CFA III – Performance Evaluation 关键词清单 Vol 6

CFA III-Performance Evaluation 关键词清单 Vol 6

Study Session 17 Performance Evaluation
Reading 33 Evaluating Portfolio Performance
中量题，有长篇简答题 & 1 套 case study 共 25 题
2 Importance of Performance Evaluation 3Q121 LOS a
2.1 The Fund Sponsor’s Perspective. 3Q121
1) Exhaustive quality control check. Not only fund performance but also sources of that
performance. 2) part of the feedback step of investment management process. 3) The increased
complexity of institutional investment management. 3Q122
2.2 The Investment Manager’s Perspective 3Q122 serve as a feedback and control loop to
monitor the proficiency of various aspects of the portfolio construction process.
3 The Three Components of Performance Evaluation 3Q122 LOS b
Performance measurement: calculates rates of return based on investment-related changes in an
account’s value over specified time periods.
Performance attribution :investigate both the sources of the account’s performance relative to a
specific investment benchmark and the importance of those sources.
Performance appraisal : draw conclusions concerning the quality (that is, the magnitude and
consistency) of the account’s relative performance. 3Q123
4 Performance Measurement 3Q123
4.1. Performance Measurement without Intraperiod External Cash Flows 3Q123
External cash flows: contributions and withdrawals to and from an account, as opposed to
internal cash flows such as dividends and interest payments 3Q123
Equation 1(no external cash flows) : 3Q123
MV 1 = 期末 market value
MV 0 = 期初 market value
Equation 2 (期初有 external cash flows): 3Q124

Equation 3 (期末有 external cash flows) 3Q124

4.3 the Time-Weighted Rate of Return 3Q126 LOS c

考虑 intraperiod external cash flows: 用 subperiods 把整个 evaluation period 断开，每个
subperiod 期末有 external cash inflow 按公式 3 计算【确保题中明确有 external cash inflow 那
些天的 market value 值包含 these cash inflows】 。Example 3 再看一遍 3Q127

Chain-linking: equation 4: rtwr = (1 + rt,1) × (1 + rt,2) × … × (1 + rt,n) – 1   3Q128

【注意结果不一定是 annualized 的】
4.4. The Money-Weighted Rate of Return 3Q128 LOS c 也称 internal rate of return.

Equation 5：MV1 = MV0(1 + R)m + CF1(1 + R)m–L(1) + … + CFn(1 + R)m–L(n)   3Q128

其中：m = number of time units in the evaluation period (for example, the number of days in the
month)
CFi = the ith cash flow
L(i) = number of time units by which the ith cash flow is separated from the beginning of the
evaluation period 3Q129

