You are on page 1of 386

Prof. Dr.

Bernd Scherer

Capital Markets

May 2010
Dr. Bernd Scherer
Professor of Finance, EDHEC Business School
Member of EDHEC Risk

1
Prof. Dr. Bernd Scherer

Sessions

• Session 1: The Financial System


• Session 2: Bank Management and Bank Regulation
• Session 3: Capital Markets and Their Returns
• Session 4: Equity Markets
• Session 5: Futures, Options and Principal Protection
• Session 6: Fixed Income Markets and Products
• Session 7: Currency Markets and Products
• Session 8: Credit Markets and Credit Products
• Session 9: Volatility Products
• Session 10: Risk Management

2
Prof. Dr. Bernd Scherer

Teaching Format

• The course will comprise a total of 30 hours split roughly into 10 sessions of
3 hours.

• The lecture notes are thought as a guidance to the literature. Lecture notes
can not replace studying the suggested literature, neither is reading the
literature a substitute for the course.

3
Prof. Dr. Bernd Scherer

Course Assessment

• The exam will be a take home exam, where students have to answer a set
of questions and hand in their solutions two weeks after they received the
questions.

4
Prof. Dr. Bernd Scherer

Session 1: The Financial System

1. Financial System
– How does the financial system promote efficiency?

2. Principles of Market Valuation


– Why is market valuation important in financial economics?

3. Financial Structure
– What are the basic facts of financial structure around the world?

5
Prof. Dr. Bernd Scherer

Literature

• MISHKIN (2007), The Economics of Money Banking and Financial


Markets, Pearson

• BODIE/MERTON/CLEETON (2009), Financial Economics, Pearson,

• COPELAND/WESTON/SHASTRI (2005), Financial Theory and


Corporate Policy, 4th edition

6
Prof. Dr. Bernd Scherer

1. The Financial System

• Flow of funds (direct versus indirect finance)


• Functions of financial markets (direct finance)
• Functions of financial intermediaries (indirect finance)

7
Prof. Dr. Bernd Scherer

Flow of Funds - channeling funds

Financial
Intermediaries
Indirect
finance

Lenders/Savers Borrower/Spender
Financial
Households Corporates
Corporates Markets Governments
Governments Households

Direct Finance: borrowers borrow directly from lender; no intermediaries

8
Prof. Dr. Bernd Scherer

Real Economy versus Financial Markets

• Ultimate investors sell financial assets to savers; they use the proceeds to
buy real assets (buying real assets is the same thing as investment).

• A real asset is an entity that generates a flow of goods or services over


time. Examples include land, people, factories, inventions, business plans,
goodwill with consumers, reputation. Anything that generates a flow of
goods or services counts. Key point: real assets need not be tangible.

• A financial asset is a legal contract that gives its owner a claim to payments,
usually generated by a real asset. Examples include currency ($), stocks,
bonds, bank deposit, bank loans, options, futures, etc.

• The FS is the place where savers (or, more generally, economic agents with
a surplus of funds relative to their immediate need for those funds) meet
investors (or, more generally, economic agents with a deficit of funds
relative to their immediate need for those funds).

9
Prof. Dr. Bernd Scherer

Functions of the Financial System –


Transfer Economic Resources Across Time And Space
• The first goal of the financial system (FS) is to facilitate the flow of funds
from savers (entities with a surplus of funds) to investors (entities with a
deficit of funds).

• Across Time
– By a house today rather than have to wait until to the rest of your life when you
will have save all the money.
– Go to university and pay back your fees after graduation allows you to develop
skills early in life.

• Across Space
– Channel investments to places with higher marginal returns
– Example: Provide capital to Russian oil firms either via stocks and bonds (direct
finance) or bank loans (indirect finance)

• Provide the information that allows decentralized decision making. This is


key in a market economy for allowing growth!

10
Prof. Dr. Bernd Scherer

Functions of the Financial System –


Risk Sharing

• Offload risks to markets


– Lufthansa hedges out oil risk or Porsche its Dollar risk

• Allow separation of management and ownership


– Usually people with specific skills don’t have the money and people with money
don’t have the specific skills
– Even, if they had the skills and the money would it be wise to invest all your
wealth in your own business?

• Create securities with tailored risk profiles


– Options, CDS, CDOs
– Create complete markets where all claims could be traded at minimal costs

11
Prof. Dr. Bernd Scherer

Functions of the Financial System –


Banking, Clearing and Settlement

• Clear and settle payments securely and efficiently (low costs) to


facilitate exchange of goods and services. Example: money, credit
cards, SWIFT transfers, etc.

• Reduce Transaction Costs


– Economies of scale (standardized loan contracts, instead of everybody writnig his
own contract)
– Provide liquidity (bank accounts, checks)

12
Prof. Dr. Bernd Scherer

Functions of the Financial System –


Reducing Asymmetric Information

Often one party of a deal lacks the information to make a good decision

• Adverse selection (happens before the transaction)


– How does the borrower know which are the bad risks (they will be the most keen
to get a loan)?
– How does the investor know which firms are not overvalued (those will be the
most keen to issue new equity)?

• Moral Hazard (happens after the transaction)


– How do creditors protect themselfes against risk shifting (asset substitution), i.e.
increasing the risks after the bond has been issued?

13
Prof. Dr. Bernd Scherer

2. Principles of Market Valuation

• Understand the law of one price as the main principle in finance


• Understand how information gets reflected in security prices
• Fair Value or Accounting Value

14
Prof. Dr. Bernd Scherer

Value Maximization

• Managers make decisions on behalf of shareholders they never met in


their life. How can they do this?

• Principle 1: Value maximization. Managers need only maximize value


irrespective of preferences. Example: You like oranges and hate
apples. You are offered for free a box of apples (market value of 100)
and a box of oranges (market value of 90). Which should you choose?
How do you compare apples and oranges. Answer: By price! Your
taste does not matter

• Principle 2: Financial markets provide the information needed if all


information is allowed to move into prices: no short sales restrictions,
no ban on CDS , etc. What do you do if assets are not traded? Use
comparables!

15
Prof. Dr. Bernd Scherer

Fisher Separation –
Production and Consumption

• Endowment: y1 y0

• Production along the


U2 transformation curve raises
U1 utility to higher level

• The optimal production plan


depends on utility –
Impossible for manager to
make the right choice on
behalf of shareholders

y1

y0

16
Prof. Dr. Bernd Scherer

Fisher Separation –
Production and Consumption with Capital Markets

• Introduction of capital market


increases welfare

• Current wealth can be


transformed across time
U2 U 3 y
Wo = yo + ( 1+1r )
U1
• We can separate the choice of
optimal investment from
consumption preferences –
Slope:
-(1+r) • Managers don’t need to know
shareholders preferences –
they just pick projects with
highest NPV
y1

y0

17
Prof. Dr. Bernd Scherer

Arbitrage and the Law of One Price

• Identical cash flows must have identical prices! Arbitrage ensures


relative pricing, not absolute pricing!

• Examples:

– Bonds: Coupon bond must equal the sum of implicit zero bonds
– Currency: If we know USD/EUR and EUR/YEN exchange rate we also know the
right value
– Capital structure arbitrage: When credit spreads move so should stock prices
– Derivatives: Option price equals value from replicating strategy

18
Prof. Dr. Bernd Scherer

Efficient Market Hypothesis –


Rational Expectations in Financial Markets
• Explains how all relevant information gets into prices (almost
instantaneously).

• Prices in financial markets will be set such that the optimal forecast - using
all available information - equals the securities equilibrium return (investors
get a fair compensation for risk, no more).

• This is ensured by the competition of investors (financial markets are the


oldest prediction markets).

• Implications: Abnormal returns are unpredictable, day trading is dangerous,


Crisis is unpredictable (credit crisis!), … Does not mean markets make no
errors!

• Prices move because expectations shift as new information arrives

19
Prof. Dr. Bernd Scherer

Efficient Market Hypothesis –


Rational Expectations in Financial Markets

• Markets have no memory: any sequence of past (stock) price changes


contains no information about future changes

• Trust market prices: in an efficient market, there is no way for most


investors to achieve consistently superior rates of return

• Read the entrails: prices (stock / bond) impound all available information –
example: if the firm’s bonds are offering a much higher than average yield,
one can deduce that the firm is most probably in trouble

• There are no financial illusions: investors are unromantically concerned


with the firm’s cash flows

• The do-it-yourself alternative: investors will not pay others for what they
can do equally well themselves (e.g. diversify, create leverage)

20
Prof. Dr. Bernd Scherer

Efficient Market Hypothesis –


Rational Expectations in Financial Markets
What Makes Markets Efficient?
• There are many investors out there doing research. As new information
comes to market, this information is analyzed and trades are made based
on this information.
• Therefore, prices should reflect all available public information If investors
stop researching stocks, then the market will not be efficient
(GROSSMAN/STIGLITZ argument)

Common Misconceptions about EMH


• Efficient markets do not mean that you can’t make money
• They do mean that, on average, you will earn a return that is appropriate
for the risk undertaken and there is not a bias in prices that can be
exploited to earn excess returns
• Market efficiency will not protect you from wrong choices if you do not
diversify – you still don’t want to put all your eggs in one basket

21
Prof. Dr. Bernd Scherer

Efficient Market Hypothesis –


Rational Expectations in Financial Markets
• Strong Form Efficiency: Prices reflect all information, including public and
private
– If the market is strong form efficient, then investors could not earn abnormal returns
regardless of the information they possessed
– Empirical evidence indicates that markets are NOT strong form efficient and that insiders could
earn abnormal returns

• Semi-strong Form Efficiency


– Prices reflect all publicly available information including trading information, annual reports,
press releases, etc.
– If the market is semi-strong form efficient, then investors cannot earn abnormal returns by
trading on public information
– Implies that fundamental analysis will not lead to abnormal returns

• Weak Form Efficiency


– Prices reflect all past market information such as price and volume
– If the market is weak form efficient, then investors cannot earn abnormal returns by trading on
market information
– Implies that technical analysis will not lead to abnormal returns
– Empirical evidence indicates that markets are generally weak form efficient

22
Prof. Dr. Bernd Scherer

Should Investors Look at Market or Book Value?


• Book value is a fantasy price that does not resemble market value nor can it
be transacted on. If all prices in an economy where accounting values
capital could not be allocated efficiently.

• Asymmetric recognition of losses on balance sheet (losses on active side


must be reflected, while losses on passive side must not) incorrectly
portrays hedges.

Assets Liabilities Assets Liabilities

100 million 100 million 90 million 100 million


Bond of owed to Bond of owed to
company C company C company C company C

23
Prof. Dr. Bernd Scherer

3. Financial Structure

• Basic facts about financial markets


• The role of asymmetric information
• Financial crisis and economic activity

24
Prof. Dr. Bernd Scherer

Stylized Facts About Financial Structures

1. Stocks are not the most important source of external financing


– The US stock market only accounts for 11% of external financing between 1970
and 2000; The same applies to other countries around the world.
– Why do companies rely more on other sources?

2. Issuing marketable securities is not the primary way in which businesses


finance their operations
– Why do companies not rely more on marketable securities

3. Indirect finance is many times more important than direct finance


– Why are financial intermediaries so important?

25
Prof. Dr. Bernd Scherer

Stylized Facts About Financial Structures

4. Banks are the most important sources of external funds


– Bank loans are the most important source of external finance
– What are the reasons for this and why is this share declining?

5. The Financial System is the most heavily regulated part of the economy
– Why are financial markets so regulated?

6. Only large, well established corporations have easy access to security


markets (direct finance)
– Why do smaller firms have to recur to banks?

26
Prof. Dr. Bernd Scherer

Stylized Facts About Financial Structures

7. Collateral (property/assets pledged to lender) is a prevalent feature of


debt markets
– Why is collateral such an important feature of debt markets?

8. Debt contracts are complicated legal documents that place substantial


restrictions (covenants) on the borrower.
– Why are debt contract so complicated and restrictive?

