0.590) (0.364) (0.619) Population (in 10,000s) ( ) ( ) ( ) ( ) ( ) ( ) ( ) ( ) ( ) Population Density * ( ) ( ) ( ) ( ) ( ) ( ) ( ) ( ) ( ) Median Age (0.0702) (0.0788) (0.0786) (0.0683) (0.0772) (0.0771) (0.0723) (0.0813) (0.0811) % Male (0.0633) (0.0629) (0.0640) (0.0616) (0.0617) (0.0628) (0.0652) (0.0649) (0.0660) % White (0.0117) (0.0111) (0.0115) (0.0113) (0.0109) (0.0112) (0.0120) (0.0115) (0.0118) Unemployment (0.0996) (0.0882) (0.0871) (0.0969) (0.0865) (0.0855) (0.103) (0.0910) (0.0898) Per Capita GDP 2.44e e-05* 3.20e-05* 2.14e e e e e e-05 (1.79e-05) (1.87e-05) (1.80e-05) (1.74e-05) (1.83e-05) (1.76e-05) (1.84e-05) (1.92e-05) (1.85e-05) % Bachelor's Degree ** ** ** * ** ** ** ** ** (0.0374) (0.0360) (0.0372) (0.0364) (0.0353) (0.0364) (0.0386) (0.0372) (0.0383) Northeast (0.555) (0.551) (0.543) (0.540) (0.541) (0.532) (0.572) (0.569) (0.559) Midwest (0.363) (0.366) (0.367) (0.353) (0.359) (0.360) (0.375) (0.378) (0.378) West *** ** ** ** ** * ** * * (0.385) (0.404) (0.405) (0.374) (0.396) (0.398) (0.397) (0.417) (0.418) Observations R-squared Standard errors in parentheses *** p<0.01, ** p<0.05, * p<0.1 20

21 Table 5: "Taxation" EFNA Subcomponent Score Dependent Variable = One, Five, and Ten Year Average Score House, Senate, and State Average Scores = Overall Congressional Economic Freedom (1) (2) (3) (4) (5) (6) (7) (8) (9) (10) (11) (12) (13) (14) (15) (16) (17) (18) VARIABLES One Year One Year One Year One Year One Year One Year Five Year Five Year Five Year Five Year Five Year Five Year Ten Year Ten Year Ten Year Ten Year Ten Year Ten Year Constant 6.457*** *** *** *** *** *** *** *** *** (0.121) (3.583) (0.119) (3.820) (0.114) (3.677) (0.121) (3.478) (0.117) (3.667) (0.113) (3.561) (0.120) (3.462) (0.114) (3.627) (0.112) (3.508) House 1.081*** *** *** (0.274) (0.397) (0.275) (0.385) (0.273) (0.384) Senate 0.807** ** ** (0.315) (0.395) (0.311) (0.380) (0.304) (0.375) State Average 1.142*** *** *** 0.853* (0.316) (0.470) (0.315) (0.455) (0.311) (0.448) Population (in 10,000s) ** * * ** * * * * * ( ) ( ) ( ) ( ) ( ) ( ) ( ) ( ) ( ) Population Density * * * * ( ) ( ) ( ) ( ) ( ) ( ) ( ) ( ) ( ) Median Age (0.0584) (0.0677) (0.0649) (0.0567) (0.0650) (0.0628) (0.0565) (0.0643) (0.0619) % Male (0.0511) (0.0504) (0.0510) (0.0496) (0.0484) (0.0493) (0.0494) (0.0478) (0.0486) % White ( ) ( ) ( ) ( ) ( ) ( ) ( ) ( ) ( ) Unemployment * * (0.0724) (0.0711) (0.0699) (0.0702) (0.0682) (0.0677) (0.0699) (0.0675) (0.0667) Per Capita GDP 5.62e-05*** 5.90e-05*** 6.00e-05*** 5.61e-05*** 5.91e-05*** 5.94e-05*** 5.89e-05*** 6.38e-05*** 6.36e-05*** (1.41e-05) (1.50e-05) (1.45e-05) (1.36e-05) (1.44e-05) (1.40e-05) (1.36e-05) (1.43e-05) (1.38e-05) % Bachelor's Degree (0.0295) (0.0296) (0.0296) (0.0287) (0.0284) (0.0287) (0.0285) (0.0281) (0.0283) Northeast * ** ** ** * * (0.447) (0.449) (0.434) (0.434) (0.431) (0.421) (0.432) (0.426) (0.414) Midwest ** ** ** ** ** ** * * * (0.291) (0.298) (0.295) (0.282) (0.286) (0.285) (0.281) (0.283) (0.281) West ** ** ** *** ** ** ** ** ** (0.312) (0.335) (0.329) (0.303) (0.322) (0.319) (0.302) (0.318) (0.314) Observations R-squared Standard errors in parentheses *** p<0.01, ** p<0.05, * p<0.1 21

