MSIM Journal Q3 2006

Morgan Stanley Investment Management Investment Management Journal 01 The emerging pension paradigm—part 1 Michael Peskin—Managing Director Chad Hueffmeier, CFA—Vice President The confluence of mounting corporate alarm over the cost of defined benefit pension schemes and the coming of age of derivatives is driving one of the few trends worthy of being called a paradigm shift. 07 Emerging markets domestic debt: A new opportunity in fixed income Federico Kaune, Ph.D.—Executive Director Eric Baurmeister, CFA—Executive Director While improvements in emerging markets’ macroeconomic fundamentals have been reflected in hard currency bonds, local bond markets have been slower to react and may represent good value. 15 Alpha returns and active extensions Martin L. Leibowitz, Ph.D.—Managing Director Anthony Bova, CFA—Associate Active extensions can be viewed as an “extended” form of traditional active equity management with the potential to improve alpha returns. 27 Pitfalls and risks in portable alpha implementations Jack Coates, Ph.D., CFA—Managing Director Mark Baumgartner, Ph.D., CFA—Executive Director By understanding both the real and the perceived risks involved, investors can make a more informed decision about the application of portable alpha in their portfolio. Issue 2 37 Risk budgeting in an Asset-Liability Management context Volume 2 Jan Baars, Ph.D.—Vice President Petr Kocourek—Executive Director Epco van der Lende, Ph.D.—Executive Director Determining a risk budget as an integral part of an Asset-Liability Management study provides internal consistency between long- and short-term aspects of investment policy. 59 About the authors 2006 The forecasts and opinions in this piece are not necessarily those of Morgan Stanley Investment Management and may not actually come to pass. Information in this report does not pertain to any Morgan Stanley product and is not a solicitation for any product. The views expressed are those of the authors at the time of writing and are subject to change based on market, economic and other conditions. They should not be construed as recommendations, but as illustrations of broader economic themes. All information contained within is based on past performance and is not intended to be indicative of future results. All information is subject to change. Morgan Stanley does not provide tax advice. The tax information contained herein is general and is not exhaustive by nature. It was not intended or written to be used, and it cannot be used by any taxpayer, for the purpose of avoiding penalties that may be imposed on the taxpayer under U.S. federal tax laws. Federal and state tax laws are complex and constantly changing. You should always consult your own legal or tax advisor for information concerning your individual situation. Equity securities are more volatile than bonds and subject to greater risks. Small and mid-sized company stocks involve greater risks than those customarily associated with larger companies. Bonds are subject to interest-rate, price and credit risks. Prices tend to be inversely affected by changes in interest rates. Unlike stocks and bonds, U.S. Treasury securities are guaranteed as to payment of principal and interest if held to maturity. REITs are more susceptible to the risks generally associated with investments in real estate. Note that it is not possible to invest in a market index. THIS MATERIAL IS PREPARED FOR INSTITUTIONAL INVESTOR USE ONLY. It may not be reproduced, shown or quoted to members of the general public or used in written form as sales literature; any such use would be in violation of the NASD Conduct Rules. THIS MATERIAL IS PREPARED FOR INSTITUTIONAL INVESTOR USE ONLY November 2006 Investment Management Journal • 1 Michael Peskin Managing Director Morgan Stanley Investment Management PG.01 Chad Hueffmeier, CFA Vice President Morgan Stanley Investment Management The emerging pension paradigm—part I The entire approach to providing for retirement is changing. In particular, corporate retirement benefits are moving rapidly from defined benefit (DB) to defined contribution (DC), with DB plans being frozen in the process. However, not all DB plans are being put on ice and even those that are will be around for a long time. The financial management of these ongoing plans, in particular the investment strategy, is also undergoing change. The extent of this change is dramatic—indeed, it is one of the few trends worthy of being called a paradigm shift. This shift in the investment of DB assets is global and is the product of two interacting, independent and powerful forces. The first force is specific to corporate pensions and can loosely be described as the growth of transparency and investor sophistication, leading to the proper pricing of pension plans and hence the equity pricing of plan sponsors. The second force applies to all investors and can loosely be termed the growing ability to separate alpha and beta returns. This article will describe the existing pension investment paradigm, the logic and structure supporting it, and the major challenges it faces. The existing paradigm cannot survive these challenges, and a new paradigm is in the process of emerging. We are currently analysing this emerging paradigm and the probable shape it will take. We will discuss our conclusions in a subsequent issue of the Investment Management Journal. 2 • Investment Management Journal November 2006 THIS MATERIAL IS PREPARED FOR INSTITUTIONAL INVESTOR USE ONLY PENSION COSTS COME TO LIGHT It is not surprising that, until recently, pensions passed under the radar. During the 1990s, the U.S. stock market had phenomenal performance and global interest rates were high (relative to the current environment). Consequently, by 2000, the “funded status” of DB plans—measured as the fair value of pension assets divided by reported liabilities—for the average company in the S&P 500 Index was approximately 130%, and several plan sponsors had enjoyed “contribution holidays” for several years. Furthermore, pension plans were not levying large charges on income statements; instead they were providing additional income and becoming “profit” centers. Few companies were focusing on managing pension costs. Following the 2000-2002 bear market, however, deficits replaced surpluses, pension expense replaced pension income and contribution holidays were over. Pension costs gained the attention of CFOs, creditors, analysts, investors and regulators. Many companies have chosen (or are considering) to address the issue directly by reducing or eliminating the growth in pension benefits through some form of freezing. Note that this approach slows the growth of companies’ exposure to pension cost and risk, but it does not reduce the exposure that already exists. This rethinking is occurring at the same time as the capital markets are evolving in response to the rapid growth of derivatives and the lower return world that faces us. TRADITIONAL PENSION PARADIGM allowing companies to book expected returns into corporate income in advance, while only slowly booking deviations from expectations. These regulations and standards encouraged and enabled plan sponsors to ignore short-term volatility and take what many in the industry have called a “longterm” view. Many still support this approach. TRADITIONAL PENSION INVESTMENT POLICY In the traditional pension paradigm, plan sponsors have typically followed the following approach: Identify asset classes to be considered. The first step of the process is for the plan sponsor to identify desirable asset classes and acceptable allocation ranges for each class (e.g., 20-35% for U.S. largecap stocks). Usually the asset classes are represented by passive benchmarks. Determine asset allocation. After asset classes and constraints have been identified, plan sponsors use a mixture of art and science to determine the strategic asset allocation. Modern portfolio theory has generally served as the foundation of the analytics and usually a long time horizon (five to 15 years) is used to represent the long-term view. The model identifies static long-term passive asset allocations that are expected to minimize risk for a certain level of expected reward. Traditionally, plan sponsors defined reward as the expected return on assets, and risk as the volatility of returns. Plan sponsors have increasingly incorporated plan liabilities into the definitions of risk and reward. An example is the use of expected pension surplus as a measure of return, and surplus volatility as a measure of risk. The modified process is typically referred to as asset-liability management (ALM) or liability driven investing (LDI). As previously mentioned, regulatory and accounting standards have allowed sponsors to take a long-term view and ignore short-term volatility. The long-term view has led almost all plan sponsors to rely heavily on the equity markets. As illustrated in Figure 1, the majority of S&P 500 companies have equity allocations between 55% and 75%. The primary motivation driving DB pension fund asset allocation has been to finance the benefits promised while minimizing the drain on corporate income and cash flow (cost). The core idea has been that tackling investment risk by investing heavily in equities would increase long-term returns and reduce long-term cost. This thought process and approach was widely embraced and embedded in funding regulations (e.g., ERISA) and accounting standards that allowed significant smoothing to enable equity risk taking. Accounting standards took the further step of THIS MATERIAL IS PREPARED FOR INSTITUTIONAL INVESTOR USE ONLY November 2006 Investment Management Journal • 3 FIG 1: REPORTED PENSION ALLOCATION TO EQUITIES—S&P 500 COMPANIES 80 70 Number of companies 60 50 40 30 20 10 0 Below 40% 40-44% 45-49% 50-54% 55-59% 60-64% 65-69% 70-74% 75-79% 80-84% 90-94% 95-100% Percentage allocated to equity Source: FactSet Data as of December 31, 2005 Furthermore, asset allocations do not tend to vary substantially across sectors. Figure 2 illustrates the average equity allocation by sector for S&P 500 companies. Select asset managers. After plan sponsors identify their strategic asset allocation, they select investment managers for each asset class. Multiple managers may be chosen for some classes in an attempt to further diversify styles and active management approaches. Each manager will be given a passive benchmark to manage against, and active managers will be allowed a tracking error to facilitate some risk taking to achieve excess returns. Monitor performance, hire and fire managers and revisit benchmarks. Once the asset allocation and managers are in place, plan sponsors have to monitor manager performance against the benchmarks. This process allows the plan sponsor to justify the retention or replacement of asset managers. In addition, plan sponsors have to revisit the overall FIG 2: AVERAGE EQUITY ALLOCATION 70 68 66 Equity allocation (%) 64 62 60 58 56 54 52 Basic materials Capital goods Conglom- Consumer Consumer erates cyclical non-cyclical Energy Financial Health care Services Technology Transportation Utilities Source: FactSet Data as of December 31, 2005 4 • Investment Management Journal November 2006 THIS MATERIAL IS PREPARED FOR INSTITUTIONAL INVESTOR USE ONLY asset allocation as objectives and economic environments evolve. This approach has been modified over the years but has, nevertheless, remained largely intact. The whole paradigm is now being seriously challenged. CHALLENGES TO THE EXISTING PARADIGM with 10-12 years a good estimate of their expected average maturity. Setting lower interest rates against long durations resulted in an increase in the expected pension liability of about 15-20%, even excluding the impact of future service accruals. The composition of pension asset portfolios exacerbated the decline. Equity markets fell over the same period; for example, the S&P 500 Index fell from almost 1500 to below 900 between the end of 1999 and the end of 2002. With approximately 6070% of the assets in equities, DB pension portfolios lost about 40% of their value. The bonds in pension asset portfolios were of little help—in the declining rate environment, the 30-40% portion of assets in bonds probably added only a few percentage points of gain to the portfolios, owing to the intermediate duration benchmark. The U.S. Pension Benefit Guaranty Corporation (PBGC) reported a swing from a large surplus to a large and growing deficit. This plunge in funded status got the attention of CEOs, CFOs, investors, rating agencies, Congress and equity analysts, as well as the press. The most important fallout from this is increased transparency, and changes in the funding and accounting frameworks. The capital markets are paying serious attention to pensions and the price for risk is being focused on more sharply. INCREASED TRANSPARENCY The challenges to the existing paradigm that are forcing change fall under two sub-headings. The first is with respect to pension plans’ part in the corporate capital structure, and the right way to manage these assets in the fully transparent and sophisticated world into which we are rapidly moving. The second concerns the changing landscape of the capital markets, where the expansion of derivatives is largely responsible for a growth in liquidity and completeness. These forces are reinforcing each other and we see rapid change ahead. We will start with the challenges specific to pensions within the context of the capital structure. MISMATCH RISK EXPOSED Between the end of 1999 and 2002, the funded status of the DB plans—measured as the fair value of pension assets divided by reported liabilities—for the average company in the S&P 500 Index fell from approximately 130% to about 80%. During that period, a surplus of over $300 billion, defined in market value terms, became a deficit of over $200 billion, eliminating some $500 billion of value on corporate balance sheets. The funded status of DB pensions has increased modestly since 2002, in part due to contributions into pension funds. But it remains well below 100% (and would be worse if some of the most under-funded companies had not been dropped from the S&P 500 universe). And, as discussed below, when one considers the large asset-liability mismatch at the core of most DB plans, a significant amount of the equity capital on corporate balance sheets is at risk while serving in effect to cushion pension equity exposures. The explanation for what happened in 2000-2002 is quite straightforward. Interest rates fell; the Moody’s AA corporate rate declined from 7.90% at the end of 1999 to 6.52% at the end of 2002. DB liabilities are the effective equivalent of long-duration bonds, There is a global move towards transparency in the expensing, reporting and funding of pensions. Recent U.S. funding regulations are an example. The primary purpose of the new funding rules is to tighten funding to take pressure off the PBGC while allowing strong companies to retain funding flexibility. An important effect of the new rules is that they have made contribution requirements much more transparent to the market and users of financial statements, who can now more readily estimate funding requirements. The current International Accounting Standards Board (IASB) and Financial Accounting Standards Board (FASB) projects are expected to move further in the marked-to-market THIS MATERIAL IS PREPARED FOR INSTITUTIONAL INVESTOR USE ONLY November 2006 Investment Management Journal • 5 direction, making impacts on balance sheets and income statements easier to estimate in all scenarios. This transparency is highly likely to promote more sophisticated pricing of pensions and lead to corporations managing their plans in the ways that add most value. In essence, pension investment policy will increasingly be based on corporate finance considerations and theory, rather than relying on the smoothed reported income and long-term portfolio analytics that have been the mainstay of the traditional paradigm. This is an important change and has implications that will be discussed in our next article. The change will lead to much closer matching of assets and liabilities, with investment risk being taken in a very different form to current practice. There are also several challenges that pension sponsors face in common with other investors. The most important of these are discussed below. A LOWER RETURN WORLD decreasing interest rate environment (the present value of future dividends being discounted at lower rates). At this point, it would be difficult for equity returns to be fueled by interest rate moves. If interest rates decrease significantly, it would indicate that there is either recession or deflation, and both of these environments are unfavorable for equities. RAPID GROWTH AND USE OF DERIVATIVES The last decade has seen an explosion in the development and liquidity of derivative instruments and structured products that use them. Swaps, futures and exchange-traded funds (ETFs) in particular have dramatically increased access and liquidity in markets that were previously illiquid and expensive to trade. The impact is significant in at least two direct ways: the separation of alpha and beta, and the reduction of risk premiums. Separation of alpha and beta. The ability to capture market benchmarks and various economic factors (e.g., inflation) through easily accessed and relatively cheap derivative instruments or structures has made the separation of alpha returns (excess returns not explained by market risk) and beta returns (returns for taking market risk) significantly easier. Theoretically, the price of adding beta to a portfolio is rapidly approaching zero in many liquid markets (e.g., U.S. large-cap stocks). This is rapidly changing approaches to portfolio construction. Alpha is difficult and costly to achieve—whether sought in more illiquid private markets such as real estate and private equity, or in the relatively liquid public markets. Beta can easily be adjusted using derivatives. Thus the more efficient approach to asset allocation is first to invest in the alpha engines and then to adjust the betas (sensitivities to equity markets) to the desired levels. Note that just because alpha is not explained by market risks, this does not mean that it carries