Definations Flashcards
(78 cards)
1
Q
Asset allocation
A
Refers to the choices over the way money in a fund will be allocated to different asset classes.
Eg: -To maximise return for a given level of risk
-Spreading assets can provide downside protection if one asset performs bad
2
Q
Security selection
A
Refers to choices within asset classes
Eg: -What securities are included in the fund
-According to the objective of the portfolio
3
Q
What type of asset classes are cash
A
- Notes and coins
- Bank deposits
- Government bonds
Advantages
- High liquidity
- Safe (to some extent)
Disadvantages
-Low return
4
Q
Equity
A
Types:
- Domestic
- Developed nation
- Emerging markets
Offers potentially higher return than cash but also higher risk
5
Q
Fixed income
A
- Income at regular interval
- Set at a particular rate (does not vary according to inflation etc)
- Eg: bonds
Types:
- Domestic
- Developed nation
- Emerging markets
6
Q
Nominal equity return
A
Nominal is the return before taking out inflation.
High nominal returns don’t necessarily translate into high real returns
7
Q
Real equity returns
A
The return after taking out inflation.
8
Q
Alternative investment
A
Encompass investments outside the standard asset classes of cash, equity and bonds.
Eg: private equity, art, commodities, et.
9
Q
Investment companies
A
Financial intermediaries that collect funds from individual investors and invest in a wide range of securities.
Each investor has a claim to the portfolio established by the investment company in proportion to the amount invested.
Types:
- Open-end (Mutual funds)
- Close-end
- Exchange-traded funds
- Hedge funds
- Sovereign wealth funds
- Private equity
10
Q
Asset Net Value (NAV)
A
The value of each share in the investment company.
NAV= (Assets-liabilities)/#of shares
11
Q
Close-end funds
A
Number of shares in open-end funds respond to fund inflows and outflows.
Shares trade at NAV
12
Q
Close-end funds
A
Number of shares are fixed that trade intra-day on stock exchange at market determined prices.
They do not issue redeemable shares.
13
Q
Closed-end Premium or discount
A
Percentage difference between the price and the NAV.
The size of the discount depends on the transaction cost or the hold-out problem (to arbitrage the discount, one must buy back all the shares in a fund from the shareholders. The last few shareholders will realise this and refuse to sell at the market price and demand a premium).
14
Q
Exchange traded funds (ETFs)
A
Index tracking open-end funds traded on exchanges (until the exchange is open)
Track stock, commodities, fixed income and indices.
Trade at NAV as the number of shares in the fund are not fixed.
15
Q
Hedge funds
A
Open to high net worth individuals and institutional investors.
Investors subject to lengthy lock-up periods during which the investment can not be taken out.
Lock-ups allow the fund to invest in less liquid assets.
Generally private partnerships and are thus subject to less stringent regulations (can peruse strategies that are not permitted by standard investment companies, like short selling and derivatives)
16
Q
Sovereign wealth funds
A
They are country level investments.
Invest in state savings, often derived from commodity exports or export surpluses.
17
Q
Diversified monetary authorities
A
Countries that have large pools of investment funds but choose not to create separate sovereign fund vehicles.
18
Q
Private equity
A
Any type of equity investment in which the stock is not freely tradable on a public stock market.
These funds either buy all the private equity themselves or buy all the equity in a publicly traded company and make it private.
Majority investors are institutional.
Investments made are typically illiquid.
19
Q
Pension fund assets
A
Consist mainly of employee contributions to own-company pension schemes.
20
Q
Efficient market hypothesis
A
Claims that the current price of a security reflects all relevant information.
Prices reflect information until the marginal cost of obtaining information and trading no longer exceeds the marginal benefit.
Forms:
-Weak
Prices reflect all information contained in historical data. (History of prices, history of trading volumes)
Hypothesis states that if historical data conveyed signals about future performance, all investors would have learnt to exploit signals.
-Semi-strong
Prices reflect all publicly available information regarding the firm concerned.
Prices reflect historical data, data from a firm’s prospectus, production lines, quality of management, patents held.
-Strong
Prices reflect all information available regarding the firm concerned public or private.
Prices even reflect information only available to company insiders.
Quite extreme.
Seems to require insiders to engage in insider dealing, which is illegal.
21
Q
Analysis
A
Technical analysis:
Using prices and volume information to predict future prices trend analysis.
Inconsistent with ???? form efficiency.
Fundamental analysis:
Using economic and accounting information to predict stock prices.
Inconsistent with ??? form efficiency.
22
Q
Return predictability
A
Test to see if stock returns can be predicted by using publicly available information.
Things to look at:
-Seasonal patterns
Calendar anomalies—a number of studies have found that stocks earn higher returns at certain times of the year. Returns are higher in January(January effect for small stocks), weekends (up on Friday and down on Monday), time of the day (first 45 minutes and last 15 minutes), end of month, and holidays (last trading day before a holiday).
-Predicting returns from past returns
This is by measuring the serial correlation of stock returns.
