Do ‘Labour’ Councils Lower Economic Freedom?	151

Reed finds that the political party of the governor has little effect on taxes. This latter result supports earlier findings by Besley and Case (1995). Thus, the influence of administrators’ political affiliation on economic freedom remains unsettled.

One of the more well-known works by Caplan (2001) presents a model where the government becomes a ‘leviathan’ in that it grows faster and further than the desires of its citizenry. In this model, a government wants to expand its own power since this power then becomes self-reinforcing and leads to even greater power. One must keep in mind that if one political party is more conservative, there could be real limits placed on the growth of government when this party is in power, thus thwarting the ‘leviathan’ growth potential. Moreover, regarding this idea, Campbell et al. (2007) extend this model to include real growth rates of government with mixed results. The authors state that “as political parity increases, the capacity of the minority to block the majority increases thereby encouraging more interparty logrolling (ibid: p. 579).” These results seem to counter the results by Bjornskov and Potrafke (2013).

Given the absence of data on the EFW index for the UK counties, the objective of this paper is to use data on the direct/specific components of the EFW index, such as government expenses, sound money and freedom to internationally trade, to test the idea that the political affiliation of the county councils in the UK has an appreciable impact on those sub-components. The overall economic freedom measure from this index is comprised of the following three sub-components for which data can be found across the UK counties. i) size of government, as government spending (a higher government expenditure is considered as impinging on economic freedom). ii) sound money is proxied by inflation, since inflation erodes the value of rightfully earned wages and savings, sound money is thus essential to protect property rights. When inflation is high, it becomes difficult for individuals to plan for the future and, thus, to use economic freedom effectively. iii) freedom to trade internationally, i.e. freedom to exchange is essential to economic freedom, which is reduced when freedom to exchange does not include businesses and individuals in other nations.

In reality, there can be certain reasons why a county’s voters would elect a particular individual as member of the council and these reasons can also have an influence on the types of policy changes. Thus, the relationship between the party of the council and the abovementioned dimensions of economic freedom is potentially endogenous. Therefore, the hypothesis that the traits of economic freedom, mainly less government intervention in markets, would be more commonly associated with Conservative than Labour councils, is one that must be tested with a careful empirical approach. The analysis employs a regression discontinuity design that leverages the fact that the party of the council is a deterministic function of the vote margin between the Labour candidate and the Conservative candidate. This approach allows us to compare counties which are similarly situated except for their choice of the party of the council. To foreshadow our results, they provide supportive evidence that Labour councils lead to less economic freedom. In fact, we find strong evidence that a Labour council leads to a higher size of government.

152Bulletin of Monetary Economics and Banking, Volume 21, Number 2, October 2018

The remainder of the paper proceeds as follows. Section II discusses the methodological approach and the dataset employed in this study, while Section III presents the empirical results and discussions. Finally, Section IV concludes.

II.METHODOLOGY AND DATA

A. Methodology

A starting point for several previous studies about the relationship between a council’s partisan ideology and economic freedom is a fixed-effects OLS model:

(1)

where EFWit stands for county i’s component of economic freedom under examination in year t, Dit is an indicator variable, which equals 1 if county 𝑖’s council is from the Labour Party, and 0 if the council comes from the Conservative Party. The coefficient 𝛽 is a consistent estimator of 𝐷, if other potentially confounding variables, both time-varying and time-invariant, are accounted for in Equation

(1). Time-varying confounders that might be related to both the component of economic freedom and the party of a county’s council might include the level of education in a county or the average income in that county. These and other observable variables are potentially accounted for in the vector 𝑿. Additionally, the term 𝛼 accounts for both observable and unobservable county fixed-effects, which could potentially influence the party of the county’s council, 𝛿 accounts for year-specific factors, which influence economic freedom that could be correlated with the party of a council, and finally ε denotes the error term. However, even after accounting for observable confounding variables, this fixed-effects model is still unable to account for time-varying unobservable factors that could influence a county’s level of economic freedom and the party of its council. For example, if an unobserved shock leads to a lower size of government and also increases the likelihood that a council from the Labour Party would be elected, then the estimated effect of a council from the Labour Party on this particular component of economic freedom would also be underestimated. Thus, we cannot rely on Equation (1) to provide a causal estimate of a council from the Labour Party’s influence on the particular sub-component of economic freedom.