1
 CFA III – Performance Evaluation 关键词清单 Vol 6
2

注：R = money-weighted return normally on daily basis. Example 5 3Q129 可再看

4.5. TWR versus MWR 3Q129 LOS c
MWR 指 average growth rate of all money invested in an account, 而 TWR 是指 growth of a
single unit of money invested in the account. = MWR is sensitive to the size and timing of
external cash flows to and from the account, 而 TWR is unaffected by these flows.==>当外部现
金流足够大时，二者区别大【在后期回报率飙升之前充入外部现金流，MWR 终值>TWR; 在
后期回报率飙升之前 withdraw 外部现金流，MWR 终值<TWR】 。3Q130
用 TWR 来评估经理表现：当组合经理无法控制外部现金流 size & timing 时，用 TWR 可避开
external cash flows 对回报率计算的影响。反之用 MWR（比如 private equity investment）
3Q130-131
4.6. The Linked Internal Rate of Return 3Q131
Linked Internal Rate of Return (LIRR) method: TWR should be approximated by calculating the
MWR over reasonably frequent time intervals and then chain-linking those returns over the
entire evaluation period. 3Q131 Example 7 3Q132
4.7. Annualized Return 3Q132
If the total investment period > one year, you must take the geometric mean（几何平均数/开根
根=number of years in measurement period） Example 8 3Q132
4.8. Data Quality Issues 3Q133 LOS d
1) For accounts invested in illiquid and infrequently priced assets, reported rate of return 不可靠.
2) Matrix Pricing: 用近似的 securities 来估算 illiquid ones 市值. 3Q133
3) Account valuations should be reported on a trade-date accounting, including accrued interest
and dividends. 3Q133
5 Benchmarks 3Q133
5.1 Concept of a Benchmark 3Q134 LOS e
Equation 6: Portfolio = Benchmark + Active Management Decisions. 3Q134
Equation 7: Portfolio = Market index + Manager’s investment style + Active Management
Decisions. 3Q134
其中：Manager’s investment style = Benchmark – Market index. 3Q134 比如 market index 是
S&P 500，而 benchmark 可能是 S&P 500 中的权重股指数。
5.2. Properties of a Valid Benchmark 3Q135 LOS f 以下 7 点尽量背熟
1) Unambiguous: 2) Investable: 3) Measurable: 4) Appropriate:
5) Reflective of current investment opinions. 6) Specified in advance. 7) Owned.
5.3 Types of Benchmarks 3Q136 LOS f
1. Absolute: 缺点：not investable. 3Q136
2. Manager Universes: 缺点：除了 measurable，其他都不符合条件。3Q136
3. Broad market indexes: 优点：基本符合 benchmark 条件，缺点：manager’s style may deviate
considerably from the style reflected in a market index. 3Q137
4. Style indexes:四个基本指数：1) large-capitalization growth, 2) large-capitalization value, 3)
small-capitalization growth, and 4) small-capitalization value。优点：基本符合 benchmark 条件，
缺点：1）Some style indexes contain weightings in certain securities and economic sectors that
are much larger than what many managers consider prudent. 2）the definition of investment style
implied in the benchmark may be ambiguous or inconsistent with the investment process of the
manager being evaluated. 3Q137
5. Factor-Model-Based: equation 8: Rp = ap + βpRI + εp   3Q137 (one-factor model)

2
 CFA III – Performance Evaluation 关键词清单 Vol 6
3

注：Rp =periodic return on the market index; RI = return on market index; εp = residual or
nonsystematic element. βp = the sensitivity of the returns on the account to the returns on the
market index; ap = “zero factor” term. 3Q138
A normal portfolio: a portfolio with exposures to sources of systematic risk that are typical for a
manager, using the manager’s past portfolios as a guide. 3Q138 优点：capture systematic sources
of return that affect an account’s performance. 缺点：1) not easy to obtain, 2) expensive to use
3) ambiguous:使用同一组 factor exposures 可以 build multiple benchmarks 且 composition is not
specified. 3Q138
6. Returns-Based: 使用 1) the series of a manager’s account returns 2) the series of returns on
several investment style indexes over the same period.然后通过 allocation algorithm to solve for
combination of investment style indexes that most closely tracks the account’s return. 3Q138
优点：基本符合 benchmark 条件，特别适合当 info 只有 account returns 时。缺点：1）like the
style indexes, too much weighting in certain securities and economic sectors that a manager
might find unacceptable. 2）需要 many months of observation 来建立 style exposure patterns.
如果经理 rotate among style exposures，patterns 就很难区分。3Q139
7. Custom Security-Based: 优点：1）完全符合 benchmark 条件 2）effectively allocate or budget
risk across teams of investment managers. 缺点：1)expensive to construct and maintain. 2)不由
published indexes 组成，potential lack of transparency. 3Q139
5.4. Building Custom Security-Based Benchmarks 3Q139 LOS g
1. Identify prominent aspects of the manager’s investment process. (analysis of past portfolios)
2. Select securities consistent with that investment process.
3. Devise a weighting scheme for the benchmark securities, including a cash position.(equal
weighting or capitalization weighting) 3Q140
4. Review preliminary benchmark and make modifications.
5. Rebalance benchmark portfolio on a predetermined schedule. 3Q140
5.5. Critique of Manager Universes as Benchmarks 3Q140 LOS h
不足：1）median acct can’t be specified in advance (only ex post basis) 2) not investable
3) ambiguous: 即使 evaluation period concludes，median manager identity remains unknown，
无法分析其组合结构无法测 appropriateness. 4) fund sponsors terminate underperforming
managers, 所以该法 subject to survivor bias. 3Q141
5.6. Tests of Benchmark Quality 3Q142 LOS i
1. Systematic biases: Over time, there should be minimal systematic biases or risks in the
benchmark relative to the account. 测试方法：1）计算 historical beta of account relative to
benchmark ≈ 1 on average 3Q14 2）correlation between A = (P – B) and S = (B – M)a
good benchmark will display a correlation between A and S that is not statistically different from
zero. 3）E (difference) = account portfolio – market index  a manager’s style (S) is in favor
relative to the market, both the benchmark and the account outperform the market. a good
benchmark will have a statistically significant positive correlation coefficient between S and E.
3Q142
2. Tracking Error: the volatility (standard deviation) of an account’s returns relative to a good
benchmark should be less than the volatility of the account’s returns versus a market index or
other alternative benchmarks. 3Q142
3. Risk Characteristics: factors that systematically affect the returns on many securities. The
objective of a good benchmark is to reflect but not to replicate the manager’s investment process.

