CESifo Economic Studies Advance Access originally published online on June 18, 2007
CESifo Economic Studies 2007 53(2):294-328; doi:10.1093/cesifo/ifm011
ACE versus CBIT: Which is Better for Investment and Welfare?
The authors are affiliated with Center for Economic Studies, University of Munich and CESifo, CES, Schackstr.4, 80539 Munich, Germany. e-mail: Radulescu{at}lmu.de; Stimmelmayr{at}lmu.de
 |
Abstract |
|---|
 Top Abstract 1 Introduction2 The ACE versus...3 The model4 Model calibration5 Simulation results6 Sensitivity analysis7 ConclusionAppendixReferences |
|---|
This article analyses the switch to an Allowance for Corporate Equity (ACE) or to a Comprehensive Business Income Tax (CBIT) type of tax system starting from the present German tax system. We show that in case an ACE type of reform is financed by an increase in the VAT and not in the profit tax, it might be preferred to a CBIT even in the context of an open economy. Moreover, the required exogenous increase in the profit tax rate cannot ensure revenue neutrality on its own due to the negative general equilibrium effects it triggers on the whole economy. For a CBIT, the exogenous reduction in the tax rates on corporate and non-corporate profits leads to better results than when we allow for an endogenous change in the VAT. The best results arise when the CBIT is accompanied by a provision for immediate write-off and a lower profit tax or when the ACE with no additional capital gains taxation on household side is financed by an increase in the VAT. (JEL-Classification: C68, D58, D92, E62, H25.)
Key Words: Capital income taxation • computable general equilibrium modelling • welfare analysis
|
1 Introduction |
|---|
|
TopAbstract1 Introduction2 The ACE versus...3 The model4 Model calibration5 Simulation results6 Sensitivity analysis7 ConclusionAppendixReferences |
|---|
In most OECD countries, the tax law provides for a deduction of debt interest when computing the profit tax base while the opportunity cost of equity capital is not accounted for. Therefore, debt financing exhibits a significant advantage compared to equity financing via retained earnings or new share issues. In order to equalise the opportunity cost of debt and equity, tax experts have designed two polar reform proposals, namely the Comprehensive Business Income Tax (CBIT) and the Allowance for Corporate Equity (ACE).1 The CBIT was developed by the US Treasury Department at the beginning of the 1990s (US Department of the Treasury, 1992) whereas the ACE was elaborated by the IFS Capital Taxes Group (1991) at the same time. The idea of an imputed interest on equity was, however, originally advanced by Boadway and Bruce in 1984.
Particularly with regard to the present discussions in various countries, which aim at reforming the tax systems in one way or the other, this topic seems to be an up-to-date issue. For instance, the German Council of Economic Advisors (GCEA 2005, 2007) suggested the introduction of a Dual Income Tax (DIT) combined with an ACE for Germany, and for Switzerland, Keuschnigg and Dietz (2006) designed a growth oriented DIT which also features elements of an ACE, as well. Additionally, the German ruling coalition's most recent reform proposal (BMF 2006) advances the idea of a lower corporate tax of only 30 percent and the partial abolition of debt interest deductibility thus moving in the direction of a CBIT. A similar idea was also put forward in the US by the President's Advisory Panel on Tax Reform (2006), which proposed for the taxation of large businesses a profit tax of 30 percent and the abolishment of debt interest deductibility.2
In this article, we develop a computable general equilibrium (CGE) model, IfoMod, to analyse and quantify the efficiency and welfare implications of the two antipodal reform proposals. To our knowledge, a comparative quantitative assessment of this kind is still lacking for the two tax reform proposals. Our paper tries to fill this gap by first comparing the neutrality properties and second by computing the efficiency and welfare effects of each approach.3 The applied CGE model resembles a two country model where the home country consists of a two sector economy and an infinitively lived agent on the household side. Based on the premise that each tax reform should be implemented on a revenue neutral basis, the introduction of the ACE requires either an increase in the corporate income tax rate, an increase in another tax such as the VAT or a reduction in government transfers, since it envisages a narrower tax base compared to the currently existing tax systems. In contrast, the corporate tax rate can eventually be reduced under the CBIT due to the broader tax base this proposal brings about.
Our simulations results show that the ACE/ACNE4 proposal can hardly be financed by an increase in the tax on corporate and non-corporate profits, since a rise in the profit tax rate has considerable negative economy-wide repercussions within the dynamic general equilibrium framework which leads to capital decumulation and a shrinking tax base for all other taxes as well. Therefore, the ACE/ACNE proposal needs to be accompanied either by reduction in transfers or increase in any other tax rate. Under the CBIT, the decline in the corporate tax rate—which may accompany such a reform—has a significant positive effect on economic growth and welfare.
Concerning the different simulation scenarios performed, the best results with regard to the change in macroeconomic variables and welfare are achieved under the CBIT combined with a provision for an immediate write-off. Under this scenario, the large increase in the capital stock and labour demand boosts long-run GDP by around 10 percent. Second, an ACE/ACNE with no additional capital gains taxation on the household level also produces positive welfare results. In detail, the long run capital stock increases by 20.5 percent leading to a significant upward swing in GDP by 9.1 percent. To ensure a comparable treatment of different types of capital income on household level we also consider scenarios that consider capital gains tax as well. These simulations, however, as well as the CBIT with no special depreciation allowances do not perform well and even lead to welfare losses.
A last but interesting comparison is drawn between the ACE and the current German system of capital income taxation.
If we allow for a proportional reduction of all capital income tax rates by 38 percent such that the costs of implementing this reform or the ACE/ACNE are equivalent in present value terms, the former reform leads to slightly better results in terms of welfare. While welfare rises by 0.78 percent in terms of GDP under the former reform, it increases by just 0.73 percent in case the ACE/ACNE is implemented. Thus, a simple cut in tax rates seems to be more efficient than an overhaul of the tax system as implied by the ACE.
