CFA Book 5 Copy Flashcards
trading | market mircostructure
structure and process of a market that affects securities’ pricing in relation to intrinsic value and the ability of managers to execute trades
trading | market order
executed at best price
• time certain (immediate), price uncertain
trading | limit order
set price
• price is certain (equal to or better), time (execution) is uncertain
trading | bid-ask spread
- ask - bid
- aka inside bid-ask spread or market bid-ask spread (ie the best bid and best ask)
trading | inside and market bid and ask
inside and market both mean the ‘best’ as in best bid or best ask or best bid-ask spread
trading | midquote
- the avg of the ask and bid
- the mid-point between the bid and ask
trading | effective spread given a transaction
- for buy: 2 * (execution (buy) price - midquote)
- for sell: 2 * (execution (sell) price - midquote)
trading | price improvement and price impact
- price improvement: trade was executed within the bid-ask
- price impact: trade was executed outside of the bid-ask
trading | avg of bid-ask and effective spread
- bid-ask is unweighted
- effective can be weighted or unweighted
trading | what do dealers do
they make the bid-ask with limit orders (in a limit order book)
• dealers = market makers
trading | securities markets offer 3 things
- liquidity
- transparency
- assurity of completion
trading | 3 type of securities markets
- quote-driven (aka dealer markets): investors and dealers
- order-driven: investors trade with investors
- brokered markets: investors use brokers to find counter-parties
trading | quote-driven market
• aka dealer market
• dealers make bid-ask spreads
• dealers needed in:
◎ low liquidity markets
◎ negotiated terms (eg. swaps)
• closed-book market: only brokers can see the order book
trading | order-drive market
- more competition >> better customer prices
- no intermediary dealers
- no guaranteed liquidity
- 4 types: electronic crossing network, auction, automated auction, hybrid
trading | ECN
electronic communications network
• automated auction market
• exp: NYSE/ARCA, Paris Bourse
trading | 3 types of order-driven
- electronic crossing network
◎ no price discovery
◎ periodic trading (crosses) of batch orders - auction
◎ yes price discovery
◎ continuous and/or batch (eg. open/close) - automated auction
◎ price discovery
◎ ECN
◎ continuous
trading | hybrid market
combination of parts of other 3 market types
• exp: NYSE
trading | dealer vs broker
- dealer is a trader that takes the other side (and risk) of a trader
- broker is a go-between representing the customer
trading | broker traits
- represent the order
- find counter-parties
- provide secrecy
- other services (eg. record keeping)
- support the market
trading | market: liquidity, trasparency, assurity
- liquidity traits: narrow bid-ask, depth, resilience
- liquidity inputs: many buyers/sellers, diverse opinions, convenient access, integrity
- transparency: accessible quotes and order confirmations
- assurance: clearing bodies and brokers guarantee both sides
trading | explicit costs
- commissions
- taxes
- stamp duties
- fees
trading | implicit costs
- bid-ask spread
- market, price impact
- opportunity costs
- delay costs (slippage)
trading | VWAP
- metric to measure trading costs
- volume weighted avg price
- based on trade prices and volumes during the period (usually a day)
trading | measuring trading costs
can use VWAP or Implementation shortfall
trading | implementation shortfall
• metric to measure trading costs
• = (paper gain - actual gain) / paper investment
• actual trading vs paper trading
• actual trade price vs decision trade price (aka arrival price, strike price)
◎ decision price has no costs
• 4 categories
◎ explicit costs
◎ realized p/l
◎ delay, slippage cost
◎ missed trade opportunity
• can be share weighted
• can be dollar or % of decision price
trading | implementation shortfall: cost breakdown
- commission
- realized p/l = (actual - latest possible) / decision * % of order
- delay p/l = (latest possible - decision) / decision * % of order
- missed opportunity = (ending - decision) / decision * % of order
- ALL THESE COSTS ARE DIV BY ORIGINAL SHARE * ORIGINAL DECISION PRICE!!!