In order to overcome the endogeneity issue for unobservable shocks, we follow Lee et al. (2004) and Lee (2008) and employ a regression discontinuity approach. This approach relies on the fact that the party of the council in power is a deterministic function of the vote margin. That is:

(2)

where vit is the difference between the vote share of the first and second place candidates and vit>0, if Dit = 1 and vit < 0, if Dit = 0. We assume that unobservable variables that are not absorbed by the county fixed-effects vary with vote margin. In other words, once we condition on vote margin, 𝛽 can be estimated without bias, which can be expressed formally with:

Do ‘Labour’ Councils Lower Economic Freedom?	153

(3)

where 𝑓(vit) is a flexible function of the vote margin. A final assumption that must be made is that (vit) is continuous. Requiring that the vote share function is continuous from both directions as it approaches the zero-vote margin position, allows us to approximate the level of economic freedom on either side of the Labour vote margin. We must make this approximation, because even the closest votes over our sample period have margins of hundreds of votes. We rely on both a linear and cubic control function in our estimation of Equation (3).

The use of the regression discontinuity method thus allows us to compare the components of economic freedom of councils from the Labour Party who just barely won their election with councils from the Conservative Party who just barely won theirs. This is an important point, because the estimation of Equation

(3)does not allow us to comment on the level of economic freedom in a county, where the council of either party won by a relatively large margin. Councils can win elections, because of past policies, because of past economic performance, or because of loyalty to the particular political party, amongst other reasons. These reasons can have implications for the types of policies the council will enact that contribute to a county’s level of economic freedom. Moreover, these reasons can affect not only who wins the race for council, but by how much (that is, the win margin).

The validity of the regression discontinuity approach hinges on the requirement that once we control for vote margin and other observable time- varying confounding variables, a county’s economic freedom score does not change as we cross the threshold from a Conservative council to a Labour council, except for the impact of the council’s ideology. In order to assess this validity, we follow a recommendation by Lee and Lemieux (2010) and regress each of the potentially confounding variables on the Labour council variable. If these variables are continuous across the 50%-50% cutoff, then the coefficient on Labour council should be close to zero.

B. Data

The starting point for this project is the UK county-level annual data, spanning the period 2010-2016 period. These years are chosen based on data availability and primarily on the fact that counties experienced significant reforms prior to 2010. In particular, the UK local (county) governments are elected bodies that operate in territorially bounded geographical areas, employ professional career staff, and generally receive over two-thirds of their income from the UK central government. They are multi-purpose authorities delivering services in the areas of education, social care, land-use planning, waste management, public housing, leisure and culture, and welfare benefits. In the UK, prior to the consolidations that took place in 2009, there were 386 county governments of five types, 32 London boroughs, 36 metropolitan boroughs, and 46 unitary authorities, mostly in urban areas, delivering all of the services listed above. In the rural areas, by comparison, there were 34 county councils administering education and social services, as well as 238 district councils providing welfare and regulatory services.

Do ‘Labour’ Councils Lower Economic Freedom?	155

Table 1.

Descriptive Statistics

The table provides summary statistics for seven variables. SD denotes standard deviation, minimum statistics (Min.) and maximum (Max.) statistics together with their mean are noted.

Variable	Mean	SD	Min.	Max.
Labour Council	0.126	0.501	0.000	0.219
Personal Income (2010 £, thousands)	29.56	6.48	15.29	44.78
Percent with Bachelor's Degree	0.157	0.046	0.068	0.317
Percent in Metro Area	0.257	0.191	0.168	0.273
Inflation (%)	1.27	0.16	1.15	1.38
Net Service Expenditure Per Capita	9,367.5	1.079	9,35	9,385
Trade Openness (as % of GDP)	0.65	1.02	0.54	0.78

The table provides summary statistics for seven variables. SD denotes standard deviation, minimum statistics (Min.) and maximum (Max.) statistics together with their mean are noted.

