SEPIO – CES, U. Paris 1 – 29 mars 2011, 16h, salle 17 – MSE, 106-112 bd de l’Hôpital, Paris 13e
Political ideology and economic freedom
in the US states
Christian Bjørnskov*
Aarhus School of Business, Aarhus University
Niklas Potrafke+
University of Konstanz
March 15, 2011
WORK IN PROGRESS – COMMENTS WELCOME
Abstract:
This paper empirically investigates how political ideology influences economic liberalization as
measured by the size of government, tax structure and labor market regulation. Our dataset of
economic freedom indicators compiled by the Fraser Institute covers 48 US states over the 1981-2005
period. We employ several indicators of political ideology that have been used in the literature and
introduce a new index that considers regional differences in Democratic and Republican Party ideology.
The results suggest that (1) Republican governors and Senates dominated by Republicans have been
more active in economic liberalization than Democrats, (2) coding ideology with the new index gives
rise to more clear-cut results, (3) divided government hardly counteracted ideology-induced economic
policies, even though disagreement between the governor and the Senate had some counteracting
effects.
Keywords: economic freedom, taxation, regulations, ideology, panel data
JEL Classification: O51, P16, R11, R50
* Aarhus School of Business, Aarhus University, Department of Economics, Frichshuset, Hermodsvej 22, DK-8230
Åbyhøj, Denmark. Email: [email protected].
+ University of Konstanz, Department of Economics, Box 138 D-78457 Konstanz, Germany, Email: [email protected].
1
1. Introduction
While political polarization between leftwing and rightwing parties and electoral cohesion have
declined in several OECD countries such as Germany or Japan, both still play a great role in the United
States. The influence of political ideology on economic policy-making appears to have become clearer
after the Presidential elections in November 2008. For example, President Obama worked to increase
the role of government in the economy by introducing compulsory health care insurance, tighter
regulations of the financial sector, and quasi-nationalizing parts of the auto industry. Several American
voters nevertheless disagreed with Obama’s economic policies and demonstrate against a compulsory
health care insurance and increasing public debt (REF PEW). In the midterm elections in November
2010, a majority consequently voted for the Republicans. With a Republican majority in the House, i.e.,
divided government, President Obama cannot implement his preferred policy without a substantial
number of Republican votes. Common sense therefore predicts that divided government moderates
ideological policy influences.
That Democrat governments attempt to implement more expansionary economic policies and
divided government results in counteracting effects appear to be ’conventional wisdom’. While political
economy models describe how government ideology and institutional characteristics such as divided
government are expected to influence economic policy-making, median voter models suggest that
ideological positions are unlikely to ensure majorities in elections. The conception of ideology-induced
policies also contradicts common wisdom in several OECD countries that voting for a leftwing or
rightwing party does not make a difference. The public debate often transmits the median-voter notion
that it does not matter whatever party one votes for because all politicians will implement nearly the
same policy. In Germany, for example, the leftwing Social Democrats and the rightwing Christian
Democrats have indeed implemented quite similar economic policies since 1990.
Ideological phenomena in the United States and its consequences are therefore worthwhile to
investigate in more detail since focusing on federal states overcomes a number of practical problems.
2
Several studies have shown that political ideology influences economic policy-making in the United
States. At the federal level and across the US states leftwing / Democrat governments seem to have
pursued more expansionary fiscal policies than rightwing / Republican governments by increasing
public expenditures and tax burdens (e.g., Blomberg and Hess, 2003, Reed 2006, Rose 2006, Chang et
al. 2009, Alt and Lowry 1994). The result of ideology-induced fiscal policies across the US states is not
only meaningful because states have the power to choose different policies and institutions. In
particular, different economic policy-making under leftwing and rightwing governments reflects
manifold preferences in the electorate and shows that politicians are not forced to provide policy
platforms that gratify the preferences of the median voter. Yet, the influence of government ideology
on more encompassing measures of economic policy has been ignored in the empirical political
economy literature in the US states. Against the background of sustained interest in the role of political
ideology in US economic policy, this is a surprising omission. We therefore use the index of “economic
freedom” developed by Karabegovic et al. (2003) which includes three components – the size of
government, the tax structure, and labor market freedom – to investigate whether ideology-induced
effects across the US states can also be shown for the more encompassing measures of economic
liberalization. The economic freedom indicators by Karabegovic et al. (2003) are primarily based on
fiscal policy measures and thus focus on government intervention in the public sector. In contrast to
the cross-country economic freedom indicators by the Fraser Institute (e.g., Gwartney et al. 1996 and
2009) only the labor market component relates to regulation policies.
A challenging issue is how to measure political ideology. When measuring parties’ and politicians’
ideological position, one faces three main problems: 1) the comparability of scales across countries; 2);
the potential multidimensionality of political positions and 3) the stability of scales across time and
space. By restricting our attention to the United States, we partially circumvent the first problem on the
comparability of scales across countries. The second problem on the potential multidimensionality of
political positions is also of less concern, as suggested in the pioneering work by Poole and Rosenthal
3
(1991, 2001, 2007). While ideology in some countries is, in fact, a multidimensional concept, Poole and
Rosenthal show that the vast majority of decisions taken in Congress can be placed on a left-to-right
scale. In this paper, we therefore explicitly deal with the third problem, the stability of scales across
time and space. This is pertinent since the positions of the two American parties have not been stable,
but have grown apart in recent decades. Likewise, political ideologies in the Democratic and Republican
parties are not homogenous across the US states. For example, Southern Democrats are more
conservative than Democrats on the East Coast and have also historically differed from the rest of the
party (Poole and Rosenthal, 2007). We therefore use the DW-Nominate data by Poole and Rosenthal
to approximate these differences. Specifically, we use state and party-specific positions of members in
Congress to scale the ideological position of parties in politicians’ home states. The DW-Nominate data
allow us to measure ideological positions with more precision than previous studies.
Veto players can counteract ideology-induced economic policy-making. In the United States,
divided government plays an intriguing role. Divided government occurs when the governor is
ideologically distinct from the majority of either chamber (House and Senate) of Congress. To be sure,
scholars have investigated the influence of divided government on economic policy-making at the
federal level and in the US states. Several studies have, however, ignored to distinguish between the
three types of divided government: 1) situations in which the governor is from party A, but both
chambers are dominated by party B, i.e. a situation with divided government but a unified congress; 2)
situations in which the governor and the majority in the House belong to the same party, but face a
Senate majority of party B (approval division); and 3) situations in which the governor and the majority
in the Senate belong to the same party, but face a House majority of party B (proposal division). In this
paper, we investigate whether these three types of divided government have counteracted ideologyinduced economic-policy making.
We use the new ideology measure that considers regional differences of the political positions of
the Democratic and Republican Party, respectively, to empirically investigate how political ideology
4
influenced economic liberalization in the US states. The dependent variables are economic freedom
indicators compiled by the Fraser Institute covering the 50 US states over the 1981-2005 period. We
perform all analyses using both the new index and the standard Democrat/Republican dummy
indicator of party identity that have been used in the literature and compare them with the new
ideology indices (ideology of the governor, House and Senate). We also separate divided government
into its three main types and examine which types block ideological influences.
The results suggest that (1) Republican governors and Senates dominated by Republicans have
been more active in economic liberalization than Democrats, (2) coding ideology with the new index
gives rise to more clear-cut results, (3) divided government hardly counteracted ideology-induced
economic policies, even though disagreement between the governor and the Senate had some
counteracting effects.
The rest of the paper is organized as follows: Section 2 discusses the related literature on
ideology-induced economic policy-making across the US states and formulates the hypotheses to be
tested. Section 3 presents our indices on political ideology in the US states. Section 4 presents the data
on economic freedom and specifies the empirical model. Section 5 reports and discusses the estimation
results, and investigates their robustness. Section 6 concludes.
2. Ideology-induced policies across the US states
2.1 Ideology-induced policies
Investigating the influence of government ideology on economic policy-making is one of the
prominent topics in public economics and political economy. Political economic theories describe why
and how political ideologies are expected to be associated with economic policies. The partisan
approach focuses on the role of party ideology and shows the extent to which leftwing and rightwing
politicians will provide policies that reflect the preferences of their partisans. According to the Michigan
School, leftist parties appeal more to the labor base and promote expansionary policies, whereas
5
rightwing parties appeal more to capital owners, and are therefore more concerned with reducing
inflation (cf., Converse 1964). The partisan approach also often assumes that the economy can be
described by a (short-run) Phillips-Curve-tradeoff and that politicians are able to exploit the tradeoff
strategically by fiscal and monetary policies. With respect to short-term economic performance, the
partisan models provide clear-cut predictions: leftist parties seek (or will accept) higher rates of inflation
to get lower unemployment and faster growth, rightwing parties seek (or will tolerate) higher
unemployment and slower growth to obtain lower inflation. This basic pattern holds for the classical
partisan approach (Hibbs 1977) and also for the rational approach (Alesina 1987).
Partisan theory implies that leftwing and rightwing governments have different preferences as to
the size and scope of government, the proper means to achieve shared goals and, thus, with respect to
economic policy: leftwing governments favor more government intervention, more income
redistribution and the use of expansionary fiscal and monetary policies. In contrast, rightwing
governments traditionally believe in the free market and favor less government intervention.1
Scholars have extensively examined to what extent and in which policy areas government
ideology has influenced economic policy (e.g., Alesina et al. 1997, Tavares 2004, Bjørnskov 2008). The
empirical results show that government ideology influenced privatization and deregulation policies in
industrialized countries as predicted by partisan theories. Rightwing governments have typically been
more active in privatizing and deregulating product markets (see, for example, Bortolotti et al. 2004,
Potrafke 2010). In contrast to privatization and deregulation policies, government ideology hardly
influenced fiscal policies in OECD countries after 1990. Before the 1990s, leftwing governments in
OECD countries tended to spend more on welfare expenditures. On the one hand, rightwing
governments also increased public spending and public debt. A prime example is Germany where the
conservative chancellor Helmut Kohl did not continue his fiscal consolidation from the 1980s but
1 However,
another reason for manipulating economic policies is electoral motives. In this paper, we nevertheless focus on
the influence of political ideology and do not investigate electoral cycles. Interested readers may find recent evidence of
electorally motivated policies in, e.g., Shi and Svensson (2006).
6
dramatically increased spending after the German Unification in 1990. On the other hand, leftwing
politicians such as Tony Blair in the United Kingdom or Gerhard Schröder in Germany also
implemented quite market-oriented fiscal and social policies in the end of the 1990s. Empirical studies
investigating ideology-induced fiscal policies in OECD countries thus show that electoral cohesion
declined.2
In the United States, political ideology has played an important role in fiscal policy at the federal
level (cf. Blomberg and Hess, 2003; Haynes and Stone, 1990; Alesina and Sachs, 1988; Krause and
Bowman, 2005). Confirming traditional partisan theory, many studies at the state level also find that
leftwing politicians pursued more expansionary fiscal policy than rightwing politicians, although results
in the literature are not consistent.3 For example, the results by Chang et al. (2009) suggest that the
growth rate of government spending was higher under Democratic governors over the 1951-2004
period. Besley and Case (1995a) likewise find that taxes and government spending was higher under
Democratic governors over the 1950-1986 period even if the incumbent Democrat was ineligible for
reelection because of term limits. Alt et al. (2002) also find that Democratic governors collected higher
general revenues and spend more per capita over the 1986-1995 period. Two studies report no
evidence of partisan effects. Rose (2006) suggests that the party composition of the state governments
did not significantly influence per capita general expenditures over the 1974-1999 period. Primo (2006)
also does not find ideology-induced government spending over the 1969-2000 period.