trading | measuring costs with implement shortfall: market adjustment
- Calc Ri = alpha + beta * Rm; assume alpha = 0
- subtract Ri from prev calced % cost to find ‘market adjusted implementation shortfall’
• negative cost is good
trading | cost measure: VWAP pros/cons
• pros:
◎ easy to understand
◎ easy to calc
◎ quickly applied to trading decisions
◎ good for small trades, non-trending mkt
• cons:
◎ bad for large volume trades
◎ possible to game
◎ not include delayed/unfilled order
◎ no include mkt movement/trade volume
trading | cost measure: implementation shortfall pros/cons
• pros:
◎ see cost of implementation
◎ see time vs price trade off
◎ decompose costs
◎ can be used in optimizer
◎ no gaming
• cons:
◎ novel to traders
◎ much data and analysis
trading | costs non-linearly corr with
- liquidy: volume, mkt cap, spread, price
- size of trade vs liquidity
- trading style: more aggressive >> higher costs
- momentum (buying stocks costs more when mkt going up)
- risk
trading | cost forecasting: econometric model
- can use econometric model regression to forecast cost and adjust size (ex ante)
- can use to evaluate trades (ex post)
- inputs: security liquidity, trade size, trade style, momentum, risk
- based on microstructure
trading | trader types (as opposed to tactics)
- information-motivated: time sensitve >> mkt order
- value-motivated: price sensitive >> limit order
- liquidity-motivated: time sensitive >> mkt order
- passive: price sensitive >> limit order
- dealer/MM
- day trader
trading | trade tactics
- liquidity at any cost: information
- costs-not-important: information, liqudity
- need-trustworthy-agent: not information
- advertise-to-draw-liquidity: not information
- low-cost-whatever-the-liquidity: passive and value
trading | algorithmic
• automated trading using algorithms
• 1/4 of all trades
• 3 types:
◎ logical participation strat: simple, implemenation shortfall
◎ opportunistic strat: increases when liquidity increases
◎ specialized strat: various
trading | algorithmic: simple logical participation
• goal: minimize trading impact by breaking trade up into smaller parts
• 3 types:
◎ VWAP: trade volume weighted
◎ TWAP: done evenly over time
◎ % of volume: trades % of total volume until trade is finished
trading | algorithmic: implementation shortfall strategy
- goal: minimize opportunity costs thru trading early >> trading early in day to avoid not executing
- aka arrival price strategy
trading | best execution
• security acquisition/liquidation the best way
• similar to prudence (security selection)
• 4 traits:
◎ cannot be judged independently
◎ not known ex ante
◎ only know ex post over time and over multiple trades
◎ relationships and practices are important as is diligence
trading | CFAI Trade Management Guidelines
• 3 Parts:
◎ processes
◎ disclosures
◎ record keeping
• maximize portfolio value using best execution
• procedures and policies >> measure and manage
• provide clients: info and conflicts
• record keeping: policies/compliance and disclosures
trading | adverse selection risk
- faced by dealers/mm
- those most willing/anxious to trade against the dealers bid or ask may be the most well-informed (ie information traders)
rebalance portfolio | causes
- investor situation/goals
- capital markets changes (eg. business cycle)
- portfolio asset values
• riskier assets will become larger % over time
rebalance portfolio | calendar rebalancing
rebalancing at pre-determined, regular intervals
• pros: disciplin
• cons: portfolio can become too unbalanced between rebalancing
rebalance portfolio | percentage-of-portfolio rebalancing
• aka PPR, percent range rebalancing, interval rebalancing
• tolerance bands, corridors = T +/- (T * P)
◎ T = asset allocation %
◎ P = max % chg
• P is specific to each asset class
rebalance portfolio | PPR: corridors
• 5 factors
1. high trans costs >> higher corridor
2. higher risk tolerance >> higher
3. higher corr with other assets >> higher
4. higher asset vol >> lower
5. higher other portfolio assets vol >> lower
• overall: sell winners, buy losers
• concave returns
rebalance portfolio | rebalance all the way?
