Abstract
The Selected Issues paper provides an estimate of the output gap and potential output for Italy, and examines the sensitivity of the results to assumptions regarding employment and productivity growth. The paper focuses on the labor market more directly by examining the linkages between wage bargaining systems, regional wage differentiation, and regional unemployment disparities. It also provides an assessment of the government’s tax reform program, including its potential to increase incentives for employment and investment.
IV. Tax Reform in Italy50
A. Introduction
73. Soon after coming to office in 2001, the new Italian government embarked on a program of fundamental tax reform. This envisages both a substantial reduction in the overall tax burden and far–reaching changes in the structure of the tax system. These reforms come soon after—and largely reverse—extensive reforms introduced by the previous government between 1996 and 1998.
74. Both the precise timing and many of the details of the reform remain to be specified,51 the government having made it clear that the former is contingent on the development of the wider fiscal position. Nevertheless, the broad outlines of what the government aims to achieve in its current term of office are clear, with profound changes are planned in all of the major taxes except the VAT. Table 1 shows the broad revenue implications of the key reforms included in the program.
Revenue Effects of the Main Components of Reform
This is the first step in the planned elimination of the IRAP, which currently raises about €30 billion (around 2.5 percent of GDP).
Revenue Effects of the Main Components of Reform
| Revenue Gain | ||
|---|---|---|
| In billions of euro | In percent of GDP | |
| Reform of company taxation from 2003 | +2.285 | 0.2 |
| Exclusion of 20 percent of labor costs from IRAP 1/ | 2.28 | 0.2 |
| Reform of capital income taxes | -1.55 | 0.1 |
| Reform of PIT (full effect) | -18.00 | 1.5 |
This is the first step in the planned elimination of the IRAP, which currently raises about €30 billion (around 2.5 percent of GDP).
Revenue Effects of the Main Components of Reform
| Revenue Gain | ||
|---|---|---|
| In billions of euro | In percent of GDP | |
| Reform of company taxation from 2003 | +2.285 | 0.2 |
| Exclusion of 20 percent of labor costs from IRAP 1/ | 2.28 | 0.2 |
| Reform of capital income taxes | -1.55 | 0.1 |
| Reform of PIT (full effect) | -18.00 | 1.5 |
This is the first step in the planned elimination of the IRAP, which currently raises about €30 billion (around 2.5 percent of GDP).
75. This chapter provides an account and evaluation of the main elements of this reform program.52 The bulk of the chapter deals in turn with the major tax categories affected: onlabor income (Section B), the corporate income tax (Section C), capital income taxation (Section D), and the regional tax on value-added (Section E). Section F considers the case for a reorientation of the reform program to produce a system closer to the “Nordic” model, to which it bears some similarity. Section G concludes.
B. Taxation of Employment Income
76. At the heart of the reform program is a substantial reduction in the level and progressivity of the personal income tax (Imposta sul reddito delle persone fisiche, the IRPEF), which is essentially a tax on labor income (separate rules and schedules apply to dividends and other forms of capital income, as discussed in Sections C and D below).
The reforms
Before reform
77. Employment income is currently taxed at the state (i.e., national) level at the five marginal rates shown in Table 2 There are also regional surcharges piggy-backed on to the state tax. These are set at 0.9 percentage points, but the regions are also able to—and, as a relatively recent development, now do—increase this by up to a further 0.5 points (and to set rates that vary with the level of income). 53
Current Rate Schedule of the IRPEF
Current Rate Schedule of the IRPEF
| Taxable Income (In thousands of euros) | Marginal Rate |
|---|---|
| 0-10 | 18 |
| 10-15 | 24 |
| 15-31 | 32 |
| 31-70 | 34 |
| 70+ | 45 |
Current Rate Schedule of the IRPEF
| Taxable Income (In thousands of euros) | Marginal Rate |
|---|---|
| 0-10 | 18 |
| 10-15 | 24 |
| 15-31 | 32 |
| 31-70 | 34 |
| 70+ | 45 |
78. The IRPEF is levied on an individual basis. While tax is in principle payable from the first euro of earnings, a large number of (nonrefundable) credits—up to 80, on some counts—are available to reduce or eliminate liability. A basic credit for “dependent workers” (that is, employees) is available against all earned income, in an amount that falls as earnings increase. This implies that, for a single person, no tax is payable below an annual income of around Є6,000. Different income–related credits apply to professional income, and further credits are available for a dependent spouse and to pensioners (all these being income–related), for children, and (in amounts corresponding to a tax rate of 19 percent, and subject to maxima) for spending on health, education, mortgage interest, and life insurance.
79. For the self-employed, that part of business income corresponding to a normal return on capital was taxed at the reduced rate of 19 percent, in line with the dual income tax applied to companies (described in Section C below). The rest was subject to tax at ordinary rates (with the return on capital included in calculation of the appropriate average rate).
The reform
80. It is proposed that by the end of the government’s current term there be only two marginal rates of IRPEF, 23 and 33 percent. The point at which tax would become payable has not yet been specified, but it has been announced that the higher rate will apply only to taxable income in excess of €100,000. This is high: it has been estimated that the higher rate would apply to less than 1 percent of taxpayers.
81. The system of credits is to be replaced by one based on allowances—that is, amounts to be deducted from the tax base—that decline with the level of earnings. But the number, level and withdrawal rates of these allowances (including the basic tax–free amount) remain to be specified.
82. The process of reforming the IRPEF is to begin in 2003, with cuts amounting to Є5.5 billion (0.5 percent of GDP).
Assessment
83. The impact of the IRPEF reform on incentives to work—a central concern in evaluating any reform of personal income taxation—is complex. Article 3 of the legge delega commits the government to ensuring that no individual loses from the reform, which (all else equal) means that the average rate of IRPEF—tax paid relative to earned income—will in all cases fall, or at least not rise. The “income effect” of the reform is thus unambiguously to reduce work effort: with higher after–tax income at their initial level of earnings, people can afford to work less. For the labor supplied by those employed to increase, this will have to be offset by the “substitution effect” arising from a reduction in the marginal effective tax rate on labor income: such a reduction makes “leisure”—to be interpreted broadly as time not spent earning taxed income—more expensive in terms of the after–tax income foregone, and so should lead workers to take less of it.
