week 8 Flashcards
Investment Cycle
o Asset Allocation
o Portfolio Construction
o Implementation
o Portfolio Attribution
o Asset Allocation
Distributing investment dollars across stocks, bonds, cash and other investment vehicles in order to achieve a target level of return within a specified level of risk exposure and tolerance
o Portfolio Construction
Selecting the actual instruments to hold in each asset classes
o Implementation
Selecting an appropriate broker/dealer
Execution
Now inc. specification algorithms and algorithmic trading rules
o Portfolio Attribution
Performance analysis
• Evaluate performance of the fund to distinguish between market movement and skilled decision-making ability
Transaction Cost Analysis
Difficult to measure trading costs independent of investment decisions
Why Transaction Cost Analysis
o For buy side traders; to determine whether brokers are working effectively on their behalf
o To confirm they are getting value for their brokerage commissions
o To determine whether
they should be trading more or less aggressively
Trading aggressively leads to different transaction costs
o To better understand why they are competing with other traders who are trading on the same information
o To inform their portfolio managers about liquidity conditions in various securities
Evaluating Investment portfolio Performance
o Measure the increase in asset or portfolio value relative to the initial asset value during a given time frame o Returns can be measured Dollar-weighted return Time-weighted average return Need to think of portfolio
o Returns adjusted by or compared to a relevant benchmark
o Refers to stock selection, based on these ratios we will evaluate the stock performance
Sharpe ratio
Treynor Ratio
Jensen Measure
Evaluating Investment Portfolio Performance
o Market timing refers to the ability of the investor to shift their portfolio’s betas
o Suppose we have a risk-free asset and risky asset
Market downturn -> allocate more money in the risk-free asset
• Shifting returns closer to Rf
Market upswing -> more money allocated into the risky asset
• Increase portfolio beta
Ways to examine market timing
o Quadratic variable approach Increase the beta when the market is good, decrease when its bad Yg is expected to be positive o Dummy Variable approach D=1 when its in a bad market
Do we expect gamma q to be negative when the market is bad?
• Quadratic variable approach
Expect gamma q to be negative in a positive market
Transaction cost components
o Explicit costs Contractual costs which include • Brokerage commissions • Clearing and settlement cost • Taxes / stamp duties Easiest to measure
o Implicit costs Price effects • Bid/ask spreads • Market impacts Hard to measure
o Missed trade opportunity costs
Informed traders lose profits when their orders go unfilled
Unfilled orders may also hurt uninformed traders
Hard to measure
o To measure these costs, need to estimate the price without the trade occurring
Two approaches
o Benchmark comparison
Difference between the benchmark price and traded price
Benchmark price: arrival price, VWAP, close price and others
Many different benchmark prices
o Implementation shortfall
Most effective means of measuring transaction costs
Involves assessing the impact of trading on portfolio returns by computing the difference between the net returns on a paper portfolio and those on a real portfolio
o Looking at these two, they are essentially the same as it is comparing two prices
A general transaction cost measurement framework
o Difference between the trade price and a price benchmark may contain information about transaction cost
o For a buy
TC/Unit = TP-BP
o For a sell
TC/unit = BP-TP
o The ideal price benchmark would estimate the price that would have been observed if not for the trade
Your trade is not included
No one can confidently specify a price not including your trade
Common Price Benchmarks
o Arrival price – spread midpoint at the time of trade
o VWAP – Average price around the trade
o Closing or opening prices
The Spread Midpoint Benchmark Price
o Estimate the benchmark price by the average of the bid and ask at the time of the trade
o Refers to the arrival price
The cost of a buy at the ask (or a sell at the bid) is one half of the spread
o Issue: the spread midpoint benchmark does not indicate whether the trade is well timed
o Doesn’t necessarily show that you traded at a better price
o Other issues
Large orders may require many trades between which quotes typically trade
Total transaction cost will be underestimated if a different midpoint benchmark is used for each trade
Brokers may defer the trade until it can be done when the spread is narrow
• May end up with worse trade price for the investor
Volume-weighted average price
o The price benchmark is the volume-weighted average price for the day of the trade
o Negative transaction cost in this scenario with vwap shows positive returns
You are trading at a lower price than the average traded price for the day
o Advantages
Simple summary statistic of all trades in a day
Transaction cost measure can be computed from volume-weighted average cost trade confirmations
o Issues
If you are the only buyer (or seller) in the interval
• Transaction cost estimate is zero
The selection/implementation decomposition is burred
• Momentum traders estimate positive costs
• Contrarian traders estimate negative costs
Inherent drawbacks of VWAP
• Trading strategy induces it to be positive/negative if you are a momentum/contrarian trader
Momentum/Contrarian
o Buy in an upward trend
o Sell in a downward trend
o E.g. join the trading session at a later time when there is momentum in an upward trend
At the end of the day, you most likely end up with a positive cost due to the average trading price being very low at the start of the day
o Contrarian
Buy in a downward trend, sell in an upward trend
The closing price benchmark
o Estimate transaction costs by the difference between the trade price and the closing price on the day of the trade
o Easy to observe
o Closing price advantages
Closing prices are obtained easily
o Issues
Informed traders estimate negative transaction costs
• Selection and implementation decomposition blurred
• Informed trader knows where the price will end up -> have a negative flow
• You are not a good trader, you just know the private information
Brokers can execute orders only at or very close to the close to have zero transaction cost estimates
• Brokers tend to trade at the end of the session because they know that’s their benchmark
• Tend to have lower transaction costs due to trading price being very close to the market close price
Implementation shortfall measures
o Transaction costs, if the order trades
o Lost opportunity costs – if the order does not trade
Failure to submit realistic orders
o Timing
The change in quote midpoint between decision and first execution
o Market impact
The difference between the first execution quote midpoint and the average trade price
o Opportunity cost
The loss suffered if the order did not execute
o Price impact
Trade amount * (Trading price – first trading price)
Implementation Shortfall: advantages and disadvantages
o Advantages
The implementation shortfall separates portfolio selection performance from implementation performance
• Does not favour brokers who handle informed order flow
Brokers cannot game the implementation shortfall measure
o Issues
The implementation shortfall requires decision time and order size data
• These data are not generally available
It estimates large opportunity costs for unrealistically large orders
General observations
o Statistical properties of measures
All transaction cost estimation methods produce noisy results when applied to single trades
o Average transaction costs, measured over many transactions, are more reliable indicators
Averaging does not solve bias or gaming problems
o Traders who offer liquidity tend to have negative transaction costs
o Only implementation shortfall method measures lost opportunities of trades not done
Evaluating Trading performance
o Brokerage evaluation requires baseline data for similar problems
Difficult problems are more expensive to solve than simple ones
Comparisons across brokers are meaningful only when they are given similar trading problems
o Expected trading costs should be integrated into portfolio composition decisions
The trader should advise the manager of what is feasible
o A well-informed trader should be willing to incur transaction costs
