144 However, we argue that the data employed here are far better suited for an analysis of partisan interests than just case studies or revenue data and that our results are therefore more reliable. 3.7 The role of divided government: partisan governments in a bicameral legislature 3.7.1 General approach and related literature The German federal government is not unrestricted in tax policy.259 Approval of tax reforms in the second chamber, the Bundesrat, is necessary in all major taxes with exception of mineral oil taxes and the tobacco taxes. Therefore, the Bundesrat has a far-reaching ability to veto tax bills. As the federal government and the state governments are dominated by the two major parties (the CDU/CSU and the SPD), the majority constellations in the two chambers can either lead to divided (different partisan majorities in both chambers) or undivided government (majority of the parties of the federal government in the Bundesrat).260 If party majorities are decisive for policy-making,261 the importance of the second chamber as a veto-player would therefore depend on whether we are in a situation of divided or undivided government. If the government parties do not have a majority in the second chamber, the Bundesrat, the opposition parties could use the veto-player position of the Bundesrat to block legislation. This has led to a “gridlock hypothesis” in case of divided government, which is very prominent in the literature on German federalism.262 In the empirical literature based on legislative statistics, which uses indicators as e.g. the number of effectively vetoed legislative proposals or the prolongation of legislation in case of divided government, there is so far only limited evidence for the gridlock thesis.263 259 See the discussion in part V.3.1. 260 See for a general discussion of bicameralism and divided government Persson/Roland/Tabellini (1997), Riker (1992), Rogers (1998), Tsebelis (1995), Tsebelis/Money (1997), Alesina/Rosenthal (1996), Brennan/Hamlin (1992) and Diermeir/Myerson (1999). For the differences in between effects of bicameralism and of divided government see Elgie (2001). 261 Please note that the importance of party majorities is not directly linked to partisan interests discussed under the partisan politics approach. Party majorities can matter even if parties do not adhere to distinctive ideological goals. For example, policy proposals might be blocked not based on their content, but because they are brought in by the opposition. 262 See e.g. Scharpf (1985) or Braeuninger/Koenig (1999). See for the gridlock thesis in tax policy in Germany especially Ganghof (1999) and Wagschal (2005), but also Renzsch (2000) and Zohlnhoefer (1999). With respect to international comparisons of the effects of divided government/bicameralism refer to Hallerberg/Basinger (1999). 263 See Burkhart/Manow (2006), Bauer (1998), Lehmbruch (2000) and Renzsch (2000). In a recent contribution Burkhart and Manow (2006) employ an “auto limitation process” (based on Vanberg (1998)): They argue that federal governments refrain from confrontation with the 145 3.7.2 Hypotheses We can directly derive three important hypotheses on German tax policy depending on divided and undivided government. In case of divided government tax policy changes via tax reforms should be less frequent and smaller than in case of undivided government (divided government hypothesis I – “gridlock hypothesis”). This should be the case especially with respect to joint taxes where approval of the Bundesrat is necessary (divided government hypothesis II). With respect to the federal taxes (especially mineral oil and tobacco taxes) we should see larger changes in case of divided government because these taxes are the only ones which can be effectively altered without approval of the Bundesrat (divided government hypotheses III). Finally, we can derive a fourth hypotheses based on partisan theory (see part V.3.6). Partisan policy preferences of governments should come out especially clearly in times of undivided government. Based on the partisan preferences of CDU/CSU- and SPD-led governments (derived in part V.3.6.2) we would expect that undivided SPD governments increase wage and income taxes and reduce indirect taxes (like the VAT or the tobacco and the mineral oil taxes) while the opposite should hold for undivided CDU/CSU governments (divided government hypotheses IV). 