27
Prof. Dr. Bernd Scherer

Transaction Costs (TC)

• The Problem: Fixed transaction costs crowd out low volume business. If you
have only 1000 Euros invest you can not get diversify across international
assets, commodities, hedge funds in your portfolios. Lower income
population gets cut out! How can financial intermediaries help to reduce
TC?

• Economies of Scale (reduction in TC per dollar of investments)


– Bundling of transactions (mutual funds)
– One legal master document from the best lawyers available can be used for all
personal loans

• Expertise (learning curve)


– IT services (call centre / online banking, etc.)

28
Prof. Dr. Bernd Scherer

The Lemon Problem – Adverse Selection (Hidden Information)

• The original lemon paper (AKERLOF, 1970): Potential buyers of used cars can
not evaluate the true value of a car. There is asymmetric information (seller
knows the true quality of his car) .
– Buyer suspects car is bad (lemon) and offers a low price
– Lower prices crowd out the seller of good cars
– Only bad cars are on offer – market does not function

• Adverse selection costs in finance (MYERS & MAJLUF, 1984) arise because two
types of firms might raise capital
• Firms that need to finance positive-NPV projects (starting the project is good for
existing investors, new investors get required rate of return)
• Firms whose claims are overvalued (raising capital transfers wealth from new investors
to existing investors!)

29
Prof. Dr. Bernd Scherer

The Lemon Problem

• Existing managers act in the interest of existing shareholders (because of


compensation contracts) and issue claims whenever they are overvalued or a
positive-NPV project emerges

• Outside investors cannot distinguish between overvalued firms and those with
positive-NPV projects
– New investors would have, on average, a lower than required return
– Potential investors understand this, and offer to pay LESS so that the earn
their required rate of return on average
– underinvestment (positive-NPV projects are not realized because
necessary capital can only be raised at additional cost)

• The lemon problem explains why marketable securities are not the
primary source of funding (Fact #2)
• It also explains why there is a pecking order (prefer internal funds over
bonds over equity as the degree of information release is different)

30
Prof. Dr. Bernd Scherer

Numerical Example

• Suppose the management knows the intrinsic value of its assets (a) as well
as the NPV of a new project (b). These values are not known to capital
markets. In order to fund the project the company needs an amount of I.

• Let P and P* be the market value of the old stockholders’ shares before and
after issuance. Note that P does not need to equal a. In fact we have argued
that companies are reluctant to finance new projects if their stocks are
undervalued P < a.

• The total value of the firm after issuance is the sum of the money raised
from new shareholders and the market value of old stockholders’ shares (I
+ P*). Note that any price drop due to adverse selection is already included
in P*.

31
Prof. Dr. Bernd Scherer

Numerical Example

• Old shareholders now own only a fraction of company assets. The company
will issue new stock if the old shareholders’ participation on assets after
issuance, namely I + a + b, is larger that the 100% participation in a. The
condition for issuance becomes

• In order for us not to need to specify the mechanism to determine P* in


equilibrium, we simply assume that a >P = P*. The company can issue
shares at P = P*. Suppose a company trades at P = 50 and can issues shares
at p* = 50. However, it knows the intrinsic value of its assets is a = 90 and it
has a positive NPV project of b = 10, which needs funding of I = 20.

32
Prof. Dr. Bernd Scherer

Numerical Example

• In this setting, issuing equities is not good idea for old stockholders. While
the value of the company would eventually increase to 120 (90 + 10 + 20),
they get only a share of 71.4% (50/70), which totals 85.71.

• Not issuing equities and giving up the positive NPV project gives them 100%
in the intrinsic value of 90. The result of this form of adverse selection is
underinvestment.

• Positive-NPV projects are not realised because necessary capital can be


raised only at additional cost.

33
Prof. Dr. Bernd Scherer

How to overcome the lemon problem?

• A. Private information production: private companies (rating companies,


etc.) sell information that identifies good and bad companies.
– Limited success due to free rider problem (copy other peoples action without
paying for the information)

• B. Government Regulation: forced disclosure


– Does not always work well (see ENRON)
– (Allow insider trading)

• C. Financial intermediation
• Banks build expertise in evaluating good and bad firms, acquire funds from
depositors and start lending to good firms.
• Similar model for cars (used car dealership, guarantees on used cars, BMW sells
used BMWs)

34
Prof. Dr. Bernd Scherer

How to overcome the lemon problem?

• This explains facts #3 and #4 (dominance of non market finance).


– Financial intermediaries produce private information that they can directly profit from
– Banking market needs competition!

• This also explain fact # 6


– Large established firms go to markets (as there is already a lot of information on
them) while small firms go to banks

• D. Collateral
– Collateral (positive equity) protects the lender in times of default.
– Example: Mortgage market

• This explains facts #7

35
Prof. Dr. Bernd Scherer

The Moral Hazard Problem – Hidden Action

• Definition: Engage in activities that are undesirable for the purchaser of the
issued security
– Bondholders are hurt if management increases business risks.
– This helps shareholders and hurts bondholders. Why?

• Costly state verification: moral hazard increases transaction cost (


monitoring costs)

• Venture capital: manager and owner are the same person, solves
defrauding equity investors, but what about bondholders?

• Debt Contracts

36
Prof. Dr. Bernd Scherer

Influence of Moral Hazard on Debt Markets

• Collateralization

• Covenants
– Discourage undesirable behaviour (loan money must only be used for particular
project, business risk must remain the same)
– Encourage desirable behaviour (no further leverage allowed, no external funding
of pension plans)
– Keep collateral value (required maintenance until loan is paid off)
– Information provision

• Moral hazard is a feature of capital markets separation of ownership and


management) that can only be mitigated.

37
Prof. Dr. Bernd Scherer

Liquidity

• Liquid assets are more valuable than illiquid assets. If you buy an illiquid asset and
transform into a liquid assets you made a profit!

• Information is the key to liquidity. If the owner knows more than the buyer he will
have difficulties to sell at a fair price. Result: adverse selection.

• How to create liquidity:


– Provide information to increase liquidity: detailed prospectus, rating
– Share losses: own stake, reputation loss
– Split asset into two assets: one safe and therefore liquid and one illiquid and
highly risky assets.

38
Prof. Dr. Bernd Scherer

Conflicts of Interest

• Multiproduct financial intermediaries enjoy economics of scope (lower the


cost of information production for each service by applying on resource to
all services)
– However, this comes at a cost: conflicts of interest: which product to sell? Best
advice or highest profit?

• Why do we care about conflicts of interest?


– Reduction in the quality of information in financial markets increases asymmetric
information problems and hence stops financial markets to work!

39
Prof. Dr. Bernd Scherer

Types of Conflicts of Interest

• Underwriting and Research


– Investment banks sell research (evaluate prospects of a company) on companies
as well as underwriting (buy stocks from company to sell it to the market) equity
issuance
– Where do you make more money from? How independent is research? How
much arm twisting is going on?

• Auditing and Consulting


– Do auditors skew their opinions to win consulting business?
– Do auditors pass inside information to consulting unit?
– Do auditors criticize the systems created by their own consulting unit?

40
Prof. Dr. Bernd Scherer

How Important is the Financial System to Promote Growth?

• Collateral and restrictive covenants are important tools to solve adverse


selection and moral hazard.
– Developing / ex-communist countries lack the legal system and efficient bankruptcy
procedures
– This increases TC (information costs) to solve adverse selection
– Collateral can not easily be seized: moral hazard prevails

• State owned banks do not manage their capital efficiently


– Politically important projects get preferred conditions
– Loans for the boys,…

41
Prof. Dr. Bernd Scherer

Financial Crisis and Economic Activity -


Factors Causing Crisis

• Increase in interest rates


– With higher rates the good risks are crowded out while the bad risks will still want to borrow.
The resulting adverse selection will make it unattractive to provide loans. Aggregate activity
falls.
• Increase in Uncertainty
– Stock market crash, bank failure makes it more difficult to identify the good borroers. Credit
market freezes
• Deterioration of Balance Sheets
– Deterioration in collateral
• Banking Sector Deterioration of Balance Sheets
– Increases adverse selection and moral hazard
• Government Imbalances
– Capital flight, black economy …

42
Prof. Dr. Bernd Scherer

Depository Institutions

• Create financial assets (deposits) to fund real assets (houses, etc)

• Key function: asset transformation: transform low liquidity left side of balance sheet
(assets) into high liquidity right side (deposits)
– Deposit insurance
– State guarantees
– Reputation, risk management, financial capital

• Depository institutions bundle functions of the financial system


– Create financial assets
– Enhance liquidity
– Risk management, sharing

43
Prof. Dr. Bernd Scherer

Investment Banks

• The securities industry is involved in brokerage services for both the


primary and secondary securities markets.
• The key functions of securities firms are:
– Securities Underwriting: Marketing and distributing new issues of debt and
equity for corporations.
– Brokerage and Market Making: In market making, a securities dealer holds an
inventory of a security and stands ready to buy the security from anyone at the
bid price, and sell the security at the ask price. Other firms act strictly as brokers,
that is, agents that conduct trades on behalf of customers.
– Trading: This activity is similar to market making, although the firm may ben
engaged in speculating about the direction of price change.
– Mergers & Acquisitions (M&A): Assisting firms wanting to expand by buying
another company, or wanting to sell out to another company. This entails finding
merger partners, provide advice about deal pricing and structure, underwriting
securities to facilitate a deal, and so on.

44
Prof. Dr. Bernd Scherer

Fund Management

• Mutual funds are FSF that pool the resources of individuals and firms and invest those
funds in a large, diversified portfolio of financial assets.

• Money market mutual funds, Bond funds, Stock funds, Pension funds

• Mutual funds are required by regulation to disclosure their investment objectives in a


prospectus, to disclose their holdings at regular intervals, and are limited in the
degree of short selling that they can do.

• Hedge funds: These are pooled investment vehicles like mutual funds but they are
very lightly regulated. Hedge funds usually raise funds from wealthy investors and also
from pension funds, University endowments and state funds. They avoid most
regulations on disclosure requirements and are unrestricted in the kinds of assets that
they can hold. Also, hedge funds face no restrictions on short positions.

45
Prof. Dr. Bernd Scherer

Session 2: Banking Theory

1. Bank Management
– Understand the general principles of bank management

2. Banking Regulation
– Understand the credit crisis and the current debate on bank regulation

46
Prof. Dr. Bernd Scherer

Literature

• MISHKIN (2007), The Economics of Money Banking and Financial


Markets, Pearson

• MATHEWS/THOMPSON (2008), The Economics of Banking, Wiley

47
Prof. Dr. Bernd Scherer

1. Bank Management

• Asset Transformation – liquidity creation

Assets Liabilities

Required Reserves 10 Deposits 100


Excess Reserves 90

Assets Liabilities

Required Reserves 10 Deposits 100


Loans 90

• Transform deposits into longer term loans, hold 10% cash buffer reserve
requirement. Money is created in the banking system!

48
Prof. Dr. Bernd Scherer

Liquidity Management

• How to deal with deposit outflows?


• Hold excess liquidity reserves

Assets Liabilities

Reserves 20 Deposits 100


Loans 80 Bank Capital 10
Buffer Securities 10
against sale Insure
of illiquid against
assets default

• After a 10 million outflow in deposits the bank has still enough reserves (10 million >
10% x 90 million)
Assets Liabilities

Required Reserves 10 Deposits 90


Loans 80 Bank Capital 10
Securities 10

49
Prof. Dr. Bernd Scherer

Liquidity Management
• What happens if reserves are not high enough? Example: 10 million reserve loss.
Assets Liabilities

Required Reserves 10 Deposits 100


Loans 90 Bank Capital 10
Securities 10

Assets Liabilities

Required Reserves 0 Deposits 90


Loans 90 Bank Capital 10
Securities 10

• The bank has a reserve requirement of 9 million, (10% from 90 million), but no
reserves. What can it do?

