Sei sulla pagina 1di 42

Technical University of Munich

Investment and Risk Management

Essay
Author: Supervisor:
Daniel Kühner Professor Luis Seco

April 14, 2017


Contents
1 Introduction 2

2 The World of Hedge Funds 2


2.1 Who is allowed to invest in hedge funds? . . . . . . . . . . . . . . . 2
2.1.1 United States . . . . . . . . . . . . . . . . . . . . . . . . . . 2
2.1.2 European Union . . . . . . . . . . . . . . . . . . . . . . . . . 3
2.2 Hedge fund strategies . . . . . . . . . . . . . . . . . . . . . . . . . . 3
2.2.1 Event driven . . . . . . . . . . . . . . . . . . . . . . . . . . . 3
2.2.2 Merger arbitrage . . . . . . . . . . . . . . . . . . . . . . . . 4
2.2.3 Relative value arbitrage . . . . . . . . . . . . . . . . . . . . 4
2.2.4 Global macro . . . . . . . . . . . . . . . . . . . . . . . . . . 5
2.2.5 Equity hedge . . . . . . . . . . . . . . . . . . . . . . . . . . 6
2.2.6 Equity market neutral . . . . . . . . . . . . . . . . . . . . . 6
2.2.7 Fixed income convertible arbitrage . . . . . . . . . . . . . . 7

3 Convertible Bond Arbitrage 7


3.1 A trading example . . . . . . . . . . . . . . . . . . . . . . . . . . . 7
3.2 Set-up . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7
3.3 The optimal Sharpe ratio . . . . . . . . . . . . . . . . . . . . . . . . 11
3.4 Introducing leverage . . . . . . . . . . . . . . . . . . . . . . . . . . 11
3.5 Introducing a maintenance margin . . . . . . . . . . . . . . . . . . . 12

4 Empirical analysis 13
4.1 Best and worst daily returns . . . . . . . . . . . . . . . . . . . . . . 16
4.2 Key statistics on a rolling time window . . . . . . . . . . . . . . . . 18
4.3 Correlations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 28
4.3.1 Correlations in turbulent times . . . . . . . . . . . . . . . . 30
4.4 Portfolio Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35

1
1 Introduction

2 The World of Hedge Funds


Hedge funds pool capital from accredited individuals or institutional investors to
invest in a variety of asset classes, often with complex portfolio-construction and
risk-management techniques. Legally, hedge funds are usually structured as limited
partnership, limited liability company, or similar vehicle. Hedge funds are generally
distinct from mutual funds as their use of leverage is not capped by regulators and
distinct from private equity funds as the majority of hedge funds invest in relatively
liquid assets. Since hedge funds are available only to accredited investors and can
not be offered or sold to the general public, hedge funds can generally avoid direct
regulatory oversight and operate with greater flexibility. As of February 2011, 61%
of worldwide investment in hedge funds comes from institutional sources Pensions
and Investments (2011). Table 1 displays the world’s largest hedgefunds by assets
under management (AUM) as of June 30, 2016 (Pensions and Investments (2016)).

Company Assets in million USD


Bridgewater Associates 102,930
AQR Capital Mgmt. 63,030
Man Group 46,300
Millennium Mgmt. 33,302
Millennium Mgmt. 33,302
Renaissance Technologies 32,000
The Baupost Group 29,200
Elliott Management 28,834

Table 1: The world’s largest asset management companies.

2.1 Who is allowed to invest in hedge funds?


2.1.1 United States

In the United States, the term ”accredited investor” is defined in Rule 501 of
Regulation D of the U.S. Securities and Exchange Commission (SEC). As a natural
person, the networth by the time the shares of a hedge fund are acquired needs to

2
be at least 1,000,000$, excluding the value of the individual’s primary residence,
or the annual income has to exceed 200,000$ in each of the two most recent years
or joint income with a spouse exceeding 300,000$ and there needs to be sufficient
ground for the assumption of the same income level for the current year. As an
institutional investor (legal person), certain criteria considering size and capital
needs to be fulfilled as well.

2.1.2 European Union

The Markets in Financial Instruments Directive (MiFID) defines the term of an


’elective’ professional client. To qualify as a retail investor, two out of the three
subsequent criteria need to be met:

• The client has carried out trade transactions of at least 50,000 EUR at an
average frequency of 10 per quarter over the previous four quarters.

• The size of the client’s financial instrument portfolio, defined as including


cash deposits and financial instruments, exceeds 500,000 EUR.

• The client works or has worked in the financial sector for at least one year
in a professional position which requires knowledge of the transactions or
services envisaged.

2.2 Hedge fund strategies


2.2.1 Event driven

Event driven hedge funds maintain positions in companies currently or prospec-


tively involved in corporate transactions of a wide variety including but not limited
to mergers, restructurings, financial distress, tender offers, shareholder buybacks,
debt exchanges, security issuance or other capital structure adjustments. In the
U.K., around 7% of all hedge funds follow an event driven strategy according to the
research house Preqin (Preqin (2013)) making it the smallest fraction of strategies.
Corporate transactional events can generally be categorized into distressed secu-
rities, merger arbitrage, and special situations. In the case of distressed securities,
hedge funds often invest in the bonds of companies facing financial distress which
are thus trading at substantial discount. Most mutual funds are not allowed to to

3
hold securities that have defaulted, making the supply even greater and depress-
ing prices. Due to financial restructurings or reorganizations debt investors often
obtain control during the bankruptcy process, which gives their investment equity
like characteristics. They can influence or control the reorganization process by
for example liquidating assets and recapitalizing healthy parts of the business.
Hedge funds pursuing a merger arbitrage strategy are investing in companies which
are being acquired or seem to be likely targets for acquisitions. Usually the stock
keeps trading below the offered price due to uncertainty whether the deal will go
through. Specialized hedge funds may have superior information or more experi-
ence in analyzing the odds for the deal.
Special situations are events that impact the value of a company’s stock, including
restructurings, spin-offs, share-buy-backs, security issuance/repurchase or asset
sales. In the healthcare sector, event driven investments can also be bets whether
a drug gets approved by the FDA.