3.7.3 Empirical analysis Within the period covered by our data-set264 we observe 13 years of undivided government265 and 26 years of divided government. CDU/CSU governments disposed of a majority in the second chamber in 10 years (1965-1966, 1983-1990), while undivided SPD-led government did not occur. The grand coalition government disposed of a majority in the second chamber in between 1966-69.266 Figure 66 shows the political development, tax revenue developments and fiscal effects of tax opposition in case of divided government when the opposition has a clear majority in the second chamber. Only if the majorities are uncertain (neither government nor opposition majority in the second chamber), controversial legislative proposals are pushed forward. For this hypothesis they find evidence based on new legislative statistics (see Burkhart/Manow (2006), p. 18). However, this approach is only justified in studies which have to rely on data on the legislative process itself (as e.g. the number of bills which are vetoed) as a direct indicator for policy outcomes is missing. In our case we dispose of such an indicator (fiscal effects of tax reforms) and can therefore concentrate on an analysis of the direct influence of divided government. 264 1965 to 2003 if we integrate only those years covered completely. 265 We apply the official definition of a majority of votes for the “R-states” in the Bundesrat. Rstates are those states in which the state government consists only of the parties governing on the federal level. 266 See as well the discussion of political development in Germany in part V.3.1. 146 reforms. At first sight it seems that changes in the revenue structure were rare in times of undidivided governments and that fiscal effects of tax reforms were lower than on average. U nd iv id ed go ve rn m en t U nd iv id ed go ve rn m en t 0% 5% 10% 15% 20% 25% 19 50 19 53 19 56 19 59 19 62 19 65 19 68 19 71 19 74 19 77 19 80 19 83 19 86 19 89 19 92 19 95 19 98 20 01 20 04 TOTAL Wage and income tax VAT Corporate profit and local trade taxes Mineral oil taxes Tobacco taxes Property taxes Tax revenues/ GDP -4% -2% 0% 2% 4% 19 65 19 68 19 71 19 74 19 77 19 80 19 83 19 86 19 89 19 92 19 95 19 98 20 01 INCREASES/GDP REDUCTIONS/GDP Fiscal effects of tax reforms/GDP TAX REFORMS UNDER DIVIDED AND UNDIVIDED GOVERNMENT DATA OF ADOPTED REFORMS - 1965-2003 Own calculations based on: Federal Ministry of Finance (2004), Federal Statistical Office(2007).Data on adopted reforms. Figure 66: Tax reforms and tax revenue developments under divided and undivided government What do we find if we analyze our four major hypotheses on the effects of divided government on tax reform in detail? We start with the overall effects of reforms in the upper part of Figure 67. We see that in case of an undivided government (the third group of columns in Figure 67), the fiscal effects of increases and reductions were smaller than on average. This indicates that the total change of the tax system (which can be measured by the sum of the value of increases and reductions) was smaller in case of undivided government. Furthermore, we find that especially tax increases were smaller in case of undivided government (on average 0.2% per GDP) than on average (on average 0.4% of GDP). This indicates that undivided governments refrained from tax increases. Now one could argue that our results are strongly affected by reunification. As tax reform activity was pushed upwards during the years after reunification and a divided government was in place in these years, our results might be dis- 147 torted.267 Therefore, we have excluded the years after reunification and have calculated fiscal effects only from 1965 to 1990 (see the lower part of Figure 67). Although the difference in between average tax reforms and tax reforms under undivided government becomes more moderate, the general observation continues to be valid: Even in the time period from 1965 to 1990 divided governments were less active in tax policy and especially tax burden increases were lower. TAX REFORMS UNDER DIVIDED AND UNDIVIDED GOVERNMENT DATA OF ADOPTED REFORMS - 1965-2004 -0,8% -0,6% -0,4% -0,2% 0,0% 0,2% 0,4% 0,6% 0,8% AV ER AG E DI VI DE D UN DI VI DE D LE FT RI GH T GR AN DC OA L INCREASES REDUCTIONS BALANCE TAX REFORMS: 1965-1990 -0,8% -0,6% -0,4% -0,2% 0,0% 0,2% 0,4% 0,6% 0,8% AV ER AG E DI VI DE D UN DI VI DE D LE FT RI GH T GR AN D CO AL IT IO N INCREASES REDUCTIONS BALANCE TAX REFORMS: 1965-2004 Own calculations based on: Federal Ministry of Finance (2004), Federal Statistical Office(2007). Fiscal effects of reforms/GDP Fiscal effects of reforms/GDP Data on adopted reforms. Figure 67: Influence of divided government on tax reforms Based on our second hypothesis we expect tax policy changes to be rarer and smaller in times of divided government especially with respect to joint taxes (where approval of the Bundesrat is necessary). The two most important joint taxes are the wage and income tax and the VAT. We have to keep in mind that VAT became a joint tax only in 1969 and was assigned to the federal level before. Throughout the period in our data-set when VAT was a federal tax (1964-1969) we were in a situation of undivided government. Our