Fiscal imbalances and asymmetric adjustment under Labour and Conservative governments in the UK

Research in Economics 68 (2014) 208–213

Contents lists available at ScienceDirect

Research in Economics journal homepage: www.elsevier.com/locate/rie

Fiscal imbalances and asymmetric adjustment under Labour and Conservative governments in the UK Christos Kollias, Stephanos Papadamou n, Iacovos Psarianos Department of Economics University of Thessaly, Korai 43, 38333 Volos, Greece

a r t i c l e in f o

abstract

Article history: Received 16 August 2013 Accepted 4 June 2014 Available online 19 June 2014

This paper examines whether the political colour of an incumbent government affects the speed at which fiscal imbalances are corrected in the case of the UK. Using quarterly data, we examine whether Conservative or Labour governments are more prone to operate under a soft budget constraint and vis-à-vis i.e. to adhere to a hard budget strategy. The tests, using quarterly data, cover the period 1961–2011 and the results reported herein reveal differences in the speed at which fiscal imbalances are corrected by Labour and Conservative governments. The former are more inclined to operate under a soft budget constraint whereas the latter under a hard budget constraint. & 2014 University of Venice. Published by Elsevier Ltd. All rights reserved.

Keywords: Fiscal policy UK VEC models Government ideology

1. Introduction In broad terms, it is generally assumed that liberal and left leaning governments will tend to be more inclined to adopt a soft budget strategy and thus step-in to correct deficits only when they exceed a certain threshold, while Conservative governments would tend to exhibit more fiscal prudence and by and large be more deficit averse (Eslava, 2011). This hypothesis is tested in the case of the UK using quarterly data over the period 1961–2011. The use of quarterly data allows for better statistical inferences to be drawn since it offers the opportunity to more accurately depict in the time series the change in office following a general election and hence model and estimate the differences in fiscal behaviour characteristics between Conservative and Labour governments in the UK. We do that by examining particular functional forms of asymmetric adjustment back to equilibrium once fiscal imbalances are observed. The current environment of acute fiscal imbalances faced by a number of European countries brings to the forefront issues associated not only with the generation of fiscal deficits but also of the correction of such imbalances whenever they arise and the degree of adherence to fiscal rules (Bird and Mandilaras, 2013; Ioannou and Stracca, 2014; Schalck, 2012; Eslava, 2011; Holmes, 2011; Payne, 2003; Thornton, 2009). The dynamics that govern the intertemporal nexus between the two budget aggregates – expenditures and revenues – have been the subject of extensive debate and research (Apergis et al., 2012; Vamvoukas, 2012; van der Ploeg, 2010; Zapf and Payne, 2009; Chang et al., 2002; Ewing et al., 2006; Kollias and Paleologou, 2006; Giavazzi et al., 2000). Payne (2003) comprehensibly surveys and reviews the issues and the associated steadily expanding body of literature. In particular, as he notes, two hypotheses describe the relationship in question. The one proposes that tax and spending decisions are made independently of one another. Hence, an empirical investigation

n

Corresponding author. Tel.: þ30 24210 74963, þ 30 6937210506; fax: þ 30 24210 74772. E-mail addresses: [email protected], [email protected] (S. Papadamou).

http://dx.doi.org/10.1016/j.rie.2014.06.001 1090-9443/& 2014 University of Venice. Published by Elsevier Ltd. All rights reserved.

