The Taxation of Industrial Foundations in Sweden (1862–2018)

Sweden was a poor agricultural country until industrialization leveled off in the second half of the 19th century. A small number of successful entrepreneurs took the lead and established family business groups controlling a large share of Swedish industry [see, e.g., 7, 9, 11, 12, 34, 39,40,41,42,43, 47]. Because of their economic significance, they have received attention from policy makers and analysts investigating their influence [e.g., 23, 24, 57]. However, no systematic examination of their influence emerged until the early 1960s when the government launched an inquiry to investigate the control of Swedish industry [51], the so-called concentration inquiry (Koncentrationsutredningen).

The inquiry found that 17 ownership spheres controlled one-third of the largest firms’ capital and that one-fifth of all private sector employees were employed in firms controlled by these ownership spheres (excluding bank and insurance companies). Fourteen of these spheres were controlled by family business groups.

See Andersson et al. [2] for the importance of family firms in Sweden.

The control of the business group was typically organized as a four-level control-ownership pyramid. The family is at the top of the pyramid. It controls one or several industrial foundations, the second level, through board representation. The foundation(s), in turn, is(are) dominant owner(s) in a listed, closed-end investment fund, representing the third level, that owns the listed and/or unlisted firms at the bottom of the pyramid. For instance, the Knut och Alice Wallenbergs Stiftelse is the most important industrial foundation in the Wallenberg sphere. It controls the closed-end investment fund Investor, which in turn controls firms such as Atlas Copco and Ericsson [e.g., 25, pp. 525–527]. In combination with differentiated voting rights (by means of dual-class shares) and pyramid building, several companies could be controlled with a relatively small amount of capital [15].

Hence, by controlling the industrial foundation(s), the family is the ultimate decision maker in a large business group without any, or negligible, personal ownership. The foundation's charter, which states that the board appoints its successors, guarantees the founding family's control over the foundation and family business group over generations. When the family business group is controlled by direct ownership, there is only a three-level control-ownership pyramid; the family directly owns the closed-end investment fund.

Three of the 17 ownership spheres identified in SOU 51 still exercise significant control over Swedish industry: the Ax:son Johnson, the Klingspor and Stenbeck, and the Wallenberg family business groups. The Wallenberg family business group is still the most influential of all business groups in Sweden, and the Ax:son Johnson family business group is one of the most influential business groups. The Wallenberg family business group is controlled by the fifth generation of family members and the Ax:son family business group by the fourth generation, which makes them the oldest of the large ownership spheres in terms of generational control. This indicates that industrial foundations may be effective in creating and sustaining large family business groups. This may be explained by the motives for establishing an industrial foundation, mentioned above: First, when tax favored, industrial foundations have a financial advantage; second, the risk of inheritance division is reduced; and third, heirs are hindered from squandering their inheritance.

Family governance was long considered inferior to managerial governance [see, e.g., 4, pp. 13–14]. A growing separation between ownership and control in the largest American corporations led to support for this view. It was confirmed in the late 1970s that the proportion of managerial enterprises among the 200 largest corporations in the United States had increased from 50% in 1929 to more than 80% [3, 22]. Later research examining corporate ownership around the word indicates that family control seems to be the general form of corporate governance and is prevalent among listed firms [e.g., 5, 10, 33]. This is also the case in Sweden, despite deregulations of financial markets and increased globalization, which have increased the possibilities for convergence toward the Anglo-American model of corporate control [19].

Research also finds that corporate control differs significantly across countries [e.g., 37]. For instance, industrial foundations play an important role as controlling owners in several countries, especially in northern Europe [e.g., 61, 62]. This is especially notable in Denmark, where foundation-controlled companies constitute approximately 70% of stock market capitalization [60], and in Sweden, where a few influential family business groups, notably the Wallenberg group as discussed above, have used foundations to exercise far-reaching control over Swedish industry. An explanation proposed, but not systematically examined, is that the tax system has favored industrial foundations over direct individual ownership. Next, we, therefore, turn our interest to the taxation of industrial foundations.

Industrial foundations do not have to pay taxes on capital income, such as dividends, interest, and capital gains. They have also been exempted from taxes on wealth, inheritance, and gifts (when applicable for natural persons). However, they pay taxes on real estate, property income, and business income (rörelseinkomst). These rules have evolved over time because of changes in statutory laws and case law (rättspraxis).

Case law is the set of decisions of courts that can be cited as precedent.

