Risk Management and Performance Evaluation using an Overlay Approach

Transcription

1 Risk Management and Performance Evaluation using an Overlay Approach 1

2 Structure of Desjardins Global Asset Management (DGAM) Global Asset Allocation Group $2.5 billions Financial Engineering Group Software Designing Group Alternative Investment Group $4.5 billions Canadian Equity Group $30 millions Fixed Income and Securities Lending Group $18 billions 40 professionals / 25 administrative support Long-term agreements (research / software development / some aspects due diligence) 2

3 Our less traditional investment products Funds of hedge funds Alternative Strategy Fund Diversified Alternative Startegy Fund Opportunist Alternative Strategy Fund - Momentum Under test Global Tactical Asset Allocation model (GTAA) 2CTAs Long-short equity fund To be launched in June-July 2004 Global Insurance Linked Fund Equity Options Fund ETPP: seven asset classes principal guaranteed activity managed retail product Some current projects High dividend yield low tracking error fund Money market plus strategy 3

4 Some of our objectives and principles Define appropriate and efficient benchmarks for our clients: Out-performing which benchmark? Make the best use of risk budget/capital Simplify the portfolio management process: Allow the managers to concentrate on decisions that matter Make the client and manager aware of the level of risk that is tolerated Provide a better value added measurement context Do not spend 90% of effort on 10% of value added 4

5 The Case of a Life Insurance Company The situation: Very long duration portfolio (> 12) Benchmark was a combination of 9 SM indices + 2 in-house indices (very short and super long) Benchmark was recalibrated on a monthly basis to mimic the cash flow structure of the liabilities of the firm as closely as possible Portfolio had 600 securities and benchmark had 400 The issues: Managers spent significant amount of time tracking the sources of the tracking error in the portfolio The best solution did not accurately reflect the cash flow structure of the liabilities Benchmark was only recalibrated once a month. There were significant issues with large intra-month cash flow transfers in and out of portfolio 5

6 The Case of a Life Insurance Company Matched portfolio (allocation committee) Canada (--) % Provinces and Municipalities (--) % Overlay portfolio (actively managed) Transactions: Duration Convexity Credit Illiquid Credit (--) % Liquid Credit (--) % Short-term arbitrage Volatility Credit Committee Risk Budget (Value at Risk) 6

7 Matched portfolio Tight investment policies (no deviation beyond 1% tolerated) per asset group Specific corporate benchmark Cash flow matched using an actuarial approach that seeks to minimized the use of C3 reserves: Portfolio can be adjusted any time to meet new cash flow requirements Allocation committee meets semi-annually to review the parameters of the portfolio 7

8 Overlay portfolio Valeur à risque - profit/perte quotidien du 2 septembre au 31 décembre 2003 $ $ $0 -$ $ profit/perte valeur 5ième centile valeur 95ième centile Based on a CAN $35 millions yearly risk budget Manager compensation based on specific levels of value creation to usage of risk (information ratio) 8

9 Information ratio Financial Analyst Journal Actions Grosse capitalisation Actions Valeurs Actions Croissance Actions Petite Capitalisation Obligations Gestion sectorielle 1 er quartile Médiane Journal of Portfolio Management Actions Grosse capitalisation Actions Petite Capitalisation Obligations US Obligations Internationale 1 er quartile Médiane US performances not easily replicable in a Canadian context 9

10 Ex ante Attribution Analysis Value added IV Quarter Cumulative performance (4 months) Investment strategy $ p.b. $ p.b. Duration , ,3 Convexity , ,2 Credit , ,0 Arbirage , ,0 Volatility , ,7 Total , ,2 Attribution analysis is simplified by the fact that managers must specify the intent of the trade at the time it is executed Managers' views rarely require more than 20 positions to be expressed 10

11 The Case of our Canadian Equity Team DGAM's mandate is not to manage all asset classes internally The mandate is to advise on the management of global portfolios and use the best expertise available (internally / externally) We believe we can manage Canadian equity Internal team We do not believe we can manage global equity and even some US equity segment as efficiently Search for the best external managers Internal team that we acquired was built on a sector expertise and operated much like a hedge fund (in-depth fundamental research / ranking process / etc. / constant revaluation of ranking / etc.) 11

12 Management of Canadian Equities Indexed equity (derivatives) + Management team will only trade sectors in which they believe to have a strong expertise Sector base long-short dollar neutral equity portfolio: -Technology -Resources -Retail -etc Calibration of tracking error specified by client / varies over time 12

13 Benchmark construction Global benchmark: Fixed income (credit / treasury) Equities Hedge funds arbitrage strategies Hedge funds momentum strategies Real estate Uncorrelated Insurance linked risks Timber (under consideration) Currencies Specific benchmark for each asset class 90% effort is asset classes allocation - 10% securities selection Security selection effort Where value added is significant Transferred externally whenever relevant 13

14 Objectives in benchmark construction Maximize expected return: Adjusted for regulatory capital or insurance related buffer Adjusted for downside risk tolerance over specific horizons Adjusted for tolerance to tracking error against the competition: Analyzing the competitors' portfolios Client must determine the level of tracking error he is willing to tolerate against the competitors' portfolios (another application of VAR) Actual management is influenced by views in relation to benchmark and by positioning of the competitors (whenever information is available) ETPP: The only constraint was insurance related buffer 14

15 Benchmarks / management process / tracking issues (Hedge funds) No good benchmark available: Date is unreliable (data biases) Portfolio mandates are too specific. performance measurement is unreliable Currently: sell volatility targets (how much volatility do you want to buy (annual tracking error: 3% -4% -5% ) Building benchmark that is investment policy specific and correct for data biases: Building a data base Adjusting for policy (strategy, diversification of fund per strategy, etc) Sampling 10,000 portfolios Position your performance within distribution of monthly returns (establishing quartile positioning) Next step: risk-adjusted distribution Most clients overlay hedge funds on portfolio of fixed income having maturity ranging from 1 month to 3 years 15

16 Benchmarks / management process / tracking issues (Equities) Use of 9 indices (usually) Optimal level of currency hedging integrated to benchmark Invested in multi-manager portfolios: Volatility being equal, prefer higher tracking error on smaller allocation to equity than smaller tracking error on larger allocation to equity Targeting index + 4 to 6%/year Minimizing downside risk: Using a high turnover diversified portfolio of short-term call options for (US market) Efficient use of risk capital (up to 8 times less) Managed by a US bases hedge fund managers for DGAM Tracking error is greater when performance range is narrow 16

17 L/S Stock Portfolio Daily Risk Monitoring We monitor the whole portfolio, each managers' "sub-portfolio", all long-short pairs 17 and the main single positions

18 Conclusion We believe in simplifying portfolio management processes (focus) We believe in concentrating efforts where value added is greatest We strongly believe in active management & research We believe in risk-adjusted performance valuation We believe in the efficient use of risk budget and capital 18