Investment Risk Measurement

Document Sample
Investment Risk Measurement Powered By Docstoc
					                                                                     03/2010/#28
          journal
            the	              journal of financial transformation	




          Cass-Capco Institute Paper Series on Risk




Recipient of the APEX Awards for Publication Excellence 2002-2009
Editor
Shahin Shojai, Global Head of Strategic Research, Capco

Advisory Editors
Cornel Bender, Partner, Capco
Christopher Hamilton, Partner, Capco
Nick Jackson, Partner, Capco

Editorial Board
Franklin Allen, Nippon Life Professor of Finance, The Wharton School,
University of Pennsylvania
Joe Anastasio, Partner, Capco
Philippe d’Arvisenet, Group Chief Economist, BNP Paribas
Rudi Bogni, former Chief Executive Officer, UBS Private Banking
Bruno Bonati, Strategic Consultant, Bruno Bonati Consulting
David Clark, NED on the board of financial institutions and a former senior
advisor to the FSA
Géry Daeninck, former CEO, Robeco
Stephen C. Daffron, Global Head, Operations, Institutional Trading & Investment
Banking, Morgan Stanley
Douglas W. Diamond, Merton H. Miller Distinguished Service Professor of Finance,
Graduate School of Business, University of Chicago
Elroy Dimson, BGI Professor of Investment Management, London Business School
Nicholas Economides, Professor of Economics, Leonard N. Stern School of
Business, New York University
José Luis Escrivá, Group Chief Economist, Grupo BBVA
George Feiger, Executive Vice President and Head of Wealth Management,
Zions Bancorporation
Gregorio de Felice, Group Chief Economist, Banca Intesa
Hans Geiger, Professor of Banking, Swiss Banking Institute, University of Zurich
Wilfried Hauck, Chief Executive Officer, Allianz Dresdner Asset Management
International GmbH
Pierre Hillion, de Picciotto Chaired Professor of Alternative Investments and
Shell Professor of Finance, INSEAD
Thomas Kloet, Senior Executive Vice-President & Chief Operating Officer,
Fimat USA, Inc.
Mitchel Lenson, former Group Head of IT and Operations, Deutsche Bank Group
David Lester, Chief Information Officer, The London Stock Exchange
Donald A. Marchand, Professor of Strategy and Information Management,
IMD and Chairman and President of enterpriseIQ ®




Colin Mayer, Peter Moores Dean, Saïd Business School, Oxford University
Robert J. McGrail, Executive Managing Director, Domestic and International Core
Services, and CEO & President, Fixed Income Clearing Corporation
John Owen, CEO, Library House
Steve Perry, Executive Vice President, Visa Europe
Derek Sach, Managing Director, Specialized Lending Services, The Royal Bank
of Scotland
John Taysom, Founder & Joint CEO, The Reuters Greenhouse Fund
Graham Vickery, Head of Information Economy Unit, OECD
Norbert Walter, Group Chief Economist, Deutsche Bank Group
                                 Articles




                                 Best practices
                                 for investment risk
                                 management



         Jennifer Bender
   Vice President, MSCI Barra

            frank nielsen
Executive Director, MSCI Barra




                                 Abstract
                                 A successful investment process requires a risk management
                                 structure that addresses multiple aspects of risk. In this paper, we
                                 lay out a best practices framework that rests on three pillars: risk
                                 measurement, risk monitoring, and risk-adjusted investment man-
                                 agement. All three are critical. Risk measurement means using the
                                 right tools accurately to quantify risk from various perspectives.
                                 Risk monitoring means tracking the output from the tools and flag-
                                 ging anomalies on a regular and timely basis. Risk-adjusted invest-
                                 ment management (RAIM) uses the information from measurement
                                 and monitoring to align the portfolio with expectations and risk
                                 tolerance.