The remaining part of the article is structured as follows: The next section presents a short comparison of the two proposals. Section 3 highlights the main features of the CGE model while Section 4 presents the simulation results as well as detailed economic interpretations. In Section 5, we perform a sensitivity analysis, and the last section concludes.
|
2 The ACE versus the CBIT |
|---|
|
TopAbstract1 Introduction2 The ACE versus...3 The model4 Model calibration5 Simulation results6 Sensitivity analysis7 ConclusionAppendixReferences |
|---|
Under the CBIT neither the incurred interest on debt nor any imputed return on equity may be deducted against the profit tax base. Even though this reform proposal induces a higher cost of capital, it might be advisable for a government to adopt it, if it can apply a lower profit tax rate due to the broader tax base (Cnossen 2000). Such an outcome might be especially applicable for a small open economy which is characterised by high capital mobility and which has to survive in the context of international tax competition.5
Under the assumption that the two reform proposals (and especially the ACE) can be financed by a change in any other tax rate, such as the VAT for instance, as well, it is however, no longer clear-cut that the CBIT is the preferred alternative if a country wants to attract investors both by offering low statutory tax rates and a narrow tax base for corporate profits.
As opposed to the CBIT, the ACE provides, in addition to the deduction of incurred debt interest, a deduction of an imputed return on equity capital against the profit tax base (Bond 2000; Cnossen 2000; Devereux and Freeman 1991).6 Accordingly, the introduction of an ACE implies a significant reduction in the user cost of capital for any investment project relying on equity capital, compared to the present tax systems enforced in all OECD countries.7 In essence, an ACE will exempt the cost of raising finance at the company level from taxation, and will only subject the profits exceeding a ‘normal rate of return’ to taxation (IFS Capital Taxes Group 1991).8
A further advantage of such a tax system is that any schedule of depreciation allowances, i.e. providing for an immediate write-off or allowing for geometric digressive depreciation over the life of an asset, does not change the present value of tax payments (Devereux and Freeman 1991; Bruckner, Gassner and Riener-Micheler 2000). Allowing for accelerated depreciation, for instance, reduces, on the one hand, the tax base of the profit tax in the current period. On the other hand, it also reduces the accounting stock of capital and thus the base which is multiplied with the protective interest to compute the ACE in the next period (Keen and King 2002). Therefore, such a reform increases the efficiency of a tax system.9
Nevertheless, the ACE has its disadvantages too. The main critique raised with regard to the ACE addresses the fact that the narrowing of the tax base has to be accompanied by a higher tax rate to achieve a certain tax revenue (Isaac 1997). Such an outcome is less desirable in a world of high capital mobility where a higher statutory tax rate has a negative signalling effect for multinational firms. In this context, the question might arise, as to why the statutory tax rate is not simply reduced in order to provide a clear signal to international investors. The sole reduction in the statutory tax rate, however, does not address the issue of financial neutrality, since debt finance would continue to be at an advantage compared to equity finance.
Moreover, since the ACE is basically just a tax on economic rents, firms that earn just or less than the minimum required rate of return pay little or even no tax, while the most profitable enterprises will face the highest tax bills.10
|
3 The model |
|---|
|
TopAbstract1 Introduction2 The ACE versus...3 The model4 Model calibration5 Simulation results6 Sensitivity analysis7 ConclusionAppendixReferences |
|---|
The following section introduces a detailed description of the dynamic CGE model IfoMod which is applied to assess the quantitative impact of introducing the ACE/ACNE or the CBIT reform proposal. IfoMod is a two-country model and thus incorporates international portfolio holdings with regard to business and government bonds.11 We follow the empirical evidence of the home bias and restrict the cross ownership of domestic equity capital in the sense that the shares of domestic firms are only owned by domestic residents. Moreover, assets are assumed to be imperfect substitutes yielding different net of tax rates of return. Since the residence principle is adopted, gross domestic and foreign interest rates equalise. Accordingly, international portfolio capital investments are affected by the different rates of return prevailing in the domestic and foreign economies, which in turn depend on the prevailing tax rates.12
3.1 Production
Optimal investment behaviour of both corporate (C) and non-corporate (N) firms is derived from an intertemporal investment model with convex adjustment costs. The household sector is modelled in the spirit of the traditional Ramsey model of an infinitely lived agent, since we mainly focus on the efficiency and welfare effects of the tax reform and not on distributional issues. Beside the government, the Rest of the World (RoW) is the model's fourth building block which completes the general equilibrium framework.13
We start with a basic neoclassical, linear homogenous production technology, Yf = F(Kf,Lf), with capital, Kf and labour, Lf, as input factors.14 The price of the output good, Y f, is normalised to unity and additionally each firm incurs adjustment costs of Jf(If,Kf) representing declining marginal returns of capital formation.15
Both firm types hire labour and accumulate capital to maximise their value. Thereby, corporate profits are subject to double taxation, first, by the profit tax, 
P,C, which is levied on the firm level and thereafter on the personal level by the dividend tax, D, in case of distributed profits16 or by the capital gains tax, G,f, in case of profit retention. Profits of non-corporate firms are taxed only once, namely at the personal level at the income tax rate, P,N, of the firm owner.