C. Kollias et al. / Research in Economics 68 (2014) 208–213

209

should not be able to trace any causal ordering. The second hypothesis, postulates some sort of interdependence between the two. This interdependence may be manifested in three different ways: fiscal synchronisation which empirically would yield results pointing to bidirectional causality, or alternatively, unidirectional causality in the form of either spend-and-tax or tax-and-spend (inter alia: Payne, 2003; Paleologou, 2013; Owoye and Onafowora, 2011; Vamvoukas, 2012; Kollias and Makrydakis, 2000; Koren and Stiassny, 1998). However, given that the two budget aggregates may have different time profiles, imbalances may very well emerge even under a fiscal synchronization regime. Hence, the way in which such imbalances are corrected by fiscal authorities emerges as an interesting theme to investigate empirically in order to shed light into the adjustment mechanisms that are set into motion in such cases. This adjustment can be achieved either through the expenditure channel or revenues, or both. Furthermore, what is also of interest is the speed at which this adjustment takes place since this can offer evidence on whether a government operates under a hard or a soft budget constraint. In the latter case, fiscal authorities would be inclined to step-in and correct deficits only when they exceed a certain threshold. A number of studies have focused on the issue of the adjustment process of fiscal imbalances which offers insights on the degree of adherence to fiscal rules and whether policy makers and governments operate under a hard or a soft budget constraint (Saunoris and Payne, 2010; Andrikopoulos et al., 2004; Arestis et al., 2004; Bajo-Rubio et al., 2004; Cipollini et al., 2009; Brunila et al., 2003; Cipollini, 2001; Garcia and Henin, 1999; Baghestani and McNown, 1994; de Haan and Sturm, 1994). As Paleologou (2013) observes, when it comes to fiscal policy behaviour, a soft budget constraint may be driven by the political and ideological underpinnings of the incumbent government or by opportunistic considerations such as to improve popularity and support from the electorate. The median voter model has been extensively used to describe political processes and economic policy platforms and decisions as politicians and parties compete for the median voter's vote (Belke and Potrafke, 2012; Milanovic, 2000; Ahmed and Greene, 2000; Holcombe and Caudill, 1985; Goff and Tollison, 2003). Belke and Potrafke (2012) note that the partisan approach is based on the assumption that politicians provide policies that reflect the preferences of their clienteles. For instance, it has been shown that increases in government spending produce distributional effects in the form of reduced inequality (Ramos and Roca-Sagales, 2008; Milanovic, 2000). Given that parties on the left of the political spectrum invariably exhibit greater sensitivity to issues of equitable income distribution, such an effect may render them more lax in the use of this fiscal instrument vis-à-vis a policy of fiscal vigilance. Similarly, given its economic stimulus potential, governments can yield to the temptation to use public spending as a stabilization and economic revitalization tool to prop-up and spur economic activity and employment possibly at the expense of fiscal judiciousness. A number of papers, using non-linear and asymmetric models, have examined the speed of adjustment when it comes to fiscal imbalances (inter alia: Saunoris and Payne, 2010; Legrenzi and Milas, 2012; Paleologou, 2013; Payne and Mohammadi, 2006; Easaw and Garratt, 1999, 2000, 2006; Legrenzi and Milas, 2012; Arestis et al., 2004). Within this strand of the literature, this paper revisits the issue by specifically allowing in the relevant tests for the political colour of the incumbent government in the UK with the introduction of a dummy variable that captures such political differences. Hence, it is hoped that possible differences in the response of Conservative and Labour administrations will be unveiled. In a nutshell, the hypothesis examined is whether imbalances are corrected at a rate that depends on the party's political preferences and ideological affiliation. Or, from a different perspective, the aversion or disposition towards deficits that different parties – Conservative or Labour – in office exhibit is comparatively assessed. Moreover, we examine whether the fiscal instrument (taxes) is flexible enough to adjust symmetrically when above or below its equilibrium level with expenditures given left – (Labour) and right – (Conservative) wing governments. In the section that follows the methodology, data and findings are presented and discussed while section three concludes the paper.

2. Data, methodology and findings As mentioned above, the question at hand is whether the political colour and ideological foundations of the incumbent government affect the speed at which fiscal imbalances are corrected once they appear. The literature postulates that Conservative governments will generally tend to be more fiscally vigilant vis-à-vis administrations that have a more liberal, left tilting political agenda (Eslava, 2011; van der Ploeg, 2010). Hence, the latter will have the propensity to exhibit more fiscal slackness and move to correct deficits at a slower adjustment pace. The data set used in this paper, consisting of government revenues and expenditures expressed as shares of GDP, is drawn from the National Statistics of the UK. It is of quarterly frequency which allows for better statistical inferences to be drawn since the quarterly nature of the time-series allows us to depict in a more accurate and time-wise realistic manner the influence exerted on the fiscal adjustment process by the political colour of the party – Labour or Conservative – in office. The sample period covers the first quarter of 1961 to the second quarter of 2011 which gives a total of 201 observations. During this period, there were a total of thirteen general elections that brought or maintained into power either the Conservative or the Labour party with the exception of the current coalition government if the Lib-Lab pact of the late 1970s is not counted as a formal government coalition. In the estimations that follow, the political dummy variable d is introduced as follows: ( 1; Labour Party in office d¼ 0; Conservative Party in office