Positive serial correlation means that positive returns tend to follow positive returns. Momentum type of property)
Negative serial correlation means that’s positive returns tend to be followed by negative returns (a reversal of correction property).
Conrad and Kaul (1988) and Lo and MacKinlay (1988) examined weekly returns of NYSE stocks and found positive serial correlation. However, this study demonstrates weak price trends over short periods but the evidence does not clearly suggest the existence of trading opportunities after transactions costs.
Jagadeesh and Titman (1993) examined the reruns of recent winners and losers and sorted them on past three to twelve months performance. Compared their performance with the next three to twelve months and found out that winners remained winners and losers remained losers (momentum effect).
De Bondt and Thaler (1995) found out that losers eventually outperformed winners in the following three years at an average of 25%.
-Predicting returns from firm characteristics
Certain easily accessible form characteristics may br useful to predict future abnormal returns. Eg. Firm size, price earnings ratio, book to market ratio.
Fama and French (1992) Shi’a that high book to market ratio firms (known as value firms) earn more on average than low book to market firms (known as growth or glamour firms).
Fama and French (1992) present a model with a market factor, book to market factor and a size factor. They say that the book to market ratio and the size factor capture the risk not taken into account by the CAPM model. (Risk firms that are small with high book to market ratios are risky and this risk is not captured by the beta of the CAPM model).
-Predictability of market returns
Several studies have documented the ability of easily observed variables to predict aggregate market returns.
Fama and French (1988) show that the return on the aggregate stock market tends to be higher when the dividend yield is high.
Campbell and Shiller (1988) find that earnings yields can predict market returns.
Keim and Stambaugh (1986) show that the yield spread between high and low grade corporate bonds can also help to predict broad market returns.
23
Q
Speed tests
A
Concerned with the speed with which information announcements are reflected in prices. (Efficient markets should reflect information in prices instantaneously).
Event studies are financial studies that allow us to normally test the impact of announcements on returns. (Focus on a particular event such as an announcement of a merger).
CAPM provides us with a way of calculating normal or expected returns.
Expected return= risk free rate+beta*market excess return
This difference between the return predicted by CAPM and the return observed after an announcement is used as a measure of abnormal returns.
Foster, Olsen, Shevlin (1984) found out a relationship between the size of the earnings surprise and the size of abnormal returns earned after the earnings announcement. Concluded that firms that make the most positive earnings surprise announcements have the largest possible drift in returns (information is slowly absorbed into prices). (Inconsistent with market efficiency).
24
Q
CAPM equation
A
E(ri)-rf=Bi[E(rm)-rf]
SML slope= E(rm-rf)
As we cannot see the expected returns, we might use realised average returns as a proxy for expected returns. Equation then becomes:
Avg(ri-rf)=Bi[Avg(rn-rf)]
25
SML in realised returns-beta space
The SML has slope equal to the average excess market return.
If CAPM holds points plotting security betas against average excess returns should plot along the realised return security market line.
Avg(ri-rf)=gama0+gama1Bi+error
To estimate the security market line regression (SML) we first need to decide on:
- which securities we wish to consider
- the relevant time window
After this we need two sets of input :
- betas
- average excess returns
26
Estimating betas
CAPM: E(ri)-rf=Bi[E(rm)-rf]
To estimate the beta of a security, we need to interpret the CAPM equation as a time series relationship.
Beta in the CAPM equation tells us how sensitive the expected security excess returns are to market excess returns. If we want to estimate this sensitivity, we can do this by tile series regression.
Problem: we can’t observe expected returns, so use realised returns instead.
Problem: we can’t see the market portfolio which consists of all the wealth that all households have invested including property, human capital, financial wealth, etc. So use a braid market indies like the FTSE-100 (assuming returns on all assets will be approximated by stock market returns).
We will therefore need to estimate the following equation:
ri,t-rf,t=alphai+Bi(rstock index,t-rf,t)+Ei,t
{alphai is the standard regression intercept and Bi is the standard regression slope}
This will provide us an estimate of the alpha which should be 0 if CAPM is true, and it’s beta which measures the security’s sensitivity to the market.
27
Estimating the SML
Lintner (1965) tested the CAPM by estimating betas using the annual data on 631 NYSE stocks fit 10 years, 1954-1963.
He found out that the intercept gama0 was highly statistically significant and 12.7%.
The average market risk premium was 16.5% per year. However, his estimate of using the SML, gama1 was 4.2% which is too low.
This, concluded that the SML was flatter than it should be.
Lintner’s SML diagram.
Lintner’s results meant that low beta stocks earned more than they should have according to CAPM and vice versa for high beta stocks.
Disappointing results.
28
Implications of measurement error in betas
Well known result from statistics that if the independent variables are measured with error then the intercept from a regression will be biased upwards, and the slope coefficient will be biased downwards.
Thus, measurement errors in betas in stage 1 of the regression may be biasing our estimates in stage 2.
29
Portfolios solution
Combining securities into portfolios diversifies away most of the firm specific part of returns enhancing the precision of estimates of beta in stage 1.