Capital accumulates over time whenever gross investment, 
, exceeds the depreciation of the existing capital stock, 
:17
|
| (1) |
Concerning debt policy, we assume that interest payments on debt include an additional premium m(bf), e.g. agency costs of debt, which are a function of the debt capital ratio, bf = Bf/Kf, of the firm. These agency costs are increasing in b f, reflecting the larger risk of bankruptcy prevailing if the debt capital ratio of a firm rises and therewith the non-tax cost of default.18 Debt accumulates according to:
|
| (2) |
, is given by the existing stock of debt, 
, and new debt, 
incurred. Net of tax profits, 
, are defined as output 
less adjustment costs 
, wage payments 
, depreciation , interest and agency costs of debt 
and the tax liability of the firm 
:
|
| (3) |
P,f, has to be interpreted as a source tax on corporate profits. The tax parameters z1 and z2 denote whether the incurred interest on debt and equity capital is tax deductible or not. In case both parameters are set to zero, z1 = z2 =0, neither the interest on debt nor the imputed return on equity, denoted by iE, are tax deductible as proposed under the CBIT. As opposed to that, if we set z1 = z2 =1 , we model an ACE/ACNE since both the interest on debt and the imputed return on equity are tax deductible. Under the present German tax rules, z1 =1 and z2 =0 holds, implying that only interest payments on debt are tax deductible. The other tax parameter, z3 represents the tax allowances for net investments INf.19
Net investments of a firm,20 
, can either be financed via a reduction in distributions/dividends and thus out of retained earnings 
, via issuing new equity 
, or externally via new debt, . Accordingly the cash flow identity states:21
|
| (4) |
Finally, the net of tax return on corporate and non-corporate equity is given by net of tax dividend payouts and net of tax capital gains:
|
| (5) |
3.2 Financial behavior
From the firm's optimality and envelope conditions (Appendix A3) we derive the following equations describing the firms' optimal financial behaviour:22
|
| (6) |
A similar neutrality property with respect to the source of finance is also achieved in case the policy reform follows an ACE/ACNE. Under this scenario both the imputed rate of return and debt interest are tax deductible:
|
| (7) |
P,fiE and the possibility to deduct debt interest incorporates an advantage of the size P,fiBH. However, since a larger indebtedness increases the debt-asset ratio, b f, additional agency costs of 
occur and diminish the advantage of debt finance. The preference for a particular source of finance under the ACE/ACNE also depends on the magnitude of the imputed rate of return, iE. Neutrality is achieved only insofar as iE = iBH, assuring that the imputed return equals the interest rate paid on debt [see Equation (7)]. The higher the chosen value for iE, the lower is the cost of equity and the larger is the incentive to draw on retentions to finance investments vis-à-vis new debt. As a consequence the optimal debt-asset ratio will shrink.
To summarise, the introduction of the CBIT increases the cost of debt finance while the cost of equity remains unchanged. Therefore, the preference for debt finance, which still prevails in nearly all OECD countries, would be reduced and should induce a significant reduction in the firms's debt asset ratios. Under an ACE/ACNE, the possibility to deduct an imputed return from the tax base lowers the cost of equity while the cost of debt stays unchanged. Accordingly, the preference for equity financing should increase under the ACE/ACNE and thus bring about a decline in the debt asset ratio as well.
To evaluate the effects of a marginal change in the tax rates on the financial decision of a firm, we analyse the change in the cost of equity stemming from a marginal change in the tax rate under consideration. Therefore, we compute the percentage change in the cost of equity analogous to:23
|
|
to avoid division by zero. Since the return rVf is given exogenously, the effects of a marginal change in the tax rates on the cost of equity is given by:
|
| (8) |
According to Equation (8), under the CBIT the cost of equity is only affected by the capital gains tax, such that a one-percentage point increase in the capital gains tax rate also leads to a one-percentage point increase in the cost of equity capital.
Under the ACE/ACNE, besides the capital gains tax, the profit tax and the level of the imputed return can also induce a preference for either debt or equity finance. Thus, both an increase in the profit tax and a higher imputed return on equity24 reduce the cost of equity capital and stimulates a preference for equity finance, implying a reduction in the debt asset ratio:
|
| (9) |
The overall effect on the financial behaviour resulting from an increase in the profit tax rate is, however, less clear cut under the ACE/ACNE. On the one hand, the increase in the profit tax rate reduces the cost of equity as already discussed. On the other hand, it also reduces the cost of debt since debt interest is also tax deductible. Therefore, the dominating effect will depend on the relative difference between the imputed return and the interest rate on firm debt. If both rates are equal, the two sources of finance are affected in an identical way. If the imputed return is, however, higher (lower) than the return on debt, there will be a preference for equity (debt) finance.
3.3 Investment behavior
From the optimality and envelope conditions derived in Appendix A3, we also derive the cost of capital formulae for corporate and non-corporate firms by substituting in the firm-specific tax parameters:25
|
| (10) |
. In a world with taxation, the cost of capital denotes the minimum required pre tax rate of return an investment has to earn to be still profitable after taxation. The resulting cost of capital formula varies according to the different views on dividend taxation assumed.26 The presented model incorporates the "Old View" as well as the "New View" on dividend taxation since both new share issues and retained earnings are available as a source of finance.27
CBIT
Under the CBIT the tax parameters z1 and z2 are zero such that neither the interest on debt nor an imputed return on equity is tax deductible. Consequently, the cost of capital for corporate and non-corporate firms, respectively, becomes:28
|
| (11) |
|
| (12) |
P,C, and the personal income tax, P,N, respectively, has a positive impact on these firms' investment behaviour.29 The economic implication of a decrease in the corporate tax rate is thus straight-forward. If the corporate tax rate decreases, returns stemming from real investments are more heavily taxed compared to those from a financial investment that is not subject to the corporate tax rate. Hence, the cost of capital declines resulting in more real investments. The size of this effect will be larger for firms endowed with much equity and smaller for highly indebted firms.