84. The best way to gauge the likely balance between these effects is by simulating the effects of the reform using a micro–data based model of labor supply. This is difficult to do, however, while important details of the reform remain to be specified. In particular, what matters for labor supply is not the statutory marginal rate structure (which has been specified) but rather the marginal effective rate of tax on labor income (METL). The latter will reflect, in particular, the pattern of income–related allowances (as yet unknown), the point being that the removal of an allowance as income rises is an implicit tax on that income. Up to the point at which all income–related allowances are removed, METLs will be higher, and perhaps significantly so, than statutory rates. Moreover, given the government’s declared intention of focusing allowances more closely on lower incomes than is the present system of credits, the addition to the METL from this source is likely to be more concentrated in the lower reaches of the earnings distribution than at present.
85. While any assessment of the IRPEF reform at this stage is thus inevitably impressionistic, two points stand out in terms of the effect on those in employment:
At low levels of earnings, the METL is unlikely to fall significantly, and may quite plausibly increase slightly. Income and substitution effects then act in the same direction, and labor supply will fall.
At middle and high income levels, the METL is likely to fall. For higher income levels—at which income–related allowances and credits are likely to be exhausted—the fall seems certain to be substantial, in the order of 10 percentage points. This creates the possibility of increased work effort from such groups, though of course the income effect is also large at the top of the earnings distribution. Though views on the issue continue to differ, the empirical evidence on labor supply responses does not suggest that the impact on labor supply—or indeed on taxable income more generally—will be very marked.54
86. The nature and strength of these effects depend on the rapidity with which allowances are withdrawn as income rises. One set of estimates,55 based on the assumption that allowances will be exhausted at an income level of about Є28,000, has METLs rising by about 2 percentage points in the lower half of the income range—a modest effect—and falling, generally by a much larger amount, in the upper half.
87. There are other dimensions of labor supply decision, however, that may be more powerfully affected than the effort of the employed. In particular, lower average tax rates decrease the attractions of working in the informal rather then the formal sector. They also tend to reduce unemployment by reducing replacement rates—the ratio, that is, of net income out of work to that in work—and boost the demand for labor by reducing the cost to employers of paying any given after–tax wage. These effects on the decision of whether or not to participate in the formal sector could plausibly be more powerful than those on the effort supplied by those already in the formal sector.
88. The reduced progressivity of the IRPEF could also affect the level of unemployment through its effect on the bargaining between unions and firms in such a way as to worsen unemployment. For in trading off their objectives of high employment and high net wages for their employed members, unions may recognize that a high marginal tax rate increases the employment costs of achieving any increase in net wages (because it increases the gross cost to employers of financing that increase). However, although there is increasing empirical evidence that progressivity has indeed proved good for employment in a number of countries,56 the only empirical investigation of this issue for Italy fails to identify such a link.57
89. The impact of the reform on the distribution of real income will also depend on details that have yet to be determined, not least the choice of threshold below which no tax will be payable. Nevertheless, the broad thrust is clear–cut. For the full reform as sketched in the legge delega, the proportionate gain to the better off will be far greater than that to the least well–off. At very high income levels, the average and marginal rates are essentially the same, so that the reduction in the average rate of tax is in the order of one–quarter. Baldini and others (2002) estimate the foil IRPEF reform will increase the Gini measure of after–tax inequality by about 3.7 percent.58 For the immediate future, however, the government has announced that the first step on the IRPEF reform, in 2003, will concentrate on reducing the burden on the least well–off. It remains to be seen how this will be done (the task not being easy, since many kinds of tax reduction that benefit the least well–off—such as an increase in tax–free amounts—tend also to benefit the better off).
90. One of the main outstanding details, as already noted, is the pattern of income–related allowances that will replace the present system of credits. In itself, this change achieves very little, since the effects of the prereform income–related credits can be exactly replicated by an appropriately chosen pattern of income–related allowances under the new one.59 For example, for the vast majority of taxpayers, who will face a marginal rate of 23 percent under the new system, a tax credit equal to 19 percent of the relevant expenditure (such as on health spending) is equivalent to a deduction of 19/23 (approximately 83 percent) of that expenditure against the PIT. The key question in respect of this aspect of the reform is thus whether it will prove an opportunity to obtain a significant reduction in the number of special provisions, whose effect is to increase METLs and—especially—to make the system more opaque. Given the commitment to ensure that there are no losers from reform, achieving this simplification will not be easy.
C. The Corporate Tax (IRPEG)
The reforms
91. The reform affects almost all the key features of the corporate tax (Imposta sul reddito delle personne giuridiche, the IRPEG).
The “dual income tax” 60
Before reform 61
92. At the heart of the corporate tax system was the dual income tax (DIT), introduced in 1998.62 The central purpose of this system was to partially redress the bias toward the use of debt finance implicit in the normal practice of allowing interest payments on debt, but not any of the return to shareholder, as a deduction against corporation tax. To this end, the DIT taxed at a reduced rate an amount corresponding to a normal return on equity capital 63—that is, retained profits and new subscriptions—put into the enterprise after September 30,1996. This was an unusual system, though not quite unique.64
93. The notional return on equity was taxed at 19 percent, and other profits at 36 percent.65 In the final year of its operation—and as part of an intended transition to a situation in which the allowance was given on all equity, not just that injected since 1998—firms were allowed an uplift of 40 percent on the equity deduction, creating the “super–DIT.” 66
The reform
94. The incoming government quickly froze the DIT: the reduced rate is now available only for equity increases up to 30 June 2001.67 Moreover, the benefit of the reduced rate is now removed when earnings are distributed to shareholders, since distributions from such earnings are no longer deemed to have paid tax at the ordinary rate. This remaining allowance is, in any event, to be abolished.
95. The rate is to be reduced from 36 to 34 percent in 2003, in line with the planned top marginal rate of the IRPEF.
Depreciation and investment incentives
Before reform
96. For investments in tangible assets, depreciation in each of the first three years may be taken at twice the ordinary rate laid down in the tax law. From 1999, the “Visco incentive”—named after the then Minister of Finance—enabled companies to charge at the reduced rate of 15 percent an amount equal to the lesser of the increase in equity that year and the amount of net investment in business assets. This was in addition to the normal provisions of the DIT.
