Categoría del artículo: Article
Publicado en línea: 19 may 2025
Páginas: 95 - 112
Recibido: 19 ene 2024
Aceptado: 03 may 2024
DOI: https://doi.org/10.2478/ntaxj-2024-0003
Palabras clave
© 2024 Frederik Zimmer, published by Sciendo
This work is licensed under the Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License.
The relationship between the taxation of the company and the taxation of its shareholders is a core issue in tax policy and, at the same time, a difficult one. The main pattern is that the company is taxed currently for its income, whereas shareholders are taxed for dividends when they are distributed. However, how this is done varies considerably, historically and in a comparative perspective. All kinds of solutions appear, from taxing both the company and the dividends without any special rules—often referred to as the “classical system”—to full integration of the taxation of companies and shareholders; in other words, from full economic double taxation to single taxation of distributed company profits. Most systems today seem to be somewhere in between these two extremes. If integration is wanted, taxation can be reduced at the company or the shareholder level. It is most common to reduce the taxation of the shareholder, not least because by a reduction of the taxation of the company, such as a deduction for distributed dividends, it would be difficult to carry out a corresponding taxation of non-resident shareholders.
Searching for the rationale behind such integration rules leads to a more fundamental question: why tax companies at all? The answer to this question may have an impact on the relationship between the taxation of the company (provided it is taxed) and its shareholders. Historically, it has sometimes been assumed that there are certain benefits connected with doing business in the form of a company (the company as a juridical person, the benefits of shareholders’ limited responsibility for the company’s debt) and that these benefits could be a basis for taxing companies as such, leading to economic double taxation.
However, this so-called benefit principle is not in high-esteem these days because there is normally no clear relationship between the assumed benefits and the income of a company. Instead, modern economic theory assumes that the company tax in the end is born by individuals. So why not tax the shareholders instead of the company? This could be done in at least two ways. Firstly, the shareholders could be taxed currently, as is the general rule for partnerships and under controlled foreign company rules (CFC rules). Such rules can be carried out in practice where there are only a few shareholders, but this would be much more difficult with many shareholders and in particular when the shares are traded a lot (typically on a stock exchange). It would also be difficult to tax nonresident shareholders currently. Secondly, company income could be taxed on the shareholder level at the time of distribution of dividends. This method, however, would provide the taxpayers with the option to postpone taxation of company income as long as they can and wish, and it would create lock-in effects. It would also raise transfer pricing and related issues (shareholders could benefit from the company in other forms than through dividends). Taxing foreign shareholders would be difficult for both legal (tax treaties) and practical reasons (NOU 2014).1 From this perspective, the company tax can be regarded as a pre-payment of the shareholders’ taxes. In addition, there is the question of the incidence of company tax. The reasoning above assumes that the company tax is in the end borne by the shareholders. However, that may not be so, wholly or partly. A generally accepted view is that a company tax is not born by the company’s shareholders alone but wholly or partly by others, notably the employees of the company (through lower salaries) or customers (through higher prices, if possible).2
The focus of this article is how the relationship between company and shareholder taxation is dealt with in current Norwegian tax law compared to rules in other Nordic countries. It turns out that the Norwegian rules differ from those in other Nordic countries not only in their practical design but also in the basic considerations on which they are based. Therefore, the article attempts to provide a critical analysis of these considerations as a basis for legal policy discussions. The ambition is not to make conclusions concerning the Norwegian rules nor to discuss alternatives.
The next section provides an overview of the company/dividend taxation in Nordic countries. Thereafter, Norwegian rules on the shielding allowance and its rationale are presented and investigated to answer the question asked in the article’s title and provide a basis for legal policy discussions.
This article discusses situations where the shareholder is an individual. For company shareholders other considerations apply; this is briefly explained in section 2.5.1 below.
This implies that the total of the company and the dividend tax amounts to 44.98 percent (41.05 percent in the case of non-listed companies). This is considerably lower than the top marginal tax rate for earned income, and therefore an additional set of rules apply in order to prevent income shifting from earned income to capital income. These rules apply when the shareholder (or a relative) is active in the company to a considerable extent. If this is the case, dividends and share capital gains exceeding certain limits and below certain maximum amounts are taxed as earned income.
In
On unlisted shares, the rules are rather complicated. First, it must be decided whether the dividend is above or under eight percent of the mathematical value of the shares. If the dividends are under eight percent of the mathematical values, then there is a difference depending on whether the dividend is under or above 150,000 euro. If the dividend is under that amount, then 25 percent of the dividend is taxable as capital income (where the tax rates are the same as for capital income on listed shares). The remaining 75 percent is tax-free. The 150,000 euro limit applies to all dividends from unlisted shares that a taxpayer receives on different unlisted shares in the same year. When the dividend amount exceeds 150,000 euros, then 85 percent of the dividend is taxable capital income, whereas 15 percent is tax-free, parallel to what applies to dividends from listed shares.
If the dividend is higher than eight percent of the mathematical value, the same rules as described above apply to the dividend up to eight percent of the mathematical value. For dividend amounts exceeding the eight percent limit, 75 percent is regarded as earned income and taxed according to progressive tax rates. The remaining 25 percent is tax-free.
The pattern that emerges is that dividends are taxable as capital income, but some of them are tax-free (the higher the dividend amount). For high dividends, 75 percent is regarded as earned income to prevent income shifting.