50
Prof. Dr. Bernd Scherer

Liquidity Management – How To Cover For A Loss in Reserves

1. Borrow in the interbank market (What happened in the credit crisis?)


Assets Liabilities

Deposits 90
Required Reserves 9
Borrowing 9
Loans 90
Bank Capital 10
Securities 10

2. Sell securities. Liquidity crisis becomes a solvency crisis when assets need to be sold
on fire sales terms.

3. Borrow from the Central bank using securities as collateral (not all securities qualify!)

4. Sell loans (costly due to adverse selection, particular in a crisis)

• Banks will hoard cash (excess reserves) when the costs of from making up
for deposit outflows are high.

51
Prof. Dr. Bernd Scherer

(Bank) Asset Management

• Buy securities with high return and low risk


• Avoid default
• Invest in liquid assets
• Diversify

52
Prof. Dr. Bernd Scherer

(Bank) Liability Management

• Deposits

• Selling short term paper CD

• Ensure funding is not too volatile: Short term funding might cease in a crisis
(difficult to rollover) while long term funding is expensive.

• COCOs (Contingent Convertible Bond) Bonds that convert into equity in


times of market stress (for example if the issuer is a bank the condition may
be that it will convert to equity if the issuer's tier one capital falls below a
limit. The automatic conversion will then re-capitalise the bank)

53
Prof. Dr. Bernd Scherer

Capital Adequacy Management

• Compare two differently capitalized banks (capital/asset ratio of 10% versus 4%

Assets Liabilities Assets Liabilities

Required Reserves 10 Deposits 90 Required Reserves 10 Deposits 96


Loans 90 Bank Capital 10 Loans 90 Bank Capital 4

• Assume a 5% loan loss


Assets Liabilities Assets Liabilities

Required Reserves 10 Deposits 90 Required Reserves 10 Deposits 96


Loans 85 Bank Capital 5 Loans 85 Bank Capital -1

• Bank capital lessens the chance a bank will become insolvent

54
Prof. Dr. Bernd Scherer

Bank Capital

• ROA (return on assets) tells us how efficiently a bank is run


net profit after taxes
ROA =
assets
• ROE (return on equity) tells us how well the owners are doing (in an
accounting sense)
net profit after taxes
ROE =
Equity
Assets
= ROA ⋅
Equity

• Lower bank capital yields higher returns for bank owners (ROE); trade-off
versus safety
• Bank need to hold capital for regulatory reasons as bank default imposes an
externality on the economy.

55
Prof. Dr. Bernd Scherer

2. Banking Regulation

• Categories of Banking Regulation


– Deposit Insurance
– Restrictions on bank assets
– Capital requirement
– Chartering
– Risk management
– Disclosure
– Consumer protection
– Restriction on Competition

56
Prof. Dr. Bernd Scherer

History of Banking Regulation

• Response to 1929 crash; US created deposit insurance and separated investment


banking from deposit taking
– Regulator wants to protect depositors and ensure financial stability as a prerequisite for
growth
– Regulatory capital aims at avoiding bank failure

• Economic view: regulation is unnecessary


– Imprudent banks find it difficult to attract capital and might experience runs
– Shareholders are disciplined by market forces
– Banks set their economic capital to avoid default (in most situations) as frictional
bankruptcy costs are high

• Regulator wants large regulatory capital (rather safe than sorry)


– Affects allocation efficiency (shareholder need fair return)
– Regulation assumes static exposure while banks can react
– Banks can not manage systemic risks; this should be done by the central bank

57
Prof. Dr. Bernd Scherer

Deposit Insurance

• Suppose a bank looks troubled. Short term finance will not get rolled over and the
creditors withdraw funds, deposits, nobody wants to trade with that bank; bank needs
to sell illiquid securities at fire-sale prices and liquidity crisis becomes solvency
problem.

• Deposit insurance creates a moral hazard problem. Depositors do not spend


information costs on banks if deposits are explicitly insured. There is no market
discipline.

• Deposit insurance also creates an adverse selection problem. Bad risks are likely too
benefit from deposit insurance so its difficult to price.

• Too big to fail (banks take on even bigger risks and morph into “too big to save”)

58
Prof. Dr. Bernd Scherer

Basel I: Definition of Capital

• Motivation in 1988: Strengthen the financial system and global standards


• For commercial banks a minimum capital is needed as buffer against large
unexpected losses
• Tier I capital
– Book equity (difference in liquidation value of assets and liabilities)
– Includes capital injection, retained earnings, …
• Tier II capital (supplementary capital)
– Undisclosed reserves, valuation reserves
– Subordinated debt with more than five years maturity
– Subordination means it is a buffer for depositors
– Limited to a maximum of 50% of bank capital

59
Prof. Dr. Bernd Scherer

Basel I: Risk Weighted Assets-Cooke Ratio

• The initial proposal was only concerned about credit risk


– Banks must maintain risk capital of at least 8% of risk weighted assets
– Risk weighted assets are assets weighted by a risk factor
Weights
(%) Assets

0 OECD Government, Cash, Gold, …


Claims on OECD banks, US
20 government Agencies

50 Uninsured residential mortgage


All other claims such as all
100 corporate bonds, EMD,…

• Example: 100 million AAA bonds and 100 million B bonds and 100 million US treasury bonds
amount to risk weighted assets of 200 million. For this the bank needs at least 16 million in
capital from which at least 4 million must be tier 1

60
Prof. Dr. Bernd Scherer

The 1996 Amendment – Trading versus Loan Book

• Include market as well as credit risks

• Distinguishes between trading and banking (loan) book


– Loan book: not marked to market: continue with risk weightings
– Trading book: daily marked to market and for most large banks risk capital was
calculated with internal models that accounted for correlations and proprietary risk
models

market − risk − capital = k ⋅ VaR + SRC

– VaR is the maximum of today’s and the last 60 days trailing VaR, k is a multiplier set by
the regulator and SRC is a specific risk charge (to account for idiosyncratic risks)

• Total capital needed is now 8% on credit risk weighted assets plus market risk weighted
assets)

61
Prof. Dr. Bernd Scherer

Basel II - Regulator tries to catch up

• Three Pillars
– Regulatory requirement (risk based capital charges)
– Supervisory review (expanded role for supervisors)
– Enforcing market discipline (disclosure rules, that yield to large shareholder
and depositor scrutiny)

• Include operational risks

• Finer buckets for credit risks, i.e. rating based approach


– Incentive to come up with rating agency sponsored fake ratings
– “AAA” bonds that have yields like BBB bonds

• Credit crisis suggests we are not done yet: Securitization!

62
Prof. Dr. Bernd Scherer

Problems with Basel I: Regulatory Arbitrage

• Economic and regulatory risk capital have not been aligned

• Insensitivity to credit risks that bordered to madness (same risk weighting


independent of Rating)
– Incentive to lower credit quality to align regulatory and economic capital
more closely

• Accelerated Securitization repackage loans as tradable securities and hold them


in trading book where the same assets get a lower risk weighting
– Groundwork for today’s crisis

63
Prof. Dr. Bernd Scherer

Why Regulation?

• Response to market failure ; External effects


– Systemic risks
– Government bailout put (too big to fail)
• Deposit insurance
– Liquidity gap between assets and liabilities (sounds familiar?)
– Bank run becomes self fulfilling
– Regulation needed because of double moral hazard
– Bank side: increase the risk and put it to the deposit insurance company if bust
– Customer side: don’t check banks and simply go for highest rates

64
Prof. Dr. Bernd Scherer

The Current Credit Crisis

• The crisis started off as a solvency crisis. Falling (less strongly rising) house prices
meant that subprime borrowers could no longer borrow against rising house as a
result of rising interest rates.
– tightening standards rising foreclosures, falling value of collateral, …

Source: MARTINO/DUCA (2007), The Rise and


Fallof Subprime Mortages, Economic Letter

65
Prof. Dr. Bernd Scherer

The Current Credit Crisis

• The solvency crisis created liquidity problems.

• Investors did not know how the losses (everybody knew were there) were distributed
among banks. CDS made it difficult to figure out which banks took the hit – a massive
adverse selection problem stopped the interbank market as a source of liquidity.

Weakenend balanced sheets


House price decline
& adverse selection

Breakdown of interbank market


Anticipation of firesales

66
Prof. Dr. Bernd Scherer

How Was The Crisis Fought?

• Paulson Plan – provide liquidity. Fed, ECB pumped massive liquidity into the
financial system
– Banks took the cash and hoarded it. Nobody wanted to lend to a failing
institution

• Collapse of AIG, Lehman – It became clear this was a solvency crisis, Banks
had to be recapitalized (stress test) to start lending again.
– Invest directly into banks
– Takeover failed institutions (Fannie and Freddie are US government
companies)
– Expand liability insurance

• Banks started to lend again

67
Prof. Dr. Bernd Scherer

Long Run Consequences

• Fewer and larger banks


– Survivors enjoy oligopoly rents

• Too big to fail


– Total loss in market discipline! Moral hazard is encoraged

• Long run inflation

• Debasing of Currencies

• Loss of Civil Liberties

68
Prof. Dr. Bernd Scherer

Session 3: Risk and Return in Capital Market Returns

1. Historical Equity Risk Premium

2. Risk and Return Calculations

3. Equity Risk Premium Puzzle

4. Alternative Equity Risk Premium

5. Inflation and Asset Returns

69
Prof. Dr. Bernd Scherer

Literature

• MISHKIN (2007), The Economics of Money Banking and Financial


Markets, Pearson

• BODIE/MERTON/CLEETON (2009), Financial Economics, Pearson

• CORNELL (1999), The Equity Risk Premium, Wiley

70
Prof. Dr. Bernd Scherer

We Could All Be Rich Fallacy

• „ An investor that would have invested


200 years ago one Euro into Equities
would (assuming a 10% return per
Stock annum) sit on 189 million Euros today
Market
With 5% return (bonds) tbis would
accumulate to 17.000 Euro instead.
This shows the advantage of equities
Real Factories Financial and the advantage of a investing
Assets Human Assets pensions into equities. The whole
Capital, etc.
society should do this.
Bond
Market • Why is this statement wrong?
– Survivor bias
– Economic leverage changes and so change
the returns on equity
• The real side and the financial side of
the economy are tightly linked.

71
Prof. Dr. Bernd Scherer

Understanding Returns

• Capital Market History I

72
Prof. Dr. Bernd Scherer

Understanding Returns

• Capital Market History II


– 1928 - 2009

73
Prof. Dr. Bernd Scherer

A Poor Mans Valuation Model

• Given that financial assets just represents claims against real assets we can plot
financial assets against real assets to see whether the former are inflated.

Bubble

74
Prof. Dr. Bernd Scherer

Sources of Return in Global Markets

• Expected Productivity of Capital Goods (goods that are used to produce other goods)
– Mines, roads, canals, machinery, power stations, patents
– The higher the productivity the higher the interest rate (=slope of production function)

• Uncertainty about the productivity of capital goods


– Equities represent a claim on to the return on capital
– Risk premium rises

• Time Preferene
– The greater the preference for current consumption, the higher the interest rates in the economy

• Risk Aversion
– Rising risk aversion leads to a larger equity risk premium

75
Prof. Dr. Bernd Scherer

Calculating RIsk (Volatility)

• „Square root of time rule“

P   P P P
ln  n  = ln  n n−1 ... 1 
 P0   Pn−1 Pn−2 P0 
 P  P  P  P
= ln  n  + ln  n−1  + ln  n−2  + ⋯ + ln  1 
 Pn−1   Pn−2   Pn−3   P0 
= rn + rn−1 + rn−2 + ⋯ + r1
Var (∑ )
r = Var ( rn ) + Var ( rn−1 ) + ⋯ + Var ( r1 )
i =1...n i

= nVar ( r )
= nσ 2

• For high frequency observations the estimation error on volatility goes to


zero. In this sense we can observe volatility, which is assumed
unobservable for low frequency data (daily, weekly, ...)
• Note that the same is not true for returns!

76
Prof. Dr. Bernd Scherer

Basics of Long Term Risks and the Central Limit Theorem (CLT)
• The CLT says that for independent (no autocorrelation in first or second
moment) and identically (can be non-normal as long as they have finite
variance) distributed random variables
– their sum is normal
– their product is log-normal

• If you look at diversified (little „factor risk) stock market investments over a
long period of time the distribution of log returns is likely to be close to
normal; this is not the case for short term risks on concentrated portfolios!