2.2.2 Merger arbitrage

Merger arbitrage is a substrategy of event driven and involves usually long short
trades with stocks of companies involved in a merger or acquisition or which are
likely targets for such. Usually after an acquisition/ merger has been announced,
the stocks of the company being acquired appreciate in price due to the ownership
premium and the share price of the acquiring company falls due to the risks whether
the paid price is justifiable and the investment pays off.

2.2.3 Relative value arbitrage

Relative value arbitrage strategies take advantage of relative discrepancies in price


between securities. The price discrepancy can occur due to mispricing of securi-
ties compared to related securities, the underlying security or the market overall.
Strategies in this category typically have very little or no directional market expo-
sure to the market as a whole which is achieved by pairs of long and short positions
in correlated securities.

4
2.2.4 Global macro

Hedge funds following a global macro strategy take positions based on the analysis
or in anticipation of macroeconomic events. This usually involves forecasts and
analysis of interest rates, international trade, political regime changes and other
broad systemic factors.

Commodity trading advisor (CTA) A Commodity trading advisor (CTA)


is the US financial regulatory term for an individual or organization who provides
advice and services related to trading in futures contracts, commodity options
and/or swaps. In 2010, Dr. Galen Burghardt, adjunct professor at the University
of Chicago’s Booth School of Business, found a correlation of 0.97 between a subset
of trend following CTAs and a broader CTA index from the period 2000-2009,
indicating that speculative technical trend following had been dominant within
the CTA community (Dr. Galen Burghardt (2010)).

”Systematic currency” - the example of the Black Wednesday One im-


portant substrategy is ”systematic currency”, where the fund tries to be long in
appreciating currencies or short in depreciating ones. One famous example is
George Soros’ bet against the pound sterling in 1992, which we want to study in
further detail.
On October 8 1990, Margaret Thatcher entered the GBP into the European Ex-
change Rate Mechanism (ERM) at 2.95 Deutsche Mark. This meant that the
British government would guarantee that the exchange rate of the pound would
fluctuate not more than 6% against other member currencies. Hence, if the ex-
change rate ever neared the bottom of its permitted range, DM 2.773, the gov-
ernment would be obliged to intervene. However, the conditions by the time of
entering the ERM were not at all favorable for the UK: inflation was at 8.1%,
three times as high as in Germany. The Bank of England had basically two in-
struments to keep the exchange rate within the guaranteed boundaries: adjusting
the base interest rate and buying or selling pound (open market operations). On
September 15 1990, the president of Germany’s Bundesbank Helmut Schlesinger
gave an interview where he stated that two currencies within the ERM could
come under pressure. It was clear that the two currencies Schlesinger referred to
were the British pound and the Italian lira. The interview was published without

5
Schlesinger’s permission and set the starting shot for currency traders. George
Soros with his Quantum fund had already been building up huge short positions
in pound in the months leading up to September 1992. On September 16, known
as the Black Wednesday, Soros through his Quantum fund sold short more than
10 billion $ in pounds. The BoE first tried to fight the pressure by raising the in-
terest rate from 10% to 12% to increase the pound’s attractiveness on September
16. As this had little effect, a prospective rate hike to 15% was announced only
a few hours later which again did not relieve the downward pressure. At 7 p.m.,
the Chancellor of the Exchequer Norman Lamont called for a press conference
where he announced that the UK would leave the ERM and let the pound float
again. The GBP depreciated about 15% against the Deutsche Mark and about
25% against the USD in the following 5 weeks. Soros’s Quantum fund made a
profit of 1 billion USD and brought him the name ”The Man Who Broke the Bank
of England”.

2.2.5 Equity hedge

Equity hedge includes a wide variety of strategies where stocks are involved. The
”hedge” usually refers to long-short pair trades. One example which intersects with
relative value and equity market neutral would be pairs trade where a hedge fund
goes long e.g. an automotive company and meanwhile shorts another automotive
company which is overvalued on a relative basis.

2.2.6 Equity market neutral

Equity market neutral is often quite similar to equity hedge, but it focuses par-
ticularly on the issue of yielding profits independently of what the market overall
is doing. Usually it involves pair trading as well, which we outlined in the section
above. As we examine the correlations later on, we will see that the HFRX Equity
Market Neutral index achieves to stay uncorrelated with the overall market (mea-
sured through the S&P 500 TR), even during distressed times when correlations
often ”heat up” and increase.

6
2.2.7 Fixed income convertible arbitrage

Fixed income convertible arbitrage is a strategy where the hedge fund usually
enters a long position into the convertible bonds of a company. These bonds can
be converted into a fixed number of shares of the same company. These securities
could be replicated by a normal bond plus a call option on the issuing company.
However, since the convertible bonds are often quite illiquid, the spread includes
a premium to recompense for the illiquidity. But since market pressure forces
funds to have ever more permissive redemption periods (often below 30 days),
this illiquidity also posts a major threat to hedge funds which for example caused
severe losses during the financial crisis for convertible arbitrage funds. We will
describe the details of this strategy in the respective section where we will also go
through a trading example.