In
It is important to underline that the Norwegian system differs in principle from the Danish even if they both operate with two tax rates for dividends– Denmark 27 and 42 percent, Norway zero and 37.84 percent. The Danish system is a traditional progressive tax rate system where the tax rate depends on the aggregate amount of an individual’s dividend income in a certain year. The Norwegian system, by contrast, refers to the income from every share regardless of the total amount of the individual’s dividend or other income, and it depends on whether and to what extent the dividend from the share in question exceeds the shielding allowance.
Further, the Norwegian system differs from the systems in Sweden and Finland. In Sweden, no share income is tax-free, but complicated rules apply to draw the borderline between capital income and earned income; in addition, a somewhat lower tax rate applies for capital income from unlisted shares. In Finland, as in Norway, parts of the dividend are tax-free but in very different ways. In Finland, a certain percentage of the dividend (varying with the circumstances) is tax-free. In Norway, the tax-free amount–the shielding allowance–is calculated as part of the yield of each investment.
On this basis, it is concluded that the Norwegian rules on dividend taxation differ significantly from the rules in the other Nordic countries. In addition, the Norwegian system seems to be unique also in a broader international comparison. This asks for a more detailed explanation of the system, how it works, and not least its rationale. This is the focus of section 3.
It is worth noting that this overview shows that the rules for taxing dividend distributions in the Nordics are deeply influenced by the so-called Nordic Dual Tax System, which is characterized by rather low and proportional tax rates on capital income and progressive tax rates on earned income. This system provides a possibility for shifting earned income to capital income by distributing dividends instead of paying (higher) salaries. The taxation of dividends may be used as an instrument to counter such income shifting, as will be illustrated in the description of the Norwegian rules in the following.
The current Norwegian rules went into effect in 2006. A brief historical approach is appropriate to understand their rationale and function.
Before the tax reform in 1991–which introduced the Nordic Dual Tax System in Norway—the rules on the taxation of companies and dividends were somewhat complicated, with different rules for municipal taxation and state taxation. Under municipal taxation, old rules from 1911 had been retained, according to which the company was taxed on its profits, and the dividend was tax-free in the hands of the shareholders. For state taxation purposes, however, dividends were taxable to the shareholders. The company was entitled to a deduction for the distributed income, but this was restricted to the company’s net income in that year.4 The state taxation rules were a single taxation system of distributed company profits for dividends distributed the same year, but they had elements of economic double taxation for other dividends.
The tax reform of 1991–in force from 1992–was first and foremost a company and capital tax reform. The company tax rate was reduced to 28 percent (down from ca. 50 percent) combined with a considerable widening of the tax base. Neutrality in capital taxation was considered important, and this resulted in a proportional and rather low tax rate (by the standards of that time) for capital income of 28 percent. However, the tax rates for earned income were still progressive, with a top marginal tax rate slightly below 50 percent. This tax rate structure followed the Nordic Dual Tax System. Neutrality in the taxation of capital income was considered to require single taxation of company profits. For this purpose, an imputation system was introduced. On certain conditions, shareholders were entitled to full imputation, and considering that the tax rate was the same for company profits and for dividend income–28 percent–the practical result was to make dividends taxfree in the hands of the shareholders.5
Such rules would make it extremely profitable to distribute dividends instead of paying (higher) salaries to employees who were also shareholders. To counter such tax planning, a complicated so-called division model was introduced (not only for companies but also for sole proprietorships). This model applied if at least 2/3 of the shareholders were active in the company. The company income was divided into a personal income part (assumed to reflect the value of work) and a capital income part. The personal income part was allocated to the active shareholders/employees according to their shareholding. Personal income was taxed according to the same progressive taxes as earned income, regardless of any distribution of dividends. This system was internationally unique at that time and not likely to have been copied later.6
After approximately ten years, this system broke down. There were mainly two reasons for this. Firstly, as expected in 1991, the division rules became the object of widespread tax planning and–most importantly– lobbying. The rules were liberalized (to make life easier for “family companies”), and the result was that it became rather easy to get around them.7 Therefore, income that should have been taxed progressively as earned income under the division model was considered capital income. Secondly, the development of EU/EEA tax law implied that the imputation rules (making distributed dividends in practice tax-free) could not be upheld. The reason was that the imputation credit was granted only to dividend distributions from companies resident for tax purposes in Norway to shareholders resident for tax purposes in Norway. Thus, they did not apply to cross-border dividends (both ways). Further, it was considered too complicated to extend the imputation rules to cross-border dividends.
Consequently, the imputation rules for dividends (and the parallel rules for share capital gains) had to be abolished, and the division model had to be reformed or replaced by something else to prevent income shifting.
This cleared the way for the introduction of the socalled
Firstly, the (re)introduction of the dividend tax would solve the income-shifting problem that the division model was intended to take care of.9 The sum of 28 percent company tax and 28 percent dividend tax (on the remaining 78 percent of the company profits) amounted to a combined tax of 48.16 percent, which was close to the level of the top marginal tax rate of earned income (exclusive of employers’ social security contribution). Thus, income shifting from earned income to capital income would not save much tax.