210

C. Kollias et al. / Research in Economics 68 (2014) 208–213

Table 1 Unit root and cointegration tests, UK 1961Q1–2011Q1. Variables

ADF

Ng–Perron test statistics MZa

MZt

MSB

REV Δ(REV) EXP Δ(EXP)

 3.010  20.381**  2.319  17.878**

 7.105  96.256**  10.376  33.218**

 1.834  6.935**  2.274  4.067**

Hypothesized no. of CE(s)

Eigenvalue

Trace statistic

0.05 Critical value

None * At most 1

0.069 0.017

17.778 3.381

15.495 3.841

MPT 0.258 0.072** 0.219 0.122**

12.911 0.952** 8.803 2.787**

p-Value

Max-Eigen statistic

0.05 Critical value

p-Value

(0.0223)** (0.066)*

14.396 3.381

14.265 3.841

(0.047)** (0.066)*

nn n

Denotes evidence against the null hypothesis at the 0.05 level. Denotes evidence against the null hypothesis at the 0.10 level.

Table 2 Estimates of alternative error-correction models for ΔREV in the UK. (i) Linear model 1

C ECTt  1 (1  D)*ECTt  1 DECTt  1 (1  D)*(ECTt  1) þ D*(ECTt  1) þ (1  D)*(ECTt  1)  D*(ECTt  1)  ΔREVt  1 ΔREVt  2 ΔREVt  3 ΔREVt  4 ΔREVt  5 ΔEXPt  1 ΔEXPt  2 ΔEXPt  3 ΔEXPt  4 ΔEXPt  5 D1973 Diagnostics Adj. R2 Q-Stat (16) p-value Q-Stat (16) squared residuals p-value ARCH(16) test p-value

(ii) Asymmetric model 2

(iii) Asymmetric model 3

Coefficient

t-Statistic

Prob.

Coefficient

t-Statistic

Prob.

Coefficient

t-Statistic

Prob.

0.00053  0.13959 – – – – – –  0.44373 0.04793 0.11918 0.11109 0.06967 0.15829  0.01779 0.01535  0.12457  0.09989  0.02706

0.87  3.84 – – – – – –  6.15 0.63 1.56 1.45 0.96 1.86  0.21 0.18  1.47  1.19  3.08

0.38 0.00*** – – – – – – 0.00*** 0.53 0.12 0.15 0.34 0.06* 0.83 0.85 0.14 0.23 0.00***

0.00049 –  0.15901  0.10529 – – – –  0.44903 0.04603 0.11492 0.11039 0.06608 0.16322  0.01357 0.02957  0.11790  0.10342  0.02722

0.78 –  3.56  1.75 – – – –  6.14 0.60 1.49 1.41 0.90 1.90  0.16 0.34  1.36  1.22  3.07

0.43 – 0.00*** 0.08 – – – – 0.00*** 0.55 0.14 0.16 0.37 0.06* 0.88 0.73 0.18 0.22 0.00***

0.001274 – – –  0.278751  0.100114  0.072756  0.12821  0.459381 0.035691 0.109654 0.110059 0.07005 0.155936  0.023821 0.016085  0.115754  0.101235  0.026724

1.26 – – –  3.45  1.04  1.10  1.27  6.30 0.46 1.41 1.41 0.96 1.82  0.28 0.19  1.34  1.20  3.03

0.21 – – – 0.00*** 0.30 0.27 0.21 0.00*** 0.64 0.16 0.16 0.34 0.07* 0.78 0.85 0.18 0.23 0.00***

21.18%

21.03% 0.80 0.14 0.27

22.04% 0.68 0.12 0.22

0.81 0.31 0.46

nnn n

Denotes evidence against the null hypothesis at the 0.01 level. Denotes evidence against the null hypothesis at the 0.10 level.