                                                                                                        37
           Best practices for investment risk management




           The last 18 months have brought risk management to the forefront        ■■ Manages risk for normal times but is cognizant of and aims to be
           and highlighted the need for guidance on best practices for inves-         prepared for extreme events.
           tors. Many institutional investors were surprised by the violent mar-
           ket moves during the current crisis. Some have argued that current      We developed this framework with institutional investors and
           risk management practices failed when they were needed most, and        their risk management challenges in mind, but this framework
           with multi-sigma events extending across formerly uncorrelated          can be adapted easily to the requirements of asset managers. In
           asset classes, investors have questioned the very meaning of the        the first section of the paper, we describe a framework that takes
           term ‘well diversified portfolio.’ What does sound risk management      into account the three guiding principles. In the second section of
           mean for plans, foundations, endowments, and other institutional        the paper, we illustrate this framework in more detail and provide
           investors? How should these institutions think about best practices     examples for its implementation.
           in risk management? We start with three guiding principles:
                                                                                   Three pillars for risk management
           1 Risk management is not limited to the risk manager; any-              Risk management has evolved rapidly over the last few decades,
             one involved in the investment process, from the cio to               marked by key developments like the adoption of the 1988 Basel
             the portfolio managers, should be thinking about risk — risk          Accord and significant episodes like the U.S. Savings and Loan crisis,
             management should not be limited to an after-the-fact reporting       the collapse of LTCM, the ‘Dot.com’ bust, and the recent financial
             function but must be woven into the investor’s decision-making        crisis. However, the degree to which various risk methodologies
             process, whether it is the asset allocation decision or the process   and practices have been implemented by institutional investors has
             for hiring managers. Those responsible for asset allocation and       varied. In particular, there remains a wide range in how market par-
             management should be risk managers at heart and consider risk         ticipants (pension plans, endowments, asset managers, hedge funds,
             and return tradeoffs before making investment decisions.              investment banks, etc.) have integrated the risk management func-
           2 if you cannot assess the risk of an asset, maybe you should           tion. Best and Reeves (2008), for example, highlight the divergence
             not invest in it — for those institutions invested in alternative     in risk management practices between buy-side and sell-side institu-
             asset classes, such as private equity and hedge funds, or who         tions, the latter being subject to greater regulatory pressure.
             have exposure to complex instruments, such as derivatives and
             structured products, the risk management requirements have            Our goal is to establish a framework for sound market risk man-
             greatly increased. These investors need a framework for manag-        agement for institutional investors. We rely on three pillars:
             ing risk that far exceeds what was needed for the plain vanilla       risk measurement, monitoring, and management (or risk-adjusted
             stock and bond investing that prevailed only ten years ago. We        investment management, RAIM). Risk measurement refers to the
             argue that one should assess one’s risk management capabilities       tools institutional investors use to measure risk. Risk monitoring
             before making the decision to invest in certain asset types.          focuses on the process of evaluating changes in portfolio risk over
           3 proactive risk management is better than reactive risk                time. RAIM refers to how investors may adjust their portfolios in
             management — being prepared for unlikely events is perhaps            response to expected changes in risk. Robust risk management
             the most important lesson learned from the recent crisis. This        integrates all three areas.
             applies to both market and nonmarket risks such as counter-
             party, operational, leverage, and liquidity. Addressing this issue    The risk manager’s toolkit may include a variety of measures captur-
             transcends the simple use of the output of models and tools. It       ing different views of risk. Figure 1 illustrates one way of categorizing
             requires an institutional mindset that analyzes the global eco-       the suite of tools needed. We distinguish between risk measures for
             nomic outlook, understands the aggregate portfolio exposures
             across asset classes, and is willing to use the model output intel-
             ligently to align the portfolio structure with the plan sponsor’s                 Alpha (active risk)                          Beta (total risk)
             assessment of the risks that may impact the portfolio.                       normal              Extreme                normal                Extreme
                                                                                    Tracking error       Stress testing        Asset class           Stress testing asset
                                                                                                         active bets           volatility/beta       classes
           In this paper, we propose a risk management framework that:
           ■■ Is aligned with the investment objectives and investment hori-        Contribution to      Active contribution   Contributions to      Total contribution
              zon.                                                                  tracking error       to tail risk          total risk            to tail risk