After reform
97. No change is proposed in the basic system of depreciation allowances. The Visco incentive, however, has been removed and replaced instead by “Tremonti bis” (named after the new Minister of Finance). Under the terms of this scheme—which closely parallels a similarly–named incentive offered in 1994—96—firms are entitled to immediate expensing of 50 percent of the amount by which qualifying investments exceed the lowest of the averages of such expenditures over any four of the preceding five years. The range of qualifying assets is wide, encompassing most investments in new business assets, including intangibles and training. Tremonti bis is to expire at the end of 2002.
The treatment of dividends
Before reform
98. Dividends were taxed by a full imputation system. That is, shareholders—personal or corporate—were entitled to a credit against their own tax liability on the dividend equal to the corporate tax underlying that dividend.68. The imputation credit was fully refundable, so that tax–exempt investors, for example, were entitled to a refund (so long as the underlying corporate income had indeed borne tax). To ensure that the benefit of the DIT provision was not undone at the stage of taxing distributions, this rate applied irrespective of the rate applied to the underlying profits.
99. These arrangements ensured, in principle, that dividends passed between companies (whether linked in a group or not) without any additional payment of tax. At personal level, the imputation method was mandatory for dividends received from substantial participations, and optional for dividends from registered shares or quotas. The alternative was final withholding at 12.5 percent, which would be the preferred option—at a corporate tax rate of 36 percent—for all those whose marginal rate of personal rate of income tax was less than 44 percent.69
The reform
100. Imputation has been abolished. Instead, some as yet unspecified fraction of dividends received will be taxed as ordinary income in the hands of the final shareholder. This is similar to the approach adopted by Germany in its 2000 reform, with half of the dividend now taxed at the personal level.
101. Within groups of companies, the payment of dividends will cease to be a taxable event. For dividends received from companies resident in the EU but outside the group, 5 percent of the amount received will be taxable (with no credit for any underlying tax paid).
Treatment of company asset revaluations and groups
Before reform
102. Taxation of gains realized by companies on their assets, including shares in other companies, are in principle fully taxable (and losses fully deductible), but there are important exceptions. Gains on disposal of interests in companies are subject to a substitute tax at the reduced rate of 19 percent. Revaluations of depreciable assets are taxed at 19 percent, and of nondepreciable assets at 15 percent. The step up of basis allowed in the former case may make this an attractive option for assets, such as intangibles, for which depreciation is relatively rapid: the initial tax charge is then more than offset by an increase in subsequent depreciation allowances.
103. While there was no provision for group taxation under the IRPEG, devices were available by which tax losses could be transferred to companies able to make immediate use of them. In particular:
As noted above, companies were able to deduct against IRPEG reductions in the book values of their holding in other companies (irrespective of the extent of that holding).
The imputation system enabled losses to be absorbed by paying dividends into a loss-making company (perhaps artificially created), since there would be no immediate liability to offset the credit received. This was often attractive, in particular, in relation to holding companies, which typically have expenses but little income. Tax would be due when the loss making company returned to profitability, but this could be some way in the future.
The reform
104. Groups of companies are to be given the option (irrevocable for 3 years) of filing a single consolidated return. It remains to be determined exactly how a group is to be defined for this purpose, but the expectation is that a direct or indirect holding of 51 percent will be required.
105. A general participation exemption is also to be introduced, under which gains on disposals of shares in other companies are to be tax–exempt. While similar provisions have been introduced in a number of EU countries in recent years (most prominently by Germany but also, for example, by the United Kingdom), the Italian scheme is relatively extensive in that it applies irrespective of the extent of the shareholding.70
International aspects
Before reform
106. The imputation credit was not available in respect of dividends received from nonresident companies, which were generally taxable at ordinary rates for both companies and persons. For substantial participations, however, the system for companies was close to outright exemption: for holdings of over 25 percent of firms resident in other EU countries, only 5 percent of the dividend was taxable; 71 for substantial participations outside the EU, only 60 percent was taxable (with the proportion planned to increase to 95 percent) for countries on a “white list” of countries offering broadly similar levels of taxation to the Italian and with which Italy has concluded an information sharing agreement. To a first approximation, Italy thus offered source–based taxation in respect of the worldwide earnings of companies resident in Italy.
107. The imputation credit was payable to nonresident shareholders of companies resident in Italy only under treaty (as was the case with the France and the United Kingdom). Dividends paid to nonresident individuals were instead subject to withholding at 27 percent or lower treaty rate.
The reform
108. Subsidiaries of Italian companies resident abroad may be included in the groups described above, with credit given for taxes paid abroad. For companies outside the group, the 95 percent exemption is extended to all dividend receipts other than those from designated tax havens.
Assessment
109. The corporate tax system has potentially powerful effects on incentives to invest. These depend not only on the statutory rate of corporation tax but also on provisions regarding depreciation and the treatment of financial costs. Such effects, which are complex, are conveniently assessed in terms of two summary statistics:
The first is the marginal effective rate of corporate taxation (MECT). This is the proportionate amount by which the before–tax rate of return that an investment must earn in order to pay the underlying investor the after–tax return they require exceeds that required after–tax return. If the tax system were nondistorting—that is, were effectively a charge only on profits in excess of a normal return—the MECT would be zero. If the MECT is positive, on the other hand, then some investments that would be privately profitable in the absence of tax are not profitable in its presence, and so will not be undertaken.
The second is the average rate of effective corporation tax (AECT), which is the proportionate amount by which the after–tax profits on some particular project exceed the before–tax profits. For a nondistorting tax that bore only on super–normal profits, the AECT would be the same as the statutory rate.72 The AECT is especially relevant in assessing the attractiveness of a particular location for internationally footloose investments: all else equal, after–tax profits are highest when an investment is located in the country offering the lowest AECT.
Table 3 reports MECTs and AECTs for Italy, both before and after–reform,73
Statutory and Effective Rates of Corporation Tax in the EU
Including surcharges and average local taxes.
From Tables 7 and 8 of European Commission (2001), except postreform figure for Italy, kindly provided by Silvia Giannini.
Assuming a pretax return of 20 percent, and calculated as AECT = τ + (MECT)x (1-τ)C /0.2, where τ is the statutory rate of tax and C the user cost of capital. This methodology is described in the Appendix to Thakur and others (2002).