77
Prof. Dr. Bernd Scherer

Central Limit Theorem and Options

• From one period to multiperiod portfolio insurance

• Investors with long horizon overpay for something they dont need (receive)

78
Prof. Dr. Bernd Scherer

Basic Risk Calculations

Calculate the probability that an asset looses more than 10% in value by
year end.

– Change your return calculations from discrete to continuous returns and calculate
expected value and volatility.
– For example
 P  P  µ = 9.71%
ln ( 1 + R ) = ln  1 + 1 − 1  = ln  1 
 P0   P0  σ = 19.48%
– Convert the target return of -10% in a continuos return of ln(1-10%)
– Calculate the standardised difference between continuous mean and target return
 −10.54% − 9.71% 
 = −1.04
 19.48% 

– Calculate the coresponding value from the cumulative standard normal; in EXCEL: Normdist(-
1.04; 0; 1; 1)=14.9%

79
Prof. Dr. Bernd Scherer

The Equity Risk Premium Puzzle

• The key to investment risk is the positive correlation with consumption; assets
that pay off well if consumption is down (your marginal utility from extra wealth is
high) should require a lower risk premium (recession hedges).

r −c
= γ ⋅ σ ( ∆c ) ⋅ corr ( ∆c, r )
σ

• This is a genral asset pricing model as we do not need the covariance of asset
returns with a market portfolio

• For realistic values of risk aversion, consumption volatility and correlation the
derived SHARPE-ratio is inconsistent (factor 100) with the observed SHARPE-ratio
– Main culprit: consumptioni too smooth: too low volatility and correlation

80
Prof. Dr. Bernd Scherer

Measuring Stock Market Performance - Stock Market IndIces

• We need a measure of stock market performance

• Market value weighted indices are an excellent choice


– low cost, zero to low turnover, transparent, standardised
– Represent capital marketequilibrium, (by definition)
– Poor mans CAPM, difficult to beat
– Only portfolio that we can invest into without distorting relative prices, only portfolio all investors
can invest

• Some problems: not all market value indices are ideal


– Some are too narrow and underdiversified
– Inefficient
– Inclusion rules are momentum driven Momentum
– Exclusion resembles stop loss

81
Prof. Dr. Bernd Scherer

Equities, Bonds, Bills in the US

• High realized risk premium with considerable risk


– Some sharpe drawdowns are not even visible (1987) due to the annual view

Risk Premium
Stocks - T.Bills Stocks - T.Bonds
60.00%
1928-2009 7.53% 6.03%
1960-2009 5.48% 3.78% 40.00%

2000-2009 -1.59% -5.47%


20.00%

Standard Error

Annual Return
Stocks - T.Bills Stocks - T.Bonds 0.00%

1928

1933

1938

1943

1948
1953

1958
1963

1968

1973

1978

1983

1988

1993
1998

2003
2008
1928-2009 2.28% 2.40%
1960-2009 2.42% 2.71% -20.00%

2000-2009 6.73% 9.22%


-40.00%

t-value
Stocks - T.Bills Stocks - T.Bonds -60.00%

Years
1928-2009 3.31 2.51
1960-2009 2.26 1.39
2000-2009 -0.24 -0.59

82
Prof. Dr. Bernd Scherer

Styles

• Sorting equities on stock specific charactersitics turned out to earn superior risk
adjusted returns not explained by the market portfolio in the CAPM.
– The most popular characteristics are value and size
– FAMA/FRENCH use these portfolios to create a portfolio based asset pricing model

• Size: Small stocks outperform lage stocks


• Value: Bad firms (low growth, i.e. Low price to book) outperform good firms (high
growth companies)

• Although the empirical evidence is far from clear it is more likely that these return
advantages are a compensation for risk.

83
Prof. Dr. Bernd Scherer

Returns on Value and Size

• Value stocks materially 40

underperform growth stocks


in times of economic crisis 35

– Is value loading up on defqult


30
risk

FAMA/FRENCH/CARHART Factor Returns


– Are value stocks just the sickest 25
puppies in the stock universe? VALUE

20

15
VALUE SIZE
(HML) (SMB)
Mean 0.403 0.237
10
Volatility 3.590 3.336
SIZE
t-value 3.558 2.249
Sharpe-ratio 0.389 0.246 5
Skewness 1.838 2.187
Kurtosis 15.557 22.233
Minimum -13.450 -16.850 0
JB-Test (p-value) 0.000 0.000 1930 1940 1950 1960 1970 1980 1990 2000
TIME
Table 1: Summary for monthly returns on investment styles, 1927 to 2009

84
Prof. Dr. Bernd Scherer

Introduction Into Expected Utility And Long Term Asset Allocation

• Consider the following game. One coin is beeing tossed. If it turns out head the
player doubles his initial holding and can decide to continue or stop playing. If it
turns out number the player looses everything. What is the expected value of the
game?

0.5 ⋅ 2 + 0.25 ⋅ 4 + 0.125 ⋅ 16 + ⋯ = ∞

• The expected value of this game is infinite. However most player would offer very
little to participate in this lottery. This is called St. Petersburg Paradox.

• How can we solve this?

0.5 ⋅ ln ( 2 ) + 0.25 ⋅ ln ( 4 ) + 0.125 ⋅ ln ( 8 ) + ... = z


value = exp ( z )

85
Prof. Dr. Bernd Scherer

Log-Utility

• Assume investors prefer more to less at a decreasing rate.

• One example is log-utility U = ln (W )

• This utility function exhibits constant relative risk aversion (The optimal percentage
allocation into the risky asset remains the same whatever the level of wealth)

10

8
U til ity

4
0 1000 2000 3000 4000 5000 6000 7000 8000 9000 10000
Wealth

86
Prof. Dr. Bernd Scherer

Example

• An investor with log-utility can participate into a lottery. He will win 100 with
50% likelihood and he will win 1000 also with 50% likelihood. What is his
expected utility?
U = 0.5 ln ( 100 ) + 0.5 ln ( 1000 ) = 5.76

• How much would you have to pay the investor to not engage into this bet?
In other words: what safe profit does the investor need to give up his right
to participate in the lottery?

exp ( 5.76 ) = 316.23

• We need to pay 316.23 for sure. This is called the security equivalent.

87
Prof. Dr. Bernd Scherer

Samuelson: Irrelevance of Investment Horizon

• Unter the following assumptions SAMUELSON has shown that the optimal allocation into equities
does not depend on the investment horizon
– Investoren exhibit CRRA
– Equiyt returns show no patterns, i.e. No mean reversion or aversion
– Investors do not have human capital

• 4 Period Example (Market goes up 33% or falls 25% with equal probability)
237.04
177.78
133.33 133.33
100.00 100.00
75.00 75.00 0.25 ⋅ ln ( 177.78 ) + 0.5 ⋅ ln ( 100 ) + 0.25 ⋅ ln ( 56.25 ) = 4.605
56.25
42.19

Periode 0 1 2 3
Nutzen 4.605 4.605 4.605 4.605

88
Prof. Dr. Bernd Scherer

Assets And Inflation

Risk Factor Coefficients h = 12 h = 24 h = 36 h = 48 h = 60


• Regression of five year αmarket ,h 7.51 14.81 22.80 31.44 39.45
1.37 1.52 1.79 2.09 2.30
rolling returns for major Market
βmarket ,h -0.82 -0.79 -0.77 -0.78 -0.72
-0.77 -0.85 -1.02 -1.19 -1.27
asset classes. R2 0.02 0.03 0.05 0.07 0.07

αHML,h 4.74 11.32 18.00 23.71 27.73


1.46 1.72 2.03 2.19 2.30
HML
• Cash is best hedge. βHML,h 0.08
0.12
-0.09
-0.14
-0.11
-0.18
-0.06
-0.12
0.02
0.04
R2 0.00 0.00 0.00 0.00 0.00

-1.99 -8.26 -16.68 -25.93 -36.54


• Commodities only provide SMB
αSMB,h
-0.74 -1.35 -1.68 -1.99 -2.44
0.57 1.02 1.32 1.56 1.78
significant short term βSMB,h
1.10 1.59 1.93 2.41 3.22
R2 0.02 0.08 0.16 0.25 0.37
hedge.
αCOM ,h -8.07 -4.40 -0.26 4.63 12.51
-1.17 -0.46 -0.02 0.34 0.82
Commodity 2.70 1.47 1.01 0.69 0.35
βCOM ,h
• Equities show significant R2
1.99
0.10
1.40
0.06
1.12
0.04
0.89
0.03
0.51
0.01
negaative correlation
αCASH ,h 2.92 5.75 8.69 11.54 14.19
5.46 4.12 3.63 3.29 3.09
CASH βCASH ,h 0.61 0.63 0.63 0.64 0.66
6.07 4.79 4.18 3.81 3.67
R2 0.39 0.40 0.41 0.42 0.44

Table 1: Inflation beta for FAMA/FRENCH factors, commodities and cash for 1926:7 to
2009:12. For each return series I run a BOODOKH/RICHARDSON (1993) OLS regression
with overlapping data

89
Prof. Dr. Bernd Scherer

Inflation and Asset Returns – Empirical Evidence

• Stylized evidence
for alternative
time horizons

90
Prof. Dr. Bernd Scherer

What Causes The Negative Correlation Between


Inflation and Equities?

• In the US, there is substantial empirical evidence that high inflation is associated with
a high equity risk premium and declining stock prices.

• The Proxy Hypothesis: This view, more fully put forward by Fama (1981), argues that
the relationship between high rates of inflation and future real economic growth
rates is negative.

91
Prof. Dr. Bernd Scherer

Session 3: Equity Products

1. Valuation of Single Stocks

2. Access to Equity

3. Stock Portfolios

4. Equilibrium Returns

92
Prof. Dr. Bernd Scherer

1. Equity Valuation

• Dividend Discount Model


• Earnings versus Dividends
• Impact of Dividend Policy
• Real Life Methods: Shillers P/E

93
Prof. Dr. Bernd Scherer

Dividen Discount Model

• A DDM is a model that determines the value of a stock by discounting the


expected future cash dividends at the risk adjusted discount rate k (more
on this rate later in this chapter)
D1 + P1
Po =
1+k

• But how big is P1? Repeated substitution yields us


D2 + P2
D1 +
Po = 1 + k = D1 + D2 + P2 = ...
1 + k (1 + k )
2
1+k
D1 D2 D3
= + 2 + 3 + ...
1 + k (1 + k ) (1 + k )

Dt
=∑
t =1 (1 + k )
t

• In general: discount cash flows to equity holders using costs of equity or


cash flows to the firm using WACC

94
Prof. Dr. Bernd Scherer

Dividen Discount Model

• For a constant dividend growth g we can simplify this to

D1
Po =
k −g

• In the case of constant dividend growth, what is the expected percentage


change in stock prices (after dividends have been paid out)?

P1
−1 = ?
Po
D2

D (1 + g )
P1 = 1
k −g 1 (1 + g )
=P
D1
k −g

P1 P (1 + g )
−1 = o −1 = g
Po Po

95
Prof. Dr. Bernd Scherer

Earnings

• Let us redefine
∞ ∞ ∞
Dt Et It
Po = ∑ t = ∑ (1 + k ) −∑
(1 + k ) (1 + k )
t t
t =1 t =1 t =1

• The value of the firm is the present value of earnings minus the present
value of the fraction of earnings that needs to get reinvested to keep the
business going.
– Do not double count earnings!
– In an expanding industry net investments are positive
• Let us break up the firm into two parts: (a) the present value of current
earnings as a perpetuity (no growth) and the present value of growth
opportunities.
E1
Po =
+ NPV of future investments
k
• How do value and growth stocks differ?

96
Prof. Dr. Bernd Scherer

Do Dividends Matter?