III.EMPIRICAL ANALYSIS

A. Baseline Results

Table 2 provides the results from Equation (3) using the entire sample of counties and a linear vote margin control function. Recall that an important assumption made in Section II was that observable and unobservable time-varying confounding variables vary continuously with the vote margin. Given this assumption, the inclusion of the control variables in the vector 𝑿𝒊𝒕 in Equation (3) is not necessary. However, as pointed out by Fredricksson et al. (2013), if these control variables are not endogenous to the council’s party affiliation, their inclusion can potentially make the estimate of the effect the council’s party more precise.

The findings document that with respect to the component of the government size (net service expenditure per capita-Column 1) there is statistical (at 1%) evidence that a Labour council can implement policies that lead to higher size of government. Moreover, the size of the estimated coefficient indicates that a Labour council leads to a size of government that is higher by approximately 0.47, which economically is a substantial effect. In particular, the standard deviation of the size of government is 1.08, implying the average effect of a Labour council on this sub- component of economic freedom is 43.4% of the standard deviation. With respect to the other two sub-components of economic freedom, the results indicate that a Labour council contributes 35.6% to higher inflation and 35.4% to weaker trade openness.

The control variables included in this regression specification also display statistically significant relationships across the three specifications of economic freedom. In particular, personal income per capita is positively correlated with all three sub-components of economic freedom, while the higher the percentage of

156Bulletin of Monetary Economics and Banking, Volume 21, Number 2, October 2018

people with a college degree, the higher the government size and trade openness are, and the lower inflation is. Finally, there is an indirect association between the percentage of population living in metropolitan counties and the size of government, and a positive association between this part of the population and both inflation and trade openness.

Table 2.

The Effect of a Labour Council on Economic Freedom (Linear Version)

The table reports the linear estimates of a Labour council on economic freedom. Figures in brackets denote p-values. Finally, *: p≤0.10; **: p≤0.05; ***: p≤0.01.

		Net Service		Trade
	Variables	Expenditure	Inflation
				Openness
		Per Capita
		(1)	(2)	(3)
	Labour Council	0.468***	0.057***	-0.361***
		[0.00]	[0.00]	[0.00]
	Personal Income Per Capita(-1)	1.094***	1.659***	2.344***
		[0.00]	[0.00]	[0.00]
	Percent with College Degree(-1)	0.318**	-0.094*	1.028***
		[0.03]	[0.06]	[0.00]
	Percent of Population in Metro Areas(-1)	-0.084**	0.127**	0.562***
		[0.05]	[0.04]	[0.00]
	Adjusted R2	0.73	0.68	0.82
	No. Observations	1,128	1,128	1,128

Table 3 repeats the previous analysis, but this time it employs a non-linear (cubic) vote margin control function. The new findings provide strong statistical support to those reported in Table 2.

Table 3.

The Effect of a Labour Council on Economic Freedom (Non-Linear-Cubic Version)

The table reports the non-linear estimates of a Labour council on economic freedom. Figures in brackets denote p-values. Finally, *: p≤0.10; **: p≤0.05; ***: p≤0.01.