Prior to the 2000 reform, which is generally believed to have increased the MECT but lowered the AECT (see, for instance, Keen, 2002).
Comprising the standard rate of 37 percent and IRAP of 4.25 percent.
Statutory and Effective Rates of Corporation Tax in the EU
| Statutory Rate of Corporation Tax 1/ | MECT 2/ | AECT 3/ | ||
|---|---|---|---|---|
| Austria | 34.0 | 20.9 | 38.4 | |
| Belgium | 40.2 | 22.4 | 44.5 | |
| Denmark | 32.0 | 21.9 | 36.8 | |
| Finland | 28.0 | 19.9 | 32.4 | |
| France | 40.0 | 33.2 | 47.5 | |
| Germany 4/ | 52.35 | 31.0 | 57.7 | |
| Greece | 40.0 | 18.2 | 43.3 | |
| Ireland | 10.0 | 11.7 | 13.0 | |
| Italy | ||||
| Before reform | 41.255/ | -4.1 | 40.7 | |
| After reform | 33 | 17.9 | 34.3 | |
| Luxembourg | 37.45 | 20.7 | 41.5 | |
| Netherlands | 35.0 | 22.6 | 39.8 | |
| Portugal | 37.4 | 22.5 | 42.0 | |
| Spain | 35.0 | 22.8 | 39.8 | |
| Sweden | 28.0 | 14.3 | 31.0 | |
| United Kingdom | 30.0 | 24.7 | 35.7 | |
Including surcharges and average local taxes.
From Tables 7 and 8 of European Commission (2001), except postreform figure for Italy, kindly provided by Silvia Giannini.
Assuming a pretax return of 20 percent, and calculated as AECT = τ + (MECT)x (1-τ)C /0.2, where τ is the statutory rate of tax and C the user cost of capital. This methodology is described in the Appendix to Thakur and others (2002).
Prior to the 2000 reform, which is generally believed to have increased the MECT but lowered the AECT (see, for instance, Keen, 2002).
Comprising the standard rate of 37 percent and IRAP of 4.25 percent.
Statutory and Effective Rates of Corporation Tax in the EU
| Statutory Rate of Corporation Tax 1/ | MECT 2/ | AECT 3/ | ||
|---|---|---|---|---|
| Austria | 34.0 | 20.9 | 38.4 | |
| Belgium | 40.2 | 22.4 | 44.5 | |
| Denmark | 32.0 | 21.9 | 36.8 | |
| Finland | 28.0 | 19.9 | 32.4 | |
| France | 40.0 | 33.2 | 47.5 | |
| Germany 4/ | 52.35 | 31.0 | 57.7 | |
| Greece | 40.0 | 18.2 | 43.3 | |
| Ireland | 10.0 | 11.7 | 13.0 | |
| Italy | ||||
| Before reform | 41.255/ | -4.1 | 40.7 | |
| After reform | 33 | 17.9 | 34.3 | |
| Luxembourg | 37.45 | 20.7 | 41.5 | |
| Netherlands | 35.0 | 22.6 | 39.8 | |
| Portugal | 37.4 | 22.5 | 42.0 | |
| Spain | 35.0 | 22.8 | 39.8 | |
| Sweden | 28.0 | 14.3 | 31.0 | |
| United Kingdom | 30.0 | 24.7 | 35.7 | |
Including surcharges and average local taxes.
From Tables 7 and 8 of European Commission (2001), except postreform figure for Italy, kindly provided by Silvia Giannini.
Assuming a pretax return of 20 percent, and calculated as AECT = τ + (MECT)x (1-τ)C /0.2, where τ is the statutory rate of tax and C the user cost of capital. This methodology is described in the Appendix to Thakur and others (2002).
Prior to the 2000 reform, which is generally believed to have increased the MECT but lowered the AECT (see, for instance, Keen, 2002).
Comprising the standard rate of 37 percent and IRAP of 4.25 percent.
Rates of Taxation Under Nordic Dual Income Tax Systems, 2000
Highest marginal rate, including both national and regional taxes.
Rates of Taxation Under Nordic Dual Income Tax Systems, 2000
| Corporate Income Tax | Interest | Dividends | Labor Income 1/ | |
|---|---|---|---|---|
| Finland | 29 | 29 | 0 | 50.13 |
| Norway | 28 | 28 | 0 | 49.3 |
| Sweden | 28 | 30 | 30 | 57 |
Highest marginal rate, including both national and regional taxes.
Rates of Taxation Under Nordic Dual Income Tax Systems, 2000
| Corporate Income Tax | Interest | Dividends | Labor Income 1/ | |
|---|---|---|---|---|
| Finland | 29 | 29 | 0 | 50.13 |
| Norway | 28 | 28 | 0 | 49.3 |
| Sweden | 28 | 30 | 30 | 57 |
Highest marginal rate, including both national and regional taxes.
110. There is no doubt that the reform has substantially increased the MECT (averaged over the different ways in which an investment can be financed)—and this is so even assuming full elimination of the IRAP. Prior to reform, the MECT was significantly negative, reflecting in particular the deduction under the super–DIT not merely of equity investments but of those investments uplifted by 40 percent. Indeed Italy had the lowest MECT—and to that extent, the strongest tax incentives for investment—of any EU country. With the removal of the DIT, and of imputation (which reduces the cost of investments financed by new equity, as explained below), the MECT is now firmly positive—although still moderately low by EU standards. The discouragement to investment through this route is plain.
111. The AECT, on the other hand, does seem to be noticeably reduced by the reform. This effect is probably over–stated by the inclusion of the IRAP in the prereform calculation 74 with an implicit assumption that whatever replaces it will not affect effective tax rates), but also reflects the substantial fall in the statutory rate of tax, which is a more central determinant of the AECT—especially for very profitable projects—than it is of the MECT. The reform thus seems to do little harm to the attractions of Italy as a site for foreign direct investment, and may do some good.