It can be shown that if there is a single factor driving returns, the variance of the portfolio residuals should be one twentieth the variance of the residuals of the average stock.
If the betas of the portfolios are clustered then we won’t have a good idea of the shape of the SML over the entire range of betas. The best thing to do then would be to construct portfolios where stocks are not randomly allocated but are ranked according to betas and then put into portfolios.
Diagrams for constructing portfolios based on size of beta.
30
Testing CAPM using portfolios
Fama and MacBeth (1973) use this methodology to test the CAPM using monthly data for every month January 1935- June 1968.
The estimate the following equation:
to= gama0+ gama1Bi+ gama2Bi^2+ gama3sd(ei)
- r0 measures the intercept and should equal rf
- r2 measures the potential non-linearity of the return
- r3 measures the explanatory power of non-systematic risk
Fama and MacBeth observed that gama3 (non-systematical risk) fluctuated monthly and was generally insignificant. This is consistent with the hypothesis that the non-systematic risk is not rewarded by higher average returns.
B^2 denoted by gama2 was insignificant which is consistent with the hypothesis that the expected return-beta relationship is linear.
Their estimated security market line was too flat. This can be see from the fact gama0-rf in the regression estimates is on average positive and that gama1 is on average less than the average excess market return.
31
Fama French three factor model
Average returns are higher for high book to market equity ratios and stocks of small firms than predict ted by the CAPM maybe because size and book to market ratio are proxies for exposure to sources of systematic risks not captured by the CAPM beta.
The systematic factors in the Fama-French model are size, book to market equity ratio and the market index.
The size factor is the differential return in small firms versus large firms. This factor is called SMB (Small minus Big).
The book to market factor or value factor is the return on high vs low ratio called HML.
Calculated monthly to generate monthly factor returns.
Model:
E(ri,t)-rf,t=ai+bi(E(rM,t)-rf,t)+siE(SMBt)+hiE(HMLt)
- coefficients bi,si and hi are the betas of the factors also called loadings.
- intercept ai should be 0 if these factors fully explain asset returns.
The model helps us to explain the cross section of asset returns.
The existence of a value premium and a size premium means that fund managers can improve their performance by tilting their portfolios in favour of value stocks or small stocks (but will not show their skills). It is therefore common when testing for whether a given fund manager is actually adding value to take away that component of his fund’s returns attributable to its size and value exposures.
32
How do we calculate returns on the size factor and book to market factor
1) Take all stocks in year 2000 (eg)
2) Sort these sticks into:
- bottom 30% of size stocks (small (2000) stocks)
- top 30% of size stocks (big (2000) stocks)
3) In year 2001, calculate the average return in 2001 on small and big (2000) stocks
4) In year 2001, sort stocks again.
5) In 2002, calculate the average returns on small and big stocks
6) Calculate the SMB In each year (2001 and 2002)
Calculate HML the same way
33
Testing Fama French model
Davis, Fama and French (2000) test the model
- Their six premium is the difference in returns between the smallest and the largest third of firms.
- Their book to market factor is the difference in returns between high and low book to market firms.
- For their market factor they take the value weighted return of all stocks traded on US national exchanges to compute the excess returns on the market portfolio relative to the risk free rate measured as a return on one month treasury bill.
They constructed 9 portfolios by sorting firm into three size groups (small medium, big) and three book to market groups (High, medium, low).
Nine portfolio matrix
For these 9 portfolios, DFF estimate the Fama French model as a first pass regression over the 816 months between 1929 and 1997.
do-rf=ai+bi(rM-rf)+siSMB+hiHML+ei
- Findings showed that the intercept of the regression are small and statistically insignificant.
- Large R^2 statistics indicate that returns are well explained by the 3-factor model.
- Can also be seen that the coefficients of the size and book to market factors are all generally significant.
34
Interpretation of the 3-factor Fama French
- Possibility that size and book to market factor proxy for risk are not captured by CAPM
- Another explanation attributes these premises to some sort of investor irrationality or behavioural biases
Risk based interpretation
Petkova and Zhang (2005) try to explain why high book to market stocks earn more on average than low. Their innovation is to allow for time varying betas. They show that high book to market stocks are more exposed to the market in market downturns and low exposed in market up states. This is because value firms on average have greater amounts of tangible capital. Value firms have a lot of investment already in place and it is difficult for them to reverse out of their current situation (downturn) (overcapitalised). Growth firms can deal better with downturns by deferring growth plans. In booms, growth firms will respond strongly by increasing their capacity.
Behavioural interpretation:
- Believe growth firms have recently performed well so it will continue. Recent good performance is overweighted (market value inflated). However, growth firms tend to do worse because of analyst over-optimism.
- Chan, Karceaki and Lakonishok (2003) make a case of over optimism.
35
Momentum
Jegadesh and Titman (1993) show that stocks with high recent performance continue to earn higher average returns over the next 3-12 months then stocks with low recent performance.