Even if this part of the reform implied that Norway went from single to economic double taxation of distributed company income (with some modification, see below), this fundamental change in approach was not focused upon in the preparatory works. The focus there was almost exclusively on the need to replace the division model with new rules to fight income shifting more efficiently.10 This is remarkable since the obligation for individual shareholders to pay taxes on dividends was made general, regardless of whether income shifting was a practical possibility. The duty to pay dividend tax was introduced also for listed shares and on parts in investment funds even if income shifting is virtually out of the question in such cases. This implies that the introduction of the dividend tax had wider functions than to counter income-shifting. Furthermore, these new rules reintroduced the difference in taxation between dividends and interest. Even if the shielding allowance reduced this difference somewhat compared with a full classical system, it would still be profitable to provide loans instead of more equity to the company. Therefore, to counter tax planning, a rule was introduced that stated that interest on loans from individuals (whether shareholders or not) to a company shall be taxed almost as dividends (economic double taxation with a shielding allowance).11
Secondly, the introduction of the so-called shielding allowance was based on neutrality considerations. Before discussing neutrality, it is necessary to provide a short outline of the rules of the shielding allowance.
The shielding allowance is supposed to reflect an approximate, risk-free yield of the shareholder’s investment in the shares. The allowance is calculated on each share in a company, even if the taxpayer owns more shares in the same company (the significance of this is explained below). The allowance amounts to a so-called shielding allowance basis multiplied with an approximately risk-free interest rate. As a point of departure, the shielding allowance basis is the taxpayer’s cost price on each share. Thus, a taxpayer who owns two or more shares in the same company may have different cost prices on the shares and, by implication, a different shielding allowance basis for each share. Since the rules are not perfectly symmetrical (see below), the shielding interest rate is slightly above the risk-free rate.
The shielding allowance is deducted from the dividend before calculating the tax.12 If the dividend amount is lower than the shielding allowance, the socalled unutilized shielding allowance can be carried forward to be deducted in future dividends on the same share or in capital gains on the sale of that share a later year.
In addition, the unutilized allowance can be added to the cost price of the share. This means, firstly, that the shielding allowance basis increases, which implies that the discounted value of the shielding allowance remains the same even if it has effect only a later year. Secondly, adding an unutilized shielding allowance to the cost price of the share means a reduction in the taxable capital gain on the sale of the share. However, to mitigate tax avoidance, there is one important exception: when a share is sold, the unutilized shielding allowance can only bring a taxable capital gain down to zero, as it cannot increase or establish a loss. This means that for the unutilized shielding allowance to be fully deductible, the taxable capital gain on the sale of the share must amount to at least the same sum as the unutilized shielding allowance on that share. If the capital gain is lower, then only parts of the unutilized shielding allowance will be effective. If the share is sold with a loss, the total unutilized shielding allowance loses its value without being utilized.
Furthermore, an unutilized shielding allowance cannot be transferred to another share that the taxpayer owns, not even to shares in the same company. The reason for this strict rule is the danger that without such a rule, many “year-end transactions” would emerge for tax planning reasons. The right to the shielding allowance is allocated (for practical reasons) to the individual who owns the shares at the end of the year. Not all taxpayers have the right to a shielding allowance, and this most notably applies to companies and most nonresident individuals. Therefore, it would be profitable tax-wise for companies and non-resident shareholders to sell their shares to resident individuals right before the end of the year and buy them back afterwards. Of course, such transactions can often be countered by the general anti-avoidance rule. However, this problem was considered so important that a stricter rule was needed. This, however, implies a lack of symmetry, which means that the shareholder cannot be certain that he/she will enjoy the shielding allowance in full. As already mentioned, this is compensated with a somewhat higher shielding interest rate than the risk-free interest rate (in practice, for many years, by adding 0.5 percent to the risk-free interest rate).13
In its pure form, these rules mean that the shareholder is guaranteed to uphold the economic value of the allowance. This implies, as economists use to underline, that a part of the risk connected to the investment is transferred from the shareholder to the state. This, in turn, implies that when a dividend tax with a shielding allowance is introduced, the investor can increase his investment in shares (for instance, by moving investments from bonds to shares), thereby upholding the same risk profile in his investment portfolio and by implication the expected profits connected with the risk profile, as before the dividend tax was introduced. This also means that the investor’s required rate of return will not increase as a result of the introduction (or enhancement) of a dividend tax with a shielding allowance.14 Furthermore, this means that even a high dividend tax (with a shielding allowance) will not lead to less investment in the shares or reduced access to capital for the company15 However, this does not apply in full in the current Norwegian system because an unutilized allowance cannot be transferred to other shares, as explained above.
This brings us to the hard questions, not least for non-economists, because the rationale for the shielding allowance is deeply embedded in economic theory. The core of the theory is that the allowance attains various forms of neutrality effects.
At the outset, it is worth noting again that the neutrality arguments discussed in the following apply only where the investor in question is an individual. In cases where the investor is a company, the participation exemption rules17 usually apply. Since there is no dividend tax in these cases, there are no effects of the kind that the shielding allowance shall compensate for in order to establish neutrality. In cases where the participation exemption does not apply to company investors a dividend tax is triggered. There is, nevertheless, no right to a shielding allowance.