ACE
In the case of an ACE, the tax parameters z1 and z2 are equal to one in order to assure that both debt interest as well as the imputed return on equity are tax deductible. The explicit form of the cost of capital under the ACE/ACNE states:30
|
| (13) |
P,f/
P,f(1-bf)iE if an ACE/ACNE is considered. Differentiating the cost of capital formulae given in (13) with respect to the corporate and personal income tax rate we get the same results as under the CBIT proposal:31
|
| (14) |
|
| (15) |
To summarise: as expected, the overall cost of capital increases under the CBIT, since the incurred interest on debt is no longer tax deductible. Hence the cost of debt finance is raised and therewith the overall cost of capital. Due to the higher cost of capital, a lower number of profitable investments that offer the minimum required rate of return are available. As a consequence, fewer investment is carried out, implying that each firm operates at a lower capital intensity. In the case of an ACE/ACNE the imputed return on equity lowers the cost of equity finance and therewith the cost of capital. Accordingly, the minimal required return an investment project has to earn declines and thus a larger number of investment projects are profitable. Consequently, capital accumulates and the economy ends up with a larger total stock of capital under the ACE/ACNE.
3.4 Welfare
As a measurement for welfare, we apply the equivalent variation that specifies the differences in expenditures with respect to the before and after tax reform utility levels U0 and U1, under the pre reform price structure p0:
|
| (16) |
After the equivalent variation has been transformed into an annuity flow, yEV(1-
H)EV , the "change in welfare in percent of GDP" (or life-time income) is finally computed by the ratio of the annuity stream yEV and GDP (or life-time income), according to:
|
| (17) |
|
4 Model calibration |
|---|
|
TopAbstract1 Introduction2 The ACE versus...3 The model4 Model calibration5 Simulation results6 Sensitivity analysis7 ConclusionAppendixReferences |
|---|
In order to solve any numeric CGE model, and therefore to be able to evaluate any tax reform, functional forms have to be specified and several behavioural parameters have to be applied. In this context, the calibration implies that the parameters for various elasticities and major economic variables have to be consistent with the empirical evidence and that the initial steady-state values reflect empirical data.
All behavioural parameters used in this model are standard results confirmed by the empirical literature. The most important ones are summarised in Table 1.
|
The real annual growth rate of the German economy is assumed to be 1 percent, which is a fair estimate for Germany after re-unification. Economic depreciation reaches 10 percent of the capital stock and the adjustment speed towards the new steady state is determined by the half-life of investment. In accordance with the study of Cummins, Hassett and Hubbard (1996), we take a value of 8.0, implying that during the following 8 years after the policy shock half of the long-run increase in the capital stock is accumulated33. For the constant elasticity of substitution (CES) production function, the applied elasticity of factor substitution of 0.8 is based on Bundesbank (1995). A value ranging between 0.3 and 1.3 for West German industries was also computed by Roskamp (1977). There is extensive empirical literature dealing with the estimates of the elasticity of substitution between labour and capital. A survey of these studies is provided by Chirinko (2002).34 Concerning the elasticity of the debt-asset ratio, we follow Gordon and Lee (2001), who estimate that a 10 percentage point decrease in the corporate tax rate leads to a reduction in the debt-asset ratio by 3–4 percent. The labour supply elasticity is set equal to 0.37,35 in order to represent an average of empirical estimates for different age and sex groups (Fenge, Übelmesser and Werding 2002). Moreover, this value lies within the range of 0.2 and 0.43 which, in the opinion of many economists, are appropriate values for compensated labour supply elasticities for men and women, respectively (Fuchs, Krueger and Poterba 1998). Given the fact that the macroeconomic effects of capital income taxation are very sensitive to the choice of the value of the intertemporal elasticity of substitution (King and Rebelo 1990; Summers 1981), the value of this parameter has to be set with great care. Our value of 0.4 is based on Flaig's (1988) empirical research for Germany, and is just one percentage point lower than the values applied by Böhringer, Boeters and Feil (2004), Keuschnigg and Dietz (2004) or by Valkonen (1999).
|
5 Simulation results |
|---|
|
TopAbstract1 Introduction2 The ACE versus...3 The model4 Model calibration5 Simulation results6 Sensitivity analysis7 ConclusionAppendixReferences |
|---|
Starting from the present German tax system we perform a number of simulations that depict either the introduction of an ACE/ACNE or of a CBIT. Revenue neutrality is assured either by a change in the VAT rate or in the profit tax. The status quo of the present Germany tax system—as presented in Table 2—serves as a starting point for the different reform scenarios.
|
Even though the statutory corporate tax rate amounts to just 25 percent in Germany, the effective corporate tax burden adds up to as much as 38.3 percent if the local trade tax and the solidarity surcharge are also considered. On the household level, theoretically all types of income, including capital and labour income are merged and then taxed at a single progressive personal income tax rate. The latter reaches a top marginal tax rate of 42 percent, or 44.3 percent if we account for the solidarity surcharge.36 In Germany, different kinds of capital income including dividends and capital gains face, however, a specific tax treatment: Dividends are taxed according to the ‘half income principle’,37 implying a lower tax burden on dividend income and capital gains are completely tax exempt. The top personal income tax rate of 44.3 percent just applies to interest income. Regarding labour income taxation, we assume an average annual labour income of
20 814 for the representative individual, which translates into an average labour tax rate of 28 percent, or 29.5 percent if we add the solidarity surcharge.