3
 CFA III – Performance Evaluation 关键词清单 Vol 6
4

即 good benchmark risk exposures 相对于 managed portfolios 时高时低；但如果一直高（或低） ，

就存在 systematic bias. 3Q143
4. (Benchmark) Coverage: 也叫 coverage ratio: market value of the jointly held securities [即
benchmark 和 actively managed portfolio 都持有] as a percentage of total market value of the
portfolio 3Q143
5. (Benchmark) Turnover: proportion of benchmark’s market value allocated to purchases during
a periodic rebalancing of benchmark. Because benchmark should be investable, turnover should
not be so excessive as to preclude a passively managed portfolio strategy. 3Q143
6. Positive Active Positions: An active position = an account’s allocation to a security -
corresponding weight of same security in benchmark；negative active positions 使 benchmark 不
能反映经理投资 approach. 3Q143
5.7. Hedge Funds and Hedge Fund Benchmarks 3Q143
Hedge funds: expose investors to a particular investment opportunity while minimizing (or
hedging) other investment risks that could impact the outcome. #143
因为 long-short position 可能使 initial market value=0(active weights sum = 0)，公式 1 不能用，
推荐公式 9：3Q144

rv = rp – rB   where rν = value-added return; rp = portfolio return; rB = benchmark return

1) Given historical returns and holdings of a long-short portfolios, we could use either
returns-based or security-based benchmark building approaches to construct separate long and
short benchmarks. 3Q145
2) Ambiguity of hedge fund manager opportunity sets has led to use of the Sharpe ratio to
evaluate hedge fund manager performance. 3Q145
6 Performance Attribution （定量分析）3Q146
Macro attribution: carried out on fund sponsor level. 3Q146
Micro attribution: carried out on investment manager level. 3Q146
6.3 Macro Attribution Inputs 3Q148 LOS k
1. Policy Allocations: normal weightings to the asset categories within the fund and normal
weightings to individual managers within the asset categories. Normal 指 long term investment
objectives & constraints oriented. 3Q148 Exhibit 3 实例展示。3Q149
2. Benchmark Portfolio Returns: Exhibit 4 实例展示。3Q149
3. Returns, Valuations, & External Cash Flows. 3Q149
1)on a return-only metric requires fund turns at individual manager level.
2)extended to include a value-metric approach, then account valuation and external cash flow
data are needed not only to calculate accurate rates of return, but also to compute the value
impacts of the fund sponsor’s investment policy decision making. 3Q149 Exhibit 5 实例展示。
3Q150
6.4. Conducting a Macro Attribution Analysis 3Q150 LOS l
以下 6 点被当成 investment strategies，each investment results are compared to cumulative
results of previous levels. 随数字变大，有 increasing volatility & complexity.  Macro
attribution calculates incremental contribution that choice to move to next strategy
produces.3Q150 分析按照以下 6 点数字顺序逐步叠加剖析。
注:Each decision-making level represents a valid benchmark (investment alternative). 3Q150
1)Net Contributions: (contribution or withdrawl) invested at a rate of zero. 3Q151