3 Convertible Bond Arbitrage


3.1 A trading example

3.2 Set-up
Let’s consider the following set-up.: In the market there is a convertible bond for
which we observe a price of 80$ today (t = 0). It grants the holder the right to
convert it into 10 shares of the issuing company any time she or he wishes to do
so. The bond has a coupon of 10% p.a. on its face value which is currently equal
to its market value of 80$. The coupon is payed semiannually which thus implies a
payment of 4$ twice a year. The shares of the issuing company currently trade at
7$. We as investors want to execute a classic convertible arbitrage trade by being
long in one bond and short in m shares of the issuing company. Our aim is it to
optimize the Sharpe ratio with respect to m, the number of shares in our short
position. At this point, we want need to clarify and explain the mechanics of a
margin account in general, as well as make assumptions regarding our example.

• Initial margin requirements: The Federal Reserve Board Regulation T (12


CFR §220 – Code of Federal Regulations, Title 12, Chapter II, Subchapter A,
Part 220 (Credit by Brokers and Dealers)) requires a minimum initial margin

7
for long as well as for short positions of 50% for equity-like instruments. (For
completeness, we need to add that it requires also an absolute amount of
2000$ in the first place.) In case of a long position, this means that if an
investor intends to buy shares worth 100$, he needs at least 50$ of his own
money (which we refer to as equity in the following) and can borrow the
additional amount of 50$ from his broker. If the investor wants to sell short
shares worth 100$, he needs an initial amount of 50$ in addition to the
proceeds from the selling of the borrowed shares, so the minimum
required amount on his margin account sums up to 150$. Brokers are allowed
to require margins above the thresholds of Regulation T, so we assume for our
example that the broker requests an initial margin of 100%. We should add
here that convertible bonds are legally not considered equity-like securities,
thus margin requirements below 50% would be possible.

• Analogy between the margin account and a company’s balance sheet: We want
to use this opportunity to establish a basic analogy here which might help
understand the functionality of a margin account - namely that a trader’s
margin account can be seen quite in the same way than a company’s balance
sheet. The balance sheet total corresponds to the total market value of
all positions. These are usually partly financed by equity, partly by debt,
borrowed from the brokerage firm. Any loss cuts into the equity and any
gain benefits the equity. The equity is recalculated usually at least on a
daily basis. If the equity drops below a minimum required percentage of
total positions, there are two possibilities: if feasible, it might make sense to
inject fresh money, which corresponds to a company raising equity. If this
can not be accomplished, a liquidator - in this case the broker - will sell all
assets which in the context of trading means close all positions in order to
recover the outstanding debt.

• Maintenance margin requirements: If the trader’s positions go against him,


they keep eating into his equity whilst the debt of course remains untouched,
which might eventually cause the equity as share of the total market value of
all positions to drop below a predefined threshold. This threshold is referred
to as the maintenance margin. When equity as share of all market positions
drops below the maintenance margin, the trader receives a margin call which

8
offers him a last chance to remargin his account, i.e. inject fresh capital.
If he can’t meet this obligation, the broker will liquidate some positions.
Regulation T requires a minimum maintenance margin of 25%. To bring
the analogy from above into play again, the ”maintenance margin” for a
company’s equity as share of the balance sheet is zero, so if this point is
reached, the company is obliged to file bankruptcy or inject fresh equity
in the very last second. If the initial margin would be just as high as the
maintenance margin, even a tiny negative market movement would trigger a
margin call directly after opening a position.

• Borrowing costs: The costs for borrowing shares in order to execute the short
selling is also known as the stock loan fee. It is charged by the brokerage
firm to the client who is borrowing the shares.

• Interest on the margin account: First, we have to distinguish between the


deposit rate and the lending rate

– Lending interest rate For the debt borrowed from the broke, usually
a non-negligible interest rate is charged. For example Interactive Bro-
kers LLC, according to research firm Preqin amongst the top 10 prime
brokers servicing hedge funds in the world, charges 2.16% for amounts
up to 100,000$ and still 0.91% for amounts above 3,000,000$ (checked
February 28, 2017).
– Deposit interest rate The deposit interest is paid to you on a positive
cash position by the broker. These rates are typically far lower than
the lending rate. Interactive Brokers for example is paying 0.16% for
USD deposits above 10,000$ and 0 for amounts below. For deposits
in EUR, zero interest is paid for deposits up to 100,000 EUR and for
amounts above 100,000 EUR the deposit rate even drops substantially
into negative territory with -0.609%.

For our example of an convertible arbitrage strategy, let’s make the following
assumptions: Our broker charges a margin interest of 2% p.a. on borrowed money
and pays zero interest on any cash deposits. The stock loan fee is 2% as well. The
initial margin requirement is 100%.

9
Our position is designed to be held for half a year before liquidating it. We consider
three possible states of the world for t = 1/2, each equally likely to materialize.

price bond price share


Scenario 1 100 14
Scenario 2 80 7
Scenario 3 30 0

Let’s first calculate the necessary equity for our trade. With an initial margin
of 100% we need 80$ to purchase the bond and m*7$ in equity to sell short m
shares, each trading at 7$. Thus V0 = 80 + 7 ∗ m is our initial equity.