Scholars have also examined the influence of legislature ideology on fiscal policies in the US
states. Reed (2006) finds that tax burdens 1960-2000 were higher when Democrats controlled the state
legislature compared to when Republicans were in control but that the political party of the governor had
little effect. In a similar vein, the results by Besley and Case (2003) show that when Democrats
controlled the House, states had higher taxes and expenditures over the 1960-1993 period. On the
2
Government ideology does not only concern economic policy issues, but also policy issues such as immigration (Llavador
and Solano-García 2011) or international relations (Potrafke 2009).
3 Scholars have examined how government ideology influenced economic policy-making across counties in other federal
states such as Canada (e.g., the companion paper by Bjørnskov and Potrafke 2011).
7
other hand, Besley and Case (2003) find that Democrat governors pursued different labor market
policies than their Republican counterparts.
Several factors constrain the influence of ideologically motivated politicians and parties.
Institutional features such as the influence of interest groups, checks and balances and divided
government are likely to counteract ideology-induced effects in economic policy. For this reason,
partisan politicians will probably implement their preferred policies incrementally, step by step over the
legislative period. It is not likely that a newly elected government can pursue its most preferred policies
from the beginning of the legislative period. This suggests investigating the influence of government
ideology and electoral motives on the changes in economic policy. In addition, specific institutions may
still limit the room for ideology-induced policy making.
In federal states such as the United States, both chambers of parliament decide on economic
policy. When political majorities in the two chambers differ, governments are not always able to
implement their preferred policies. The institutional feature most commonly explored in studies of US
policy-making is that of divided government: when the governor is ideologically distinct from the
majority of either chamber (House and Senate) of Congress (cf. Krehbiel, 1996). Theoretically, by
balancing the influence of different ideologies, divided government may thus lead to policy
convergence (Alesina and Rosenthal 1996).4 Yet, even this feature varies considerably across the US
states and over time. Divided government has been comparatively rare in South Dakota and Maryland
in recent decades while New York and Delaware have had divided governments in the entire period we
consider in this paper.
Political scientists and political economists have been divided on the merits, drawbacks and
consequences of divided government. Most of the literature has focused on divided government as a
source of gridlock, i.e. on an “inability to change policy” (Saeki 2009). Within this view, divided
government would imply a situation in which politicians are unable to solve salient problems because
4
The model by Alesina and Rosenthal (2000), however, predicts that ´checks and balances` induce more divergence in
platforms, especially when uncertainty is high and the legislature is more powerful than the executive.
8
solutions proposed by the president are likely to be blocked by either the House or the Senate. With the
exception of Nebraska – the only state with a unicameral political system – this situation could also
arise in the state legislatures in which Congress can block gubernatorial initiatives. In this view, even in
the face of severely adverse shocks, politicians would be unable to take effective action.
Yet, another view holds that legislative gridlock caused by divided government contributes to
policy stability. With divided government, legislatures can only decide on whatever compromise both
parties can agree on and what is, in some either economic or political sense, objectively necessary. The
compromise and identification of necessary steps in turn depend on leader characteristics and other
institutional features and lead to both a decrease in policy production as well as a decrease in
subsequent policy repeal (Chiou and Rothenberg 2003; Michael, 2010). Parties anticipating divided
government can also have strong incentives to adopt more polarized policy positions (Alesina and
Rosenthal 2000). In addition, divided government implies that political oversight is improved as
partisans gain an incentive to balance the ideological bias of their counterpart (Fox and van Weelden,
2010).
Most of the studies on divided government have predominantly focused on the federal level (e.g.,
Calcagno and Lopez 2011). Similar gridlocks are likely to occur across the 49 two-chamber states, and
to have implications for the influence of political ideology on policy making. In particular, when
exploring policy changes, situations with divided government would seem to exclude any partisan
influences while unified governments would be more able to shift government spending, tax policy and
institutional characteristics in ideological directions. Of the comparatively few studies to explore this
situation, Alt and Lowry (1994) find clear evidence that states with divided government respond
differently to economic shocks than those with unified government (see also Lowry et al., 1998; Alt et
al., 2002).
The focus on the state level nonetheless allows us to go conceptually further than most of the
divided government literature and distinguish between types of divided government. Specifically, three
9
types of divided government exist: 1) situations in which the governor is from party A, but both
chambers are dominated by party B, i.e. a situation with divided government but a unified congress; 2)
situations in which the governor and the majority in the House belong to the same party, but face a
Senate majority of party B (which we term approval division); and 3) situations in which the governor
and the majority in the Senate belong to the same party, but face a House majority of party B (proposal
division). Situations 2) and 3) are both cases with a split legislature while case 1) is a split-branch
situation (cf. Alt and Lowry, 1994).
We observe overall divided government, proposal division and approval division in our data. In
2005, for example, the governors Janet Napolitano (Democrat (D), Arizona) and Mike Huckabee
(Republican (R), Arkansas) faced overall divided government: their party did not have the majority in
either parliamentary chamber. Other governors, such as Tim Pawlenty (R, Minnesota) and Phil
Bredesen (D, Tennessee), only faced approval division while George Pataki (R, New York) and Brian
Schweitzer (D, Montana) faced proposal division. Since 1981, proposal division has been common in
Delaware and North Dakota while approval division has been common in Indiana and New York.
Overall divided government in which the governor is from party A, but one or both chambers are
dominated by party B is extensively analyzed in the literature (e.g. Baron and Ferejohn, 1989; Krehbiel,
2000). Most studies have ignored whether situations with only proposal or approval division could
differ from a situation of split-branch government. Yet, at least two reasons exist to consider the
consequences of these situations to be different.
A first reason is that the main role of state Houses is to pass legislation. Any proposals for state
legislation and policy therefore need to be put forward and approved in the House, which would lead
to a situation in which the ideological influence of the House is substantial. Senates have to approve of
legislation and policies, thereby providing them with the potential power of blocking proposals. Given
that proposing legislation is influenced more by ideological considerations than blocking legislation we
would expect to find situations with proposal division more likely to alleviate ideological influences
10
than situations with approval division. An example is to publish a proposal only because of signaling
reasons even if it is known with certainty to not pass the House. A second reason is that senators are
elected for longer periods than representatives, and thereby are likely to adopt a somewhat longer time
horizon in policy-making. As such, given that much ideological policy-making aims at securing shortrun goals, senators would be likely to behave in a less ideological way. This second reason thus also
provides a rationale for expecting that proposal division is more likely to alleviate ideological policymaking than approval division. A possible exception would be situations in which ideological proposals
may have long-term consequences, in which case one would expect them to be more likely to be
blocked by an ideologically opposed Senate.
2.2 Institutional background of economic policy-making in the US states
A set of other institutional features could also block or change ideological influences. The overall
institutional frameworks adopted by the 50 US states have several similar characteristics. All, with the
exception of Nebraska, have bicameral political systems, all have elected governors, and all except
Louisiana adhere to a British common law judicial system. However, due to the substantive powers
delegated to the states, many potentially important institutional details vary across the United States.
The institutional limitations and powers delegated to the states are outlined in the Tenth
Amendment to the US Constitution and further regulated by the Commerce Clause. The Tenth
Amendment states that “The powers not delegated to the United States by the Constitution, nor
prohibited by it to the States, are reserved to the States respectively, or to the people”, implying that
institutional and policy choices not explicitly delegated to the US Congress are within the power of the
state legislatures. These powers are further delineated by what is usually known as the Commerce
Clause, which was originally supposed to ensure that the federal government only regulates matters
affecting interstate commerce. However, the interpretation of what defines interstate commerce
widened markedly during the 20th century (see, for example, Wiseman and Ellig 2007). With the change
11
in the legal interpretation adopted by the US Supreme Court, the influence of the federal government
also increased.
While the last voting restrictions were eliminated in the early 1960s, state regulations of campaign
finance laws, districting, and voter registration requirements and procedures vary. Certain states require
that voters register more than a month prior to elections (e.g., Georgia), while Maine and North Dakota
allow voters to register on Election Day, and Wyoming does not recognize that voters even need to
register. Some states have also introduced laws, such as California’s well-known ‘Proposition 13’,
intended to make it more difficult to increase government spending. 15 states had by 2005 imposed
supermajority requirements on new taxation in order to curb the increasing state budgets and takings.5
Most states also have some form of requirement that the legislature must submit a balanced
budget or similar balanced budget requirements, as those introduced in Massachusetts in 1998 and
California in 2004. By 2005, only eight states – Hawaii, Indiana, Missouri, New Hampshire,
Pennsylvania, Vermont, Virginia, and Washington – had not imposed clear requirements on either the
governor or the state legislature (NASBO, 2007).6 Likewise, governors in the majority of states have the
power to veto specific line items and appropriations, thus further putting a formal cap on expenditures
and tax increases.7
Against this background, a significant question is how states have expanded their economic reach
so considerably. While 17 states did not even raise income taxes in 1950 (Besley and Case 2003), three
states raised more than 10% of Gross State Product in taxes by 2005. Similarly to the variation in
budgetary institutions, labor market regulations vary widely across the 50 states. Minimum wages and
the attendant legislation vary in general as well as in detail. While minimum wages are regulated through
5 These states are Arizona, Arkansas, California, Colorado, Delaware, Florida, Kentucky, Louisiana, Mississippi, Missouri,
Nevada, Oklahoma, Oregon, South Dakota, and Washington (Knight, 2000). On restrictions on spending and taxation see
also Poterba (1994, 1995), Holtz-Eakin (1988), Hagen (1991).
6 In principle, Idaho only has weak formal requirements. However, NASBO (2007: 34) notes that for the governor to submit
an unbalanced budget would be “political suicide”. Other states have some form of statutory requirements with uncertain
consequences.
7 The consequences of these institutional differences are highly disputed, and it is questionable whether such formal
requirements have actual effects (cf. Poterba 1996, Knight 2000; Primo 2006). See, for example, Besley and Case (2003) and,
Lowry (2008) for surveys of fiscal policies and institutions in the US states.
12
political agreement spanning several years in Illinois, Montana introduced an automatic annual cost-ofliving adjustment in 2007 (Fitzpatrick and Perine, 2008).8 These labor market regulations are also
ideologically contested with cross-country studies supporting that labor market policies are affected by
government ideology (Botero et al. 2004); the same appears to be the case across the Canadian
provinces (Bjørnskov and Potrafke 2011).
A final question is which level of government is most relevant for partisan policy-making. The
literature has primarily explored the effects of the ideological position of the governor, yet ideological
positions in the House and the Senate are also likely to influence partisan policy-making (Alt and
Lowry, 1994, Reed, 2006). The role of the government as well as the lower chamber is to initiate
legislation. The role of the upper chamber (Senate) is to approve or disapprove of proposals from
government and the House. As such, even though the Senate may not have to directly turn down a
piece of legislation to exert its influence. The Senate could affect policy making indirectly if some
legislation is proposed in the case that the governor or representatives deem its chances to pass Senate
to be too low. The actual ideological influence of the three levels of government is thus an outcome of
a complex game of political interaction.
2.3 Hypotheses
Following the implications of the related literature on partisan politics in the US states and
government ideology and economic liberalization, we form a set of directly testable hypotheses. The
hypotheses to be investigated in the following are:
1. Republican governors have been more active in economic liberalization than Democratic governors.
8
On the influence of the US president in the legislative process see, for example, McCarty and Poole (1995). The empirical
results by McCarty and Poole (1995) suggest that the US president has less influence in the legislative process than was
intended by theoretical predictions.
13
2. Republican dominated legislatures have been more active in economic liberalization than Democratic dominated
legislatures.
3. Divided governments counteract ideology-induced policies.
4. Counteraction varies across types of divided government.
To test these hypotheses, we first need a measure of government and parliamentary ideology that
is consistent and comparable across time and states.