4
 CFA III – Performance Evaluation 关键词清单 Vol 6
5

2)Risk-Free Asset: 3Q151

3)Asset Categories: the Fund’s beginning value and external cash flows are invested passively in a
combination of the designated asset category benchmarkspure index fund approach.
Equation 10(return-metric):
注 ： rAC = incremental return contribution of Asset Category
investment strategy, rCi = return on the ith asset category,
rf =risk-free return, wi = policy weight assigned to asset category i,
and A = the number of asset categories. 3Q151-152
4)Benchmarks: represent managers’investment styles. 3Q152misfit return (style bias)
=benchmark return – asset category return 注：benchmark return = 1-4 步 ttl return；asset
category return = 1-3 步 ttl return. 3Q153
Equation 11(return-metric):
注: rIS =incremental return contribution of the Benchmarks
strategy, rBij = return for the jth manager’s benchmark in asset
category i, rCi = return on the ith asset category[market index],
wi = policy weight assigned to the ith asset category, wij = policy weight assigned to the jth
manager in asset category i, and A and M are the number of asset categories and managers.
3Q152
5)Investment Managers or Value of Active Management: further deviation from benchmark:
managers’ active bets based on actual portfolio.Equation 12(return-metric): 3Q153
注 ： rIM =incremental return contribution of Active
Management strategy, rAij=actual return on the jth manager’s
portfolio within asset category i
6)Allocation Effects: Allocation Effects incremental contribution is a reconciling factor = Fund’s
ending value - value calculated at the Investment Managers level(1-5 步 ttl value) 3Q154
源于：fund sponsors deviate from their policy allocations to asset categories & managers.3Q154
6.5. Micro Attribution Overview 3Q154 LOS k & L
该小点下的“account” = a specific portfolio invested by a specific investment manager which we
will refer to as the “Portfolio.” 3Q154
Equation 13 : Rv = = Active Return = Rportfolio –Rbenchmark = 展开如下
注: Wpi = proportions of the actual Portfolio invested in security i;
Wbi = proportions of the Portfolio invested in security i;

Equation 14 : Equation 13 变形式 simplest form of micro performance attribution: a

security-by-security attribution analysis 3Q155

注：rB = return on the Porfolio’s benchmark 3Q155

用 Equation 14 做 micro 分析缺点: large number of securities in a well-diversified portfolio makes

the impact of any individual security uninteresting. 3Q155
更有效的 Micro 分析手段：Factor model of returns: return on a security (or portfolio of securities)
is sensitive to the changes in various factors. 3Q156 最简单的形式：one factor model:
Rp = ap + βpRI + εp

5
 CFA III – Performance Evaluation 关键词清单 Vol 6
6

6.6. Sector Weighting/Stock Selection Micro Attribution 3Q156 LOS k & l

Equation 15:
注：wpj = Portfolio weight of sector j；wBj = benchmark weight of sector j
rpj = Portfolio return of sector j；rBj = benchmark return of sector j
S = number of sectors 3Q157
Equation 16: Equation 15 变形式 3Q158 rB = returns on overall portfolio’s benchmark 3Q158

Example 13 The Pure Sector Allocation Return for Consumer Nondurables 3Q158 可再看一遍
Example 14 The Within-Sector Allocation Return for Technology 3Q159
Example 15 The Allocation/Selection Interaction Return for Technology 3Q159
Allocation to a particular security adds not only to that security weight but also to the sector
weight to which the security belongs, unless an offsetting adjustment to securities within that
sector is made, security selection decisions can drive sector-weighting decisions.159
6.7. Fundamental Factor Model Micro Attribution 3Q160 LOS m
Fundamental factors: with respect to common stocks, a company’s size, its industry, its growth
characteristics, its financial strength. 3Q160
评估伊始要确定：1）Portfolio Exposures(Risk factors Active Exposure=portfolio expo-normal
expo) 2）benchmark to the factors of the fundamental factor model (normal exposure) 3Q160
注：fundamental factor model 中最后一项 specific(unexplained) return component which can’t
be explained by factor model is attributed to investment manager. 3Q161
Example 16 Fundamental Factor Model Micro Attribution 3Q 160-161 可再看一遍。
6.8. Fixed-Income Attribution 3Q161 LOS n & o
1. Sector weighting/stock selection approach to mirco attribution 适用于 fixed-income
2. 将 economic sectors such as energy 改成 market segments 诸如 government bonds, MBS 等
3Q162
External Interest Rate Environments: fixed-income portfolios have higher return rates in periods
of falling interest rates (bond value 增幅大)and lower return rates in periods of rising interest
rates.
Active Management: 3Q164： 在 IPS 规定范围内， manager can adjust portfolio’s interest-sensitive
characteristics in anticipation of forecasted yield curve & spread changes. 3Q164
Ttl return of a fixed income portfolio’s attributions = external interest rate effect + active
management effect: 3Q165
一、external interest rate effects总体表现同 passive default free bond portfolio (benchmark)
1）return on the Default-free benchmark assuming no change in forward rates (expected return)
2）return due to forward rates change (unexpected return) 3Q165
二、active management effects: 一般无对比 benchmark 3Q165 以下计算公式理解，考试会
直接给出计算结果，会分析即可。
1）Interest rate management effect: how well manager predicts interest rate changes. To calculate