           ■■ Tackles multiple aspects of risk and is not limited to a single       Active exposures
              measure like tracking error or Value at Risk (VaR).                   Benchmark misfits    Maximum active        Sources of return/    Maximum
                                                                                                         drawdown              exposures             drawdown,
           ■■ Measures, monitors, and manages exposures to economic and
                                                                                                                                                     contagion effect
              fundamental drivers of risk and return across asset classes to
              avoid overexposures to any one risk factor.                           Figure 1 – Structure for risk measurement and monitoring




38 – The                 journal of financial transformation
Best practices for investment risk management




normal and extreme times as well as risk measures that relate to                             is available. While largely unpredictable, the impact of such shocks
absolute losses or losses relative to a benchmark. On one hand, insti-                       can be analyzed using stress tests. Below we list a number of stress
tutional investors need to manage the total risk of their investments,                       test categories that investors might employ on a regular basis to
which means protecting themselves from asset-liability deficits,                             assess the immediate impact on the portfolio as well as the change
declines in broad asset classes, and more generally, any losses large                        in impact over time.
enough to make it difficult to meet the investor’s obligations. On the
other hand, institutions need to manage the risk of managers under-                          systemic shock
performing their benchmarks, which involves monitoring the tracking                          ■■ Liquidity shock
error and performance relative to the assigned benchmark.                                    ■■ Leverage shock

To assess future risks, it is essential to measure and monitor risk                          Macro shock
both at the aggregate level and at the factor level. For risk mea-                           ■■ Interest rate shock
surement, most institutional investors measure aggregate portfolio                           ■■ Oil price shock
risk with volatility or tracking error, which rely on individual volatili-
ties and correlations of asset classes and managers. However, while                          Market-wide shock
volatility, tracking error, and correlations capture the overall risk of                     ■■ Market-wide decline in equity prices
the portfolio, they do not distinguish between the sources of risk,
which may include market risk, sector risk, credit risk, and interest                        Targeted shock
rate risk, to name a few. For instance, energy stocks are likely to                          ■■ U.S. value stocks hit
be sensitive to oil prices, and BBB corporate bonds are likely to                            ■■ Japan growth stocks hit
be sensitive to credit spreads. Sources of risk, or factors, reflect
the systematic risks investors are actually rewarded for bearing                             Our discussion of stress testing segues naturally into the problem
and are often obscured at the asset class level [Kneafsey (2009)].                           of managing tail risk, or the risk of some rare event occurring.
Institutional investors can decompose portfolio risk using a factor                          Whereas stress tests do not address the likelihood of extreme
model to understand how much return and risk from different asset                            shocks occurring, other methods for analyzing tail risk do. This
classes or managers resulted from prescribed factor exposures in                             recent period of turmoil has acutely highlighted both the impor-
the past1, or how much risk to expect going forward.2                                        tance of managing tail risk and the inadequacy of generic tail risk
                                                                                             measures, such as parametric VaR.
Risk monitoring enables institutions to monitor changes in the
sources of risk on a regular and timely basis. For instance, many                            While the simplest measure of parametric VaR assumes that
well diversified U.S. plans saw a growing exposure to financial sec-                         returns are normally distributed, more sophisticated methods
tor, housing, and credit risk from 2005-2006. While risk managers                            for calculating VaR do not. These span a wide range of modeling
may not have foreseen a looming correction, the ability to monitor                           choices that may rely on parametrically or non-parametrically
these exposures would have at least alerted them to the risks in the                         specified non-normal distributions, or Extreme Value Theory. For
event of a correction.                                                                       a more detailed discussion on the latter, we refer to Barbieri et al.
                                                                                             (2009) and Goldberg et al. (2009). Other measures of tail risk, such
Portfolio decomposition plays an important role in stress testing.                           as Expected Shortfall (conditional VaR) and Maximum Drawdown,3
Here, the sources of risk are stressed by the risk manager to assess                         seek to capture a different and potentially more relevant facet of
the impact on the portfolio. Stress testing is flexible in enabling risk                     tail risk. In general, turbulent times highlight the need for moni-
managers to gauge the impact of an event on the portfolio. The                               toring appropriate tail risk measures. Such times also call for the
stress scenario might be real or hypothetical, commonplace or rare,                          frequent reporting of exceptional developments, i.e., reporting that
but stress tests are used typically to assess the impact of large and                        highlights unusual changes in risk measures or increases in expo-
rare events. Scenarios can come in different flavors, such as macro                          sure to certain factors.
shocks, market shocks, or factor shocks. The intuition behind stress
testing for market risk can be applied to nonmarket or systemic                              Before we move on to the third pillar, RAIM, it is important to point
risks, such as leverage and liquidity risk. When leverage and liquid-                        out that risk monitoring requires the necessary IT and infrastruc-
ity shocks occur, as in 2008, it may result in unexpected increases                          ture resources for support. First, accurate data is essential, as is
in investment commitments for which no immediate funding source                              sufficient coverage of the assets held in the portfolio. Delays in a