3. Data
3.1 Measuring political ideology across the US states
Several pitfalls are associated with measuring political ideology, as outlined in Castles and Mair’s
(1984) pioneering paper. Measuring political ideology consistently across time and space involves
assessing the dimensionality of ideology, choosing a scale of ideology common to all units of
observation, and in most cases making the implicit or explicit assumption that ideology is scaleinvariant across time. Scholars have employed two measures for political ideology in the US states:
governor ideology (Republican / Democrat) and the ideological position of the legislature (e.g., Reed
2006) – that is the political ideology of the House and the Senate. Most studies exploring evidence
across the 50 states treat scale issues as resolved by assuming that the positions of the Democrat and
Republican parties do not change over time, or change so consistently across the states that all changes
are picked up by a joint time trend. Most studies assume that there are no material differences between
party positions across the states. For example, scholars assume that Massachusetts Democrats and
Illinois Democrats share the exact same ideology, and that Arizona Republicans and Maryland
Republicans do so too. Alt et al. (2002) is a notable exception.
Party positions are likely to differ across the US states. We follow Berry et al. (1998) in employing
political positions in the US Congress to estimate state party positions. We deviate from Berry et al.
14
(1998) in employing Poole and Rosenthal’s (2006) DW-NOMINATE data on political positions in the
US Congress to relax the standard assumption that members of specific parties hold the same
ideological positions across all US states. We furthermore distinguish between the political ideology of
governors and the two chambers of parliaments.
Poole and Rosenthal (2006) deal with the dimensionality question – whether the set of political
decisions taken by Congress is distributed along a single or multiple ideological dimensions – by
exploring roll-call votes in the US Congress. Their study shows that Gerring’s (1997: 975) definitional
assessment that a set of values “becomes ideological only insofar as it specifies a concrete program, a
set of issue-positions” holds for the two American parties. The vast majority of votes can be placed on
a uni-dimensional left-to-right scale, which they define as between -1 and +1. In the 1981-2005 period,
which we analyze in this paper, an assumption of uni-dimensionality is thus warranted.9 However,
Poole and Rosenthal’s data also exhibit two specific features: 1) that party positions have not remained
stable, but have moved apart; and 2) that representatives of the same party from different states can
place themselves in quite different positions on a left-right scale of political ideology.
Our identifying assumption is that the political positions taken by members of Congress from
different states reflect the ideological positions of their party competing for state government. This
relaxes the standard assumption that ideological party positions do not vary across states or over time.
We place the Democratic Party in every individual state at the average ideological position of Democrat
representatives from this state, and do the same for the Republican Party, and use these positions to
calculate our ideology index. We acknowledge that our identifying assumption implies that the
ideological positions of federal representatives in Congress reflect state-specific voter preferences and
state party positions to a sufficient degree. Congressmen (and state parties) need to gratify the
preferences of the voters in the individual states to become elected. For example, voters in the Midwest
9 Poole
and Rosenthal document that two dimensions of American politics arose in specific periods. In particular in the late
1950s and early 1960s, the civil rights question formed a separate dimension distinguishing Southern Democrats from the
rest of the country.
15
are more conservative than voters on the East Coast. Election-motivated as well as ideology-induced
politicians will consider the state-specific differences of the electorates. We therefore believe that our
index is more precise than mere party identity measures because it considers varying ideological party
positions across states or over time.
As suggested by Figures 1 and 2, the choice of how to measure ideology is not innocuous. Figure
1 shows the overall shift of party ideological positions by plotting the average Democrat and
Republican positions between 1981 and 2005: average policy positions between Republicans and
Democrats became more polarized over time. The index illustrated in Figure 1 allows the Democrat
and Republican parties at the state level to take up more than a singular point on the left-to-right line.
For example, averaging the positions of members of Congress of the two parties across the 106th to
110th Congress, the state averages of Democrat positions vary between -.553 (Massachusetts) and -.154
(Oklahoma). State averages of Republican positions vary between .209 (Connecticut) and .729
(Colorado). The positions furthermore vary over time, and show that parties’ ideologies may change.
For example, Democrats became more leftwing over time in states such as Arizona (moving from a
DW-score in 1981 of 0.019 to score in 2005 of -0.528), New Mexico (DW-score in 1981: 0.268; DWscore in 2005: -0.506) and Virginia (DW-score in 1981: 0.090; DW-score in 2005: -0.352). In other
states such as Hawaii, Idaho and Wyoming, the DW-score for Democrats did not change. Republicans
became more rightwing over time in states such as Arizona (DW-score in 1981: 0.323; DW-score in
2005: 0.705), Indiana (DW-score in 1981: 0.252; DW-score in 2005: 0.660) and Tennessee (DW-score
in 1981: 0.220; DW-score in 2005: 0.620). The DW-score for Republicans did not change in states such
as Hawaii, Idaho and Wyoming. In Oregon, Republicans even became more leftwing over time (DWscore in 1981: 0.665; DW-score in 2005: 0.405). As such, our index of ideology across the US states is
substantially more flexible than those used in most existing studies.
To illustrate the structure of our index, Figure 2 shows the average ideology of the state
parliaments across the 1981-2005 period, measured as either the average ideology of seats in state
16
Houses and Senates or state governors, and employing either a standard -1/ +1 coding of party identity
or the coding of party ideology based on Poole and Rosenthal’s (2006) DW-NOMINATE placements.
Figure 2 also illustrates the discrepancy between the ideology of the incumbent governments and the
average positions of the parliaments. We follow Bjørnskov (2008) in calculating the average position of
the state parliaments as the average party identity / ideology, weighted with the shares of seats occupied
by either party. We place independent members at 0, unless we know them to be ideologically extreme,
in which cases we place them at either -1 or 1.10
The choice of calibrating state-specific party positions makes a substantial difference, and the
span of party ideological positions is quite wide. The most leftwing Democratic Party position in our
dataset is that of Massachusetts in recent years (at an index of -.55) while the Republican Party in
Massachusetts in 1990-91 also takes the most leftwing position in our dataset (index -.09). Conversely,
our procedure places the Republican Party in Colorado in 2005 as the most rightwing (index .72) while
the Democratic Party in New Mexico in 1981 takes the most rightwing position for that party (index
.27).
The data on political ideology cover 48 states. The two exceptions for which there are no data are
Alaska and Nebraska. Alaska is excluded because it has not sent a Democratic representative to
Washington since 1981, which implies that we cannot calculate an ideology score for Alaska
Democrats. Nebraska is excluded because of its special status in the United States by being the only
state with a unicameral political system, which is moreover broadly non-partisan. Following the related
literature we cannot categorize Nebraskan politics in partisan terms.
3.2 Data on economic freedom in the US states
10
We for example place all libertarian members of any state legislature at the rightwing extreme position of 1. We likewise
place members of the Green Party at the leftwing extreme of -1.
17
We use the dataset on economic freedom in the US states first introduced by Karabegovic et al.
(2003). This dataset is available for the 1981-2005 period and contains yearly data for the 50 US states
(balanced panel).
The index of economic freedom includes three components: 1) the size of government,
composed of general government consumption expenditure (% of GDP), total provincial transfers and
subsidies (% of GDP), and social security payments (% of GDP); 2) the tax structure, measured as an
index equally weighting tax revenue (% of GDP), the top marginal tax rate and its applied income
threshold, indirect tax revenue (% of GDP) and sales taxes (% of GDP); and 3) labor market freedom,
measured as the extent of minimum wage legislation, government employment (% of total provincial
employment) and union density. Each of the subcomponents enters with equal weights in the three
components of the index. The construction of these indices follows a political logic as they are pooled
into measures of expenditure policy, revenue policy, and labor market policy. The overall indicators are
scaled to take on values between 0 (minimum of economic freedom) and 10 (maximum of economic
freedom).11
Figure 3 illustrates that average economic freedom in the analyzed sectors was quite
pronounced in the 1980s (maximum 7.18 in 1989), but remarkably decreased in the beginning of the
1990s. Since 1993, economic liberalization has proceeded steadily for some years, but declined again in
2000. Over the 1993-2005 period, changes have been made in most of the areas covered by the
indicators, but most spectacularly in the size of government, and to a lesser extent, in labor market
freedom. Figure 5 illustrates that overall economic freedom was most pronounced in states such as
Florida, Tennessee and Texas (overall indices on average 8.01, 8.27, 8.06) and less pronounced in states
such as Maine, New York, Rhode Island and West Virginia (overall indices on average 5.88, 5.51, 5.80,
5.48). In Texas, however, overall economic freedom steadily declined in the 1980s from 9.0 in 1980 to
8.0 in 1988. In New York, economic liberalization was proceeding rapidly till the end of the 1990s. In
11 For further details on the construction of the economic freedom indicators, as well as the primary data, see Karabegovic
et al. (2003).
18
Mississippi, the overall index dropped from 7.8 in 1981 to 6.6 in 2005, an effect mostly driven by state
intervention in the government sector (Figure 6). In particular, Figure 6 also shows a dramatically
decreasing size of government index (i.e. rapid growth of the government sector) in Alaska (8.9 in 1981
and 3.8 in 2005). In New Mexico, the size of government indicators decreased from 8.6 in 1981 to 5.9
in 2005. The tax structure has been liberalized in states such as Delaware (Figure 7). US labor markets
became more flexible over time (Figure 8). The labor market freedom index increased on average from
6.22 in 1981 to 6.78 in 2005 with labor market deregulation proceeding particularly quickly in states
such as Michigan, Minnesota, West Virginia and Wisconsin. The figures show that economic freedom
has varied over time and across the US states. We will therefore investigate the influence of political
ideology on economic freedom in an error-correction model that considers variation over time and
across states.
3.3 Correlation between economic freedom and political ideology
In order to illustrate the association between political ideology and economic freedom across
the US states, we present correlations between the averaged economic freedom indices and averaged
ideology of the governors. One can see with the naked eye that considering ideological differences
across the US states results in a positive association between economic freedom and Republican
governors (Figure 8), whereas economic freedom hardly appears to be associated with the common
Democrat/Republican dummy variables (Figure 9). Measuring ideology appears to play an important
role.
4. Empirical model
The base-line error correction panel data model that investigates hypothesis 1 and 2 has the following
form:
19
∆ Economic Freedom Indexijt = α1 Ideologyikt-1 + α2 ∆ Ideologyikt +Σl δl1 Xilt-1 + Σl δl2 ∆ Xilt + ηi + εt +
uijt
with i=1,…,48; j = 1,…, 4; k=1,…,9; l=1,…,11; t=1,...,24
(1)
where the dependent variable ∆ Economic Freedom Indexijt denotes the first difference of economic
freedom index j. We distinguish between the four indicators of “Overall economic freedom”, “Size of
government”, “Takings and discriminatory taxation” and “Labor market freedom”. Ideologyikt
describes the ideological orientation of either the governor, House or Senate as discussed in the
previous section. In order to test whether the coding of the ideology variables (common measures
versus the DW measure) has an influence of economic freedom we distinguish between six different
ideology variables: the common Democrat/Republican coding for the governors, majority in the House
and Senate and our new coding for the governors, majority in the House and Senate, respectively. In
either way, we include (1) the level of the ideology variable in period t-1 which is the long-term effect
of ideology on economic freedom and (2) the first difference of the ideology variable which is the
short-term effect of ideology on economic freedom. Σl Xilt contains eleven exogenous economic and
institutional control variables that we also include in levels in period t- 1 (long-term effect) and in first
differences (short-term effect). We include the dependency ratio (persons aged below 15 and above 65
as a share of total population), women as a share of total population, blacks as well as Hispanics as a
share of total population, total population, employment, the consumer price index (CPI), the number
of dollars received back for every dollar spent from the fiscal equalization system, a supermajority and a
balanced budget dummy variable. We expect economic freedom to decrease the higher is the number
of transfer receivers and the tighter are institutional restrictions. The appendix provides descriptive
statistics of all variables included. Lastly, ηi represents a fixed state effect, εt is a fixed period effect and
uijt describes an error term.