112. The fall in the statutory rate of corporation tax—which has long been higher in Italy than in most EU countries, and indeed has been increased there while it fell elsewhere—may have beneficial effects even apart from any impact on investment. This is because it reduces the incentives to use transfer pricing and financial arrangements to artificially shift profits to jurisdictions offering lower tax rates. But while this should strengthen Italian corporate tax revenues, it is unclear how strong the effect will be: there remain many jurisdictions offering still lower statutory rates (not least Ireland) through which paper profits can be routed.
113. Another area in which the reform is also likely to have a significant impact is on the way in which firms finance themselves. The elimination of the deduction for the implicit cost of equity finance makes retention finance more expensive, and the replacement of imputation by partial taxation of dividends at personal level is likely to raise the cost if finance by new share issues. The net impact is thus to increase the attraction to firms of financing themselves by debt rather than equity.75
114. In abolishing imputation and instead simply taxing dividends at the personal level (at a reduced rate), Italy is following the recent lead of the United Kingdom, Germany and Ireland. And for the same reasons. The merit of imputation is that it mitigates the disincentive to new share issues that the taxation of dividends otherwise implies.76 In an economy open to capital movements, however, this becomes a less significant concern, since the marginal shareholder—who determines how much the firm must earn on its investments—may be resident abroad, in which case the imputation credit is generally not available. While it would be possible to overcome this limitation by extending the credit to nonresidents—and Italy does this under some of its double tax treaties—the revenue cost of doing so may well exceed the efficiency gains. Moreover, imputation itself offers opportunities for tax reduction strategies: indeed it is in principle possible under a scheme like the former Italian one to avoid all tax on corporate earnings.77 Also, given that there have been signs in recent years that imputation systems which do not extend credits to all EU residents may be in violation of Union rules on nondiscrimination, a strong case can be made for moving away from imputation.
115. The adoption of participation exemption is also in line with developments elsewhere in the EU, most prominently by Germany in its landmark 2000 reform. It has a strong tax policy rationale, the exemption of intercorporate capital gains being akin to the normal exemption of intercorporate dividends.78 Not adopting participation exemption might also run the risk of firms (and tax base) relocating to countries that do.
116. The prospective consolidation provisions should provide a useful rationalization in an important aspect of corporate taxation. Under the prereform system, firms could already enjoy many of the benefits of consolidation—but the means of doing so were needlessly complex, and the opportunities may even have been overgenerous. Apart from establishing some coherence, clear rules enabling consolidation may serve to overcome some of the inefficiencies widely felt to be associated with the small firm size in Italy.
117. It is in the area of international taxation, however, that the proposed consolidation rules may have their most marked effect, since they imply a fundamental change in the tax treatment of Italian subsidiaries abroad. Under the prereform regime, the earnings of such companies located elsewhere in the EU were essentially exempt from Italian tax. For profitable subsidiaries, exemption is more favorable than the proposed treatment, which would imply that tax is ultimately paid at the Italian rate (even on undistributed earnings). In respect of loss–making subsidiaries elsewhere in the EU, on the other hand, the ability to use those losses to reduce Italian tax is a source of gain from the reform. For subsidiaries located outside the EU, but subject there to a sufficiently high rate of tax to enjoy the exemption of 60 percent of dividends, the change implies heavier taxation of profits wherever the foreign tax rate exceeds about 23 percent.79 The implications are thus complex, and will prompt extensive tax–replanning. In some important respects, however, the effect seems likely to be to discourage Italian direct investment abroad.
D. Taxation of Interest and Capital Gains
The reforms
Before reform
118. Italy has a long tradition of taxing interest income by final withholding levied at flat rates.80 An advantage of this approach—which differs from that of the conventional “global” income tax, which applies a progressive rate structure to the sum of all kinds of income—is that it does not rely on taxpayers identifying their interest income to the authorities. These “substitute” taxes applied only to forms of interest income specifically listed in the statute, and were for many years levied at rates that varied quite substantially across assets. By 2001, however, only two distinct rates applied: 27 percent on bank and post office accounts, and on long–term bonds; and 12.5 percent on other debt instruments.
119. In 1998, Italy adopted an innovative approach to the taxation of capital gains. The conventional approach is to tax gains as they are realized—that is, when the underlying asset is sold. This though leads to a “lock–in” problem: investors holding an accrued gain have an incentive to defer selling the asset since they can thereby postpone their tax liability (and, conversely, investors holding an accrued loss have an incentive to realize it quickly). The 1998 reforms in Italy represented the most thorough attempt that any country has yet made to mitigate this distortion of holding period decision. It involved two different methods applied to different classes of asset.81 For managed funds, capital gains were “marked to market,” meaning that tax is charged on all gains that have accrued within the tax year, whether or not they had actually been realized. In this case there is, in principle, no lock–in. For other gains, tax is charged at realization but with an adjustment to broadly offset the advantage likely to have been enjoyed by deferring realization of the gain. The intellectual origins of this latter approach—the “equalizer”—date back to Vickrey (1939); but this was the first significant practical application. Capital gains, so calculated, were subject to the same two rates of substitute tax as interest income, depending on the nature of the asset (with the 27 percent rate applying for holdings that represent a significant share in the companies’ total equity).
After reform
120. All financial income is to be taxed at 12.5 percent, eliminating the higher rate.
121. The equalizer was abolished in September 2001. Marking to market for managed funds remains for reasons discussed below, but the government has announced its intention to remove this too. All gains are thus to be taxed on realization.
Assessment
122. Unifying the rates of substitute tax removes a clear source of distortion, for which there was no clear rationale, completing a process of convergence that has been underway for some years. The main issue that it poses relates to the gap between this common rate and the rate of corporation tax, and is discussed in Section F below.
123. The case for undoing the innovative capital gains tax regime is less clear–cut. Though far from being theoretically ideal—both the equalizer and the mark–to–market schemes treated gains and losses asymmetrically, for example—this was an attempt to address what is widely seen as one of the main weaknesses of the usual approach to taxing capital gains.
124. The equalizer suffered from being intrinsically complex—or at least unfamiliar. For many taxpayers, the adjustment made to offset the benefits of deferral was far from simple, even for tax practitioners; moreover, the precise nature of that adjustment varied over different assets. Clearly unpopular—no doubt in part because of its complexity, but also because it removed a tax advantage—the equalizer also faced legal challenges, the argument being that it violated notions of ability to pay enshrined in the constitution. Nevertheless, substantial costs (of, for example, developing appropriate information systems) were incurred in implementing the system, perhaps in the order of Є200 million.82 With time and some educational effort, taxpayers might have become more comfortable with the system.