We start the empirical analysis by testing for cointegration between revenues and expenditures. To this effect, we estimate a vector error correction model (VECM) (Johansen, 1988) of the form κ1

Δyt ¼ ∑ Γ ι Δyt  i þ Πyt  1 þ c þεt ι¼1

where yt ¼ ½REV; EXP0 is the set of nonstationary I(1) variables, εt  niid (0, Σ) c is a drift parameter, and Π is a (p  p) matrix of the form Π¼αβ0 where α and β are (p  1) matrices of full column rank, with β containing the one cointegrating vector and α carrying the corresponding loadings in the vector. The VAR approach is preferred here as it allows modelling both revenues and expenditures as potentially endogenous. We set the lag length k ¼5 based to the Akaike information criterion. Worth pointing out is that the lag length identified by the Akaike criterion more or less corresponds to the average duration of a government term.1 As the null of cointegration between the tax and the expenditure shares of GDP cannot be rejected at conventional levels of statistical significance, we proceed to estimate the corresponding error-correction model. This is done 1

We thank an anonymous referee for this astute observation.

C. Kollias et al. / Research in Economics 68 (2014) 208–213

211

Table 3 Estimates of alternative error-correction models for ΔREV in the UK by excluding the recent financial crisis (2007q3–2011q1). (i) Linear model 1

C ECTt  1 (1  D)*ECTt  1 D*ECTt  1 (1  D)*(ECTt  1) þ D*(ECTt  1) þ (1  D)*(ECTt  1)  D*(ECTt  1)  ΔREVt  1 ΔREVt  2 ΔREVt  3 ΔREVt  4 ΔREVt  5 ΔEXPt  1 ΔEXPt  2 ΔEXPt  3 ΔEXPt  4 ΔEXPt  5 D1973 Diagnostics Adj. R2 Q-Stat (16) p-value Q-Stat (16) squared residuals p-value ARCH(16) test p-value

(ii) Asymmetric model 2

(iii) Asymmetric model 3

Coefficient

t-Statistic

Prob.

Coefficient

t-Statistic

Prob.

Coefficient

t-Statistic

Prob.

0.00063  0.13737 – – – – – –  0.46931 0.02422 0.12418 0.13685 0.06144 0.20470 0.02638 0.03966  0.15300  0.12217  0.02613

0.99  3.65 – – – – – –  6.18 0.30 1.55 1.69 0.81 2.27 0.29 0.44  1.68  1.38  2.93

0.32 0.00*** – – – – – – 0.00*** 0.76 0.12 0.09* 0.42 0.02** 0.77 0.66 0.09* 0.17 0.00***

0.00049 –  0.16240  0.08600 – – – –  0.47811 0.01798 0.12069 0.13435 0.06166 0.21401 0.03783 0.04527  0.14820  0.12613  0.02640

0.78 –  3.61  1.36 – – – –  6.26 0.22 1.50 1.66 0.81 2.36 0.41 0.50  1.63  1.42  2.96

0.43 – 0.00*** 0.17 – – – – 0.00*** 0.82 0.13 0.10 0.42 0.02** 0.68 0.62 0.11 0.16 0.00***

0.001687 – – –  0.302453  0.108256  0.060966  0.083594  0.49097 0.006503 0.115395 0.132361 0.062912 0.212239 0.034284 0.033322  0.143506  0.120704  0.025918

1.62 – – –  3.69  1.11  0.91  0.74  6.45 0.08 1.41 1.62 0.83 2.36 0.38 0.37  1.58  1.37  2.93

0.11 – – – 0.00*** 0.27 0.36 0.46 0.00*** 0.94 0.16 0.11 0.41 0.02** 0.71 0.71 0.12 0.17 0.00***

22.6%

22.7% 0.71 0.14 0.26

23.9% 0.70 0.18 0.33

0.84 0.42 0.57

nn n

Denotes evidence against the null hypothesis at the 0.05 level. Denotes evidence against the null hypothesis at the 0.10 level.