1. The naive view: yes as it affects the cash flows to be discounted in the
DDM. Earlier cash flows are more valuable.

2. The Modgliani/Miller (no frictions) view: no as investors are indifferent to


receiving dividends and reinvestment and keeping the money in the firm.
This assumes a world without frictions (in particular no differential tax
treatment)

3. Real world: yes due to capital market frictions.


– Differential taxes
– Information effects from issuing new equity
– Information content of dividends (increase in dividends is a good sign)
– Investment banking fees

97
Prof. Dr. Bernd Scherer

Shiller: Cyclical price/earnings ratio (CAPE)

• CAPE compares share prices with average earnings during the past decade,
rather than to the most recent year's earnings.
– This evens out bumps in earnings multiples caused by the profit cycle, and has proved to be a
great market timing vehicle - highs and lows for this metric have overlapped almost perfectly
with highs and lows for the market.

50 20
2000
45 18
1981
40 16
1929
35 14

Long-Term Interest Rates


30 1901 12
Cyclically Adjusted PE

1966
25 Price-Earnings Ratio 10
22.04
20 8
1921 You are here
15 6

10 4
Long-Term Interest Rates
5 2

0 0
1860 1880 1900 1920 1940 1960 1980 2000 2020

Year
http://www.econ.yale.edu/~shiller/data.htm

98
Prof. Dr. Bernd Scherer

2. Access to Equity

• What Kind of Equities can Firms offer?


• Shareholder Value and Corporate Governance
• IPOs and Underpricing

99
Prof. Dr. Bernd Scherer

Types of Equity

• Common Stock – share of ownership that usually includes voting rights


(Exemptions are dual class common stock)

• Preferred Stock – financial instrument that fives holders a claim on the


firms earnings that must be paid before dividends on common stocks can
be paid. Dividends are fixed in advance.

• Warrants – long term call options on the firms stock

100
Prof. Dr. Bernd Scherer

Shareholder Value

• Maximize the market value of the firm

• Social welfare is maximised when all firms in an economy attempt to


maximize their own total value

• Firm management is disciplined by the market of corporate control (firms


that don’t maximize shareholder value are taken over by firms that will do)

• When is shareholder value a bad idea?


– If monopolies exist on input or output markets
– If there are negative externalities

• Instead of abolishing shareholder value, we should address the


externalities: regulation and competition policies, internalize external costs

101
Prof. Dr. Bernd Scherer

Private Equity

• Long term equity capital for not publicly listed firms


• Objective: Exit (going public) with a profit
• Long term horizon
• Sales pitches claim PE to provide an extension of the traditional
equity universe (Large cap, small cap, etc.) ? Questionable as same
risk factor runs through all equity products
• Motives
– Diversification (questionable)
– Premium for illiquidity (probably for venture but not for buyouts, i.e. you do not
earn an extra return from buying something liquid and transform it into
something illiquid. The opposite is true)
– Reduction of agency costs (management and owner are the same)
– Reduction of negative effect of information asymmetries with capital markets

102
Prof. Dr. Bernd Scherer

Going Public

• Efficient stock markets allocate capital to its best use and therefore create
growth

• Liquid equity markets allow investors to diversify (reduction of capital


costs) and venture capitalists to exist. This is a competitive advantage for
young firms.

• Benefits
– Lower capital costs due to better access to markets, increased liquidity and the fact that
owners can now diversify better
– employees can sell their stocks
– External monitoring

• Disadvantages
– Expensive: 10% in lawyer, investment banking and underwriting fees + 15% underpricing (see
next section) = 25%
– Information revealed to competition

103
Prof. Dr. Bernd Scherer

Initial Public Offering (IPO) Underpricing

• Most companies that go public do so via an initial public offering of shares to


investors.

• Logue (1973) and Ibbotson (1975), documented that when companies go


public, the shares they sell tend to be underpriced, in that the share price
jumps substantially on the first day of trading.
– Since the 1960s this ‘underpricing discount’ has averaged around 19% in the United States,
suggesting that firms leave considerable amounts of money on the table. Underpricing has
tended to fluctuate a great deal, averaging 21% in the 1960s, 12% in the 1970s, 16% in the
1980s, 21% in the 1990s, and 40% in the four years since 2000 (reflecting mostly the tail-end of
the late 1990s internet boom).

• Clearly, underpricing is costly to a firm’s owners: shares sold for personal


account are sold at too low a price, while the value of shares retained after
the IPO is diluted. In dollar terms, IPO firms appear to leave many billions ‘on
the table’ every year in the U.S. IPO market alone.

104
Prof. Dr. Bernd Scherer

Acesssing Equities – Initial Public Offering (IPO)

• Empirical Evidence for the US

Hot issue periods, driven by available capital

105
Prof. Dr. Bernd Scherer

IPO Underpricing – Information Asymmetries

• Perhaps the best-known asymmetric information model is Rock’s (1986) winner’s


curse, which is an application of Akerlof’s (1970) lemons problem.
• Rock assumes that there are two type of investors. Informed investors bid only
for attractively priced IPOs, whereas the uninformed bid indiscriminately. This
imposes a ‘winner’s curse’ on uninformed investors: in unattractive offerings,
they receive all the shares they have bid for, while in attractive offerings, their
demand is partly crowded out by the informed.
• Thus, the return uninformed investors earn conditional on receiving an allocation
is below the simple average underpricing return.
– In the extreme case, the uninformed are rationed completely in underpriced IPOs and receive 100
percent allocations in overpriced IPOs, resulting in average returns that are negative.
• When conditional expected returns are negative, uninformed investors will be
unwilling to bid for IPO allocations, so the IPO market will be populated only with
(equally) informed investors.
– Rock assumes that the primary market is dependent on the continued participation of uninformed
investors, in the sense that informed demand is insufficient to take up all shares on offer even in
attractive offerings

106
Prof. Dr. Bernd Scherer

IPO Underpricing – Information Asymmetries

• This requires that conditional expected returns are non-negative so that the
uninformed at least break even.

• In other words, all IPOs must be underpriced in expectation. This does not
remove the allocation bias against the uninformed – they will still be crowded
out by informed investors in the most underpriced offerings – but they will no
longer (expect to) make losses on average, even adjusted for rationing.

• Beatty and Ritter (1986) argue that as repeat players, investment banks have an
incentive to ensure that new issues are underpriced by enough lest they lose
underwriting commissions in the future. Investment banks thus coerce issuers
into underpricing. Of course, they cannot underprice too much for fear of losing
underwriting market share.

107
Prof. Dr. Bernd Scherer

3. Stock Portfolios

• How Do we Measure Risk and Return?


• Basic Portfolio Theory
• Basic Risk Decomposition

108
Prof. Dr. Bernd Scherer

Expected Returns and Risks

• Suppose you are given two assets with the following scenarios and
respective scenario probabilities

Event Probability Return A Return B


1 0.167 10 5
2 0.167 20 -5
3 0.167 -15 -4
4 0.167 30 4
5 0.167 5 7
6 0.167 -15 -3

• What are expected returns, risks (standard deviation of returns) and


covariances?

109
Prof. Dr. Bernd Scherer

Some Simple Calculations

• Expected return for asset A


1 1
E (RA ) = ∑ i =1 pi RA,i = ∑
6 6
R = (10 + 20 + ... − 15) = 5.83
i =1 A,i
6 6

• Variance for asset A


1
pA,i (RA,i − E (RA,i )) =
2
∑ (10 − 5.83) + ... + (−15 − 5.83)  = 18.28
6
Var (RA ) =
2 2
i =1
6 

• Covariance between asset A and Asset B

Var (RA ) = ∑ i =1 pA,i (RA,i − E (RA,i ))


6 2

1
= [(10 − 5.83)(5 − 0.67) + ... + (−15 − 5.83)(−3 − 0.67)] = 80.5
6

110
Prof. Dr. Bernd Scherer

Why Use Standard Deviation to Measure Risks?

• This is the right approach if investors only care about standard deviation to
measure risks.
— Returns are multivariate normal
— Preference (utility functions) only care about variance

• Both assumptions are violated in practice. Still practitioners (and


academics) continue to look at standard deviations. Why?
– Deviations from normality are unstable and difficult to estimate
– Lack of credible alternative (many will tell you about better risk measures, but they will not
tell you the side effects)
– Central limit theorem will help us in many applications (large n, diversified and factor neutral
portfolios)

111
Prof. Dr. Bernd Scherer

Portfolio Risk and Return

• Some basic statistics


RP = wARA + wB RB
E (RP ) = E (wARA + wB RB ) = wAE (RA ) + wB E (RB )
Var (RP ) = Var (wARA + wB RB ) = wA2Var (RA ) + wB2Var (RB ) + 2wAwBCov (RA, RB )

Event Probability Return A Return B Portfolio (50:50)


1 0.167 10.0 5.0 7.5
2 0.167 20.0 -5.0 7.5
3 0.167 -15.0 -4.0 -9.5
4 0.167 30.0 4.0 17.0
5 0.167 5.0 7.0 6.0
6 0.167 -15.0 -3.0 -9.0

Return 5.83 0.67 3.25


Standard deviation 18.28 5.24 10.44

112
Prof. Dr. Bernd Scherer

Diversification I

• Less than perfect correlation is desirable

20

18

16
Portfolio Standard Deviation

14

12

10

4
What is the return on this portfolio?
2

0
100%
95.00%
90.00%
85.00%
80.00%
75.00%
70.00%
65.00%
60.00%
55.00%
50.00%
45.00%
40.00%
35.00%
30.00%
25.00%
20.00%
15.00%
10.00%
5.00%
0.00%
Weight in Aset B

correl = 0 correl = 1 correl = -1

113
Prof. Dr. Bernd Scherer

Mean Variance Diagram For Asset A and B

• Provides feasible risk return combinations


• We would go upwards left (but can not) while we could go downward right
(but should not)

5
Portfolio Return

0
0 5 10 15 20 25
Portfolio Risk (Standard Deviation)

114
Prof. Dr. Bernd Scherer

Risk Contributions

• What is the marginal contribution to risk for an asset?

∑ i =1 wi2σi2 + ∑ i ≠`j wi w j σij


n
σ2 =
dσ2
= wi σi2 + ∑ i ≠`j w j σi j
dwi

• Covariance risk is most important for an assets contribution to portfolio


risk!
• How does this relate to the beta of a security?

Cov ( ri , rp ) Cov ( ri , ∑ i =1 wi ri )
n
wi σi2 + ∑ i ≠`j w j σi j
βi = = =
σp2 σp2 σp2

115
Prof. Dr. Bernd Scherer

The Limits of Diversification

σp2 = n ( n1 ) σ 2 + n ( n − 1)( n1 ) ρσ 2 = + (1 − n1 ) ρσ 2
• Assume equal return 2 2 σ2
correlation volatility n

• What is the minimum


σp2 = 0 + ρσ 2
volatility for large number of
assets?

• Restriction on correlations:
σ2
+ (1 − n1 ) ρσ 2 > 0, σn > − (1 − n1 ) ρσ 2
2
weighted random variables σp2 = n
can not have negative
variance ρ > − n 1−1

µ S
• What is the maximum Sharpe Sp = 1
= 1
Ratio? ( σ2
n + (1 − n1 ) ρσ 2 ) 2
( n1 + (1 − n1 ) ρ )2
1
lim S p = 1 S
n →∞ ρ 2

116
Prof. Dr. Bernd Scherer

4. Equilibrium Returns

• Theory behind CAPM


• Determinants of the Risk Premium on market Portfolio
• Using the CAPM in Portfolio Selection
• Using the CAPM in Valuation
• Empirical Validity

117
Prof. Dr. Bernd Scherer

The CAPM in Brief

• Fundamental idea: in equilibrium the capital markets reward investors only


for taking on systematic, i.e. un-diversifiable risk.
– In equilibrium assets can not be diversifying and offer high returns at the same time (beware
those pitch-books)
– Investors hold inefficient portfolios at their peril
– The risk premium for an asset is unrelated to the assets stand alone risk

• For the CAPM we need two assumptions


– All investors agree in their forecasts of expected returns, risks and correlations and tehrefore
optimally hold the same relative proportion of individual stocks (note that this can be relaxed if
we are willing to assume the capital market is complete)
– No capital market frictions

• Result
– All investors hold the market portfolio (in varying proportions with cash due to their risk
aversion)
– Practically investors should hold a money market fund and an index fund (ETF)

118
Prof. Dr. Bernd Scherer

Capital Market Line (CML)

• THE CML represents the best risk reward combination available to all
investors
E (rm ) − c
E (r ) = c + ⋅ σr
E r , E (rm )
( ) σ m

Market portfolio

E (rm ) − c

c
σm
σr , σm

• The slope of the CML is the famous SHARPE ratio named after Nobel price
winner Bill SHARPE. Interestingly it is most of the time misapplied to
individual investments!

