1 - Introduction Flashcards
If markets are efficient, only way to get higher returns is….
Taking on more risk
Proofs against EMH
Arbitrage Opportunities
Compensation for collecting information (think not everybody has access to all info)
Market crashes and bubbles (“overpricings”)
Trading volume and volatility too high (constant mispricings?)
Existance of investment funds and quant strategies
What is Systematic/Quant investing?
Strategies with high returns (seem to in backtests) without taking on too much risk
Constant portfolio rebalancing (systematic), based on observable characteristics
Stat used for Alpha statistically significant?
t-stat > 2
Sharpe Ratio =
E (R - Rf) / vol (R - Rf)
Information Ratio =
alpha / vol. residuals (in a model)
Information Ratio (against benchmark) =
E (R - Rb) / vol. (R - Rb)
Sortino Ratio =
E (R - Rf) / vol. downside
vol. downside =
vol for returns below MAR (minimum acceptable return)
Drawdown =
(HWM - P) / HWM
Hedge Fund Leverage =
Long Positions / NAV
Gross Leverage =
(Long + Short Positions) / NAV
Net Leverage =
(Long - Short Positions) / NAV
P&L =
(Return x Investment on Long) - (Return x Investment on Short) + Financing
Financing =
Cost of loan from Prime Broker to support Longs
Interest earned on cash collateral (shorts) held by securities lender (rebate)
Interest earned on additional cash in money market products
Robustness Tests
In-sample vs. Out-of-sample
Cross-Section: Assets classes and Geographies
Parameters: Robust to changes in parameters? think interest rate
Backtest vs. Live Performance: Transaction costs and fees
Capacity constraints: Performance decreases with more AUM, think market movements
Other practicalities in Evaluating Trading Strategies
Geometric vs. Arithmetic returns
Estimating future volatility (“easy”) and correlations (“hard”)
Annualizing performance
High Water Mark (can only charge fees when above it)
Adjust for liquidity (regressing considering lagged periods of entrance)
Needs to perform Backtest
Universe of securities
Signals
Trading Rules:
Objective function (maximize this, minimize that)
Frequency of trading
Implementation Shortfall (bid-ask spreads)
Constraints (position size, country, industries. often imposed by clients)
Risk Model: Target risk level and rebalanced based on it.
Transaction costs
Data and time lags in backtesting
Data must be available at time (account for release dates)
Not realistic to assume you can trade at closing price
More realistic to use closing price one or two days later
Signal Research:
Credible Hypothesis:
- Rationale
- Why market not pricing already?
- What information inefficiency or liquidity provision are we targeting?
Test Economic Mechanism:
- Is rationale risk-based or friction-based?
- Can signal forecast future fundamentals and not only returns? Are other participants making systematic errors?
Additivity:
- Does it add anything beyond market factors? Value, momentum, carry…
Measure of Transaction Costs:
TC (volume-weighted average price) =
Price execution - Price vwap
Compares price at which you traded to average price at which other people traded on the same day
Investors tend to compare earnings yield to bond yield. What does this ignore
Earnings Yield -> Real Returns (account for inflation)
Bond Yield -> Nominal Returns
What are cyclically adjusted earnings?
Account medium-term fluctuations in terms of earnings
Shiller’s CAPE takes 10-yr average of inflation adjusted earnings
Bond returns =
Yield to Maturity - Capital Appreciation due to Yield Change
YTM = yield if everything goes “normal”
Cap. App. = Modified Duration x Yield Change
Mod. Duration = Bond sensitivity to rate changes, how much bond valuation changes if interest rate up 1%
This is for the case we want to sell the bond, if we’re just holding to maturity, excess returns will go down, but impact isn’t that big.