• Scenario 1:
In this scenario, we would convert our bond into 10 shares. We therefore
do not receive the semiannually paid coupon of 4$. On our account, we still
have 7*m$ proceedings from selling the borrowed shares plus 7*m$ of initial
margin. We could use this money to buy back m shares at 14$ each in order
to cover our short position. Finally, the stock loan fee due is 2% ∗ 0.5 ∗ m ∗ 7$.
This adds up to:

V1.1 = 10 ∗ 14 + m ∗ 7 ∗ 2 − m ∗ 14 − 0.02 ∗ 0.5 ∗ 7 ∗ m


= 140 − 0.07m

• Scenario 2:
In this scenario, we receive the coupon payment of 4$ since we will not convert
the bond. The rest of the calculation is analogously to the one above:

V1.2 = 80 + 4 + m ∗ 7 ∗ 2 − m ∗ 7 − 0.02 ∗ 0.5 ∗ 7 ∗ m


= 84 + 6.93m

• Scenario 3:
Obviously the company has defaulted with a recovery rate of 37.5%, reducing
the bond’s market value to 30$ and the share price to zero. The coupon is

10
not paid here.

V1.3 = 30 + m ∗ 7 ∗ 2 − m ∗ 0 − 0.02 ∗ 0.5 ∗ 7 ∗ m


= 30 + 13.93m

3.3 The optimal Sharpe ratio


We optimized the Sharpe ratio with a brute force approach, running through 20,000
values for m between zero and 20, thereby assuming that any non-integer number
of shares could be bought or sold. For each value of m, we calculated the return
in scenario i ∈ {1, 2, 3} as
ri = V1.i /V0 − 1.

The Sharpe Ratio was then calculated by averaging the returns of the three sce-
narios and dividing it by the standard deviation of the returns. (Note that this
implicitly assumes a risk-free rate of zero.) The results are displayed subsequently.

Optimal Sharpe ratio 7.130


Optimal m (number of shares to short) 7.857
Optimal short ratio 68.749%
Optimal expected return 3.049%
Optimal std. dev. of returns 0.428%

3.4 Introducing leverage


In the next step, we reduce the initial margin requirements thus introducing lever-
age. Let’s assume a required initial margin of 50% on long positions and 80% on
short positions. To buy the bond, we take 40$ of our own money (equity) and bor-
row the additional 40$ from our broker. To sell short 7.857 shares, we need 7.857 ∗
7 ∗ 80% = 43.999 in equity. Whereas in previous set-up 80 + 7.857 ∗ 7 = 134.999$
were necessary, the required equity is now reduced to 40 + 43.999 = 83.999$. We
continue to calculate the final value of our position for all three scenarios. The
leverage causes both, the average return as well as the standard deviation of the
returns to increase, but the rise in standard deviation exceeds the rise in average

11
return, with negative implications for the Sharpe Ratio. The results are summa-
rized in Table 2.
Scenario 1
Sell shares from conversion 10*14 140
Buy back shares to close the short -7.857*14 -109.998
Return 40$ plus interest to broker -(40+0.5*2%*40) -40.400
Pay stock loan fee to broker - 0.5*2%*7.857*7% 0.549
Initial equity plus proceedings from short on margin account 1.8*7*7.857 98.998
Total 88.050
Return 4.822 %

Scenario 2
Receive semiannual coupon 4
Sell bond 80
Buy back shares to close the short -7.857*7 -54.999
Return 40$ plus interest to broker -(40+0.5*2%*40) -40.400
Pay stock loan fee to broker - 0.5*2%*7.857*7% 0.549
Initial equity plus proceedings from short on margin account 1.8*7*7.857 98.998
Total 87.049
Return 3.631 %

Scenario 3
Sell bond 30
Close the short 0 0
Return 40$ plus interest to broker -(40+0.5*2%*40) -40.400
Pay stock loan fee to broker - 0.5*2%*7.857*7% 0.549
Initial equity plus proceedings from short on margin account 1.8*7*7.857 98.998
Total 88.048
Return 4.820 %

3.5 Introducing a maintenance margin


Up to now, we did not ask the question whether the trader will receive a margin
call. Let us now assume that the maintenance margin requirement is 25% for short
and for long positions. This equals the maintenance margin requirements of the
NYSE and NASD 1 .
1
http://www.investopedia.com/ask/answers/05/shortmarginrequirements.asp

12
Sharpe ratio 6.437
Expected return 4.424%
Std. dev. of returns 0.687%

Table 2: Key statistics of the trade after introducing leverage.

In Scenario 1, the value of the short position is -7.857*14=-109.998$. The required


maintenance margin equals 125% of the short position, which is 137.497$. The
initially posted equity plus the proceedings from selling short are 180% of 7*7.857
which equals 98.998$. Thus, in scenario 1 the trader will receive a 38.499$ margin
call. Plus there is interest on the loan as well as stock loan fee to pay, so in total
the account needs to be remargined with 39.449$.
For Scenarios 2 and 3, the initially posted equity plus the proceeds from short
selling are above the required maintenance margins, thus the trader only needs to
pay the fees plus interest, but does not receive a margin call.

4 Empirical analysis
In the following section we analyze return data of different hedge fund strategies
and compare them to stock market indexes and to the government bond market.
We include indexes of seven hedge fund styles into our examination. The data
is provided from Hedge Fund Research, Inc. (HFR). According to them, HFR
”utilizes a UCITSIII compliant methodology to construct the HFRX Hedge Fund
Indexes. The methodology is based on defined and predetermined rules and objec-
tive criteria to select and rebalance components to maximize representation of the
Hedge Fund Universe” (Hedge Fund Research, Inc. (2017)). However, it should
be noted that there are indications of a negative self-selection bias: For example
when comparing return statistics of the HFRI and the HFRX indexes which are a
subset of the HFRI, HFRX performs worse. HFR requires the funds within HFRX
to disclose a lot of information. Successful hedge funds might not want to conform
with this obligation whereas less successful funds might appreciate the additional
advertisement and might be less afraid to disclose parts of their trading strategy.
The seven indexes we include are:

• HFRX Equity Hedge Index

13
Net equity exposure can range broadly. Typically, net exposure is above 50%
either long or short.

• HFRX Equity Market Neutral Index


Typically, net equity market exposure is kept below 10% long or short.