125. Marking to market is, if anything, theoretically preferable to the equalizer scheme. Its introduction happened to come, however, at an unfortunate time, being followed by a prolonged depression of share prices. This has exposed an unforeseen technical problem with the scheme that will be of interest to other countries considering more widespread use of marking to market; and which, paradoxically, now makes it more difficult to remove.
126. The problems arises from the substantial accumulation of losses by managed funds since the introduction of the scheme. Contrary to the textbook prescription, these do not give rise to an immediate payment by the government, but instead to tax losses that are carried forward without interest. While this in itself reduces the present value of the credits below that required for symmetric treatment of gains and losses, the effect is exacerbated by the practice of paying the value of the credit in full to those who withdraw their monies from the fund. In this way, withdrawals reduce the value of the fund’s assets by the difference between the nominal value of the tax credit and its present value in the hands of the firm. Given the extensive amount of the credits now outstanding—some industry sources put this at Є5 billion—and restrictions on the ability to use tax losses on one fund against gains on another, this difference can be substantial.
127. The problem is not intrinsic to marking–to–market. It would be avoided if the credit received by those withdrawing funds were paid by the government (and so extinguished) rather than being paid from the fund and carried thereafter on its books, or if losses in the fund were given a immediate tax credit symmetrically with the tax paid on gains. The more general feature of accruals taxation to which it points is the potential pro–cyclicality of revenues, reinforcing automatic stabilizers but raising the question of the extent to which governments are willing and able to deal with the associated risks.
128. While the underlying design problem can in principle be fixed—and arguably should be in respect of future withdrawals—the overhang that has been built up is so large that the government would be reluctant to cash out all outstanding credits. Indeed there is no obvious case for giving such priority to these over other unrelieved losses. But there may be a case for increasing the value of these credits, perhaps by increasing the extent to which they can be used against gains other than on the fund concerned, or by carrying them forward at interest.
E. The IRAP
The reform
Before reform
129. Another innovation of the late 1990s was the introduction of the IRAP (Imposta regionala sulle activita produttive), payable by business on, broadly speaking, the amount by which their sales exceed the sum of their material purchases and depreciation. The base is thus essentially the firm’s value–added, so that—although not popularly recognized as such—the IRAP is a particular kind of VAT. It is, to be precise, an origin–based income–type VAT administered by the subtraction method.83 Note too that, although not calculated as such, the identity between the firm’s sources and uses of funds means that the base of the IRAP is, in essence, the sum of wages and profits.
130. The IRAP is a regional tax (albeit collected by the state),84 and replaced a local income tax,—from which partnerships, professionals and farmers were generally exempt—a tax on net assets, and other charges. The central rate is 4.25 percent, but regions can vary this, in either direction, by 1 percentage point, and may differentiate the rate by sector. Significant use is now made of this power, typically with lower rates applied to agriculture and higher rates to financial activities.
131. Revenue from the IRAP is substantial, in the order of Є30 billion (about 2.5 percent of GDP), and financing about one–quarter of all regional spending.
132. No other country applies a tax of this kind at regional level.85
The reform
133. The new government has made a strong commitment to eliminate the IRAP. The first step is to be taken in 2003, with 20 percent of labor costs excluded from the base. It still remains to be decided how the revenue raised by the IRAP will be replaced.
Assessment
134. The IRAP has, in principle, some attractive features. It is simpler than the mix of taxes it replaced. The record–keeping requirements—the essential requirement being the difference between receipts and purchases—are fairly minimal, and indeed already likely to be required for the IRPEG or IRPEF. And the potential base is broad, enabling a relatively low rate. Viewed as a regional tax, moreover, it could be argued to have merit as a benefit tax—in so far as local public spending is reflected in increased private profits and/or wage earnings—and as bearing on factors, especially labor, that will be relatively immobile in response to likely interregional differences in the tax rate. The IRAP even received favorable treatment under double tax arrangements, with that part of the tax corresponding to a tax on profits being regarded as creditable, for instance, against corporation tax due in the United States (which is not true of the corresponding component of the base of the national VAT). Not least, the IRAP succeeded in raising considerable amounts of revenue.
135. One weakness of using the subtraction method for a regional VAT is that interregional differences in the rate of tax will lead to the tax bearing in part on intermediate purchases by firms. A firm in a region requiring it to charge IRAP at 4 percent is also in effect obtaining a credit of 4 percent on its purchases, so that if it buys from a firm in a region charging 5 percent some of the input tax will “stick.” This means that the IRAP would act to some degree as a turnover tax, so distorting production decisions: the firm might find it preferable to buy from a firm that is less efficient but is in a region charging a lower tax rate. With interregional rate differentials in the order of 2 percent, however, the effect was unlikely to be severe.
136. Despite these apparent attractions, the IRAP seems to have proved unpopular. It is not entirely clear why. Part of the reason may be that it brought into tax groups that had previously enjoyed significant exemptions from regional taxation. It also came to be seen as largely another tax on labor, as indeed it is—but so too, in essentially the same way, is the national VAT, which is not often criticized on these grounds.
137. The key unanswered question, however, is that of how the IRAP will be replaced. Either other forms of regional taxation will need to be developed or grants from the center will have to be increased—neither of them easy options. Further uncertainty is added by a recent legal decision implying that the regional governments have substantially more discretion in the taxes they deploy than had previously been supposed.86
F. Toward a Nordic Dual Income Tax?
138. A striking and problematic aspect of the Italian tax system is the wide gap between the rate of corporation tax and the rate of tax on financial income. Though already present in the prereform system, this feature is highlighted by the unification of the rate of tax on financial income at 12.5 percent, some 20 percentage points below the corporate tax rate. Such a gap is problematic because it creates opportunities for riskless tax arbitrage, with the private sector effectively able to extract payments from government. By lending Є100 to an Italian corporation at an interest rate of 10 percent, for instance, an individual will become liable to tax of Є1.25 on interest received; but in deducting the interest payment of Є10 against corporation tax at a rate of 33 percent the company reduces its tax liability by Є3.3. The net effect of the transaction is thus to improve the cash position of the private sector by Є2.05, all at the expense of government.
