Rethinking risk management in portfolio construction

Deutsche Bank Group DB Advisors Rethinking risk management in portfolio construction Conventional risk management processes often fail to address the risks involved in portfolio design and construction. Stephen Johnson Global Head of Fixed Income Investment Process Management explains how to build an effective risk framework around this key stage of the investment process The governing objective of risk management is to remove or limit exposures that are likely to result in unacceptable outcomes. In this article, we describe how to achieve that objective more effectively in the portfolio / product design and construction stage of the investment process. [In the previous paper in this series 1, we focused on risk management in research and signaling.] Weaknesses in conventional approaches Like research and signaling, portfolio design and construction can contribute to and occasionally drive outcomes that are unacceptable. It is therefore critical that the risk management framework explicitly addresses this stage of the investment process. The conventional risk framework is substantially incomplete In many investment firms, risk management is conducted primarily by portfolio managers. Using daily risk reports which typically detail total portfolio risk and how each position contributes to the total the portfolio manager adjusts exposures in line with risk allowances. 1 Managing risk in the research process, March 2010; see www.dbadvisors.com Larger firms often also have an independent risk management group, tasked with monitoring and reporting on the risk activities of the portfolio managers. These two elements constitute the full scope of a conventional risk management process. We believe this framework is substantially incomplete. For one thing, while it is somewhat effective in managing immediate portfolio risk relative to client objectives and risk allowances, it does not monitor risk-taking practices. To illustrate why this matters, consider the following scenarios: A portfolio manager allocates a significant portfolio weight to a credit that is not favored by the research analyst A portfolio manager regularly reduces their total credit position by selling the largest positions, as this results in the fewest transactions A portfolio manager facing a large loss in a position doubles its weight because the current price is more attractive than the purchase price A portfolio manager only has the time and expertise to use a limited number of the analysts ideas, resulting in a narrowly invested portfolio All of these practices should be monitored and evaluated although they are potential causes www.dbadvisors.com DB Advisors 1

of bad outcomes, each could be carried out in a manner that still satisfies the risk allowances of the client. Beyond the conventional approach In broad terms, the following improvements can often be made to risk management as it relates to portfolio design and construction. Identify possible winners and losers There is rarely sufficient clarity as to which risk exposures are potential winners and which are likely losers. Conventional risk management therefore indiscriminately changes risk exposures without regard to the possible profit of the eliminated position. To the extent possible, the risk management process should assist the portfolio manager to cut positions that, other things equal, are less likely to generate profit. Monitor risk-taking practices, as well as risks Often, risk management focuses on the immediate risk picture, identifying only whether a portfolio is in compliance and how each investment contributes to total risk. For an investment firm to get better at managing risk, risk-taking practices must be continually measured and studied to flag ineffective procedures. Develop a clear basis for intervention When risk-related problems are identified, the basis for making portfolio changes is often unclear and poorly defined. The risk management process should provide a clear basis for making changes to risk-taking practices. In too many instances, only the portfolio manager is authorized to do so. To achieve these improvements, we believe that the risk management process should include additional steps beyond those applied in the conventional approach. These steps, which apply to every stage of the investment process, can broadly be grouped into the following areas: Defining objectives Establishing investment disciplines Creating a process for control and quality management Structuring the organization to support risk management They are explained fully below. Define objectives The first step in building an effective risk framework is to define explicitly all of the objectives of risk management. Example risk management objectives for the portfolio design and construction stage are: Invest signals in a timely manner when sent by analyst Build signal exposures in a risk efficient manner, consistent with client guidelines Manage the portfolio according to client and process authorizations Manage according to process disciplines Improve disciplines and effectiveness Control measures which are used to identify undesirable practices should also have defined objectives, such as: Control for excess portfolio exposures and risks Control for slow execution Control for poor implementation Establishing investment disciplines Investment discipline is the term we use to describe the rules that regulate working practices. In product design and portfolio construction, example disciplines may include: Create exposures using a broad and balanced range of alpha sources Create exposures with reference to the expected skill level of the analyst Ensure positions abide by targets and tolerances Disciplines are important to risk management for several reasons. First, by ensuring that standardized procedures are followed, they support the goal of making each step of the research process transparent and measurable and therefore open to analysis by the management team. Data generated via disciplines also enables the quality management and control mechanism (see below) to work effectively. Finally, disciplines can be used to help constrain or affect investment risk. For example, we employ disciplines in our product design process to www.dbadvisors.com DB Advisors 2