• HFRX Event Driven Index


Event driven managers maintain positions in companies currently or prospec-
tively involved in corporate transactions of a wide variety including but not
limited to mergers, restructurings, financial distress, tender offers, share-
holder buybacks, debt exchanges, security issuance or other capital structure
adjustments.

• HFRX ED: Merger Arbitrage Index


As a subsection of event driven strategies, merger arbitrage strategies typi-
cally have over 75% of positions in announced transactions over a given mar-
ket cycle, exposure to situations where no formal announcement has been
made is typically low.

• HFRX Macro/CTA Index


While both macro and equity hedge managers may hold equity securities,
the overriding investment thesis is predicated on the impact movements in
underlying macroeconomic variables may have on security prices, as opposed
to equity hedge, in which the fundamental characteristics on the company
are the most significant and integral to investment thesis.

• HFRX Relative Value Arbitrage Index


Relative value investment managers who maintain positions in which the
investment thesis is predicated on realization of a valuation discrepancy in
the relationship between multiple securities.

• HFRX RV: Fixed Income Convertible Arbitrage Index


Strategies employ an investment process designed to isolate attractive op-
portunities between the price of a convertible security and the price of a
non-convertible security, typically of the same issuer.

14
These seven indexes for hedge fund strategies are compared against the German
DAX, the S&P 500 Total Return Index and an index for the German Bundesan-
leihe. For an introduction, we will describe these three indexes briefly.

The DAX The DAX is a blue chip index consisting of the 30 major German
companies trading on the Frankfurt Stock Exchange. According to Deutsche Börse,
DAX represents Germany’s largest 30 companies in terms of market capitalization
as well as order book volume. Two versions of the DAX are being published, the
more commonly quoted is the performance index which measures total return (i.e.
reinvestment assumption for dividends). The other version is a price index and
thus more similar to commonly quoted indexes in other countries. The constituent
companies are weighted according to the market capitalization of the shares in
public float. If for a given company different type of stock is traded - like common
stock and preferred stock - the more liquid type is used for the calculation. The
worst years for the DAX were 2002 (-44%) due to the burst of the Dot-com bubble
and 2008 (-40%) due to the global financial crisis. When calculated backwards,
the best years since 1960 were 1985 with a plus of 66% and 1967 (+51%).

The S&P 500 Total Return Index The Standard & Poor’s 500 consists of the
500 largest U.S. companies with stock listed on the NYSE or NASDAQ. Due to its
comprehensiveness it is considered a good representation of the American economy.
The weighting is again proportional to the market capitalization of free floating
stock and therefore considerably different from the also frequently quoted Dow
Jones Industrial Average, which is a price-weighted index. The normally quoted
S&P 500 is calculated without reinvesting dividends, but to ensure comparability,
we use the total return index which reinvests dividends.

The Bundesanleihe The German Bundesanleihe (often just called the ’Bund’ )
is a debt security issued by Germany’s federal government. Bunds are auctioned
with initial maturities of 10 and 30 years. Bunds are highly liquid and are accepted
by the European Central Bank as collateral for credit operations. Germany’s
Bundesbank acts as a market maker at the German exchanges. The Euro-Bund-
Future is one of the most traded futures contract in the fixed income universe. To
fulfill a physical delivery of the contract, Bunds with a time-to-maturity between

15
8.5 and 10.5 years can be chosen. In our data, at each point in time we look at
the price of the most recently issued Bund with 10 original maturity of 10 years.

Index avg. return std. dev.


DAX 13.30 % 21.91 %
S&P 500 TR 10.86 % 18.88 %
Merger Arbitrage 4.39 % 4.21 %
Bunds 3.28 % 7.04 %
Event Driven 3.23 % 4.86 %
Macro CTA 0.85 % 6.27 %
Relative Value Arbitrage 0.84 % 4.12 %
Equity Hedge 0.83 % 6.42 %
Equity Market Neutral 0.06 % 3.95 %
Fixed Income Convertible Arbitrage -2.19 % 6.30 %

4.1 Best and worst daily returns


The window of time in our examination is April-01-2003 to December-30-2016.
We observe quite low average returns for the hedge funds but also lower volatility.
Figure 1 shows for the DAX, the S&P500 Total Return, the Bund and seven hedge
fund indexes their 10 best and worst days. One can observe that the most extreme
returns are far smaller in absolute terms for the hedge funds compared to the two
stock market indexes. Another noteworthy observation is that for the DAX, 5 out
of 10 of the best daily returns and 7 out of 10 of the worst returns took place
between Sep-2008 and Dec-2008, after the investment bank Lehman Brothers filed
for bankruptcy on September 15, 2008. For the S&P500 Total Return index the
concentration is even more extreme with only two days in the best and two days in
the worst 10 being outside of these three months. The -6.82% return of the DAX
on June-24-2016 was caused by the Brexit vote the day before. The two extremely
negative DAX returns on Aug-10-2011 and Aug-18-2011 happened during a period
where the European sovereign debt crisis was boiling up dramatically. On Aug-
10-2011, rumors were spread that France could be downgraded, The FTSE Mib,
Italy’s benchmark index had reached a new two-year low and the Bank of England
cut its growth forecast for the UK.