119
Prof. Dr. Bernd Scherer

The Risk Premium on The Market Portfolio

• The risk premium on the market portfolio is the great unknown in CAPM
calculations

• We can think of the risk premium on the market portfolio be driven by the
index of risk aversion (within the economy) denoted by A as well as the
riskiness of the market portfolio (its variance)

E (rm ) − c = Aσ 2

• Empirically we can try to capture variation in A by variations in Wealth


(stock market capitalization). If wealth is falling risk aversion is likely to rise.

120
Prof. Dr. Bernd Scherer

Beta and Risk Premium on Individual Securities

• The risk of an efficient portfolio is measured by standard deviation

• In a CAPM world the risk of an individual security is measured by its beta,


i.e. the standardized the covariance between a securities returns and
market returns
σim σi σm ρi σi
βi = = = ρi
σm2 σm2 σm

• This makes sense as the diversifiable risk diversifies away and the risk
contribution to the market portfolio equals its beta (the sum of betas add
up to 1)

• The expected return from an individual security is measured by the SML


(security market line)
E (ri ) − c = βi E (rm ) − c 

121
Prof. Dr. Bernd Scherer

CAPM Implementation – Beta Calculation

1. Create (download from a data bank) time series of project (security)


returns.

2. Download a time series of risk free rates (one month bills)

3. Download a series of returns for a portfolio that proxies the market


portfolio (usually a broad based cap weighted index)

4. Regress the time series of security excess returns against the time series of
market excess return (include an intercept or your result will be biased)
– At this stage you can add Bayesian procedures to deal with parameter instability or mean
reversion in betas if you wish

5. Take the beta and calculate the respective benchmark or hurdle


rate

122
Prof. Dr. Bernd Scherer

Using the CAPM in Portfolio Selection

• Diversification Advice
– Diversify your portfolio to get rid of unsystematic risks
– Best to hold a combination of cash and market portfolio

• Investment advice
– The CAPM provides a breakeven return
– Ensure that you have good reason to believe the market is out of equilibrium and
inefficient enough not to have returned to the fair price.

• Measure the performance of your portfolio manager


– Regress the risk premium of your fund against the risk premium of the market
portfolio to see whether the fund earns in excess (alpha) of the rturn rquired to
compensate for systematic risks.

123
Prof. Dr. Bernd Scherer

Other CAPM Uses

• Capital Costs
– Calculate the beta of a project to calculate the required hurdle rate (costs of
equity, i.e. return requirement for the provider of equity capital)
• Regulation
– Calculate the fair return for a regulated entity
• Court
– Calculate the stocks return in the absence of insider trading, information release,
etc.

124
Prof. Dr. Bernd Scherer

Empirical Evidence

• Small Caps earn more than they should under the CAPM; this lead to new
model that added a small CAP and a value premium (FAMA/FRENCH model)

Size Decile 10
Small Cap Stocks

125
Prof. Dr. Bernd Scherer

5. Execution Costs

• Components
• Approaches to Reduce Execution Costs

126
Prof. Dr. Bernd Scherer

Components of Execution Costs

• Commissions, Fees, Taxes


– Commissions are generally negotiable; some are paid for particular services: research

• Market impact
– Bid, ask spread widening as larger orders are placed
– VWAP (volume weighted average price as benchmark for trading desk
– Issues with VWAP: close to useless if the trade in question represents a large trading volume,
does not take opportunity costs into account

• Opportunity Costs
– Loss or gain in performance as a result of failure to trade a decision
– Very important for high frequency trading with rapid information decay

127
Prof. Dr. Bernd Scherer

Approaches to Reduce Execution Costs

• Internal Crossing
– Not always possible, difficulty to set fair transaction price

• External Crossing
– Might take too long (opportunity costs)

• Principal trade
– Trade though a dealer that guarantees price. Dealer acts as a principal, i.e. acts
on firm price.
– Fast, but with large hidden execution costs

• Agency trade
– Trader works the order against a trading benchmark

128
Prof. Dr. Bernd Scherer

Session 4: Futures, Options, Principal Protection

1. Futures

2. Options

3. Principal Protection

129
Prof. Dr. Bernd Scherer

Literature

• BODIE/MERTON/CLEETON (2009), Financial Economics, Pearson

130
Prof. Dr. Bernd Scherer

1. Futures

• How to use futures?


• Arbitrage relations
• What information can and can not be inferred from futures?

131
Prof. Dr. Bernd Scherer

Forwards versus Futures

• Forward contract
– Parties agree to exchange an item at a delivery price agreed now.
– The forward price is defined as the delivery price that equates the market value of the
contract with zero.
– The face value of the contract is the quantity specified times the forward price.
– The buyer (seller) of a forward is called long (short)
– Forward contracts are bespoke agreements betweeen two parties (traded OTC)

• Futures contract
– Exchange traded contracts
– Exchange specifies exact underlying (quality, etc), contract size, delivery point
– Easy to close out, i.e. To terminate the position befor delivery date.

132
Prof. Dr. Bernd Scherer

Margin Account

• EXAMPLE DJIA 3 month Futures, Futures price 10220 (allows you to buy 10 „shares“ of
the DJIA portfolio)
– Initial margin: USD 4000, i.e. 4%
– Maintenance margin: USD 3000, i.e. 3%

• Daily realization of profits and losses minimizes the probability (and severity) of
counterparty default.
• Futures markets are used by individuals whose credit rating is costly to be chsecked
• Forward marketis OTC between counterparties with high and easy to verify credit
ratings.

133
Prof. Dr. Bernd Scherer

Financial Futures

• Main consideration: buying a stock (index) today or buying it today at the


futures price for delivery in a month must (in the absence of dividends) be
the same.

• Example spot price at 100 with r% interest rate.


– Action 1: Buy indext today at 100. Cash Flow in a year is S1.
– Action 2: Buy future todat at F. Put F/(1+r) in the bank to yield F at delivery. You
receive a stock worth S1 and pay F from your accumulated cash account. Yor cash flow
in year 1 is again S1.

• This yields the following pricing relationships


Ft ,t +1 = S (1 + r1 )
Ft ,t +1 = (S − PV (Dividend ))(1 + r1 )

Futures price is lower
as there is an advantage to hold
the stock

134
Prof. Dr. Bernd Scherer

Future Pricing Example

• Assume the gold price to be 300 USD, storage costs are 2% and interest
rates are 8%.
• Return on stored gold is
S1 − S 0
rgold = −s
S0

• Return on synthetic gold (buy futures and invest PV of futures price into
cash)
S1 − F
rgold = +c
S

• By arbitrage both returns must be the same

F = (1 + c − s ) S = (1 + 8% − 5%) 300 = 330

135
Prof. Dr. Bernd Scherer

Economic Function of (Commodity) Futures Markets

• Reallocation of exposure to commodity risk: Producer wants to off load


price risk (to speculator, so speculation is socially usefall after all!)

• Let S be the spot price (of wheat for example) and Cj the storage costs of
the j-th producer. As usual F is the Future price.

Cj < F −S
 ⇒ carry the wheat (store and sell later )
¨spread

• The future and forward markets create an economic structure where the
most cost efficient distributor would do the physical storage.
• Example: Suppose you are a distributor of corn (you store it!). The spot pice
is 3 USD/bushel and the futures price for 1 month delivery is 3.10
USD/bushel. Your costs of carrying (storing) corn is 0.15. WHat should you
do.

136
Prof. Dr. Bernd Scherer

The Role of Speculators

• Speculators motive is to make profits from futures trades. Speculators


spend time and money to obtain and process information that allows them
to forecast prices: Is this socially useful?

• Speculators
– Provide liquidity. For every hedger there must be a speculator to take the opposite
position
– Price discovery. Speculators ensure that relevant information enters the price
 Example: US housing market: If speculators would have been able to sell short
thet bubble might not have developed and popped earlier with less severe
consequences

137
Prof. Dr. Bernd Scherer

Relation Between Commodity Spot and Futures Prices

• General arbitrage condition (futures price can not exceed spot price by
more than cost of carry, cost of carry can vary over time)
F −S ≤C

• Commodity is not in storage (stock out). The above holds as strict


inequality. In this case the forward price reveals some information about
the spot price, i.e. There is information in the future price that is not already
in the spot price.

• Commodity is in storage: Spot and future price are perfectly linked by


arbitrage. The above condition holds with equality. Future and spot price
contain the same information.
– We can estimate the cost of carry. If the cost of carry is negative (backwardation)
there are benefits from holding the commodity physically (convenince=ence yield)

138
Prof. Dr. Bernd Scherer

Backwardation

• Backwardation means: spot prices are higher than futures prices, and/or
prices for near maturities are higher than for distant.
– Backwardation occurs if convenience yield exceeds storage cost (inventories are low)

139
Prof. Dr. Bernd Scherer

Minimum Variance Hedging (Hedging with Basis Risk)

• The hedge ratio is the number of futures used to hedge a unit exposure to the
risk of spot price. With basis risk, how do we choose optimal hedge ratio?
– Basis risk arises from mismatch of underlying asset (an investment bank shorts index futures
to hedge the risk in underwriting a large stock issue) or mismatch of maturity (roll over short
contracts to hedge long term risks) Basis risk leads to imperfect hedge. Use linear regression
(minimizes basis risk)!
∆S = a + b∆F + e

140
Prof. Dr. Bernd Scherer

2. Options

• Binomial Trees
• BLACK/SCHOLES Formula
• Greeks (Sensitivities)
• Delta Hedging
• Where do we find option payoffs?

141
Prof. Dr. Bernd Scherer

Options

• An option is a contract that gives a contracting party the option to


buy (sell) something at a prespecified exercise price (strike)
– Options are so called contingent claims, i.e. their payoff dependon the outcome
of an uncertain event.
– The simplest building blocks are European (can only be exercised at maturity)
calls and puts. We will focus on these.

• Options come in many forms not only as financial options


• Equity is a long call on the companies assets with strike equal to the principal
value of outstanding bonds
• Company is short a put option on the pension fund surplus

142
Prof. Dr. Bernd Scherer

Quotes and Moneyness

• Example
Call = max [S − X , 0]
Put = max [X − S , 0]

Expiry of option contract Intrinsic value: 30 - 27.5 = 2.50 = max[S-X,0]


Time value: 3.26 - 2.5=0.76
What factors will affect the time value?

143
Prof. Dr. Bernd Scherer

Payoffs

• Insure a portfolio with a protective put (buy put option to cut off the
downside

• It looks like stock plus put (left picture) give you the same payoff than cash
plus call with same strike (right picture).
– Arbitrage: If two assets A and B, have same future payoffs with certainty, then they should sell
for the same price now.
– Option pricing theory is about relative pricing (makes no statement whether the stock is priced
correctly)

144
Prof. Dr. Bernd Scherer

Put-Call Parity

• IBM stock currently sells for $34 per share. There are call and put
options available with a strike price of $36 and an expiration date
of one year. The price of the IBM call option is $5, and the risk-free
interest rate equals 4%. What is the price of the put option?

36
P + 34 = 5 +
1.04

Bond ¨
PV of strike

P = 5 + 34.62 − 34 = 5.62

• This parity is enforced by arbitrage and holds empirically very well.

145
Prof. Dr. Bernd Scherer

One Period Binomial Model

• Assume a two period example. The share price of 100 today can move to 115 or fall to
95. What is the value of a call on the stock price with strike 100, i.e. the value of
max[S-100,0] today?