22 Table 6: "Taxation" EFNA Subcomponent Score Dependent Variable = One, Five, and Ten Year Average Score House, Senate, and State Average Scores = "Taxation" Congressional Economic Freedom Sub-Index (1) (2) (3) (4) (5) (6) (7) (8) (9) (10) (11) (12) (13) (14) (15) (16) (17) (18) VARIABLES One Year One Year One Year One Year One Year One Year Five Year Five Year Five Year Five Year Five Year Five Year Ten Year Ten Year Ten Year Ten Year Ten Year Ten Year Constant 6.723*** *** *** *** *** *** *** *** *** (0.125) (3.682) (0.127) (3.677) (0.125) (3.680) (0.123) (3.538) (0.124) (3.518) (0.123) (3.538) (0.121) (3.565) (0.120) (3.504) (0.121) (3.564) House 0.451** * (0.219) (0.236) (0.215) (0.227) (0.212) (0.229) Senate (0.632) (0.671) (0.618) (0.642) (0.600) (0.639) State Average 0.852** * (0.391) (0.414) (0.385) (0.398) (0.379) (0.401) Population (in 10,000s) * ** * * ** * * ** * ( ) ( ) ( ) ( ) ( ) ( ) ( ) ( ) ( ) Population Density ** * ** ** * ** ( ) ( ) ( ) ( ) ( ) ( ) ( ) ( ) ( ) Median Age (0.0582) (0.0563) (0.0577) (0.0560) (0.0538) (0.0555) (0.0564) (0.0536) (0.0559) % Male * * * (0.0513) (0.0509) (0.0503) (0.0493) (0.0487) (0.0484) (0.0496) (0.0485) (0.0487) % White ( ) ( ) ( ) ( ) ( ) ( ) ( ) ( ) ( ) Unemployment (0.0794) (0.0720) (0.0768) (0.0763) (0.0689) (0.0739) (0.0769) (0.0687) (0.0744) Per Capita GDP 5.35e-05*** 5.18e-05*** 5.30e-05*** 5.39e-05*** 5.26e-05*** 5.38e-05*** 5.59e-05*** 5.36e-05*** 5.60e-05*** (1.43e-05) (1.45e-05) (1.43e-05) (1.38e-05) (1.38e-05) (1.37e-05) (1.39e-05) (1.38e-05) (1.38e-05) % Bachelor's Degree (0.0308) (0.0300) (0.0311) (0.0296) (0.0287) (0.0299) (0.0298) (0.0286) (0.0301) Northeast ** ** ** * * * (0.453) (0.454) (0.449) (0.435) (0.434) (0.432) (0.438) (0.432) (0.435) Midwest ** * ** ** * ** ** ** (0.294) (0.356) (0.304) (0.283) (0.341) (0.292) (0.285) (0.340) (0.294) West *** *** *** *** *** *** *** *** *** (0.307) (0.306) (0.306) (0.295) (0.293) (0.295) (0.297) (0.292) (0.297) Observations R-squared Standard errors in parentheses *** p<0.01, ** p<0.05, * p<0.1 22

23 Table 7: "Labor Market Freedom" EFNA Subcomponent Score Dependent Variable = One, Five, and Ten Year Average Score House, Senate, and State Average Scores = Overall Congressional Economic Freedom (1) (2) (3) (4) (5) (6) (7) (8) (9) (10) (11) (12) (13) (14) (15) (16) (17) (18) VARIABLES One Year One Year One Year One Year One Year One Year Five Year Five Year Five Year Five Year Five Year Five Year Ten Year Ten Year Ten Year Ten Year Ten Year Ten Year Constant 6.844*** 10.68*** 6.974*** 8.452*** 6.884*** 8.926*** 6.833*** 10.48*** 6.976*** 8.199*** 6.878*** 8.676*** 6.893*** 9.951*** 7.051*** 7.186** 6.942*** 7.804*** (0.0965) (2.904) (0.0821) (2.907) (0.0851) (2.796) (0.0927) (2.782) (0.0777) (2.767) (0.0807) (2.651) (0.0935) (2.897) (0.0761) (2.802) (0.0800) (2.679) House 0.504** 0.745** 0.566** 0.768** 0.618*** 0.890*** (0.220) (0.322) (0.211) (0.308) (0.213) (0.321) Senate 0.796*** 0.937*** 0.861*** 0.960*** 0.966*** 1.151*** (0.218) (0.301) (0.206) (0.286) (0.202) (0.290) State Average 0.769*** 1.209*** 0.845*** 1.242*** 0.937*** 1.468*** (0.237) (0.357) (0.224) (0.338) (0.223) (0.342) Population (in 10,000s) * * * * ( ) ( ) ( ) ( ) ( ) ( ) ( ) ( ) ( ) Population Density e e e