no risk. However, alpha strategies’ risks are idiosyncratic (i.e., not systematic) and thus virtually uncorrelated with market risks or with each other. This allows diversification of risk and leads to higher riskadjusted returns. Any thinking on prospective returns usually starts with treasury rates because they can provide investors with a near “riskless” return over extended periods. In the current interest rate environment, treasury rates establish a low risk-free return base. Investors expect the returns of “risky assets”, such as equities, to be in excess of treasury rates. In general, the term “risk premium” describes the excess returns that investors expect to compensate them for taking the risk. Unfortunately, equity risk premiums are not easily measured. We may however be able to gain an insight from the credit market where credit spreads are at historical lows. It seems likely to us that the compression of credit spreads has been caused by increasing investor demand for return rather than a reduction in credit risk, and it is reasonable to expect that investor demand for return has also reduced equity risk premiums. Furthermore, between 1981 and 1999, equity prices grew much more rapidly than earnings (i.e., price/ equity (PE) ratios increased). The appreciation of PE ratios in this period can be attributed to the 6 • Investment Management Journal November 2006 THIS MATERIAL IS PREPARED FOR INSTITUTIONAL INVESTOR USE ONLY Risk premiums. Expected returns of risky asset classes can be broken into a rate of return on government bonds plus a risk premium. The risk premium may be further thought of as a premium for each kind of risk embedded in the investment. In many markets, a key part of the risk premium is associated with liquidity risk. If the liquidity of an investment class is improved by the growth of a derivative market, then the risk premium associated with that market may be permanently reduced as the liquidity premium shrinks. The effect is not easy to measure in real time and may only be known after many years. Consider corporate bonds, where the traditional spread over U.S. treasury bonds has been well in excess of the default losses. The gap is partly explained by a liquidity premium because of the historically higher costs of trading corporate bonds. The advent and rapid growth of the credit default swap market has allowed a separation of the credit default risk from the corporate bond and thus, indirectly, increased the liquidity of the corporate bond market. This could permanently reduce the spread between corporate and government bonds. Investors who rely heavily on corporate bonds for spread may need to rethink their portfolios. PROLIFERATION AND GROWTH OF ALTERNATIVE ASSET CLASSES The full impact of the emergence of these strategies is not yet known, but it seems reasonable to assume that markets will become even more efficiently priced. In particular, weakness and corporate inefficiency is likely to be more heavily punished than previously. This will put even more pressure on corporations to manage their pension plans as efficiently as possible, given pension transparency and the increasing sophistication of short sellers. Additionally, the growth of private equity and leveraged buyout funds’ massive war chests will apply additional pressure for management to efficiently manage their companies (including their pension plans). Hedge funds are also migrating into the private equity area in search of alpha. These war chests and the high costs of remaining public are likely to fuel an increase in private equity and a decrease in public equity. This may also erode public market returns. For instance, some of the returns in an equity index would have accrued as weak managements were replaced and companies turned around. Such turnarounds are now more likely to occur in the private equity markets as private equity firms take over poorly managed companies. CONCLUSIONS The search for alpha has led investors to move into more complex investment strategies and less conventional and more illiquid areas. This has complicated the task of the CIO. It also indirectly complicates the task of the CEO and CFO with respect to two trends. First, the rapid growth and development of longshort investment strategies has greatly increased the number of investment alternatives available to plan sponsors. While these strategies have been with us for a long time, and have usually been associated with hedge funds, a growing number of traditional managers have started offering 130/30 funds and other variations of long-short investing. The combination of the challenges discussed is leading to a fundamental change in how DB assets are invested. Some of these changes are already captured in ALM and LDI approaches, which attempt to match assets much more closely with liabilities. There is, however, more to this puzzle than matching because the degree of risk taking and the kinds of risks taken are also of great importance. We will discuss this in our next article. THIS MATERIAL IS PREPARED FOR INSTITUTIONAL INVESTOR USE ONLY November 2006 Investment Management Journal • 7 Federico Kaune, Ph.D. Executive Director Morgan Stanley Investment Management PG.07 Eric Baurmeister, CFA Executive Director Morgan Stanley Investment Management Emerging markets domestic debt: A new opportunity in fixed income A decade of sound macroeconomic management has brought a secular reduction in the risks of emerging market (EM) investments, as cautious financial policies and a benign international environment have driven a marked improvement in economic performance. Over the last few years, most EM countries have benefited from strong global growth and low developed market interest rates, as well as windfall revenues from high commodity prices and growing trade volumes. Many governments have utilized these benign international conditions to accelerate public sector de-leveraging. As a result, a majority of EM countries have experienced a sharp decline in their public debt levels, illustrated in Figure 1, and a sharp increase in their international reserve levels. 8 • Investment Management Journal November 2006 THIS MATERIAL IS PREPARED FOR INSTITUTIONAL INVESTOR USE ONLY FIG 1: TOTAL PUBLIC DEBT IN SELECTED COUNTRIES 110 100 Turkey 90 % of GDP 80 Brazil 70 Malaysia 60 50 2002 2003 2004 2005 2006 2007 Source: Central banks, Morgan Stanley Investment Management Data as of August 31, 2006 At the same time, most EM countries have mitigated chronic balance sheet currency mismatches by paying down external debt and replacing it with domestic debt. In order to develop their local markets, many have advanced second stage reforms such as removing taxes and other regulatory impediments, improving the transparency and liquidity of their financial markets and issuing a variety of bond types more suited to international investors. The improved macro performance and market reforms have lowered risk premiums in most emerging markets. This phenomenon has been most notable in hard currency denominated assets, while local currency fixed income valuations have generally lagged fundamental macroeconomic improvements. At the same time, local sources of demand for EM bonds have grown rapidly as pension funds, banks and individuals have repatriated funds to invest in their domestic markets. FIG 2: A HIGH- QUALITY ASSET CLASS Average ratings of countries in the JPMorgan Global Bond Index—Emerging Markets 100% 90 80 70 60 50 40 30 20 10 0 Dec-02 Apr-03 Aug-03 Dec-03 Apr-04 Aug-04 Dec-04 Apr-05 Aug-05 Dec-05 Apr-06 Aug-06 Investment Grade B Rated BB Rated Source: S&P, Morgan Stanley Investment Management Data as of August 31, 2006 THIS MATERIAL IS PREPARED FOR INSTITUTIONAL INVESTOR USE ONLY November 2006 Investment Management Journal • 9 FIG 3: DOMESTIC DEBT AS A SHARE OF TOTAL PUBLIC DEBT IN 2002 AND 2006 90 85 80 % of total debt 75 70 65 60 55 50 2002 Turkey Malaysia Brazil 2006 Source: Central banks, Morgan Stanley Investment Management Data as of August 31, 2006 INCREASING SIZE AND IMPROVING CREDIT FUNDAMENTALS Voluntary local currency financing is becoming increasingly available to EM countries, reflecting the vast recent improvement in macroeconomic fundamentals and in the institutional framework supporting domestic financial markets. Lower and more predictable inflation rates, improved fiscal discipline, declining public debt ratios and the adoption of more flexible exchange rates have helped EM countries to develop domestic fixed income markets. Until recently, these markets were either non-existent or closed to foreign investors. But removal of capital controls and other administrative impediments, as well as a more transparent regulatory environment, have helped to open domestic bond and FX markets to foreign capital. Additionally, the last decade’s micro-level reforms enacted in the pension, social security and financial sectors have created natural domestic demand. Governments have been able to rely, increasingly, on private domestic financing, which in turn has provided a self-reinforcing source of financial market stability. As EM countries lessen their reliance on international capital markets and hard currency denominated debt, rating agencies have rewarded them with upgrades. Those who do not follow these markets closely may be surprised to discover that the average rating of a broad local-markets index is Baa1/BBB+ (Figure 2). EVOLUTION OF DOMESTIC MARKETS In the early stages of public sector financial reform, EM governments may only be able to issue shortterm local currency bonds linked to the nominal exchange rate, the inflation rate or an interest rate benchmark. These indexing mechanisms may allow governments to tap into a new local source of financing, but do not diminish debt market vulnerability to external and financial shocks. 10 • Investment Management Journal November 2006 THIS MATERIAL IS PREPARED FOR INSTITUTIONAL INVESTOR USE ONLY FIG 4: LOCAL TREASURY YIELD CURVES IN SELECTED COUNTRIES 10 South Africa (left-hand scale) 9 Yield to maturity (%) Yield to maturity (%) 5.5 8 Thailand (right-hand scale) Mexico (left-hand scale) 5.0 7 6 1 Month 3 Months 6 Months 1 Year 3 Years 5 Years 7 Years 10 Years 4.5 20 Years Source: Central banks, Morgan Stanley Investment Management Data as of August 31, 2006 As the private sector gains more confidence in government commitment to a set of sustainable policies, the quality of local financing gradually improves. More mature local markets allow for longer-term fixed-coupon bonds, without the indexation features described above. As the share of non-indexed local currency denominated bonds increases, the vulnerability of public debt to external and financial shocks declines, strengthening fundamentals and further increasing government access to domestic financing. Several local emerging markets have experienced these positive dynamics in the past five years (Figure 3). A GOOD VALUE PROPOSITION? The opportunity set for relatively open, transparent and liquid domestic fixed income markets has increased dramatically over the last decade. Instruments include (fixed and zero coupon) nominal bonds and (inflation, exchange rate or interest rate) indexed bonds. Many countries have lengthened the average maturity of their domestic bond issuance and have issued bonds with 20- to 30-year maturities (Figure 4). In addition, many EM countries have made settlement of trades in their domestic bonds possible through international clearing corporations. While the improvement in macroeconomic fundamentals has been swiftly reflected for hard currency denominated bonds in the form of lower country risk premiums, the same has not been true for local bond markets. Due to the more limited investor base and other technical factors, local bond markets have been slower to react and therefore may represent a good value proposition. THIS MATERIAL IS PREPARED FOR INSTITUTIONAL INVESTOR USE ONLY November 2006 Investment Management Journal • 11 A NEW ASSET CLASS FIG 5: JPM GBI- EM DIV BROAD INDEX COMPOSITION 10.0% Brazil 3.4% Colombia 10.0% Mexico 0.7% Chile 10.0% S. Africa 0.8% Russia 5.2% Turkey 1.5% Slovakia China 10.0% India 10.0% Indonesia 2.1% Malaysia 8.7% Thailand 5.9% Czech Republic 5.2% Hungary 6.5% Poland 10.0% Source: JPMorgan Chase, Morgan Stanley Investment Management Data as of August 31, 2006 FIG 6: CHARACTERISTICS OF EMERGING MARKETS DEBT INDICES Based on month-end data from December 2002-August 2006 JPM GBI-EM Div Broad (Local currency bonds shown in US$) Characteristics Yield to maturity (%) Interest rate duration (years) Spread duration (years) Sovereign spread (bps) Market capitalization (US$ bn) Total return (%) 2003 2004 2005 One year Three years Year to date 15.70 16.87 6.32 5.77 11.26 3.86 25.68 11.74 10.50 9.00 12.09 5.21 7.66 4.11 — — 589 6.76 6.99 7.03 197 290 JPM EMBIG (Hard currency bonds shown in US$) The creation in June last year of the JPMorgan Global Bond Index—Emerging Markets (JPMorgan GBI-EM), a natural extension of the existing JPMorgan Global Bond Index (JPMorgan GBI) for developed economies, is likely to facilitate strategic benchmarked allocations to this new and expanding investment universe. The JPMorgan GBI-EM closely follows the methodology of the family of JPMorgan GBI indices, which are widely used by investors in developed markets. This index includes domestic currency bonds issued by low- or middle-income countries with markets that are easy to access and where no major impediments exist for foreign investors. Figure 5 shows the index’s composition by country, while Figure 6 demonstrates some average return characteristics. Information in this article about the returns and characteristics of local currency bonds are based on the JPMorgan GBI-EM. However, the universe of investable local EM bond markets is much broader and includes upwards of 30 countries. As their domestic markets evolve, we expect these countries to be included in the index over time. BENEFITS OF LOCAL EMERGING MARKETS While local markets are still inefficient and expanding, the opportunity set continues to grow and diversify. It is in both issuers and investors’ interests that these nascent markets expand at a prudent pace. Issuers rightly prefer gradual expansions of their domestic bond markets, as they do not want to subject themselves to rapid or frequent capital inflows/outflows. Investors concur because they do not want supply of local debt to outstrip demand. Although rapid expansion is unattractive, EM debt denominated in local currency offers significant attractions to both issuers and investors, and both groups benefit from their continued expansion. December 2002-August 2006—annualized figures (%) Total return Volatility Sharpe ratio 11.84 7.39 1.24 14.66 7.29 1.63 Source: JPMorgan Chase, Morgan Stanley Investment Management Past performance is no guarantee of future results 12 • Investment Management Journal November 2006 THIS MATERIAL IS PREPARED FOR INSTITUTIONAL INVESTOR USE ONLY FIG 7: HISTORICAL MULTI- YEAR PERFORMANCE RECORD FOR VARIOUS ASSET CLASSES Total returns shown in US$ (%) 14 12 10 8 6 4 2 0 1 Year JPMorgan GBI-EM JPMorgan EMBI Global ML High Yield 3 Years JPMorgan GGBxUS S&P 500 Source: JPMorgan Chase, Morgan Stanley Investment Management Past performance is no guarantee of future results Data as of August 31, 2006 For issuers • Borrowing in their own currency shifts currency risk away from the country towards investors in the developed world, who may be better able to manage it. • A local currency debt market allows a government to diversify its funding across a yield curve. This shift has been facilitated by the great success the EM countries have had over the past decade in lowering inflation. As a result, EM countries are able to issue long bonds in their own currencies. • A government yield curve opens the door for private market issuance and the resulting deepening of local capital markets and financial institutions. This improves financial stability and may make crises less likely. For investors • As risk premiums in local markets must compensate for both credit and currency risks, the potential for currency gains adds to the attraction of higher yields relative to hard currency denominated bonds. • Local EM markets may well deliver greater diversification benefits to investors as they may be less prone to contagion, and returns are more subject to local macroeconomic and policy outcomes. THIS MATERIAL IS PREPARED FOR INSTITUTIONAL INVESTOR USE ONLY November 2006 Investment Management Journal • 13 FIG 8: ROLLING THREE- MONTH VOLATILITY Based on month-end data from December 2002-August 2006 (%) 35 30 25 20 15 10 5 0 Feb-00 Aug-00 JPM GBI-EM Feb-01 Aug-01 Aug-02 Aug-02 Feb-03 Aug-03 Feb-04 Aug-04 Feb-05 Aug-05 Feb-06 S&P 500 Aug-06 JPM GBI JPM 10Yr Treasury JPM EMBI Global Source: JPMorgan Chase, Morgan Stanley Investment Management HIGH RETURNS Undervalued assets in conjunction with improving credit fundamentals have supported the strong total returns for all EM assets over the last few years (Figure 7). However, returns for domestic currency debt lagged the strong returns in EM external debt and equity markets in 2005. As the investor base for EM local currency debt continues to expand, we would expect this pattern to reverse. VOLATILITY SIMILAR TO OTHER FIXED INCOME ASSET CLASSES markets, deregulated the public sector and removed impediments to savings and investment. In most EM countries, flexible exchange rate policies have been in place for a relatively short period of time and, as such, historical analysis of EM local bond market returns and volatility is only meaningful if limited to the last few years (Figure 8). While it is true that this period of unprecedented global liquidity has dampened financial market volatility, we expect the extensive improvements in both economic policies and their transparency to limit volatility even during less favorable legs of the economic cycle. However, we must caution that there may be temporary spikes in volatility during future global events. Emerging countries are well into the process of transitioning to open, market-oriented economies. However, only recently, and in some cases only partially, have they liberalized trade and financial 14 • Investment Management Journal November 2006 THIS MATERIAL IS PREPARED FOR INSTITUTIONAL INVESTOR