reinforce the principle that product alpha source usage should be broad and balanced. Furthermore, we evaluate the approach to balancing alpha source use, as different methodologies can create different expectations regarding product outcomes. Consider the two product designs below, the first of which is balanced from an expected alpha contribution standpoint, while the second is balanced from a risk contribution standpoint. R Product design 1 Balanced by expected alpha contribution Alpha source Product design 2 Balanced by expected risk contribution Alpha source Active weight Active weight Alpha (bp) Alpha (bp) Risk (bp) 1 148.00 23.48 118.58 2 157.70 10.04 117.09 3 79.70 33.27 118.64 4 202.10 14.18 118.46 5 249.80 14.11 118.55 6 155.10 11.98 117.97 7 70.10 38.88 118.64 8 66.40 20.59 118.18 9 139.90 16.38 118.81 Alpha target total 180.00 Information ratio 0.46 Risk (bp) 1 131.50 20.61 105.28 2 332.70 20.94 248.12 3 48.90 20.39 72.75 4 266.20 20.10 156.21 5 369.80 20.13 175.56 6 238.30 18.27 181.70 7 36.60 20.47 61.94 8 66.40 20.00 118.58 9 169.40 20.39 143.47 Alpha target total 180.00 Information ratio 0.36 Both products expect to produce roughly 180 basis points of annual alpha. Importantly, the product that builds the alpha expectation through exposures designed by weighting information ratios provides a much better information ratio for about the same expected alpha. This example illustrates how disciplines in product design can help to make sure that products are risk efficient. Creating controls and alerts Controls provide a mechanism for immediately adjusting positions or processes when risk levels approach unacceptable levels. They also give the management team a sound basis for considering intervening in the investment process to adjust practices, where appropriate. For controls to be effective, systems must capture sufficient data and generate control reporting. Control reporting can be thought of as exception reports designed to flag problems. Reports can take different forms, but usually they relate a process measure to targets or tolerances. The sample control report below measures rate movements and warns when the actual movement suggests changes to product return, assuming constant investment skill. The report can also be useful in flagging changes to the volatility environment changes that could be quite harmful when combined with ineffective decision making. Sample control report Shaded cells indicate actual performance significantly different from forecasts Forecast target Actual target Alpha source alpha (bp) alpha (bp) German 10yr rates 9.1 2.75 US 10yr swap spread 2.3 2.1 Europe 10yr swap 10.4 16.3 spread We also advocate a system of alerts, which can be thought of as user-defined controls and decision aids. The idea is to provide the investment team with tools to create their own alerts to help them change risk positions and practices. An alert system is a useful testing ground for potential new investment disciplines or controls if alerts prove effective, they can www.dbadvisors.com DB Advisors 3

be considered for wider use. However, without a fairly high level of transparency in the investment process, the control and alert mechanism cannot function properly. Managing investment quality Quality management is a technique more commonly used in engineering circles than investment firms. It involves estimating outcomes and then comparing actual outcomes against the estimates, with the goal of continually refining a process so that it produces results that are consistently closer to expectations. To create a functioning quality management system, the investment firm must first estimate the expected outputs of each step in the investment process: including the output from each research signal; the portfolio exposure to the signal; and the alpha impact of the signal on portfolio returns. In portfolio (or product) quality management, the estimated contribution of each alpha signal to the portfolio should factor in expectations of signal effectiveness, market volatility, and implementation. The next step is to measure actual outcomes so that the quality management team can evaluate them versus expectations. Below is a sample quality management report. Alpha source Target alpha Sample quality report Actual alpha Influence of skill market position A 10 10-2 0 2 B 8 6 0-2 0 C 10-3 -8 0-5 D 10 20 6-2 6 Total 38 33-4 -4 3 Armed with this detailed analysis, the quality management team can alter or remove practices that lead to bad outcomes. We believe that quality management is a key driver of improvements in risk management skill. Executed correctly, the quality management team can isolate common forms of investment process failure. Importantly, a properly functioning quality management system should facilitate gradual adjustments to process execution over time, rather than only when a problem becomes critical. Organizational design and tasking Organizational design is perhaps the most common reason why risk management is ineffective. Among many others, typical problems include: Allowing for or requiring selfauthorization or control Assigning control responsibilities to individuals without any executive authority Concentrating process operation and management on too few people, usually the portfolio manager We recommend separating the operation of the investment process i.e., the functions carried out by analysts, traders and portfolio managers from management of the process. This allows the investment team to focus on delivering consistent performance, while the management team concentrates on objectively monitoring the investment process, maintaining controls and quality management mechanisms, and intervening where appropriate to strengthen risktaking practices. A final key aspect of organizational design is to ensure that the various disciplines, control responsibilities and quality management procedures are assigned to an appropriate team member, and aligned with other process steps in a manner that leads to logical execution. Conclusions Portfolio construction and design is a critical stage in the investment process, and one that requires thorough risk management. Too often, risk management approaches only address the risk currently in the portfolio, without adequately considering how and why that risk was taken on in the first place, nor how to improve the risktaking skills of the investment team. To be effective, risk management in portfolio design and construction requires much more than standard risk reports and a risk committee. With clearly defined objectives and supported by robust data and analysis, the risk management framework should shape and continually monitor www.dbadvisors.com DB Advisors 4

the investment process, with the goal not only of managing immediate risk, but also of improving the process and the risk-taking skills of the investment team. commitment from the investment firm to create a culture of risk management that demands excellence in all risk-related decisions. In our view, a methodical and systematic approach to risk management can deliver these improvements. But critical to success is a strong Stephen Johnson is a Managing Director at DB Advisors and Global Head of Institutional Fixed Income Investment Process Management at DB Advisors www.dbadvisors.com DB Advisors 5