The economic reasoning starts with realizing that a tax on dividends increases the financial costs of the company, which is not intuitively obvious. This is because a tax on dividends implies that a possible investor will require a higher dividend than without a dividend tax (the investor’s so-called required rate of return, in Norwegian
Of course, certain conditions must be fulfilled for this neutrality effect to take place. The reasoning here takes as a point of departure the consideration of an individual who plans to invest in a company. If the investment has the form of a loan to the company, the investor will receive interest income. Interest on the loan is deductible for the company (as a general rule) and taxable for the creditor/investor. Investment in shares leads to income in the form of dividends. Dividends are taxable to the shareholder and not deductible for the company. Therefore, without a shielding allowance, the yield of an investment in shares would be taxed both on the company and shareholder level. Thus, the taxpayer has an incentive to provide loans to the company instead of investing in shares. A shielding allowance will provide single taxation also for dividend income up to the value of the allowance (though the taxation takes place on the company level because the distribution of dividends is not deductible, whereas taxation of interest takes place on the investor/creditor level).
However, investors of course often plan for a higher yield than can be obtained by providing a loan. This requires investing in equity. Perhaps surprisingly, the theory implies that it is not necessary to extend the right to the shielding allowance to this higher yield. This is because a higher yield is difficult to obtain by providing loans. Thus, if the investor goes for a higher yield, providing a loan is not an alternative; the only alternative is to invest in shares. Therefore, the investor will have an incentive to invest in shares even if dividends exceeding the shielding allowance are taxed heavier than interest. Even after tax on dividends exceeding the shielding allowance, the income is higher than after-tax income from loans. The conclusion of this reasoning is that the shielding allowance does not need to cover the yield on equity resulting from the risk taken by equity investments as compared with providing loans. A shielding allowance based on the riskfree yield is sufficient to obtain investment neutrality and also financing neutrality seen from the company’s point of view because the dividend tax does not make it necessary to increase the dividend distribution to obtain investors in shares.18
These considerations throw some light over the different reasoning in 1991–when the argument was that all dividends received by individuals should be tax-free in order to reach single taxation of distributed company profits– and the argument introduced in 2004 that dividends should be tax-free only to an amount reflecting the riskfree yield of the investment. In 1991, the perspective was that
This reflects a shift in, and a narrowing down of, the neutrality concept. The rules of 2004 focus on the neutrality of financing the company with equity or loans. The 1991 rules, however, focused on the neutrality of all capital income, also including income from investments outside the companies (such as in real estate). For instance, if a dividend tax is introduced or increased in an investment in shares under a shielding allowance regime, it may nevertheless be profitable over providing loans, as explained above. However, the increased dividend tax–and, by implication, the lower profits after tax–may induce the investor to invest in something other than companies, such as directly in real estate. As for a yield exceeding the shielding allowance, an investment in shares is taxed both on the company and the shareholder (investor) level, whereas a direct investment in real estate is taxed only on the investor level. Even if also the risk-free yield on a direct investment in real estate is taxed, the investor may nevertheless find such an alternative more profitable.19
In sum, the reintroduction of the tax on dividends has had the effect of not only bringing back the different tax treatment of equity and loans,20 but it has also led to different taxation of dividends and other forms of capital income more generally.
It is generally assumed that the reasoning above concerning the effect of dividend tax does not apply to all financing of companies. Of course, the reasoning does not apply if the company funds itself with loans or with its own retained profits. Further, it is commonly agreed that it does not apply to all equity financing either. To the extent that the company funds itself in international capital markets, the cost of capital is decided in international markets and is therefore not affected by the dividend tax in a small economy like Norway’s. Therefore, even if the Norwegian dividend tax increases the required rate of return for Norwegian investors, this will not affect the financial costs of the company because the company will choose to finance itself internationally. Therefore, in such cases, a shielding allowance is not needed to counter the companies’ increased capital cost because the Norwegian dividend tax will not affect this capital cost. This is also the general view among the supporters of the shielding allowance.21 Of course, if the dividend tax is not followed by expected increase in dividend distribution, a Norwegian investor may want to invest in something other than the shares of the company. In a well-functioning capital market, however, the company will nevertheless be able to fund itself without increasing the capital costs.
An important question is how far-reaching this insight concerning the funding in international markets is. It is assumed, also by people who strongly favors the shielding allowance system, that the effect of this allowance is very low for listed shares, even if they are listed on the Norwegian stock exchange, if the value of the shares is first of all decided on international capital markets.22
The importance of international capital markets is the basis for a critique of the shielding allowance model, which has been forwarded by Swedish economists Lindhe and Södersten (2012).23 These authors argue that in a small open economy (like Norway’s), “the internationally determined rate of return requirement on large company shares may have a very substantial impact on the rate of return requirement for small company shares”,24 where large companies refer to companies that are internationally traded, and small companies refer to companies that are domestically traded. They conclude: “Such a link considerably weakens the effect of the rate of return allowance [i.e. shielding allowance] also for small domestically owned companies and even imply a detrimental effect on small company investment incentives.”
This insight is an argument for exempting shares in listed companies from the shielding allowance rules, which would have been a significant advantage from an administrative point of view. This was considered but rejected in Meld. St. 11 (2010–2011) p. 82. The basic consideration was that such an exception would create uncertainty for the investor because he/she could not be certain that the shielding allowance would be granted in the future, especially if the company became listed. In addition, taxpayers might avoid the restriction by owning listed shares through non-listed companies.
