  |
  |
 |
 |
 Managing Risk in a Concentrated Market Click here to download this file (PDF file. Adobe Acrobat Reader required) Managing Riskin a Concentrated Market (Canadian Equity Benchmark Alternatives) June, 2000 Contents
I. INTRODUCTION William M. Mercer Limited has been commissioned by the Industry Practices Committee of the Pension Investment Association of Canada (PIAC) to prepare a research paper on managing risk in a concentrated equity market. In light of the recent increase in the concentration of the Canadian stock market (the "Nortel effect"), institutional investors are seeking guidance in developing appropriate policies to deal with this issue. While this paper will provide information helpful to plan sponsors, it is not intended to deal directly with the issue of whether and at what level plan sponsors might wish to set a single issuer concentration limit for their fund. The Nortel effect was the result of the rapid growth in the capitalization of Nortel Networks Corporation. By the end of 1999, Nortel shares made up more than 15% of the capitalization of the TSE 300. Today, Nortel's weight exceeds 25%, following BCE's "spin off" of its stake in Nortel. We are very pleased to be asked to assist plan sponsors across the country in making informed decisions on this important matter. The research paper was prepared based on a number of guiding principles.
This research paper presents four alternative approaches for consideration in managing risk within the concentrated Canadian equity market:
This paper includes a brief description of the approach, followed by a more detailed discussion of the issues particular to each approach, an overview of the advantages and disadvantages, and ends with the implications of the approach in terms of implementation. The paper focuses primarily on defined benefit (DB) plans. However, a separate section outlines the special considerations that need to be addressed by sponsors of defined contribution (DC) plans. This paper does not provide specific criteria for the selection of alternatives, although advantages and disadvantages of each alternative is presented. Possible criteria for selection of an alternative might include, investment beliefs and theoretical support, ease of implementation, and overall practicality and cost effectiveness of the approach. II. EXECUTIVE SUMMARY The following chart summarizes the advantages and disadvantages for each of the four alternative approaches presented in this paper, and the investment belief upon which each of these approaches is based. BENCHMARK ALTERNATIVES IN A CONCENTRATED MARKET
III. BACKGROUND The size of Nortel relative to the TSE 300 increased considerably from mid-1999 through 2000. By May 2000, Nortel comprised approximately 27% of the TSE 300. Historical Nortel Networks weights in the TSE 300 (excluding holdings by BCE)
This presents a significant issue for pension plan sponsors in Canada. The issue of prudence with respect to the appropriate investment in one security is a basic consideration for plan sponsors. Many have self-imposed market value limitations on the holding of a single security, either at the asset class or total fund level (typically 5% to 10%). Various legislative and regulatory limitations governing segregated and pooled pension funds also limit book and market weights of the size presented by Nortel. HISTORICAL PERSPECTIVE Concentration in the Canadian marketplace is not a new phenomenon. What is new is the magnitude of this concentration within a single stock. There was a period of time during the 1960s when Inco held a dominant position in the TSE 300 Index. From 1978 to 1982, Dome Petroleum's weight in the TSE 300 rose from 1% to 6% - and then fell to below 1% by 1984. BCE's weight was just over 10% for a brief time in the mid-1980s before levelling off to the 4% to 8% level for the next ten years or so. BCE's weight exceeded 10% again during 1999 and early 2000. However, this was primarily due to its partial ownership of Nortel. Following the "spin off" of its Nortel stake, BCE's weight on the TSE 300 declined to about 3%. Most of the industry discussion in Canada regarding market concentration has centred on stock-specific risk. Of note, however, is the dominant position of the industrial products sector, which represents over 40% of the TSE 300. Although not an official index sector, the "technology sector" more accurately depicts the current sector concentration issue in Canada. Technology stocks are estimated to comprise roughly 45% of the TSE 300. Dominant sector positions in the TSE 300 are also not limited to recent times. Notable high sector concentrations in the past include the following: ¨ Oil & Gas 1980-81 (20% to 24%); ¨ Gold & Precious Metals 1993-96 (10% to 12%); and ¨ Financials 1997-98 (14% to 18%). Should plan sponsors be concerned with both stock and industry concentration levels? For example, if managers were limited to 10% in Nortel but would like to preserve their Canadian technology weight, would they be forced to invest in stocks of lesser quality? Managing sector-specific risk is an important element of an overall risk management process, particularly in Canada where performance in any given year can deviate significantly from one sector to the next. However, this paper will concentrate on the stock-specific risk issue, recognizing that this is the most pressing matter, and the fact that the dominance of the technology sector in Canada has largely been driven by one stock. Whether the Nortel concentration ultimately goes away or not, the point is that we have seen concentrated markets in Canada before (albeit not at these current levels), and we will in all likelihood see them again. Perhaps one positive result of this recent concentration is that it has caught the attention of everyone in the industry. It is galvanizing plan sponsors, investment managers and consultants into action - action that will hopefully lead to better investment and risk management of pension assets. Even if the current concentration does go away, there will be policies and procedures in place to deal with this type of situation if it occurs again sometime in the future. REGULATORY ENVIRONMENT The relevant regulatory provisions of provincial and federal pension legislation govern pension fund investments. In general, these regulations restrict a pension fund from investing more than 10% of its assets in any one security (on a book value basis). Mutual funds are regulated by National Instrument 81-102 (formerly known as National Policy 39), prohibiting these types of funds from owning more than 10% of a single security measured by market value at the time of purchase. Pooled funds also have a 10% single stock limitation based on market value. These restrictions have been criticized of late. Some interest groups - both plan sponsors and investment managers - are currently lobbying for change in legislation, calling for either an increase or the removal of these caps. Those seeking change have a range of rationales: the perceived inappropriateness of limits as large, global firms dominate the economic landscape; the perceived arbitrariness of the "10% rule", and the inability of fiduciaries to set prudence standards based on their own particular objectives and risk tolerance. Another consideration is that most "prudent" rules of diversification would probably be market value limits, not book value limits, since market value will measure total exposure to market risk at any given time. Prudence with respect to single security levels is difficult to deal with in broad terms since it can be affected by plan-specific factors. For example, a sponsor of a very well funded plan may decide that its risk tolerance is high compared to an unfunded plan, and therefore able to bear higher concentration. On the other hand, most would probably agree that a ten stock portfolio (possible with a 10% maximum holding) is not prudent. We may indeed be suffering from regulatory inertia, and lobbying for changes may or may not be successful. However, whatever strategy that is adopted must recognize prevailing legislation. Perhaps the legislative requirement most germane to the market concentration issue in Canada is the foreign property limit. The natural solution to reducing reliance of an increasingly concentrated domestic equity market is to invest abroad. For example, consider a pension fund with a traditional 60%/40% equity/debt asset mix policy. If the pension fund resides in a market where 25% of the local stock index is represented by one security, the stock-specific risk of the fund can be diminished considerably by increasing the foreign equity exposure. This is consistent with the increasing globalization of economies and capital markets.