6
 CFA III – Performance Evaluation 关键词清单 Vol 6
7

this return, each security in the portfolio is priced as if it were a default-free security. 公式 =
aggregate return of these repriced securities – return of entire Treasury universe. 注：可进一步
分为 returns due to duration, convexity, & yield curve shape change. 3Q165
2）Sector/quality effects: manager’s ability to select “right” issuing sector and quality group.
计算过程（概念化） ：1.reprice each security in the portfolio using average yield premium in its
respective category 2. 第一步结果计算 a gross return 3. sector/quality effect return = the gross
return – (external interest rate effect + interest rate management effect) 3Q165
3）Security Selection effect: how return of a specific security within its sector relates to average
performance of the sector. Security selection effect for each security = total return of a security
- all the other componentsportfolio security selection effect = market-value weighted average
of all the individual security selection effects. 3Q165
4）Trading activity: effect of sales and purchases of bonds over a given period. 公式=total
portfolio return - all the other components. 3Q165
Exhibit 14.Performance Attribution Analysis of Two Fixed-Income Managers for the Windsor
Foundation, Year Ending 31 December 20xx 3Q165-167.分析可以再看一遍
7 Performance Apprasial （定性分析）3Q167
7.1. Risk-Adjusted Performance Appraisal Measures 3Q168 LOS p & q
1. Ex Post Alpha(也称 Jensen’s alpha): 基于 ex post Security Market Line横轴是β=COV
i,mkt ，纵轴是 expected return on asset i 3Q168
Equation 17:

Equation 18: equation 17 变形式,成为 linear

regression for a period of t, 3Q168
2. Treynor Measure: 也源自 ex post SML，是 SML 的 slope
Equation 19: 注：分子两个因子都是 average values over the evaluation
period. 分母是 expected value. 3Q169
小结： 1)ex post alpha 和 Treynor measure 有 same assessment 结论。
3Q169
2) only consider beta (market risks)
3. Sharp Ratio: 3Q169 Equation 20:
1) 基于 ex post CML (capital market line)  a skilful manager’s acct return 将高于
CML 2) portfolio standarddeviation include both beta & alpha(active management)
4. M2 : 同 sharp ratio, 基于 ex post CML
Equation 21: M2 measures what the account would have returned if it had taken on the same
total risk as the market index. 3Q169
注：1）A skillful manager’s M2 value > market index return.

5. Information ratio: Equation 22: 3Q170

分母=active return = long position in the account is funded by a short position
in the benchmark. 3Q170
分子=active risk = standard deviation of difference between acct returns &

7
 CFA III – Performance Evaluation 关键词清单 Vol 6
8

benchmark returns. 3Q170  所以只有 alpha related risk.