		Net Service		Trade
	Variables	Expenditure	Inflation
				Openness
		Per Capita
		(1)	(2)	(3)
	Labour Council	0.496***	0.069***	-0.388***
		[0.00]	[0.00]	[0.00]
	Personal Income Per Capita(-1)	1.138***	1.594***	2.137***
		[0.00]	[0.00]	[0.00]
	Percent with College Degree(-1)	0.279**	-0.073*	0.874***
		[0.04]	[0.08]	[0.00]
	Percent of Population in Metro Areas(-1)	-0.089**	0.146**	0.528***
		[0.05]	[0.03]	[0.00]
	Adjusted R2	0.75	0.66	0.83
	No. Observations	1,128	1,128	1,128

Do ‘Labour’ Councils Lower Economic Freedom?	157

B. Validity of the Regression Discontinuity Design

Even though the evidence shows no causal impact of a council’s party on the economic freedom witnessed by a county, it is important to show that the regression discontinuity approach used in this paper is valid. This approach can only be used if the researcher is confident that time-varying county-level characteristics, which could potentially be correlated with the party of the council, are not systematically different under a Labour council that is barely elected compared to a Conservative council that is also barely elected. If any of these county-level characteristics are different on either side of the vote margin cut-off, then it would be imprudent to state that the council’s partisan affiliation is the cause of any differences between economic freedom.

A common way to test the validity of the regression discontinuity design is to regress potentially confounding variables on the Labour council variable. This method, proposed by Lee and Lemiux (2010), examines whether these variables are discontinuous at the vote margin cut-off. It is important to note that this validity test can only examine whether the observable variables, are discontinuous at the vote margin cut-off. We have no way of testing whether unobservable county-level characteristics are discontinuous at the cut-off. Tables 4 and 5 (the linear and non- linear versions, respectively) give us confidence that the other variables that are plausibly related to whether a county has a Labour council are similar as the vote margin moves from Labour to Conservative.

Table 4.

Validity of the Regression Discontinuity Design (Linear Version)

The table presents the results of testing the validity of the regression discontinuity design in its linear version by regressing potentially confounding variables on the Labour council variable. Figures in brackets denote p-values.

	Personal	Percent with	Percent of
Variables	Income Per	College	Population in
	Capita	Degree	Metro Areas
	(1)	(2)	(3)
Democratic Council	-0.00048	-0.00036	-0.00018
	[0.28]	[0.39]	[0.55]
Adjusted R2	0.28	0.22	0.14
No. of Observations	1,128	1,128	1,128

Table 5.

Validity of the Regression Discontinuity Design (Non-Linear Version)

The table presents the results of testing the validity of the regression discontinuity design in its non-linear version by regressing potentially confounding variables on the Labour council variable. Figures in brackets denote p-values.

	Personal	Percent with	Percent of
Variables	Income Per	College	Population in
	Capita	Degree	Metro Areas
	(1)	(2)	(3)
Democratic Council	-0.00048	-0.00036	-0.00018
	[0.28]	[0.39]	[0.55]
Adjusted R2	0.28	0.22	0.14
No. of Observations	1,128	1,128	1,128

158Bulletin of Monetary Economics and Banking, Volume 21, Number 2, October 2018

IV. CONCLUSION

The definition of economic freedom emphasized the importance of limited government with respect to the government size, sound money (inflation) and trade openness sub-components of economic freedom. Previous research for the case of the US had found evidence that Democratic governors enact higher income taxes and that Republican governors are associated with less labor market regulation. However, researchers had also found evidence that the party of the governor had little effect on important policy outcomes which contribute to economic freedom.

Given this motivating definition and the unsettled nature of the previous research, we hypothesized that Labour councils in the UK would enact policies that would lead to lower economic freedom compared to Conservative councils. The analysis tested this hypothesis using county-level data in the UK, spanning the period 2000–2016 and employed a regression discontinuity approach in order to overcome potentially endogenous relationships between the party of the council and other factors that could affect the choice of council and the types of economic policies that are implemented.

The results provided strong statistical (and economic) support to the hypothesis. Over the time period studied, we found strong evidence that a Labour council led to lower economic freedom compared to a Conservative council. The conclusion one could draw from these results is that the partisan ideology could affect economic freedom through certain sub-channels, such as government size, sound money and the freedom to internationally trade. In other words, partisan ideology can have substantial effects on a county’s voter.

REFERENCES

Besley, T., & Case, A. (1995). Does Electoral Accountability Affect Economic Policy Choices? Evidence from Gubernatorial Term Limits. Quarterly Journal of Economics, 110, 669-797.