1 Performance attribution, the attribution of realized returns to a set of exposures           assets, both equity and fixed income, together with commodities, hedge funds, and
  times factor returns, can provide valuable insight to risk managers seeking to iden-         currencies, are then combined into a single risk model. This makes it suitable for a
  tify where their investments or managers added value. In addition, it can highlight          wide range of investment purposes, from conducting an in-depth analysis of a single-
  similarities between asset groups or managers’ strategies in a way that is far more          country portfolio to understanding the risk profile of a broad set of international
  informative than looking at historical inter-manager correlations alone.                     investments across several asset classes.                                              39
2 The Barra Integrated Model (BIM) is such a multi-asset class model for forecasting         3 VaR captures the expected loss at some threshold, while Expected Shortfall captures
  the asset- and portfolio-level risk of global multi-asset class portfolios or plans. The     the expected loss once that threshold has been exceeded. Maximum drawdown is
  model begins with a detailed analysis of individual assets from 56 equity markets and        defined as the largest drop from a peak to a bottom in a certain period.
  46 fixed income markets to uncover the factors that drive their risk and return. The
           Best practices for investment risk management




           risk manager’s ability to view changes in holdings, prices, or char-                                                                       Risk measurement         Risk monitoring           RAiM
           acteristics are often caused by infrastructure limitations. In some                                                               Total    Volatility               Monitor sources of        Limit exposure to
           cases, data may not be readily available, or the resources required                                                                                                 volatility                biggest sources of




                                                                                                                                   normal
                                                                                                                                                                                                         volatility
           to collect data from custodians or individual managers may be
                                                                                                                                             Active   Tracking error           Monitor sources of        Limit exposure to
           prohibitively expensive. In addition, hard-to-model assets, such as                                                                                                 tracking error            biggest sources of
           complex derivatives, hedge funds, and private equity, can pose a                                                                                                                              tracking error

           challenge for even the most advanced systems. In sum, institutions                                                                Total    Stress tests/tail risk   Monitor expected          Implement portfolio
                                                                                                                                                      measures                 shortfall of the total    insurance plan
           should consider the costs of implementing the necessary risk man-                                                                                                   plan