To show this they take all stocks that trade in the US markets and sort them by prior 6 month return.
They show that momentum cannot he explained by market risk or Fama French factors.
Harvey and Siddique (2000) argue that momentum may be a proxy for soreness in returns as momentum winner portfolios have more negative skewness than loser portfolio (they fall later)
Pastor and Stambaugh (2003) argue that momentum may be a proxy for liquidity risk.
36
Carhart’s 4-factor model
Carhart (1997) adds an additional factor to the findings of Jegadesh and Titman to add the three factors of Fama-French. This factor is called MOM (Momentum) and it captures Jagadesh and Titman momentum anomaly.
ri,t-rf,t=ai+bi(rM,t-rf,t)+siSMB+hiHML+miMOM+ei
His momentum factor is constructed monthly as the equally weighted average return of firms with the highest 30% prior year return minus he equally weighted average returns of firms with the lowest 30% prior year returns.
Claims that his pricing model improves the pricing error of the three factor model.
37
Geometric Average return
(1+rG)^n=[(1+r1)x(1+r2)+...(1+rn)]^1/n
38
Arithmetic average return
(x-y)/n
39
Time-weighted returns
Giving equal weights to each unit of time that a return is earned.
Money weighted returns take into account the amount of money we have in each period.
If we take money out before good returns, then the money weighted returns respond by decreasing and vice versa for poor returns.
Money weighted returns measure investor timings and not manager timing.
Calculated as the same way as Internal Rate I’d Return (IRR)
AUM at beginning+PV(inflows)=PV(outflows)+AUM liquidation value (AUM at the end of year)
PV= (flow)/(1+MWR)^t
40
Peer group adjustments
Evaluating performance based on average return is not sufficient. Returns must be adjective for risks before meaningful comparison. This can be done by comparing their rates if return by those of other investment funds with similar characteristics. (High yield bonds with other fund’s high yield bonds).
After doing so, the average returns of each fund within the universe are ordered and each portfolio manager receives a percentage ranking depending on relative performance within the comparison universe.
Chart.
However, these rankings ca be misleading as there can be a large difference in the risk of funds within a group. Also, it may but always be easy to determine the right comparison peer group.
Cooper, Giles and Rau (2005) examined the gaming of mutual funds sectors by US mutual funds. They found out that the funds may change their names to make them look good or to be in sectors that are currently getting a lot of flow at present.
41
Sharpe Ratio
A method of measuring risk adjusted performance.
[Avg(rp)-rf]/sdp
It can be understood as a rewards to risk ratio.
Numerator
-reward (risk premium)
Denominator
-risk
If we plot portfolios in expected return-volatility space, the ratio of portfolio P denotes the slope of the capital allocation line through P. The capital allocation line for any portfolio denotes the set of expected return and risk combinations.
42
M^2=RAP measure
It focuses on total volatility as a measure of risk proposed by Graham and Harvey (1994). It has the easy interpretation of of a differential return relative to the benchmark index.
How to calculate:
- calculate weights first by dividing Standard deviation of the market by the standard deviation of the portfolio (sdmarket/sdportfolio). This is for p.
- Then (1-p) for risk free assets.
- then (risk free rated risk free weight)+(pxaverage return of portfolio)
Assumptions (also for Sharpe ratio):
- investor has mean-variance preferences
- this means the objective function of the investor is increasing in the average fund return and decreasing in the fund volatility.
- we are measuring risk as total volatility
- we are assuming total volatility of the investment matters to the investors
- if this is the case, then we are assuming that the investor is investing in the fund in an I diversified manner as idiosyncratic risk is being taken into account
43
Treynor ratio
Like the sharpe ratio, it gives the excess return per unit risk but it uses systematic risk instead of total risk.
It is calculated as the ratio of the average excess fund return to the fund beta.
Tp=(rbarp-rf)/Bp
A reward to risk ratio
(It measures risk based on beta instead of volatility)
The higher the ratio, the better the fund’s performance.
We are assuming that the investor is well diversified as the systematic risk or beta is taken into account.
44
Jenson’s alpha
It measures the average return on the portfolio over and above that predicted by the CAPM given the portfolio’s beta and average market return.
Jenson’s measure is the portfolio’s alpha value.
It measures how far (in return units) a security is from a security market line.
How to calculate:
Jenson’s alpha= what P has earned-what P should’ve earned
What P warna is the average return of the portfolio
What P should’ve earned if CAPM holds is from the CAPM equation (rf+Bp(rbarm-rf))
Basically:
Alpha=rbarp-[rf+Bp(E(rM)-rf)
45
Information ratio
It measures ratio of a portfolio’s alpha to its residual standard deviation.
Information ratio= alphap/sd(ep)
The residual sd or risk of a fund is the standard deviation of the residuals derived from a regression of the returns of the portfolio on the market or benchmark portfolio. It is also known as the tracking error as it reflects the extent to which the portfolio fails to track the benchmark return.