results are presented in Figure 68. We find that the fiscal effects of wage and income tax and VAT reforms were smaller in times of undivided governments. With respect to the VAT the effects would have been even stronger if we would have 267 See discussion of the data-set in part III.2 148 excluded 1965 to 1969 (when the VAT/sales tax was a federal tax) as the most important changes took place under the grand coalition (1966-1969). Therefore, we have to turn down our second hypothesis as well. There was less overall change in the both main joint taxes in case of undivided government. And again we see that undivided governments refrained especially from tax increases in the wage and income tax as well as in the VAT. -0,60% -0,50% -0,40% -0,30% -0,20% -0,10% 0,00% 0,10% 0,20% 0,30% AV ER AG E DI VI DE D UN DI VI DE D LE FT RI GH T RG T+ UN DI V GR AN D CO AL IT IO N INCREASES REDUCTIONS BALANCE WAGE AND INCOME TAX REFORMS -0,15% -0,10% -0,05% 0,00% 0,05% 0,10% 0,15% AV ER AG E DI VI DE D UN DI VI DE D LE FT RI GH T RG T+ UN DI V GR AN D CO AL IT IO N INCREASES REDUCTIONS BALANCE REFORMS (VAT) TAX REFORMS UNDER DIVIDED AND UNDIVIDED GOVERNMENT DATA OF ADOPTED REFORMS - 1965-2003 Fiscal effect of reforms/GDP Fiscal effect of reforms/GDP Own calculations based on: Federal Ministry of Finance (2004), Federal Statistical Office(2007).Data on adopted reforms. Figure 68: Influence of divided government on tax reforms in consumption taxes Our third hypothesis is that we expect to see larger changes in case of divided government in the federal taxes (especially mineral oil and tobacco taxes) as these taxes have been the only ones which could be effectively altered without approval of the second chamber. This time the data seem to speak in favor of our hypothesis. Average increases in tobacco and mineral oil taxes were higher in times of divided government (see Figure 69). However, we have to keep in mind that this is as well in line with our general finding of lower activity in tax reforms under undivided government and with our finding that undivided governments refrained from tax increases. 149 0,00% 0,01% 0,02% 0,03% 0,04% 0,05% 0,06% 0,07% 0,08% 0,09% AVERAGE DIVIDED UNDIVIDED LEFT RIGHT RGT+UNDIV GRAND COALITION INCREASES REDUCTIONS BALANCE -0,02% -0,01% -0,01% 0,00% 0,01% 0,01% 0,02% 0,02% 0,03% 0,03% 0,04% AVERAGE DIVIDED UNDIVIDED LEFT RIGHT RGT+UNDIV GRAND COALITION INCREASES REDUCTIONS BALANCE REFORMS (TOBACCO) REFORMS (MINERAL OIL) TAX REFORMS UNDER DIVIDED AND UNDIVIDED GOVERNMENT DATA OF ADOPTED REFORMS - 1965-2003 Fiscal effect of reforms/GDP Fiscal effect of reforms/GDP Own calculations based on: Federal Ministry of Finance (2004), Federal Statistical Office(2007).Data on adopted reforms. Figure 69: Influence of divided government on tax reforms in mineral oil and tobacco taxes Finally, we expected that partisan preferences come out most clearly in times of undivided government (hypothesis IV). In Figures 67 to 69 we see that undivided right governments reduced the overall tax burden stronger than on average and abstained from increases in the wage and income tax. This speaks in favor of our first two partisan hypotheses (see the discussion in part V.3.6.2). However, increases in the VAT and the mineral oil and tobacco taxes were as well smaller than on average. Our observations raise the question, whether the effects under right undivided governments were caused by the partisan orientation of the government or by the institutional factor of undivided government. As undivided left government did not occur in our data-set, this is hard to decide. However, there is some further evidence which might be important to decide on the causes of the observed reform effects. If partisan preferences would have been decisive, we would expect that tax increases would be smaller in wage and income taxes but larger in VAT under undivided than under divided right-wing governments. But wage and income tax as well as VAT increases are lower under undivided than under divided right governments. We see this as further evidence pointing at the importance of undivided government (and not partisan preferences) for tax reform patterns. 150 3.7.4 Summary of results The thesis of legislative gridlock in case of divided government is very prominent in the literature on the German bicameral system – especially with respect to tax legislation where approval of the second chamber is mandatory for most of the tax laws. However, we found strong evidence against the thesis of legislative gridlock. Overall change of the tax system via tax reforms was substantially smaller in times of undivided government. This results especially from smaller average tax increases under undivided government. Furthermore, we saw less change in the wage and income tax and the VAT (where approval of the second chamber is necessary) as well as with respect to the mineral oil and the tobacco tax (where no approval is necessary) in times of undivided government. Finally, we showed that there was no evidence for a stronger influence of partisan preferences in case of undivided governments. Taken together we conclude that not legislative gridlock under divided government has been important for tax policy, but that undivided governments refrained especially from tax increases. 