Reading 3.3 Flashcards
A factor (in multifactor pricing model) represents a
Factors should have a ____ with ____, and should show that the ____ over long periods.
unique source of return or premium that is not highly correlated with other factors (i.e., the return cannot be fully explained by other factors).
sound economic foundation
rigorous supporting research
premiums persist
multi-factor models are more successful at explaining ____ and generally produce better estimates of _____
systematic returns
idiosyncratic returns
Idiosyncratic returns refer to
the portion of an investment’s return that is attributed to factors specific to that particular investment, rather than general market movements.
Multi-factor models may be expressed in ex-ante form, describing expected returns, as:
Multi-factor models may be expressed in ex-ante form, describing realized returns, as:
Factors used in multi-factor models may include
- market portfolio
- spread between small and large stock returns
- spread between liquid and illiquid stock returns
Multi-factor models are primarily used for:
estimate the variance-covariance matrix of asset returns in order to inform investors about the portfolio exposures
What is MARGINAL UTILITY in investing?
The utility of gaining more capital
why investors tend to derive higher marginal utility from returns when the market performs poorly, compared to when it performs well?
Investors tend to derive higher marginal utility from returns when the market performs poorly compared to when it performs well due to the concept of diminishing marginal utility.
Diminishing marginal utility suggests that as individuals acquire more of a good or benefit, the additional satisfaction or utility gained from each additional unit decreases.
There are 3 main categories of factors that drive asset returns & give examples for each:
- Macroeconomic factors - drive asset returns throughout the entire economy and across asset classes. Examples: productivity, inflation, credit, economic growth, and liquidity.
- Fundamental, style, investment, or dynamic factors - based on companies’ fundamental traits that are empirically identified as key drivers of investment returns across companies, industries, and sectors. Examples: value, size, momentum, quality, and low volatility.
- Statistical factors - drive asset returns within an entire economy, asset class, or sector, and are identified based on empirical traits (rather than style or economic traits). For example, statistical factors have been identified in bond returns.
- Principal component analysis is an approach that identifies statistical factors. For instance, it may find that most return differences between assets are explained by 3-5 components.
However, with this analysis, the factors’ economic identity and the economic cause of the relation between the factor and asset returns are not typically known.
2 classifications of asset pricing models
- Theoretical models - Factors are derived from known relations and facts about financial economics. There is typically a logical explanation for theoretically derived factors.
- Empirical models - Factors are based on historical observation and analysis (using sound statistical techniques).
► For instance, a model may describe stock returns in terms of market-to-book ratios, based on observing a correlation between asset returns and market-to-book ratios.
Setting up an empirical return model involves the following 3 steps:
- Calculate the excess return (by subtracting the risk-free rate from the security’s historical return) to be used as the model’s dependent variable.
- Select potential factors to be used as the independent variables.
- Use statistical analysis to identify the factors that are significantly correlated with the returns.
A tradable asset is a ___
position that can be easily established and liquidated.
The Fama-French model expands the single-factor CAPM and describes asset returns in terms of three factors:
- Market portfolio - spread between returns of the market portfolio and the risk-free asset (same as CAPM)
- Value (book-to-market) - spread between returns on high book-to-market ratio stocks and returns on low book-to-market stocks.
- Size - spread between the returns on small stocks and those on large stocks
ex-ante form of the Fama-French model is given by (formula)
The Fama-French model (1992) is an ____ model used to explain returns of ___
empirical multi-factor
traditional equities and equity-oriented alternative investments
Fama-French-Carhart model (Carhart 1997) extends the Fama-French model by adding a ____. This is motivated by the empirical observation that stock performance may be explained by whether the stock has ____
momentum factor
recently increased or decreased in value
The ex-ante form of the Fama-French-Carhart model is given by (formula)
The Fama-French and Fama-French-Carhart models describe historical equity returns better
than single-factor models
Fama-French five-factor model adds 2 factors to the original Fama-French
model:
- Robust minus weak factor - It distinguishes firms by their reported accounting profitability, with robust firms having higher accounting profits as a proportion of equity.
- This factor represents “the average return on … robust operating profitability portfolios minus the average return on … weak operating profitability portfolios”.
- Conservative minus aggressive factor - It distinguishes firms by the rate of reported corporate asset investment, with conservative firms exhibiting lower rates of investment in corporate assets.
- This factor represents “the average return on … conservative investment portfolios minus the average return on … aggressive investment portfolios”.