20
The base-line error correction panel data model that investigates hypothesis 3 and 4 has the
following form:
∆ Economic Freedom Indexijt = α1 Ideologyikt-1 + α2 ∆ Ideologyikt
+ β1 Divided Governmentimt-1 + β2 ∆ Divided Governmentimt
+ γ1 Ideologyikt-1* Divided Governmentimt-1
+ γ2 ∆ Ideologyikt-1* ∆ Divided Governmentimt
+Σl δl1 Xilt-1 + Σl δl2 ∆ Xilt ηi + εt + uijt
with i=1,…,48; j = 1,…, 4; k=1,…,6; l=1,…,8; m=1,…,3 t=1,...,24
(2)
The model described in equation (2) is similar to the model in equation (1). We now also include
the Divided Governmentimt variable which is a dummy variable that takes on the value one when
government in a state was divided and zero otherwise. We distinguish between three different kinds of
divided government (section 2): 1) the governor is from party A, but both chambers are dominated by
party B, 2) ‘proposal division’ and 3) ‘approval division’. “Ideologyikt* Divided Governmentimt” is an
interaction term between the individual ideology variable and the divided government dummy variable.
We expect that ideology-induced economic policies are counteracted under divided government and
thus, the coefficients of the interaction terms to have a negative sign. We again include the level of the
divided government variable in period t-1 and its interaction with the respective ideology variables in
levels in period t-1 (long-term effect) and the first difference of the divided government dummy
variable and its interaction with the first difference of the respective ideology variable (short-term
effect).
We now turn to our choice of estimation procedure. We estimate the model with feasible
generalized least squares and implement heteroscedastic and autocorrelation consistent (HAC) NeweyWest type (Newey and West 1987, Stock and Watson 2008) standard errors and variance-covariance
21
estimates, since the Wooldridge test (Wooldridge 2002: 176-177) for serial correlation in the
idiosyncratic errors of a linear static panel-data model implies the existence of unrestricted serial
correlation for all the four economic freedom indicators. In the robustness tests section we also
describe results when the models are estimated by a dynamic bias corrected estimator.
5. Results
5.1 Basic Results
Table 1 shows the regression results when the common Democrat/Republican ideology
measures are used. We include the political ideology variables of the governor in columns (1) and (4),
political ideology of the House in columns (2) and (5), and political ideology of the Senate in columns
(3) and (6). In columns (1) to (3) we present results without control variables. The control variables in
columns (4) to (6) mostly display the expected signs. The coefficient of the first difference of the
dependency ratio has a negative sign and is statistically significant at the 1% level in column (4) and at
the 5% level in columns (5) and (6). A higher share of persons aged below 15 and above 65, i.e.
effectively outside the labor market, thus had a negative effect on economic freedom in the short run
while the level of the dependency ratio in period t-1 (the long-run effect) does not turn out to be
statistically significant. The level of the share of females in period t-1 has a positive sign and is
statistically significant at the 5% level in columns (5) and (6) but does not turn out to be statistically
significant in column (4). The coefficient of the first difference of the share of females – the short-run
effect – does not turn out to be statistically significant. The coefficient of the level of the share of
Hispanics in period t-1 (long-run effect) does not turn out to be statistically significant, whereas the
coefficient of the first difference of the share of Hispanics has a positive sign and is statistically
significant at the 10% level, although only in column (4). Likewise, the share of blacks, total population
and supermajority requirements do not turn out to be statistically significant across all equations.
Employment had a strong positive effect on economic freedom in the short run: the coefficient of the
22
first difference of employment is statistically significant at the 1% level in columns (4) to (6). However,
employment did not influence economic freedom in the long run, indicating that employment effects
are transitory. Higher consumer prices also had a negative influence on economic freedom in the short
run: the coefficient of the first difference of the consumer price index is statistically significant at the
1% level in columns (4) to (6). Again, the long-run effect of higher consumer prices on economic
freedom does not turn out to be statistically significant at conventional levels. Finally, fiscal equalization
had a positive influence on economic freedom in the long-run although the effect is minor: the
coefficient of the level of fiscal equalization transfers received is statistically significant at the 1% level
in columns (4) to (6) and indicates that economic freedom increased by about eight percent of a
standard deviation when fiscal equalization transfers received increased by one standard deviation (.3).
In the short run, fiscal equalization transfers received did not influence economic freedom.
The results in Table 1 show that the common ideology measures hardly influence overall
economic freedom. The coefficient of the first difference of governor ideology has the expected
positive sign and is statistically significant at the 10% level in column (1) but lacks statistical significance
in column (4). The numerical meaning of the effect in column (1) is that increasing governor ideology
by one standard deviation induces an increase of the economic freedom indicator of about 10% of the
standard deviation in the short run. The long-run effects of governor ideology also do not turn out to
be statistically significant. Average ideology in the House does not influence overall economic freedom:
the coefficients of the level in period t-1 and the first difference of the common House ideology
measure do not turn out to be statistically significant (columns 2 and 4). The coefficient of the first
difference of Senate ideology has the expected positive sign and is statistically significant at the 5% level
in column (6) but lacks statistical significance in column (3). The numerical meaning of the effect in
column (6) is that increasing Senate ideology by one standard deviation induces an increase of the
economic freedom indicator of about 10% of the standard deviation in the short run. The long-run
effects of Senate ideology also do not turn out to be statistically significant.
23
Table 2 provides the results with the alternative DW ideology measures. As expected, the
influence of the control variables does not change compared to the results in Table 1. However, in
contrast to the results presented in Table 1, Table 2 reveals more pronounced ideology-induced effects
employing the more precise DW ideology measures. The coefficient of the first difference of governor
ideology again has the expected positive sign and is statistically significant at the 5% level in column (1)
and at the 10% level in column (4). The long-run effect of governor ideology is statistically significant at
the 10% level in column (1), but lacks statistical significance in column (4). Ideological differences may
be statistically significant but so small that they are without economic or political significance (cf.
McCloskey, 1985). The numerical meaning of the effects in column (1) and column (4) is that
increasing governor ideology by one standard deviation induces an increase of the economic freedom
indicator of about 10% of the standard deviation in the short run. The effects of Republican majorities
in the House as measured by the DW variables are not robustly associated with economic freedom.
Conversely, Republican majorities in the Senate as measured by the DW variables strongly influenced
economic freedom in the short-run: the coefficient of the first difference of the ideology variable is
statistically significant at the 5% level in column (3) and at the 1% level in column (6). The numerical
meaning of the effects in column (3) and column (6) is that increasing Senate ideology by one standard
deviation induces an increase of the economic freedom indicator of about 30% of the standard
deviation in the short run. This effect is thus also numerically more pronounced than the effects with
the common measure of Senate ideology. In the long run, however, Republican majorities in the Senate
as measured by the DW variables do not have a lasting influence.
The findings when employing the DW measure thus suggest that political ideology, more
precisely measured, influences economic freedom. Table 3 therefore provides more specific
information by showing the results for the sub-indicators of economic freedom. To save space, we only
report the coefficients and t-statistics of the ideology variables that belong to regressions with the entire
set of control variables included; the main findings with the control variables do not change.
24
The choice of the ideology measures also matters for ideology-induced effects on the sub
indicators of economic freedom. The long-run effect of ideology in the House as measured by the
common variables is shown to have an unexpected negative influence on economic freedom in the
“Size of Government” sector, which is statistically significant at the 10% level. No other coefficients of
the ideology variables as measured by the common variables turn out to have a significant influence on
the “Size of Government” indicator. By contrast, Republican majorities in the House and Senate as
measured by the DW variable increased economic freedom in the “Size of Government” sector. The
effects are statistically significant at the 1% level and 10% level.
Measuring ideology also matters for ideology-induced effects on economic freedom in the
“Taxation” sector: the ideology of the Senate as measured by the common ideology variable appears to
have a negative long-run effect on economic freedom in the “Taxation” sector, although the effect is
only significant at the 10% level. Yet, the coefficient of the ideology of the Senate as measured by the
DW ideology variable (long-run effect) does not turn out to be statistically significant. In the short-run,
on the contrary, changes to the average ideology of the Senate as measured by the common and the
DW variables have a positive influence on economic freedom in the “Taxation” sector, which is
substantially larger using the DW measure.
Finally, regarding labor market freedom we find no statistically significant ideology-induced
effects when we use the common ideology measures. By contrast, Republican majorities in the House
and the Senate as measured by the DW variables have a positive influence on labor market freedom in
the long run. The ideology-induced effect of the House is statistically significant at the 10% level and
the ideology-induced effect of the Senate is statistically significant at the 1% level.
5.2 Effects with divided government
Overall, the estimates suggest that Republican governors and Senates have been more active in
promoting overall economic freedom and deregulating the size of government sector. These findings
25
occur clearly with the DW measures of government ideology, yet the findings may still reflect
substantially heterogeneous effects, depending on institutional characteristics. In particular, divided
government is expected to counteract ideology-induced policies: when a Republican governor faces a
partly or fully Democratic dominated legislature, he or she may well be prevented from implementing
market-oriented policies (and vice versa) (cf., Alt et al., 2002). We have therefore included a divided
government variable that takes on the value one with divided government and zero otherwise. We also
include an interaction term between the ideology variables and the divided government dummy
variables. The findings are summarized in Table 4.
Including a dummy variable for divided government suggests first that divided government is
associated with lower labor market freedom in the short run, but not in the long run. Only in the case
of area 3 – labor market freedom – do the results with divided government differ substantially from
those without. Average ideology in the Senate is still statistically significant, yet this effect evidently only
arises when government is not divided. With divided government, Senate ideology does not influence
labor market freedom.
Overall divided government hides two different forms of divided government that could have
different consequences. The results in Tables 5a, 5b and 5c refer to the analysis similar to Table 4, but
we now distinguish between proposal and approval division. Table 5a reports the results of exploring
the effects of governors’ ideology. In the short run, a change of governor ideology toward the political
right is significantly associated with more overall economic freedom, but only when neither type of
divided government occurs. We find no significant short-run effects on the separate indices and no
effects in the long run. However, the importance of separating the two types of division becomes clear
when exploring the effects of governor ideology on taxation. The results show that without division,
there is no association between governors’ ideological positions and changes in the tax burden.
Conversely, with proposal division, we find a clear negative effect (coefficient -.1059, std. dev. 0434)
26
while the effect turns significantly positive with approval division (coefficient -.0912, std. dev. 0316).
Governor ideology is still associated with labor market freedom when government is unified.
Tables 5b and 5c report the results when using the average ideology of the House and the
Senate. While the results in Table 5b are similar to those in Table 4, the results in Table 5c with Senate
ideology differ. In the short run, Senate ideology is significantly associated with changes in government
size, although not with proposal division, but with the largest coefficient when ideological shifts are
accompanied by approval division. The short run effects on tax burdens are not affected by either type
of divided government. On the contrary, in the long run we find that Senate ideology is strongly
associated with labor market freedom when either there is unified government or proposal division
(coefficient .1707, std. dev. 0502). With approval division instead, there is no significant association.