However, as we know that high book to market ratio and small firms earn higher average returns, and funds managers take advantage of this, we need to take this (size, value and market factor) into account for the Information Ratio. We can instead use the Fama French 3 model to calculate the alpha, which will then be called 3 factor alpha.
46
What determines the size of mutual funds industry in each country?
Khorana, Servaes and Tufano (2005) week to explain why are differences in sizes of the mutual fund industry across countries.
They find that the size of the mutual fund sector is bigger in countries with:
-Stronger rules and regulations
-Higher levels of education
-Higher levels of wealth
-A greater prevalence of defined contributions rather than defined benefits in the pension system of the country.
47
What types of funds can investors select
Mutual funds can be categorised in various ways:
- By asset class (may mix asset classes)
- Whether they are active or passive
- Income or growth funds
- Whether they are accumulation or distribution funds
- By their mutual fund sector
48
Fund sectors
The Investment Management Association (IMA), the UK fund industry body, classified UK funds into sectors.
- The classification system depends on mix of assets invested in by funds (eg. Funds in the UK all companies sectors must invest 80% if assets in UK equities)
- The classification system of IMA is accepted by all market participants including fund managers, fund rating agencies (the performance is ranked in IMA sectors)
- IMA forces its classification system
IMA classification system diagram
49
Who runs mutual funds
Funds are typically a part of a management group or Fund Family
E.g.Aberdeen Asset Management, Fidelity International, Scottish Widows, etc.
50
Mergers between fund families
The number of fund families in the UK has declined partly due to various mergers between asset management companies. These mergers may take place because fund families believe there are economies of scale in fund management.
- 2009 BNP Paribas acquires Fortis
- 2009 Blackrock acquired Barclays Global Investors
51
What makes mutual funds tick
Contracts are negotiated by the board of directors.
- The legal council provides guidance on complying with relevant regulations
- The transfer agent tracks shareholder positions and answers their questions
- The distributor sells shares to investors, directly or through intermediaries
- The management company (also called fund sponsor, fund complex, or fund family) creates the fund and does the investing (or hires a sub advisor to do it)
- The fund accountant maintains the fund’s books and computes the NAV nightly
- The custodian holds the securities in the fund’s portfolio
- The auditor helps the board to evaluate the accuracy of financial statements and the adequacy of its operational procedures
52
How are mutual funds distributed
```
Mutual fund shares are sold to retail investors in various ways:
-Directly
•Mail
•Phone
•Internet
```
-Through Investment advisors
•Independent
These are free to sell funds from any family. They are laid mainly in commission in a way that the fund family will split the front-end load with them. This may create a conflict of interest as they may recommend funds that will pay them the highest commission.
•Affiliated
These are employed to sell only one company’s funds. They are paid partly on commission and partly based on salary from the fund management company they sell for
-Through fund supermarkets
This is a recent introduction to the fund distribution process. Their selling proposition is not to provide fund advice but to provide access to funds more cheaply than other channels. They can do this as they pass most of the front-end load commission on to the investor. This is made into a viable business model by selling enough volume.
RDR introduced in 2013
Financial advisors will now have an upfront fees rather than commission. According to Deloitte 5.5 million people will cease to seek advisors. Advisors will have access to only the really wealthy clients.
53
Mutual fund fees
Mutual fund investors may pay 3 types of fee:
- Initial fee/Front-end load (%of money invested): pays for distribution expenses e.g. sales commission.
- Annual Free (% of assets under management): pays for fund expenses including administration and marketing, service charges and payments to fund managers.
- Selling fee /back-end load/contingent deferred sales charge (%of money redeemed): recoup distribution expense, discourage redemptions
54
Mutual fund investment restrictions
-Typically need to be diversified and plain vanilla
-Limited leverage (30%max)
-Limited derivative use
-Exchange-traded assets (since need to be able to establish NAV daily)
-No short selling
-Diversification requirement
•Max 5% of portfolio in one security
•Max 10% of voting shares of a given company
55
Mutual fund disclosure
They are required to file a prospectus and disclose their portfolio holdings regularly (In US quarterly with a 45 day delay. A few choose to disclose monthly)
Implications of disclosure
- Helps protect against investor lawsuits
- Creates incentives for portfolio managers to window dress portfolios by adding top performing securities just before disclosure
- Enables other investors to free-ride on disclosed information without paying management fees.
56
New rules for mutual funds
- Soft dollars have been regulated against
- 4PM hard cut off (a trade order must arrive at the mutual fund by 4pm to be executed)
- Fair value pricing (When computing fund NAV, Prices of the securities in closed markets are adjusted to reflect the likely impact of big price changes expected when the market opens)
- Redemption frees proposed (A redemption fee of 2% or more to be charged on mutual fund shares sold within 1 week of purchase. These fees would be paid not to the management company but directly into the fund to compensate the investors for potential losses due to rapid trading)
57
Mutual fund performance
Two types of tests:
-Evidence if abnormal performance
•Test for market efficiency
•Tells us whether fund managers add value or not
-Evidence of actively managed fund manager skill
•Testing for persistence in fund performance
Two crucial data issues to look at:
1) Fees
2) Data bias
We also need to decide which performance measures to use.