Why are the Fama-French models are not appropriate for alternative assets that are not focused on public equities?
because alternative assets have factor exposures that differ considerably from
those of traditional assets
There are two issues to consider when using empirical multi-factor models for return attribution or forecasting expected returns:
- FALSE IDENTIFICATION OF FACTORS - Randomly identified factors cannot be used to predict future returns. Factor identification needs to be based on theoretical reasoning and careful statistical testing
- DIFFERENTIATING FACTOR CORRELATION FROM FACTOR CAUSATION
For many alternative investments, investing in a market-weighted portfolio of alternative assets is not possible since many alternative investments are ____ and, thus, their ____ risks are not easily diversified. This suggests that expected returns of alternatives likely depend on ____, so a ____ model cannot adequately describe these returns
privately held
idiosyncratic
multiple factors
single-factor
Factor investing is a relatively new asset allocation approach that allocates based on ____ as opposed to the ____
exposure to certain risk factors
traditional allocation to asset classes
Ang (2014) presents 3 important observations related to factor investing:
- Factors matter (not assets) - they are key to factor investing. Asset allocators should view assets merely as a means of accessing factors.
- Assets are bundles of factors (that reflect different risks and rewards) = asset’s risk premium represents a package of risk premiums offered by factor exposures and can be estimated using the factor exposures.
- Different investors should use different factors
Idiosyncratic risk refers to
the risk that is specific to an individual asset or investment and cannot be diversified away by investing in a larger portfolio. It’s also known as unsystematic risk or specific risk.
Two well-recognized equity factors are:
1) Momentum - constructed by building two equally weighted portfolios based on equity prices, one with stocks that performed well during a period and one with stocks that performed poorly during the period. This procedure is repeated each period to attain two return series.
► A strategy of going long the portfolio of winners and short the portfolio of losers results in the return to the momentum factor. This is a new factor if the return to this strategy cannot be explained by other factors.
2) Value factor
► The value factor is constructed in a similar way except the portfolios are constructed by dividing stocks based on the ratio of their book values to their market values, where stocks with high book-to-market ratios are value stocks, and those will low book-to-market ratios are growth stocks. Two equally weighted portfolios are built each period, one with stocks with above-average book-to-market ratios and the other one with stocks with belowaverage book-to-market ratios.
► A strategy of going long the portfolio of above-average book-to-market ratios and short the other portfolio results in the return to the value factor.
3 important observations about the momentum and value factors:
- Both factors have significant positive mean returns over the long time period.
- Both factors are volatile, so they may periodically generate negative returns.
- The two factors are negatively correlated and are not highly correlated with market risk.
10 well-known risk premiums identified by academic and industry research (and strategies for investment using these premiums)
- Value premium - strategy involves buying stocks with high book-to-market ratios and shorting stocks with low book-to-market ratios.
- Size premium - strategy involves buying small-cap stocks and shorting large-cap stocks.
- Momentum premium - strategy involves taking long positions in past winners and short positions in past losers.
- Liquidity premium - strategy involves buying illiquid assets and selling similar but otherwise liquid assets.
- Credit risk premium - strategy involves buying bonds with low credit quality and selling bonds with high credit quality.
- Term premium - strategy involves buying longterm bonds and selling short-term bonds.
- Implied volatility premium - strategy involves being short the implied volatility of options (e.g., create a market-neutral position using short positions in out-of-the-money puts and short positions in stocks) and being long the
realized volatility of options by delta hedging the underlying stock. - Low volatility premium - strategy involves buying low volatility (or beta) stocks and selling high volatility (or beta) stocks.
- Carry trade - strategy involves buying bonds denominated in currencies with high interest rates and selling bonds denominated in currencies with low interest rates.
- Roll premium - strategy involves buying commodities that are in backwardation and selling commodities that are in contango.
On what is the implied volatility premium strategy based?
on the observation that implied volatility tends to exceed realized volatility, which creates an opportunity for volatility sellers
Contango & Backwardation definition (in futures market)
A better approach than passively allocating to factors is to
apply tactical asset allocation to factors, assigning higher weights to factors that offer better risk premiums
factors that generate attractive returns in good and bad times are not
factors but arbitrage opportunities
academic studies have shown that factors that perform poorly in bad times
generate ___ returns in normal times
attractive
momentum crashes are?
when assets that recently overperformed experience extremely poor performance relative to other assets
While many factors are tradable, they may not be fully ____
implementable
what is alpha?
Excess returns earned on an investment above the benchmark return when adjusted for risk
The alpha of an investment product (factor investment strategy) can be estimated as the
intercept of a regression of the product’s excess return (excess return compared to the benchmark) against the returns of traded factors (investment into a certain factor.
What are the 2 issues related to factor investing and why do they occur?
- No benchmark for passive factor investing, because:
- factor investing involves long/ short strategies
- unclear how a diversified portfolio of factors should be passively weighted (e.g., equally or based on volatility). - No passively managed factor portfolios, because
- direct factor investing requires factor-based portfolios to be actively managed
- Investors may adopt a buy-and-hold strategy if enough products (e.g., ETFs) are available that replicate each factor, but the portfolio still needs to be regularly rebalanced.
Why is Direct factor investing is not possible for all institutional investors?