USE ONLY FIG 9: CORRELATIONS OF EMERGING MARKETS LOCAL DEBT VS. OTHER ASSET CLASSES Based on month-end data from December 2002-August 2006 JPM GBIEM Div Broad JPM GBI-EM Div Broad JPM EMBI Global ML High Yield Lehman Gov't JPM GGBxUS S&P 500 NASDAQ MSCI EMF Lehman Aggregate ML BBB Corp JPM EMBI Global ML Lehman High Gov't Yield JPM GGB xUS S&P NASDAQ 500 MSCI Lehman EMF Aggregate ML BBB Corp 1.00 0.68 0.55 0.38 0.46 0.49 0.34 0.65 0.43 0.53 — 1.00 0.65 0.76 0.60 0.30 0.18 0.40 0.81 0.85 — — 1.00 0.31 0.39 0.47 0.36 0.40 0.40 0.57 — — — 1.00 0.71 -0.06 -0.15 -0.02 0.99 0.93 — — — — 1.00 0.09 0.03 0.11 0.71 0.76 — — — — — 1.00 0.86 0.74 -0.01 0.13 — — — — — — 1.00 0.62 -0.12 0.02 — — — — — — — 1.00 0.04 0.14 — — — — — — — — 1.00 0.95 — — — — — — — — — 1.00 Source: JPMorgan Chase, Morgan Stanley Investment Management GOOD DIVERSIFICATION EM local currency investments provide good diversification benefits, as multi-year correlations between the JPMorgan GBI-EM and other markets are relatively low. Notably, correlations with equities and higher-rated fixed income asset classes are particularly low, as shown in Figure 9 above. CHALLENGES AND OPPORTUNITIES AHEAD We believe that the outlook for EM economies remains bright. While the risk premiums in external debt markets reflect this outlook, we believe that domestic currency denominated markets offer an excellent opportunity to purchase undervalued assets and to diversify risk relative to other fixed income asset classes. Many EM countries are now in the midst of multi-decade reform processes, with the dual goals of raising growth rates and delivering macroeconomic stability through more market-oriented policies. Relatively strong economic performance over the last few years across the emerging markets has resulted in increased reserve levels, reduced debt levels and widespread credit rating upgrades. This success has allowed many EM countries to shift their borrowing away from external to domestic markets at a prudent pace. We believe that local markets may provide the best opportunity for fixed income investors looking to benefit from the ongoing structural changes in these countries. THIS MATERIAL IS PREPARED FOR INSTITUTIONAL INVESTOR USE ONLY November 2006 Investment Management Journal • 15 Martin L. Leibowitz, Ph.D. Managing Director Morgan Stanley Equity Research PG.15 Anthony Bova, CFA Associate Morgan Stanley Equity Research Alpha returns and active extensions The relaxation of the long-only constraint within equity portfolios and the subsequent move into an active extension “120/20 strategy” can lead to material improvements in both alpha returns and alpha/tracking error volatility (TEV) ratios. These potential benefits can be estimated by combining an alpha ranking system, a position weighting function and a tracking error model. There is empirical evidence that the structure of a wide range of active portfolios can be approximated by exponentially declining alpha rankings and position weightings. These alpha/weighting models can be used to explore how active extensions (and active portfolios in general) can generate alpha returns subject to prescribed risk limits. Additional benefits from active extension portfolios include the ability to offset unproductive correlations and to facilitate specific pair trades between long and short positions. Such offsets can sharpen the intended risk exposures and lead to higher alpha/TEV ratios. Moving into a risk-controlled active extension will generally not have a significant impact on fund level volatility. Since most U.S. institutions’ portfolios are overwhelmingly beta dominated, any incremental TEV will be submerged by this beta effect. Active extensions that provide positive alphas can therefore significantly increase the fund’s total return with only a minimal impact on the overall volatility. This article was first published by Morgan Stanley Research, August 31, 2006. 16 • Investment Management Journal November 2006 THIS MATERIAL IS PREPARED FOR INSTITUTIONAL INVESTOR USE ONLY The preconditions for realizing any of these benefits are a credible basis for producing positive alphas in both long and short portfolios, a high level of risk discipline, an ability to minimize and/or offset unproductive correlations, and an organizational ability to pursue short extensions in a benchmarkcentric, cost-efficient fashion. BENEFITS OF ACTIVE EXTENSIONS The removal of the long-only constraint enables active extension portfolios to provide material improvements in both the cumulative alpha and in the alpha/TEV ratio.21-23 Additional benefits of the active extension portfolio include the ability to offset unproductive correlations and to facilitate specific pair trades between long and short positions.24-26 The interest in 120/20-type active extension strategies has grown significantly as both investment sponsors and asset managers have sought higher levels of positive alpha. The acceptance of these strategies has also been enhanced because, with appropriate risk control, they can be viewed as an “extension” of traditional equity management rather than as a quantum leap into the more constrained space allocated to alternative assets. For a base case example with correlation effects, we use a long-only portfolio that provides an alpha of 1.85% and an alpha/TEV ratio of 0.54. Without any offsets, a 20% extension increases the alpha to 3.08% and the ratio to 0.63. However, the TEV-reducing effect of offsets between the longs and shorts can raise this alpha/TEV ratio into the 0.75-0.90 range. The preconditions for realizing any of these benefits are a credible basis for producing positive alphas in both long and short portfolios, a high level of risk discipline, an ability to minimize and/or offset unproductive correlations, and an organizational ability to pursue short extensions in a benchmarkcentric, cost-efficient fashion. ALPHA RANKING AND PORTFOLIO WEIGHTING MODELS In a benchmark-centric long-only portfolio, the ability to take significant underweight positions is limited to those few stocks with very large market capitalizations in the benchmark. By allowing a limited facility to short stocks within a riskcontrolled framework, active extension strategies open the door to a fresh set of underweight positions in lesser-cap stocks. These active extension strategies are also referred to as 120/20 portfolios. This term came from the early implementations that allowed up to 20% of the portfolio to be shorted with the proceeds from the short sales used to purchase 20% additional longs. Hence, the portfolios maintained their 100% net long exposure with gross footings of 120% long and 20% short. More recent active extension launches have been in the 130/30-140/40 range as managers have gained more comfort with this product. A growing body of studies has addressed the potential performance benefits that can be obtained by loosening the standard long-only constraint.1-20 Active extensions are based upon relaxation of this long-only constraint but have special features that maintain the basic risk characteristics of benchmarkcentric long-only funds: • The percentage sold short is offset by reinvestment in beta-equivalent new longs so as to preserve the original beta target. • The overweight and underweight positions are structured so as to keep the tracking error within reasonable bounds. The standard measure for the value added from active management is the alpha/TEV ratio. The first step in analyzing this ratio is to create an alpha ranking model.27 Portfolio managers generally have some formal or informal process for classifying their prospective active positions into a descending THIS MATERIAL IS PREPARED FOR INSTITUTIONAL INVESTOR USE ONLY November 2006 Investment Management Journal • 17 FIG 1: ALPHA RANKING MODEL 6 5 Original long-only fund 4 Alpha 3 2 1 0 0 5 10 15 20 25 30 35 Number of positions Source: Morgan Stanley Research sequence based upon their level of conviction. Alpha ranking models can be used to approximate such classifications. These alpha ranking models may take a variety of forms depending on the style of the investment fund and/or perceived opportunities in the market. Our base case ranking model is based on an exponential alpha decay with a beginning alpha of 5% that declines to 2.24% at the 25th position (Figure 1). In theory, with a constant residual volatility, the optimal weighting for each position should be proportional to its alpha ranking. The actual sequential weights seen in practice provide empirical evidence that portfolios are at least roughly structured along these lines. Figure 2 displays the sequence of position weights for a sample of long-only funds. The top line represents the gross weights while the bottom line shows the active weights, i.e., the difference between the gross weight and benchmark weight. FIG 2: GROSS AND ACTIVE WEIGHTING FUNCTIONS FOR LONG- ONLY PORTFOLIOS 4.5 4.0 3.5 3.0 Exponential weighting model 2.5 2.0 Gross weights 1.5 1.0 0.5 0.0 0 5 10 Number of positions Source: Morgan Stanley Research Weight (%) Active weights 15 20 25 18 • Investment Management Journal November 2006 THIS MATERIAL IS PREPARED FOR INSTITUTIONAL INVESTOR USE ONLY FIG 3: LONG- ONLY TEV: CORRELATED VS. UNCORRELATED 5.0 4.5 4.0 3.5 3.0 ρ L = 0.05 TEV 2.5 2.0 1.5 1.0 0.5 0.0 0 5 10 15 20 25 30 ρL = 0 35 Number of positions Source: Morgan Stanley Research The middle dotted line is an exponential weighting function that begins at a 3% weight. This theoretical weighting function has approximately the same decay rate as the alpha ranking model in Figure 1. CORRELATION EFFECTS ALPHA/TEV RATIOS The three factors that affect the TEV are the residual volatilities of each position, the portfolio weightings, and the correlations that exist between the positions. Using the exponential weighting model from Figure 2, our baseline long-only portfolio was constructed to have 25 “pro-active” positions with a net “activity level” of 50%. The remaining “non-pro-active” component of the portfolio serves as a source of funds as well as helping to maintain the fund’s target beta. After the 25th position, the active weight remains constant for any additional long positions added to the portfolio. Figure 3 compares the TEVs for the long-only portfolio under assumed pairwise correlations ( L) of zero and 0.05. In calculating the TEVs, we assume throughout a constant residual volatility of 23% for all active positions. For the 25-position long portfolio, the TEV increases from 2.38% for the uncorrelated case to 3.46% for a 0.05 correlation. It only takes a slight increase in pairwise correlation to generate significant increases in the TEV. The alpha ranking model can be combined with the exponential weighting function to generate a cumulative portfolio alpha. Figure 4 displays this portfolio alpha as a function of the TEVs from Figure 3. In the uncorrelated case, the alpha continues to rise as more positions are added to the portfolio while the TEV stays within a 2-2.5% range. This suggests that under the assumption of a zero correlation, more positions should continue to be added to the portfolio since the cumulative alpha increases faster than the portfolio’s TEV. This is also evidence that in order to get TEVs greater than 3% (which is what is observed in actively managed longonly equity portfolios), there must exist some degree of positive correlations among the long positions.28 With a 0.05 correlation, the alpha and TEV increase at nearly the same rate as more positions are added to the portfolio, leading to a roughly constant alpha/ TEV ratio. In this situation, the investor may sacrifice a higher TEV (if it can be tolerated) for the higher alpha portfolio even without any corresponding improvement in the alpha/TEV ratio itself. THIS MATERIAL IS PREPARED FOR INSTITUTIONAL INVESTOR USE ONLY November 2006 Investment Management Journal • 19 FIG 4: LONG- ONLY ALPHA VS. TEV: CORRELATED AND UNCORRELATED 3.0 2.5 N = 50 Portfolio alpha 2.0 N = 25 1.5 ρ L =0 1.0 ρ L =0.05 0.5 0.0 0 1 2 TEV (%) Source: Morgan Stanley Research 3 4 5 THE SHORT EXTENSION The ability to take short positions provides access to a fresh set of underweights. In the following analysis, these new underweights are assumed to have alphas that coincide with the corresponding long-only alpha model, less some given shorting cost. Figure 5 schematically depicts a 20% short extension. The short portfolio is based on an alpha ranking model that follows the original long-only 5% declining alpha ranking model but with a 0.5% reduction to account for shorting costs. This model also assumes that the short portfolio follows the same exponential weighting model as the long portfolio. The proceeds generated by the shorts are then reinvested into the new long positions. The proceeds from the shorts could theoretically be reinvested to increase the weight invested in the highest-alpha long positions. However, most portfolios will have already established their maximum allowable weight in these high-ranked FIG 5: ALPHA RANKINGS FOR SHORT POSITIONS 6 5 1. Original longs 4 Alpha 3 2. 20% Short extension 3. 20% Reinvested longs 2 1 0 0 10 20 Number of positions Source: Morgan Stanley Research 30 40 50 20 • Investment Management Journal November 2006 THIS MATERIAL IS PREPARED FOR INSTITUTIONAL INVESTOR USE ONLY FIG 6: ACTIVE EXTENSION IMPACT ON PORTFOLIO ALPHA 3.5 New longs 3.0 Shorts 2.5 Portfolio alpha 2.0 Long only 1.5 1.0 0.5 0.0 0 10 20 Number of positions Source: Morgan Stanley Research 30 40 50 positions, so we take the more conservative approach and reinvest the proceeds starting with the 26th ranked long position. Note that because of the front-end loading from the exponential weightings, the 20% extension is achieved with only eight new short positions. The 20% funds are then reinvested into new longs from the 26th to 42nd position, where the position weight was assumed to be fixed at the 1.2% minimum. Figure 6 displays cumulative portfolio alpha as a function of the total number of positions for the long-only and a 120/20 portfolio. The biggest boost in alpha comes from the eight new short positions that come from the early part of the alpha ranking model. The 17 new longs consist of the tail end of the alpha ranking model and weighting function. Since these are lower-grade alpha sources, these new longs do not provide as significant a benefit as the new shorts. FIG 7: WEIGHTING FUNCTIONS FOR ACTIVE EXTENSION PORTFOLIOS 4.0 3.5 3.0 Exponential weighting model Weight (%) 2.5 2.0 1.5 Active longs 1.0 0.5 Active shorts 0.0 0 10 20 Number of positions Source: Morgan Stanley Research 30 40 50 THIS MATERIAL IS PREPARED FOR INSTITUTIONAL INVESTOR USE ONLY November 2006 Investment Management Journal • 21 FIG 8: LONG AND SHORT EXPOSURES FOR ACTIVE EXTENSION AND MARKET NEUTRAL FUNDS 160 140 120 Long weight (%) 100 80 60 0 % net long 40 20 0 0 20 40 Short weight (%) Source: Morgan Stanley Research 100 % net long Active extension Market neutral 60 80 100 FIG 9: LONG AND SHORT BETAS FOR ACTIVE EXTENSION AND MARKET NEUTRAL FUNDS 1.6 1.4 Active extension 1.2 Long beta 1.0 0.8 0.6 0.4 Net beta = 0 0.2 0.0 0.0 0.2 0.4 0.6 Short beta Source: Morgan Stanley Research Net beta = 1 Market neutral 0.8 1.0 1.2 1.4 As an empirical test for how these exponential models apply to actual portfolios, Figure 7 displays the weighting functions for an admittedly small sample of active extension funds that have reported their holdings to the the U.S. Securities and Exchange Commission (SEC). As with the long-only funds, both the longs and shorts in the active extension funds follow a pattern that can be approximated as an exponential decay. The exposures and beta values for this sample of active extension funds along with a sample of SECreporting market neutral funds are shown in Figure 8 and Figure 9. It can be seen that the active extension funds are quite closely aligned to the target 100% net exposure and have betas that remain close to 1. Not surprisingly, the market neutral funds all have net long exposures and net betas close to zero. 22 • Investment Management Journal November 2006 THIS MATERIAL IS PREPARED FOR INSTITUTIONAL INVESTOR USE ONLY FIG 10: CORRELATION EFFECTS: TEV VS. SHORT WEIGHT 10 ρ L = ρS = 0.05 8 ρ L, S = 0 ρ L, S = -0.03 6 TEV ρ L, S = -0.05 ρ L = ρS = ρ L, S = 0 4 2 0 0 20 40 Short weight (%) Source: Morgan Stanley Research 60 80 100 OFFSETTING LONG- SHORT CORRELATIONS The basic long-only correlation model applies when all positions have a common pairwise correlation. Just as a positive correlation can have a material TEV-increasing effect, so the opportunity for offsetting negative correlations can act as a major TEV-reducing factor. In theory, such offsets could be present within the long portfolio itself. However, for the sake of simplicity, only positive correlations are assumed to exist within the longs and within the shorts, while the offsetting negative correlations are assumed to occur only between the shorts and longs. Figure 10 shows the impact of various correlation effects. The top three lines all assume a positive 0.05 correlation within the longs and within the shorts but differ in the short-to-long correlations. The TEV-reducing effect of these offsets is clearly evident as the -0.05 offset curve moves towards the uncorrelated case. FIG 11: ALPHA/TEV VS. SHORT WEIGHT 1.2 ρ L = ρS = ρ L, S = 0 1.0 0.8 ρ L, S = -0.05 ρ L, S = -0.03 Alpha/TEV ρ L, S = 0 0.6 ρ L = ρS = ρ L, S = 0.05 0.4 0.2 0.0 0 20 40 Short weight (%) Source: Morgan Stanley Research 60 80 100 THIS MATERIAL IS PREPARED FOR INSTITUTIONAL INVESTOR USE ONLY November 2006 Investment Management Journal • 23 FIG 12: ALPHA VS. TEV FOR VARIOUS OFFSET CORRELATIONS 5.0 Short weight Uncorrelated case ρ L, S = -0.03 ρ L, S = -0.05 ρ L, S = 0 Correlated cases 4.0 40% 20% 3.0 Alpha 2.0 0% Long-only portfolios 1.0 0.0 0 1 2 3 4 TEV Source: Morgan Stanley Research 5 6 7 8 The alpha ranking models from Figure 5 can be combined with the TEVs in Figure 10 to calculate the alpha/TEV ratios at various short weights. With positive correlation of 0.05, the information ratio for the basic 25-position long-only portfolio ratio was 0.54. With varying short extensions added from that point, the long-short correlation can be seen to play a key role. With a zero offset correlation, the short extension provides only a modest increase in the information ratio to 0.63. However, with an offsetting -0.05 correlation (Figure 11), the short extensions can raise the information ratio to 0.92 for short weights in the 40-60% range. With the more moderate offset of -0.03, the ratio reaches a peak value of around 0.74 for short weights of 30-50%. For the uncorrelated situation, the short extension improves the alpha/TEV ratio from 0.78 for the long-only portfolio to a peak value of 1.10 for short weights in the 60-80% range. Clearly, any active extension strategy is critically dependent on an efficient facility for selecting, implementing, and maintaining the short portfolio. If shorting costs become too high, the resulting alpha degradation would eliminate any benefits from the short extension. The alpha/TEV ratio is an important metric but it may not always serve as a sufficient gauge of portfolio value. It also makes sense to look separately at the two components of this ratio. Figure 12 presents the results from active extension in alpha versus TEV space. If a fund had a maximum TEV it was willing to tolerate, the extension could add a significant alpha increment to the return from the long-only portfolio (even with a zero offset between internally correlated longs and shorts). FUND LEVEL EFFECTS For typical asset allocations, it is well known that the TEV from a moderate-sized component portfolio is likely to have only a minimal impact on the overall fund volatility.29 Therefore, many institutional portfolios may care more about the portfolio alpha than the alpha/TEV ratio. Figure 13 provides an example of how active extensions can affect performance characteristics at the overall fund level. The first column represents the passive benchmark portfolio with a 60% equity/ 40% bond allocation. The second column replaces the 60% passive equity benchmark with 60% active long-only equity. 