in order to understand how quarter-to-quarter changes in taxes adjust the economy back toward a sustainable long-run equilibrium, conditional on government spending which resulted weakly exogenous in the investigation.2 We can interpret this error-correction model as the fiscal authorities' reaction function to budgetary disequilibria, in the sense of deviations of taxes and spending from their long-run sustainable equilibrium. Initially, the short-run adjustment of taxes within a linear error-correction model is estimated. Then different asymmetric error correction models are investigated. The first one is obtained by taking the deviations of the cointegrating vector ECTt  1 around its mean value, and partitioning them into two components one occurring during left governments and one during right governments (denoted by dnECT t  1 and ð1  dÞnECT t  1 , respectively). Then, the second model investigates whether adjustment to long run equilibrium differs under different colour governments during positive and negative components of the ECT (the relevant variables denoted by dnECT tþ 1 and ð1  dÞnECT tþ 1 , respectively dnECT t 1 and ð1  dÞnECT t 1 , respectively). Preliminary analysis using different unit root tests suggests that both the revenue and the expenditure series are nonstationary in levels. More specifically, the Augmented Dickey–Fuller and the Ng–Perron tests in Table 1 provide evidence that our variables of interest are all I(1). Therefore we proceed by testing for cointegration between revenues and expenditures. Looking at both the Trace statistic and Maximum-Eigenvalue statistic in the lower part of Table 1, we find evidence for one cointegrating relationship between our two variables. Table 2 presents the results from the estimation of the linear and asymmetric vector error correction models for Δ(REV). Column (i) indicates a significant speed of adjustment in the linear model. Results for the asymmetric error-correction model based on the political colour of the government are reported in column (ii) of Table 2. The results indicate that the speed of adjustment varies depending on whether a left – Labour – or a right – Conservative – party is in office. More specifically, our findings indicate that the tax adjustment to the long run relationship is faster and significant only under a Conservative government. Moreover, according to (iii) asymmetric model when taxes are higher than the equilibrium relationship Conservative governments are more responsive, reducing them faster compared to when they are lower to equilibrium. The point estimates suggest that when taxes are higher than equilibrium, they fall rapidly only under Conservative governments. On the other hand, when taxes are lower than equilibrium, they increase slowly. This result implies that Conservative governments are more flexible concerning tax decisions. Hence, our results point to downward flexibility of taxes under Conservative governments. To account for the possibility of any effects stemming from the oil crisis a dummy variable is included in our model. This dummy variable turned out to be statistically significant in all models, implying the existence of lower tax revenues during 2 Our investigation provided evidence against the significance of the coefficient of the error -correction term in the equation of the expenditures in first differences. Results are not presented for reasons of brevity, but are available upon request by the authors.

212

C. Kollias et al. / Research in Economics 68 (2014) 208–213

Fig. 1. Asymmetric adjustment and political colour.

that period. Finally, in the lower part of the table diagnostic tests provide evidence against autocorrelation and heteroskedasticity. The p-values of the Q statistic on the level and the squared residuals are larger than the 0.05. The same holds for the ARCH test for 16 lags. Furthermore, in order to examine whether the observed relationships were affected by the recent financial crisis, we re-estimated our models by limiting the sample to the pre-crisis period (1961q1– 2007q2). As a first and general observation, our main results still hold. Secondly, it is interesting to observe that the faster statistically significant adjustment towards equilibrium relationship under Conservative government is more evident in the asymmetric models 2 and 3. The weak adjustment towards long run relationship under Labour government becomes appreciably weaker (Table 3). Fig. 1a and b plots the asymmetric adjustments against the cointegrating vector, respectively. On the one hand, we notice from Fig. 1a where the asymmetric adjustments of model (ii) against the cointegrating vector is plotted that for large disequilibrium deviations, correction back to equilibrium becomes stronger once a Conservative government is in office. On the other, hand and in order to assess the differences amongst positive and negative disequilibrium deviations under different political parties, we plot the asymmetric adjustments of model (ii) against the cointegrating vector. Our findings support the proposition that when the Conservative party is in office, it shows a stronger preference for higher revenues' flexibility under positive disequilibrium deviations versus negative.