What percentage of a pension fund in Canada should be invested in non-domestic securities? Few would support the notion that a Canadian pension fund should invest 98% of its assets outside of Canada simply because the Canadian equity markets comprise only 2% of the world market. However, taking into consideration the Canadian-denominated liabilities of DB plans, and the "home country" bias that prevails in most global markets, one-half to two-thirds of a Canadian pension fund's equity portfolio invested outside of Canada seems quite reasonable. Acceptable synthetic foreign equity strategies exist to allow Canadian pension funds to obtain foreign equity market allocation while meeting the book value requirement of the Income Tax Act. However, derivative-based strategies also require plan sponsors to educate themselves about appropriate risk controls which may be new to some certain plan sponsors. It is interesting to note that while Canada is a relatively small market from the equity market capitalization perspective, we have anywhere from the fourth to sixth largest accumulation of pension assets in the world. Among the major pension asset markets, Canada is the only country to have a foreign investment restriction. Pension Fund Capital Around the World in 2000 (Top Eight Pension Assets Countries)
Source: Intersec Research Corporation, Pensions & Investments (as reprinted in Pension Fund Excellence - Ambatchtsheer/Ezra). This paper is not a platform for change with respect to the foreign property limit. However, any solution requires a balance between prudence and public policy requirements. Suffice to say that underlying the four alternative strategies presented in this paper is the implicit assumption that plan sponsors would first increase their foreign equity exposure in dealing with the concentration in the Canadian equity market. Generally speaking, from a historical perspective, too great a focus on the TSE 300 has not served Canadian pension plan sponsors well. It would be ironic if a skewed Canadian equity index pushes sponsors into more foreign equities. The two persuasive reasons for more foreign equity investing - risk reduction through diversification and the potential for higher returns - have always been present and prudent. BENCHMARKS AND INDICES Integral to addressing the concentrated Canadian market, is the appropriate benchmark to use for performance measurement purposes. Nortel's recent concentration has forced many sponsors to recognize this distinction. The characteristics of a good benchmark are generally accepted to be as follows:
Benchmarks are used to measure a manager's performance. Indices, on the other hand, are primarily used to measure the performance of a certain market. This distinction may appear to be fine. However, there are examples of this in today's environment. The Dow Jones Industrial Average, for example, has been used as a measure of U.S. market performance for the last 100 years, but almost no one uses it as a benchmark for investment managers. As a result of the distinction between indices and benchmarks, the TSE 300 Total Return Index may still represent the performance of the broad Canadian equity market, but some investors may decide that it is not an appropriate benchmark. If, for example, one views the criteria of being "investable" as complying with prevailing legislation, then given our current legislative environment, the TSE 300 may not be investable and therefore not a good benchmark. Others may argue that the TSE 300 remains the best indicator of broad equity market movements, and will continue to use it subject to legislative compliance. Finally, the issue of benchmarking must be addressed at different levels within a pension fund. Each manager must be given a benchmark. Most asset classes also need a benchmark in order to perform attribution analysis. The same holds true for the total fund. The appropriate benchmark chosen by a plan sponsor may be different at various levels. For example, most pension funds use the S&P 500 as a benchmark for their U.S. equity holdings at the overall fund level. However, if a multi-manager structure is employed, the managers may very well have style- and/or capitalization-specific benchmarks in order to measure their performance more effectively. In Canada, a plan sponsor may adopt the TSE 300 as a fund or asset class benchmark, perhaps capped at a specific percentage. However, the managers - depending on their strategy (active vs. passive) and style (core vs. value vs. growth, etc.) - may also have the TSE 300 as a benchmark, but capped at varying levels, where in aggregate the managers' capped levels add up to the overall fund capped percentage. WHAT IS AN APPROPRIATE MAXIMUM LEVEL FOR A SINGLE STOCK?<br> This question is at the heart of risk management. It is also a question that naturally arises when markets become more concentrated. At this time, there does not appear to be a consensus answer - neither domestically nor abroad - within the pension investment industry. There are several sources, however, which are outlined later in this section that do address this question. For a specific pension plan, there are many factors that may influence the appropriate single-stock maximum, including:
For example, Scotia Capital prepared a research paper on this topic earlier this year, supporting a 7% maximum single stock exposure, as a percent of a total Canadian equity portfolio. In the paper, the authors state that "Going beyond 7% results in a return/risk ratio that is worse than the initial diversified index, thus we would not advocate randomly overweighting a stock beyond 7% - even if the index is overweight - unless the investor believes the potential returns for this stock really justify the dramatic increase in risk. We believe that a reasonable benchmark should have a return/risk profile that minimizes non-systematic risk for the average investor that by definition can not benefit from superior stock selection." It should be noted that the Scotia Capital paper calls for regulators to increase the current 10% limit to at least double their 7% limit. The explanation given is that "investors can potentially benefit from the increased returns that can be had from less than optimal portfolio diversification combined with superior stock selection." Another source on this issue is the report entitled 'Rebuilding Pensions - Recommendations for a European Code of Best Practice for Second Pillar Pension Funds" authored by Pragma Consulting. This report is explored in more detail in the next section. Recommendation 24 of this report states that "There should be an obligation to diversify, which implies that no more that 4% of the equity part of any pension fund can be invested in a single stock . . ." This recommendation arises from 125 responses to a related question put to pension plan sponsors around the world. Finally, many plan sponsors have single stock limits articulated their investment policy statements, ranging from 5% to 10%. Some plan sponsors have relied on pension legislation to set these limits (although legislative limits are at the total fund level), while others have set these limits using either a common sense approach to prudence or some level of quantitative analysis. In any event, the 10% maximum has served as a guide for sponsors in setting plan-specific limitations. Therefore, there is no consensus approach to setting appropriate single stock maximum. The three approaches briefly outlined above result in limits ranging from 4% to 10%. However, some other sponsors may be comfortable with a higher percentage. IV. GLOBAL DEVELOPMENTS The concentration in equity markets is by no means limited to Canada. In fact, only the largest equity market (U.S.) has escaped the global concentration phenomenon. Even Japan is relatively concentrated, with the top three stocks representing almost 20% of the market at the end of 1999. (This level of concentration in Japan has continued to increase in 2000) We have spoken to our colleagues in the U.K., New Zealand, Australia, Ireland and the Netherlands, and they have provided insights as to how investors in their respective countries are dealing with market concentration. The following chart summarizes the level of concentration within the largest global equity markets:
(As of December 31, 1999) What is the effect on volatility in increasingly concentrated stock markets? Research conducted by Sanford Bernstein (and others) shows that market concentration does indeed materially impact volatility, as shown in the chart below. And where would Canada plot on this chart? The historical annual volatility of the TSE 300 has been approximately 16% (1957 to 1999). Volatility over the last 10 years has actually been lower (15.5%) than long-term volatility. Therefore, from a historical perspective, Canada would plot slightly below the regression line, given that the weight of the largest three stocks is approximately 31%.
The fact that volatility is positively correlated with market concentration levels is the predominant risk management problem to be addressed by plan sponsors. Noteworthy is Italy and Japan's absolute volatility, indicating that concentration in and of itself is not the sole source of volatility. Conversely, higher concentration does not always imply higher volatility (e.g. Australia and the Netherlands). Continuing our analysis of global equity market concentrations, the chart below illustrates more recent figures for the United Kingdom, Germany and the Netherlands:
All three of these markets have experienced increasing market concentration during 2000. In addition, market concentration has increased in France, Switzerland and Italy, where the top 15 stocks now account for 67%, 93% and 94% within each of these markets, respectively. New Zealand is also experiencing equity market concentration similar to Canada, with NZ Telecom comprising 28.8% of the NZSE40 index. Given these concentration levels around the world, one would assume that plan sponsors in many countries would all be affected in terms of overall risk management. However, this is not the case. Consider the following average asset mixes of pension plans in various countries: Average Asset Mix of Pension Funds
* May include group insurance assets, GIC's and other assets. Apart from the U.K., and possibly the Netherlands, this increase in market concentration has not had a significant impact on plan sponsors' overall risk profiles in most European countries. Moreover, Germany and France are not 'mature' pension markets. A point of clarification on Germany and France is needed. German plans typically account for their pension plans via book reserves. As such, not many pension plans in Germany actually have large pools of assets to invest - the data in the above chart is an average of a relatively small sample. The French situation is much the same, although for different reasons. Most of the pensions in France are provided through the public pension scheme. With the exceptions of those offered by large organizations, private pension plans are a relatively recent phenomenon. Until recently, they were typically group insurance plans or book reserve plans. Therefore, only a few countries' experiences provide further context to the Canadian situation. Given these developments, we will look at two countries (New Zealand and the U.K.) that will give insights as to how others have dealt (or are dealing with) market concentration. NEW ZEALAND New Zealand has one stock, NZ Telecom Ltd., which represents 29% of the market capitalization of the local stock index. Prior to July 1991, the average level of domestic equities in New Zealand pension plans was 30%, and the most common limit on a single company holding was 5% of the total fund. Therefore, a publicly-listed company would only have to be approximately 17% of the domestic market for this restriction to prevent a market weight of this security in the fund. The concentration problem in New Zealand started after July 1991, when NZ Telecom was initially listed and soon grew to 20% of the index. It has since continued to experience strong returns and has grown to its current levels. The approach adopted by New Zealand plan sponsors has been as follows:
UNITED KINGDOM Canada is not the only country to consider a capped index. In the U.K., the WM Company (a performance measurement company) has launched and made available to clients a 5% capped version of the U.K. index. FTSE has also produced a capped index (10%). To our knowledge, these indices have not to date received wide acceptance in the pension community. Recommendations coming out of two major U.K. consulting firms have included increasing foreign equity holdings or adopting a new multi-national index as a benchmark for many of the listed companies that have been the source of the U.K.'s concentration problem, while leaving the other stocks in the existing indices as they are. No consulting firms appear to be supporting the capped indices. However, it seems investors other than pension plans have created a demand for capped indices in various countries. In an article by Sandy Ratray of Goldman Sachs (April 17, 2000), German, Dutch and Finnish investors are mentioned as having created sufficient demand for capped indices and that these capped indices now do exist. EUROPEAN COMMISSION While not directly related to the issues under discussion, the report "Rebuilding Pensions - Recommendations for a European Code of Best Practice for Second Pillar Pension Funds" should again be referenced. This report, which was produced by Pragma Consulting and can be found on the European Commission's website, contains the results of 125 institutions' responses to a very detailed questionnaire on pension issues. Respondents were from across the European Union (EU) as well as the United States of America. The objective was to provide recommendations for a Code of Best Practice for European Pension Funds. While many readers of the document may find the entire report useful, the following four excerpts, as they relate to investment restrictions, are particularly relevant given the mandate of this paper.