24 • Investment Management Journal November 2006 THIS MATERIAL IS PREPARED FOR INSTITUTIONAL INVESTOR USE ONLY FIG 13: ALPHA ENHANCEMENT WITH MINIMAL VOLATILITY INCREASE Benchmark portfolio U.S. equity passive benchmark U.S. equity active—long only Active extension—40% short weight U.S. bonds Total Total beta Expected return Alpha TEV Total volatility Source: Morgan Stanley Research Active long only — 60% — 40% 100% 0.65 6.96 1.11 2.08 11.36 Correlated AE case without offset — 40% 20% 40% 100% 0.65 7.36 1.51 2.62 11.48 Correlated AE case with offset — 40% 20% 40% 100% 0.65 7.36 1.51 2.26 11.40 60% — — 40% 100% 0.65 5.85 — — 11.17 The move of the 60% equity allocation from a passive index into long-only active management increases the total volatility modestly from 11.17% to 11.36%. However, the total portfolio return increases by a significant 1.11%, i.e., 60% of the active alpha of 1.85%. The next two columns show the effect of moving 20% of the active equity into a correlated active extension with a short weight of 40%. (To be conservative we have assumed that the TEVs of the active extension and long-only active equity are fully correlated.) Without any long-short offset, the portfolio return increases by 0.40% to 7.36%, with the portfolio volatility only moving from 11.36% to 11.48%. With offsets, the return remains at 7.36% while the volatility declines to 11.40%. The reason that there is not a significant increase in the portfolio volatility in Figure 10 is because this portfolio (as with most U.S. institutions’ portfolios) is beta dominated and any additional tracking error volatility is submerged by this beta effect.30-32 Moving to an active management posture or to an active extension will generally not have a significant impact on the overall volatility of the fund—one beta-dominated asset is just being replaced with another. The only question then becomes whether these active management processes can reliably generate positive levels of expected alpha. CONCLUSIONS Active extension can be viewed as an “extended” form of traditional active equity management that has the potential to materially improve both portfolio alphas and alpha/TEV ratios by: • Creating access to a fresh crop of active underweight opportunities. • Reinvesting the short proceeds in productive new longs (even if they are of lower alpha rank). • Providing offsets that reduce unproductive correlations and facilitate return-enhancing pairing opportunities. The preconditions for realizing any of these benefits are a credible basis for producing positive alphas in both long and short portfolios, a high level of risk discipline, an ability to minimize and/or offset unproductive correlations, and an organizational ability to pursue short extensions in a benchmarkcentric, cost-efficient fashion. THIS MATERIAL IS PREPARED FOR INSTITUTIONAL INVESTOR USE ONLY November 2006 Investment Management Journal • 25 1 Robert D. Arnott and David J. Leinweber, “Long-Short Strategies Reassessed”, Financial Analysts Journal, September/October 1994. Peter L. Bernstein, “The Points of Inflection Revisited”, Economics and Portfolio Strategy, January 2006. John S. Brush, “Comparisons and Combinations of Long and Long/Short Strategies”, Financial Analysts Journal, May/June 1997. Roger Clarke, Harinda de Silva and Steven Thorley, “Portfolio Constraints and the Fundamental Law of Active Management”, Financial Analysts Journal, September/ October 2002. Roger Clarke, Harinda de Silva and Steven Sapra, “Toward More InformationEfficient Portfolios”, Journal of Portfolio Management, Fall 2004. Roger Clarke, Harinda de Silva and Steven Thorley, “Performance Attribution and the Fundamental Law”, Financial Analysts Journal, September/October 2005. Roger Clarke, “Portfolio Constraints and the Fundamental Law of Active Management”, Society of Quantitative Analysts Half-Day Fall Seminar: Advances in Optimization and Portfolio Construction, November 2005. Richard C. Grinold, “The Fundamental Law of Active Management”, Journal of Portfolio Management, Spring 1989. Richard C. Grinold and K. Eaton, “Attribution of Performance and Holdings”, Worldwide Asset and Liability Modeling, 1998. Richard C. Grinold and R. Kahn, “The Surprising Large Impact of the Long-Only Constraint”, Barclays Global Investors Investment Insights, May 2000. Richard C. Grinold and R. Kahn, “The Efficiency Gains of Long-Short Investing”, Financial Analysts Journal, November/December 2000. Richard C. Grinold and R. Kahn, “Active Portfolio Management: Quantitative Theory and Applications”, 2000. Richard C. Grinold, “Implementation Efficiency”, Financial Analysts Journal, September/October 2005. Bruce I. Jacobs and Kenneth N. Levy, “More on Long-Short Strategies”, Financial Analysts Journal, March/April 1995. Bruce I. Jacobs and Kenneth N. Levy, “20 Myths About Long-Short”, Financial Analysts Journal, September/October 1996. Bruce I. Jacobs and Kenneth N. Levy, “Enhanced Active Equity Strategies”, Journal of Portfolio Management, Spring 2006. Bruce I. Jacobs, “The Long and Short on Long-Short”, Journal of Investing, Spring 1997. Bob Litterman, “Are Constraints Eating Your Alpha?”, Pensions & Investments, March 2005. Harry M. Markowitz, “Market Efficiency: A Theoretical Distinction and So What?”, Financial Analysts Journal, 2005. Richard O. Michaud, “Are Long-Short Equity Strategies Superior?”, Financial Analysts Journal, November/December 1993. Martin L. Leibowitz and Anthony Bova, “Short Extension Portfolios; An Exploration of 120/20 Concept”, Portfolio Analysis Note, January 18, 2006. Martin L. Leibowitz and Anthony Bova, “An Integrated Analysis of 120/20 Short Extension Strategies”, Portfolio Analysis Note, July 27, 2006. Simon Emrich, “Alpha-Beta Separation and Short Extension Portfolios”, Quantitative and Derivatives Strategy, Morgan Stanley, June 2006. Martin L. Leibowitz and Anthony Bova, “Correlation Effects in Short Extension “120/20 Portfolios”, Portfolio Analysis Note, January 15, 2006. Kenneth Winston, “Buy-Side Risk Management”, Morgan Stanley, March 2006. Kenneth Winston and Thomas Hewett, “Long-Short Portfolio Behavior with Barriers Part 1: Mechanism”, Morgan Stanley, July 2006. Martin L. Leibowitz and Anthony Bova, “Alpha Ranking Models and Short Extension 120/20 Strategies”, Portfolio Analysis Note, May 16, 2006. Martin L. Leibowitz and Anthony Bova, “The Tracking Error Gap”, Portfolio Analysis Note, May 18, 2006. Martin L. Leibowitz, “The -Plus Measure in Asset Allocation”, Journal of Portfolio Management, Spring 2004. Martin L. Leibowitz and Anthony Bova, “Beta-Based Asset Allocation: A Summary”, Portfolio Analysis Note, November 30, 2005. Martin L. Leibowitz and Anthony Bova, “Allocation Betas”, Financial Analysts Journal, July/August 2005. Martin L. Leibowitz and Anthony Bova, “Beta-Based Asset Allocation: A Summary”, Portfolio Analysis Note, November 30, 2005. 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 26 • Investment Management Journal November 2006 THIS MATERIAL IS PREPARED FOR INSTITUTIONAL INVESTOR USE ONLY Important disclosures The securities/instruments discussed in this article may not be suitable for all investors. This article has been prepared and issued by Morgan Stanley primarily for distribution to market professionals and institutional investor clients. Recipients who are not market professionals or institutional investor clients of Morgan Stanley should seek independent financial advice prior to making any investment decision based on this article or for any necessary explanation of its contents. This article does not provide individually tailored investment advice. It has been prepared without regard to the individual financial circumstances and objectives of persons who receive it. Morgan Stanley recommends that investors independently evaluate particular investments and strategies, and encourages investors to seek the advice of a financial advisor. The appropriateness of a particular investment or strategy will depend on an investor’s individual circumstances and objectives. You should consider this article as only a single factor in making an investment decision. This article is not an offer to buy or sell any security/ instrument or to participate in any trading strategy. In addition to any holdings disclosed in the section entitled “Important disclosures on subject companies”, Morgan Stanley and/or its employees not involved in the preparation of this article may have investments in securities/instruments or derivatives of securities/ instruments of companies mentioned in this article, and may trade them in ways different from those discussed in this article. Derivatives may be issued by Morgan Stanley or associated persons. Morgan Stanley makes every effort to use reliable, comprehensive information, but we make no representation that it is accurate or complete. We have no obligation to tell you when opinions or information in this article change. With the exception of information regarding Morgan Stanley, reports prepared by Morgan Stanley research personnel are based on public information. Facts and views presented in this article have not been reviewed by, and may not reflect information known to, professionals in other Morgan Stanley business areas, including investment banking personnel. Important disclosures on subject companies The information and opinions in this article were prepared by Morgan Stanley & Co. Incorporated and/ or one or more of its affiliates (collectively, “Morgan Stanley”) and the research analyst(s) named on page 15. Morgan Stanley policy prohibits research analysts from investing in securities/instuments in their MSCI sub industry. Analysts may nevertheless own such securities/ instruments to the extent acquired under a prior policy or in a merger, fund distribution or other involuntary acquisition. Morgan Stanley is involved in businesses that may relate to companies or instruments mentioned in this article. These businesses include market making, providing liquidity and specialized trading, risk arbitrage and other proprietary trading, fund management, investment services and investment banking. Morgan Stanley trades as principal in the securities/instruments (or related derivatives) that are the subject of this article. Morgan Stanley may have a position in the debt of the Company or instruments discussed in this article. THIS MATERIAL IS PREPARED FOR INSTITUTIONAL INVESTOR USE ONLY November 2006 Investment Management Journal • 27 Jack Coates, Ph.D., CFA Managing Director Morgan Stanley Alternative Investment Partners PG.27 Mark Baumgartner, Ph.D., CFA Executive Director Morgan Stanley Alternative Investment Partners Pitfalls and risks in portable alpha implementations In 1984, in a paper entitled “Defining Risk”, a group of researchers suggested that perceived risk can be separated into two components: unknown risk, or aversion to uncertainty, and dread risk, or the fear of severe consequences.1 Anything that is new or not well understood has a high unknown risk component. Anything perceived to be uncontrollable or to have potentially catastrophic consequences has a high dread risk component. As an example, compare the risk perception of two energy sources: coal power and nuclear power. Coal power scores low on unknown risk and dread risk measures, while nuclear power elicits high scores on both. The authors would like to thank Brian Erickson, Vice President and senior administrative officer of Morgan Stanley Alternative Investment Partners, and Dominick Carlino, Senior Associate and portfolio specialist for Liquid Markets portfolios at Morgan Stanley Alternative Investment Partners, for their contributions to this article. 28 • Investment Management Journal November 2006 THIS MATERIAL IS PREPARED FOR INSTITUTIONAL INVESTOR USE ONLY Unfortunately, portable alpha also tips the scales on both perceived risk measures as well. It is viewed as a new and complex approach, and it involves the use of both leverage and derivatives, which are generally thought to increase the risk of severe losses in a portfolio if an unexpected event occurs. But is perception reality? When events occur that are sensational or dramatic, they can be strongly etched in the minds of those who note the event. Then, when questions regarding similar situations arise in the future, in the absence of countervailing data people draw on the only sources of information available to them—their awareness of prior events. Perhaps spectacular and well-publicized failures like Long-Term Capital Management and Orange County served to anchor a fear of leverage and derivatives in investors’ minds. The media also regularly contribute to this perception by reporting extreme losses caused by mistakes or misuse of financial tools. Portable alpha implementations are admittedly complex—there are many moving parts, subtle interactions and sophisticated financial instruments involved. However, understanding and quantifying the risks involved is essential to making an informed decision about how to use portable alpha in your portfolio. Only then can perception be separated from reality. EXPLORING THE RISKS We define portable alpha as a financial engineering methodology that seeks to add low correlation sources of return (alpha) to a portfolio while maintaining the combined portfolio’s desired systematic (beta) exposures. A common objective of such a strategy is to outperform a passive index beta by a meaningful amount while incurring very little incremental risk. To achieve this, the approach typically uses leverage and derivatives. An example is investing in a fund of hedge funds (the alpha engine) while reserving a small portion of the cash to obtain exposure to a fixed income or equity index (the beta source) using a total return swap. There are three major steps in implementing a portable alpha strategy: selecting an optimal alpha engine, maintaining desired beta exposures and, last but not least, obtaining stakeholder support (e.g., the investment committee and/or the plan sponsor) for the strategy. There are pitfalls and risks to be avoided in each of these areas—illustrated in Figure 1. The remainder of this article will be devoted to exploring these risks, both real and perceived. RISKS IN SELECTING AN OPTIMAL ALPHA ENGINE Arguably the most important piece of a portable alpha solution, the alpha engine, is also typically the most opaque and least understood. The objective when selecting an alpha engine for a portable alpha application is to find a consistent, sustainable, positive source of excess return (alpha) that has a low correlation with the beta to which you are transporting it. This is much easier said than done. THIS MATERIAL IS PREPARED FOR INSTITUTIONAL INVESTOR USE ONLY November 2006 Investment Management Journal • 29 FIG 1: PORTABLE ALPHA PITFALLS AND RISKS CLASSIFICATION Portable alpha pitfalls and risks Alpha engine Beta exposures Stakeholder support Sub-par alpha engine Inappropriate engine for objectives High or volatile embedded betas Mis-application of leverage Insufficient attention to overlay management Unintegrated approach Mis-specifying risk Failing to educate the committee Opportunity costs Source: Morgan Stanley Alternative Investment Partners The main risks associated with selecting an optimal alpha engine are: • Selecting a sub-par (e.g., inconsistent, unsustainable, negative or sham) alpha engine. • Using an inappropriate alpha engine to meet desired objectives. • Mis-estimating the magnitude and volatility of betas embedded in the alpha engine. Selecting a sub-par alpha engine. Many pundits equate a manager’s alpha with a manager’s skill and use the terms interchangeably. This is only partially correct. In order to be truly recognized as skillful, a manager must deliver positive, consistent and sustainable alpha. These three attributes are hallmarks of a skillful manager. A common pitfall in selecting an alpha engine is to choose a manager based on a high expected alpha without consideration of the reliability of the manager’s alpha delivery. Tracking error is one measure of this consistency. High tracking errors invalidate the attractiveness of high historical alphas because they reduce the likelihood that these high alphas will be achieved going forward. An investor who chooses a manager with a high tracking error relative to the expected alpha risks being surprised when their manager regresses to the mean in subsequent periods, delivering sub-par returns. Another error of judgment is to seek an alpha engine with a higher volatility based on the assumption that the volatility will eventually translate into higher returns. This is a “beta universe” assumption, where additional risk premiums are believed eventually to result in additional return over the long term. This relationship does not hold in the alpha universe. That is, a manager with low skill will not be expected to deliver alpha going forward simply because they employ a higher volatility approach. Using this manager in a portable alpha application risks transporting only volatility, rather than alpha, to a portfolio. Even managers with low volatilities and consistent track records of alpha delivery must be subjected to further scrutiny. Managers who appear to be generating alpha may in fact be just providing beta masquerading as alpha. This can occur with managers who are simply leveraging fairly priced assets, or with those using highly episodic strategies that pay a fair premium for tail risk and/or optionality (e.g., selling hurricane insurance). These excess returns are not alpha, and they expose investors to a variety of less obvious risks. 