We have calculated marginal effects on the influence of political ideology given the different
types of divided government. Starting with the effects of governor ideology, we identify three effects
significant at the five percent level: 1) a negative effect of (Republican) ideology on taxation with
proposal division; 2) a positive effect of ideology on taxation with approval division; and 3) a positive
effect on labor market freedom without divided government. Changing governor ideology by one
standard deviation in the long run induces a decrease in situation 1) of 10% of a standard deviation and
an increase of 10% in situation 2). In other words, across a four-year electoral cycle, both will result in a
cumulative change of taxation by about half a standard deviation. Conversely, the effect of governor
ideology on labor market freedom is small, as a one standard deviation change only increases the annual
rate of change by roughly two percent of a standard deviation. Across a four-year period, this becomes
a modest ten percent change.
Turning to the effects of changes in House ideology, only short-run changes in government size
are robust. For situations with unified Congress or proposal division, a one standard deviation change
in House ideology is associated with an increase of the government size index (a decrease of
government size) by about 6-8 percent of a standard deviation; with approval division, this effect is 14
27
percent of a standard deviation. However, as noted, these are mere short-run effects and the results
provide no evidence of persistent changes.
Likewise, we find a significant short-run effect of changes in Senate ideology on government
size of approximately the same size. Conversely, the influence of Senate ideology on labor market
freedom is numerically important. With either unified Congress or proposal division, a one standard
deviation change in Senate ideology is associated with an average annual change in labor market
freedom of about eight percent of a standard deviation. Evaluated across a four-year electoral cycle, this
amounts to a cumulative change in the level of labor market freedom of approximately two standard
deviations. In other words, long-run changes of Senate positions on labor market freedom – a policy
choice with likely long-un consequences – are of clear political and economic significance.
The results in Tables 5a, b and c suggest that divided government only under some
circumstances counteracted ideology-induced policies, yet under other circumstances may have
multiplied them. Furthermore, which particular type of divided government a state faces turns out to be
important.
5.3 Robustness of the results
We checked the robustness of the results in several ways. Economic liberalization in one state is
likely to be influenced by economic liberalization in neighboring states (Case et al. 1993, Besley and
Case 1995b). We have therefore estimated a model in growth rates and included a spatially lagged
dependent variable that considers geographical neighbors (Tables 5 and 6). The spatial weight matrix is
row-normalized. The spatially lagged dependent variables are statistically significant at the 1% level for
the overall economic freedom index and the two sub indicators on the size of government and takings
and discriminatory taxation, but do not turn out to be statistically significant for labor market freedom.
Including the spatially lagged dependent variables does not change the inferences regarding the
ideology variables.
28
We have estimated the model in growth rates using a dynamic estimation procedure. We
employ a GMM estitmator (Blundell-Bond 1998) because the common fixed-effect estimator is biased
when including a lagged dependent variable. The instruments are collapsed as suggested by Roodman
(2006). Applying this procedure ensures not to use invalid and too many instruments (see Roodman
2006 and 2009 for further details). The lagged dependent variable has a positive sign and is statistically
significant at the 10% level when the overall economic freedom indicator is used but does not turn out
to be statistically significant in the other three specifications when the subindicators of economic
freedom. Both ideology variables – the common Democrat/Republican dummy variable and the PR
ideology measure – lack statistical significance at conventional levels in these dynamic panel data
models.
A caveat applying to all panel data models concerns potential endogeneity of the explanatory
variables. It is, however, not reasonable to expect that government ideology is influenced by changes in
economic liberalization in the US states. Good instrumental variables for government ideology are
simply not available, consistent with the prevalent view that voters decide based on predominantly noneconomic information (e.g., Nannestad and Paldam 1997). Moreover, instrumenting ideology with the
help of lagged government ideology would not be reasonable because ideology is highly persistent.
The reported effects could depend on idiosyncratic circumstances in the individual states. We
have therefore tested whether the results are sensitive to the inclusion/exclusion of individual states.
Results are not sensitive to the inclusion/exclusion of individual states.
6. Discussion and conclusions
Whether and to what extent political ideology influences fiscal policy-making and institutional
choices is a major question in public economics and political economy. We have used the economic
freedom indicators by the Fraser Institute to investigate how political ideology influenced economic
liberalization as measured by the size of government, tax structure and labor market regulation in the
29
US states over a 25-year period. To measure political ideology, we have at first employed the common
Democrat/Republican indicators for the governors, House and Senate that have been used in the
literature. The results show that Republican governors and legislatures dominated by Republicans
hardly promoted economic liberalization in the 1981-2005 period. Political ideologies in the
Democratic and Republican party are not homogenous across the US states. For example, Southern
Democrats are more conservative than Democrats at the East Coast. We have therefore used the DWNominate data by Poole and Rosenthal to approximate these differences. The results show that
ideology-induced effects are more pronounced with the new ideology coding than with common
Democrat/Republican dummy variables: Republican governors and, in particular, Senates dominated
by Republicans promoted economic freedom. The result that a more precise measure of political
ideology reveals stronger ideology-induced effects is significant for empirical tests of partisan politics.
Divided government is often expected to counteract ideology-induced economic policy-making.
Our findings show however that overall divided government in the US states hardly mitigated the
influence of political ideology. In contrast to the related literature on the counteracting effects of
divided governments, we also distinguished between three different types of divided government:
overall, approval and proposal division. The results show that the type of division that governments
face has an influence. While we find evidence of ideological influences on labor market regulations
when state governments were unified, we find no such evidence for legislatures with either type divided
government when looking at governors’ ideology and no evidence with approval division when looking
at Senates’ average ideology. With tax policy, we find effects of governor ideology in opposite
directions, depending on whether the state has proposal division (a negative effect) or approval division
(a positive effect). This specific characteristic of the United States – a federal state with two-chamber
systems in all but one state and considerable legislate decentralized power – therefore yields results that
are both similar to cross-country results and provide more interpretative nuance. The distinction
30
between the three types of divided government may explain the mixed results from previous studies
exploring ideological influences across the states.
Acknowledgements
We are grateful for comments from Felix Bierbrauer, Martin Hellwig, David Dreyer Lassen, Leandro de
Magalhaes, Christian Traxler and participants of the 2010 meetings of the Public Choice Society in
Monterey, the 2010 meetings of the European Public Choice Society in Izmir, the 2010 Silvaplana
workshop in Pontresina, the Research Seminar at the University of Lucca in November 2010 and the
Research Seminar at the Max Planck Institute for Research on Collective Goods in Bonn in December
2010 . All remaining errors are of course ours.
31
References
Alesina, Alberto (1987). Macroeconomic policy in a two-party system as a repeated game.
Quarterly Journal of Economics 102, 651-678.
Alesina, Alberto, Nouriel Roubini and Gerald D. Cohen (1997). Political Cycles and the Macroeconomy.
(Cambridge, MA: MIT Press).
Alesina, Alberto and Howard Rosenthal (1996). A theory of divided government.
Econometrica 64(6), 1311-1343
Alesina, Alberto and Howard Rosenthal (2000). Polarized platforms and moderate policies with checks
and balances. Journal of Public Economics 75, 1-20.
Alesina, Alberto and Jeffrey D. Sachs (1988). Political parties and the business cycle in the United
States, 1948-1984. Journal of Money, Credit and Banking 20(1), 63-82.
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.
Alt, James E., David D. Lassen and David Skilling (2002). Fiscal transparency, gubernational approval,
and the scale of government: evidence from the states. State Politics and Policy Quarterly 2, 230250.
Baron, David P. and John A. Ferejohn (1989). Bargaining in Legislatures.
American Political Science Review 83, 1181-1206.
Berry, William D., Evan J. Ringquist, Richard C. Fording and Russell L. Hanson (1998). Measuring
citizen and government ideology in the American States, 1960-93.
American Journal of Political Science 42(1), 327-348.
Besley, Timothy and Anne Case (1995a). Does electoral accountability affect economic policy choices?
Evidence from gubernational term limits. Quarterly Journal of Economics 110(August), 769-798.
Besley, Timothy and Anne Case (1995b). Incumbent behavior: vote-seeking, tax-setting, and yardstick
competition. American Economic Review 85(1), 25-45.
Besley, Timothy and Anne Case (2003). Political institutions and policy choice: evidence from the
United States. Journal of Economic Literature 41(4), 7-73.
Bjørnskov, Christian (2008). The growth-inequality association: government ideology matters.
Journal of Development Economics 87(2), 300-308.
Bjørnskov, Christian and Niklas Potrafke (2011). Political ideology and economic freedom across
Canadian provinces. Eastern Economic Journal, forthcoming.
Blomberg, S. Brock and Gregory D. Hess (2003). Is the political business cycle for real?
32
Journal of Public Economics 87(5-6), 1091-1121.
Blundell, Richard W. and Stephen Bond (1998). Initial conditions and moment restrictions in dynamic
panel data models. Journal of Econometrics 87(1), 115–143.
Bortolotti, Bernado, Marcella Fantini and Domenico Siniscalco (2004). Privatisation around the world:
evidence from panel data. Journal of Public Economics 88(1-2), 305-332.
Botero, Juan, Simeon Djankov, Rafael la Porta, Florencio C. Lopez-de-Silanes and Andrei Shleifer
(2004). The regulation of labor. Quarterly Journal of Economics 119(4), 1339-1382.
Calcagno, Peter T. and Edward J. Lopez (2011). Divided we vote. Public Choice, forthcoming.
Case, Anne, James Jr. Hines and Harvey S. Rosen (1993). Budget spillovers and fiscal policy
interdependence: evidence from the States Journal of Public Economics 52(3), 285-307.
Castles, Francis G. and Peter Mair (1984). Left-right political scales: some ‘expert’ judgement.
European Journal of Political Research 12, 73-88.
Chang, Chu-Ping, Yoonbai Kim and Yung-hsiang Ying (2009). Economics and politics in the United
States: a state level investigation. Journal of Economic Policy Reform 12(4), 343-354.
Chiou, Fang-Yi and Lawrence S. Rothenberg (2003). When Pivotal Politics Meets Partisan Politics.
American Journal of Political Science 47, 503-522.
Converse, Philip E. (1964). The nature of belief systems in mass publics. Critical Review 18, 1-74.
Fitzpatrick, John J., Jr. and James L. Perine (2008). State labor legislation enacted in 2007.
Monthly Labor Review, January, 3-31.
Fox, Justin and Richard Van Weelden (2010). Partisanship and the effectiveness of oversight.
Journal of Public Economics 75, 1-20.
Gerring, John (1997). Ideology: a definitional analysis. Political Research Quarterly 50(4), 957-994.
Gwartney, J., R. Lawson and W. Block (1996). Economic freedom of the world: 1975-1995.
The Fraser Institute, Vancouver.
Gwartney, J., R. Lawson, H. Grubel, J. De Haan, J.-E. Sturm and E. Zandberg (2009).
Economic freedom of the world: 2009 annual report. The Fraser Institute, Vancouver.
Hagen, Jürgen von (1991). A note on the empirical effectiveness of formal fiscal restraints.
Journal of Public Economics 44(2), 199-210.
Haynes, Stephen E. and Joe A. Stone (1990). Political models of the business cycle should be revived.
Economic Inquiry 28(3), 442-465.
Hibbs, Douglas A. J. (1977). Political parties and macroeconomic policy.
American Political Science Review 71(4), 1467-1487.
Holtz-Eakin, D. (1988). The line item veto and public sector budgets: evidence from the States.
33
Journal of Public Economics 36(3), 269-292.
Karabegovic, Amela, Samida, Dexter, Schlegl, Chris M and FredMcMahon (2003). North American
economic freedom. European Journal of Political Economy 19(3), 431-452.
Knight, Brian G. (2000). Supermajority voting requirements for tax increases: evidence from the states.
Journal of Public Economics 76(1), 41-67.
Krause, George A. and Ann O’M. Bowman (2005). Adverse selection, political parties, and policy
delegation in the American federal system. Journal of Law, Economics & Organization 21(2),
359-387.