V. IDENTIFICATION OF ALTERNATIVES GENERAL DISCUSSION Prudent pension asset management involves concepts of diversification, risk management and maximizing risk-adjusted returns. Sophisticated tools available today (such as asset/liability modelling) enable the measurement of risk and thus the development of approaches to manage such risk. From an asset/liability perspective, an investment portfolio will reflect the split between fixed income and equities that best meets the objectives of the plan sponsor. Asset/liability techniques can further help in structuring the bond portfolio (term and economic structure) to align better the growth between pension assets and liabilities. The appropriate domestic/foreign equity split can be developed by the use of efficient frontier analyses. An efficient frontier shows how different asset classes can be combined to optimize returns and minimize risk. This reflects expectations (or historical information) regarding returns, volatility and correlation of major asset classes. Users of efficient frontiers had to be careful in the past in setting assumptions, as historical volatility and correlations have varied over time. Given the current market concentration, a user of efficient frontier models must be even more careful when setting volatility levels. It is reasonable to assume that the expected volatility for Canadian equities will increase versus historical levels, and assumptions in efficient frontiers should be consistent with this current market reality. In Section II (Background), we discussed the issue of prudent levels of holding for a single stock as well as the issue of foreign equity content. As previously stated, increasing foreign equity exposure within a pension fund would increase diversification at the total fund level and reduce the relative holdings in Nortel at the same time. The options presented in this paper assume that the process of establishing the proper level of foreign equity within the pension fund has been completed by the plan sponsor. The levels of foreign equity will be dependent on the nature of plan liabilities, financial position of the plan, demographics of the membership, and expectations with respect to interest rates, equity returns, market volatility and correlations. A plan sponsor's beliefs with respect to risk will drive the decision as to which alternative is selected. RISK MANAGEMENT Risk management is fundamental to the ongoing operation of a pension plan, transcending almost all decisions that sponsors need to make to ensure the prudent and effective management of their pension funds. Within the context of this report, the discussion of risk management is confined to security concentration, although some of the concepts raised may apply to other aspects of plan management. Risk, with respect to single-stock concentration, can be managed from four different levels:
There is no reason why a plan sponsor must choose only one of these levels. Moreover, there is no right or wrong approach to selecting a level from which to manage risk or single-stock concentration. The selection is driven by the specific situation of a plan sponsor (e.g., the financial position of the plan may impact the selected level(s)). However, what is important is that the level(s) selected be consistent with the plan sponsor's investment beliefs, and guide the ultimate selection of the alternatives for managing risk in the Canadian concentrated market. For example, if a plan sponsor chooses to manage risk from the total fund level, then any or all of the following may be inferred regarding the sponsor's investment beliefs:
The list is not exhaustive. Similar lists could be prepared for the other levels. They demonstrate that what may seem like a fairly simple decision may actually imply a lot about a plan sponsor. It is important, then, that a plan sponsor clearly articulates the selection of the level(s) from which risk will be managed, as well as the rationale for this selection (i.e., a clear articulation of investment beliefs). ALTERNATIVE OPTIONS Alternatives for managing risk in a concentrated market are:
Analyses of the alternatives follow. All are worth considering; some, however, may be difficult to implement at this time. Implementation may depend on the ability of custodians to track single holdings across managers and portfolios, and to aggregate holdings at the total fund level to facilitate appropriate portfolio construction responses. The next four sections will:
VI. ALTERNATIVE 1: USE THE TSE 300 DESCRIPTION Canadian equity mandates would continue to be benchmarked against the TSE 300. DISCUSSION It is important to recognize that there is more than one way of arriving at this alternative. In fact, one can arrive at this choice for Canadian equities from any of the risk management levels outlined in Section V. From the total fund or total equity level, this decision could be based on the view that, given the plan sponsor's current Canadian equity exposure level, single-stock risk does not need to be managed. For example, assume a plan sponsor with a 10% total fund limit (at market value) for a single holding adopts a policy mix that contains 20% Canadian equities, and the sponsor manages risk from a total fund perspective. In this case, unless a stock reaches a 50% weight in the TSE 300, the sponsor will not need to adjust the benchmark. For sponsors managing risk from the specific asset class and/or manager levels, this alternative could also be chosen because the sponsor believes that the TSE 300 index is the best measure of activity for the broad Canadian equity market ("the market is what the market is"). Of course, this alternative may only be chosen if the resulting stock concentration does not contravene any applicable legislation. ADVANTAGE The advantage of this option is reasonably clear - simplicity and ease of communication. Given that sponsors selecting this option are probably already using the TSE 300 as the benchmark, no changes need to be made (with respect to manager benchmarks), although the reasons behind this status quo approach should be articulated. DISADVANTAGES The disadvantages of this option may vary, depending on the reasons for selecting it (see "Discussion" above). For plan sponsors managing risk from the total fund or total asset class level, there may be no perceived disadvantages if the sponsor believes that the concentration risk in Canada has been managed away using other techniques such as increasing foreign equity exposure, smoothing of asset values, establishing contingency reserves, conservative actuarial assumptions, etc. Regardless of the level from which risk is being managed, there are a number of disadvantages to this approach.