The theory supporting the shielding allowance has been updated mainly in light of Lindhe and Södersten’s critique.25 NOU 2022: 20 refers to this updated analysis. A closer discussion of this goes far beyond the competence of this writer. However, the reasoning seems to be that the effect of the shielding allowance will be affected by how well diversified the investor in a domestically traded company is. The less diversified, the greater the impact of the shielding allowance on the required rate of return and, by extension, the capital cost of the company. This is because the well-diversified investor may shift investment, for instance, from bonds to shares and thereby uphold the risk and the expected rate of return in his portfolio, as explained at the end of section 3.4 above.
The Oxford International Tax Group forwards a similar argument as that from Lindhe and Södersten.26 Arguing against the view that international rates of return has no impact on “purely domestic business, even one operating in an open economy,” the Group writes:
The issue is whether an individual owning a purely domestic business makes decisions based on a required rate of return. […] Basic investment theory would suggest that she should. That is, if she can earn a rate of return of 10% on some other investment of comparable risk, then she should require the same rate of return from her own business. If she can only earn, say, 7% from her own business then she would be better off investing her money elsewhere. But—as long as she resides in an open economy – the 10% rate of return she can earn elsewhere is almost certainly determined by the equilibrium rate of return in world markets. In this case, her investment decision would still be determined by the world rate of return even if her investment is purely domestic.
The group admits, however, that there may be business owners who do not do such calculations: “Some may simply prefer to own, and work for, their own business as long they receive a reasonable income”.27 The Oxford Group seems to mean that tax policy cannot be based on the attitude of such shareholders. However, this argument is related to the so-called home bias phenomenon, which refers to the tendency of investors to prefer investments in their home country (and not only their own business), even where an investment abroad would result in higher income. This tendency should not be overlooked and gives some support for rules of the shielding allowance type.
The Oxford Group has taken the argument one step further by introducing the incidence of company taxes. The view is that there is no need for relief on the shareholder level because, while the investor must bear the dividend tax, shareholders do not ultimately bear the company tax because of the incidence of that tax.28 A possible investor may conclude the company tax implies that the investment will not reach his/her required rate of return, and he/she will instead invest somewhere else in the international market. Therefore, this investor is not worse off because of the company tax. However, the company tax must borne by someone. It may result in lower dividends for the existing shareholders, but they may behave in the same way as the possible new investors by taking their money out of the company and investing somewhere else. A further possibility is that the company tax is shifted to the customers of the company in the form of higher prices (which is difficult if the product price is set in a market that the company cannot control) or by the employees in the form of lower salaries.29 Also for this way of reasoning, the existence of an international capital market, where the rate of return is given, seems to be crucial.
Intuitively, to the extent that the company tax is ultimately borne by someone other than the shareholders, there seems to be no reason to introduce a shielding allowance to mitigate the effects of the dividend tax on shareholders.
For a long time, the incidence of company taxes has been a prominent research issue in public finance. This research has so far not provided clear results, but there seems to be a widely spread view that company taxes are, to a significant extent, shifted to the employees and perhaps the customers.30 The incidence in each case will depend on elasticity in the labor market, the strength of labor unions, etc.
Despite the disagreement on how far-reaching the effect of international capital markets is, it seems clear that granting a shielding allowance on dividends to internationally traded shares goes beyond the rationale of the allowance. In fact, it seems that granting a shielding allowance also in such cases, brings the effective tax rate down to a level below what is neutral.31 Investors in international capital markets get a shielding allowance even if that is not necessary to mitigate the effect of the dividend tax on the company’s financing costs. This seems to imply that tax revenue is unnecessarily lost.
In summary, even if one accepts the theory behind the shielding allowance in its updated form, the types of investments for which the shielding allowance is considered to have the planned effects seem to have narrowed since the system was introduced in 2004. The rules seem to benefit the company’s capital cost mainly for domestic undiversified investors.32 In addition, even that effect is contested in international economic theory. Thus, the theoretical support for the shielding allowance seems weaker today than when it was introduced almost twenty years ago.
The issue of so-called holding period neutrality (in Norwegian:
Holding period neutrality means that the tax effects for the investor are the same in economic terms regardless of whether he/she chooses to take out dividends from the (personal holding) company early or later. This runs against the intuitive view that postponing the tax liability on the dividends (and income generally) is profitable from a tax perspective and implies that using personal holding companies does not effectively postpone the tax liability. The reasoning is this:34 In a given year, an investor (who controls the company) can choose to receive dividends or not to receive dividends. If the investor chooses to receive dividends, he/she is taxed on the dividend amount. The distribution of dividends does not impact the shielding allowance on the investor’s shares in the company because the dividend is distributed from retained profits, which is not part of the cost price of the shares and, therefore, is not part of the shielding allowance basis for the shareholder. However, the shareholder–after having paid the dividend tax, perhaps with borrowed money–may want to invest the money received as dividends in the stock of a company (the same company or another). That investment will be part of the cost price of new shares and, thus, part of the shielding allowance basis, increasing the shielding allowance. On certain conditions, the gain following from the increased shielding allowance due to this investment will balance the tax paid on the distribution of dividends.35
This reasoning assumes that the shareholder is able and interested in reinvesting an amount similar to the dividend in new shares. However, this will not necessarily happen. The dividend may, for instance, be used for paying taxes or other debts, and the shareholder is not able or interested in taking up new loans for investments in shares. The taxpayer may want to invest in bonds or real estate instead of shares, and such investment would not provide any increased shielding allowance. He/she may instead want to consume (parts of) the dividend. In the opinion of this writer, this reduces the significance of the holding period neutrality.