3. Concluding remarks The dynamics that govern the intertemporal nexus between the two budget aggregates – revenues and expenditures – have been the subject of extensive debate and research (Payne, 2003; Eslava, 2011; van der Ploeg, 2010). The acute fiscal imbalances currently faced by a number of European Union countries raise research questions regarding not only the generation of fiscal deficits but also of the correction of such imbalances whenever they arise and hence adherence to fiscal rules (Ioannou and Stracca, 2014; Bird and Mandilaras, 2013; Schalck, 2012; Andrikopoulos et al., 2004; Thornton, 2009; Debrun et al., 2009). Using the UK as a vehicle of empirical investigation, this paper examined whether the fiscal imbalances adjustment process differs under governments of different political orientation and ideological foundations. Analysis using quarterly data for the period 1961–2011 revealed that in the case of the UK, Conservative governments invariably tend to operate under a hard budget constraint, exhibiting a stronger fiscal vigilance vis-à-vis Labour administrations. The findings reported herein suggest that the latter are more inclined to operate a soft budget constraint and to be more lax when it comes to fiscal imbalances as the speed at which deficits are corrected reveals.

C. Kollias et al. / Research in Economics 68 (2014) 208–213

213

References Ahmed, S., Greene, K., 2000. Is the median voter a clear-cut winner?: Comparing the median voter theory and competing theories in explaining local government spending. Public Choice 105, 207–230. Andrikopoulos, A., Loizides, I., Prodromidis, K., 2004. Fiscal policy and political business cycles in the EU. Eur. J. Politic. Econ. 20 (1), 125–152. Apergis, N., Payne, J., Saunoris, J., 2012. Tax-spend nexus in Greece: are there asymmetries? J. Econ. Stud. 39, 327–336. Arestis, P., Cipollini, A., Fattouh, B., 2004. Threshold effects in the U.S. budget deficit. Econ. Inq. 42, 214–222. Baghestani, H., McNown, R., 1994. Do revenues or expenditures respond to budgetary disequilibria? South. Econ. J. 92, 311–322. Bajo-Rubio, O., Diaz-Roldan, C., Esteve, V., 2004. Searching for threshold effects in the evolution of budget deficits: an application to the Spanish case. Econ. Lett. 82, 239–243. Belke, A., Potrafke, N., 2012. Does government ideology matter in monetary policy? A panel data analysis for OECD countries. J. Int. Money Finance 31, 1126–1139. Bird, G., Mandilaras, A., 2013. Fiscal imbalances and output crises in Europe: will the fiscal compact help or hinder? J. Econ. Policy Reform 16 (1), 1–16. Brunila, A., Buti, M., Veld, J., 2003. Fiscal policy in Europe: how effective are automatic stabilisers. Empirica 30 (1), 1–24. Chang, T., Liu, W.R., Caudill, S.B., 2002. Tax-and-spend, spend-and-tax, or fiscal synchronization: new evidence for ten countries. Appl. Econ. 34, 1553–1561. Cipollini, A., 2001. Testing government intertemporal solvency: a smooth transition error correction model approach. Manch. Sch. 69, 643–655. Cipollini, A., Fattouh, B., Mouratidis, C., 2009. Fiscal readjustments in the United States: a nonlinear time series analysis. Econ. Inq. 47 (1), 34–54. de Haan, J., Sturm, J., 1994. Political and institutional determinants of fiscal policy in the European community. Public Choice 80, 157–172. Debrun, X., Hauner, D., Kumar, M.S., 2009. Tied to the mast? National fiscal rules in the European union. J. Econ. Surv. 23 (1), 44–81. Easaw, J., Garratt, D., 1999. UK government expenditure and electoral security in the 1980s: a non-linear analysis. Econ. Lett. 62, 287–292. Easaw, J., Garratt, D., 2000. Elections and UK government expenditure