                                                                                                                                   Extreme
           agement systems when they decide in which assets to invest.                                                                       Active   Stress tests/tail risk   Monitor changes in        Ask managers to
                                                                                                                                                      measures                 potential active losses   limit exposures to
                                                                                                                                                                               if market declines        certain sources
           One consequence of the current crisis may be that investors                                                                                                         by X%                     of risk
           become more cautious when they choose their investments.
           Warren Buffett, for example, commented at his recent shareholder                                                        Figure 3 – Three pillars of risk management

           meeting on complex calculations used to value purchases: “If you
           need to use a computer or a calculator to make the calculation, you                                                    Specifically, RAIM could be used in the development of overlay
           shouldn’t buy it.” Even though that statement may be extreme, the                                                      strategies that would facilitate certain hedges, such as currency
           point is well taken. The damage that exotic, illiquid, and hard-to-val-                                                hedges, or tail risk insurance.
           ue instruments have triggered over the last 18 months highlighted
           the need to be able to assess the risks of such investments before                                                     As an example, the declines in the broad equity market last year
           money is allocated to them.                                                                                            caused many pension plans to become underfunded. Decision-
                                                                                                                                  makers may decide that their tolerance for losses should be limited
           The third pillar in our framework is RAIM, which puts risk measure-                                                    to a specific percentage. They should then decide whether that limit
           ment and monitoring outputs into action. While risk measurement                                                        should be maintained through a passive hedge or through a trigger
           provides the measures, and risk monitoring ensures that the                                                            mechanism defined by the breach of clearly defined parameters of
           measures are timely and relevant, without the ability to make                                                          a risk measure. Some pension plans started hedging their equity
           adjustments to the portfolio, this information is of limited value                                                     exposure to limit downside risk, though for many it was too late.
           for protecting the investor against losses. RAIM aligns the invest-                                                    One reason why pension plans may not have hedged their market
           ment decision-making process with the risk management function.                                                        exposure more frequently is the cost of hedging. Hedging reduces
           For instance, RAIM might be used to make portfolio adjustments                                                         the performance of the portfolio in up markets, but in periods when
           as either the correlations between assets or managers rise or the                                                      the market declines, hedging limits the downside. Figure 2 illus-
           probability of certain tail risk or disaster scenarios increases. RAIM                                                 trates a successful market hedge that includes a stop-loss plan at a
           could also facilitate the management of risks coming from certain                                                      point when assets drop below a specified level.
           sources of return, or it could aid in better diversifying the portfolio.
                                                                                                                                  All three pillars — risk measurement, risk monitoring, and RAIM — are
                                                                                                                                  indispensable to a complete risk management structure. Figure 3
                                                                                                                                  summarizes the three pillars, illustrated with specific examples.
             60%          Cumulative return                                                                        Portfolio
                                                                                                                   insurance      The Figure uses the same idea we presented before, namely, that
                                                                                                                   takes effect
             40%                                                                                                                  risk measures can be categorized by normal and extreme times and
                                                                                                                                  relative versus absolute investment objectives.
             20%


              0%                                                                                                                  implementing a market risk management framework
                                                                                                                                  In practice, the needs of institutional investors can be wide ranging,
                                                                                                     8
                                                                                           08
                                                       7




                                                                                                                   08
                     06




                                                                                    08
                                              07




                                                                    07



                                                                            8
                                       07




                                                                                                              8




                                                                                                                           9
                                7




                                                               7
                                                     -0




                                                                                                    -0
                              -0




                                                                           -0




                                                                                                                          -0
                                                            t-0




                                                                                                          t-0
                                             n-




                                                                                          n-
                     c-




                                                                    c-




                                                                                                                   c-
                                     r-




                                                                                   r-
                                                      g




                                                                                                   g
                               b




                                                                            b




                                                                                                                           b
                                    Ap




                                                                                 Ap
                                                   Au




                                                                                                Au
                    De




                                                           Oc


                                                                   De




                                                                                                         Oc


                                                                                                                  De
                                            Ju




                                                                                         Ju
                            Fe




                                                                         Fe




                                                                                                                        Fe




            -20%
                                                                                                                                  and their ideal measurement, monitoring, and managing capabili-
            -40%
                                                                                                                                  ties will differ. In this section, we illustrate the case of a hypotheti-
                                    Plan bears cost of insurance during
                                    normal markets but benefits from                                                              cal but typical U.S. plan sponsor. Although there may be additional
                                    large unexpected drops due to
            -60%
                                    systemic blow-ups
                                                                                                                                  criteria, the three critical drivers of risk management requirements
            -80%
                                                                                                                                  are as follows:

            -100%                    Uninsured portfolio                        Insured portfolio                                 1 Return requirements — the plan’s liabilities or expected pay-
                                                                                                                                    outs will influence not only the assets in which it invests but
            Figure 2 – Risk-adjusted investment management to protect against downside risk                                         also which benchmarks are used and how much it can lose over



40 – The                            journal of financial transformation
Best practices for investment risk management




  certain periods. The latter, in turn, may drive how much risk it is     tives would be small. A Type-3 plan would invest in a variety of
  willing to take and with how much exposure to certain sources of        alternative asset classes as well as complex instruments but to a
  return/risk it is comfortable.                                          larger extent than Type-2 plans.
2 investment horizon — the plan’s investment horizon, or willing-
  ness to sustain shorter-term shocks, will influence which risk          Currently, the vast majority of pension plans fall into the second
  measures are appropriate and how frequently they need to be             group. We, therefore, consider a hypothetical Type-2 plan for our
  monitored.                                                              illustration of a risk management framework. Its asset allocation
3 complexity of investments — plans that invest in difficult-to-          is as follows: equity (60%) [U.S. (36%), international (24%)], fixed
  value assets with potentially non-normal return distributions           income (U.S.) (25%), alternatives (15%) [real estate (5%), private
  or unusually high exposure to tail events require additional risk       equity (5%), hedge funds (5%)].
  measures, higher monitoring frequencies, and advanced RAIM
  capabilities.                                                           A first critical step is to adopt tools that enable the plan to mea-
                                                                          sure and monitor risk at the source, or factor level, and not just at
These criteria are naturally linked, although the degree of impor-        the aggregate level. The plan should then monitor its exposure to
tance might vary from plan to plan. For instance, return require-         different risk sources on a regular basis — monthly or quarterly at
ments may play the primary role in driving the choice of instru-          the very least. This would occur at the total portfolio level, look-
ments and asset classes for some plans, while they may play a             ing across asset classes. In addition, the plan can require from its
less important role for other plans. For some plans, the investment       managers more detailed information on risk exposures. Many plans
horizon is tied directly to their return requirements, while for oth-     receive only high-level performance summaries focusing on real-
ers, it is more a function of how much they are willing to lose over a    ized returns and tracking error. Requiring estimates of exposures
given period. Regardless, these three criteria determine the guiding      to various sources of risk is a crucial extension. For illustration, we
principles for any plan’s risk management function.                       show how this would fit in the framework we have used so far:


For example, a plan sponsor with reasonable and relatively                Total risk:
infrequent payout obligations, a large surplus, and with limited          ■■ normal periods — the plan can evaluate sources of return and
exposure to alternatives and complex instruments does not need               risk across its overall portfolio using a multi-asset class factor
short-term measures or frequent monitoring. This plan would                  model. Sources of risk can include macroeconomic and market
benefit from focusing on longer-term risk measures. Instead of               factors. An example of the type of analysis that can be done is
setting up a system to calculate 10-day VaR measures, the plan               to look at the performance of the plan’s portfolio in different
could focus on how multiyear regime shifts in different risk fac-            macroeconomic regimes. The plan could then adjust its asset
tors, such as interest rate cycles, may affect the portfolio’s value.        allocation during the next review period.
In contrast, a plan with frequent and significant expected payouts,       ■■ Extreme events — using the sources of risk for the overall port-
a limited ability to sustain short-term losses, and with substantial         folio, the plan can shock certain factors or sources of risk, i.e.,
exposure to alternatives and complex instruments would require               those likely to suffer in the event of a market dislocation, includ-
a wide variety of risk measures, frequent risk monitoring, and a             ing a systemic meltdown or series of external shocks. These
well developed RAIM process. Most plans are likely to fall between           stress tests would enable the plan to evaluate how individual or
these two extremes.                                                          multiple simultaneous shocks impact the overall portfolio (i.e.,
                                                                             tail risk and tail correlations). The plan could then establish an
Another example may help to shed light on these ideas. Below, we             action plan if asset values drop by some absolute or relative (i.e.,
group investors into one of three categories using the third criteria —      to liabilities) amount.
complexity of instruments and asset classes.
                                                                          Active risk:
A Type-1 plan invests in a straightforward allocation to equities and     ■■ normal periods — the plan can ask for reports on their sources
fixed income. Equity and fixed income allocations may be limited             of risk from all equity, fixed income, and alternatives managers,
to the domestic market, and fixed income investments are mostly              or the plan can estimate them internally. Sources of active risk
concentrated in government bonds and AAA-rated corporate. A                  should be detailed and focused on the specific risk and return
Type-2 plan may invest in equities globally, including emerging              drivers of the manager’s investment strategy. For instance,
markets. Fixed income investments may include high yield bonds               analyzing a value, small cap equity manager’s tracking error will
and mortgage-backed securities, and the plan may also invest in              focus on the active bets relative to the agreed upon benchmark,
alternatives and complex derivatives. However, as a percentage of            ideally a small cap value benchmark like the MSCI Small Cap
the plan’s total value, the investments in alternatives and deriva-          Value Index. Measuring and monitoring will focus on questions



                                                                                                                                                    41
           Best practices for investment risk management




              of active bets relative to this benchmark, for example, does the
              manager’s portfolio have a consistent small cap value bias or did                               senior investment /
              the portfolio move towards growth-oriented companies over the                                     risk committee

              last few years when value stocks underperformed? The active
              risk analysis should ensure that the hired manager is following                                                              • Initial asset/manager allocation
              his or her mandate and is not deviating from the agreed upon                                       Risk manager              • Standard risk reports
                                                                                                                                           • Red flags/exceptions reporting
              guidelines.
           ■■ Extreme events — the plan may want to stress test the impact
              of the joint underperformance of a number of active strategies
              that historically have been uncorrelated. For example, during                 Equity                Alternatives            fixed income
              the quant meltdown in August 2007, a number of return factors
              became suddenly highly correlated, leading to severe negative
              portfolio performance relative to their respective benchmarks.           Figure 4 – Organizational structure for risk management
              Such stress tests enable the plan to evaluate how shocks impact
              an entire group of managers. Other useful measures for rare
              events are tail risk and tail risk correlations of active bets across   Our example focused on a Type-2 plan. For a Type-3 plan, this illus-
              managers. Certain factors or strategies may become highly cor-          tration would also be relevant, but the requirements for measuring,
              related across asset classes or markets during crises (i.e., value      monitoring, and managing different types and sources of risk would
              or momentum across equity markets) and could lead to vastly             be more extensive. For a Type-1 plan, the extent to which it invests
              higher losses relative to their respective benchmarks than esti-        in risk management should depend on its liabilities structure and its
              mated by the tracking error.                                            short-term risk tolerance.