The roots of tax rules for foundations date back to the regulation from 1810, where the so-called pious foundations (fromma stiftelser) were exempted from taxation. The law stated that foundations were exempted from paying tax on chattels, immovables, gifts, and inheritance [56]. In the Appropriation Law (Bevillningsförordning) introduced in 1862, the tax exemption was extended to several areas such as research, education, childcare and healthcare.

The main idea behind a pious foundation was that all payouts should be used for charitable purposes. One rationale for the tax exemption was that these foundations spent money on activities that otherwise had to be financed by taxes. A foundation could have more than one purpose (and, as a consequence, use its revenues in more than one way). If only part of the foundation had charitable purposes, then these rules applied only to that part. If, for example, half of the foundation's activity had charitable purposes (as stated, e.g., in the statutes of the foundation), half of the income must be spent on charitable purposes, and this half was exempted from income taxation. A foundation with multiple purposes could, in this way, both keep some money within the foundation and spend money on non-charitable purposes without being required to pay taxes on all income.

See SOU 50 and SOU 53 for more detailed discussions.

In 1942, the legal framework was formalized and the current legal framework instituted [26]. The legislation was preceded by a long process based on a proposal from a tax committee in 1936. The rules have then remained largely unchanged. Before 1942, the main focus of the tax authorities was whether a foundation should be regarded as a pious foundation. Classification as a pious foundation was based on case law, but the case law was inconsistent because administrative courts could differ in their judgment of whether a foundation fulfilled the requirements to be tax exempt.

A main concern with the statutory law before 1942 was that it was possible for an industrial foundation to retain income and accumulate funds to be spent on charitable activities in the future that were instead spent on non-charitable activities. Although unlikely and difficult, the purpose of a foundation could be changed, or the foundation could be dissolved and liquated. Hence, there was a risk that tax-exempt income could be used for non-charitable activities (if the purpose of the foundation was changed) or could be obtained by ordinary people (if the foundation was liquidated).

There is a limited possibility to change the taxation of past income. Current tax law allows the tax authority to change the taxation of income 2 years in the past after an appeal and at most 5 years in the past if incorrect information was reported on the tax return.

The new legislation clarified that foundations supporting charitable activity should be taxable only for income from property and business activity.

At this time, the property tax had two components, local and national, and these foundations had to pay only the local part. It was argued that removing the local part would reduce municipal financing in a non-legitimate manner.

With a formal completion requirement, it would not be possible to accumulate (all or the bulk of) tax-exempted income in the foundation over time (on the grounds that it will be spent on charity sometime in a distant future). With the activity requirement, the foundation was, on the other hand, not obliged to use everything it spent (but only the main part) on charitable activities.

All activity must, however, be in line with the purpose of the foundation.

The rules were now also made binary, meaning that either the tax exemption criteria were fulfilled–and then all income (with the exception of income from property and business income) was tax exempt–or the criteria were not met–and then all income had to be taxed (as if earned by a limited company). Hence, foundations could no longer divide their income into nontaxable (the charitable part) and taxable (the non-charitable part) income. Failing to satisfy one requirement was sufficient to be fully taxable. An alternative tax rule, which would keep the tax incentives for foundations with charitable purposes in place, could be to allow foundations to deduct all expenditures for charitable purposes and then tax the residual net income in the same way as other businesses. This option was rejected for two reasons: high administrative burden for the foundation and weakened opportunities for consolidation because new investments would have to be carried out with post-tax incomes [52]. Note that the sharp reduction in the corporate income tax rate since the 1980s has made the latter argument less valid.

The statutory corporate income tax has decreased from above 50% during the 1980s to slightly above 20% (see Section 3.2).

In practice, the new rules implied that, on average, approximately 80% of the net return had to be spent every year, and of these expenditures, 90–95% must be on activities that the tax authority regards as charitable.

There have been some changes since 1942, but the idea behind the rules has remained basically unchanged. In 1964, the definition of charitable purposes was widened to include Nordic cooperation, and in 1984, the municipal taxation of legal entities was abolished. No specific changes in the taxation of industrial foundations were made as part of the major Swedish tax reform in 1990–1991. In 1999, the activity requirement was changed from mainly (huvudsakligen) to solely or virtually solely (uteslutande eller så gott som uteslutande). The tax laws for foundations were made more liberal in 2014 (including that the concept of philanthropic purposes was widened again), but these changes did not essentially change the possibility to own or control firms via foundations [14].

Importantly, no exact numbers are mentioned directly in the law. Both case law and circumstances are relevant for the exact determination of how much of the return has to be used for charitable purposes to exempt a foundation from most taxes instead of being liable for full taxation on all its net income.