The holding period neutrality seems not to be affected by whether the company is financed in international or domestic markets, contrary to what applies to the financial and investment neutrality discussed in section 3.5.
However, it turns out that, in practice, shareholders in small controlled companies (personal holding companies) are reluctant to take out dividends if it is not necessary.36 This may be so for legitimate non-tax reasons, typically that the taxpayer is satisfied with the investment or that the capital is needed in the company but not for the taxpayer’s private purposes. Even without such reasons, investors seem not to believe in the holding period neutrality if they know about it at all. The effects of the increased shielding rules are not intuitively understood. Taxpayers easily see the dividend tax, which will be triggered but not the effects on the shielding allowance resulting from a new investment. However, even if the taxpayer realizes the holding period neutrality, there may be rational arguments against distributing dividends. One important such argument is that a taxpayer who takes out a dividend and pays a dividend tax may have to borrow money in order to invest a similar amount as the dividend, because part of the dividend goes to pay a dividend tax. In practice, the interest rate on borrowed money is higher than the shielding allowance interest rate. Therefore, taking out dividends and reinvesting is, in practice, often less favorable than retaining the profits in the company.37 The resistance to distribute dividends may also be based on investors’ hope to avoid (parts of) the dividend tax (by emigrating or being able to profit from the assets for private purposes without triggering a dividend tax) or hope for a reduction of the dividend tax rate in the future.38 Either way, the treatment of retained profits in personal holding companies has been hotly debated in the later years. Critics argue that this is a great tax advantage for rich investors (only a company tax of 22 percent is paid so far, and the shareholders can use the assets of the company almost as their own, or they can emigrate before the tax liability is triggered) and that, consequently, the participation exemption rules should not apply to the profits of the company in such instances. On the other hand, it is argued that the shielding allowance rules imply that the benefits are, in fact, much smaller than they may seem, if they exist at all. Further, shareholders do not have the right to treat the company’s assets as if they were their own assets without paying dividend tax.39 This has extended to a fierce debate on exit tax rules for individuals, triggered by a significant increase in the number of rich people moving abroad, mostly to Switzerland. The main reasons for this last wave of emigration is probably the increase of both the wealth tax and the dividend tax, combined with the expectation of possibly stricter exit rules to come.40 In addition, the taxation of so-called private use of company assets has been much discussed. The existing rules on this issue seem to be rather robust in principle but difficult to apply, mainly because of valuation issues. Therefore, the Ministry of Finance in 2022 proposed more standardized and considerably stricter rules, which were immediately dubbed “the monster tax”; for instance, the tax under certain conditions would be triggered even if the owner never used the assets for private purposes.41
Closely related to the holding period neutrality is the effect of the shielding allowance on the application of the rule in GTA §10-35. This rule applies when the company has paid back paid-in capital (share capital, etc.) to the shareholders. In Norway, such paybacks are taxfree to the shareholders. However, GTA §10-35 requires the shareholders in question to reduce the cost price of the shares with the distributed tax-free amount. The rationale, of course, is to avoid a double favor for the shareholders: A tax-free payback and a lower taxable gain or larger loss on future sales of the shares (because the payback reduces the value of the shares).
The reduction of the cost price of the shares will also affect the shielding allowance base similarly; thus, future shielding allowances will be lower than without the pay-back, and consequently, future dividends and capital gains will be higher (and capital losses smaller). In addition, profits earned in the company may be taxed twice (on company level and shareholder level above the shielding allowance), whereas capital income outside the company is taxed only once. This means, at least as a main rule, that there is no tax benefit for the shareholder to receive the payback of company capital tax-free because the increase in future dividends and capital gains taxation will balance that out in discounted value. Dividends may not be distributed, and the shares may not be sold for a long time, but eventually, the increased tax will materialize.42
A similar effect has explained the Ministry of Finance’s relaxed approach to introducing the dividend tax and the shielding allowance rules as of 1 January 2006.43 The taxpayers were granted a rather long period to adapt to those new rules. Many (and mostly rich) shareholders did this by receiving as much dividend as possible from their companies before this date, then paying the amount back to the company as paid-in capital. The idea was that this amount could be paid back tax-free to the shareholder over the coming years instead of taxable dividends. However, this strategy would require the taxpayer to reduce the cost price of the shares with the received amount when payback takes place. This would not only reduce the cost price but also shielding allowance, in the same way as described above, the result being that as a point of departure, there were no taxes to be saved by this strategy.
When the shielding allowance was introduced, the tax rates for dividends and interest income were the same– 28 percent. The effect of the allowance was that the average tax rate on dividends was somewhat lower than on interest income. This could be explained by the fact that the distributed income had also been taxed on a company level.
In the later years, the dividend tax rate has increased considerably to 37.86 percent (2023).44 This gap between the general capital income tax rate and the dividend tax rate is difficult to defend. However, the shielding allowance helps to mitigate that gap by reducing the average tax on dividends.45
Of course, a lower tax on dividends compared to interest could be obtained by lowering the dividend tax rate. However, this would also lower the marginal tax rate on dividends, making the dividend tax less forceful in countering income shifting.