IMPLICATIONS Regardless of the reason for selecting this alternative, implementation is easy - most sponsors would simply continue with existing benchmarks and mandates, keeping in mind that overall asset allocation might be revised to recognize changes in market concentration. If the market subsequently becomes less concentrated, no shifts would likely be required. If the market becomes more concentrated, then those managing risk from the manager or specific class level would need to review their original decisions. SUMMARY The following summarizes the rationale and implications of selecting the "status quo" approach in managing risk at a given level:
VII. ALTERNATIVE 2: CAPPED INDEX DESCRIPTION Use an "industry standard" capped index, such as the TSE 300 Capped Index developed by Standard and Poor's or CPMS. DISCUSSION This option takes the approach of modifying the standard TSE 300 Composite Index by limiting any one stock to a maximum weight of the index. A new index is thus created which is maintained on an ongoing basis in parallel with the TSE 300 Composite Index. Using the TSE 300 Capped Index proposed by Standard and Poor's as an example, the mechanics of the process would be as follows:
The key question for those selecting this option is "What is the correct level to cap single security exposure?" The answer to this question is very plan specific. However, there may be a common approach that plan sponsors can use to establish this level. Consider the following example. The ABC plan has a 60% equity, 40% fixed income split, with 40% in Canadian equities, 10% in U.S. equities and 10% in non-North American equities. For this example, we will assume that Nortel has a 30% weight in the TSE 300 and 2% in the S&P 500. Consider the following matrix: Maximum Nortel Weight at Various Levels
Plan sponsors can develop similar matrices for their specific asset mix policy and then evaluate what level of exposure is reasonable for them. Regardless of which level risk is being managed, the main reason that this policy would be selected is a belief that, above a certain level, single stock concentration is not prudent and exposes the fund to excessive risk. ADVANTAGES Certainly the main advantage of a capped index approach is the limits it places on potential downside risk associated with excessive concentration in one stock, at both the total fund and manager level, irrespective of what total fund allocation a sponsor makes to Canadian equities. There are a number of other potential advantages. Some active investment managers are asking plan sponsors for more latitude in dealing with their investment policies. The reason? For the most part, the managers are measured against the TSE 300 for Canadian equity performance. They need to be able to over- or under-weight stocks in the TSE 300 if they are going to (potentially) outperform the index. A capped index can represent a benchmark against which investment managers can be judged reasonably fairly. With a capped index, a manager can invest fully in all elements of the index without running into constraints imposed by plan sponsors to address concerns about a concentration of risk. In this sense, it meets the criterion that a benchmark should be investable. To be effective, a benchmark should be liquid. For most funds in Canada, a capped index should be able to meet their liquidity needs. In fact, for the very largest funds, an uncapped index may itself present liquidity problems when trying to move to the market weight of a concentrated stock, which would not be the case if a capped index was employed. Two additional key attributes of a proper benchmark are that it represents a reasonable proxy for the overall market and that it represents a relatively stable risk profile. As a reasonable proxy for the market, even though a capped index may significantly underweight one stock relative to its market weight, it may be more representative of an investable market profile for most institutional investors than an uncapped index. A capped index may also exhibit a more consistent risk profile from year to year as compared with an uncapped index dominated by one stock, especially if that one stock itself is undergoing significant changes in risk profile as it develops and grows rapidly. By adopting a capped index, a pension fund prevents managers from carrying excessive concentrations in one stock. A manager who is judged against an unconstrained index may take on significant benchmark risk when he or she adopts a large underweight position in one major stock. To minimize this risk, the manager may be forced to carry a higher concentration in this stock than intended. Consider a hypothetical situation where Nortel becomes de-listed in Toronto and is only listed in the U.S. One could argue that few managers would voluntarily hold the equivalent of the current Canadian market weight in such a stock. If so, holding a high percentage of a portfolio in Nortel may be an investment decision driven not by the merits of Nortel but by the benchmark. Plan sponsors may view a capped index as a practical solution to the concentration issue for pension funds. Managers can have mandates that are clear, and the monitoring process becomes much simpler. Pension fund fiduciaries can fulfil their obligations for prudent diversification and at the same time judge and compare their managers fairly. DISADVANTAGES Some argue that a capped index is an artificial construct that is designed in an arbitrary way to meet a specific problem that can be managed in other ways. If that problem ceases to exist or changes significantly (as it has in the past with other concentration problems), then the new capped index loses its relevance. Ideally, a good benchmark should have longevity and a fundamental rationale in its construction. A capped index could fall short on both accounts if the concentration issue in Canada disappears in the future. Another disadvantage of capped indices that relates to their construction, is that it is more complicated than an unconstrained index. It is prone to adjustments both on the regular rebalancing and at unpredictable times when the constraint boundaries are breached. This can add increased amounts of volatility to the market as indexers and other investors attempt to manage against the index. In addition, by focusing on a capped index, Canada may be out of step with the rest of the world. As mentioned previously, the concentration issue is not unique to Canada and has occurred in the majority of world markets. However, even though capped indices already exist in other countries, they have yet to be embraced by institutional investors. Another argument against a capped index approach relates to the efficient market theory. This theory would support the premise that a natural and expected outcome of the markets going forward is that more successful companies will become a larger component of an index than unsuccessful companies. Therefore, by artificially reducing the impact of successful companies, a sponsor is interfering in the market and sub-optimizing investment performance. Will focusing on the capped index as the benchmark encourage investment managers to focus on an index that is sub-optimal? This