139. The scope for such arbitrage within closely–held companies is clear, and indeed rules are envisaged to restrict transactions of this kind between related parties. But exactly the same effect can be achieved by anonymous transactions through financial intermediaries. The loan described in the previous paragraph, for instance, could be run through a bank, with the de facto government subsidy of Є20.5 shared between the bank, lender and borrower. With such a clear opportunity for essentially riskless profits, one must expect the market to find some way to realize them. Nor need the exploitation of such arbitrage even be consciously recognized by market participants: it is simply a matter of decisions responding to patterns of demand and supply for funds reflecting these tax arrangements.
140. The natural way to eliminate this difficulty is by unifying the rates of tax applied to corporate and financial income. This would bring the further advantages, moreover, associated with a reduction in the IRPEF rate in itself, such as a further reduction in the effective rates of corporation tax and an easing of transfer pricing problems. Very rough calculations suggest that the two rates could be unified, whilst raising the same revenue as the anticipated postreform system, at a common rate of about 25 percent.
141. Such an alignment of rates would take the Italian system very close to a “dual income tax” of the form adopted by most Nordic countries since the early 1990s. The essence of the Nordic approach—rationalized by the increasing difficulty of taxing internationally mobile capital income—is to apply a progressive tax to labor income and a flat, low rate to all forms of capital income. This is a sharp break from the tradition of most OECD economies of treating capital and labor income alike, applying a “global” progressive tax schedule to the sum of the two. In Italy, however, there has been no such tradition, so that a Nordic dual income tax would fit well with the established practice of subjecting personal capital income to final withholding at low, flat rates.
142. Experience with the Nordic dual income tax has been broadly positive.87 Perhaps the thorniest problem that has been encountered is in the treatment of the self–employed—a particularly important group in Italy—the difficulty being to distinguish between that part of their income which is properly regarded as a return on capital and so to be taxed at the flat rate and that part which is properly taxed as labor income. Essentially the same issue arose in Italy under the DIT, since individuals’ business income was also taxed at a reduced rate in so far as it reflected a notional return on equity injected since 1996. 88 To minimize the compliance burden that this calculation might imply, small businesses could opt instead to be taxed at a flat rate of 15 percent. The same approach could of course be adopted again; and provision is any case made for the use of presumptive methods in assessing small enterprises. More fundamentally, however, the extent of this problem would be considerably less given a unified rate of the magnitude that the calculations above suggested, since the rate on capital income would then be very close to the 23 percent rate on labor income that the vast majority of individual taxpayers will face.
G. Conclusion
143. The program of tax reform now underway in Italy will involve a sizable cut in the tax burden that is likely to support employment growth. But, in addition, the reform represents a fundamentally different approach to tax policy compared to that followed by the previous government. The latter, with the introduction of the DIT, the equalizer and use of marking–to–market, and the IRAP, had focused on reducing a number of tax distortions (the bias in favor of debt finance, the lock–in problems of the capital gains tax, and the dependence of regional governments on a variety of taxes with relatively narrow bases). The new approach, while reintroducing a number of distortions, focuses more on establishing a system that is both simpler and closer to those found elsewhere in the EU.
144. The move towards a simpler system has clear merits. The personal income tax, for instance, is in clear need of restructuring and simplification, and the introduction of consolidation provisions promises a useful rationalization of the taxation of businesses. However, it is not clear that the reforms of the late 1990s, which had the merit of reducing tax distortions, involved any insurmountable technical problem, or that taxpayers were unable to come to terms with the various novelties with which they were faced. Careful management of the transition to the new system will be especially important given that this reform comes so soon after those of the late 1990s.
References
Alworth, Julian, Giampaolo Arachi, Roni Mamui and Roberto Viola 2002, “The Taxation of Income from Capital: 1990–2001,” Mimeo, Università Luigi Bocconi.
Auerbach, Alan, 2001, “Taxation and Corporate Financial Policy,” NBER Working Paper No.8203.
Bordignon, Massimo, Silvia Giannini and Paolo Panteghini, 2001, “Reforming Business Taxation: Lessons From Italy?” International Tax and Public Finance, vol.8, 191210.
Baldini, Massimo, Paolo Bosi and Massimo Matteuzzi, 2002, “Effetti Distributuvi della Proposta di Reforma dell’Irpef: Prime Valutazioni “in corse d’ opera,’” Mimeo, CAPP, Università di Modena e Bologna.
Brunello, Giorgio, Maria Laura Parisi and Daniela Sonedda, 2002, “Labor Taxes and Wages: Evidence from Italy,” Mimeo, University of Padova.
Cnossen, Sijbren, 2000, “Taxing Capital Income in Nordic Countries,” in Taxing Capital Income in the European Union: Issues and Options for Reform, ed. by Sijbren Cnossen, PP. 180–213. (Oxford University Press).
Deidda, Antonia, and Christian Grabbe, 2002, “The Reform of the Italian Tax System,” in International Tax Review, vol.30, pp. 307–14.
Ebrill, Liam, Michael Keen, Jean-Paul Bodin and Victoria Summers, The Modern VAT, (International Monetary Fund: (Washington D.C)
European Commission, 2001, Company Taxation in the Internal Market (European Commission: Brussels).
Goolsbee, Austan, 1999, “Evidence on the High–Income Laffer Curve from Six Decades of Tax Reform,” Brookings Papers on Economic Activity, vol.2 pp. 1–47.
Holmlund, B. and A.-S. Kolm, 1995, “Progressive Taxation, Wage Setting and Unemployment—Theory and Swedish Evidence,” Swedish Economic Policy Review, vol.2 pp. 423–460.
Keen, Michael, 2002, “Tax Reform in Germany,” International Tax and Public Finance, forthcoming.
Keen, Michael, and Fabio Schiantarelli, 1991, “Corporation Tax Asymmetries and Optimal Financial Policy,” Oxford Economic Papers, vol.43, pp. 280–291.
Lindhe, Tobias, Jan Södersten and Ann Öberg, 2002, “Economic Effects of Taxing Closed Corporations under a Dual Income Tax,” Mimeo, Uppsala University.
