Types of Return Attribution
Performance attribution aims to explain the sources of a portfolio's returns relative to a benchmark. The methodology chosen depends on data availability, desired accuracy, and the complexity of the investment strategy. There are three primary types of return attribution:
Definition: Returns-based attribution uses the historical time series of a portfolio's total returns (and often the returns of relevant market factor indices) to infer the sources of performance. In MantaRisk our risk factor model to determine the portfolio's exposure via a sophisticated Lasso regression analysis.
Primary Inputs:
- Portfolio's periodic total returns (e.g., daily, weekly, monthly).
- Periodic returns of selected market indices.
Advantages:
- Useful When Holdings Data is Unavailable or Unreliable: Particularly valuable for analyzing strategies where detailed holdings information is not accessible, frequently changes, or is difficult to obtain in a timely manner (e.g., some hedge funds, private equity, or when analyzing external managers with limited transparency).
- Provides a "Top-Down" View: Can give a good overview of how a portfolio's returns have historically correlated with broad market factors or investment styles.
Disadvantages:
- Less Accurate and Specific: As it infers exposures from returns, it doesn't directly analyze the manager's actual investment decisions (specific security selections or transaction timing).
- Relies on Historical Correlations: Assumes that past relationships between the portfolio's returns and factor returns will continue. This may not hold true, especially if the manager changes strategy.
- Vulnerable to Data Manipulation or "Window Dressing": Because it doesn't look at actual holdings, it can be misled.
Typical Use Cases:
- Evaluating opaque investment strategies (e.g., some hedge funds, funds-of-funds).
- High-level style analysis of asset managers.
- When detailed holdings or transaction data is not available.
Definition: Holdings-based attribution analyzes a portfolio's performance based on its constituent holdings at specific points in time (e.g., beginning of the month, end of the month). It calculates the return of these holdings over a period, assuming a buy-and-hold strategy between valuation points, and then compares these to benchmark holdings and returns to determine allocation and selection effects.
Primary Inputs:
- Portfolio holdings (security identifiers, quantities, market values) at the beginning of each measurement period.
- Benchmark holdings at the beginning of each measurement period.
- Security-level returns for both portfolio and benchmark constituents.
Advantages:
- More Accurate than Returns-Based: Provides a more direct link between investment decisions (what was held) and performance.
- Identifies Allocation and Selection Effects: Can clearly distinguish between the impact of sector/asset class weighting decisions and security selection decisions.
- Commonly Used: A widely accepted method in the industry.
Disadvantages:
- Doesn't Fully Account for Intra-Period Transactions: Assumes a buy-and-hold strategy between valuation dates. The impact of trades made during the period (and their exact timing and price) is not precisely captured. This can lead to a "residual" or unexplained portion of the return.
- Accuracy Dependent on Valuation Frequency: The shorter the interval between holdings snapshots (e.g., daily vs. monthly), the more accurate the attribution, but this also increases data requirements.
- Timing/Trading Effect is an Approximation: Any difference between the holdings-based attribution result and the actual portfolio return is often labeled as a "trading effect" or residual, which is an estimate of the impact of intra-period activity.
- Can Be Misleading for High-Turnover Strategies: If the portfolio changes significantly between valuation dates, HBA may not accurately reflect the true drivers of performance.
Typical Use Cases:
- Attribution reporting for traditional equity and fixed-income portfolios.
- Manager evaluation where full transaction data is not available but periodic holdings are.
- Providing a reasonably accurate breakdown of performance drivers when daily transactions are too cumbersome to process for all portfolios.
Definition: Transaction-based attribution is the most granular and potentially most accurate method. It incorporates all portfolio transactions (buys, sells, corporate actions, cash flows) at their exact time and price, along with holdings, to precisely calculate the contribution of each investment decision.
Primary Inputs:
- Initial portfolio holdings.
- All transactions (security, quantity, price, date/time, transaction costs).
- All cash flows (contributions, withdrawals).
- Security-level returns and benchmark data.
- Corporate action details.
Advantages:
- Highest Level of Accuracy: Provides the most precise explanation of performance because it accounts for the exact timing and cost of all investment decisions.
- Fully Reconciles to Portfolio Return: When implemented correctly, there should be no unexplained residual; the sum of attribution effects should match the actual portfolio return.
- Captures Impact of Trading: Explicitly measures the impact of trade execution and timing.
- Diagnostic Tool: Can help identify errors in pricing or transaction recording because it requires full reconciliation.
- Reflects True Manager Activity: Best reflects the dynamic nature of active portfolio management.
Disadvantages:
- Most Complex and Data-Intensive: Requires complete, accurate, and timely data for all holdings, transactions, and cash flows.
- Difficult and Costly to Implement: The data requirements and reconciliation efforts can be substantial, making it the most challenging and expensive method to set up and maintain.
- Requires Robust Systems: Needs sophisticated systems to process and manage the large volume of detailed data.
Typical Use Cases:
- Detailed performance analysis for actively managed portfolios where precision is paramount.
- Internal performance review by portfolio managers to understand the exact impact of their decisions.
- Client reporting for sophisticated investors who demand a high level of transparency and accuracy.
- Fixed-income attribution, where the timing of cash flows and yield curve changes are critical.