58
How should net of performance be calculated (mutual fund)
When comparing the post fee performance of funds with direct investment, it is typical to subtract the annual management fee from returns but not the front/back-end load. This is because direct investment involves transaction costs. Not deducting front/back-end loaf is roughly equivalent to those transaction costs. (Argued that front/back-end loss are equal to purchase and sale transactions costs).
59
Data issue1: Role of fees in assessing performance
Fees needs to be adjusted properly when measuring fund performance. If we look at fund performance we can either do so
•before taking fees into account or
•after taking fees into account.
To understand if fund managers have skills we should look at performance before taking fees into account -gross returns.
To understand if managers add value for investors we should look at performance after taking fees into account -net returns.
Fees can have a big effect on fund performance especially in the long run.
60
Survivorship bias (mutual funds)
Funds not surviving biases performance measurement (funds that die are generally bad performers).
If we were to measure performance between 1980 and 2009, data will only be available for firms alive today. Data vendors selling data are Morningstar & Lipper etc.
If we measure performance using funds alive for the whole period (1980-2009) from the standard database, we would have issues as it would overestimate past performance. This is because we would not have the data of the dead firms.
We would be using a survivorship biased dataset. As funds that close underperform, using a survivorship biased data set will lead to an over estimate of mutual fund performance.
61
Taking into account survivorship bias (mutual fund)
To do so, we can add data of the dead funds in the sample.
62
Which performance measure to use (mutual funds)
Literature looks at risk adjusted performance of mutual funds.
Other than the Sharpe Ratio, most studies use an alpha based measure as they realise that most investors will generally invest in more risky assets. Alphas are a natural way of telling us if funds are adding value as they tell us whether funds are earning more than they should given their risks.
The empirical evidence in the asset pricing area suggests that there are size and value premia. If we are to take these factors into account then we might choose to estimate alphas using a three factor model with market size and value factors.
If we believe that fund managers should not receive credit for momentum strategies then we might add the 4th factor MOM.
63
Mutual fund performance
Fama and French (2008) examined the issue of mutual fund performance. They used the survivorship bias free US equity mutual fund data from 1962-2006 from the CRSP mutual fund database.
They formed equally weighted and value weighted portfolios of the funds in their sample and regressed the equally weighted and value weighted returns on the 3 Fama French factors or 4 Carhart factors.
Their findings showed that fund managers do not add value.
64
Persistence of mutual fund performance
Zero performance could be either because mutual fund performance is just random or there are some fund managers with skill and some that are very poor therefore on average we get zero.
Carhart(1997) tested for this using his 4 factor model and said there may be a small group of managers with/without skills. The rest of the performance is mainly by chance. However his graph is informal.
Fama French conduct more formal tests for it as well using the data on all US equity funds between 1984-2006. They sorted funds into deciles:
- Based on the t-statistic of average net of market gross returns in the previous 60 months
- Based on the t-statistic for the 3 factor alpha estimated over 60 months using gross returns
Results:
- Sorts on last t-statistics of average returns suggest there is no persistence.
- Sorts on t-statistics of three alpha model suggests that there is evidence of persistence however, it is period specific.
- The results overall show there is evidence of persistence but it is relatively weak.
65
Sorts on 60-month three factor alpha estimates (mutual funds)
There is a case for sorting funds on alpha estimates, since;
- They are more precise than average returns
- Sorts on alpha estimates may focus better on the effects of information about individual stocks
Downside of using alpha sorts
-They May in part focus on the benchmark model’s problems in explaining the averages returns on some stocks
To check the robustness of their results, Fama and French run tests for the more recent 1993-2006 period.
66
The smart money effect (mutual fund)
Mutual fund investors pick the active fund managers that do well (money is smart). They provide rationale for investing in active funds.
To test for Smart Money Effect;
-Obtain data on flows
-relate net flows to subsequent performance
-Then look and combine things such as;
•Do negative net flow funds underperform on average
•Do high flow funds underperforms low flow
•Do positive net flow funds outperform on average
•Do high flow funds outperform low funds
Gruber takes sunset of US domestic equity mutual funds 1985-1994.
He estimates net quarterly flows. Finds out:
-Positive net flow funds outperform on a risk adjusted basis by 2.9bp/month one quarter ahead
-Negative net flow funds avoid underperformance of 7.1bp/month on a risk adjusted basis one quarter ahead
-Combining the two outperformance 14.7bp/month in a risk adjusted basis one quarter ahead
However he does not test for statistical significance
Zheng conducts tests of the smart money effect using US domestic equity mutual funds 1970-1993. Then estimates net quarterly flows. Later, risk-adjusts using Fama French 3 factor model. Finds out statistically significant excess return difference between positive and negative flow portfolios. Therefore concludes that investor have fund selection ability.
Sapp and Tiwari(2004) argue that the smart money effect is due to prior studies’ failure to account for momentum.