Lockwood, Ben, and Alan Manning, 1993, “Wage Setting and the Tax System: Theory and Evidence from the United Kingdom,” Journal of Public Economics, vol.52 pp. 1–29.
Petretto, Alessandro, undated, “La Delega sulla Riforma Fiscale e la Finanza degli enti Decentrati,” Mimeo, Università di Firenze.
Sørensen, Peter Birch, 1994, “From the Global Income Tax to the Dual Income tax: Recent Tax reforms in Nordic Countries,” International Tax and Public Finance, vol.1, pp.57–79.
Thakur, Subhash, Michael Keen, Balaze Horvath, and Valerie Cerra, 2002, The Swedish Welfare State: Can the Bumblebee Keep Flying? (International Monetary Fund: Washington D.C.), forthcoming.
Vickrey, William, 1939, “Averaging Income for Income Tax Purposes,” Journal of Political Economy, vol.39, pp. 379–97.
Prepared by Michael Keen.
The guiding principles of the reform, and some specific objectives, are set out in a legge delega (still not yet finally approved by parliament) giving the government power to reform the tax system along the lines it describes. A subsequent Ministry of Finance’s technical report on the legge delega provides somewhat more detail, especially on the expected revenue implications. The medium–term financial program published in July 2002 outlines the tax changes intended for 2003.
The focus is on those aspects of the package that are most fundamental to the overall tax system, so that several significant but relatively minor items—such as the introduction of a tonnage tax and a unified service tax—are not considered here. Some are discussed in Deidda and Grabbe (2002). Nor does this chapter discuss recent tax amnesty schemes, though the associated erosion of the credibility of tax policy may damage the credibility of the wider reform program.
See Petretto (undated), who indicates that Veneto and Marche actually set rates in excess of the maximum 1.4 points allowed.
See, for example, Goolsbee (1999).
Kindly provided by Paulo Bosi.
See, for instance, Lockwood and Manning (1993) on the United Kingdom; and Holmlund and Kolm (1995) on Sweden.
The abolition of the inheritance and gift tax by the new government will tend to reinforce this increase in inequality.
Denoting pretax income by Y and expenditure attracting credit by E, net income under the prereform system is T (Y)—C (Y,E), where C denotes the credit received. With a postreform tax schedule of τ and allowances of A, net income is τ(Y - A (Y,E)). Setting A (y,E) = τ-1 [T (y)-C (y,E)], net income is the same in both cases.
Confusingly, the terra “dual income tax” is used in two quite different senses: here it refers to there being two rates of corporate taxation; more usually (as in Section F) it refers to distinct tax treatment of capital and labor income along the lines of the Nordic model. The system described here would be better described as a “partial Allowance for Corporate Equity (ACE)”, an ACE being a system that entirely excludes the normal return on equity from tax.
By this is meant, here and elsewhere, the system as it existed at the election of the current government.
The system applied to banks and financial institutions only from 2000.
The notional return on equity was fixed annually by the Ministry of Economy on the basis of state and private bond returns plus a risk allowance of 3 percent. Latterly, it was 7 percent.
Brazil and Austria have similar schemes (with differences in the precise definition of the equity base on which an imputed return is allowed); between 1994 and 2000, Croatia took this approach to its logical conclusion and adopted a full–blown ACE.
Until 2001, there was a minimum average tax rate of 27 percent.
The uplift was 20 percent in 2000.
The grandfathering provisions are asymmetric: while the equity allowance is not increased by subsequent increases in equity it is reduced by subsequent reductions (associated, for example, with the repurchase of shares).
With a corporate tax of 36 percent, for instance, the credit was 56.25 percent (=0.36/(1-0.36)) of the dividends received,
The comparison is between 12.5 percent withholding and, on the other hand, additional personal tax at the rate Tp on the underlying gross dividend (1/(1 - 0.36) less the imputation credit of 0.36/(1 - 0.36), an amount of (Tp-0.36)/(l -0.36).
The benefit is denied, however, in respect of shares held for less than a year. Some such provision is standard, as a means of ensuring that those dealing in shares by way of trade do not escape tax on their capital gains.
With no credit for foreign taxes paid on the underlying income source.
Precise definitions (and hence properties) of the AECT vary: see the Appendix to Thakur and others (2002).
These calculations do not take account of the Visco and Tremonti incentives.
This is dictated by the use here of the calculations in European Commission (2001), which have the merit of being comparable across countries. Conceptually, inclusion of the IRAP in these calculation is not attractive: if it is to be included, so should national VATs.
Giannini estimates, for instance, that the MECT for equity financed investments will be 31.5 percent after the reform, but -30.7 percent for debt finance.
The tax treatment of dividends does not affect the incentive to retain earnings (so long as that treatment does not change over time): by retaining, a company avoids the dividend tax today, but must pay it in the future when it pays the dividends without which the firm would have no intrinsic value to shareholders. (On this (no longer very) “new” view of dividend taxes, see Auerbach, (2001).
See Keen (2002) for further discussion of the case for participation exemptioa
Assuming all profits net of foreign tax to be distributed, tax paid under present arrangements (leaving aside any withholding tax) is Tf + Ti(0.4)(l-Tf), where Tr and Tj denote respectively the foreign and Italian tax rates; under consolidation, it simply TI The calculation assumes Ti = 0.33.
For companies, interest is taxable at ordinary rates and the withholding tax is creditable.
Alworth and others (2002) provide a detailed description and assessment of these schemes.
It is origin–based in the sense that exports are not relieved of the IRAP nor imports brought into it, and income–type in the sense that sense that an effective deduction is not given for the immediate full cost of an investment but only for depreciation. On this terminology, and alternative types of VAT, see Ebrill and others (2001).
For companies active in several regions, revenues are allocated with reference to the proportion of total labor costs incurred in each.
Japan implements a subtraction–based VAT at national level. In the United States, Michigan and New Hampshire have state–level VATs, but levied on an addition basis (that is, on the sum of wages and profits). While some countries (notably Germany) share the revenue from national VAT across the regions, no gives the regions discretion over rates or base of the kind enjoyed under the IRAP..
See Box 2 the Staff Report.
See, for example, Sørensen (1994) and Cnossen (2000).
Income subjected to this reduced rate was however taken into account in determining the tax payable on the labor component of income.