The shielding allowance rules have been criticized for being complicated. It required the establishment of a register of the taxpayers’ cost price of every share in Norwegian companies and non-resident companies listed on the Oslo Stock Exchange. The initial costs of establishing that register were considerable, but the costs of running it are considered to be rather modest. Shares in companies that are not in the register (for instance, non-resident companies not listed on the Oslo Stock Exchange and non-resident investment funds) can be complicated to handle for the taxpayer. It is reported that many taxpayers abstain from the right to allowance because the allowance is often small and sometimes smaller than the compliance costs.
The rules are complicated not only in practice but also in the law book, admittedly more so for parts in transparent companies (partnerships and limited partnerships) and sole proprietorships, where parallel shielding allowance rules apply. In day-to-day practice, however, shielding allowance rules seem to raise few interpretation issues.
Experience shows that taxpayers struggle with understanding the rationale and effects of the shielding allowance rules. This seems to apply in particular to their effects on the holding period neutrality. In addition, the theory that the investor can neutralize a (increased) dividend tax by increasing the risk profile of the portfolio is certainly not universally accepted. This is important because the rules are intended to work through how they influence taxpayers’ motivation. The limited effects of the rules over many years due to the low interest rates (and, by extension, the shielding allowance interest rate) are probably also important for most taxpayers’ lack of interest in the shielding allowance. That may change with the increased level of interest rates. However, the shielding allowance will probably remain rather small for most investors.
Further, in some cases, the shielding allowance reduces the robustness of the tax system. The system presupposes that companies’ profits have been fully taxed somewhere in the chain of companies (but not necessarily in the company that distributes the dividends). However, that is not always the case.47 If the profits were taxable in a company resident in another state, the tax rate may have been lower than the Norwegian tax rate and/or the rules for computing the income more lenient. Admittedly, such cases will often trigger the Norwegian controlled foreign corporation (CFC) rules, but that is not the case if less than 50 percent of the shares or capital in the company is controlled by Norwegian residents or the efficient company tax rate is above to thirds of the efficient company tax rate in Norway. Thus, the CFC rules will not apply to most portfolio investments in lowtax countries. Furthermore, foreign company income tax may have been saved through avoidance transactions, which would not have been accepted in Norway, or there may have been outright tax evasion on the company level. Capital gains on shares are covered by the participation exemption also when they reflect expected future profits, which may not be effectively taxed. The taxation of dividends secures at least a onelevel taxation of such benefits. However, the shielding allowance is granted on the distribution of dividends to non-company investors also in cases where the income has not been fully taxed on company level. This implies that the shielding allowance, in combination with the participation exemption rules, may partly lead to double nontaxation of distributed profits.
The shielding allowance’s
In regards to financing and investment neutrality, it is agreed in economic theory that the allowance is not needed for investors in companies operating in international financial markets because their financial cost is set in international markets, which are not affected by the Norwegian dividend tax. Therefore, the core focus is on the domestic investor, and the idea with the shielding allowance is to avoid that the dividend tax increases the domestic investors’ required rate of return and, thus, increases the financial cost for the company.
However, non-Norwegian economists have doubted the significance of this difference between international and domestic investors, pointing to the fact that domestic investors may also be influenced by the rate of return in international markets, as discussed in section 3.5.2. This article will not attempt to discuss who is right on this issue–the Norwegian48 or the non-Norwegian economists. Either way, the doubt that has been raised on this issue weakens the arguments in favor of the shielding allowance method. Granting the allowance to investors in companies that raise equity in international markets seems to make little sense and is an argument to restrict the allowance to dividends from unlisted shares (based on the assumption that non-listed companies have mostly domestic investors while this is not the case with investors in listed companies). Admittedly, as already mentioned in section 3.5.4, this may raise new problems. These problems, however, may seem somewhat exaggerated.49
It is striking that there seems to be no empirical evidence on how the shielding allowance works in practice, primarily whether and to what extent it helps to keep the financial costs for the companies down. This is remarkable, considering that the rules have now been in force for almost 20 years.50Some practical experience is made concerning the holding period neutrality, and they point in the direction that investors are not motivated by it. Profits are held back in the companies, presumably at least partly to avoid the dividend tax.
In addition, the conclusion of the evaluation in 2011 is surprisingly muted. As long as neither empirical nor theoretical evidence can exclude that a dividend tax without a shielding allowance would be unfortunate, it would involve a considerable risk to repeal it.51 Also, in NOU 2022: 20 (p. 196), the committee makes substantial reservations. The research results on the damaging effects of dividend taxation–and, by implication, the necessity of a shielding allowance–differ considerably because results are very sensitive to the assumption on which analyses are based.52
All this adds up to skepticism as to the existence and extent of the shielding allowance’s significance for companies’ financial costs. In addition, as outlined in section 3.8 above, the rules are complicated (but, admittedly, raise rather few interpretation issues), they reduce the system’s robustness, and—not least—their rationale is difficult to understand for taxpayers.
The extent to which the holding period neutrality is realized by the rules, is discussed. Anyway, in the opinion of this writer, holding period neutrality is much less important from a policy point of view than financial and investment neutrality.
One aspect of the rules that seems to have been overlooked is that their effects are vulnerable to changes in tax rates. The carry-forward of unutilized shielding allowances implies that the allowance’s tax value will depend on the dividend tax rate in the year in which it is utilized, not in the year in which it was established.53 For the holding period neutrality the tax effects of the reinvestment of the dividend amount will depend on the development of tax rules during the holding period.