16
Figure 1: 10 best and 10 worst daily returns between April-01-2003 and December-
30-2016.

17
4.2 Key statistics on a rolling time window
First, we take a view on several key statistics for each index. Therefore, we go
through our time series with a 2-year rolling window.

Mean return The first key statistic we take a look at is the mean of daily re-
turns. The first black vertical line indicates Monday Sep-15-2008, the day Lehman
Brothers Holdings Inc. filed for bankruptcy after a weekend of dramatic negotia-
tions whether the investment bank could be overtaken by Barclays or should be
bailed out by the US government, which both did not happen as we all know.
The second black bar indicates the date two years later, when the effects of the
financial collapse start vanishing out in our two year rolling window analysis. The
second black bar marks Aug-10-2011, a day when the German DAX had its worst
fall since 2008 during the height of the European sovereign debt crisis. Again, the
second black bar marks the date two years later. One can see how especially the
financial crisis after 2008 dragged down the average daily returns of both, DAX
and S&P 500 TR. Soon after these extremely negative returns in autumn 2008 van-
ish out from our window, returns shoot up again, only to be dragged down again
about 10 months later in August 2011, when Europe was hit by the sovereign
debt crisis with full vehemence. It is also visible that the effect on the S&P 500
TR was smaller. The Bund is more stable as one would expect for a sovereign
bond. After the breakout of the financial crisis, the Bund profits from sharply
falling ECB rates: Between Oct 2008 and April 2009, the ECB’s main refinancing
operations (MRO) were lowered from 3.75% to 0.25%, making the higher yielding
Bund’s a rather attractive investment. The HFRX Equity Hedged and even more
the HFRX Market Neutral indexes are by far more stable than the stock market
indexes, which is again not very surprising. However, even the HFRX Market Neu-
tral index seems to be impacted slightly by the overall trends. Looking at Figure
3, the dramatic decline of average returns for HFRX Fixed Income Convertible
Arbitrage catches the eye. In the run-up of the financial crisis 2008, hedge funds
specialized in convertible arbitrage had increased their leverage - credit of an es-
timated volume of about 120 to 150 billion USD was added to their equity of an
estimated 35 billion USD. In the aftermath of the Lehman bankruptcy, the hedge
funds’ prime brokers had to call back a large amount of their outstanding credit,

18
forcing convertible arbitrage funds to get rid of their convertibles in fire sales,
causing immense pressure on the anyway rather illiquid market for convertibles
and sending prices southwards (Gerald Braunberger, FAZ (2010)).

Figure 2: Average return, 2 year rolling window.

19
Figure 3: Average return, 2 year rolling window.

Volatility - detecting the big quant meltdown Secondly, we examine volatil-


ity, the standard deviation of daily returns, again on a 2 year rolling window. For
the DAX and the S&P 500 TR, the Lehman collapse in September 2008 and the
sovereign debt crisis in August 2011 seem to be the most influential events. The
first one has a stronger impact on the S&P 500, the second on the DAX, which is
somehow comprehensible since each represents the respective home market where
the crisis originated. For the HFRX Equity Market Neutral, both the financial
crisis after 2008 as well as the European sovereign debt crisis seem to have little
effect. However, we can detect another turbulent phase, beginning on July-31-
2007, a day which is marked by the red dot. When checking Figure 1, we can
see that for the HFRX Market Neutral index, three of the best as well as three

20
of the worst daily results within the total 14 year period fall into the two weeks
between 31-July-2007 and 13-August-2007. Analyzing the news flow during these
days, it becomes clear that the crash in this area of the market was probably initi-
ated by substantial liquidations of the Global Alpha fund, a 9 billion USD market
neutral, quantitative hedge fund run by Goldman Sachs. Global Alpha was down
about 8% in the week ended July 27 2007 which could have caused redemptions of
investors. Many hedge funds have relatively permissive redemption periods, allow-
ing investors to withdraw their money within 30 days or even less. The pressure
was created because many market neutral hedge funds have relatively similar long
and short positions. Adding up to this, leverages are often quite high, so if even
smaller losses occur, the funds are forced to liquidate positions of considerable
size, increasing downward pressure and potentially frightening investors and caus-
ing redemptions which then leads to the necessity to further liquidate assets in a
positive feedback loop (Alistair Barr, MarketWatch (2010)). Another factor which
might also have played a role is that markets are typically less liquid in August.
Looking at Figure 5, we can see that volatility increased also significantly in July
2007 for the HFRX Macro CTA, HFRX Merger Arbitrage and HFRX Relative
Value. However, these spikes start a little bit earlier. The red dot on the Macro
CTA chart marks 06-July-2007, the other two red dots mark 20-July-2007.

21
Figure 4: Volatility, 2 year rolling window.

22
Figure 5: Volatility, 2 year rolling window.

Skewness Skewness is a measure of the asymmetry of the probability distribu-


tion of a random variable about its mean. Speaking about asset returns, a positive
skew indicates ’fatter’ right tail compared to the left tail, i.e. returns substantially
above the mean occur relatively more frequently compared to returns substan-
tially below the mean. Since returns enter the formula with the power of three,
the weight of outliers is increased considerably. It is important to stress that the
skewness is always relative to the mean, i.e. a positive skew during a period does
not tell whether returns themselves were positive. This becomes clear for the DAX
in Figure 6. As it can be seen in Figure 2, average rolling returns in the 2 years
between 15-September-2008 and 15-September-2010 are mainly negative, however
the skewness shoots up in this period. This effect can be explained again with

23
a look at Figure 1: 8 of the 10 best daily returns fall again into the years 2008
and 2009, amid them 3 incredibly high returns above 10% only in October and
November 2008. The 11.3% plus on 13-October-2008 was the highest daily return
in the history of the DAX since its inception in 1988. Investors took faith that the
world would not come to an end after the G7 presented a coordinated action plan
against the crisis and the German government announced guarantees for financial
institutions via a 500 billion EUR bailout package. Interesting is also the massive
but extremely short spike of the HFRX Equity Market Neutral index which occurs
shortly after 31-July-2007 (marked by red dot) due to the turmoils described in
the ”Volatility” paragraph. It is astounding that the spike disappears largely only
a few days later. The reason is that the spike is almost solely caused by the outlier
on 09-August-2007 (-3.09%) which gets compensated only 4 days later through the
outlier on 13-August-2007 (+3.10%).

