Performance Evaluation

Question 1

Returns-Based, Holdings-based, Transactions-Based Performance Attribution

Answer 1

Returns based = regresses total port returns against major risk factors (eg systematic risk, size, value) to ID active bets of mgr and their impact o active returns

–> uses less data than HBSA

–> less window dressing, because looks at avg exposures over the regression period than HBSA which is a current snapshot

Holdings-based = uses bgn of period portfolio holdings to assess active sector/stock selection bets of the mgr and their contribution to active return

–> doesnt adjust for portfolio changes made during evaluation period –> output might not match overall port returns for mgr with higher turnover –> mismatch = “timing” or “trading” effect –> fix using shorter evaluation periods

–> requires MORE data

–> subject to WINDOW DRESSING

Transactions-based = improves on holdings based by including imact of any rades executed during evaluation period

selection of what method is right depends on nature of availabele data, invmt process of mgr, and need for accuracy

Question 2

Style Return

Answer 2

benchark return (B) - market index return (M)

NOTE: the market index return can be the benchmark return if the fund’s benchmark is the same as the relevant market index, in which case style return = 0%

BEWARE OF RfR BEING DISTRACTOR

Question 3

Misfit Active Return and True Active Return

Answer 3

if a mgr uses the wrong index for a benchmark, and there’s one that more accurately reflects the fund’s invmt universe,

Misfit Active Return = the difference between the CORRECT benchmark and the wrong one

True Active Return = is just the correct active return now using the new appropriate benchmark

BEWARE OF RfR BEING DISTRACTOR

Question 4

Sharpe Ratio

Answer 4

Return - RfR

_______________________________

StDev

Drawback: denominator doesn’t differentiate between volatility that is upside vs. downside –> penalizes all volatility, even “good” volatility

Question 5

Treynor Ratio

Answer 5

similar to sharpe but denominator is BETA

–> so only considers SYSTEMATIC risk rather than total risk

–> universe of appropriate benchmarks = limited to only the ones that assume efficient markets

–> only useful when evaluating portfolios that have systematic risk and DO NOT have unsystematic risk, i.e. WELL-DIVERSIFIED PORTFOLIOS

Question 6

Information Ratio (IR)

Answer 6

measures port performance against benchmark, but accounts for differences in risk

numerator = diff btwn mean returns of the portfolio and the benchmark, respectively

DENOMINATOR = TRACKING RISK, aka variability in portfolio performance vs that of the benchmark

Question 7

Appraisal Ratio

Answer 7

measures ratio of active return to volatility of the residual term, also referred to as the standard error of the regression

(^both derived from factor-based regression)

A mgr with a higher AR generates more active return per unit of active risk

–> so higher AR = stronger active mgr

Note: the std error of the regression is the volatility of the error term in a factor-based regression –> represents the volatility of mgr returns that is NOT explained by regression factors —> it’s generated by mgr taking active bets reltaive to factors of the model

THEREFORE: AR is basically the same as the IR but the difference is that AR uses factor-based regression

SEE EXAMPLE IN BOOK Mod 26.4

Question 8

CAPM

Answer 8

Expected return of manager = RfR + Beta(market return - RfR)

–> alpha can be calculated as diff between return of security / portfolio and the CAPM fair return

Question 9

Sortino Ratio

Answer 9

only considers stdev of DOWNSIDE risk (doesnt account for “good” volatility)

–> more apprpriate for non-normal return distributions

–> positive/negative skewed strategies = lower sharpe, but only the negatively skewed strategy = lower Sortino

–> better for hedge funds or options

Drawback: comparability problem: what MAR is specific to each investor and is subjective

FORMULA:

Expected return - Minimum expected return

____________________________________________

target semideviation

Question 10

Properties of a valid Benchmark

S A M U R A I

Answer 10

• S pecified in advance
• A ppropriate
• M easurable
• U nambiguous
• R eflective of mgr’s investment opinions (be familiar w securities and factor exposures)
• A accountable (mgr must accept ownership of the benchmark)
• I nvestable

Question 11

TWR vs MWR

Answer 11

Time-weighted returns = unaffected by ECFs (removes the effect of them), better measure of true returns under the assumption that the CLIENT controls the ECF timing, exception if mgr dictates when client may add or subtract funds; if manager controls, you MUST use money-weighted

Money-weighted returns = includes ECFs

if there is a huge inflow before a large increase in performance due to mkt value change, MWR will be higher