Revenue neutrality via a change in the VAT rate
The first set of simulation results is presented in Table 3. Within these simulations, revenue neutrality is achieved via an adjustment of the VAT rate, since we do not find any profit tax increase that fully compensates for the introduction of an ACE. Moreover, we set the imputed return on equity equal to the rate of debt interest to ensure financial neutrality. The two different scenarios under the ACE/ACNE vary with respect to the personal tax on capital income.
|
Scenario 1 depicts the case of an ACE/ACNE for Germany and thus addresses the problem of financial neutrality at firm level. The taxation at the personal level remains unchanged, implying a zero tax on capital gains and a tax of around 22 percent on dividend income. Under this constellation of tax rates, interest income faces, however, an enormous tax burden compared to dividends and capital gains, which might distort the investment decision in favour of firm shares vis-à-vis firm bonds. To avoid this distortion, we abolish the half imputation principle in Scenario 2 and apply in addition a capital gains tax of 22 percent to assure that all types of capital income are taxed equally.38
Under the CBIT proposal, we also consider two reform alternatives: Scenario 3 regards solely the introduction of a CBIT as usually stipulated in the literature. Furthermore, Scenario 4 introduces an extended reform, which also provides for an immediate write-off for investment expenditures to ensure investment neutrality.
As depicted in Table 3, the introduction of the ACE/ACNE accompanied by the half-income principle of dividend taxation on the household level (Scenario 1) achieves the best results in terms of welfare. Due to the possibility to deduct an imputed return from the profit tax base, the cost of capital decreases by 6.3 percent and 4.3 percent for corporate and non-corporate firms, respectively (Table 4). As a consequence, the investment activity and capital accumulation are enhanced. In detail, the capital stock increases by around 20 percent for the whole economy, inducing an increase in labour demand of corporate firms by 5 percent and an increase in aggregate labour supply of 1.7 percent. Since this reform Scenario 1 is, however, rather costly, the VAT rate has to be increased in the long-run steady state by 5.1 percentage points to ensure revenue neutrality. Therefore, the positive welfare effects induced by the increased capital accumulation and labour demand are counterbalanced by the rather large increase in the VAT rate such that overall welfare in terms of life-time wealth rises by only 0.1 percent.
|
Scenario 4, the CBIT combined with a provision for an immediate write-off, turns out to be the second best reform alternative in terms of welfare. The provision of the immediate write-off induces a decline in the cost of capital by 21.0 percent and 4.7 percent for corporate and non-corporate firms, respectively. Despite that, the non-corporate sector faces a capital decumulation by nearly 30 percent (Table 4), which can be explained by simple general equilibrium effects. Due to the huge demand for capital by the corporate sector—following the drastic reduction in the cost of capital and thus in the effective marginal tax rate—combined with the sheer size of the corporate sector, investments by the non-corporate sector are crowded out. As a consequence, the increase in the capital stock of 16 percent in Scenario 4 falls short of the accumulation of capital in Scenario 1 when the ACE/ACNE is introduced.39 Accordingly, also the welfare effects turn out less positive compared to Scenario 1 even though the reform raises a larger amount of tax revenue from the profit tax such that the VAT rate can even be lowered by 2.4 percentage points in the long-run steady state.
The comparison of Scenario 1 and 4 raises the question of whether the positive welfare results occur due to the shift of the tax burden from corporations to households via the VAT. This is one possible explanation, since we have to keep in mind that, on the one hand, the VAT just distorts the labour-leisure decision by affecting current real wages. The profit tax, on the other hand, is more harmful to the economy in a world of high capital mobility: Capital flight results in a decreasing capital intensity, and therefore the marginal productivity of labour drops as well. Domestic real wages go down such that the burden of the corporate tax is ultimately borne by labour. All these implicit negative effects can be avoided if the tax burden is shifted from the corporate to the consumption tax.
In this context, another interesting comparison refers to the introduction of an ACE versus a simple reduction in all capital income tax rates. For this purpose, we perform an additional simulation where we allow for a proportional reduction in all capital income tax rates. Thereby, the magnitude of the cut in the capital income tax rates is determined by the present value of the costs which could be incurred if the ACE/ACNE proposal was introduced.40 In case of the ACE/ACNE, welfare as a percentage of life-time income or GDP increases by 1.27 percent or by 0.73 percent, respectively, while the same two measures increase by a slightly larger extent, namely by 1.36 percent and 0.78 percent for the reform scenario envisaging a proportional cut in all capital income tax rates. Therefore, a simple cut in tax rates seems to be more efficient than a overhaul of the tax system as implied by the ACE.
Table 4 also depicts the results for the second ACE/ACNE Scenario. Under this Scenario 2, the cost of capital declines only within the non-corporate sector, since corporate sector firms, which mainly rely on retained earnings as investment funds, suffer from the abolition of the half imputation system as well as the introduction of the capital gains tax of 22 percent. Therefore, the cost of capital as well as the effective marginal tax rate of corporate firms rise by 14.4 percent and 19.7 percent, respectively.41 According to the change in the sector-specific cost of capital, capital accumulates within the non-corporate sector and decumulates within the corporate sector. In total, the economy-wide capital stock increases by 9.6 percent, as reported in Table 3. Thus, the negative, economy-wide repercussions induced by the increase in the dividend and capital gains tax demand an even larger increase in the VAT rate, which has to be raised by 6.5 percentage points to ensure that the government budget is balanced. Consequently, the labour supply declines (–0.5 percent) and therewith domestic consumption (–0.7 percent), leading to a decline in welfare of 1.1 percent in terms of total wealth.