30 • Investment Management Journal November 2006 THIS MATERIAL IS PREPARED FOR INSTITUTIONAL INVESTOR USE ONLY FIG 2: THE PORTABLE ALPHA SPECTRUM “Optimal” portable alpha zone Aggressive active Optimal active Too much risk Alpha potential (E(r)) Too little return Incremental active Pure indexing Tracking error (σ) Source: Morgan Stanley Alternative Investment Partners Using an inappropriate alpha engine. Even if an investor has managed to find a true, consistent, positive, sustainable source of alpha, they may face risks because they have not considered their objectives fully when choosing the alpha source. The alpha engine must provide enough return to meaningfully impact the investor’s portfolio, while avoiding imprudent exposure to risks. Figure 2 shows a variety of portable alpha strategies in the context of an active risk/active reward spectrum. At the low volatility/low return end of the spectrum are “incremental active” strategies. These strategies offer a nominal additional expected return over the index, or the ability to decrease portfolio volatility slightly. They can be useful but are not products that allow meaningful shifts in allocations among asset classes within a portfolio. At the other end of the spectrum are the “aggressive active” alpha strategies. These strategies seek to generate maximum amounts of alpha with little regard to the risk incurred. Alpha is a scarce resource in modern, highly efficient markets, and delivering significant amounts of it often involves implementing a more concentrated strategy. However, when choosing a concentrated strategy for their alpha engine, investors must also be willing to accept the significant concomitant idiosyncratic risk. For many investors, however, a middle ground may be more appropriate. This is the “sweet spot” in the portable alpha spectrum, and it achieves two key objectives—first, it is expected to provide enough incremental alpha to allow meaningful shifts in allocations within portfolios (for example, from equities to fixed income) while maintaining or even increasing target returns; second, it should admit only a modest amount of risk to the portfolio, permitting a significant reduction in the overall risk budget if desired. Mis-estimating embedded betas. The third area where potential risk may be incurred in alpha engine selection is in incorrectly assessing the implicit betas embedded within the alpha engine. Ideal alpha engines have consistently low embedded betas, but as with any portfolio the betas can be expected to change over time. Investors who do not consider embedded betas when selecting an alpha engine risk overpaying for that portion of their return. Research has shown that even the average so-called “market neutral” hedge fund has significant embedded beta exposure.2 Some may even have exposures as high as 40%.3 Paying hedge fund fees for this type of beta exposure diminishes the benefit of a portable alpha implementation. THIS MATERIAL IS PREPARED FOR INSTITUTIONAL INVESTOR USE ONLY November 2006 Investment Management Journal • 31 An even greater risk faced by investors is the difficulty involved in accurately estimating these embedded betas. Beta is notoriously tricky to measure and predict, and there is a necessary element of art in its estimation. The fact that art is involved does not preclude the use of quantitative techniques in assessing beta, and in reality some subjectivity is necessitated by the plethora of methods that are employed in the process. For instance, how do you define beta? What timeframe do you consider in your estimates? Do you use unlevered or assetweighted data? Single or multiple regressions? Univariate or stepwise regressions? Principal component analysis? Stochastic betas? Lagged betas? Upside/downside betas? All of these techniques don’t necessarily have to be used in the estimation of embedded betas, but an investor who isn’t able to forecast betas accurately assumes additional risk due to a higher likelihood of unwanted beta exposures in their portable alpha strategy. For instance, ignoring the embedded beta in the aforementioned market neutral hedge fund when implementing the beta overlay would result in a combined beta exposure of 140% for the portable alpha implementation, well above the target beta of 100%. Solutions for selecting alpha engines. To properly address these risks, investors must thoroughly assess the quality of their alpha engine. First, they must set return targets that provide the ability to meet their stated objectives and/or meaningfully adjust their portfolios. Then they must find and access alpha sources that provide a high expectation of continued alpha delivery going forward. And, lastly, they must properly assess the embedded betas within the alpha engine. Selecting an optimal alpha engine is only half the battle. Establishing and maintaining desired beta exposures efficiently and effectively is the next challenge faced by portable alpha investors. RISKS IN MAINTAINING DESIRED BETA EXPOSURES As the “passive” player in the portable alpha equation, the beta source typically plays second fiddle to the alpha engine in considerations. However, the beta source is an equally important contributor to the portfolio—in fact, in most portable alpha applications, the majority of the expected returns and volatility will very likely be attributable to the beta exposure, not the alpha engine. Although portfolio risk (defined as overall portfolio volatility or alternatively as surplus volatility) should generally be the paramount risk in an investor’s mind, often other perceived risks take center stage when portable alpha is considered. One perceived risk is the additional downside risk assumed to be added due to the use of leverage and derivatives in a portable alpha approach. In reality, portable alpha may actually reduce investors’ downside risk. Leverage and derivatives. The fear of leverage and derivatives is not without precedent—the failures of Long-Term Capital Management and Orange County have demonstrated that if misused, leverage and derivatives can be dangerous. Portable alpha strategies do employ leverage and derivatives, but not in the traditional sense. Instead they are used in a more sophisticated application: to obtain exposure to different, or diversifying, asset classes. The question “How much leverage are you using?” is typically asked by those seeking to gauge the riskiness of a particular investment approach. However, this question in isolation does not provide the complete picture. In order to properly characterize the effect of leverage on a portfolio, there are two key additional questions that must be asked along with the question “How much leverage are you using?” They are: • What is the expected volatility of what the leverage is being used to obtain? • What is the expected correlation of what the leverage is being used to obtain? 32 • Investment Management Journal November 2006 THIS MATERIAL IS PREPARED FOR INSTITUTIONAL INVESTOR USE ONLY FIG 3: THE “LEVERAGE ENVELOPE” FOR VARIOUS COMBINATIONS OF VOLATILITY/CORRELATION 4 2.0 0% correlation 50% correlation 100% correlation Relative volatilty of overall portfolio 1.5 1.0 0.5 -50% correlation 0.0 0.0 0.2 -75% correlation 0.4 -100% correlation 0.6 0.8 1.0 Relative volatility of asset transported to portfolio Source: Morgan Stanley Alternative Investment Partners These two dimensions are critical in the assessment of the incremental risk leverage brings to a portfolio. If the correlation of what is being added to the portfolio with leverage is 1.0 (e.g., the identical asset) and the volatility is of the same magnitude, then the amount of added risk is proportional to the amount of leverage used—this is the traditional application of leverage which simply scales the risk of an existing bet. However, if the leverage is being used to add an asset to the portfolio that (a) is not perfectly correlated with or (b) has a lower relative volatility compared with the existing portfolio, a different result is obtained. The relative volatility obtained in either case is lower than would have been realized with a perfectly correlated asset of similar volatility. Thus, risks associated with traditional applications of leverage can be mitigated by leveraging with lower volatility and lower correlation asset classes, as in typical portable alpha approaches. Figure 3 shows the effect of leveraging a portfolio across a range of volatilities and correlations. The highlighted region delineates the “leverage envelope”—the opportunity set for various combinations of volatility and correlation. Note that it may be possible to reduce the relative volatility of the overall portfolio by leveraging with lower volatility or less correlated assets. As an example, consider a portable alpha application where the returns from a fund of hedge funds with an expected volatility of 3.0% is transported to an equity index with an expected volatility of 15.0%— hence, the fund of hedge funds has a volatility of 0.2 relative to the portfolio to which it is transported. Assuming also that the correlation between the fund of funds and the index is 20.0%, the expected volatility of the combined portable alpha solution would only be 15.9%.5 This case is indicated by the yellow point just above the 1.0 line in Figure 3— from a statistical standpoint, it represents an immaterial increase in expected risk. Downside risk. There is an unreasonable level of concern that downside risk is increased when using portable alpha strategies. We define downside risk as the probability of a portfolio realizing a loss—a common concern for many institutional investors. For pension plans, negative returns increase the likelihood of needing to obtain funding from the sponsoring corporation. And endowments and foundations must continually grow assets to keep pace with inflation, even as required spending policies consume capital and erode returns. THIS MATERIAL IS PREPARED FOR INSTITUTIONAL INVESTOR USE ONLY November 2006 Investment Management Journal • 33 FIG 4: DOWNSIDE RISK COMPARISON 4 Equity index Probability Portable alpha 0 Return Source: Morgan Stanley Alternative Investment Partners Figure 4 illustrates how portable alpha can be used to potentially reduce downside risk. The chart shows an assumed returns distribution for a hypothetical equity index portfolio and the returns distribution of a portable alpha solution applied to the same equity index portfolio. The downside risk in each case is shown as the shaded area under the curves to the left of zero. Because the mean of the portable alpha distribution is significantly shifted to the right due to the additional returns provided by the alpha engine but accompanied by just a nominal widening of the spread due to the slight increase in volatility, the effect of the approach is actually to lessen the downside risk of the portfolio. This result is somewhat counterintuitive from a traditional perspective—it demonstrates an application of significant leverage and derivatives in a portfolio to potentially reduce downside risk rather than increase it. Portfolio management. In order to properly maintain the desired beta exposure, a portfolio manager must carefully monitor and manage several aspects of the portable alpha implementation. Diligent portfolio oversight, including rebalancing and overlay management, is essential for reducing the risk of not achieving the portfolio’s objectives. Rebalancing: There are many moving parts in any portable alpha implementation. The alpha engine, the embedded beta and the beta overlay are all changing simultaneously. In addition, asset flows may be entering and leaving the portfolio and correlations among asset classes may be shifting. It is very easy to develop “drift” in your portable alpha solution if the moving parts are not continually monitored and adjusted. There are costs and risks associated with rebalancing, and there are costs and risks associated with not rebalancing. These must be weighed against each other and optimized. It is critical to ensure that there is a robust and systematic rebalancing process to keep the portfolio aligned with objectives while minimizing cost. Overlay management: Risk may be incurred through mismanagement of the derivative overlay. Although institutional investors are warming to the use of derivatives in portfolios, there are a number of issues that must be dealt with when gaining exposures through these means. There are basis and cash settlement risks associated with derivatives. Investors should be aware of these risks before undertaking a portable alpha program. Also, if swaps are used, it is important to mitigate timing and counterparty risks through active management of the swap program. 34 • Investment Management Journal November 2006 THIS MATERIAL IS PREPARED FOR INSTITUTIONAL INVESTOR USE ONLY Another risk entailed in managing the overlay is sacrificing returns due to significant cash drag, or, on the opposite end of the spectrum, paying more in financing costs than required because inadequate cash is used to satisfy counterparties. Effective cash management is another essential factor in reducing risk in a portable alpha implementation. These portfolio management techniques are straightforward, there is nothing mysterious about them and they are becoming more commonplace as more derivatives are used in institutional portfolios. But they do require oversight and management, which in turn requires resources. The implementation risk inherent in these vehicles arises if they are not given the proper attention prior to, during, and after implementation. Solutions for maintaining desired beta exposures. One way to address these issues is to outsource the management of a portable alpha program to a third party provider. However, there has been a boom in the number of portable alpha offerings in the marketplace, also representing a range of associated risks. At the riskier end of the spectrum are the “bolt-together” portable alpha solutions that simply combine alphas and betas with little concern for the actual exposures, embedded betas, interactions or risks. At the lower-risk end of the spectrum are solutions that account for the risks associated with each element of the portable alpha solution, and are highly focused on managing and mitigating not only the risks associated with each element, but also how they interact in the overall system. Whether you decide to do this yourself or outsource it to a third party, evaluate the approach holistically and make certain you are comfortable with the level of risk associated with your portable alpha implementation. A highly integrated approach focused on alpha, beta and the nuances of their interrelationship, is essential for effective risk management. OBTAINING AND MAINTAINING STAKEHOLDER SUPPORT Without stakeholder support, it is difficult to execute and sustain a portable alpha implementation. Two of the more common pitfalls in the process of obtaining and maintaining stakeholder support are not properly specifying what risks you are attempting to mitigate, and/or not properly educating your investment committee or board regarding portable alpha. Properly specifying risk. How should risk be defined? Overall volatility? Downside risk? Surplus risk? Underfunded risk? Spending risk? The answer is: it varies depending on objectives. However, the era of risk being thought of as overall volatility is nearing an end. Now there is increased focus on asymmetric return distributions, fat tail risk, assetliability mismatch and/or not achieving spending policy requirements. Portable alpha is an answer to these new types of risks—it breaks a number of traditional compromises in investment space and provides additional flexibility to address the real risks in a portfolio. Ensure that the solution you choose meets the true needs and objectives of your portfolio. Education. We have examined both perceived and actual risks in this article. Education is the key to dismissing fears of perceived risks and instead focusing discussions on the actual risks involved. You should draw upon your managers, consultants and trusted advisors to help you with the education process and set correct expectations regarding a portable alpha approach. THIS MATERIAL IS PREPARED FOR INSTITUTIONAL INVESTOR USE ONLY November 2006 Investment Management Journal • 35 FIG 5: PORTABLE ALPHA MAY BE APPLIED TO VARIED EXTENTS WITHIN A PORTFOLIO Basic use Use in portion of single asset class (e.g., fixed income) One danger you may face is that your committee or board may believe portable alpha is guaranteed to provide excess returns in all circumstances. Even the best alpha engines experience volatility, and a shortterm drawdown—especially if it occurs just after implementation—can create a predicament for those supporting the strategy. To prevent this and ensure commitment to the approach, make sure it is written into your investment policy statement and adhered to in the long run. Solutions for obtaining and maintaining stakeholder support. Typically, risk is considered as the by-product of initiating some action in a portfolio, however, there are opportunity costs associated with inaction as well. Alpha