Profits made by corporations controlled by industrial foundations are subject to corporate income tax. Figure 1 depicts the evolution of the marginal corporate income tax rate from 1862 to 2018. Corporate taxes were paid to the state (national government) and, until 1985, to the municipality (local government). Corporate taxation was progressive between 1903 and 1939, and the figure shows the highest and lowest statutory tax rates during this period.

Figure 1

In the first 50 years of our study, corporate tax rates were low (below 13%) compared to later tax rates. The highest marginal tax rate increased sharply after World War I. The lowest marginal tax rate increased markedly in 1939 when the system was made proportional. The statutory tax rates continued to increase during the post-war period and exceeded 50% by the mid-1950s. The 1990–1991 tax reform decreased the statutory tax rate to 30%. The tax rate was lowered in four subsequent steps, reaching 22% in 2013. Between 1984 and 1990, an additional “profit sharing tax” on corporations was levied to finance the so-called wage-earner funds (löntagarfonder).

It has been estimated that this tax increased the statutory corporate tax rate by approximately 5 percentage points [1]. However, there was a fear among businessmen that the rules might be sharpened. Unimplemented proposals with the purpose of transferring private ownership to the funds–which had been suggested before the formal rules came into effect–was seen as a threat to business by many owners [18, pp. 352–354]. This effect is not included in the METR because the King–Fullerton framework does not take business or political risks into account.

There have been ample opportunities to reduce the statutory corporate tax by allowances and grants–particularly between 1939 and 1991, when the effective corporate tax rate could be substantially lower than the statutory corporate tax rate [48, 49]. The tax reform in 1990–1991 abolished most of these options, thus making the statutory and effective corporate tax rate much more equal.

See Lodin [35, chapter 7] for further discussion about the design of the new corporate taxation.

The inflation rate varied, with few exceptions, between −5% and +5% until World War I, but it was zero, on average, and the price level was virtually stable (see Figure 2). Inflation peaked during World War I and was close to 50% in 1918. Deflation followed the war with a policy to restore the price level to their pre-war levels, and deflation was nearly 20% in 1921. Sweden also experienced deflation at the end of the 1920s and the beginning of the 1930s. On average, the price level was roughly stable for approximately 80 years between 1862 and 1939. Inflation peaked again during World War II and during the Korea boom in the 1950s. In addition, inflation was moderate during the 1950s and 1960s and rarely exceeded five percent. It increased during the 1970s and 1980s and occasionally exceeded 10%. The central bank was granted independence, price stability was the prime goal of monetary policy, and an inflation target to keep inflation at approximately 2% was established in the 1990s. Inflation fell and was approximately 1%, on average, between 1994 and 2018.

Figure 2

The King and Fullerton [31] framework is a standard method for measuring the METR on investment projects in the nonfinancial corporate sector in both research and practice, for example, by the OECD. The framework accounts for all capital income taxes, corporate taxes, wealth taxes, and inflation that concern the investment decisions of the owner. Consequently, it accounts for the source of finance–new share issues, retained earnings, or debt–because capital incomes are taxed differently. Similarly, the METR will depend on ownership because different categories of owners are treated differently in tax law.

The METR is formally the difference between the preand post-tax real rate of return of a marginal investment project, divided by the pre-tax real rate of return. For example, if the pre-tax real rate of return on an investment project is 10% and the post-tax real rate of return is 6%, the METR will be 40% ((10-6)/10).

However, note that the METR is not simply an addition of corporate and owner-level taxation adjusted for inflation. It is an equilibrium model that is supposed to be solved where (1)

the present discounted value of the profits from the investment must equal the cost of the investment,

Using the King–Fullerton framework and considering the rules and evolution of the tax system as presented in Section 3, we can calculate the METR for industrial foundations, with new share issues, retained earnings, and debt as sources of finance for the investment.

In the King–Fullerton framework, investments in machinery, buildings, and inventories are analyzed. In this study, we will analyze investments in machinery. We adopt to the standard assumptions of 20 percent marginal rate of return and 10 percent rate of exponential depreciation using the fixed-p model as described in King and Fullerton [31].

The corporate income tax rate is straightforward to incorporate when the corporate income tax system is proportional. We will use the top tax rate when the system is progressive (1903–1939).

Using, for example, the lowest or the average of the highest and lowest tax rates in 1903–1939 would not change our general conclusions.

The capital income tax rate is first set to zero, because industrial foundations are exempt from paying tax on their capital income. This is in line with the analysis performed in earlier studies [30, 31, and, for Sweden, 48, 49].

However, industrial foundations are obliged to pay out the bulk of their capital income (less capital gains) for charitable purposes, as described in section 3. This imposes a cash flow effect that weakens the ability to maintain control over the “sphere companies” and, hence, provides a negative incentive for entrepreneurs to use industrial foundations as a control vehicle. In fact, this effect parallels the cash flow effect caused by the personal capital income tax on dividends and interest. This cash flow effect has not been discussed or considered in previous analyses. To illustrate the impact on the incentives to control firms through direct individual ownership or through industrial foundations, we will make a complementary calculation of the METR where the requirement to donate a large part of the return for charitable purposes is treated as a tax. Although not formally correct, this calculation will capture the cash flow effect and further our understanding of the incentives to use industrial foundations to control companies.

A tax is formally defined as compulsory unrequited payments to general government.

This complementary calculation requires an assumption regarding how large a share of its net income the foundation is obliged to donate. As described above, no exact numbers are mentioned in the statutory law, and both case law and the specific circumstances of the foundation are relevant for the exact determination of how much of the income has to be used for charitable purposes. Case law after World War II implies that, on average, approximately 80% of the net return has to be spent on charitable causes; we will use this percentage in our calculations for the whole period.

The wealth tax rate is set to zero, because industrial foundations are exempted from the wealth tax. Actual inflation rates are used in the calculations, as presented in Section 3.3.

There are special tax rules that must be accounted for during the period, for example, the Annell deduction, the investment fund system, a special additional allowance given between 1976 and 1978 and in 1980, and the SURV (Skatteutjämningsreserv, tax equalization reserve). These allowances lower the effectiveness of corporate taxation in different ways. The Annell deduction, however, only reduces corporate tax liability when new share issues are the source of finance. Between 1939 and 1951, the immediate write-off of investments was possible. These rules and how they are incorporated are described in Wykman [64]. Finally, the model assumes that no (further) tax changes will occur and that all tax allowances for investments can always be used.

Figure 3 describes the METR with new share issues and retained earnings as a source of finance.

As control is exercised through ownership, debt is a less relevant source of finance for industrial foundations. The METR for debt is presented in Johansson et al. [28].

Figure 3

In the ordinary METR calculations, the income tax for the foundation is set to zero. In a strict sense, this is a true interpretation because donating a part of one's income cannot be equated with a tax. However, as discussed above, it can be argued that the METR calculated in this way does not correctly capture the incentive effects and that it may be misleading. The requirement to donate the bulk of the net income for charitable purposes has a negative cash flow effect similar to a dividend tax. This effect is not addressed in the ordinary King–Fullerton framework, but the METR can be recalculated to include this effect, as discussed in Section 4.2.

Figure 4 depicts the results when including this cash flow effect. In the case of new share issues, the METR generally fluctuates around 100% and 150%.

In the case of new share issues, the potential investor will require the return net of the dividend tax to be equal to the alternative investment, corresponding to the nominal interest rate net of the interest tax, that is, the investor will value the investment as if the investor will be remunerated through dividends only (see Wykman 64 for a more detailed description).

Figure 4

The favorable treatment of retained earnings over new share issues favors incumbent, well-established, and mature firms, which historically have generated profits, in contrast to new entrants that lack retained earnings to use. Industrial foundations also generally prefer to finance investments with retained earnings to avoid the risk that ownership will be diluted, which is possible when using new share issues.

The METR for new share issues in Figure 4 should, however, be considered a maximum ceiling for two reasons: the donation requirement could be somewhat lower than 80%, and the company may not distribute all profit as dividends but rather reinvest it.

As the foundation does not pay tax on capital gains, a complementary analysis is to decompose the true return on ownership into dividends and price changes on the underlying stocks, that is, capital gains, and use that as the basis for the calculation of the incentives. The share of dividend yields of the return on the public stock exchanges for the period 1870–2012 is, on average, approximately 40% [63], and a recalculation of the METR using this number is shown in Figure 5.

Figure 5

The METR fluctuates around 20–50% until World War II (ignoring the spikes). After the war and until the tax reform in 1990–1991, the METR fluctuates around 50–85%. After the tax reform, the METR decreases to approximately 40–50%. Under these assumptions, the METR will be lower and not exceed 100% (ignoring the spikes during World War I), even if the negative cash flow from donating the bulk of the dividends for charitable purposes is included.