Scenario 3, the pure CBIT, which is not accompanied by a provision for immediate write-off, does not produce favourable results either. Since the advantage of the debt interest deductibility is abolished, the cost of capital increases by almost 10 percent for corporate firms and by about 22 percent for non-corporate firms. As a result, the capital stock decumulates within the entire economy by 10.2 percent and GDP shrinks by 5.3 percent. Moreover, labour demand declines and accordingly gross wages and consumption. Thus, it is not surprising that this reform induces a decline in welfare by 1.2 percent in terms of life-time income or 0.7 percent in terms of GDP. This negative outcome occurs even though the increased profit tax revenue allows a reduction in the VAT rate by 1.3 percentage points from initially 16–14.7 percent. Thus, the negative impact of a pure CBIT on capital accumulation and accordingly on gross wages (if the profit tax rate is not reduced after the reform) is so large that it can not be compensated by a lower consumption tax. Once again, the harmful consequences of profit taxation compared to the rather modest effects triggered by lower consumption taxation are striking.
Revenue neutrality via a change in the profit tax
The simulation that envisages the introduction of an ACE/ACNE financed by an exogenous increase in the profit tax rate shows that this rate can not be high enough to compensate for the costs of the reform. Even for very large values of the profit tax, lump sum transfers to households have to decline in order to balance the government budget.42 This result occurs since our dynamic general equilibrium model captures a wide range of effects and economy-wide repercussions. Accordingly, a very high profit tax rate has substantial negative effect on investments, and therefore on capital accumulation, labour demand and GDP. As a consequence of this economic drop down, the tax bases of other taxes shrink as well and an adjustment in transfers is required in order to finance the reform. If we performe our analysis in a two-period framework such that the budget is balanced in the second period, our results show that the introduction of the ACE/ACNE (above Scenario 1) needs to be accompanied either by an increase in the corporate tax rate by 15.2 percentage points to a level of 53.5 percent or by a simultaneous increase in the profit tax that applies to corporate and non-corporate firms. Under the latter alternative, the tax rate on corporate profits has to increase from 38.3 to 46.5 percent and the tax rate on non-corporate profits from 45.4 to 54.5 percent, other things being equal.
If the CBIT without an immediate write-off (Scenario 3) is implemented within the two period framework, the amount of additional tax revenue collected is enormous and second period tax rates can be reduced to a large extent without violating the governmental budget constraint. In detail, the profit tax rate can be cut by almost 10 percentage points, implying a second period corporate tax rate of just 27.5 percent and a tax rate of only 35.5 percent on non-corporate profits. In case only the corporate tax rate is cut, it can be reduced ceteris paribus to 20.5 percent. Such a computation, however, neglects the economy-wide repercussions that can be captured by our dynamic, multi-period general equilibrium model. Therefore, we perform a last simulation of the CBIT proposal and take this time an exogenous adjustment in the corporate tax as a means to ensure revenue neutrality into account. In case the CBIT is introduced without (with) an immediate write-off, the uniform profit tax can be reduced to a value of 38.0 (30.0) percent.
The first three columns of Table 5 show the results of implementing a pure CBIT without any special depreciation allowances. The overall message is clear: Even though the usual line presented in the literature states that a CBIT might have positive effects if it is accompanied by a reduction in the corporate tax rate, the simulation results show that the opposite is true. First of all, the cost of capital for corporate firms does increase since the after-reform uniform profit tax rate is still too high and does not compensate the tax base broadening coming along with the CBIT. The new tax level is, however, lower for the non-corporate sector such that these firms face a cost of capital decline. Nevertheless, the overall, economy-wide capital stock drops by 9 percent in the long-run since the capital decline in the corporate sector is the key driving force due to its sheer size. The picture, however, changes when we allow for an immediate write-off. In this case, there is a tremendous decline in the capital costs for both types of firms, since the uniform profit tax rate is reduced to just 30.0 percent. As a result, capital accumulates and the long-run, steady-state capital stock grows by 20 percent. The increase in labour demand (+3.0 percent) induces an upswing in wages and accordingly also disposable income and consumption rise. Welfare in terms of total wealth increases by 0.3 percent (or by 0.2 percent in terms of GDP) and is even higher than in the ACE/ACNE Scenario 1 (Table 3).
|
To summarise, the different ACE/ACNE and CBIT scenarios simulated in this chapter can be ranked in the following way: The best results are achieved under the CBIT accompanied by an immediate write-off and if revenue neutrality is achieved by an exogenous change in the profit tax rate (Scenario 4*). When the VAT is adjusted in order to balance the government budget, the ACE/ACNE proposal with no additional capital gains taxation (Scenario 1) leads to the most significant welfare gains.
|
6 Sensitivity analysis |
|---|
|
TopAbstract1 Introduction2 The ACE versus...3 The model4 Model calibration5 Simulation results6 Sensitivity analysis7 ConclusionAppendixReferences |
|---|
In order to check the robustness of the performed simulations, a sensitivity analysis is of crucial importance. In our case, we focus especially on the debt-asset ratio, the corporate tax rate as well as the imputed return on equity since these parameters are, among other things, of major importance for the derived results.
To start with, we reconsider Scenarios 1–4 and check how these policy shocks affect the economic aggregates if we start from a comparatively lower corporate tax rate. Instead of the corporate tax rate of 38.3 percent currently prevailing in Germany, we start from a statutory tax rate of 25 percent on corporate profits. Such an experiment is of interest especially for countries like the new EU Member States, which are characterised by much lower corporate tax rates and for which accordingly different reform scenarios might be appealing. The long-run changes in the key economic figures are depicted in Tables 6 and 7.
|
|
If the corporate tax is reduced to just 25 percent, the results under the ACE/ACNE Scenario 1 worsen substantially. One explanation for this result refers to the fact that the additional tax advantage accompanying the deductibility of an imputed return from the profit tax base (which is now lower than in case we start from a higher profit tax) is more than offset by the disadvantage coming along with the higher VAT rate of 21.5 percent, instead of the initial 16 percent. This increase in the VAT rate is, however, necessary to balance the government budget. As a consequence, welfare decreases only by 0.3 percent of GDP in Scenario 1.