Krehbiel, Keith (1996). Institutional and Partisan Sources of Gridlock: a Theory of Divided and
Unified Government. Journal of Theoretical Politics 8, 7-40.
Krehbiel, Keith (2000). Party Discipline and Measures of Partisanship.
American Journal of Political Science 44, 212-227.
Llavador, Humberto and Angel Solano-García (2011). Immigration policy with partisan parties.
Journal of Public Economics 95, 134-142.
Lowry, Robert C. (2008). Fiscal policy in the American states. In Virginia Gray and Russell L. Hanson
(eds.) Politics in the American States: A comparative analysis. 9th edition, CQ Press, Washington, D.
C., 287-315.
Lowry, Robert C., Alt, James E. and Karen E. Ferree (1998). Fiscal policy outcomes and electoral
accountability in American states. American Political Science Review 92(4), 759-774.
McCarty, Nolan M. and Keith Poole (1995). Veto power and legislation: an empirical analysis of
executive and legislative bargaining from 1961-1986. Journal of Law, Economics & Organization,
11(2), 282-312.
McCloskey, Donald N. (1985). The loss function has been mislaid: the rhetoric of significance tests.
American Economic Review, 75 (2), 201-205.
Michael, Ragusa Jordan (2010). The lifecycle of public policy: An event history analysis of repeals to
landmark legislative enactments, 1951-2006. American Politics Research, 38(6), 1015-1051.
Nannestad, Peter and Martin Paldam (1997). From the pocketbook of the welfare man: a pooled crosssection study of economic voting in Denmark, 1986-92. British Journal of Political Science 27(1),
119-136.
NASBO. 2007. Budget Processes in the States; all editions 1981- 2005. National Association of State Budget
Officers, Washington DC.
Washington, DC: National Association of State Budget Officers.
Newey, Whitney K. and Kenneth D. West (1987). A simple, positive semi-definite, heteroskedasticity
34
and autocorrelation consistent covariance matrix. Econometrica 55, 703-708.
Poole, Keith T. and Howard Rosenthal (1991). On Dimensionalizing Roll Call Votes in the U.S.
Congress. American Political Science Review 85, 955-960.
Poole, Keith T. and Howard Rosenthal (2001). D-NOMINATE after 10 years: a comparative update to
Congress: a political-economic history of roll-call voting. Legislative Studies Quarterly 26(1), 5-29.
Poole, Keith T. and Howard Rosenthal (2007). Ideology and Congress.
Transaction Publishers: Piscataway, NJ.
Poterba, James M. (1994). State responses to fiscal crises: the effects of budgetary institutions and
politics. Journal of Political Economy 102(4), 799-821.
Poterba, James M. (1995). Capital budgets, borrowing rules, and state capital spending.
Journal of Public Economics 56(2), 165-187.
Poterba, James M. (1996). Budget institutions and fiscal policy in the US states.
American Economic Review 86(2), 395-400.
Potrafke, Niklas (2009). Does government ideology influence political alignment with the US?
An empirical analysis of voting in the UN General Assembly. Review of International Organizations
4(3), 245-268.
Potrafke, Niklas (2010). Does government ideology influence deregulation of product markets?
Empirical evidence from OECD countries. Public Choice 143(1-2), 135-155.
Primo, David M. (2006). Stop us before we spend again: institutional constraints on government
spending. Economics & Politics 18, 269-312.
Reed, W. Robert (2006). Democrats, republicans, and taxes: evidence that political parties matter.
Journal of Public Economics 90(4-5), 725-750.
Roodman, David (2006). How to do xtabond2: an introduction to “Difference” and “System”
GMM in Stata. Working paper 103, Center for Global Development.
Roodman, David (2009). A note on the theme of too many instruments.
Oxford Bulletin of Economics and Statistics 71(1), 135-158.
Rose, Shanna (2006). Do fiscal rules dampen the political business cycle? Public Choice 128(3), 407-431.
Saeki, Manabu (2009). Gridlock in the Government of the United States: Influence of Divided
Government and Veto Players. British Journal of Political Science 39, 587-607.
Shi, Min and Jakob Svensson, J. (2006). Political budget cycles: Do they differ across countries and
why? Journal of Public Economics 90(8-9), 1367-1389.
Stock, James H., Watson, Mark W. 2008. Heteroskedasticity-robust standard errors for
fixed-effects panel-data regression. Econometrica 76(1), 155-174.
35
Tavares, José (2004). Does left or right matter? Cabinets, credibility and fiscal adjustments.
Journal of Public Economics 88, 2447-2468.
Winters, Richard (1976). Party control and policy change. American Journal of Political Science 20, 597-636.
Wiseman, Alan E. and Jerry Ellig. 2007. The politics of wine: Trade barriers, interest groups, and the
Commerce Clause. Journal of Politics 69, 859-875.
Wooldridge, Jeffrey (2002). Econometric Analysis of Cross Section and Panel Data.
(Cambridge, MA: MIT Press)
36
Figure 1. Ideological changes in the US states, party-level, 1981-2005
37
Figure 2. Ideological changes in the US states, state-level, 1981-2005
38
6
6.5
7
7.5
8
Figure 3. Average aggregated economic freedom indicators. 1981-2005. 50 US states.
1980
1985
1990
1995
2000
2005
year
overall
takings_and_discrim_taxation
39
size_of_government
labor_market_freedom
Figure 4. Overall economic freedom indicator. 1981-2005. Individual US states.
Alaska
Arizona
Arkansas
California
Colorado
Connecticut
Delaware
Florida
Georgia
Hawaii
Idaho
Illinois
Indiana
Iowa
Kansas
Kentucky
Louisiana
Maine
Maryland
Massachusetts
Michigan
Minnesota
Mississippi
Missouri
Montana
Nebraska
Nevada
New Hampshire
New Jersey
New Mexico
New York
North Carolina
North Dakota
Ohio
Oklahoma
Oregon
Pennsylvania
Rhode Island
South Carolina
South Dakota
Tennessee
Texas
Utah
Vermont
Virginia
Washington
West Virgina
Wisconsin
Wyoming
5 6
7
8
9
5
6 7
8
9
5
6
7 8
9
5
6
7
8 9
5
6 7
8
9
5
6
7
8
9
Alabama
1990
2000
2010
1980
1990
2000
2010
1980
1990
2000
2010
1980
1990
2000
2010
1980
1990
2000
2010
1980
1990
2000
2010
5
6
7
8
9
1980
1980
1990
2000
2010
1980
1990
2000
2010
Year
Graphs by state
Figure 5. Size of government sub indicator. 1981-2005. Individual US states.
Alaska
Arizona
Arkansas
California
Colorado
Connecticut
Delaware
Florida
Georgia
Hawaii
Idaho
Illinois
Indiana
Iowa
Kansas
Kentucky
Louisiana
Maine
Maryland
Massachusetts
Michigan
Minnesota
Mississippi
Missouri
Montana
Nebraska
Nevada
New Hampshire
New Jersey
New Mexico
New York
North Carolina
North Dakota
Ohio
Oklahoma
Oregon
Pennsylvania
Rhode Island
South Carolina
South Dakota
Tennessee
Texas
Utah
Vermont
Virginia
Washington
West Virgina
Wisconsin
Wyoming
4
6
8
10
4
6
8
10
4
6
8
10
4
6
8
10
4
6
8
10
4
6
8
10
Alabama
1990
2000
2010
1980
1990
2000
2010
1980
4
6
8
10
1980
1980
1990
2000
2010
1980
1990
2000
2010
Year
Graphs by state
40
1990
2000
2010
1980
1990
2000
2010
1980
1990
2000
2010
1980
1990
2000
2010
Figure 6. Takings and discriminatory taxation sub indicator. 1981-2005. Individual US states.
Alaska
Arizona
Arkansas
California
Colorado
Connecticut
Delaware
Florida
Georgia
Hawaii
Idaho
Illinois
Indiana
Iowa
Kansas
Kentucky
Louisiana
Maine
Maryland
Massachusetts
Michigan
Minnesota
Mississippi
Missouri
Montana
Nebraska
Nevada
New Hampshire
New Jersey
New Mexico
New York
North Carolina
North Dakota
Ohio
Oklahoma
Oregon
Pennsylvania
Rhode Island
South Carolina
South Dakota
Tennessee
Texas
Utah
Vermont
Virginia
Washington
West Virgina
Wisconsin
Wyoming
4
6
8
10
4
6
8
10
4
6
8
10
4
6
8
10
4
6
8
10
4
6
8
10
Alabama
1990
2000
2010
1980
1990
2000
2010
1980
1990
2000
2010
1980
1990
2000
2010
1980
1990
2000
2010
1980
1990
2000
2010
4
6
8
10
1980
1980
1990
2000
2010
1980
1990
2000
2010
Year
Graphs by state
Figure 7. Labor market freedom sub indicator. 1981-2005. Individual US states.
Alaska
Arizona
Arkansas
California
Colorado
Connecticut
Delaware
Florida
Georgia
Hawaii
Idaho
Illinois
Indiana
Iowa
Kansas
Kentucky
Louisiana
Maine
Maryland
Massachusetts
Michigan
Minnesota
Mississippi
Missouri
Montana
Nebraska
Nevada
New Hampshire
New Jersey
New Mexico
New York
North Carolina
North Dakota
Ohio
Oklahoma
Oregon
Pennsylvania
Rhode Island
South Carolina
South Dakota
Tennessee
Texas
Utah
Vermont
Virginia
Washington
West Virgina
Wisconsin
Wyoming
4
6
8
10
4
6
8
10
4
6
8
10
4
6
8
10
4
6
8
10
4
6
8
10
Alabama
1990
2000
2010
1980
1990
2000
2010
1980
4
6
8
10
1980
1980
1990
2000
2010
1980
1990
2000
2010
Year
Graphs by state
41
1990
2000
2010
1980
1990
2000
2010
1980
1990
2000
2010
1980
1990
2000
2010
9
Figure 8. Average aggregated economic freedom and common Democrat/Republican governor
coding.
Overall Economic Freedom
6
7
8
Tennessee
Florida
Virginia
Texas
New Hampshire
Delaware
Arizona
South
Carolin
Alabama
Louisiana
South Dakota
North
Carolin
Colorado
Indiana
Missouri
Nebraska
Nevada
Mississippi
Utah
Kansas
Wyoming
Maryland
Arkansas
Connecticut
Illinois
Iowa
Massachusetts
Kentucky
Oklahoma
Idaho
Pennsylvania
New JerseyNorth Dakota
New Mexico
Vermont Minnesota
OhioWisconsin
Washington
Oregon
Alaska
Michigan
California
Hawaii
Montana
Georgia
Maine
Rhode Island
5
York
West New
Virgina
-1
-.5
0
.5
Common Democrat/Republican coding
1
9
Figure 9. Average aggregated economic freedom and DW Democrat/Republican governor coding.
Overall Economic Freedom
6
7
8
Tennessee
Texas
Florida
New Hampshire
Virginia Delaware
Arizona
South Carolin
Alabama
Louisiana
South Dakota
NorthIndiana
Carolin
Georgia
Colorado
Missouri
Nevada
Mississippi
Utah
Wyoming
Maryland
ArkansasKansas
Connecticut
Illinois
Iowa
Massachusetts
Kentucky
Oklahoma
Idaho
Pennsylvania
North
Dakota
New Jersey
New Mexico
Vermont
Minnesota
Ohio
Washington
Wisconsin
Oregon
Michigan
California
Hawaii
Montana
Maine
Rhode Island
5
New York
West Virgina
-.4
-.2
0
.2
PR Democrat/Republican
42
.4
Table 1: Regression Results.
Error Correction Model
Heteroskedastic and autocorrelation consistent (HAC) Newey-West type standard errors.
Dependent variable: First difference of the economic freedom indicators (overall).
Common ideology measures (Republican).
(1)
Ideology Governor t-1
∆ Ideology Governor
Ideology House t-1
∆ Ideology House
(2)
(3)
0.0083
[1.59]
0.0163*
[1.95]
(4)
(5)
(6)
0.0054
[1.09]
0.0122
[1.53]
0.0058
[0.85]
-0.0004
[0.05]
Ideology Senate t-1
0.0057
[0.85]
0.0013
[0.20]
-0.0069
[0.97]
0.0102
[1.33]
∆ Ideology Senate
Dependency ratio t-1
0.5083
[0.59]
-4.8002***
[3.08]
4.4064
[1.24]
-4.0618
[0.78]
0.3049
[0.58]
2.4311*
[1.99]
-0.7197
[0.90]
4.7961
[1.49]
-6×10-10
[0.09]
7×10-9
[0.06]
0.1971
[0.43]
6.2192***
[7.97]
-0.0028
[1.67]
-0.0127***
[3.08]
0.1935***
[3.33]
-0.0331
[0.33]
0.0338
[1.42]
0.0084
[0.19]
0.0301*
[1.72]
0.0325
∆ Dependency ratio
Females t-1
∆ Females
Hispanics t-1
∆ Hispanics
Blacks t-1
∆ Blacks
Population t-1
∆ Population
Employment t-1
∆ Employment
Consumer Price Index t-1
∆ Consumer Price Index
Fiscal Equalization t-1
∆ Fiscal Equalization
Supermajority t-1
∆ Supermajority
Balanced Budget t-1
∆ Balanced Budget
43
0.5568
[0.65]
-4.2533**
[2.66]
7.4346**
[2.15]
-1.179
[0.19]
0.5555
[0.97]
2.8976
[1.40]
-0.7712
[0.90]
4.2553
[1.29]
-5×10-9
[0.64]
6×10-9
[0.06]
0.2134
[0.45]
6.1422***
[8.15]
-0.0024
[1.45]
-0.0132***
[3.21]
0.1886***
[3.35]
-0.0237
[0.23]
0.0294
[1.26]
0.0027
[0.06]
0.0293
[1.61]
0.0424
-0.0054
[0.91]
0.0143*
[2.00]
0.6754
[0.81]
-4.0575**
[2.50]
7.3059**
[2.11]
-1.7843
[0.30]
0.5543
[0.96]
3.2479
[1.55]
-0.768
[0.90]
3.5261
[1.10]
-3×10-9
[0.43]
1×10-8
[0.13]
0.268
[0.57]
6.1953***
[8.51]
-0.0027
[1.52]
-0.0135***
[3.09]
0.1876***
[3.31]
-0.0195
[0.20]
0.0309
[1.44]
-0.0053
[0.12]
0.0284
[1.62]
0.0418
[0.92]
[1.09]
Fixed state effects
Yes
Yes
Yes
Yes
Yes
Fixed period effects
Yes
Yes
Yes
Yes
Yes
Observations
1152
1143
1143
1152
1143
Number of states
48
48
48
48
48
R-squared (overall)
0.29
0.29
0.29
0.21
0.19
Notes: Absolute value of t statistics in brackets; * significant at 10%; ** significant at 5%; *** significant at 1%
44
[1.09]
Yes
Yes
1143
48
0.20
Table 2: Regression Results.
Error Correction Model
Heteroskedastic and autocorrelation consistent (HAC) Newey-West type standard errors.
Dependent variable: First difference of the economic freedom indicators (overall).
DW ideology measures (Republican).
(1)
Ideology Governor t-1
∆ Ideology Governor
Ideology House t-1
∆ Ideology House
(2)
(3)
0.0280*
[1.91]
0.0467**
[2.34]
(4)
(5)
(6)
0.0214
[1.60]
0.0384*
[1.93]
0.1116
[1.19]
0.1847
[1.31]
Ideology Senate t-1
0.1631*
[2.01]
0.1687
[1.32]
0.0035
[0.05]
0.3038**
[2.64]
∆ Ideology Senate
Dependency ratio t-1
0.5503
[0.64]
-4.7313***
[3.08]
4.626
[1.30]
-3.6149
[0.69]
0.2851
[0.54]
2.4235*
[1.98]
-0.7713
[0.97]
5.0452
[1.53]
-3×10-10
[0.05]
9×10-9
[0.09]
0.2004
[0.44]
6.2085***
[7.90]
-0.0029*
[1.73]
-0.0126***
[3.09]
0.1925***
[3.33]
-0.0328
[0.33]
0.0342
[1.46]
0.0093
[0.21]
0.029
[1.66]
0.0312
∆ Dependency ratio
Females t-1
∆ Females
Hispanics t-1
∆ Hispanics
Blacks t-1
∆ Blacks
Population t-1
∆ Population
Employment t-1
∆ Employment
Consumer Price Index t-1
∆ Consumer Price Index
Fiscal Equalization t-1
∆ Fiscal Equalization
Supermajority t-1
∆ Supermajority
Balanced Budget t-1
∆ Balanced Budget
45
0.7181
[0.87]
-4.2875***
[2.74]
7.8908**
[2.21]
-0.0319
[0.01]
0.6413
[1.12]
3.22
[1.55]
-0.7986
[0.91]
4.6231
[1.38]
-6×10-9
[0.86]
6×10-9
[0.04]
0.2416
[0.52]
6.1857***
[8.22]
-0.0017
[1.09]
-0.0123***
[2.89]
0.1934***
[3.50]
-0.0287
[0.29]
0.0243
[1.10]
0.0055
[0.12]
0.0299*
[1.77]
0.0434
0.089
[1.17]
0.3206***
[2.83]
0.7858
[0.93]
-4.3062**
[2.60]
7.8148**
[2.22]
-0.7028
[0.12]
0.5232
[0.91]
3.4640*
[1.78]
-0.8743
[1.02]
4.0471
[1.24]
-5×10-9
[0.76]
-1×10-8
[0.10]
0.2617
[0.56]
6.1882***
[8.38]
-0.0019
[1.07]
-0.0126***
[3.03]
0.1867***
[3.31]
-0.03
[0.30]
0.0292
[1.32]
0.0013
[0.03]
0.0297*
[1.81]
0.0385
[0.90]
[1.15]
Fixed state effects
Yes
Yes
Yes
Yes
Yes
Fixed period effects
Yes
Yes
Yes
Yes
Yes
Observations
1152
1143
1143
1152
1143
Number of states
48
48
48
48
48
R-squared (overall)
0.29
0.27
0.30
0.21
0.16
Notes: Absolute value of t statistics in brackets; * significant at 10%; ** significant at 5%; *** significant at 1%
46
[0.98]
Yes
Yes
1143
48
0.18
Table 3: Regression Results.
Error Correction Model
Heteroskedastic and autocorrelation consistent (HAC) Newey-West type standard errors.
Dependent variable: First difference of the economic freedom indicators (Subindicators).
(1)
(2)
(3)
Common ideology measures
Area 1
Size of Government
Ideology Governor t-1
∆ Ideology Governor
0.0068
[1.07]
0.0155
[1.32]
Ideology House t-1
∆ Ideology Senate
Area 2
Taxation
0.006
[0.69]
0.0188
[1.52]
0.0191
[0.84]
0.0407
[1.43]
Ideology Senate t-1
Area 3
Labor Market Freedom
Ideology House t-1
∆ Ideology House
0.1207
[0.95]
-0.025
[0.10]
-0.0198*
[1.75]
0.0292**
[2.30]
∆ Ideology Senate
∆ Ideology Governor
0.028
[0.27]
0.3576*
[1.96]
0.0118
[1.03]
0.0027
[0.21]
∆ Ideology House
Ideology Governor t-1
0.1375
[1.27]
0.5882***
[3.12]
-0.0067
[0.69]
0.0111
[0.98]
Ideology House t-1
(6)
0.0289
[1.64]
0.051
[1.59]
Ideology Senate t-1
∆ Ideology Governor
(5)
DW ideology measures
-0.0131*
[1.74]
0.0001
[0.01]
∆ Ideology House
Ideology Governor t-1
(4)
0.0041
[0.77]
-0.0026
[0.44]
-0.0227
[0.22]
0.5505***
[2.91]
0.0188
[1.23]
0.0033
[0.20]
0.0139
[1.63]
0.0121
[1.34]
Ideology Senate t-1
0.1852*
[1.76]
0.019
[0.11]
0.0071
0.1992***
[0.93]
[2.71]
∆ Ideology Senate
0.0049
0.1485
[0.49]
[1.25]
Notes: Absolute value of t statistics in brackets; * significant at 10%; ** significant at 5%; *** significant at 1%
47
Table 4: Regression Results.
Error Correction Model
Heteroskedastic and autocorrelation consistent (HAC) Newey-West type standard errors.
Dependent variable: First difference of the economic freedom indicators.
DW ideology measures (Republican).
Overall Divided Government.
(1)
0.0299
[1.42]
0.0353*
[1.81]
-0.0122
[1.53]
0.0104
[0.79]
(2)
Area1
Size of Government
-0.0031
[0.09]
0.0449
[1.49]
-0.0191
[1.26]
0.0238
[1.32]
(3)
Area2
Taxation
0.045
[1.35]
0.0377
[1.45]
-0.0166
[1.24]
0.0082
[0.33]
(4)
Area3
Labor Market Freedom
0.0313
[1.29]
0.0095
[0.60]
0.0008
[0.10]
-0.0254**
[2.58]
-0.0096
[0.39]
0.0603
[1.23]
-0.0373
[0.95]
-0.0234
[0.76]
-0.0009
[0.04]
0.1664**
[2.02]
0.1712
[1.40]
-0.0095
[1.13]
0.0129
[0.96]
0.0013
[0.05]
0.1573
[1.38]
0.6087***
[3.34]
-0.0146
[1.05]
0.0284
[1.49]
-0.0035
[0.09]
0.1527
[1.14]
-0.0337
[0.14]
-0.0144
[1.01]
0.0085
[0.34]
-0.0018
[0.08]
0.1559
[1.56]
0.0182
[0.10]
0.0032
[0.43]
-0.0230**
[2.22]
-0.0137
[0.32]
-0.0503
[0.70]
-0.0712
[0.92]
0.0512
[1.04]
0.1309
[0.57]
0.1233
[1.59]
0.3382***
[2.83]
-0.0107
[1.37]
0.0116
[0.86]
0.0017
[0.01]
0.0659
[0.57]
0.3684*
[1.94]
-0.015
[1.11]
0.0263
[1.44]
0.5188
[1.10]
0.0793
[0.77]
0.6045***
[3.14]
-0.0155
[1.19]
0.0085
[0.34]
-0.344
[1.62]
0.1807**
[2.58]
0.1374
[1.10]
0.0014
[0.18]
-0.0248**
[2.43]
-0.0639
[1.47]
-0.0753
[1.09]
-0.1914**
[2.59]
0.029
[0.60]
Overall
Ideology Governor t-1
Ideology Governor
∆ Ideology Governor
Divided Government t-1
∆ Divided Government
Ideology Governor t-1*
Divided Government t-1
∆ Ideology Governor*
∆ Divided Government
Ideology House t-1
Ideology House
∆ Ideology House
Divided Government t-1
∆ Divided Government
Ideology House t-1*
Divided Government t-1
∆ Ideology House*
∆ Divided Government
Ideology Senate t-1
Ideology Senate
∆ Ideology Senate
Divided Government t-1
∆ Divided Government
Ideology Senate t-1*
Divided Government t-1
∆ Ideology Senate*
∆ Divided Government
0.1289
0.0465
0.3082
-0.1826
[0.79]
[0.15]
[1.16]
[0.95]
Notes: Absolute value of t statistics in brackets; * significant at 10%; ** significant at 5%; *** significant at 1%
48
Table 5a: Regression Results.