As pointed out above (section 3.3), the main purpose of (re)introducing the dividend tax was to prevent income shifting. For this purpose, the
No other Nordic country has introduced a similar system to Norway’s. Admittedly, all Nordic countries, Norway included, have a classical system as a point of departure. Some other Nordic countries accept parts of the dividends from tax liability but in other ways than Norway (In Finland, a percentage of the dividends is tax-free; in Iceland, there is a tax-free amount). Therefore, to paraphrase the title of this article, the Norwegian system stands out as an oddity in a Nordic comparison.
However, Norway stands out not only among the Nordics; recent comparative tax law literature indicates that none of the following countries have a system similar to Norway’s,54 including the UK, the US, Germany, France, Italy, Canada, Turkey, New Zealand, Japan, China, India, Australia, and Brazil. When the shielding allowance system was first introduced in NOU 2003: 9, it created some international attention, and it was assumed that other countries would copy it.55
However, this has not happened. This may be taken as an indication that other countries have not been convinced by the theory behind the shielding allowance and/or have decided not to go for it for administrative or other reasons.
The latest evaluation of the shielding allowance rules was carried out in NOU 2022: 20 pp. 196. The majority of the committee (all seven economists and one lawyer) favored retaining the rules, with only minor changes. A minority (three lawyers, among them this writer) proposed to abolish the rules. This, combined with the fact that economists in the Ministry of Finance seem to strongly favor them, almost certainly means that the shielding allowance rules will remain in force in the years to come. This means that Norway will still stand out as an oddity in a Nordic and international comparison also in the future.
However, there is more: The majority of the tax committee (six economists) proposed to extend the shielding allowance system to (in principle) all capital income.56 The context is the following.
The committee had a very broad mandate, which also included the tax rate structure for capital income generally. Under the Nordic Dual Income Tax introduced in 1991, all capital income was taxed with the same proportional tax rate of 28 percent.57 In the reform of 2004, dividends exceeding the shielding allowance were taxed at 28 percent. During the reduction from 2014 through 2019 of the company tax rate, from 28 to 22 percent, the general capital tax rate (for unconvincing reasons) followed the company tax rate down to 22 percent. However, to prevent income shifting, the reduction of the company tax rate had to be compensated with an increased dividend tax. Therefore, by 2019, the dividend tax rate had been increased to 31.68 percent. Two further increases of the rate, in 2022 and 2023 ending at 37.84 percent, were not motivated by a lower company tax rate but by a political wish to increase the tax on dividends for revenue and distributional reasons, and to prevent income shifting even better.
The result is a great wedge between the general capital income tax rate and highest tax rates on earned income (22 percent vs. 47.4 percent in 2023). Furthermore, the difference between the general capital income tax rate and the tax rate on dividends has become greater than reasonable (even taking the shielding allowance for dividends into consideration).
On this background, the committee unanimously agreed that the general tax rate for capital income should be increased. Whereas the minority (four lawyers and one economist) proposed a rather conventional tax rate structure, the majority proposed–as already indicated–an extension of the shielding allowance system to capital income generally. The main features of this so-called extended shielding allowance model are the following.
Firstly, all capital income is taxed at the tax rate of 22 percent. This can be regarded as a parallel to company taxation in the company/dividend taxation rules. Secondly, the remaining 78 percent of the capital income is granted a shielding allowance calculated similarly to dividends. The rest, after the allowance deduction, is then taxed at a tax rate that may be the same as the dividend tax rate, but they could also be lower.58
If the capital tax rate is the same as the dividend tax rate, the total tax effects will be the same (both on margin and average) regardless of whether the investment is made through a company (investment in shares of a company which owns a real estate, for instance) or directly in the asset in question (the real estate in the example). If the tax rate is set lower, then investment directly in the asset will be favorable. However, even in that case, the system is claimed to be neutral as to whether the investment is financed through equity or loans because of the deduction of interest on a tax basis.59
In a way, this system implies that the capital tax rate is no longer proportional. However, it is not progressive in the traditional sense (tax rates increase with the taxpayer’s income). Instead, the tax increases with the yield of every capital asset. To the extent that the yield exceeds the risk-free (alternative) income, the tax rate increases, even if the total income of the taxpayer is low. One may ask whether this is in accordance with the ability to pay principle. Certainly, this is the system under current rules for dividends, but the issue becomes more important when such rules apply to all capital income.60 In addition, such high marginal taxes cannot be defended by the purpose of countering income shifting from earned income to capital income (dividends), since there is no question of earned income involved.
A marginal capital tax rate of around 48 percent is very high compared with the current tax rate of 22 percent, even if the average tax rate will be lower because of the shielding allowance. For assets, typically real estate, which the taxpayer (or parents, etc., because of the continuity principle on inheritance and gifts) have owned for a long time, the cost price and therefore the shielding allowance may be very low. This issue is probably more important for real estate than for shares because real estate value tends to increase almost uninterrupted while share values generally fluctuate more.
On another level, the practical difficulties connected with the extended shielding allowance model are great. In parallel to the share register under the current shielding allowance rules for dividends, every capital asset would have to be included in such a register, with information on the current owner, his/her cost price, etc. Some assets can be left outside, but that will require rules for delimiting which assets. The
Whether the majority’s proposal will ever see the day of light remains to be seen. If that happens, Norway will be an even greater oddity in a Nordic and international tax context.