CAIA L2 - 5.2 - Benchmarking and Performance Attribution Flashcards
List
Attributes of an Effective Benchmark
(Bailey Criteria)
5.2 - Benchmarking and Performance Attribution
SUMO ISA
1. Specified in advance
2. Unambiguous - specification of assets and their weights
3. Measurable - market data readily available
4. Opinion - manager opines on deviations
5. Investible - all assets must be investable
6. Style - bench reflects manager’s style
7. Agreed - manager agreed to be evaluated against it
5.2 - Benchmarking and Performance Attribution
Quote
An appropriate benchmark for LAI -
Liquid Alternative Investment
5.2 - Benchmarking and Performance Attribution
peer group benchmark is
preferred (of same style)
vs
long-only equity indices
Drawback: may not be investable + 5 other factors. Usually satisfy measurability
5.2 - Benchmarking and Performance Attribution
Formula
Excess return on a fund
(in a single-factor model)
R’i’ - R’f’
5.2 - Benchmarking and Performance Attribution
Excess return on a fund
= return attributable to systematic risk βi (R’m’ – R’f’) +
return attributable to idiosyncratic risk ε’i’
(R’i’ – R’f’) = βi (R’m’ – R’f’) + ε’i’
Ex:
In using the ex post form of the capital asset pricing model (CAPM), an asset has a risk-free rate of 3%, a beta of 1.3, and a realized return of 10%. The market portfolio return was 8%. What is the asset’s idiosyncratic return?
10% = 3% + [1.3 (8% – 3%)] + idiosyncratic return
idiosyncratic return = 10% – 3% – 6.5% = 0.5%
5.2 - Benchmarking and Performance Attribution
Formula
Excess return on a fund
(in a time series single-factor model)
R’it’ - R’f’
What happens to the formula for CAPM holds?
5.2 - Benchmarking and Performance Attribution
Excess return on a fund
= return attributable to manager value added
return attributable to systematic risk +
return attributable to idiosyncratic risk
(R’it’ – R’f’) = α’i’ + β’i’ (R’mt’ – R’f’) + ε’it’
In order to CAPM holds = α = ε = 0; β = same of true CAPM beta
5.2 - Benchmarking and Performance Attribution
List
Reasons why
CAPM-based risk analysis
is inappropriate for alternative investments
5.2 - Benchmarking and Performance Attribution
Remermber: Alternative investments are multiperiod, illiquid, non-normal + investor-specific holding
(investors can concentrate because of specifics of liabilities)
1. Multiperiod issues. CAPM is a one-period model, and it assumes that all investors have the same one-period time horizon. For CAPM to hold in a multiperiod scenario, we would have to assume that the asset return distributions are static; if they fluctuate over time, additional risk factors would be introduced. While these multiperiod issues could affect traditional asset classes also, it would be more pronounced due to the dynamic nature of alternative investments, especially those alternative investments with unique distributions (e.g., structured products).
2. Non-normality. CAPM assumes that return distributions are normal. Normal distributions are characterized by only two parameters: mean and variance. If the distribution of asset returns is not normal, additional parameters such as skewness and kurtosis would need to be considered.
Skewness and kurtosis is especially relevant for the many alternative investments that have asymmetric payoffs. Even in revised models that account for skewness and kurtosis, these parameters are unlikely to be stable over time for alternative investments. Hence, alternative investments are subject to additional risk factors beyond the market risk.
3. Illiquidity. CAPM assumes frictionless markets, which would allow investors to diversify and change asset weights without any cost. In the actual alternative investment universe, transaction costs and taxation considerations can be significant (e.g., in real estate).
4. Investor-specific holdings. CAPM assumes cost-free diversification and hence no reward for bearing diversifiable, idiosyncratic risks. Institutional investors (e.g., pension funds) may have large holdings of illiquid and undiversified real estate. Similarly, the liability profile of some of these investors may be heterogeneous (i.e., short-term for P&C insurance companies and long-term for life insurance companies and pension plans). Additionally, there can be multiple risk exposures for liabilities, making a single risk factor model like CAPM inadequate.
5.2 - Benchmarking and Performance Attribution
Differentiate
3 Generations of
commodity indices
5.2 - Benchmarking and Performance Attribution
1. First generation commodities indices - Long only + NO dicretionarity on contango/backwardation (usually energy)
2. Second generation commodities indices - Long only + Lessen contango effect / exploit backwardation
3. Third generation commodities indices - Long/short + active management to gain in both scenarios
5.2 - Benchmarking and Performance Attribution
List
Two categories of
commodity trading advisors (CTAs)
best suitable benchmark for each
5.2 - Benchmarking and Performance Attribution
Trend following
* Best benchmark: Passive trend-following indices (algorithmic index)
Nontrend following
* Best benchmark: peer group; pushback: availability of quality peer group data (not investable)
’–
* Both have low correlations with traditional asset classes
* CTA Excess returns: < 50% from passive exposure - More than half of the historical excess returns of trend-following CTAs is not explained by passive exposures (i.e., beta return is less than half of the total excess return).
5.2 - Benchmarking and Performance Attribution
List
Two approaches to
creating benchmarks
for private equity funds
5.2 - Benchmarking and Performance Attribution
1. Listed asset-based benchmarks. This approach creates an index (e.g., LPX 50) consisting of shares of publicly traded private equity firms or business development companies. A drawback of this approach is that the index performance is driven by the general partner’s cash flows (from fund performance and fees) and not on the performance of the underlying PE assets (i.e., the LP’s cash flows). The statistical characteristics of the return distribution of publicly traded PE funds is similar to that of other public equities and not to that of private equity investments.
2. Public market equivalent (PME) method. The Long-Nickels (LN) method of PME focuses on calculating two IRRs: one based on the cash flows of the private equity fund and the other based on the same cash flows invested in a hypothetical portfolio at a public market index rate of return. Like the IRR, PME is a cash-weighted metric that compares the performance of a private equity investment to that of an investment in a public market index.
5.2 - Benchmarking and Performance Attribution
Formula, Define
Cap rate
Cap rate spread
Cap rate limitations
5.2 - Benchmarking and Performance Attribution
Cap rate = NOI / value
NOI => preleverage income ( => cap rate = unleveraged rate of return)
NOI = recent, current, or forecasted
value = transaction price or appraised
‘–
Cap rate spread = cap rate - US10Y
‘–
Cap rate limitations =
1. ignores any growth potential in income
2. ignores any capital appreciation/depreciation
5.2 - Benchmarking and Performance Attribution
List
3 Approaches
to Benchmarking
Noncore Real Estate
5.2 - Benchmarking and Performance Attribution
- Observed cap rates for similar styles of real estate. Unfortunately, unlike that of core real estate, the data availability for value-added and opportunistic properties may be unreliable.
- Absolute hurdle rates. This approach fails to consider different interest rate/inflation environments.
- Risk premium based on relative (to core real estate) risk level differences. This approach is the most appropriate due to the limitations of the other two methods.
5.2 - Benchmarking and Performance Attribution