- Investors tend to put heir money into ex-post best performing funds
- These funds have disproportionate holdings of ex-post beta performing stocks momentum stocks
- Thus, after buying into winning funds, investors benefit from momentum returns in winning stocks
To test this reasoning Sapp and Tiwari calculate abnormal performance following money flows with and without accounting for the momentum factor using net quarterly flows. They find that inclusion of the momentum factor in the performance procedure eliminates outperformance of high flow funds
67
Hedge funds
Considered a part of the alternative investment class which includes real estate, venture capital, private equity, etc. The basic idea is investment pooling. I besties buy shares in these funds which invest the money on their behalf.
To bypass the Security Acts of 1933, hedge funds are restricted to having only accredited investors like instructions, high net worth, companies. (Who can fend for themselves).
As they are not registered companies:
-They don’t have to disclose their portfolios (transparent holdings). It is designed to protect unsophisticated investors (must provide periodic information on portfolio composition). They have to report their holdings to the Securities and Exchange Commissions and must have audited statements. Secrecy although helps protect against imitation, it also limits access to fund risk (Long term capital management eg).
-Can Pursue complex strategies:
Hedge funds make extensive use of derivatives, leverage and short selling. Mass selling of hedge fund strategies would be difficult because hedge fund strategies are too complex for for the typical mutual fund investor to understand.
-Can charge asymmetric incentive fees
Hedge funds normally charge incentive fees, which is nearly always asymmetric. This means that the fund earns fee when they do well but don’t have to pay investors when they do bad.
-Have redemption restrictions
Hedge funds often impose lock-up periods as long as several years in which the investment can not be withdrawn. These limitations limit the liquidity of investors but enable the fund to invest in illiquid investments (higher returns) without worrying about meeting the redemption demand.
68
What hedge funds do
Most hedge funds try to find trades that are almost arbitrage opportunities—pricing mistakes in the markets that can produce low risk profits. When the fund finds assets that are mispriced, they device hedges for their position so that the fund will benefit form the correction of the mispricing but be affected by little less.
Eg. Long-Term Capital Management specialised in identifying mispriced bonds. It would sell overvalued bonds short and hedge its position against interest rate risk and any other risks. The return of the fund would only depend on the corrections on the mispricing of the bonds and not on interest rate changes.
However, not all positions hedge funds take are hedged either because of high costs or due to intrinsic difficulties in hedging against some risks.
As judge funds seek inefficiencies in the financial markets and attempt to correct them, they play an important role as the bring the asset prices closer to their fundamental values.
69
Hedge funds and market efficiency
By selling short, the hedge fund portfolio is not effected by the changes in the market as a whole.
However, if the stock market falls sharply then the mutual fund would lose and the hedge fund would not.
This makes mutual fund manager less keen on taking such risks and limits their efficiency.
Diagram pg 23 lecture 10.
70
Hedge fund styles
CA-TASS(Credit Suisse/Tremont Advisors Shareholder Services) maintains one of the most comprehensive databases on hedge fund performances.
It divides the funds into 11 categories:
- Convertible Arbitrage (Hedged investing in convertible securities, typically long convertible bonds and short stocks)
- Dedicated short bias (net short position, usually in equities, as opposed to pure short exposure)
- Emerging markets (Goal is to exploit inefficiencies in emerging markets. Typically long-only as short selling isn’t feasible in such markets)
- Equity market neutral (Commonly uses long/short hedges. Control for industry, size, sector, and other exposures, and establishes market neutral positions designed to exploit some market inefficiency. Commonly involves leverage)
- Event driven (Attempts to profit from situations such as mergers, bankruptcy, etc).
- Fixed-income arbitrage (Attempts to profit from price anomalies in related interest rate securities. This includes US vs Non US bonds arbitrage, interest rate swap arbitrage, etc)
- Global Marco (Involves long and short positions in capital or derivative markets around the world. Portfolio positions reflect views in broad market conditions and major economic trends)
- Long/Short equity hedge (Equity-oriented positions on either side of the market)
- Managed futures (uses financial, currency or commodity futures)
- Multistrategy (Opportunistic choice of strategy depending on outlook)
- Fund of funds (Funds allocates its cash to several other hedge funds to be managed)
71
Style analysis for hedge funds
It uses regression analysis to measure the exposure of a portfolio to various factors or asset classes.
The betas on a set of factors determine the fund’s exposure to each source of systematic risk.
A market neutral fund will have no sensitivity to the market index.
Directional funds have significant betas in the factors that the fund bets on.
Hasanhodzic and Lo (2007) used 6 factors to run a style analysis by regressing hedge fund returns on these factors:
- Equity market conditions (return on the S&P 500)
- Foreign Exchange rates (US dollars index returns)
- Interest rates (return on the Lehman Corporate AA bond index)
- Credit conditions (return gap between the Lehman BAA index returns and the Treasury index returns)
- Commodity markets ( Goldman Sachs commodity index returns)
- Volatility (change in the CBO volatility index (VIX))
The findings presented the factor exposure of 1610 hedge funds. The estimated factor betas seem reasonable in the terms of the fund’s stated style. Equity market neutral funds have uniformly low factor betas as we expected. Dedicated short bias funds have substantial negative betas on the S&P index.