Error Correction Model
Heteroskedastic and autocorrelation consistent (HAC) Newey-West type standard errors.
Dependent variable: First difference of the economic freedom indicators.
DW ideology measures (Republican).
Divided Government House (Proposal and Approval Division).
(1)
(2)
Area1
Size of Government
(3)
Area2
Taxation
(4)
Area3
Labor Market Freedom
.0249
[.0196]
.0368**
[.0184]
-.0128
[.0084]
.0069
[.0083]
.0021
[.0158]
.0156
[.0129]
-.0094
[.0301]
.0458
[.0296]
-.0173
[.0152]
.0019
[.0120]
.0199
[.0166]
.0217
[.0144]
.0139
[.0299]
.0409
[.0259]
-.0129
[.0133]
.0081
[.0172]
.0030
[.0360]
.0021
[.0287]
.0552**
[.0229]
.0133
[.0147]
.0024
[.0108]
.0059
[.0099]
-.0252
[.0169]
.0105
[.0159]
.0324
[.0237]
.0355
[.0349]
-.0407
[.0377]
-.0800**
[.0305]
-.0577***
[.0219]
-.0096
[.0236]
.0379
[.0504]
.0315
[.0318]
-.0125
[.0256]
-.0242
[.0309]
.0914**
[.0421]
-.0152
[.0285]
Overall
Ideology Governor t-1
∆ Ideology Governor
Proposal Division t-1
Ideology Governor
Approval Division t-1
∆ Proposal Division
∆ Approval Division
Ideology Governor t-1*
Proposal Division t-1
Ideology Governor t-1*
Approval Division t-1
∆ Ideology Governor*
∆ Proposal Division
∆ Ideology Governor*
∆ Approval Division
.0144
.0255
-.1199***
[.0207]
[.0234]
[.0411]
Notes: Standard errors in brackets; * significant at 10%; ** significant at 5%; *** significant at 1%
49
.0276
[.0289]
Table 5b: Regression Results.
Error Correction Model
Heteroskedastic and autocorrelation consistent (HAC) Newey-West type standard errors.
Dependent variable: First difference of the economic freedom indicators.
DW ideology measures (Republican).
Divided Government House (Proposal and Approval Division).
(1)
(2)
Area1
Size of Government
(3)
Area2
Taxation
(4)
Area3
Labor Market Freedom
.1634**
[.0702]
.1292
[.1212]
-.0156*
[.0091]
.0085
[.0083]
.0033
[.0175]
.0189
[.0136]
.1547
[.0093]
.5540
[.1832]
-.0186
[.0149]
.0084
[.0119]
.0203
[.0185]
.0287*
[.0164]
.1477
[.1102]
-.1054
[.2517]
-.0189
[.0164]
.0082
[.0174]
.0029
[.0371]
.0072
[.0299]
.1488
[.0999]
.0177
[.1891]
-.0008
[.0118]
.0061
[.0100]
-.0259
[.0173]
.0100
[.0165]
-.0549
[.0548]
-.1127
[.0799]
-.0872
[.1094]
-.0397
[.0619]
.0487
[.0454]
.0935
[.0847]
-.0113
[.0948]
.0843
[.0551]
.1604
[.1244]
-.1257
[.2656]
.4256
[.3345]
.0541
[.1865]
Overall
Ideology Governor t-1
∆ Ideology Governor
Proposal Division t-1
Approval Division t-1
Ideology House
∆ Proposal Division
∆ Approval Division
Ideology Governor t-1*
Proposal Division t-1
Ideology Governor t-1*
Approval Division t-1
∆ Ideology Governor*
∆ Proposal Division
∆ Ideology Governor*
∆ Approval Division
.3344
.4576
.7394
[.2753]
[.3909]
[.4861]
Notes: Standard errors in brackets; * significant at 10%; ** significant at 5%; *** significant at 1%
50
-.4775*
[.2642]
Table 5c: Regression Results.
Error Correction Model
Heteroskedastic and autocorrelation consistent (HAC) Newey-West type standard errors.
Dependent variable: First difference of the economic freedom indicators.
DW ideology measures (Republican).
Divided Government House (Proposal and Approval Division).
(1)
(2)
Area1
Size of Government
(3)
Area2
Taxation
(4)
Area3
Labor Market Freedom
.1193
[.0734]
.3292**
[.1071]
-.0141
[.0086]
.0059
[.0082]
.0057
[.0184]
.0125
[.0145]
.0764
[.1112]
.3728*
[.1894]
-.0146
[.0145]
.0042
[.0121]
.0233
[.0179]
.0198
[.0154]
.0701
[.0969]
.5746***
[.1693]
-.0189
[.0150]
.0063
[.0164]
.0034
[.0381]
.0004
[.0304]
.1648**
[.0649]
.1377
[.1179]
.0008
[.0119]
.0034
[.0101]
-.0245
[.0184]
.0095
[.0176]
-.0528
[.0674]
-.0419
[.0499]
-.1860
[.1138]
.0059
[.0592]
.0117
[.0605]
-.0069
[.0593]
-.0119
[.0982]
.0502
[.0619]
-.1124
[.2279]
-.1316
[.3831]
.0975
[.3285]
-.1785
[.3353]
Overall
Ideology Senate t-1
∆ Ideology Senate
Proposal Division t-1
Approval Division t-1
Ideology Senate
∆ Proposal Division
∆ Approval Division
Ideology Senate t-1*
Proposal Division t-1
Ideology Senate t-1*
Approval Division t-1
∆ Ideology Senate*
∆ Proposal Division
∆ Ideology Senate*
∆ Approval Division
.2869*
.3549
.4098*
[.1494]
[.233]
[.2309]
Notes: Standard errors in brackets; * significant at 10%; ** significant at 5%; *** significant at 1%
51
-.0904
[.1802]
Appendix: Data description and sources
Descriptive Statistics. Means of the Economic Freedom indicators by state
state
Alabama
Alaska
Arizona
Arkansas
California
Colorado
Connecticut
Delaware
Florida
Georgia
Hawaii
Idaho
Illinois
Indiana
Iowa
Kansas
Kentucky
Louisiana
Maine
Maryland
Massachusetts
Michigan
Minnesota
Mississippi
Missouri
Montana
Nebraska
Nevada
New Hampshire
New Jersey
New Mexico
New York
North Carolin
North Dakota
Ohio
Oklahoma
Oregon
Pennsylvania
Rhode Island
South Carolin
South Dakota
Tennessee
Texas
Utah
Vermont
Virginia
Washington
West Virgina
Wisconsin
Wyoming
Total
Overall
7.68
6.22
7.78
7.12
6.19
7.49
7.07
7.85
8.01
7.53
6.18
6.78
7.07
7.48
7.04
7.15
7.00
7.66
5.88
7.12
7.01
6.19
6.34
7.28
7.46
6.09
7.36
7.34
7.90
6.64
6.50
5.51
7.54
6.64
6.33
6.94
6.26
6.68
5.80
7.71
7.61
8.27
8.06
7.16
6.35
7.86
6.33
5.48
6.28
7.14
6.97
Area1
7.25
5.78
7.98
7.79
6.32
7.98
7.64
8.57
8.21
8.27
7.10
7.49
7.44
8.24
7.52
7.92
7.52
7.51
6.31
7.40
7.26
6.31
6.85
7.10
8.20
6.26
8.50
8.20
8.47
7.37
7.04
5.87
8.04
7.22
6.26
7.56
6.43
6.66
5.75
7.51
8.36
8.32
8.57
7.77
6.63
8.34
6.78
5.65
6.80
7.49
7.36
Area2
7.84
7.23
6.82
7.02
6.12
7.28
7.14
8.16
7.26
7.39
5.88
6.32
7.38
7.76
6.94
6.58
7.09
7.59
5.46
7.16
7.31
6.79
6.19
6.65
7.82
6.30
6.93
7.31
8.30
6.64
6.25
5.48
7.48
6.36
6.59
6.75
6.73
7.27
5.54
6.99
7.87
8.16
8.04
7.16
6.01
7.75
6.54
5.54
6.18
7.14
6.93
52
Area3
7.98
5.66
8.55
6.54
6.11
7.22
6.47
6.82
8.56
6.98
5.56
6.51
6.37
6.45
6.68
6.94
6.38
7.90
5.90
6.80
6.50
5.46
6.00
8.13
6.34
5.70
6.62
6.50
6.90
5.86
6.21
5.18
7.13
6.35
6.17
6.48
5.59
6.11
6.10
8.63
6.58
8.33
7.57
6.54
6.42
7.48
5.64
5.23
5.84
6.81
6.62
Frequency
25
25
25
25
25
25
25
25
25
25
25
25
25
25
25
25
25
25
25
25
25
25
25
25
25
25
25
25
25
25
25
25
25
25
25
25
25
25
25
25
25
25
25
25
25
25
25
25
25
25
1250
Descriptive Statistics.
Variable
Observations
Mean
Std. Dev.
Min
Max
1250
6.97
0.75
4.90
9.00
1250
7.36
0.98
3.80
9.70
1250
6.93
0.82
4.60
9.20
1250
6.62
0.91
4.20
9.10
1250
1207
1207
-0.02
-0.22
-0.12
0.99
0.97
0.99
-1
-1
-1
1
1
1
Karabegovic et al.
(2003)
Karabegovic et al.
(2003)
Karabegovic et al.
(2003)
Karabegovic et al.
(2003)
Own collection
Own collection
Own collection
1200
0.03
0.36
-0.54
0.72
Own collection
1191
0.00
0.17
-0.45
0.52
Own collection
1191
0.00
0.17
-0.44
0.55
Own collection
1250
0.59
0.49
0
1
Own collection
1250
0.58
0.49
0
1
1250
0.61
0.49
0
1
1250
0.10
0.09
0.00
0.37
1250
0.06
0.08
0.00
0.44
1250
0.51
0.01
0.47
0.52
Overall economic freedom
index
Size of government
(sub index)
Takings and discriminatory
taxation (sub index)
Labor market freedom
(sub index)
Ideology, governors
Ideology House
Ideology Senate
Ideology, governors
(Poole & Rosenthal)
Ideology House
(Poole & Rosenthal)
Ideology Senate
(Poole & Rosenthal)
Divided Government
(common)
Divided Government
(Proposal division)
Divided Government
(Approval division)
Blacks (as a share of total
population)
Hispanics (as a share of total
population)
Females (as a share of total
population)
Age 65 and older (as a share
of total population)
Aged 15 and younger (as a
share of total population)
Population
Employment
GDP per capita
Consumer Price Index
1250
0.12
0.02
0.03
0.18
1250
1250
1250
1250
1250
0.22
5205564
0.56
142.14
27800.21
0.02
5667141
0.06
31.40
7337.872
0.18
418493
0.37
87.20
14705.49
0.33
3.58E+07
0.73
212.70
60390.92
Supermajority
1250
0.21
0.41
0
1
Balanced Budget
No Carry Over
1250
957
0.80
0.74
0.40
0.44
0
0
1
1
Source
Own collection
Own collection
Census Bureau
Census Bureau
Census Bureau
Census Bureau
Census Bureau
1250
1.12
0.29
0.57
Fiscal Transfers
Note: * is constructed on the basis of information drawn from the BLS (2010).
53
2.33
BEA (2010)
BEA (2010)
BEA (2010)
BLS (2010)*
Knight (2000) and
NASBO
NASBO
NASBO
Tax Foundation
(2010)