could lead to investment returns for pension funds that are less than they should expect. Nortel is such a large weight of the TSE 300 because it is a highly successful company with a large global organization that is well placed to perform above the market norm in the future. If other Canadian stocks surge above the 10% level in the future, it will probably be due to global competitiveness as well. By reducing Nortel's impact artificially, are we lowering the benchmark and our expectations unnecessarily? This may lead in turn to a situation where the uncapped TSE 300 Index would outperform the capped TSE 300. Managers who are free to invest in the uncapped index would have a greater opportunity to outperform the capped index. This would no doubt confuse peer comparisons (or at least force the creation of multiple Canadian equity universes to account for capped and uncapped managers) and may also further complicate the issue of determining the potential value added of active management. A final disadvantage is that by adopting a capped index as its equity benchmark, a pension fund is solving a potential problem of over-concentration of risk at the individual investment manager level. There are other methods of solving the problem. These include emphasizing more global exposure, or dealing with the issue at a policy level. By following the capped index course, the fund may overcorrect its concentration problem because only part of the total fund is invested in Canadian equities. For example, if 35% of the fund is allocated to Canadian equities, a capped index approach (say 10%) would essentially target 3.5% of the total fund as the appropriate exposure to a single issue. Following this approach may be overly cautious in regards to the issue of concentration of risk at the total fund level. IMPLICATIONS There are a number of implementation issues to be addressed if the benchmark is changed. Investment managers must be given time to adjust their portfolios. Otherwise, rapid shifts will incur transaction costs and a discontinuity in performance that will need to be explained into the future. If implementation is conducted at the manager level, the plan sponsor will not need to worry about rebalancing at the total fund level. If, on the other hand, a capped index approach is implemented at the total fund or asset class level, sponsors will need to spell out responsibilities and procedures for an effective rebalancing policy to deal with those times when the capped weight is breached. Finally, the cap level may vary if risk is being managed at different levels. For instance, if a plan sponsor decides the appropriate maximum cap level is 15% of the Canadian equity portfolio, and employs one passive Canadian equity manager (25% of equity portfolio), and two active Canadian equity managers (75% of portfolio - split evenly between the two active managers), and assuming a 30% index weighting to Nortel, the manager-specific caps could be as follows:
However, if the active managers are capped at 10% and are measured against a 10% capped index, they are not allowed to overweight Nortel. Some sponsors have already resolved this issue by providing a capped index benchmark at a percentage lower than the maximum permitted fund allocation to a specific stock. For example, if a manager is capped at 15% of one stock, perhaps a 10% capped index might be a more meaningful benchmark. SUMMARY The following summary is the rationale and implications of selecting a capped index approach to manage risk at a given level:
VIII. ALTERNATIVE 3: SEPARATE ALLOCATION TO NORTEL DESCRIPTION Use separate allocations for Nortel and the TSE299. There are two ways to approach this alternative strategy. First, as described above, the manager is given a Nortel weighting to maintain within a range and a TSE 299 for the remainder of the portfolio. Second, the manager could be given a portfolio to manage that only has a TSE 299 benchmark and is prohibited from purchasing Nortel stock. The Nortel exposure is managed on a total fund level at whatever threshold within which the sponsor is comfortable, but is separate from the Canadian equity portfolio. The discussion focuses on the first approach, but many of the points raised in this section are also applicable to the second. This option would involve a Canadian equity mandate made up of x% of Nortel stock and (100-x%) of TSE 299. Also, ranges around both allocations would normally be specified. Therefore, a Canadian equity mandate could be described as follows:
DISCUSSION This option requires the computation of a new benchmark, the TSE 299. This is not currently available, although should be relatively straightforward to generate. In addition, total returns on Nortel should become readily available. Some consider Nortel large enough to warrant a separate allocation within the portfolio. Of course, the reason Nortel's weighting is so large in many Canadian pension funds is due to the large allocation to Canadian equities overall. The need to treat Nortel as a separate "asset class" diminishes as foreign equity holdings rise. Few in Canada would need to concern themselves with this matter with other large global players (e.g., Cisco, Nokia, Vodafone) simply because their weights at the total fund level are relatively small. However, given many plan sponsors' high allocation to Canadian equities, what is the correct allocation to Nortel? There have already been a number of studies conducted on this subject of treating Nortel as a separate allocation. If one supports this alternative, the consensus approach is to model Nortel as a separate asset class using assumptions as to future return, risk and correlation relationships to determine quantitatively, when combined with the other asset classes, what risk-adjusted return pattern is acceptable to the plan sponsor at the total fund level. One caveat to this approach relates to the setting of these forward-looking Nortel assumptions. We all can agree that Nortel is a very different company than it was five to ten years ago (even two to three years ago, for that matter). Therefore, reduced reliance on historical risk/return characteristics and more reliance on recent history and global experiences in other concentrated markets may be advisable. The plan sponsor will need to decide whether the allocation between the TSE 299 and Nortel should be managed on an active or passive basis. This decision will depend on the sponsor's overall approach to asset mix management. Plan sponsors will also need to determine if this alternative approach is consistent with their philosophy towards risk management in terms of what level of risk the sponsor emphasizes. Finally, if sponsors adopt this approach to managing risk in a concentrated market, they will need to review their decision-making process to ensure this approach is consistent with their current governance structure. ADVANTAGES Plan sponsors who wish to raise the Nortel allocation decision to the policy level may find this option appealing. The sponsor assumes the decision as to the target allocation to Nortel. This is primarily a risk control exercise that can be conducted using input from the sponsor, the investment managers and consultants. The ultimate decision on the target allocation will reflect concerns on benchmark risk and views on a prudent approach to equity investments. The manager does not need to identify the quantity of single stock they wish to hold for benchmark risk management purposes. Also, the investment managers' hands are not entirely tied with respect to Nortel. They are allowed to take higher or lower positions on a strategic basis and be rewarded for it. Nortel aside, the managers can apply skills to the remainder of the Canadian stock pool in a manner that would be consistent with their current investment approach and style. Evaluating the manager performance is also relatively easy. The manager can be evaluated on two levels:
Finally, this approach works well in a multi-manager structure. For example, the approach can be applied to each manager by assigning target weights and ranges for Nortel and the TSE 299. DISADVANTAGES Some investment managers will have concerns about this approach. Managing to the TSE 299 will require difficult transitions for many of them. Pure bottom-up managers will have an easier time with this approach. Those using a top-down approach will have some implementation difficulties. Some plan sponsors may perceive this approach as involving a level of precision that is beyond the scope of pension fund management. If we separate out Nortel, why not other securities? What level do we set to determine if a stock warrants treatment as a separate allocation within a portfolio? In addition, this approach may have to be revisited on a frequent basis - at least as often as regular policy reviews. If concentration levels within the Canadian equity market change materially (either up or down), the market shifts will trigger even more frequent reviews. Will this approach create more work than the value or benefits that may (or may not) be gained? Finally, does Nortel meet the criteria to be treated as a separate "asset class"? As outlined previously, distinct criteria exist in terms of defining what constitutes an acceptable benchmark. The same holds true for what defines a separate asset class (distinct return, risk and correlation characteristics for a prolonged period of time). Nortel's weight in the index alone may not be sufficient justification for treatment as a separate allocation in a portfolio. IMPLICATIONS Conceptually this strategy is reasonably straightforward, but a number of implementation issues need to be addressed. First, the purpose of this approach is simply to restrict the holdings of a single stock. In its generic form, it could be triggered when one stock exceeds a certain threshold, reverting back to the TSE 300 approach otherwise. These changes should be discussed with the manager beforehand so as to make the mandates manageable. Clear policies and procedures would need to be developed with respect to the frequency of the reviews of this approach and at what threshold are separate allocations triggered. Tolerance levels would be required to ensure the approach and benchmarks are not changed too frequently when a stock hovers above or below the chosen threshold level. Second, what should the ranges be around the target allocations to Nortel and the TSE 299? Some sponsors may prefer broad ranges to avoid too frequent rebalancing. Others may opt for tight ranges and regular rebalancing programs to eliminate any market timing component. The ultimate discussion should link back to the plan sponsor's investment belief system with respect to active and passive management and the potential value added from short-term "asset class" shifts. Finally, this approach may not make sense for those plan sponsors managing risk from the manager or Canadian equity level. It appears to be a more consistent approach for those managing concentration risks from the total fund level. SUMMARY The following summary is the rationale and implications of selecting this approach to manage risk at a given level:
IX. ALTERNATIVE 4: GLOBAL APPROACH DESCRIPTION Use global approach to manage overall equity portfolio. This option redefines the universe of stocks open to equity managers and can be viewed in one of two ways. The first is to view the world from the global sector perspective and define investment limits with respect to these sectors. Managers would then implement their strategy with reference to these sectors and make appropriate geographical weighting decisions on this basis. The second view assumes that the portfolio of the Canadian investor can be built by considering the entire global equity market. Various benchmarks such as the Morgan Stanley Capital International (MSCI) World can be used for this purpose. Under the first approach, the goal is to define a process where geography remains important, but global sector considerations provide overall guidance. Under the second approach, the goal is to reduce reliance on the Canadian equity market. Given that we have already discussed the benefits of reducing reliance on the Canadian market earlier in this paper, we will focus this section on the first of these two approaches. A more detailed description of this global approach is as follows:
DISCUSSION Globalization is overtaking all industries. Consider, for example, recent, large mergers in car manufacturing, telecommunication, technology, pharmaceuticals, and others. Many brokerage houses now analyze companies in terms of their global competition as opposed to the historical focus on local market competitors. Qualitative measures also support this view; local markets are becoming more correlated to each other and so are correlations between local and global sectors. The natural extension is that portfolio managers will consider their "investable universe" as the global equity marketplace and define strategies accordingly. This would eventually lead to a sector/industry approach along global lines. Although the investment industry is beginning to view the global equity markets divided along industry/sector lines, there currently exist few pure global sector products. If this approach were adopted, it would probably involve a combination of bottom-up and top-down portfolio construction from both a regional and industry/sector perspective. For this reason, of the four alternative approaches outlined in this research paper, this global approach is the most difficult to implement at this time because global sector products and benchmarks are not readily available. However, as the investment industry gradually moves to a more global sector framework, this alternative will become easier to implement. Despite current implementation difficulties, we believe it is important to consider future strategies that may indeed transform the current landscape of the institutional investment industry. To this end, there would be a number of advantages and disadvantages associated with a global strategy. ADVANTAGES
IMPLICATIONS Adopting this approach amounts to a significant shift in investor attitude towards equity investing. Even if the Nortel issue were to disappear, investment behaviour would have changed to embrace broader equity universes. SUMMARY The following summary is the rationale and implications of selecting this approach to manage risk at a given level.