268 Governments did not use their majority in the second chamber to expand government revenues by strong tax increases, but quite the contrary is shown by the data: the extent of tax increases is smaller for undivided government. Apparently, governments abstain from large tax increases in times they have the political power to implement them. An explanation for this behavior is that governing parties can be blamed exclusively by the voters for tax increases in times of undivided government while an agreement of the opposition in the second chamber under divided government leads to a shared responsibility for tax burden increases. This limits the potential losses in electoral support for the government vis-àvis the opposition in the next elections. 268 This has an important implication for an alternative explanation of our findings. The theory of log-rolling (see e.g. Tullock (1959)) would argue that an additional chamber, which has to agree to policy changes, increases the size of the necessary majority coalition. As all members of the coalition try to profit from the agreement, it has to be expected that overall net tax burden increases (increases minus reductions) are higher in times of divided government (I owe this point to Charles B. Blankart). Empirically, however, our data shows higher increases in times of divided governments but a similar net effect (increases minus reductions) in case of divided and undivided government. Therefore, the evidence speaks more in favour of the argument of undivided governments fearing to be punished for tax increases than for the logrolling argument. 151 3.8 Combined testing of different approaches In the previous parts we reviewed several polit-economic approaches separately. However, some of them can be combined (as especially the partisan politics approach, the opportunistic approach269 and the approach of divided government) and tested simultaneously. What do we find if we test these approaches statistically based on our data-set? We have tested variations of the following equation using OLS: RILE EL tt4t3t20 ????? ++++=? tT With: 0? , x? Constant, coefficients ?Tt Fiscal effects of tax reforms (net fiscal effects, fiscal effects of increases, fiscal effects of reductions) ELt Election year variable (dummy) LEt Left-wing government (dummy) RIt Right-wing government (dummy) EL? Stochastic error term To make tax reforms comparable over different years we have divided the fiscal effects by nominal GDP. All employed polit-economic variables are dummies. Table 10 reports our results for the dependent variables “net fiscal effects of tax reforms”, “fiscal effects of tax increases” and “fiscal effects of tax reductions”.270 We do not find strong correlations. Most of the coefficients are not significant and we never reach adjusted R2 scores higher than 0.27.271 In our results we find no indication for a partisan hypothesis. On the other hand our analysis shows that the opportunistic approach, which includes the variable for undivided government and the election dummy, performs best with respect to the fiscal effects of tax burden increases. The coefficients for both variables are highly significant at the 1% level and the adjusted R2 score is with 0.27 high compared to other specifications. As expected by the theory of opportunistic government behavior, elections had a negative effect on the extent of tax increases before elections. Governments postponed tax increases after the elections to improve their re-election probabilities. Furthermore, our observation that undivided governments refrain from 269 Frey and Schneider (1978) have early pointed out that partisan and opportunistic approaches can be combined. Parties have an ideology to which they adhere in general. But before elections they might nonetheless behave opportunistically. 270 The Durbin-Watson test statistics for our regression (reported in table 3) do not support the hypothesis of serial autocorrelation in the error terms. 271 Partly this is affected by the fact that we are focusing here on data on fiscal effects of adopted reforms as we would expect effects of partisan constellation and divided government on adopted reforms. However, with respect to opportunistic behavior data on implemented reforms would be more suitable. See the discussion of opportunistic behavior in part V.3.5.