is a scarce resource and the early movers in this space may end up reaping most of the rewards. One way you can help your investment committee or board become comfortable with the risks (both real and perceived) is to trial portable alpha in your portfolio. Figure 5 shows a range of ways portable alpha may be applied in your portfolio, from limited to full extent. In its most basic form you can use it to increase the returns in a portion of one of your low expected alpha allocations, for example, a passive fixed income exposure. In this way, you can test it out, observe the results and demonstrate its effect to your board. Once your board or committee gains confidence in the approach, you can increase use of the strategy by transporting alpha to all of your fixed income investments; then you can enhance your use by reducing exposure to higher risk asset classes like equities. As confidence builds, you may choose to add it to all of your basic asset classes to broadly improve expected returns, or eventually apply it to your entire portfolio in a custom architecture that makes the best use of your risk budget. Increased use Use in 100% of single asset class Enhanced use Shift mix of asset classes to lower overall volatility, improve asset-liability mismatch and/or reduce spending shortfall risks Broad use Use to supplement all basic asset classes Full use Apply custom architecture to entire portfolio Equities Other Fixed income Portable alpha Private equity Real assets Source: Morgan Stanley Alternative Investment Partners 36 • Investment Management Journal November 2006 THIS MATERIAL IS PREPARED FOR INSTITUTIONAL INVESTOR USE ONLY CONCLUSION There are many risks investors must consider when evaluating a portable alpha implementation. Some are real, such as the risk of obtaining a sub-par implementation if adequate resources are not devoted to the task of sourcing alpha or carefully managing the beta overlay. And some, like the use of leverage and/or derivatives, are only perceived to be risky but may actually, counterintuitively, serve to decrease the risk in a portfolio if used prudently. Institutional investors must be certain to discern between real and perceived risks when considering the application of portable alpha in their portfolio. Anything new, unknown and perceived to have catastrophic consequences faces severe adoption challenges. Such is the case with portable alpha. However, by maintaining the status quo and avoiding the use of portable alpha, investors may actually be accepting more risk than needed. As Charles Darwin once said, “It is not the strongest of the species that survives, nor the most intelligent that survives. It is the one that is the most adaptable to change.”6 Important information This article has been prepared solely for information purposes and is not an offer to buy or sell or a solicitation of an offer to buy or sell any security or instrument or to participate in any trading strategy. This article reflects the views of the authors at the time of writing. These views may change in response to changing circumstances and market conditions. Alternative investments are speculative and include a high degree of risk. Investors could lose all or a substantial amount of their investment. Alternative investments are suitable only for long-term investors willing to forgo liquidity and put capital at risk for an indefinite period of time. Alternative investments are typically highly illiquid—there is no secondary market for private funds, and there may be restrictions on redemptions or assigning or otherwise transferring investments in private funds. Alternative investments typically have higher fees and expenses than other investment vehicles, and such fees and expenses will lower the returns achieved by investors. Fund of funds often have a higher fee structure than single manager funds as a result of the additional layer of fees. Alternative investment funds are often unregulated and are not subject to the same regulatory requirements as mutual funds, and are not required to provide periodic pricing or valuation information to investors. The investment strategies described in the preceding pages may not be suitable for your specific circumstances; accordingly, you should consult your own tax, legal or other advisors, at both the outset of any transaction and on an ongoing basis, to determine such suitability. This article is prepared for sophisticated investors who are capable of understanding the risks associated with the investments described herein and may not be appropriate for you. No investment should be made without proper consideration of the risks and advice from your tax, accounting, legal or other advisors as you deem appropriate. 1 Baruch Fischoff, Stephen R. Watson and Chris Hope, “Defining Risk. Policy Sciences”, Volume 17, 1984, pp. 123-139. Andrew Patton, “Are Market Neutral Hedge Funds Really Market Neutral?”, Working paper, London School of Economics, 2005. Dominic Clermont, “Don’t Pay for Beta”, Canadian Investment Review, Spring, 2004. Illustrative example only. Illustrative example only. Assumes normality of returns and portfolio variance according to the equation 2 = w12 12 + w22 22 + 2 w1 w2 1, 2 1 2, where w is the asset weight, is the volatility, and is the correlation between the assets. Charles Darwin, (1859) The Origin of Species, Paperback edition, University of Pennsylvania Press, Philadelphia, Pennsylvania, 2006. 2 3 4 5 6 THIS MATERIAL IS PREPARED FOR INSTITUTIONAL INVESTOR USE ONLY November 2006 Investment Management Journal • 37 Jan Baars, Ph.D. Vice President Morgan Stanley Investment Management PG.37 Petr Kocourek Executive Director Morgan Stanley Investment Management Epco van der Lende, Ph.D. Executive Director Morgan Stanley Investment Management Risk budgeting in an Asset-Liability Management context Determining an appropriate risk budget for active management has historically taken place without much regard to the determination of the investment strategy. Yet adding an alpha-generating active management strategy has important implications for the absolute risk/return trade-off that lies at the heart of finding the right long-term asset mix in the first place. In this article we show how the risk budget can be determined as an integral part of an Asset-Liability Management study, providing internal consistency between the long- and short-term aspects of investment policy. The analysis also allows for separation of alpha and beta exposures. Furthermore, we investigate how successful active management needs to be, in terms of the information ratio, for it to be worthwhile, as well as the maximum risk budget that is appropriate. 38 • Investment Management Journal November 2006 THIS MATERIAL IS PREPARED FOR INSTITUTIONAL INVESTOR USE ONLY INTRODUCTION The concept of “risk budget” has been used in recent years as a lodestar for active management of an investment portfolio. The risk budget is traditionally expressed in terms of tracking error relative to a long-term strategic benchmark, which in turn is determined by an Asset-Liability Management (ALM) study. However, there is usually a disconnect between the two, as the risk budget is determined exogenously, without regard to the strategic benchmark or the ALM study. An ALM study will determine the strategic benchmark by finding the optimal trade-off between absolute risk and return, or functions thereof. For a defined benefit pension fund, for instance, the absolute risk/return trade-off might consist of minimizing the present value of contributions in the worst case, while maximizing the expected funding ratio within this framework. The strategic benchmark selected will embody this trade-off. Yet adding active management to this benchmark introduces some tracking error, thus altering the trade-off. It is, therefore, important to quantify the impact of active management on the absolute risk/ return trade-off. In fact, given the crucial nature of this trade-off, determining how much tracking error one can add should be done within the context of an ALM study, and not as an afterthought. The risk budget and strategic benchmark are interrelated: by adding a risk budget the strategic benchmark should change. This is because the risk profile of the allocations changes, which in turn affects the trade-off. We can also calculate the information ratio that is required to make up for the additional risk that has been introduced. There is also a benefit that may be derived from separating alpha and beta. Traditionally, an ALM study has only resulted in finding an optimal mix of long-term beta exposures. Given that equity and bond market betas can be modeled with more long-term significance than alpha, this has been a reasonable approach to take.1 Now, however, we can also incorporate largely uncorrelated alpha into the portfolio at the ALM stage. In the next section we present the conceptual theoretical setting for this article. We show in particular how in the ALM context a strategic benchmark (for beta exposure) can be derived in combination with a risk budget (for uncorrelated alpha). On the basis of the above-mentioned theoretical setting we then move to an illustrative case study for the rest of the article. We model the case of a foundation with a well-defined spending pattern and a desire to accumulate capital over time. The methodology employed is more or less standard ALM practice. We define the macroeconomic environment and distribution assumptions for the asset categories used, and we explain how ALM simulations lead to the assessment of relevant probabilities. Finally, we present the case study results. One of the main concluding observations is that the appropriate risk budget for active management is indeed closely related to the long-term strategic benchmark. An increasing risk budget (up to the maximum possible) results in a lower strategic allocation to equities. However, our modeling shows that the decrease in expected foundation capital due to this more conservative strategy can be more than compensated for by implementing an uncorrelated alpha-strategy at reasonable levels of required information ratio. THIS MATERIAL IS PREPARED FOR INSTITUTIONAL INVESTOR USE ONLY November 2006 Investment Management Journal • 39 FIG 1: BASIC RISK/RETURN TRADE- OFF MRP Risk criterion Acceptable level ARP S-curve Objective Source: Morgan Stanley Investment Management CONCEPTUAL THEORETICAL SETTING The typical end result of an ALM study is presented in Figure 1. A range of alternative (beta only) investment strategies is scored against an objective function (x axis) and some downside criterion (y axis). Each point of the purple S-curve represents the score for some strategy, ranging from extremely conservative (left end) to highly aggressive (right end). One can now either simply recommend the “minimum risk portfolio” (MRP) that scores best on the downside criterion, or can set a lower acceptable downside criterion and then maximize the objective. The resulting recommendation in this case will be the “acceptable risk portfolio” (ARP), which is the point furthest to the right (maximizing the objective function) while staying above the acceptable risk level. Note that the shape of the S-curve was not chosen arbitrarily. Although stylized, this shape is typical for most practical applications where the range of alternative strategies is actually an efficient frontier, the objective being related to the upper tail of the return distributions and the downside criterion to the lower tail. A trivial example would be to take the expected return as the objective, and the worst case return at a certain confidence level as the risk criterion. Traditionally, having established the appropriate beta for the portfolio, the ALM study would be concluded at this stage. The setting of a risk budget, if any, would take place entirely separately from this exercise. Such an approach is far from perfect— the risk budget evidently should be taken into account as it affects the long-term strategic risk profile. We employ a two-step approach to overcome this: • Overlay an uncorrelated asset with volatility = 0 and = 0 and investigate the effect this has on the strategy and the risk/return trade-off. • Determine the level of alpha necessary to compensate for the additional volatility. By starting off with an uncorrelated asset possessing zero alpha and with volatility 0, we can avoid having to make assumptions on manager alpha. First of all one should realize that adding the overlay introduces new volatilities to both the outcomes for the objective and the downside criterion. This effect is conceptually represented by the circle areas2 around the S-curve in Figure 2. The radius of these circles is stylized as 0. Connecting the lower ends of this family of circles as the new downside levels leads to a transformation of the S-curve to a new curve, labeled as T0, 0(S). 40 • Investment Management Journal November 2006 THIS MATERIAL IS PREPARED FOR INSTITUTIONAL INVESTOR USE ONLY FIG 2: ADDING VOLATILITY TO THE STRATEGY WITH AN UNCORRELATED OVERLAY Risk criterion Acceptable level σ0 S-curve = T0, 0 (S)-curve T0, σ (S)-curve 0 Objective Source: Morgan Stanley Investment Management It is clear that the setting now changes. The previous MRP or ARP recommendations turn out to be irrelevant or at least sub-optimal since they were derived from the S-curve, which has been superseded by the T0, 0(S)-curve. The new MRP and ARP are shown in Figure 3. Note that the T-curve was derived from the S-curve by adding an uncorrelated asset to the original strategies. In terms of its composition the new MRP will typically be derived from a strategy close to the previous MRP, but the same certainly does not hold for the new ARP. This is derived from a certain strategy S0, which will generally be considerably more conservative than the original ARPS. One can FIG 3: easily see this also from their different positions on the relevant curves. Although by definition the new and the old ARP both have the same downside level, it is clear that in terms of the objective the old ARP is superior as it lies further to the right in the graph. This is due to our zero alpha assumption for the overlay, which has the net effect of adding volatility without any added value. The next step is to add alpha to the overlay. A positive alpha will move the new ARP to the right and upwards, since it will improve the situation on the downside and will also help in achieving the RISK/RETURN TRADE- OFF WITH ZERO ALPHA OVERLAY MRPS Risk criterion S0 ARPS ARPT 0, σ 0 MRPT Acceptable level 0, σ 0 (S) (S) = T0, σ (S0) 0 S-curve T0, σ (S)-curve 0 Objective Source: Morgan Stanley Investment Management THIS MATERIAL IS PREPARED FOR INSTITUTIONAL INVESTOR USE ONLY November 2006 Investment Management Journal • 41 FIG 4: ADDING ALPHA TO THE OVERLAY MRPS Risk criterion S0 Tα , σ (S0) 0 0 Tα, σ (S0)-curve 0 Acceptable level ARPT 0, σ 0 (S) S-curve T0, σ (S)-curve 0 Objective Source: Morgan Stanley Investment Management objective. A negative alpha will do exactly the opposite. The purple-dotted curve, labeled T , 0(S0) in Figure 4 represents this effect for a range of alphas. We highlight one particular point on this curve, namely the point where the result for the objective is equal to the original ARPS. This point, labeled T 0, 0(S0), shows superior behavior with respect to the risk criterion. Therefore, if active management is able to deliver 0 excess return within a tracking error of 0, or equivalently an information ratio of 0/ 0, it is worth having S0 as a strategy and adding an overlay with 0 volatility, instead of sticking passively to the original (more aggressive) ARPS. In each specific case, one can assess the level of the required information ratio and take decisions accordingly. Note that the portfolio T 0, 0(S0) is in fact the point on the T 0, 0(S)-curve that corresponds to the original strategy S0, as shown in Figure 5. In fact, there is a continuum of curves that we can calculate, depending on one of the parameters alpha, volatility and strategies S0, as illustrated in Figure 6. One final question that this methodology can address is whether there is a maximum risk budget. Going back to the zero alpha situation, for the sake of prudence, one can draw T-curves for increasing volatilities for the overlay. At a certain point the FIG 5: SHIFTING THE ENTIRE CURVE MRPS Risk criterion S0 Tα , σ (S0) 0 0 Tα, σ (S0)-curve 0 Acceptable level ARPS ARPT 0, σ 0 (S) S-curve T0, σ (S)-curve 0 Tα , σ (S)-curve 0 0 Objective Source: Morgan Stanley Investment Management 42 • Investment Management Journal November 2006 THIS MATERIAL IS PREPARED FOR INSTITUTIONAL INVESTOR USE ONLY FIG 6: ENTIRE FAMILY OF CURVES CAN BE GENERATED Tα , 0 (S0) 0 Tα, 0 (S0)-curve Tα, σ (S0)-curve 0 MRPS Risk criterion S0 Tα , σ (S0) 0 0 Acceptable level ARPS ARPT 0, σ 0 Tα , 0 (S)-curve 0 (S) = T0, σ (S0) 0 S-curve T0, σ (S)-curve 0 Tα , σ (S)-curve 0 0 Objective Source: Morgan Stanley Investment Management T-curve will end up tangential to the acceptable level. Going beyond that point will lead to a situation where none of the T-curve portfolios satisfies the risk tolerance criterion. The T-curve will then be entirely below the acceptable level. Hence the maximum allowed volatility max is reached when T0, (S) is tangential to the acceptable level, with only one portfolio satisfying the risk criterion. This portfolio is then the MRP and the ARP at the same time for T0, max(S) (Figure 7). Although not too hard to visualize in this stylized form, putting this concept into practice can be a lot harder. In order to make the above concept work we will have to complete all necessary ALM-related steps. We make a start in the next section. FIG 7: DETERMINING THE MAXIMUM RISK BUDGET σmax Risk criterion MRP Acceptable level ARP MRPT 0, σ max (S) = ARPT 0, σ max (S) S-curve T0, σ (S)-curve 0 T0, σ (S)-curve max Objective Source: Morgan Stanley Investment Management THIS MATERIAL IS PREPARED FOR INSTITUTIONAL INVESTOR USE ONLY November 2006 Investment Management Journal • 43 FIG 8: EXPECTED ASSET CLASS CHARACTERISTICS Correlations Asset classes Expected return 2.10% 4.50% 7.25% Expected volatility 0.76% 3.79% 17.14% Price inflation 1.00 -0.15 -0.06 European bonds -0.15 1.00 0.15 Global equities -0.06 0.15 1.00 Price inflation European bonds Global equities Source: Morgan Stanley Investment Management Research CASE STUDY— PRELIMINARIES To set the stage for the calculations, we use the main economic climate as we have currently defined it. In broad terms this economic climate consists of a global macroeconomic environment of moderate inflation and moderate GDP growth. Using our Long-Term Asset Return Model (LTARM) we determine the long-term premiums of risky asset categories. Using the GDP and inflation assumptions as inputs, as well as assumptions on earnings growth and payout ratios, the model uses a vector autoregressive Monte Carlo simulation process to provide us with a distribution of expected risk premiums. (To provide a full description of the LTARM is beyond the scope of this article.3) In order to focus on the main issue of setting the risk budget, we have used a highly simplified twoasset model consisting of just European bonds and global equities. Using the LTARM we have assigned expected returns to both European bonds and global equities, rounding them slightly (Figure 8). The long-term expected volatility and correlations are based on historical data (from January 31, 1988 to May 31, 2006) using the following indices: • Price inflation • European bonds • Global equities CPI EMU Index JPM European Government Bond Index MSCI World Index The period under consideration has a largely homogenous economic climate, which exhibits by and large the characteristics of economic growth and inflation that we foresee for the future. As such, historical data provides a decent starting point for our modeling, even though we need to take into account that some aspects will not be replicated. From current starting levels, for instance, it is not plausible to assume the same kind of secular disinflation that marked the period under consideration. 