7.2 Quality Control Charts 3Q171 LOS r
To present performance appraisal data. 3Q171
Exhibit 15 Quality Control Chart:
Cumulative Annualized Value-Added
Illustrating Manager Performance
within Expectations 3Q172
纵 坐 标 0 点 ： benchmark
performance; 3Q171 / Jagged line：
portfolio cumulative annualized
excess return(value-added return)
基于 3 个 assumeptions：1）null
hypothesis that manager has no
investment skill expected value
added return = 0即 exhibit 15 中 jagged line 与 benchmark 重合 suggests that those ex post
value-added returns have no directional biases and are entirely due to random chance. 172
2）manager’s value-added returns are independent from period to period and normally
distributed around the expected value of zero. 3Q172
3）manager’s investment process does not change from period to period.==>variability of
manager’s value-added return remains constant over time.（即可 use past performance to
estimate the dispersion of value-added return distribution）3Q172
基于以上三点及 Exhibit 15，=========================
Exhibit 16. Expected Distribution
of the Manager’s Value-Added
Returns 3Q173
Confidence band indicates the
range in which we anticipate that
the manager’s ex ante
value-added returns will fall a
specified percentage of the
time.3Q173
例 子 ： 某 一 年 80% time 的
manager value-added return 在
confidence band 里，推出以 mean =0 为中心，1.28 standard deviations around the mean
captures ex ante 80% of the possible outcomes associated with a normally distributed random
variable.==>exhibit 15 的 funnel-shaped confidence band 是累积统计 cumulative value-added
returns. 3Q173
1) Standard deviation of annualized cumulative value-added returns decreases at a rate
= (1 / square root of time) *the one-year value . 即 the standard deviation of annualized
cumulative value-added returns at two years is 1/√2 of the one-year value, at three years it
is 1/√3 of the one-year value 3Q173
2) 随时间增长 confidence band gets narrower and converge on benchmark linecumulative
annualized results lie closer to expected value of zero.(as time passes, it becomes
increasingly likely that the manager’s random positive and negative value-added returns will

8
 CFA III – Performance Evaluation 关键词清单 Vol 6
9

offset one another.) 3Q 173

7.3. Interpreting the Quality Control Chart 3Q174 LOS r
3) For a manager whose investment results are within that range (confidence band), we have no
strong statistical evidence to indicate that our initial assumption of no skill is incorrect. “We
cannot reject the null hypothesis that the manager has no skill.” 3Q174
4)超越 pierce confidence band on upside or downside, reject null hypothesis, 当穿越下档时，表
示 the manager is systematically incapable of recapturing the costs of doing business and should
be classified as an “underperformer.” 3Q175
8 the Practice of Performance Evaluation 3Q175 3
大步骤合并分析
8.2. Manager Continuation Policy 3Q178 LOS s
Exhibit 21 Manager Continuation Policy flow chart.
3Q179
Goals & guidelines with MCP:
1. Replace managers only when justified
(minimize unnecessary manager turnover)
1) Short periods of underperformance should not
necessarily mean automatic replacement.
2. Develop formal policies and apply them consistently to all managers.
3. Use portfolio performance & other info in evaluating managers.
1) Appropriate & consistent investment strategies 2) Relevant benchmark style
3) Personnel turnover 4) Acct Growth
Implement MCP process involves: 1) continual manager monitoring 2) regular, periodic manager
review (closely resembles manager selection process).
Costs of hiring & firing managers: 1) Fired managers’ portfolios must be converted to the hired
managers’ portfolios. Which requires buying and selling securities, incurring trading costs.
2) A substantial percentage of the fired manager’s portfolio may need to be liquidated in the
process of moving the assets to a new manager. 3) expenses of converting a manager’s portfolio
considers only direct monetary costs. For most fund sponsors, replacing managers involves
significant time and effort 3Q178
8.3. Manager Continuation Policy as a Filter 3Q181 LOS t
Positive value-added managers: beat benchmarks(after all fees & expenses) 3Q181
Zero value-added managers: just enough to cover their fees & expenses matching benchmark
performance.
Negative value-added managers: can’t earn back fees & expenses. 3Q181
null hypothesis : the managers under evaluation are at best zero-value-added managers 181
•Type I error—keeping (or hiring) managers with zero value-added. (Rejecting the null hypothesis
when it is correct.其实不应该 reject) < --caused by a coarse filter
•Type II error—firing (or not hiring) managers with positive value-added. (Not rejecting the null
hypothesis when it is incorrect.)= a fine 细致 filter 会导致,manager monitoring guideline
application too rigid 3Q182
小结：扩大 width of confidence band within quality control chart，减少 Type I error 同时增加
Type II errors. 3Q182