cycles in the 1980s: an empirical analysis. Appl. Econ. 32, 381–391. Easaw, J., Garratt, D., 2006. General elections and government expenditure cycles: theory and evidence from the UK. Eur. J. Political Econ. 22, 292–306. Eslava, M., 2011. The political economy of fiscal deficits: a survey. J. Economic Surv. 25, 645–673. Ewing, B.T., Payne, J.E., Thompson, M.A., Al-Zoubi, O.M., 2006. Government expenditures and revenues: evidence from asymmetric modeling. South. Econ. J. 73, 190–200. Garcia, S., Henin, P.Y., 1999. Balancing budget through tax increases or expenditure cuts: is it neutral? Econ. Model. 16, 591–612. Giavazzi, F., Jappelli, T., Pagano, M., 2000. Searching for non-linear effects in fiscal policy: evidence from industrial and developing countries. Eur. Econ. Rev. 44, 1259–1289. Goff, B., Tollison, R., 2003. The monopsony power of the median voter. Public Finance Rev. 31, 180–188. Holcombe, R., Caudill, S., 1985. Tax shares and government spending in a median voter model. Public Choice 46, 197–206. Holmes, M., 2011. Threshold cointegration and the short-run dynamics of twin deficit behaviour. Res. Econ. 65, 271–277. Ioannou, D., Stracca, L., 2014. Have the euro area and EU governance worked? Just the facts. Eur. J. Politic. Econ. 34, 1–17. Johansen, S., 1988. Statistical analysis of cointegration vectors. J. Econ. Dyn. Control 12, 231–254. Kollias, C., Paleologou, S.M., 2006. Fiscal policy in the European Union: tax and spend, spend and tax, fiscal synchronization or institutional separation? J. Econ. Stud. 33, 108–120. Kollias, C., Makrydakis, S., 2000. Tax and spend or spend and tax? Empirical evidence from Greece, Spain, Portugal and Ireland. Appl. Econ. 32, 533–546. Koren, S., Stiassny, A., 1998. Tax and spend, or spend and tax? An international study. J. Policy Model. 20, 163–191. Legrenzi, G., Milas, C., 2012. Nonlinearities and the sustainability of the government's intertemporal budget constraint. Econ. Inq. 50, 988–999. Milanovic, B., 2000. The median-voter hypothesis, income inequality, and income redistribution: an empirical test with the required data. Eur. J. Politic. Econ. 16, 367–410. Owoye, O., Onafowora, O., 2011. The relationship between tax revenues and government expenditures in European Union and n on-European Union OECD countries. Public Finance Rev. 39, 429–461. Paleologou, S.M., 2013. Asymmetries in the revenue–expenditure nexus: a tale of three countries. Econ. Model. 30, 52–60. Payne, J.E., 2003. A survey of the international empirical evidence on the tax-spend debate. Public Finance Rev. 31, 302–324. Payne, J., Mohammadi, H., 2006. Are adjustments in the U.S. budget deficit asymmetric? Another look at sustainability. Atl. Econ. J. 34, 15–22. Ramos, X., Roca-Sagales, O., 2008. Long-term effects of fiscal policy on the size and distribution of the pie in the UK. Fisc. Stud. 29 (3), 387–411. Saunoris, J.W., Payne, J.E., 2010. Tax more or spend less? Asymmetries in the UK revenue-expenditure nexus. J. Policy Model. 32, 478–487. Schalck, C., 2012. Investigating heterogeneity in European fiscal behaviours. Res. Econ. 66, 383–390. Thornton, J., 2009. Do fiscal responsibility laws matter: evidence from emerging economies suggests not. J. Econ. Policy Reform 12 (2), 127–132. van der Ploeg, F., 2010. Political economy of prudent budgetary policy. Int. Tax Public Finance 17 (3), 295–314. Vamvoukas, G., 2012. Panel data modelling and the tax-spend controversy in the euro zone. Appl. Econ. 44, 4073–4085. Zapf, M., Payne, J.E., 2009. Asymmetric modelling of the revenue-expenditure nexus: evidence from aggregate state and local government in the US. Appl. Econ. Lett. 16, 871–876.