44 • Investment Management Journal November 2006 THIS MATERIAL IS PREPARED FOR INSTITUTIONAL INVESTOR USE ONLY FIG 9: SHORTFALL PROBABILITY, HORIZON 12 MONTHS 100 90 80 70 ty (%) Probabili 60 50 40 30 20 10 0 2 26 0 42 51 59 67 7 10 14 Bond weigh 75 t (%) 83 93 -7 100 Source: Morgan Stanley Investment Management Since we are modeling just two asset categories, we can explore exhaustively all possible combinations of the two. In a more realistic setting there would be a larger number of asset categories involved, making a separate optimization necessary to shrink down the possible number of combinations that need to be examined. In this case, we can simply investigate portfolios ranging from 100% European bonds to 100% global equities. Using the expected returns shown above, we can easily calculate some characteristics analytically for these portfolios (Figure 9), such as the shortfall probability relative to a certain return target. With increasing bond weight the shortfall risk gradient becomes steeper as the target return changes. This is due to the lower volatility of the bonds, which in turn means a narrower probability distribution. For all-equity portfolios, on the left of the graph, the shortfall risk changes far less with changing target returns. Obviously this is just one way of examining the characteristics of possible asset mixes. While easy to model, it is not sufficient for reaching the kind of risk/return trade-off in which most institutional investors will be interested. After all, the true quantities of interest will be functions of these risks and returns, in combination with the actual liabilities or capital development. In order to address these more complex interactions, an analytical treatment will no longer suffice and we have to rely on Monte Carlo simulations. Ta r 34 ge t 18 re tu rn (% ) THIS MATERIAL IS PREPARED FOR INSTITUTIONAL INVESTOR USE ONLY November 2006 Investment Management Journal • 45 FIG 10: SCENARIO RESULTS CAPITAL Period ending in 2020 100% European bond portfolio 1000 900 800 700 Capital Capital 100% global equity portfolio 1000 900 800 700 600 500 400 300 200 100 0 600 500 400 300 200 100 0 0 100 200 300 400 500 600 700 800 900 1000 0 100 200 300 400 500 600 700 800 900 1000 Source: Morgan Stanley Investment Management Research So as not to detract from the purpose of this article, we have kept the case under consideration relatively simple. The modeling only encompasses a foundation with a well-defined spending pattern, where we focus on the capital development over time. Return on capital is the only assumed source of income. We set the start capital equal to 100 and the horizon at 15 years (ending in 2020). The foundation’s spending rule is as follows: • Annual spending equals 3% of average capital in the three previous years, but is not allowed to decrease in nominal terms. This puts us close to the simplest analytically solvable case. We do, however, resort to our simulation-based ALM models here as these are more generic, and the same process we show can be applied to any complex optimization problem that includes liabilities (e.g., a defined benefit pension fund). CASE STUDY—METHODOLOGY Our simulation model uses a vector auto-regressive time series model that generates internally consistent future time paths for all relevant variables, ranging from macroeconomic to asset class returns, yield curve with full-term structures and any liabilityrelated items, as well as a complete balance sheet. In this case, again, we restrict ourselves to the bare minimum to illustrate the general theme of this article. Simulating 1,000 possible future time paths, we can look at the impact on total capital growth over time for different asset mixes. For instance, the graph on the left in Figure 10 shows the 1,000 outcomes for capital of these time paths in the year 2020. Here we are assuming that the capital is fully invested in European bonds. The horizontal axis shows the sequential number of the simulated time path, while the vertical axis shows the capital accrued. 46 • Investment Management Journal November 2006 THIS MATERIAL IS PREPARED FOR INSTITUTIONAL INVESTOR USE ONLY FIG 11: HISTOGRAM SCENARIO RESULTS CAPITAL Period ending in 2020 100% European bond portfolio 300 270 240 210 180 150 120 90 60 30 0 0 130 260 390 Capital Source: Morgan Stanley Investment Management Research 100% global equity portfolio 1000 900 800 700 600 500 400 300 200 100 0 520 650 780 40 30 20 10 0 0 130 260 390 Capital 520 650 780 90 80 70 60 50 1000 900 800 700 600 500 400 300 200 100 0 The distribution of the results is fairly narrow and shows few, if any, outliers. This is to be expected with an asset class that has low volatility and a low expected return. Using the same scale, we can also show the impact of the other extreme in asset mix terms, the 100% global equity portfolio. The graph on the right in Figure 10 shows that the capital in the year 2020 will in general be higher, but also with a much wider range of possible outcomes. This is, again, consistent with a higher expected return and higher volatility. Obviously, similar graphs can be produced for all combinations of European bonds and global equities. To get a better view of these simulated outcomes, we have created histograms of this data. For the 100% European bond portfolio, the result is shown in the graph on the left in Figure 11. The distribution’s narrow profile is more clearly visible here. Using the same scale in the graph on the right, we can clearly see the difference with the 100% global equity portfolio. The gray area in the two graphs in Figure 11 shows the cumulative distribution. Comparing the two graphs we find that the 100% global equity portfolio shows a much wider range of results for capital in the year 2020 than the 100% European bond portfolio. While this means that much better outcomes are possible, there is a price to pay: the distribution extends further to the left as well. Equities do bring with them the risk of a lower overall result. THIS MATERIAL IS PREPARED FOR INSTITUTIONAL INVESTOR USE ONLY November 2006 Investment Management Journal • 47 FIG 12: CAPITAL PERCENTILES 100% European bond portfolio 450 400 350 300 Capital Capital 122.6 250 200 150 100 50 0 2005 2008 2011 2014 95% Source: Morgan Stanley Investment Management Research 100% global equity portfolio 450 400 350 300 250 200 150 100 93.5 2017 75% 2020 50 0 2005 50% 2008 25% 2011 2014 5% 2017 2020 32.7 131.0 These charts show just one snapshot in the year 2020, but it is also instructive to see how the distribution changes over time. In the chart to the left in Figure 12 we show the evolution of capital from 2005 to 2020 for a 100% European bond portfolio. Starting with a capital of 100, the gray line (labeled 50%) shows the median of the distribution in each year. The lines on either side are respectively the fifth, 25th, 75th and 95th percentiles of the distribution. That means that for the yellow-dotted line (the fifth percentile), 950 out of 1,000 future time paths resulted in capital higher than this line. Equally for the purple-dotted line (25th percentile), 750 out of 1,000 future time paths resulted in capital above that line. We can obviously produce the same chart for the 100% global equity portfolio, and the result is shown in the graph to the right. The lines fan out much more than for the 100% European bond portfolio, leading to better results on the upside, but worse on the downside. The tables in Figure 13 show the same data as the graphs in Figure 12 in numerical format. At the 50th percentile, or median, a 100% European bond portfolio will grow capital from 100 to 122.6 in 2020. The 5% worst-case outcome is a capital of 93.5. Contrast that with the 100% global equity portfolio: in the median case the capital grows to 131.0, while in the 5% worst case it shrinks to 32.7. 48 • Investment Management Journal November 2006 THIS MATERIAL IS PREPARED FOR INSTITUTIONAL INVESTOR USE ONLY FIG 13: CAPITAL OUTCOME PERCENTILES 100% European bond portfolio Percentiles 5% 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017 2018 2019 2020 100.0 95.9 94.0 93.2 93.3 92.0 92.0 92.7 92.1 92.9 92.5 92.2 92.5 92.7 93.3 93.5 25% 100.0 99.1 99.3 99.6 100.4 101.1 102.0 102.4 103.4 104.4 105.8 105.7 106.4 107.4 108.3 108.6 50% 100.0 101.3 102.7 104.2 105.3 106.4 108.2 110.2 112.5 113.3 115.0 115.9 117.7 118.8 120.3 122.6 75% 100.0 103.8 106.5 108.9 111.2 112.9 115.1 117.9 120.0 122.0 124.1 126.7 128.5 131.9 134.2 137.1 95% 100.0 107.5 112.1 116.2 120.1 122.4 125.7 130.8 133.5 137.1 140.1 143.0 146.1 150.4 155.0 158.7 μ 100.0 101.4 102.9 104.4 105.8 107.1 108.7 110.6 112.3 113.8 115.3 116.7 118.3 120.1 121.8 123.6 0.0 3.6 5.4 6.9 8.2 9.2 10.2 11.5 12.7 13.6 14.4 15.6 16.8 18.1 18.9 20.2 0.0 0.1 0.1 0.1 0.3 0.2 0.2 0.2 0.2 0.3 0.2 0.3 0.3 0.3 0.3 0.3 100% global equity portfolio Percentiles 5% 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017 2018 2019 2020 100.0 77.1 71.4 65.3 59.8 56.1 53.7 49.7 46.5 43.3 43.0 40.5 39.5 37.5 35.2 32.7 25% 100.0 91.7 89.6 86.5 86.8 86.1 83.7 82.0 84.3 83.5 83.4 81.0 80.8 80.4 78.0 77.4 50% 100.0 103.0 105.2 108.4 110.2 112.7 115.3 116.7 119.3 123.1 121.9 123.4 125.7 128.1 130.0 131.0 75% 100.0 113.9 122.4 131.1 136.1 144.1 150.8 160.1 166.8 171.8 176.7 184.3 186.3 197.3 201.7 210.2 95% 100.0 133.4 154.9 171.4 190.0 209.4 231.7 258.0 271.3 295.2 316.4 327.5 349.6 362.1 397.9 443.2 μ 100.0 103.7 107.6 111.4 115.8 120.0 124.8 129.4 134.0 138.6 142.6 145.7 150.6 156.0 160.8 166.3 0.0 17.5 25.3 33.7 41.5 50.2 57.9 66.5 72.7 80.1 88.9 93.4 103.4 116.6 126.6 133.8 0.0 0.6 0.6 0.8 1.0 1.6 1.7 1.6 1.7 1.8 2.0 1.7 1.9 2.3 2.3 2.0 THIS MATERIAL IS PREPARED FOR INSTITUTIONAL INVESTOR USE ONLY November 2006 Investment Management Journal • 49 FIG 14: INVESTMENT STRATEGY RISK ANALYSIS: BASIC RISK/RETURN TRADE- OFF 100 ARP 5% worst-case capital 80 MRP Acceptable risk level = 75 60 S-curve Trend 40 20 120 130 140 Expected capital Source: Morgan Stanley Investment Management Research 150 160 170 So far we have only looked at the extremes of the alternative asset mixes. Tables and graphs similar to the ones shown on the previous pages can be created for all intermediate points along the curve. To be able to make a risk/return trade-off, we need to look at this entire set of data and lift out some relevant measures of risk and return. In this case, we will use: • Objective: Expected outcome for capital in 2020 • Risk criterion: 5% worst-case outcome for capital in 2020 Figure 14 shows the risk/return trade-off. On the horizontal axis the expected capital is shown for all asset mixes ranging from 100% European bonds on the left to 100% global equities on the right. Note that this is exactly the S-curve from our previously explained theoretical setting. The composition of MRPS (purple triangle) is 91.9% European bonds, 8.1% global equities. Given an “acceptable risk” level of capital being at least 75 in 2020, the composition of ARPS (purple circle) would be 56.6% European bonds and 43.4% global equities. 50 • Investment Management Journal November 2006 THIS MATERIAL IS PREPARED FOR INSTITUTIONAL INVESTOR USE ONLY FIG 15: INVESTMENT STRATEGY RISK ANALYSIS: ZERO ALPHA OVERLAY 100 ARP 5% worst-case capital 80 MRP Acceptable risk level = 75 60 ARP MRP 40 Trend T0%, 5%(S)-curve 20 120 130 140 Expected capital 150 160 170 S- curve Trend Minimum Risk Portfolio S Expected return Volatility European bonds Global equities 4.7% 4.0% 91.9% 8.1% T0%, 5%(S) 4.8% 4.2% 88.9% 11.1% Acceptable Risk Portfolio S 5.8% 8.2% 56.6% 43.4% T0%, 5%(S) 5.2% 5.5% 75.8% 24.2% Source: Morgan Stanley Investment Management Research RISK BUDGETING We are now going to follow the path from the theoretical setting and introduce an overlay. We assume zero alpha, 5% volatility and zero correlation. The resulting curve is labeled T0%, 5%(S). The net effect is adding active management to the portfolio with a tracking error of 5% but zero alpha (Figure 15). The MRPs as indicated by the triangles are shifted slightly from one another, but the portfolio composition for each is almost identical, as shown in the data table. The big difference, however, is in the ARP, whose composition changes from the 57% European bonds for the beta-only purple S-curve, to 76% European bonds once tactical management has been added in the blue beta-plus-alpha curve T0%, 5%(S). Note that the MRP and ARP on T0%, 5%(S) are derived from adding the overlay to beta-only strategies with exactly that composition. The two circles in the graph show the ARPs and MRPs on each curve, and the table above shows the portfolio compositions. The expected return and volatility for the two MRPs is almost the same, but in the case of the ARPs there are again significant differences. The added volatility of the active management forces the strategy to become more conservative if the same worst-case outcome is to be attained. This is at the cost of lower expected outcomes. The expected return itself falls from 5.8% to 5.2%. THIS MATERIAL IS PREPARED FOR INSTITUTIONAL INVESTOR USE ONLY November 2006 Investment Management Journal • 51 FIG 16: CAPITAL DEVELOPMENT PERCENTILES Capital ARPS Percentiles 5% 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017 2018 2019 2020 Capital ARPT 100.0 89.8 86.5 84.2 82.7 80.1 79.6 78.9 79.7 78.0 77.2 77.0 77.5 75.8 75.4 75.0 25% 100.0 96.7 97.1 96.3 97.3 97.9 98.6 99.3 101.3 102.1 104.3 102.7 103.8 104.2 106.0 107.4 50% 100.0 102.1 103.8 106.2 109.1 110.7 113.9 116.2 117.6 120.8 122.9 126.0 127.2 130.9 132.9 134.8 75% 100.0 107.5 112.7 117.4 120.3 125.2 129.1 134.0 137.8 142.0 145.6 150.4 153.2 156.9 161.0 167.7 95% 100.0 116.8 125.6 133.2 141.3 149.3 160.0 168.0 173.0 182.7 187.7 195.2 207.6 213.7 229.1 237.1 μ 100.0 102.4 105.0 107.3 110.0 112.5 115.4 118.3 121.4 124.2 126.7 129.1 132.1 135.3 138.3 141.7 0.0 8.2 11.7 15.3 18.3 21.3 24.0 27.3 29.3 31.8 34.2 36.4 39.6 43.0 46.1 49.2 0.0 0.5 0.4 0.4 0.5 0.7 0.7 0.7 0.7 0.8 0.8 0.7 0.8 0.9 1.0 1.0 0%, 5%(S) Percentiles 5% 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017 2018 2019 2020 100.0 90.1 87.4 86.3 85.0 82.2 81.8 79.9 81.1 80.6 79.1 76.9 75.1 76.2 77.1 75.0 25% 100.0 96.8 96.9 97.0 97.2 96.9 97.5 98.2 98.9 100.3 100.4 100.4 100.6 102.1 102.5 103.0 50% 100.0 101.9 103.5 105.5 107.0 108.8 110.3 112.4 114.2 115.7 117.7 118.8 121.0 123.5 125.1 127.7 75% 100.0 106.7 111.2 114.8 117.8 120.9 124.9 129.0 133.4 135.4 138.6 141.7 145.1 147.4 150.7 155.9 95% 100.0 114.6 122.0 128.8 136.2 143.4 148.4 156.6 161.0 170.5 174.7 181.1 190.1 195.0 199.4 210.7 μ 100.0 101.9 104.2 106.0 108.3 109.9 112.0 114.6 117.2 118.9 120.9 122.8 125.2 127.7 129.9 132.5 0.0 7.3 10.5 13.0 15.8 18.2 20.7 23.4 25.2 27.5 29.4 31.5 34.0 36.1 38.3 40.8 0.0 0.1 0.2 0.2 0.3 0.4 0.5 0.6 0.6 0.6 0.6 0.6 0.7 0.7 0.7 0.7 52 • Investment Management Journal November 2006 THIS MATERIAL IS PREPARED FOR INSTITUTIONAL INVESTOR USE ONLY We can investigate the properties of these two ARPs, and these characteristics are shown in the tables in Figure 16 and graphs in Figure 17. In the S-curve of beta-only unperturbed portfolios (first table), the median outcome in 2020 is 134.8 for the capital, with the worst-case outcome at the fifth percentile equal to 75.0. That is exactly the “acceptable risk” level that we assumed earlier. The second table shows the characteristics for the ARP from the T0%, 5%(S)-curve. Here we see the clear deterioration in the median case, the value being 127.7 in 2020, as compared to the 134.8 in the beta-only case above. The worst case is by definition again 75.0 at the fifth percentile of outcomes. Under the assumption that there will be zero alpha but 5% uncorrelated volatility coming from active tactical management, we found a new ARP. It is more conservative than the original one as the added volatility adversely affects the worst-case outcomes. This is exactly what we could expect given the theoretical framework we set up earlier. But investigating the effect of active management also means having some expectation of positive alpha. Using our new ARP on T0%, 5%(S) as the starting point, we can calculate how much alpha is needed from active management to make up for the extra volatility we have introduced. By adding positive alpha to the ARP of T0, 5%(S) we can determine the characteristics of each of these combinations. The yellow curve in Figure 17 represents these portfolios, with increasing alpha as the curve moves up and to the right. Adding uncorrelated alpha to the ARP ameliorates the worst-case outcome by moving the line