DB Advisors is the brand name for the institutional asset management division of Deutsche Asset Management, the asset management arm of Deutsche Bank AG. In the US, Deutsche Asset Management relates to the asset management activities of Deutsche Bank Trust Company Americas, Deutsche Investment Management Americas Inc. and DWS Trust Company; in Canada, Deutsche Asset Management Canada Limited (Deutsche Asset Management Canada Limited is a wholly owned subsidiary of Deutsche Investment Management Americas Inc); in Germany and Luxembourg: DWS Investment GmbH, DWS Investment S.A., DWS Finanz-Service GmbH, Deutsche Asset Management Investmentgesellschaft mbh, and Deutsche Asset Management International GmbH; in Australia, Deutsche Asset Management (Australia) Limited (ABN 63 116 232 154); in Hong Kong, Deutsche Asset Management (Hong Kong) Limited; in Japan, Deutsche Asset Management Limited (Japan); in Singapore, Deutsche Asset Management (Asia) Limited (Company Reg. No. 198701485N) and in the United Kingdom, RREEF Limited, RREEF Global Advisers Limited, and Deutsche Asset Management (UK) Limited; in addition to other regional entities in the Deutsche Bank Group. This material is intended for informational purposes only and it is not intended that it be relied on to make any investment decision. It does not constitute investment advice or a recommendation or an offer or solicitation and is not the basis for any contract to purchase or sell any security or other instrument, or for Deutsche Bank AG and its affiliates to enter into or arrange any type of transaction as a consequence of any information contained herein. Neither Deutsche Bank AG nor any of its affiliates gives any warranty as to the accuracy, reliability or completeness of information which is contained in this document. Except insofar as liability under any statute cannot be excluded, no member of the Deutsche Bank Group, the Issuer or any officer, employee or associate of them accepts any liability (whether arising in contract, in tort or negligence or otherwise) for any error or omission in this document or for any resulting loss or damage whether direct, indirect, consequential or otherwise suffered by the recipient of this document or any other person. The views expressed in this document constitute Deutsche Bank AG or its affiliates judgment at the time of issue and are subject to change. This document is only for professional investors. This document was prepared without regard to the specific objectives, financial situation or needs of any particular person who may receive it. The value of shares/units and their derived income may fall as well as rise. Past performance or any prediction or forecast is not indicative of future results. No further distribution is allowed without prior written consent of the Issuer. The forecasts provided are based upon our opinion of the market as at this date and are subject to change, dependent on future changes in the market. Any prediction, projection or forecast on the economy, stock market, bond market or the economic trends of the markets is not necessarily indicative of the future or likely performance. For Investors in the United Kingdom: Issued in the United Kingdom by Deutsche Asset Management (UK) Limited of One Appold Street, London, EC2A 2UU. Authorised and regulated by the Financial Services Authority. This document is a "non-retail communication" within the meaning of the FSA s Rules and is directed only at persons satisfying the FSA s client categorisation criteria for an eligible counterparty or a professional client. This document is not intended for and should not be relied upon by a retail client. When making an investment decision, potential investors should rely solely on the final documentation relating to the investment or service and not the information contained herein. The investments or services mentioned herein may not be appropriate for all investors and before entering into any transaction you should take steps to ensure that you fully understand the transaction and have made an independent assessment of the appropriateness of the transaction in the light of your own objectives and circumstances, including the possible risks and benefits of entering into such transaction. You should also consider seeking advice from your own advisers in making this assessment. If you decide to enter into a transaction with us you do so in reliance on your own judgment. For Investors in Australia: In Australia, Issued by Deutsche Asset Management (Australia) Limited (ABN 63 116 232 154), holder of an Australian Financial Services License. An investment with Deutsche Asset Management is not a deposit with or any other type of liability of Deutsche Bank AG ARBN 064 165 162, Deutsche Asset Management (Australia) Limited or any other member of the Deutsche Bank AG Group. The capital value of and performance of an investment with Deutsche Asset Management is not guaranteed by Deutsche Bank AG, Deutsche Asset Management (Australia) Limited or any other member of the Deutsche Bank Group. Investments are subject to investment risk, including possible delays in repayment and loss of income and principal invested. For Investors in Hong Kong: Interests in the funds may not be offered or sold in Hong Kong or other jurisdictions, by means of an advertisement, invitation or any other document, other than to Professional Investors or in circumstances that do not constitute an offering to the public. This document is therefore for the use of Professional Investors only and as such, is not approved under the Securities and Futures Ordinance (SFO) or the Companies Ordinance and shall not be distributed to non-professional Investors in Hong Kong or to anyone in any other jurisdiction in which such distribution is not authorised. For the purposes of this statement, a Professional investor is defined under the SFO. For Investors in MENA region: This information has been provided to you by Deutsche Bank AG Dubai (DIFC) branch, an Authorised Firm regulated by the Dubai Financial Services Authority. It is solely directed at Market Counterparties or Professional Clients of Deutsche Bank AG Dubai (DIFC) branch, which meets the regulatory criteria as established by the Dubai Financial Services Authority and may not be delivered to or acted upon by any other person. I-016661-1 (3/10) www.dbadvisors.com DB Advisors 6