24
Figure 6: Skewness, 2 year rolling window.

25
Figure 7: Skewness, 2 year rolling window.

Kurtosis The kurtosis measures the fourth moment of a distribution. A higher


kurtosis means that the tails are fatter which can be interpreted that the proba-
bilities of both - particular high as well as particular low / negative returns are
higher. As kurtosis incorporates the returns with the power of four, outliers have
a major impact. This can also be seen in our graphs: when extreme events enter
the time window, kurtosis shoots up just to go down as soon as these events leave
the window.

26
Figure 8: Kurtosis, 2 year rolling window.

27
Figure 9: Kurtosis, 2 year rolling window.

4.3 Correlations
Figure 10 shows in a heatmap the correlations of daily returns between our indexes.
The Bund is negatively correlated with almost all other hedge fund indexes and
stock markets. With the German DAX, the correlation is the most negative one.
This is the typical relationship between a bond and the stock market in one coun-
try or currency area, since central banks normally decrease interest rates during
recessions which lifts the prices of older, higher yielding bonds. Meanwhile, stock
markets are normally in decline during recessions. The opposite holds when the
economy is recovering and growing. This mechanism gets intensified since large
institutional investors like pension funds usually shift their asset allocations. In

28
turbulent times, a ’flight to safety’ can be observed and the weight of stocks in the
portfolio is decreased in favor of AAA-rated government bonds.
The S&P 500 TR, the DAX, HFRX Equity Hedged, HFRX Event Driven,
HFRX Merger Arbitrage and HFRX Relative Value Arbitrage seem to create a
cluster where correlations are a bit higher. The highest correlation can be ob-
served between HFRX Equity Hedged and HFRX Event Driven, which leaves
open questions. HFRX Equity Hedged is also relatively strong correlated with the
S&P 500 TR, which indicates that these funds maintain net long positions in the
US stock market over the long run. HFRX Fixed Income Convertible Arbitrage,
HFRX Market Neutral and HFRX Macro CTA seem to be uncorrelated in large
parts to the other indexes (since Convertible Arbitrage is a sub index of Relative
Value Arbitrage, this correlation is natural).

29
Figure 10: Correlations of daily returns

4.3.1 Correlations in turbulent times

Usually, asset correlations are not constant over time. They tend to go up in
turbulent periods. We define three ’states’ according to which we classify our
dates:

• Noisy: A date in our observations is classified ’noisy’, if for at least two


of the 10 indexes, abnormal returns were detected. Hereby, we define the

30
return of index x on day i as being ’abnormal’ if

|ri,x | > µx + 2 ∗ σx

where µx and σx denote the index’s mean return and its standard deviation
respectively.

• Normal: For none of our indexes an abnormal return was detected.

• Turbulent: For at least four of our indexes an abnormal return was detected.

For each of these four states we analyzed the correlations. Table 4.3.1 displays
a summary of this examination. As indicated by the mean (without the diagonal
elements), correlations rise during abnormal times.

mean correlation share of observations


Normal 0.141 77.70%
Noisy 0.207 9.02%
Turbulent 0.216 2.43%
All 0.195 100%

Figure 11 presents the correlations on ”noisy” days, Figure 12 the correla-


tions on ”turbulent” days. Figure 13 finally shows the correlations on ”normal”
days, which excludes noisy and turbulent days. The ’flight to safety’ effect be-
comes intensified on noisy and turbulent days. Also, the cluster of indexes which
is correlated anyway becomes much ’hotter’. However, the HFRX Fixed Income
Convertible Arbitrage, the HFRX Equity Market Neutral and the HFRX Macro
CTA indexes manage to remain basically uncorrelated with all others even on tur-
bulent days, which highlights the important role hedge funds can play to diversify
a portfolio. One way to incorporate the changes in correlation levels into a return
model would be by a Markov switching model, where the state of the economy is
changed according to a transition matrix, which allows to model that for example
in a recession, the economy can either shift into an early recovery or remain in
recession, each with a certain probability, and for each state, a separate correlation
matrix can be used.

31
Figure 11: Correlations on ’noisy’ days.

32
Figure 12: Correlations on ’turbulent’ days.

33
Figure 13: Correlations on ’normal’ days.

Correlation on a rolling time window Figure ?? displays some selected cor-


relations on a rolling two year window. We can see that the correlation between
DAX and S&P 500 TR is the highest all over the time, but the correlation spikes
shortly after the Lehman collapse and does not come down even when the event
vanishes from the window two years later, but rather remains on this new plateau
until about August 2013. Analyzing the correlation between the DAX and Bunds,
the financial crisis does not seem to have a substantial impact, even though the

34
ECB rates were falling sharply between October 2008 and April 2009. However,
during the height of the sovereign debt crisis, when investors where fleeing to the
save haven of German government bonds, the correlation reaches its lowest val-
ues. Remarkable is the correlation between Macro CTA and the S&P 500 TR:
during the financial crisis as well as during the sovereign debt crisis, the correla-
tion even comes down. This is in accordance with our previous findings: Figure
12 states that the correlation on ”turbulent” days becomes is slightly negative
(-0.01) whereas on ”normal” days, the correlation is positive (0.14). This makes
Macro CTA attractive from an diversification point of view. HFRX Fixed Income
Convertible Arbitrage is also basically uncorrelated with the S&P 500 and we get
a similar picture, though not quite as consistent, for the HFRX Equity Market
Neutral Index.