X. IMPLICATIONS FOR DC PLANS BACKGROUND The fundamental feature of DC arrangements, whether a DC pension plan, group RRSP, DPSP or after-tax savings plan, is that the plan participants assume the investment risk. In return for assuming the investment risk, participants fully participate in the investment returns earned under the plan. By selecting the plan investment options which will impact on the retirement pension of the plan participants, the plan sponsor acts in a fiduciary capacity. Sponsors and members of DC plans may have more exposure to market concentration than DB plan sponsors. DB plans have, in theory, a constant time horizon while each DC individual account has a shrinking horizon. Market concentration is a risk issue which becomes more material as the time horizon shortens. If the stock(s) creating the market concentration experience a period of poor performance, the impact on DC members will be significant. DC plans place the onus on individuals to manage their investments within the framework provided by the plan sponsor. However, plan members may well not have the knowledge, experience or time to properly address market concentration issues. From the plan sponsor's perspective, the balance is delicate between allowing members to have control over their investments and providing too many options. Plan sponsors need to provide an effective education program for plan members to allow them to make informed investment decisions. In this concentrated environment, there are several responsibilities that need to be addressed by sponsors, including:
At present, there is no Canadian legislative equivalent to the U.S. safe harbour rules, as outlined under Section 404 (c) of ERISA. The ERISA guidelines set out a minimum threshold for investment offerings to DC plan members, to assist plan sponsors in discharging their fiduciary obligations. In absence of formal legislative guidelines, many Canadian plan sponsors have looked to the ERISA safe harbour rules for guidance. The guidelines are designed to ensure that members are given options that allow for diversification at both the asset class level and at the policy level. To comply with the guidelines, plan sponsors must offer at least three diversified market-based funds, each having materially different expected risk and return characteristics. A typical range of investment options for a Canadian DC plan might include a guaranteed fund(s), a money market fund, a bond fund, a balanced fund, a Canadian equity fund, a U.S. equity fund and an international equity fund. In some cases, multiple funds may be offered in certain asset classes. The impact of market concentration on a DC plan will be influenced by the following factors:
ALTERNATIVE INVESTMENT OPTIONS We are addressing the issue of alternative investment options only as it relates to equity investment options. These alternatives can be extended within the context of balanced options.
In this case, all existing options would remain and plans would proceed status quo, and foreign content levels would rise as regulatory limits increase. The only apparent advantage to this approach is that it requires that little be done. There is, on the other hand, more that one disadvantage:
This approach adds on an additional (capped) Canadian equity option to the DC investment program (and, possibly, an additional balanced and total equity option as well). The 'cap' would either be consistent with what has been adopted for the DB plan (if there is one), or consistent with an available benchmark (Standard & Poor's or CPMS). Most likely, the cap would be either 10%, since this is the cap level that the benchmarks have adopted, or 15% to allow managers the opportunity to overweight Nortel versus the capped benchmark. The advantages of adopting this approach are:
The plan sponsor is not making a value or prudence decision for the plan members, but rather providing additional investment vehicles to allow members to exercise their own discretion. The disadvantages of the approach are:
In this approach, the first decision required is what should be the maximum level of Canadian equity. There are clearly many ways of determining this level. The following three may have some appeal.
The same advantages and disadvantages apply as in option 2. An additional disadvantage of this approach is that in the short-term, there is a potential shortage of existing products of this nature which are available for the DC market. Structuring a customized product for a small-to-mid-sized plan could be quite expensive. Plan sponsors would select this option over option 2 if they believe:
Sponsors would select option 2 over option 3 if they believe:
The advantage of these approaches over their earlier counterparts is that they reduce complexity of implementation, administration and monitoring as well as ensuring that members with identical asset mixes would experience identical (time-weighted) rates of return. The disadvantage of these approaches is that the sponsor may be viewed as imposing their own views of prudence on the plan member. EDUCATION A critical responsibility of a DC plan sponsor is to ensure that the members receive effective investment education. This helps to ensure that plan participants are provided with the information needed to make informed investment decisions. A critical element of a DC education program would normally include the following:
Market concentration in the Canadian equity market and the potential impact on individual DC members' portfolios is an important issue for DC members to understand. Regardless of the option chosen by the plan sponsor for dealing with this issue, the following issues should be incorporated into member communications:
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