depends on mkt trends and vol
rebalance portfolio | 3 strategies
- buy-and-hold: no rebalance; linear returns
- constant mix: calendar and PPR; sell winners, buy losers; concave returns
- constant proportion portfolio insurance (CPPI): buy winners, sell losers; convex returns
rebalance portfolio | constant prorportion portfolio insurance
• aka CPPI
• target weight equities investment = M * (portfolio value - floor value) = M * cushion
◎ M = constant proportion > 1 and does not change
• if M = 1 >> buy-and-hold
• if M < 1 >> CM
• overall: buy winners, sell losers
• convex returns
rebalance portfolio | 3 strategies vs mkt movements
- mkt trends in one direction: best(CPPI, buy/hold, CM)worst
- mkt mean reverts: best(CM, b/h, CPPI)worst
rebalance portfolio | NOTE: when assessing 3 strats, think of portfolio of risk-free asset (cash) and risky asset (equities) and what happens as the mkt moves
rebalance portfolio | 3 strategies vs risk tolerance
- CPPI: absolute and relative risk tolerance pos corr with wealth (larger corr than others); dynamic cash floor that increases wth mkt decline
- B/H: absolute and relative risk tolerance pos corr with wealth; static cash floor
- CM: absolute risk corr w/ wealth, relative risk tolerance proportional to wealth (relative risk aversion); decreasing cash floor (constant %)
rebalance portfolio | causes: investor changes
- wealth
- time horizon (time passing, events)
- liquidity
- taxes
- legal and regulatory (for institutional)
evaluating performance | fund sponsor
institutions: pension fund, endowment, foundation that has multiple managers beneath it managing investments
evaluating performance | evaluation purpose from fund sponsor perspective
• feedback and control
- evaluates policy performance
- directs managers to most added and lost value
- quantifies active management and policy decision performance
- provides other strats that may be successful
- feedback on IPS implementation
evaluating performance | evaluation purpose from manager perspective
• provides details of performance and allows comparison to benchmarks
evaluating performance | evaluation components
- performance measurement: rates of returns
- performance attribution: sources of performance
- performance appraisal: causes of those sources (investment decisions, overall mkt , chance, etc)
evaluating performance | external cash flows into/out of the fund
these must be taken into account when calculating return
evaluating performance | time-weighted rate of return
- geometric mean
- compound rate of growth over multiple subperiods (delineated by external cash flows)
- does not take into account timing of external CF or funds invested during each period >> evals manager decision, not investors external CF timing
- used for GIPS and manager eval
- harder to calc than MWRR
- aka TWRR
evaluating performance | money-weighted rate of return
- IRR
- PV of CFs
- takes timing into account of external CF and funds invested during each period >> groups together manager performance and investor external CF timing
- not good manager eval unless they control external CF timing
- easier to calc than TWRR
- aka MWRR
evaluating performance | TWRR vs MWRR
- if external CFs are large and returns are volatile, TWRR and MWRR can be very different
- TWRR to measure manager’s performance because it excludes effects of investors external CF decisions (unless manager controls those decsions)
evaluating performance | matrix pricing
creating a mark for a security when none is directly available and looking at dealer quotes on a similar security
evaluating performance | 3 components of return (performance) attribution
- market
- management style
- active management
• P = M + S + A
◎ P = manager’s portfolio return
◎ M = market return (not benchmark)
◎ S = B - M; B = benchmark (proxy for style)
◎ A = P - B
evaluating performance | 7 benchmark properties
- Specified in advance
- Appropriate
- Measurable
- Unambiguous
- Reflective of managments current investment opinions
- Accountable
- Investable
• SAMURAI
evaluating performance | 7 types of benchmarks
- absolute
- manager universes
- broad market indices
- style indices
- factor-model-based
- returns-based
- custom security-based
evaluating performance | benchmark types: absolute
- absolute return
- pro: simple
- con: not investable
evaluating performance | benchmark types: manager universes
- median manager performance from broad unverise of funds
- pros: measurable
- cons: no advance ID, ambiguous, uninvestable, appropriateness unverifiable, fund sponsors have to rely on 3rd party compiler, survivor bias (upwards)
evaluating performance | benchmark types: broad market indices
- broad mkt like S&P, etc
- pros: understandable, unambiguous, investable, measurable, in advance
- cons: often used when manager’s style is not the same (eg. small cap style)