Figure 14: Correlations on a 2 year rolling window.

4.4 Portfolio Analysis


In the follow section we delve into the world of portfolio theory. We therefore start
with looking at four different scenarios.

35
1. Universe without hedge funds. This is a universe consisting of stocks and
bonds only. Stocks are represented by the DAX and the S&P 500 Total
Return, bonds by the German Bund.

2. Universe with hedge funds, stocks and bonds. As above, stocks are repre-
sented by DAX and S&P 500 TR, bonds by the Bund and hedge funds by
the seven HFRX indexes we introduced before.

Furthermore, we define two scenarios for restrictions. We define leverage here by


Debt/Equity.

a No leverage, long only. This means all weights need to be non-negative and
they need to sum up to one.

b Leverage up to 400%, short allowed. This means that the sum of absolute
weights minus one needs to be less or equal the leverage (Qian et al. (2007)).
So if we only operate with our own money, the (absolute) sum of the weights
equals one and our leverage is zero.

So in total, we have four different scenarios. For each scenario, we created ran-
dom portfolios with respect to the above mentioned restrictions out of respective
investment universe. Figure 15 displays scenario 1.a, so the portfolios all consist
of the DAX, the S&P 500 TR and Bunds without any short position and without
using leverage.

36
Figure 15: Scenario 1.a

Each portfolio is represented by a colored point. The color corresponds to


the Sharpe Ratio. We used the mean of daily returns (without annualization) for
our calculations, therefore the Sharpe Ratios are that unfamiliarly low (note that
the Sharpe is not time invariant). The efficient frontier is marked with red dots.
Figure 16 corresponds to scenario 2.a and displays random portfolios consisting of
the full investment universe including hedge funds but with the same restrictions
as in 1.a.

37
Figure 16: Scenario 2.a

Since the possible combinations rapidly, the area is by far less dense. The
graphic also could lead to the wrong impression that the achievable returns are
less than in scenario 1.a, which is of course not the case since possibilities have
increased and no new restrictions have been added, it is now just by far less
likely to ”hit” the maximum return portfolio with random weights. Therefore
we ran an optimization to obtain the maximum return portfolio for each given
portfolio volatility / standard deviation. The result of this is the efficient frontier
which is again highlighted by red dots and which is above the randomly generated
portfolios, just for the reason that the likeliness to ”hit” the maximum return
portfolios for each volatility have gone down close to zero due to the manifold
possible weight combinations. However, we can clearly see that by adding hedge
funds to the investment universe, it is now possible to construct portfolios with a
volatility close to zero which was not possible in the stock-plus-bond-only universe.
Figure 17 displays portfolios under scenario 1.b, so the universe excludes hedge
funds but we now allow for shorting and leverage.

38
Figure 17: Scenario 1.b

We are now able to produce much higher returns but at the price of increased
volatility. The best Sharpe ratios are in a similar range as in scenario 1.a, so the
relaxing of restrictions did not substantially benefit the Sharpe ratios. Figure 18
finally shows portfolios constructed from the full investment universe and with lax
restrictions.

39
Figure 18: Scenario 2.b

Again, due to the numerous possible combinations, it becomes increasingly


unlikely to randomly generate optimal portfolios, and due to the complexity, an
optimization like we ran for scenario 2.a was not possible any more. Generally, we
can see that very good as well as very bad combinations are possible here.

References
Alistair Barr, MarketWatch, 2010. Big liquidation triggers hedge-fund turmoil.
URL http://www.marketwatch.com/story/portfolio-liquidation-triggers-turmoil-amon

Dr. Galen Burghardt, 2010. Measuring the impact of trend following in the cta
space.
URL http://www.opalesque.tv/youtube/Galen_Burghardt/1

40
Gerald Braunberger, FAZ, 2010. Die auferstehung einer hedge-fonds-strategie.
URL http://www.faz.net/aktuell/finanzen/fonds-mehr/
convertible-arbitrage-die-auferstehung-einer-hedge-fonds-strategie-11023530-p3.
html

Hedge Fund Research, Inc., 2017. https://www.hedgefundresearch.com/indices/hfrx-


eh-equity-market-neutral-index.
URL https://www.hedgefundresearch.com/indices/
hfrx-eh-equity-market-neutral-index

Pensions and Investments, 2011. Institutional share growing for hedge funds.
URL http://www.pionline.com/article/20110210/ONLINE/110219980

Pensions and Investments, 2016. More hedge fund firms seeing a decline in assets.
URL http://www.pionline.com/article/20160919/PRINT/309199980/
more-hedge-fund-firms-seeing-a-decline-in-assets

Preqin, 2013. Event driven remain the top performing strategies for hedge funds.
URL http://www.valuewalk.com/2013/11/top-performing-strategies-for-hedge-funds

Qian, E. E., Hua, R. H., Sorensen, E. H., 2007. Quantitative equity portfolio
management: modern techniques and applications. CRC Press.

41

Potrebbero piacerti anche