Concerning Scenario 2, the increase in the corporate-sector cost of capital due to the abolition of the half imputation system and the introduction of the capital gains tax is now even larger such that the long-run capital accumulation is only modest compared to the results presented in Table 3. Accordingly, if we start from a lower level of the corporate tax, an even larger increase in the VAT rate is required to finance the reform. This distinct rise in the VAT rate will, however, trigger considerable losses in welfare (Table 6).
As also depicted in Tables 6 and 7, Scenario 3, the CBIT reform with no immediate write-off, leads under the new tax constellation with a lower profit tax rate of 25 percent to slightly better results compared to the base scenario presented in Table 3. Since we start from a lower corporate tax rate, the disadvantage accompanying the removal of the debt interest deductibility is now less severe. Accordingly, the cost of capital increases by 5.0 and 21.5 percent for corporate and non-corporate firms, respectively. The decline in the economy-wide capital stock amounts to 6.5 percent and is thus 3.7 percentage points lower compared to the baseline simulation (Table 3). Gross wages and thus disposable income and consumption decline, resulting in a decrease in welfare by 0.6 percent in terms of GDP.
If the CBIT is implemented in combination with a provision for immediate write-off, as it is the case in Scenario 4, the additional tax revenue collected by the CBIT is less compared to the reform scenario where we start from a 38.3 percent profit tax. Hence, the long-run, steady-state VAT rate can only be reduced by 1.6 percentage points (Table 6) such that the reform induces a slight welfare loss of 0.2 percent in terms of GDP.
The second sensitivity analysis is carried out with regard to the imputed return on equity capital, which is set below the interest rate paid on firm debt such that financial neutrality is no longer guaranteed under the ACE/ACNE. Table 8 shows how the results for the ACE/ACNE Scenario 1 change.
|
Even though this alternative does not equalise the cost of capital across the different sources of finance, the overall induced macroeconomic effects have improved compared to the baseline scenario (Table 3). Since the imputed return on equity has been reduced, the government's financing gap is also smaller. Accordingly, only a small increase in the VAT rate is sufficient, leading in turn to slightly better welfare results.
Finally, the last sensitivity analysis is performed with respect to the debt-asset ratio. This parameter is of particular importance for the CBIT proposal since it influences the outcome of such a reform to a large extent. For instance, if in some countries firms are characterised by rather low debt-asset ratios, these firms will suffer less from the introduction of a CBIT. The government will, however, also earn only little additional tax revenue from such a reform. On the contrary, if firms are highly indebted, they will be severely hit by the removal of the debt interest deductibility. Table 9 summarises the results of the sensitivity analysis where the debt asset ratio of corporate and non-corporate firms is reduced by 50 percent.43
|
The first three columns of Table 9 depict the results for Scenario 3*, a CBIT without an immediate write-off. Compared to the baseline simulation (presented in Table 3), the changes in most of the macroeconomic variables do not differ much except for the change in the VAT rate. Since the broadening of the tax base assures only little additional tax revenue—due to the reduced debt asset ratio—the VAT rate needs to be increased slightly by 0.7 percent to balance the government budget. Nevertheless, the occurring welfare effect is of the same size as the one reported under the baseline scenario in Table 3.
As opposed to this, in Scenario 4* the CBIT is accompanied by a provision for immediate write-off. In this case, the reform produces even larger positive effects. The cost of removing the debt interest deductibility is now smaller and therefore the considerable capital accumulation of 15.4 percent has positive, economy-wide repercussions. As a consequence, the reform leads to higher revenues from taxation per se such that the long-run VAT rate can be reduced by five percentage points. The overall welfare gain now amounts to 1.8 percent in terms of life-time income or 1 percent in terms of GDP.
|
7 Conclusion |
|---|
|
TopAbstract1 Introduction2 The ACE versus...3 The model4 Model calibration5 Simulation results6 Sensitivity analysis7 ConclusionAppendixReferences |
|---|
The aim of this paper is to discuss and qualify the effects of two opposed reform alternatives: the ACE and the CBIT. Under both policy proposals, the cost of capital under debt and equity finance are aligned. However, due to the deductibility of both debt interest and of an imputed return on equity, the ACE/ACNE is characterised by a much smaller profit tax base compared to the CBIT under which neither of the two is tax deductible. Therefore, opponents of the ACE/ACNE argue that a higher corporate tax rate is necessary to guarantee a certain tax revenue; this outcome is however, not desirable for a small open economy in a world of high capital mobility where the profit tax rate acts as a signalling device.
These kind of arguments are, however, invalid if we assume that the reforms are financed by any other tax, like the VAT for instance. For this case, our simulation results show that introducing an ACE/ACNE with no additional capital gains tax on the personal level has significant positive effects on investments, capital accumulation and welfare. For the opposite reform scenario, the CBIT, the results turn out to be less favourable. Under the CBIT, the cost of capital rises and induces negative consequences for investment. Accordingly, firms will reduce their labour demand such that gross wages decline. Even though the VAT rate can be reduced under the CBIT proposal, this tax stimulus is not sufficient to compensate the negative effects arising from the reform. If the CBIT is, however, accompanied by a provision for immediate write-off to ensure investment neutrality, the impact of such a reform is quite substantial and positive. Moreover, if we additionally allow for an exogenous adjustment in the profit tax instead of the VAT to finance the reform, the CBIT accompanied by an immediate write-off achieves the best results in terms of welfare. Under the ACE/ACNE, an exogenous increase in the profit tax cannot be high enough to balance the financial cost of such a reform. The rationale for this result refers to the fact that an increase in the profit tax rate also raises the tax advantage from the ACE/ACNE in the next period. Therefore, the profit tax has to be increased once again to assure a balanced government budget. This increase in the profit tax will, however, boost once again the tax advantage of the ACE/ACNE and thus in turn imply once again an increase in the profit tax rate. Moreover, the higher profit tax additionally has a substantial negative effect on capital accumulation and labour demand and thus on all other macroeconomic variables. A single rise in the profit tax rate is, therefore, insufficient to finance the introduction of ACE/ACNE in our dynamic general equilibrium framework.