evaluating performance | benchmark types: style indices
- specific portion of asset category (eg. large cap equities)
- pros: understandable, usually investable
- cons: some indices are not smart investment strats, mis-classifying style
evaluating performance | benchmark types: factor-model-based
- based on a factor model
- pros: insight into manager’s style by defining factors
- cons: unintuitive, difficult to calc, abiguous due to getting model correct
evaluating performance | benchmark types: returns-based
- regressing time series of manager returns against various style indices to determine composite style benchmark
- intuitive, meets benchmark criteria, use when only have acct returns
- cons: style indices may not reflect manager’s style, enought data, not work if manager chgs style
evaluating performance | benchmark types: custom security-based
- composite of manager’s allocations and investment process • meets benchmark criteria, continual monitoring, clear definition of allocation exposure (for sponsor allocation ease)
- cons: expensive, impossible if lack of manager transparency (hedge funds)
evaluating performance | benchmark types: constructing custom security-based
- ID important elements of manager’s process
- select securites based on the process
- weight the securities based on the process
- review and adjust
- rebalance on predetermined schedule
evaluating performance | test of benchmark quality
- systemic bias
- tracking error
- risk characteristics
- coverage
- turnover
- positive active positions
evaluating performance | benchmark test quality: systematic bias
• if corr of portfolio to benchmark are not close (ie beta ~ 1), then portfolio may have different factors than benchmark
• where A = P - B, S = B - M, P = portfolio, B = benchmark, M = market
◎ active return (A) should not be corr with style return (S)
◎ S should be corr with P-M
evaluating performance | benchmarket test quality: tracking error
std dev of P-B < std dev of B-M
P = portfolio
B = benchmark
M = market
evaluating performance | benchmarket test quality: risk characteristics
systemic risk exposure should over time be the same for portfolio and benchmark
evaluating performance | benchmarket test quality: coverage
mkt val of securities in both portfolio and benchmark / mkt val of securities in portfolio
evaluating performance | benchmarket test quality: turnover
% of benchmark’s total mkt val that is bot/sold during rebalancing
evaluating performance | benchmarket test quality: positive active position
- active position = portfolio weight of the security - benchmark weight of the security
- absence of security in portfolio >> neg active position
- large number of neg active positions is bad
evaluating performance | benchmarks for hedge funds
• difficult
◎ lack of transparency
◎ absolute return
◎ no definable style
◎ skewed returns (problem with sharpe)
• 3 possibilities:
◎ value-added return: benchmark weights sum to zero
◎ separate long/short benchmarks that are then combined
◎ Sharpe ratio
evaluating performance | performance attribution: 2 types
• macro:
◎ fund sponsor level
◎ % or dollars
◎ measure decisions of sponsor
• micro:
◎ portfolio manager level =measures value added by manager
• CFA test focuses on macro
evaluating performance | macro performance attribution: inputs
- policy allocations
- benmarket returns
- portfolio returns, valuations, external cash flows
• when using %, returns are calced at individual manager level
• when measuring in dollars, portfolio valuation and external CF are required to eval sponsor’s policies
evaluating performance | macro attribution analysis: 6 layers (levels)
- net contributions
- risk-free asset
- asset categories
- benchmarks
- investment managers
- allocations effects (plug number)
evaluating performance | macro attribution analysis: 6 layers: risk-free asset
- layer 2
- calcs the risk-free portion of the return on the beginning balance
evaluating performance | macro attribution analysis: 6 layers: asset categories
• layer 3
• calcs the asset clase return - risk-free return
• R = sum(Wa * (Ra - Rf))
◎ a = asset class
◎ R = return
◎ W = weight
evaluating performance | macro attribution analysis: 6 layers: risk-free asset
• layer 4
• calcs benchmark return - asset class return
• sum(Wa * Wb * (Rb - Ra))
◎ a = asset class
◎ b = benchmark return for each manager
◎ W = weight
◎ R = return
evaluating performance | macro attribution analysis: 6 layers: active management
• layer 5
• calcs active return - benchmark return
• sum(Wa * Wm * (Rm - Rb))
◎ a = asset class
◎ m = each manager
◎ W = weight
◎ R = return
evaluating performance | macro attribution analysis: 6 layers: allocation effects
- layer 6
- plug number to make: begin value + layers 1-5 + allocation effects = end value