Because some can’t take short positions
The merger arbitrage strategy exploits a factor related to the
risk of a merger being completed
Convertible arbitrage strategy exploits a form of the
implied volatility factor
Equity long/ short and market-neutral strategies have significant exposures to
equity market factors
Global macro strategies use the
carry trade factor
if all factors are traded and no short sale constraints exist, a portfolio constructed using risk allocation would be unlikely to ____
consistently outperform a portfolio constructed using asset allocation
4 practical issues associated with risk-factor-based asset allocation
- NOT SUSTAINABLE: Portfolio construction using return factors is unlikely to become implemented globally because some of the allocation strategies are not sustainable (i.e., not consistent at the macro level).
- For instance, all investors cannot short growth stocks or short commodities in contango. Therefore, the capacity is likely to be limited, and as more capital is allocated to factor investing, the strategies will become expensive and the risk premium will decline or disappear.
- EXTREME POSITIONS: Risk allocation entails extreme positions in some asset classes (e.g., shorting all growth stocks), which many institutional investors are not permitted to make.
- HIGH COSTS: Risk allocation does not necessarily result in asset allocations that have better returns than allocations based on asset classes. The cost of the strategy should be taken into account.
- UNREALIZABLE for ALTS: Pure risk allocation cannot be applied to alternative investments, since most alternative investments represent a bundle of risk factors.
Information about alternative assets’ risk-return profiles can be attained by analyzing ____ and this information can be valuable when allocating to ____.
Give an example:
factor exposures of alternatives
traditional assets
EX: factor exposure analysis of PE reveals its exposure to size and credit factors and, as a result of this information, investors may tilt their traditional portfolio’s exposure to other factors.
Describe the Adaptive Markets Hypothesis (AMR), proposed by Lo (2004).
What are species in the AMR?
Explains the evolution of markets based on principles of evolutionary biology, where market dynamics are driven by competition, mutation, reproduction, and natural selection.
Species are distinct groups of market participants: e.g., pension funds, retail investors, hedge funds.
According to the AMH, profit opportunities exist when ____
Explain how increase in the competition affects the market
more resources are available and competition is low
As competition increases, players with a competitive advantage adapt to survive and those that are not able to adapt perish, which reduces the level of competition and starts the cycle again
There are four key practical implications of AMR view on markets:
- Time-varying risk premiums = Risk premiums vary over time, and changes in risk premiums may be predicted using technical and fundamental analysis => strategies that allocate dynamically across time may be able to capture risk premiums.
- Market efficiency is relative = there are varying degrees of efficiency at different points in time and for different market participants => during periods when markets are less efficient, strategies that exploit market efficiencies can find opportunities.
- Adapting for success and survival = Trading strategies must be adapted as the market environment evolves
- Inevitable degradation of alpha - As a result of innovation and competition, consistent alpha opportunities degrade over time and alpha becomes beta = highly adaptive strategies may find fleeting alpha opportunities
Time-varying volatility refers to
returns with varying (i.e., non-constant) levels of true return volatility.
Equity market volatility is ____
fairly predictable
Volatility is partially predictable with ___ analysis: volatility is persistent and thus tends to be higher immediately after a ____ than after a ____
technical
high-volatility period
low-volatility period
What are the implications of the fact that Equity volatility is negatively correlated with market returns:
HIGH volatility typically occurs when market returns are negative and LOW volatility typically occurs when market returns are positive.
There are two popular approaches to modeling time-varying volatility as a stochastic process.
- Heston model (1993) = model’s key assumptions are that volatility is a continuous stochastic (random) process and it reverts towards a long-term mean.
- similar to Brownian motion (Weiner process) used for equity returns in the Black-Scholes option pricing model - Bates model (1996) - describes volatility as a stochastic process and includes a jump process for the asset price (i.e., permits the price to jump with random magnitude and at random time intervals).
Explain Stochastic modeling
Stochastic modeling forecasts the probability of various outcomes under different conditions, using random variables that account for certain levels of unpredictability or randomness.
Markov process =
process in which new returns are independent of previous returns
Hamdan, Pavlowsky, Roncalli and Zheng (2016) finding:
when assets with negatively skewed return properties are combined into a portfolio => the volatility can decrease, however, the skewness can become even more negative
Present value of an expected cash flow using a TRADITIONAL discount factor may be expressed as:
Stochastic discount factors (or ____) are
pricing kernels
different discount rates used in valuation models to discount different cash flows
STOCHASTIC DISCOUNT FACTORS PRESENT VALUE FORMULA
negative, normal and positive skewness graphs
In a multifactor world with both traditional and alternative investments, portfolio components have different types of risk that require multi-factor models that recognize the following:
- Cash flows in good times must be valued differently than cash flows in bad times.
- Investors with different constraints, time horizons, liabilities, and illiquid asset holdings need to value potential cash flows with different risk premiums.