           The above examples illustrate how a model that decomposes risk             Most plans, regardless of their specific characteristics, can take
           along its sources can help institutions evaluate different types of        some basic actions on an organizational or administrative level to
           risk across different dimensions. It can be applied similarly to real-     manage risk. Our hypothetical plan may establish a risk commit-
           ized returns in order to attribute past performance. For our hypo-         tee consisting of the CIO, risk managers, senior portfolio manag-
           thetical Type-2 plan, a suggested set of minimum components for            ers, and legal and compliance officers that meets at least once a
           risk management may include:                                               quarter to discuss the overall economic and financial environment.
           ■■ Aggregate measures of volatility and tracking error across man-         Participants can discuss their concerns regarding systemic risk
              agers and asset classes.                                                issues such as liquidity and leverage, review the results of stress
           ■■ An accurate decomposition of return and risk across asset               tests, and debate whether hedging strategies should be activated to
              classes, utilizing an integrated (across asset classes) multi-factor    address undesired exposures or potential tail risk events. The plan
              risk model.                                                             could also develop a reporting framework where the risk committee
           ■■ A stress testing framework and/or extreme risk measures for             would receive at least monthly reports on unusual developments
              understanding tail risk and tail risk correlations in the portfolio.    identified by the risk manager. Then, if the investment committee
           ■■ An appropriate set of benchmarks.                                       is sufficiently concerned about exposure to a certain segment, it
                                                                                      could ask those managers with large exposures to hedge them or
           The exact measures, monitoring frequencies, and RAIM processes             to eliminate the undesired exposures.
           the plan adopts will depend on its return requirements and expect-
           ed payouts, and its investment horizon and willingness to tolerate         Figure 4 illustrates this type of setup, where the risk manager pre-
           shorter-term losses. For instance, a plan with limited ability to          pares risk reports and recommendations for the risk committee and
           withstand short-term losses may want to build out its ability to           deliverers risk management services and advice to the different
           assess tail risk over different horizons using risk measures such as       asset class managers.
           Expected Shortfall based on Extreme Value Theory [Goldberg et al.
           (2009)], which is more conservative than parametric VaR. These             Finally, the plan may establish minimum risk management require-
           plans may also want to implement extensive stress tests across             ments for external managers. For instance, the external managers
           asset classes and within certain subcategories of investments.             could be required to demonstrate their ability to calculate tracking
           Meanwhile, plans with greater ability to withstand short-term losses       error, VaR, or other measures, as well as how risk management
           may opt for more basic tail risk measures and stress tests.                impacts their portfolio construction.




42 – The                  journal of financial transformation
Best practices for investment risk management




conclusion
Recent events have put into stark relief the inadequacy of the cur-
rent state of risk management. Much has been said about the need
for better risk management and a greater degree of risk aware-
ness in the broader investment community. Risk management is
a dynamic area, and any set of best practices are bound to evolve
over time. Here we set out to clarify some of the principles and
tools that we believe are required for a sound risk management
framework.


Specifically, we lay out a framework that rests on three pillars — risk
measurement, monitoring, and RAIM (or risk-adjusted investment
management). Each of the three domains is critical for risk manage-
ment. Risk measurement means having the right tools to measure
risk accurately from various perspectives. Risk monitoring means
observing the risk measures on a regular and timely basis. RAIM
means using the information from the measurement and monitor-
ing layers intelligently to ensure that the portfolio management
process is aligned with expectations of risk and risk tolerance.
While each pillar encompasses a different aspect of risk manage-
ment, each is indispensable to a strong risk management process.
Moreover, they are interdependent and should be aligned with the
investor’s objectives. Their interconnectedness drives the key con-
ceptual theme — that risk management and the investment process
should be fully integrated.


References
• Barbieri, A., V. Dubikovsky, A. Gladkevich, L. Goldberg, and M. Hayes, 2009, “Central
  limits and financial risk,” MSCI Barra Research Insights
• Best, P., and M. Reeves, 2008, “Risk management in the evolving investment manage-
  ment industry,” Journal of Financial Transformation, 25, 88-90
• Goldberg, L., M. Hayes, J. Menchero, and I. Mitra, 2009, “Extreme risk analysis,” MSCI
  Barra Research Insights
• Kneafsey, K., 2009, “Four demons,” Journal of Financial Transformation, 26, 18-23




                                                                                           43

				
DOCUMENT INFO
Description: Investment Risk Measurement document sample