e-05 ( ) ( ) ( ) ( ) ( ) ( ) ( ) ( ) ( ) Median Age * (0.0474) (0.0515) (0.0493) (0.0454) (0.0490) (0.0468) (0.0472) (0.0497) (0.0473) % Male (0.0414) (0.0384) (0.0387) (0.0397) (0.0365) (0.0367) (0.0413) (0.0370) (0.0371) % White ** ** ** * * ** * ( ) ( ) ( ) ( ) ( ) ( ) ( ) ( ) ( ) Unemployment * (0.0586) (0.0541) (0.0531) (0.0562) (0.0515) (0.0504) (0.0585) (0.0521) (0.0509) Per Capita GDP -2.87e e e e e e e e e-06 (1.14e-05) (1.14e-05) (1.10e-05) (1.09e-05) (1.09e-05) (1.04e-05) (1.14e-05) (1.10e-05) (1.06e-05) % Bachelor's Degree * ** * ** ** * ** ** (0.0239) (0.0225) (0.0225) (0.0229) (0.0215) (0.0214) (0.0239) (0.0217) (0.0216) Northeast * * (0.362) (0.342) (0.330) (0.347) (0.325) (0.313) (0.361) (0.329) (0.316) Midwest ** ** ** ** ** ** ** ** * (0.236) (0.227) (0.224) (0.226) (0.216) (0.212) (0.235) (0.219) (0.215) West *** ** ** *** *** *** *** ** ** (0.253) (0.255) (0.250) (0.243) (0.243) (0.237) (0.253) (0.246) (0.240) Observations R-squared Standard errors in parentheses *** p<0.01, ** p<0.05, * p<0.1 23

24 Table 8: "Labor Market Freedom" EFNA Subcomponent Score Dependent Variable = One, Five, and Ten Year Average Score House, Senate, and State Average Scores = "Labor Market Freedom" Congressional Economic Freedom Sub-Index (1) (2) (3) (4) (5) (6) (7) (8) (9) (10) (11) (12) (13) (14) (15) (16) (17) (18) VARIABLES One Year One Year One Year One Year One Year One Year Five Year Five Year Five Year Five Year Five Year Five Year Ten Year Ten Year Ten Year Ten Year Ten Year Ten Year Constant 6.942*** 11.37*** 7.018*** 8.376*** 2.434*** *** 11.19*** 7.023*** 8.132*** 2.352*** 2.280* 7.020*** 10.78*** 7.099*** 7.369** 2.140*** (0.115) (3.051) (0.0834) (2.954) (0.376) (1.554) (0.112) (2.923) (0.0791) (2.819) (0.317) (1.308) (0.114) (3.092) (0.0788) (2.942) (0.256) (0.884) House (0.219) (0.287) (0.214) (0.275) (0.218) (0.291) Senate 0.414*** 0.422*** 0.442*** 0.430*** 0.476*** 0.479*** (0.108) (0.142) (0.103) (0.135) (0.103) (0.141) State Average 1.273*** 1.388*** 1.296*** 1.383*** 1.375*** 1.539*** (0.106) (0.113) (0.0890) (0.0952) (0.0718) (0.0643) Population (in 10,000s) e e e-05 ( ) ( ) (6.12e-05) ( ) ( ) (5.15e-05) ( ) ( ) (3.48e-05) Population Density e e e ( ) ( ) ( ) ( ) ( ) ( ) ( ) ( ) ( ) Median Age ** ** * ** (0.0479) (0.0518) (0.0235) (0.0459) (0.0495) (0.0198) (0.0486) (0.0516) (0.0134) % Male ** (0.0426) (0.0384) (0.0189) (0.0409) (0.0366) (0.0159) (0.0432) (0.0382) (0.0107) % White * *** *** *** ( ) ( ) ( ) ( ) ( ) ( ) ( ) ( ) ( ) Unemployment (0.0689) (0.0546) (0.0270) (0.0660) (0.0521) (0.0227) (0.0698) (0.0544) (0.0153) Per Capita GDP -7.32e e e e e e-06* -1.26e e e-06 (1.19e-05) (1.15e-05) (5.49e-06) (1.14e-05) (1.10e-05) (4.62e-06) (1.21e-05) (1.14e-05) (3.12e-06) % Bachelor's Degree * ** * *** (0.0248) (0.0226) (0.0110) (0.0237) (0.0215) ( ) (0.0251) (0.0225) ( ) Northeast * * * (0.374) (0.348) (0.170) (0.359) (0.332) (0.143) (0.379) (0.347) (0.0964) Midwest *** ** *** ** ** ** (0.244) (0.228) (0.119) (0.234) (0.218) (0.100) (0.248) (0.227) (0.0676) West *** *** *** *** *** ** (0.257) (0.252) (0.136) (0.246) (0.241) (0.114) (0.261) (0.251) (0.0772) Observations R-squared Standard errors in parentheses *** p<0.01, ** p<0.05, * p<0.1 24