72
Hedge fund performance and downside betas
Lo (2001) examines the sensitivity of hedge fund returns to market conditions by partitioning them into periods (when S&P rose and fell). He estimated the upside betas for when it is rising and vice versa. He found out that for many strategies the downside betas exceeded the upside betas.
73
Hedge fund returns and downside betas
Investors do not want high market sensitivity when the market is going down. Lo argues that funds may write options explicitly or implicitly through dynamic trading strategies.
Pay off to funds that either:
- Hold stock and write put options
- Hold stock and write call option
Diagram pg 39
74
Hedge fund performance
If risk is measured by volatility then the hedge fund index is less volatile than then S&P 500.
Ibbotson and Chen (2005) examined the performance of 3538 funds from January 1999 to March 2004. After adjusting for various sample biases they concluded that equally weighted compound average return of the funds net of fees is 9.1%. They find that exposure to broad market indexes account fit a return of 5.4%. The return bet of fees minus the 5.4% gives us the average alpha of the fund if 3.7%. This alpha is a lot better than that if a mutual fund which is mostly negative.
Kosowski, Naik and Teo (2007) used an extremely large database from 1994-2002, to conclude that the average alpha is 5.16%. However this is insignificant. They also said that depending on the approach used, the top performing funds have average alphas between 12.7% and 16.1% per year which is highly significant.
75
Issues measuring Hedge fund performance
1)Data biases
Measured hedge fund returns may be effected by survivorship bias. Malkiel and Saha (2005) found out that attrition (closure) rate for hedge fund is higher than that if mutual fund. They claimed that hedge fund databases contain 4.4% survivorship bias. Fung and Hsieh (2000) claim that it is 3.6%.
Backfill bias is a bias that arises because hedge funds do not need to report their returns as they are not regulated. Hedge fund with seed capital will only open to public if they have good past performance to attract new clients (blackfill bias). Therefore prior performance of funds included may not be representing due to the Black full bias.
Aragon and Nanda (2013) studied the return reporting behaviour of hedge funds and concluded that they delay their bad news. To minimise the impact of bad news they may report returns in cluster usually starting off with bad and eventually turning good. Investors then don’t invest in funds that delay news. They found that funds delaying earn 4% less returns.
2)Difficulties of risk adjustment
-Time varying riskiness
Because hedge funds can go long, short, borrow, use derivatives, their exposure to risk can vary a lot over a short time period. This makes it difficult to asses these exposures based on a limited sample of monthly returns.
-Which factors
A hedge fund’s return is best viewed as a basket of derivatives with nonlinear payoffs, so what factors to use (as opposed to a mutual fund’s which is a basket of stocks so use the return on index as a factor)
3)Illiquidity premium of hedge funds
Hedge funds are illiquid investments and their returns might just be the illiquidity compensation. Aragon (2007) shows that once the lock-ups or redemption notice periods, the positive alpha of the hedge fund turns insignificant.
4)Misvaluing illiquid assets
Agarwal et all (2097) show that hedge funds may manipulate the returns by mis-valuing illiquid or non traded assets. In their paper they report a Santa effect: hedge funds report average returns in December that substantially greater than in other months. This pattern is stronger for funds that are near or beyond the threshold return at which performance fees kick in. This study suggests that illiquid assets are valued more generously in December when annual performance relative to benchmark is being calculated. If funds take advantage of illiquid or non traded assets to manage returns turn accurate Performance measurement becomes impossible.
5)A short performance window of a particular hedge fund may not indicate it’s true risk (discussed earlier)
76
Hedge fund persistence
Novikov et al (2006) conducted a study using a large database of hedge funds and concluded that half of the performance of hedge funds over a three year period spills over the next three years. Eg: a fund with an alpha of 2% can be expected to have an alpha of 1% in the next three years. Therefore it shows that investing in high alpha funds is profitable.
77
Hedge fund fee structure
Typically a hedge fund’s fee structure is a management fee of 1-2% of assets plus an incentive fee equal to 20% of investment profits beyond a specified benchmark.
Incentive fees are effectively call options on the portfolio with a strike price equal to portfolio value x(1+benchmark return).
High watermarks:
Many funds have high watermarks. They can’t charge incentive fees if the investor losses until they recover the previous high value. High watermarks have been blamed for high attrition rate of the funds.
78
Funds of funds
They are the funds that invest in other funds.
Their fees can hurt investors as they charge a fee for each outperforming underlying asset even if the overall performance of the fund of funds is poor. In doing so diversification can hurt the investor.
They operate with leverage on top of leverage of primary funds which makes returns highly volatile. If the primary funds invested by fund of funds have similar investment styles then the benefits of diversification disappear.