A last but interesting comparison is drawn between the ACE/ACNE and the current German system of capital income taxation. Therefore, we perform a simulation where we allow for a proportional reduction of all capital income tax rates such that the present value of the financial cost is identical for implementing this reform or the ACE/ACNE. In this case, our simulation results show that the proportional cut in capital income tax rates lead to slightly better results in terms of welfare compared to the ACE/ACNE. Under the former, welfare increases by 0.78 percent in terms of GDP and in case we implement the ACE/ACNE welfare increases by 0.73 percent. Accordingly, a simple cut in tax rates seems to be more efficient than an overhaul of the tax system as implied by the ACE.
To avoid the double taxation of interest income under the CBIT reform proposal, one could envisage the abolishment of interest income taxation on the household level. Nevertheless, under these circumstances, policy makers would have to think of whether to extend this generous treatment of interest income only to income received from domestic firm bonds or to domestic and foreign government bonds as well. Even though we do not quantify this last simulation exercise in this article, the overall effects will depend on which of the two opposite effects prevail. The positive effects of no interest income taxation face the negative effects of a higher VAT or a lower decline in the profit tax, which might be necessary to compensate the revenue loss.
Finally, the sensitivity analysis underlines the accuracy which policy-makers should put forward when designing a reform proposal. The outcomes of the different scenarios are very sensitive with regard to the choice of the imputed return, the calibrated debt-asset ratio or the corporate tax rate. Therefore, the magnitude of these parameters is crucial when deciding on one reform alternative or the other.
We conclude with the statement that for the comparison and evaluation of these two diametrically opposite reform alternatives it is very important to consider additional issues besides the aspect of financial neutrality on firm level. For instance, the fact whether an ACE/ACNE reform proposal in addition taxes dividends and capital gains on personal level, or if a provision for immediate write-off is also applied under the CBIT to ensure investment neutrality, are important features that change the results to a large extent. Moreover, the ultimate decision of which scenario is the best reform alternative depends on the concrete reform details and also differs from country to country, depending on the prevailing corporate tax rates, firms characteristics and macroeconomic background. Consequently, one cannot make across-the-board recommendations but has to admit that the most appropriate reform scenario depends on the given circumstances.
|
Appendix |
|---|
|
TopAbstract1 Introduction2 The ACE versus...3 The model4 Model calibration5 Simulation results6 Sensitivity analysis7 ConclusionAppendixReferences |
|---|
A1 Corporate firms
Since we refer to a mature economy, characterised by mature firms, we follow the "New View" of dividend taxation.44 However, keeping in mind the empirical evidence provided by Auerbach and Hasset (2003),45 who state that both views on the effects of dividend taxation are valid, we determine new share issues by
. New investments are largely financed by retained earnings or by new debt, BNC, and only a fixed fraction, ß, of 5 percent is financed via new share issues. This approach, however, does not apply to non-corporate firms because these have to rely on external equity to finance investments.
Plugging Equation (3) into the flow of funds equation, we derive an explicit expression for dividends DivC as output YC less labour costs wCLC, interest payments iBHBC, new shares 
, depreciation 
KC and corporate tax payments:
|
| (A1.1) |
In equilibrium, the return on equity has to equal the net of tax dividend payment and the net of tax capital gains which can be derived from holding firm shares. Hence,
|
| (A1.2) |
Here 
is the investor's required return which is necessary if the investor should be willing to hold the asset. This return is higher than the net return on firm or government bonds since it includes a risk premium.
A2 Non-corporate firms
As opposed to corporate firms, a non-corporate firm has no possibility to finance investments out of retained earnings since all profits are distributed to the owner, implying DivN = 
N. Accordingly, a non-corporate firm can only choose between new debt, BNN, and new equity injections, VNN, as a possible source of finance for its investments. Thus, the flow of funds equation for the non-corporate firm can be simplified to:
|
| (A2.1) |
The return on equity again equals dividends and net of tax capital gains:
|
| (A2.2) |
A3 Intertemporal optimisation
The firms of each sector seek to maximise their value, and therefore choose their intertemporal pattern of labour demand, investment and new debt optimally. Thereby, the firm value Vf will increase with the size of the capital stock accumulated and fall with the level of debt inherited from the past. At the beginning of each planning period, t, the capital stock and level of debt are exogenous, as they are historically predetermined from previous actions. The differential equation determining the end of period firm value is given by:46
|
| (A3.1) |
The variable denotes a tax factor which is defined as ‘one minus the respective tax rate’. Hence, the end of period market value of a firm is determined by the present value of all future net of tax dividend payments less new equity injections. The net dividend flow is discounted at the cost of equity, ref = rVf/(1- G,f), which is the required gross return on firm level.
Applying a value function of the form 
and assuming that investment is optimised from period t + 1 onwards, we can find today's optimal labour demand, investment and financial behaviour by maximising the Bellman Equation of Dynamic Programming:
|
| (A3.2) |
Moreover, defining the shadow prices of capital 
and debt 
, respectively,47 as well as the following three tax parameters for corporate and non-corporate firms:
|
| (A3.3) |
the optimality conditions concerning the control variables labour, investment and new debt state: