Investment Associations in Denmark

Sonnie Karlsson and Casper Ristorp-Thomsen, Market Operations

INTRODUCTION AND SUMMARY

Today, around one Dane in six invests through Danish investment associ ations. At the end of 2007, total assets managed by such associations were estimated to have multiplied by a factor of 10 over a decade, to more than kr. 900 billion. Investment associations thus play a key role in relation to individual savers and to society in general.

There are numerous arguments for investing through an investment association. It is easy, it enables risk diversification, and economies of scale may potentially reduce costs. However, individual investors should be aware of some special characteristics of the Danish market that affect these advantages. It is a relatively protected and concentrated market with little external competition. By far the greater share of the assets are actively managed (96 per cent) – while index-based, passive manage ment is more prevalent outside Denmark, where e.g. Exchange Traded Funds (ETFs) have gained considerable ground in recent years. On aver age, the return achieved by actively managed funds is lower than the market return due to higher costs. The few funds that might actually outperform the benchmark are very difficult to identify beforehand. What can be identified is the level of costs, which is typically higher for the active funds. Finally, it turns out that surprisingly many actively man aged funds are de facto pursuing a passive investment strategy.

This article seeks to shed light on the market for investment associ ations in Denmark. It is concluded that lack of information, financial in centives and fiscal barriers are some of the possible explanations as to why index-based alternatives constitute a relatively small share of the Danish market for investment associations.

THE DANISH MARKET FOR INVESTMENT ASSOCIATIONS: LOCAL, CONCENTRATED AND DOMINATED BY ACTIVE FUNDS

An investment association is an association that receives funds from a group of investors and then invests the funds in various securities¹. Unlike most foreign funds (mutual funds), investment associations in Denmark are owned by their members, funds are typically traded on an exchange, and they are traded at levels close to their book value.

Investment associations have existed in Denmark since 1956, but have roots back to 1928. However, they have only really taken off over the last decade. In the period 1997-2007 the aggregate assets managed by Danish investment associations thus multiplied by a factor 10, cf. Chart 1. The decline in 2008 is primarily attributable to the fall in stock prices, whereby the value of the assets has decreased.

Of the total assets managed at the end of 2007, 43 per cent was placed in equity funds, 42 per cent in bond funds, 14 per cent in mixed funds and 1 per cent in other funds. Compared with the global distribu tion – where bond funds account for approximately 16 per cent² – Denmark has a relatively high percentage of bond funds.

Moreover, the Danish market is dominated by Danish participants. Danish tax rules on investments through investment associations have a significant impact on the structure of and competition in the market, cf. the Appendix. Finally, the market is concentrated. The largest four in vestment association groups manage approximately 74 per cent of the assets.

Just over 100 foreign investment undertakings have been approved by the Danish Financial Supervisory Authority to market a total of almost 6,000 foreign funds in Denmark. The foreign funds are primarily aimed at companies and pension savers. Private investors' shares of foreign funds are generally taxed as investment income and thus at a higher rate than returns from dividend-paying Danish associations. The foreign undertak ings may achieve the same tax status as the dividend-paying Danish associations, but this requires the establishment of a reporting system so that the tax payable by Danish taxpayers can be correctly stated. Furthermore, the foreign undertakings must adjust their portfolios to the investment rules applying to dividend-paying associations. In practice the reporting system and investment rules seem to constitute a major obstacle since no foreign funds have chosen to go through the necessary procedures to obtain the same tax status as the Danish associations. Ordinary investors paying personal income tax are thus in effect left only with the option of placing their funds in Danish investment associations.

Most Danish investment associations have been initiated by Danish banks, which generally serve as distribution networks. Day-to-day oper ations are undertaken by an investment management company, while services such as distribution, trading, depositary services and advisory services in relation to portfolio structure are often purchased from the "parent bank". The distributor of investment certificates is paid commis sion, i.e. a percentage of the value of all new certificates, and the bank receives a fixed annual percentage of the assets as payment for cus tomer contact and advisory services.³,⁴ Irrespective of the actual services rendered, a fixed share is thus transferred to the bank; a payment that would not immediately seem to benefit the owners of the certificates. There are, however, some Danish investment associations that do not pay the banks for customer contact and advisory services.

CONVENIENCE, DIVERSIFICATION AND ECONOMIES OF SCALE – THE RAISONS D'ETRE OF THE INVESTMENT ASSOCIATIONS

Why invest in an investment association? It has often been argued that compared with structuring your own portfolio of securities it is easy and convenient, the investment (even a small one) is diversified, and econ omies of scale will reduce costs.

Easy and convenient is a good reason to opt for an investment associ ation rather than composing an individual portfolio of securities. Many investors want to avoid having to select the securities and are happy to leave it to professional investors by placing their money in investment associations. In addition, the latter provide easy access to markets that might otherwise be difficult to access, e.g. in developing countries, and they take care of the practical aspects of reinvesting interest and dividend, etc.

Diversification means spreading investments on a range of assets. When the prices of different assets are not influenced by the same fac tors, assets can be combined to spread the risk. An investment in equities that are negatively affected by rising oil prices, e.g. an airline company, can be combined with energy equities to reduce the "oil risk". Diversification is a strong argument in favour of investment associations – at least as regards relatively small equity investments. A private in vestor may, of course, structure his or her own diversified portfolio, but in many cases this requires a considerably greater investment than when investing in an association. Chart 2 illustrates the diversification gain, i.e. the risk reduction on diversification. The effect is strong at first, but then declines and converges towards a lower limit as the number of assets increases.

Firstly, the Chart shows that diversification is a relative term – a port folio can be "more or less" diversified. The risk reduction is thus lower on a regional portfolio than on a global portfolio. Secondly, risk converges towards a lower limit, which means that some risks cannot be eliminated by diversification. The reason is that the underlying assets display a posi tive correlation, i.e. they tend to react in the same way to the develop ment in a given factor. For example, a change in GDP growth will have a parallel impact on both IT and car equities. The risk remaining when all diversification options have been exhausted is called the systematic risk. Finally, Chart 2 shows that investors do not achieve further diversification gains by adding more funds to a well-diversified global portfolio.

In relation to investments in government or mortgage-credit bonds, the case for diversification is often weaker. The risk on government bonds is lower and the degree of correlation higher. This reduces the possibility of spreading risk, while the risk to be spread is also lower.

It might be expected that the transaction costs in an investment association would be lower than the costs of structuring and managing a similar individual portfolio. Investment associations have lower fixed costs per trade and can trade at lower marginal costs. However, total transaction costs comprise the number of trades multiplied by the cost per trade and therefore also depend on the trading volume. Moreover, developments in recent years, with web-based stock trading and in creased competition on price parameters such as brokerage and deposi tary's charges, have substantially lowered the costs to private investors of trading equity.

The associations also have various overheads, e.g. payroll costs, rent, audit and fees to public authorities, and most of them remunerate their distribution networks for providing consultancy and advice to customers in relation to buying and selling. This item – intermediation fees – ac counted for nearly half of the costs in 2007, cf. Chart 3. The second larg est cost was asset management, comprising purchase of investment ser vices from specialists within specific areas.

One of the principles behind an investment association is that many investors pool their funds and share the overheads. Economies of scale should thus mean that costs fall as the association grows. However, this is not the case, cf. Chart 3, possibly because a large part of the costs are payable to the banks as a fixed percentage of the assets.

A comparison with investment associations in other European countries shows that the level of costs in Denmark is on the low side.⁵ Never theless, there are many indications that costs could be cut further. For example, the costs of one of the largest US index funds (Vanguard 500) are as low as 15 basis points, while the large ETFs have costs in the range of 10-30 basis points. According to a cost analysis made by the Feder ation of Danish Investment Associations in June 2008, the costs of the Danish equity funds average around 75 basis points for passive units and just under 150 basis points for active units.

In theory, investing through an investment association makes good sense. Investment associations offer convenience, risk diversification and potential cost reductions. In practice, however, the associations do not seem to use their financial muscle to lower costs. Nor does investment in investment certificates automatically lead to the desired spread in risk. Particularly in many bond funds, the diversification gain is limited.

ACTIVELY MANAGED FUNDS VERSUS INDEX FUNDS

The Danish market for investment associations is dominated by "actively managed" funds. At end-2007, the assets of active funds constituted approximately 96 per cent of the market. The management strategy of these funds is to outperform a benchmark (typically a market index).

An alternative to actively managed funds are "index" funds⁶. Unlike the active funds, these funds pursue a passive strategy, the main aim being to replicate an index.

In the USA, where the tradition of investing through funds goes con siderably further back, the situation is different. There, the passive index funds have a larger share of the market (15 per cent)⁷ and the share of passive funds is rising. In the USA and various other countries there are also a number of other passive alternatives. One such alternative, EFTs, is making considerable headway internationally, cf. Box 1. In many re spects, these funds resemble the Danish index funds, which are also traded on an exchange, but ETFs often have lower costs.

EXCHANGE TRADED FUNDS	Box 1
An investment product that has become immensely popular in recent years is "Exchange Traded Funds" (ETFs). An ETF is usually an index fund, i.e. a fund repli cating a predefined index, with very low costs and is traded as ordinary equity. The large ETFs have costs in the range of 10-30 basis points. In many respects, index-based ETFs resemble the Danish index funds, whose investment certificates are typically also traded on stock exchanges. ETFs were developed by US asset managers in the early 1990s. In Europe, they have existed since 2000. Customers include both private individuals and professional invest ors, and today ETFs are among the fastest growing financial products, cf. Chart 4.
DEVELOPMENT IN ETF SALES IN THE USA AND EUROPE	Chart 4
Note: Observations up to and including September 2008 for the USA and July 2008 for Europe. Source: USA: Investment Company Institute; Europe: ETF Network and Bloomberg.
ETFs are offered in most countries – including Sweden. In Denmark, however, no actual ETFs have been established, although the Danish index funds resemble ETFs. So far no foreign funds comply with the Danish tax rules for dividend-paying funds. In other words, private individuals' gains and losses are taxed as investment income. At the moment ETFs are therefore mainly interesting for pension savers in Denmark.

Active management involves the collection and analysis of information with a view to outperforming a given benchmark. Needless to say, this involves costs (collection and analysis of information, trade costs, payroll costs, etc.), to be deducted from any excess return achieved. The costs of investing in an index fund are generally lower than the costs of investing in an actively managed fund.⁸ A passive strategy typically in volves fewer trades (trading primarily takes place when the composition of the index changes) and lower costs for collection and analysis of mar ket information.

Basically, the two strategies reflect two different views on the infor mation efficiency of the financial markets. In an efficient market, active management does not "pay off", cf. Box 2. In that case, the logical stra tegic choice is to invest in a portfolio that reduces diversifiable risk while keeping costs at a minimum. This can be done fairly easily by replicating a large market index.

Is active management worthwhile? Academic research has paid much attention to this question. In most cases, the conclusion is that on aver age the active funds achieve a lower return than the passive, market-replicating funds.⁹

It is not surprising that the average active manager achieves a lower return than a passive, market-replicating fund as the market return is a weighted average of the return on all assets. Since the active portfolios have higher average costs than the passive ones, this inherently means that the average net return on active portfolios, i.e. the return after de duction of costs, must be lower than the net return on index portfolios.¹⁰

However, this does not mean that active management never pays off. In this context it is interesting to investigate whether there are funds that persistently outperform the market. A number of studies have ad dressed this issue.¹¹ While it is not ruled out that such funds may indeed exist, the overall picture is as follows:

• on the basis of historical data it is very difficult to identify funds that systematically outperform an index,
• a large share of the funds' returns can be explained by exposure to sys tematic risk. Only a small part is attributable to the active strategy that is the argument for the higher costs.

These results imply that it is very difficult for an individual investor to select a fund that yields a higher return than a passive index fund. The probability of finding one that yields a lower return is greater.

In reality it is not only difficult to beat a well-diversified index fund; many funds are also struggling to exceed their own benchmarks. Chart 5 illustrates the cost-adjusted returns of Danish investment associations in relation to their own benchmarks, adjusted for costs of 50 basis points.¹² It is seen that, over a period of just over five years, most of the funds yield a lower return than a passive investment in the benchmark.

To get an idea of the extent to which the excess return is attributable to "pure" active performance, the excess return must be risk-adjusted. When the marketing of an investment association is entirely based on historical returns without any indication of the risk exposure, investors should be on the alert. For example, many bond funds use government bonds as the benchmark, but predominantly invest in mortgage-credit bonds, which have a different systematic risk profile.

The fact that the active funds often yield lower returns than an index partly reflects the higher costs of active management. The cost effect becomes more pronounced with the duration of the investment due to the compound interest effect. This can be illustrated by the following example: an initial investment/asset portfolio of kr. 10,000, an average annual return of 8 per cent and costs to the investment association of 1.5 and 0.5 per cent, respectively. After 25 years, around one third of the total assets have been paid in costs to the association if the annual cost is 1.5 per cent, cf. Chart 6. At an annual cost of 0.5 per cent, 12 per cent goes to costs.

A study by Bechmann and Rangvid (2006b) of the Danish market in the period 1994-2003 shows how the cost level of a fund is an indicator of the future market-adjusted return – the higher the costs, the lower the return.

Both in Denmark and internationally, active investment funds are struggling to outperform their benchmarks. The few funds that may be successful in this respect are very difficult to identify. What can be iden tified beforehand is the level of costs – which has a significant impact on the investors' ultimate returns.

WHY SO FEW PASSIVE ALTERNATIVES?

Studies of the funds' management performance go back a long way¹³, and it should not come as a surprise to the sector that active funds cap able of outperforming passive alternatives are few and far between. It is therefore a mystery why so many people invest in actively managed funds. Even in the USA, where passively managed funds are far more prevalent, the aggregate volume of the actively managed funds seems to be astonishingly large.¹⁴

On the demand side, many customers may be unaware of the histor ical evidence in favour of the passive strategy. Most of them buy invest ment certificates through their bank. Unless the bank has an incentive to advertise the index products, this information is not passed on to cus tomers. Presumably there is also a psychological aspect. Words such as "passive" and "average return" are perceived as negative by many people, while words like "specialist", "active" and "outperform the benchmark" have a far more positive ring to them. Part of the explan ation could also be that the individual investor believes him/herself cap able of selecting the winning investment association – possibly on the basis of historical returns.

On the supply side, the bank distribution networks play a major role. The banks have a financial incentive to recommend the active funds, which typically generate higher income for the bank by way of trading fees, depositary's charges and portfolio advisory services. In the case of index funds it can be difficult to argue that one fund is better at repli cating a benchmark than another. The best argument for selecting one particular index fund is the cost parameter. Suppliers thus have an incen tive to differentiate their products.

Moreover, many of the funds marketed as active funds de facto mirror a benchmark closely, cf. Chart 7. The tracking error is a measure of the standard deviation in the spread between the portfolio and benchmark returns, i.e. how close the portfolio is to the index it refers to.

It goes without saying that passive funds should have a low tracking error since their strategy is to replicate a benchmark. Active funds that have an explicit strategy aimed at – and that charge extra for – "outperforming" the benchmark, would be expected to have a some what greater tracking error than the passive ones. Most Danish index funds have a tracking error in the range of 1-4 per cent p.a., which is also where most of the active funds are found. However, the latter gen erally have higher costs.

If the tracking error is low, the fund deviates only marginally from the benchmark and can therefore be expected to have correspondingly lower costs. Chart 7 confirms the positive relationship between costs and level of activity, but also indicates a considerable spread. Thus there are very active funds with lower costs than those close to the benchmark, and likewise there are active funds with low tracking errors and costs matching those of the passive funds.¹⁵ Index funds generally have lower cost-to-activity ratios, but as the Chart shows, there are also index funds with relatively high costs. In other words, what matters is not whether the funds are designated as "active" or "passive", but whether costs are low in relation to the level of activity.

In a study based on US data, Cremers and Petajisto (2007) criticise a number of active US funds for being "closet index funds" with high costs. The average tracking error is higher in the USA than in Denmark, but the relationship between costs and tracking errors is similar. In the USA, too, many funds are marketed and recommended as active, but in reality mirror an index closely.

A passive fund is not necessarily inexpensive. Typically, trading costs are lower than in an equivalent active fund, but marketing and distribu tion account for a large share of the costs. Pressure to reduce such costs would presumably require a market situation with easy access to infor mation and greater competition, not only among the active funds, but also among the passive ones. This is not facilitated by the current tax rules in Denmark that bar private investors' access to foreign funds – such as ETFs – combined with the existing distribution channels and financial incentives.

SUMMARY AND CONCLUSION

In Denmark, actively managed investment associations are dominant. Many active funds are struggling to outperform those applying a passive strategy due to the higher level of costs. Finally, it turns out that sur prisingly many actively managed funds are, in fact, close to pursuing a passive investment strategy. Lack of information, financial incentive structures and fiscal barriers may be some of the explanations to why index-based alternatives with low costs constitute a relatively small share of the Danish market for investment associations.

LITERATURE

Anderson, Seth C. and Parvez Ahmed (2005), Mutual Funds, fifty years of research findings, Springer.

Bechmann, Ken L. and Jesper Rangvid (2006a), Danish Mutual Funds: Description, performance, and a European comparison, Chapter 3, pp. 31-62, in Performance of Mutual Funds, edited by Greg N. Gregoriou. Palgrave Macmillan.

Bechmann, Ken L. and Jesper Rangvid (2006b), Rating Mutual Funds: Construction and information content of an investor cost based rating of Danish mutual funds, Journal of Empirical Finance, Vol. 14, No. 5, pp. 662-693.

Christensen, Michael (2003), Evaluating Danish Mutual Fund Perform ance, Aarhus School of Business, Working Paper.

Cremers, Martijn and Antti Petajisto (2007), How Active is Your Fund Manager? A New Measure That Predicts Performance, Review of Finan cial Studies, forthcoming.

Fama, Eugen F. and Kenneth R. French (2008), Mutual Fund Perform ance, Dartmouth College, Working Paper.

Felstedt, Niels and Jesper Rangvid (2007), Index Investment Associations in Denmark (in Danish only), Finans/Invest, No. 7, pp. 14-20.

Gruber, Martin J. (1996), Another puzzle: The Growth of Actively Man aged Mutual Funds, Journal of Finance, No. 51.

Henriksen, Karsten (2008), Economies of scale primarily benefit the banks (in Danish only), Morningstar.

Malkiel, Burton G. (1999), A Random Walk Down Wall Street, W.W. Norton & Company, New York.

Malkiel, Burton G. (2005), Reflection on the Efficient Market Hypothesis: 30 years Later, The Financial Review, No. 40, pp. 1-9.

Sharpe, William F. (1991), The Arithmetic of Active Management, Finan cial Analysts Journal, January/February, pp. 7-9.

APPENDIX: TAXATION OF INVESTMENT ASSOCIATIONS

Danish tax rules distinguish between investor types (private individuals, companies, pension funds, etc.), investment targets (equities or bonds), and dividend-paying and accumulating investment associations.

• Private individuals (free funds): Realised gains on investments in divi dend-paying equity funds are taxed as equity income (28-45 per cent). Losses can be offset against dividends and gains on quoted shares, etc. For bond funds, gains are taxed as investment income (33-59 per cent), but only capital losses on certificates in funds investing in foreign bonds can be deducted from investment income. Returns from bond funds disbursing gains on "rubber-stamped" bonds are tax-free in come. As regards investments in accumulating investment associations (known as investment firms), both realised and unrealised gains and losses are taxed as investment income. Shares in foreign funds ("investment undertakings") are in principle taxed under the rules applying to accumulating investment associations (investment firms), i.e. as investment income (up to 59 per cent, depending on other income). If investors in foreign investment undertakings are to be taxed in the same way as those investing in Danish dividend-paying associations, these institutions must be registered by the tax author ities (SKAT) in Denmark and must regularly report amounts disbursed for taxation purposes. Finally, they must observe special Danish invest ment rules in the same way as the Danish dividend-paying associ ations.
• Pension funds: Both realised and unrealised gains are taxed at 15 per cent, while net losses are carried forward to the following year.
• Companies: Realised gains and losses are taxed as corporate income for dividend-paying funds. However, the 3-year rule for equities ap plies to gains/losses on certificates in equity funds. For accumulating funds (investment firms) both realised and unrealised gains and losses are taxed as corporate income.

For an in-depth description of the tax rules, reference is made to SKAT.