Besley, T., & Case, A. (2003). Political Institutions and Policy Choices: Evidence from the United States. Journal of Economic Literature, 41, 7-73.

Berggren, N. (1999). Economic Freedom and Equality: Friends or Foes? Public Choice, 100, 203-223.

Blais, A., Blake, D., & Dion, S. (1993). Do Parties Make a Difference? Parties and the Size of Government in Liberal Democracies. American Journal of Political Science, 37, 40-62.

Bjornskov, C., & Potrafke, N. (2013). The Size and Scope of Government in the US States: Does Party Ideology Matter? International Tax and Public Finance, 20, 687-714.

Campbell, N., Finney, R., & Mitchell, D. (2007). Hunting the Whale: More Evidence on State Government Leviathans. Southern Economic Journal, 74, 566-580.

Caplan, B. (2001). Has Leviathan Been Bond? A Theory of Imperfectly Constrained Government with Evidence from the States. Southern Economic Journal, 16, 825- 847.

Carter, J. (2007). An Empirical Note on Economic Freedom and Income Inequality. Public Choice, 130, 163-177.

Do ‘Labour’ Councils Lower Economic Freedom?	159

Compton, R., Giedeman, D., & Hoover, G. (2011). Panel Evidence on Economic Freedom and Growth in the United States. European Journal of Political Economy, 27, 423-435.

Compton, R., Giedeman, D., & Hoover, G. (2014). A Distributional Analysis of the Benefits of Economic Freedom. European Journal of Political Economy, 33, 121- 133.

Downs, A. (1957). An Economic Theory of Political Action in a Democracy. Journal of Political Economy, 65, 135-150.

Fredriksson, P.G., Wang, L., & Warren, P.L. (2013). Party Politics, Governors, and Economic Policy. Southern Economic Journal, 80, 106-126.

Garand, J.C. (1988). Explaining Government Growth in the U.S. States. American Political Science Review, 82, 837-849.

Gwartney, J., Lawson, R., & Block, W. (1996). Economic Freedom of the World: 1975- 1995. Vancouver: The Fraser Institute.

Hall, J.C., Stansel, D., & Tarabar, D. (2019 forthcoming). Economic Freedom Studies at the State Level: A Survey. In Economic Behavior, Economic Freedom, and Entrepreneurship. Eds. Cebula, R., Hall, J., Mixon, F., & Payne, J. Edward Elgar Publishing.

Imbeau, L.M., Petry, F., & Lamari, M. (2001). Left–Right Party Ideology and Government Policies: A Meta-Analysis. European Journal of Political Research, 40, 1-29.

Krehbiel, K. (1993). Where’s the Party? British Journal of Political Science, 23, 235- 266.

Lee, D.S., Moretti, E., & Butler, M.J. (2004). Do Voters Affect or Elect Policies? Evidence from the US House. The Quarterly Journal of Economics, 119, 807-859.

Lee, D.S. (2008). Randomized Experiments From Non-Random Selection in US House Elections. Journal of Econometrics, 142, 675-697.

Lee, D.S., & Lemieux, T. (2010). Regression Discontinuity Designs in Economics. Journal of Economic Literature, 48, 281-355.

Pettersoson-Lidbom, P. (2003). Do Parties Matter for Fiscal Policy Choices? A Regression Discontinuity Approach. Mimeo, Department of Economics, Stockholm University.

Reed, W. (2006). Democrats, Republicans, and Taxes: Evidence that Political Parties Matter. Journal of Public Economics, 90, 725-750.

Scully, G.W. (2002). Economic Freedom, Government Policy and the Trade-Off Between Equity and Economic Growth. Public Choice, 113, 77-96.

Winters, R. (1976). Party Control and Policy Change. American Journal of Political Science, 20, 597-636.

160Bulletin of Monetary Economics and Banking, Volume 21, Number 2, October 2018

This page is intentionally left blank