upwards in the graph, but it has an even bigger impact on the expected outcome by moving it to the right. We zoom in on the area of interest in the graph to highlight this impact. The second chart in Figure 17 shows a close-up of the region where our ARP on T0%, 5%(S) lies. The impact of adding uncorrelated alpha is fairly dramatic as the yellow curve only spans 0 alpha 0.81%. To return to an expected end capital equal to the unperturbed beta-only case (as shown by the purple circle ARP in the graph), the required alpha is 0.42%. The corresponding portfolio T0.42%, 5%(S0) is depicted by the yellow diamond. With the tracking error of 5% that we have introduced, this corresponds to a required information ratio of 0.08. To return to the same information ratio as the one originally attained by ARPS (purple circle), active management needs to generate an alpha of 0.81%, which amounts to a modest information ratio of 0.16.4 The corresponding portfolio T0.81%, 5%(S0) is depicted by the yellow square. These results are summarized in the table in Figure 17. The characteristics for these two portfolios are presented in the tables in Figure 18. THIS MATERIAL IS PREPARED FOR INSTITUTIONAL INVESTOR USE ONLY November 2006 Investment Management Journal • 53 FIG 17: INVESTMENT STRATEGY RISK ANALYSIS: ADDING ALPHA TO THE OVERLAY 100 Alpha = 0.42%, IR = 0.08 MRP Tα, 5%(S 0)-curve 5% worst-case capital 80 Acceptable risk level = 75 Alpha = 0.81%, IR = 0.16 60 ARP ARP MRP 40 Trend T0%, 5%(S)-curve 20 120 130 140 Expected capital 150 160 170 S-curve Trend CLOSE- UP 100 S-curve 95 Trend 5% worst-case capital Tα, 5%(S 0) - curve 90 85 80 75 MRP Trend T0%, 5%(S) - curve ARP Alpha = 0.42 %, IR = 0.08 MRP Acceptable risk level = 75 ARP 130 135 140 145 150 155 Alpha = 0.81% , IR = 0.16 70 120 125 Expected capital P=T , 5%(S0) Alpha 0.42% 0.81% Information ratio 0.08 0.16 Objective-criterion: E(Capital P) = E(Capital ARPS) Information ratio-criterion: E(P-S0)/ (P-S0) = E(ARPS-S0)/ (ARPS-S0) Source: Morgan Stanley Investment Management Research 54 • Investment Management Journal November 2006 THIS MATERIAL IS PREPARED FOR INSTITUTIONAL INVESTOR USE ONLY FIG 18: CAPITAL DEVELOPMENT PERCENTILES Capital T0.42%, 5% (S0) Percentiles 5% 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017 2018 2019 2020 100.0 90.6 88.2 87.5 86.6 84.1 84.0 82.6 84.2 84.0 83.0 81.1 79.6 81.2 82.9 81.4 25% 100.0 97.2 97.8 98.2 98.9 99.0 100.1 101.3 102.7 104.4 105.0 105.5 106.4 108.7 109.5 110.5 50% 100.0 102.3 104.4 106.8 108.8 111.1 113.1 115.9 118.2 120.3 123.0 124.7 127.6 130.7 133.2 136.8 75% 100.0 107.1 112.1 116.2 119.7 123.4 128.0 132.8 137.9 140.6 144.5 148.5 152.7 155.7 159.9 166.1 95% 100.0 115.0 122.9 130.3 138.4 146.3 152.0 161.1 166.3 176.8 182.0 189.4 199.6 205.5 211.1 224.1 μ 100.0 102.4 105.1 107.4 110.1 112.2 114.9 118.1 121.3 123.6 126.3 128.8 132.0 135.2 138.2 141.7 0.0 7.3 10.6 13.1 16.0 18.5 21.1 24.0 25.9 28.3 30.4 32.7 35.4 37.8 40.2 43.0 0.0 0.1 0.2 0.2 0.3 0.4 0.5 0.6 0.6 0.6 0.6 0.6 0.7 0.7 0.7 0.7 Capital T0.81%, 5%(S0) Percentiles 5% 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017 2018 2019 2020 100.0 91.0 89.0 88.6 88.1 85.9 86.2 85.2 87.3 87.3 86.8 85.1 84.2 86.4 88.6 87.5 25% 100.0 97.6 98.6 99.4 100.5 101.1 102.6 104.4 106.2 108.5 109.5 110.5 111.9 114.7 116.4 117.8 50% 100.0 102.7 105.2 108.1 110.6 113.3 115.8 119.1 122.0 124.8 128.3 130.5 134.1 137.9 141.0 145.6 75% 100.0 107.5 112.9 117.5 121.6 125.7 131.0 136.5 142.3 145.6 150.2 155.1 160.1 163.9 168.9 176.4 95% 100.0 115.4 123.8 131.7 140.4 149.0 155.4 165.3 171.5 182.8 189.1 197.5 208.9 215.9 222.7 237.4 μ 100.0 102.7 105.9 108.6 111.8 114.5 117.7 121.4 125.2 128.2 131.5 134.7 138.6 142.6 146.4 150.8 0.0 7.3 10.6 13.2 16.2 18.8 21.5 24.5 26.5 29.1 31.4 33.9 36.9 39.5 42.1 45.2 0.0 0.1 0.2 0.2 0.3 0.4 0.5 0.6 0.6 0.6 0.6 0.6 0.7 0.7 0.7 0.7 THIS MATERIAL IS PREPARED FOR INSTITUTIONAL INVESTOR USE ONLY November 2006 Investment Management Journal • 55 FIG 19: INVESTMENT STRATEGY RISK ANALYSIS: DETERMINING THE MAXIMUM RISK BUDGET 100 Alpha = 0.74%, IR = 0.13 MRP Alpha = 0.89%, IR = 0.16 5% worst-case capital 80 Acceptable risk level = 75 MRP 60 ARP 40 Tα, 5.5%(S 0)-curve Trend T0%, 5.5%(S)-curve 20 120 130 140 Expected capital Source: Morgan Stanley Investment Management Research S-curve Trend 150 160 170 A comparison between T0.42%, 5%(S0) and ARPS shows that, although they score equally with respect to the objective (expected capital in 2020), T0.42%, 5%(S0) has superior characteristics on the downside and behaves slightly worse on the upside. In the above analysis we chose to work with a fixed risk budget of 5%. Obviously the size of the risk budget can be adjusted to fit prevailing preferences and the belief in uncorrelated alpha and achievable information ratios. One might also be interested in the maximum risk budget that is appropriate in a specific case. In order to investigate, we start again with the S-curve and add an additional tactical alpha-overlay on top (Figure 19). Let us assume zero alpha, % volatility and zero correlation for this tactical overlay, and label the resulting curve T0%, %(S). We increase up to the point max, where the resulting blue curve T0%, max%(S) is tangential to the acceptable risk level. By construction, the corresponding ARP on T0%, max%(S) is equal to the MRP for T0%, max%(S). 56 • Investment Management Journal November 2006 THIS MATERIAL IS PREPARED FOR INSTITUTIONAL INVESTOR USE ONLY FIG 20: COMPARISON OF VARIOUS LEVELS OF ALPHA AND RISK BUDGET Strategy (S0) weights Risk budget ( ) 0.0% 3.0% 3.5% 4.0% 4.5% 5.0% max Capital T0%, (S0 ) 5% worst case 75 75 75 75 75 75 75 *Exp 5% best case 237 230 226 223 216 211 192 Capital T0.5%, (S0) 5% worst case Exp 5% best case Capital T1%, (S0) 5% worst case Exp 5% best case Global European equities bonds 56.6% 60.6% 62.6% 65.7% 70.7% 75.8% 43.4% 39.4% 37.4% 34.3% 29.3% 24.2% 10.1% 142 140 139 137 135 133 126 82 82 84 83 83 83 151 150 149 146 144 137 248 243 240 233 227 206 90 90 92 91 90 91 164 163 161 158 155 148 268 262 258 251 244 222 = 5.5% 89.9% * Exp = Expected Note that the = 0 top row in the above table by definition represents ARPS. Source: Morgan Stanley Investment Management Research In our case it turns out that max = 5.5%. This means that the maximum risk budget is 5.5%. This level of active management roughly corresponds with the original MRPS (purple triangle) as the beta-only strategy. Adding positive alpha, similar to our previous analysis, leads to the yellow curve showing the added value that can be achieved through this. Whether active implementation in combination with a more conservative strategy is preferable to the more aggressive beta-only strategy ARPS depends on individual preference. In the table shown in Figure 20 we list the 5% worst-case behavior, as well as the expected capital outcome and the 5% best case for several levels of risk budget and assumed alpha. For the same range of risk budgets we also list the behavior of the actively managed portfolios in case we choose alpha such that the expected capital is equal to that of ARPS (Figure 21). THIS MATERIAL IS PREPARED FOR INSTITUTIONAL INVESTOR USE ONLY November 2006 Investment Management Journal • 57 FIG 21: ALPHA- CRITERION: E(CAPITAL T , (S 0 )) = E(CAPITAL ARP S ) Strategy (S0) Weights Risk budget ( ) European bonds 56.6% 60.6% 62.6% 65.7% 70.7% 75.8% = 5.5% 89.9% Global equities 43.4% 39.4% 37.4% 34.3% 29.3% 24.2% 10.1% 5% worst case 75 76 77 80 80 81 87 Capital T Expected , (S0) 5% best case 237 233 230 230 227 224 214 Required alpha 0.00% 0.09% 0.14% 0.19% 0.31% 0.42% 0.74% 0.0% 3.0% 3.5% 4.0% 4.5% 5.0% max 142 142 142 142 142 142 142 Source: Morgan Stanley Investment Management CONCLUDING REMARKS With this we conclude our illustrative analysis of the ALM and risk budgeting choices facing this particular foundation. As one can see, the ALM driven results fit nicely with the theoretical setting originally described. Since ALM techniques can be used for much more complicated situations as well, similar analyses can also be undertaken in these cases. In doing so it is possible to relate levels of active management to the choice of the investment strategy in general, thus integrating tactical and strategic considerations in a truly consistent way. 58 • Investment Management Journal November 2006 THIS MATERIAL IS PREPARED FOR INSTITUTIONAL INVESTOR USE ONLY 1 Alpha by definition is uncorrelated with beta, and can come from a multitude of idiosyncratic sources. This makes the modeling of alpha much trickier and less reliable than that of beta. The circular shape is used here only for illustrative purposes. In any given practical case the shape will be different. Also note that we are only modeling risky asset categories with well-defined beta. Both equities and bonds possess fundamental relationships to the macroeconomic environment that makes the modeling of their betas feasible. Modeling alpha (which is often manager specific) is much harder. Our methodology for determining the risk budget circumvents this problem by adding the alpha separately. The information ratio relative to the originating strategy S0, from which the ARP on T0%, 5%(S) was derived. 2 3 4 THIS MATERIAL IS PREPARED FOR INSTITUTIONAL INVESTOR USE ONLY November 2006 Investment Management Journal • 59 About the authors Michael Peskin Managing Director Michael is head of the Global Pension Solutions Group. This group is responsible for advising the senior financial management of corporations, pension sponsors, insurers, banks and other institutions on financial strategies and solutions for addressing their asset-liability (enterprise risk) issues. Michael joined Morgan Stanley in 1988 from First Boston, where he developed and marketed pension finance products. Prior to that, Michael was a consulting actuary/ account executive at Buck Consultants. Michael has more than 30 years of retirement finance experience and has spoken and written extensively on a wide variety of asset-liability and other finance issues. He is an Associate of the Society of Actuaries and the Institute of Actuaries, a Fellow of the Conference of Consulting Actuaries and a member of the American Academy of Actuaries. Chad Hueffmeier, CFA Vice President Chad supports our Global Pension Solutions Group, consults on how to efficiently manage pension risk and develops assetliability management (ALM) models for clients. He joined Morgan Stanley in 2006 and has seven years of pension consulting and ALM experience. Prior to joining the firm, he was a pension actuary at Towers Perrin, Mercer Human Resource Consulting and Hewitt Associates. Chad received a B.S. in actuarial sciences from Maryville University of St. Louis. He holds the Chartered Financial Analyst designation, is a Fellow of the Society of Actuaries and an Enrolled Actuary. Chad is a member of the CFA Institute, Society of Actuaries and the American Academy of Actuaries. Federico Kaune, Ph.D. Executive Director Federico is a member of the Fixed Income team. He joined Morgan Stanley in 2002 and has nine years of investment experience. Prior to joining the firm, he was a senior vice president and senior economist at Goldman Sachs. Previously, he was an economist at the International Monetary Fund and a lecturer at the University of Chicago. Federico received a B.A. in economics from Universidad del Pacifico, and an M.A. and a Ph.D. in economics from the University of Chicago. He is a member of the American Economic Association. Eric Baurmeister, CFA Executive Director Eric is a member of the Fixed Income team. He joined Morgan Stanley in 1997 and has 13 years of investment experience. Prior to joining the firm, he was an associate at Citicorp and a portfolio manager at MIMCO. Eric received a B.A. in economics and government from Cornell University and holds the Chartered Financial Analyst designation. He is a member of the New York Society of Security Analysts and the JPMorgan EMBI Index Advisory Committee. Martin L. Leibowitz, Ph.D. Managing Director Martin is a member of Morgan Stanley Equity Research’s Global Strategy team. Over the past two years, he and his associates have produced a series of studies on such topics as beta-based asset allocation, the integration of active and passive alphas, and the need for greater fluidity in policy portfolios. Prior to joining Morgan Stanley, Martin was vice chairman and chief investment officer of TIAA-CREF, with responsibility for the management of over $300 billion in equity, fixed income, and real estate assets. Previously, he had a 26-year association with Salomon Brothers, where he became director of global research, covering both fixed income and equities, and was a member of that firm’s Executive Committee. Martin received an A.B. and an M.S. from the University of Chicago and a Ph.D. in mathematics from the Courant Institute of New York University. He has written over 150 articles on various financial and investment analysis topics, and has been the most frequent author published in both the Financial Analysts Journal as well as the Journal of Portfolio Management. Martin has also published four books, Investing (1992), Return Targets and Shortfall Risks (1996), Franchise Value (2004) and a revised edition of his study on bond investment, Inside the Yield Book (2004). The first edition of Inside the Yield Book was published in 1972, went through 21 reprintings, and remains a standard in the field. 60 • Investment Management Journal November 2006 THIS MATERIAL IS PREPARED FOR INSTITUTIONAL INVESTOR USE ONLY Anthony Bova, CFA Associate Anthony is an associate with Morgan Stanley Equity Research’s Global Strategy team, focusing on institutional portfolio strategy. Prior to his current role, he spent three years covering commodity chemicals at Morgan Stanley. Anthony received a B.S. in economics from Duke University and holds the CFA designation. Jack Coates, Ph.D., CFA Managing Director Jack is head of Morgan Stanley Alternative Investment Partners’ Portable Alpha team, focusing on the design and execution of comprehensive fund architectures. Jack is one of the founding principals of Morgan Stanley Alternative Investment Partners and served as co-head of the business until 2005. He has 21 years of relevant industry experience. Before joining the firm, Jack was a vice president of Weyerhaeuser Company and a managing director of Weyerhaeuser Company’s Pension Fund Investment Group, where he headed the group and pioneered a state-of-the-art alternative investing program. Jack received a B.A.E. and an M.S.A.E. in aerospace engineering and was a National Defense Education Act Doctoral Fellow at the Georgia Institute of Technology. He received an M.B.A. as a Wharton Fellow from the Wharton School at the University of Pennsylvania and a Ph.D. from the University of Washington. Jack holds the Chartered Financial Analyst designation. Mark Baumgartner, Ph.D., CFA Executive Director Mark is an executive director of the Portable Alpha team at Morgan Stanley Alternative Investment Partners, focusing on the design and execution of comprehensive fund architectures including overlays. He joined Morgan Stanley in 2006 and has nine years of relevant industry experience. Prior to joining the firm, Mark co-managed a global portfolio of equities at Quantal Asset Management, a quantitative long-short market neutral hedge fund, and served as vice president and portfolio manager at Strategy Capital, a fundamental long-short equity hedge fund. He also spent seven years as a management consultant, most recently with The Boston Consulting Group, where he specialized in new venture creation and strategy development. Mark received a B.S.E. in aerospace engineering with high honors from the University of Florida. He also received an M.S.E. and Ph.D. in aerospace engineering and a certificate in public policy, all from Princeton University. Mark holds the Chartered Financial Analyst designation. Jan Baars, Ph.D. Vice President Jan is a member of the Global Tactical Asset Allocation portfolio management team and is responsible for quantitative risk management techniques, asset-liability management and asset allocation. He joined Morgan Stanley in 1999 and has 13 years of industry experience. Prior to joining the firm, he was responsible for the development of disability products and for research on and implementation of reserving models for non-life insurance products at Zurich Insurance. He has been a professor at the Free University in Amsterdam on the faculty of mathematics and computer science, as well as at York University, Toronto in the department of mathematics and statistics. Jan received an M.Sc. in mathematics from the Free University and a Ph.D. in mathematics from the University of Amsterdam. Jan is an actuary and a member of the Dutch Actuarial Society. Petr Kocourek Executive Director Petr is a member of the Global Tactical Asset Allocation portfolio management team and is responsible for quantitative techniques and asset allocation. He joined Morgan Stanley in 1997 and has nine years of investment experience. Prior to joining the firm, he worked as a freelance scientific programmer at Delft University of Technology and before that at the Netherlands Organization for Applied Scientific Research (the TNO). Petr received a B.S. in physics from Delft University of Technology. Epco van der Lende, Ph.D. Executive Director Epco is a member of the Global Tactical Asset Allocation portfolio management team and is responsible for quantitative risk management techniques, asset-liability management and asset allocation. He joined Morgan Stanley in 1996 and has 14 years of investment experience. Prior to joining the firm, he was head of the Actuarial Research department at PVF, a large Dutch pension fund management company. Epco received a Ph.D. in mathematics from the University of Amsterdam. Please direct any comments or questions about the Morgan Stanley Investment Management Journal to: Elizabeth Ladhams Managing Editor elizabeth.crowley@morganstanley.com Morgan Stanley Investment Management (“MSIM”) is the asset management division of Morgan Stanley. Assets are managed by teams representing different MSIM legal entities with offices in New York, Philadelphia, Houston, Chicago, London, Amsterdam, Singapore, Tokyo and Mumbai. © 2006 Morgan Stanley IS06-00831I-T08/06 MS JRNL 11/06 MO954