ր max 115 − 100, 0 = 15


[ ]
ր 115
100 c
ց ց
95 max [ 95 − 100, 0 ] = 0

• Suppose we buy 0.75 shares and sell 1 option. Whatever happens to the stock price
we would arrive at

0.75 ⋅ 115 − 15 = 71.25


0.75 ⋅ 95 − 0 = 71.25

• Assuming 5% interest rate we can solve for the option price


¨risk −free payoff

71.25
= 1 + 5% ⇒ c = 7.143
0.75 ⋅ 100 −
 c
investment

146
Prof. Dr. Bernd Scherer

A more general model

• We create a riskless position (delta hedge the option) by


d ⋅ S ⋅ up − cup = d ⋅ S ⋅ down − cdown
cup − cdown
d=
S (up − down )

• Solving for the option value we get

d ⋅ S ⋅ up − cup m ⋅ cup + (1 − m )cdown


=1+r ⇒c =
d ⋅S −c 1+r
m= 1+r −down
up−down

• The option value is simply the discounted expected value of final payoffs
under risk neutral probabilities.
– The risk neutral probabilities are pseudo probabilities that fix the likelihood of up and down
movements such that the expected rate of the return on the stock equals the risk free rate.

147
Prof. Dr. Bernd Scherer

Two - period example

• Two period tree

ր 132.25 ր 32.25
115 cup
ր ց ր ց
100 109.25 ⇒ c 9.25
ց ր ց ր
95 cdown
ց ց
90.25
0

• We can use m to run Monte Carlo simulation to value the claim,


i.e. its expected value under the risk neutral distribution.

148
Prof. Dr. Bernd Scherer

Calculations

• Working backwards (sequence of one period trees)

m ⋅ cup,up + (1 − m )cup,down
cup = = 19.76
1+r
cdown = 4.41
c = 11.508

• How do we calibrate a tree to exhibit the same dispersion than the


stock price?

up = e σ t , down = e −σ t

149
Prof. Dr. Bernd Scherer

Delta Hedging

• Dynamically replicate a long call with a delta equivalent portfolio of short


stocks.

• The replicating profits can be calculated from

∑ d ( St − St ) + ∑ n =1 rn −1 ( dt St − C 0 )
T T
( max [ ST − K , 0 ] − C 0 ) −
 n =1 t
n −1 n
 n −1 n −1
 n −1
Payoff from long call rebalanced interest earned on
minus cos t from long call delta equivalent proceeds from
short stock

Call-Delta

150
Prof. Dr. Bernd Scherer

Delta Hedging (continued)

Delta-Hedged Gains
Using a Call on IBM on January 3, 2006
Closing Price DTM Call Hedging Interest
Call Price ($) 2.95 Date DTM St Volatility Delta Profits Payments
Risk Free Rate (r ) 0.05 3-Jan-06 17 82.06 0.194 0.753
Date of Maturity 20-Jan-06 4-Jan-06 16 81.95 0.191 0.750 -0.08 0.01
Strike Price (K ) 80 5-Jan-06 15 82.50 0.189 0.809 0.41 0.01
6-Jan-06 14 84.95 0.191 0.953 1.98 0.01
9-Jan-06 11 83.73 0.257 0.860 -1.16 0.02
10-Jan-06 10 84.07 0.239 0.905 0.29 0.01
11-Jan-06 9 84.17 0.211 0.944 0.09 0.01
12-Jan-06 8 83.57 0.192 0.944 -0.57 0.02
13-Jan-06 7 83.17 0.187 0.940 -0.38 0.02
17-Jan-06 3 83.00 0.167 0.993 -0.16 0.01
18-Jan-06 2 83.80 0.153 1.000 0.79 0.02
19-Jan-06 1 83.09 0.164 1.000 -0.71 0.02
20-Jan-06 0 81.36 -1.73 0.02
Delta-Hedged Gain
Dollar Gain ($) -0.20 Terminal Call Cost Hedging Interest Profit π
As a % of Spot Price S 0 -0.24% Payoff ($) C 0 ($) Profits ($) Payments ($) ($)
As a % of Call Price C 0 -6.81% 1.36 2.95 -1.22 0.17 -0.20

151
Prof. Dr. Bernd Scherer

BLACK&SCHOLES Option Pricing

• Assumptions
– Returns are lognormal and distribution is known
– No transaction costs, continuous hedging feasible
– Known and constant interest rates
– No early exercise, no dividends

• Don’t worry if you don’t like the assumptions. Most have been relaxed and
effectively dealt with in derivatives research.

• However, continuous trading (hedging) is practically the most troubling and


not quite resolved issue

152
Prof. Dr. Bernd Scherer

BLACK&SCHOLES Equation

• The B/S equation for a European call (no dividends) is

c = SN (d1 ) − e −rt XN (d2 )


Normal distribution
Probability: Look it up in statistical
Tables or in EXCEL
S 
log   + rt + 21 σ 2t
X 
d1 = , d2 = d1 − σ t
σ t

• The parameters are define as:

S : stock price
X : strike
σ : volatility (annualized )
r : interest rate
t : maturity

153
Prof. Dr. Bernd Scherer

Some Interpretation

• In the case of no volatility (σ = 0) both normal probabilities will be 1 and we get

c = S − e −rt X
• For (σ = 0) the option price will be a weighted average of stock and zero bond

• Example: S = 100, X = 110, r=0.08 (continuously compounded!), t=0.5, σ = 0.3

 100 
Probability the stock ends in the money log   + 0.08 ⋅ 0.5 + 21 ⋅ 0.32 ⋅ 0.5
 110 
d1 = = −0.1547
0.3 ⋅ 0.5
d2 = d1 − 0.3 ⋅ 0.5
N (d1 ) = 0.4384
N (d2 ) = 0.3568
c = 100 ⋅ 0.4384 − 110 ⋅ e −0.08⋅0.5 ⋅ 0.3568 = 6.13
p = c − S + e −rt = 11.82

154
Prof. Dr. Bernd Scherer

Sensitivities – Delta

• The delta expresses the share equivalent of an option: How much does the option
price move if the (underlying) stock moves?
– Rule of thumb: The delta of an ATM call is about 0.5
– The delta of a deep in the money option is one (used as directional market play with little notional
exposure)
dc dp
∆c = = N (d1 ) = 0.4384, ∆p = = N (d1 ) − 1 = −0.5616
dS dS

• If deltas are known we can aggregate them and calculate the directional exposure
of an options portfolio.
• We can also calculate an options elasticity …
 dc  S 
εc =    = delta ⋅ leverage
dS  c 
• … or its beta

 dc  S 
βc =     βS = delta ⋅ leverage ⋅ stock − beta
dS  c 

155
Prof. Dr. Bernd Scherer

Sensitivities – Delta

• We can plot the delta of an option to see how the sensitivity changes when yhe
underlying price changes.

Moves like stock 1


0.9
0.8
0.7
0.6

D e lta
0.5
0.4
0.3
0.2
0.1
Busted 0
60 80 100 120 140
Stock Price

• How might this change if volatility rises?

156
Prof. Dr. Bernd Scherer

Gamma – How fast does my delta change?

• A delta hedged position is of little comfort if the delta changes fast


• Gamma is the second derivative of the call value function and measures the
speed at which the slope changes
 dc  1 −12d12
d   e
  d 2c
0.02
γc = dS = 2 = 2π
0.018 dS dS Sσ t
0.016 1 −12(−0.1547)2
e
0.014
= 2π
0.012 100 ⋅ 0.3 0.5
= 0.0186
Gamma

0.01

0.008

0.006

0.004

0.002

0
60 70 80 90 100 110 120 130 140
Stock Price

• What does a gamma of 0.0186 mean? It means that an initial delta of


0.4384 moves to 0.4570 if the share price moves from 100 to 101

157
Prof. Dr. Bernd Scherer

Option Risk: MC Simulation and Delta Gamma Method

• Options are non-linear positions. How can we measure their risks?


– Volatility as a risk measure is insufficient as a long call will not experience a three sigma event
(a long call positions los s is limited to the call premium)

• Alternative 1: Taylor Expansion


– Problem: works only for infinitesimal changes (not what risk management is usually interested
in)

dc d 2c
∆c ≈ ∆S + 12 2 ∆S 2 + O
dS dS

• Alternative 2: Direct evaluation using MC Simulation


– Simulate distribution of stock prices in 1 month time and re-evaluate the option. This gives a
distribution of one month option prices
– Problem: computationally expensive for complex options

158
Prof. Dr. Bernd Scherer

Implied Volatility

• Implied volatility contains a risk premium over expected future realized volatility due
to the impossibility to delta hedge correctly.
• Option prices rise as the world gets riskier

159
Prof. Dr. Bernd Scherer

The many kind of options

• Performance based compensation


• Equities as options
• Corporate bonds
– Government bond + short put on operating assets
• Projects as options
– make sequel or movie, expand mining operation, etc.
• …

160
Prof. Dr. Bernd Scherer

3. Portfolio Insurance

• Who Should Buy Portfolio Insurance?


• CPPI strategies

161
Prof. Dr. Bernd Scherer

The Options Business (Cynics Guide)

• How are options sold? How do you make most money as an investment banker?

• Sell risks that clients (wrongly) dont perceive as risks! Exploit clients behavioural biases!
• Capital guarantee in high return currency that is thought to appreciate
• Sell digital payouts on dual events that never ever happenend before
• Etc. ...

• Create intransparency to hide fees and mispricing

• Create non-standard payoffs to avoid benchmarking and performance measurement!

• Some rules of thumb:


– Buy portfolio insurance If you have a higher risk aversion than the market or If you have
higher return expectation than the market
– Dont buy path dependent products when you are a long term investor? Path dependency
creates excess dispersion in final wealth

162
Prof. Dr. Bernd Scherer

Option Based Portfolio Insurance – OBPI

• Choose time frame and protection level

• Buy and forget products (no dynamic trading, but beware the credit risk!)
– Good choice if you believe realized volatility will be higher than implied volartility

• Derived from Put-Call parity: Equity plus put (protective put) equals cash plus
call (fiduciary call), Typically perceived as too expensive. Variations exist
– Capital guarantee in foreign currency
– Averaging option rather than end of period price
– Sell protection, ..

• OBPI is really selling a structured product through an asset management


distribution channel

163
Prof. Dr. Bernd Scherer

CPPI
• CPPI offers principal
protected exposure
by dynamically
allocating the initial
investment between Constant Proportion Portfolio Insurance (“CPPI”) offers principal protected exposure by
a risky and riskless dynamically allocating the investments of a portfolio between two assets: a risky asset (such as
asset an equity index or fund) and a riskless asset (such as a government bond). As the value of the
• Hence, exposure risky asset increases, the allocation to the risky assets increases; as the value of the risky asset
towards the risky decreases, the allocation to the riskless asset increases. At any given time, the minimum portfolio
asset is variable value (the “floor”) must equal the present value of the minimum amount guaranteed at maturity.
• The more the To determine the amount allocated to the risky asset, the excess of the portfolio value over the
market increases,
the more client floor (the “cushion”) is multiplied by a predetermined figure (the “multiple”).
participates in
future market
increases
• The more the
market decrease, Graphical Representation For example, an investor with a €100
Example
the more client
reduces his portfolio value, a floor of €90 and a
Assumptions multiple of 5 will allocate €50 (5 * (€100 -
exposure to the
Initial Portfolio
market Value = 100
Cushion = 10
Risky €90)) to the risky asset and €50 to the
(excess of Portfolio Assets = 50
Floor = 90 Value over the Floor) riskless asset.
(Cushion x
• The dynamic nature Multiple = 5 Multiple)
Floor = 90
of the CPPI means (present value
The floor, initial cushion and multiple are
of the
that the final minimum defined according to the investor's risk
participation level guaranteed Riskless Assets =

cannot be pre-
payoff at
maturity)
50 tolerance and are exogenous to the
(Portfolio Value
specified less Risky Assets) model.

Calculate Allocations

164
Prof. Dr. Bernd Scherer

CPPI
In theory, the allocation between risky and riskless assets is calculated continuously as the
value of the risky asset changes. An increase in the cushion allows for more funds to be
allocated away from the riskless asset into the risky asset. A decrease in the cushion indicates
there is less safety margin and that more funds will be allocated into the riskless asset. If the
CPPI level approaches the price of the riskless asset (=floor), the cushion will tend towards zero
and the investment will be fully allocated into the riskless asset.

This ensures that the minimum guarantee is paid out to the client at maturity.

Risky Assets Portfolio Value = 95


Return: -10%

Risky Risky Cushion = Risky


Assets = Assets = 5 Assets =
50 45 Calculate New 25
Allocations
Risky Asset =
cushion x
Riskless Riskless multiple Riskless
Assets = Assets = Floor = 90 Assets =
Assume no Riskless Asset
50 change 50 70
=
Portfolio Value
less Risky Asset

Portfolio Value = 100 Determine Floor and Cushion of the Portfolio Value = 95
Portfolio Value

Assumptions: Multiple = 5, Floor = 90 (assume no change)

165
Prof. Dr. Bernd Scherer

CPPI
• Performance of
CPPI Portfolio vs.
Underlying Risky
Asset

130 Price of risky asset Minimum Guarantee at


• Proportion of (starting value of 100) Maturity
Portfolio in Risky 120

and Riskless 110 Cushion


Assets 100

90
• Whenever the Price of safe Asset (Floor)
80
market increases,
the more client 70 Portfolio Value Risky Asset Performance Minimum Guarantee at Maturity Floor
participates in future 60
market increases
125%
Risky Assets Riskless Assets
• The more the market
100%
decrease, the more
client reduces his
exposure to the 75%

market
50%

25%

0%

166
Prof. Dr. Bernd Scherer

• First scenario
• As the cushion
approaches zero, the
CPPI
allocation to the risky
asset approaches
zero. In continuous
Risky Assets
time, this dynamic Return: -20%
rebalancing keeps the Risky
Risky
portfolio value from Assets = Assets =
20
falling below the floor. 25
At this point all the Floor = 90 Risk-
Risk- Calculate Floor
Risk- New less
investment is fully less (thus, no Only =
less Allocations Assets =
allocated into the Assume no Assets = cushion left) 90
Assets = change
90
riskless asset. This 70
70
ensures that the
minimum guarantee is
paid out to the client Portfolio Value = 90 Portfolio Value = 90
Portfolio Value = 95
at maturity

Risky Assets
Return: -40%
Risky
• Second Scenario Assets =
Shortfall

• The portfolio value 25 Risky


will fall below the floor Assets=15 Floor = 90 Calculate New
Allocations Risk-
if there is a very sharp Risk- (thus, no less
Floor
drop in the market Risk-less less cushion left +5 Only =
Assets =
Assets = Assume no Assets = and there is 90
before the allocation 70 Issuer contributes 5
90
change 70 a shortfall
between risky and to make up for
shortfall
riskless assets can be
rebalanced. This risk
(which is equal to the Portfolio Value = 95 Portfolio Value = 85 but at Portfolio Value = 90
any given time, it should be
cushion) is known as at least 90
the “Gap Risk”.
• Note that the “Gap
Risk” is taken by Assumptions: Multiple = 5, Floor = 90 (assume no change)
Morgan Stanley when
managing the CPPI

167
Prof. Dr. Bernd Scherer

CPPI

• Third scenario Cushion = Cushion =


Risky Assets Risky
• There is usually a Return: 80%
25 25
Assets =
maximum value Risky 45 Risky Assets
Assets = = 115
that will be (cushion x
25
allocated to the Calculate
multiple =
Risk- Floor = 125, but the
risky asset that is less 90 New
allocation is Floor =
Risk-less Allocations 90
defined as a Assume no Assets = capped by
Assets = change 70
percentage of the 100% of the
70
portfolio
Portfolio Value. value)
• Assume in the
example that the Portfolio Value = 95 Portfolio Value = 115 Portfolio Value = 115
allocation to the
risky asset is
capped at 100% of Assumptions: Multiple = 5, Floor = 90 (assume no change), Risky Asset Allocation Capped at 100% of Portfolio Value
the Portfolio Value.

168
Prof. Dr. Bernd Scherer

CPPI – Often Ignored Facts


Gap Risk
• If the risky asset falls by more than 1/multiplier without the ability to adjust
positions (overnight or when markets are closed) the portfolio will fall below
the floor value

CPPI Costs
• Costs arise from rebalancing, i.e. buying high and selling low. These costs
become larger if wither volatility or multiplier becomes larger
• Costs are path dependent, i.e. they differ from path to path

Volatility Risks
• CPPI is short vega, i.e. it will under-perform its projections if market volatility
increases heavily

Performance Measure
• Skewed distribution makes average return less likely than median return
• Performance is path dependent

169
Prof. Dr. Bernd Scherer

Economic Use of Derivatives

• Risk transfer (risk allocation, risk sharing)


– Efficient allocation of risk
• Lower costs
– Low costs access to baskets
• Limited liability
– Easier investments in risky assets; allows to protect subsistence levels
• Access to diversification
– Some asset classes can best be accessed by derivatives (commodities)
• Standardisation increases liquidity
• Price discovery
– Bubbles are less likely to develop if investors can short the underlying markets

170
Prof. Dr. Bernd Scherer

Session 5: Fixed Income Markets And Their Products

1. Perfect Replication - Bond Prices, Spot Rates, Forward Rates

2. Imperfect Replication – Duration

3. Swaps

4. Fixed Income Options

171
Prof. Dr. Bernd Scherer

Literature

• BRITTON –JONES (1999), Fixed Income and Interest Rate Derivative Analysis ,
Butterworth-Heinemann

172
Prof. Dr. Bernd Scherer

Zerobonds

• B(t,T) zerobond with maturity T at time t.


• s(t,T) spot rate with maturity T at time t.
1
B ( t ,T ) =
( 1 + s ( t,T ) )T −t

• Natural fixed income building blocks, even though they are hardly traded in the
market

173
Prof. Dr. Bernd Scherer

Forward Rates

• Forward rate is an interest rate which is specified now (t=0) for a loan that will
occur at a specified future (T1) date over the period T2-T1

Time 0 T1 T2

(1+ f ( 0,T ,T ) )
T2 −T1
Payment - -1 1 2

( 1 + s ( 0,T1 ) )T1 ( 1 + f ( 0,T1,T2 ) )T2 −T1 = ( 1 + s ( 0,T2 ) )T2

( 1 + s ( 0,T2 ) )T2
f ( 0,T1,T2 ) = T2 −T1 −1
( 1 + s ( 0,T1 ) )T1

s ( 0, T1 ) f ( 0, T1 , T2 )

s ( 0, T2 )

174
Prof. Dr. Bernd Scherer

Example

s ( 0,1 ) = 2.2% f ( 0,1, 2 ) = ?

s ( 0, 2 ) = 2.6%

( 1 + 2.2% ) ( 1 + f ( 0,1, 2 ) ) = ( 1 + 2.6% )2


( 1 + 2.6% )2
f ( 0,1, 2 ) = − 1 = 3%
( 1 + 2.2% )1

175
Prof. Dr. Bernd Scherer

Forward rates and expectations hypothesis

• From the current spot rates we can derive the path of furure spot rates (as implied
by the current spot rates)

f ( 1,2 ), f ( 1, 3 ),..., f ( 1,T )

• What happens if next periods spot rates ae eaual to this periods forward rates
(ecpectations hypothesis? All bonds earn (over next period) the same, i.e. the
curren spot rate.

B ( 1, 2 ) = 1 (
1+ f ( 1,2 ) , B 0, 2
) = 1
( 1+s ( 0,1 ) )( 1+ f ( 1,2 ) )
s ( 0,1 )
dB = B ( 1, 2 ) − B ( 0, 2 ) = 1+ f ( 1,2 ) − ( 1+s ( 0,1 ) )( 1+ f ( 1,2 ) ) =
1 1
( 1+s ( 0,1 ) )( 1+ f ( 1,2 ) )

dB s ( 0,1 )
= 1
( 1+s ( 0,1 ) )( 1+ f ( 1,2 ) ) ( 1+s ( 0,1 ) )( 1+ f ( 1,2 ) )
= s ( 0,1 )
B

176
Prof. Dr. Bernd Scherer

Couponbonds

• BC(0,Tn): price of a coupon bond at time 0


• Tn : Maturity
• Coupons: C
• Valuation (portfolio of zero bonds)

BC ( 0,Tn ) = C ⋅ B ( 0,1 ) + C ⋅ B ( 0, 2 ) + ... + C ⋅ B ( 0,Tn ) + 100 ⋅ B ( 0,Tn )

177
Prof. Dr. Bernd Scherer

Example

• You are given

Maturity 1 2 3 4
B(0,T) 0.98 0.95 0.92 0.89

• What is the value of this bond?

Maturity 1 2 3 4
B(0,T) 0.98 0.95 0.92 0.89
B(0,T)C 2.94 2.86 2.77 91.91

BC ( 0, 4 ) = 2.94 + .. + 91.91 = 100.47

178
Prof. Dr. Bernd Scherer

Yield to maturity (internal rate of return)

• Fictitious rate of return (Why?)

C1 Cn 100
−BC ( 0,Tn ) + + ... + + =0
( 1 + y ( 0,Tn ) ) ( 1 + y ( 0,Tn ) )n ( 1 + y ( 0,Tn ) )n

• Why is the yield of a ten year zero bond higher thqn the yield of 10 year coupon
bond if the term sructure is rising?

179
Prof. Dr. Bernd Scherer

Par Bonds

• Par Bonds sind Kuponanleihen, die zu pari (100) notieren, d.h. die Rendite des
Par Bonds entspricht seiner yield to maturity.
• Wie berechnet man den Kupon eines Par Bonds, bei gegebenen Nullkupon-
anleihen?

B c ( 0,Tn ) = C ⋅ B ( 0,1 ) + C ⋅ B ( 0, 2 ) + C ⋅ B ( 0, 3 ) + ⋯ + C ⋅ B ( 0,Tn ) + 100 ⋅ B ( 0,Tn )


100 = C ⋅ ∑ i B ( 0, i ) + 100B ( 0,Tn )
100 ( 1 − B ( 0,Tn ) )
C /100 = ,
∑i B ( 0, i )

180
Prof. Dr. Bernd Scherer

From Couponbonds to Zero Curves - Bootstrapping

• For a one period bonds we have y(0,1)=s(0,1)


• How can we calculate s(0,2)? Calculate B(0,2) and solve for s(0,2).

BC ( 0, 2 ) = C ⋅ B ( 0,1 ) + (C + 100 ) ⋅ B ( 0, 2 )
BC ( 0, 2 ) − C ⋅ B ( 0,1 )
B ( 0, 2 ) =
(C + 100 )

181
Prof. Dr. Bernd Scherer

Example

Price Coupon Maturity Return  BC ( 0,1 )   100 


 
 103 0
 0 
  
    
100 3 1 3.0% B =  B ( 0, 2 )  =  98  C =  3 103 0 
C  C  

     
98 3 2 3.5%  BC ( 0, 3 )  99   4
 4 104 

 
99 4 3 4.4%

BC = CB
  
 100   1 0 0   B ( 0,1 )   B ( 0,1 ) 
  
  
   
    
 = 
  ( )  =
  ( ) + ( )
  
98 3 103 0 B 0, 2 3B 0,1 103 B 0, 2 
 99   4 4 104     
      B ( 0, 3 )   4B ( 0,1 ) + 4B ( 0,2 ) + 104B ( 0, 3 ) 

B = BC C−1
 0,009708738  100   
 0 0  0,970873786 
    
=  -0,000282779   98  =  0,923178433 

0,009708738 0
   
 -0,000362537 -0,000373413 0,009615385  99   0,879074915 
 

182
Prof. Dr. Bernd Scherer

Duration

• Question: what happens to the price of a coupon bond if interest rates are
changing?

C1 C2 Cn 100
BC ( 0,Tn ) = + + ... + n +
( 1 + y ( 0,Tn ) ) ( 1 + y ( 0,Tn ) )2 ( 1 + y 0,Tn
( ) ) ( 1 + y ( 0,Tn ) )n
dBC ( 0,Tn )
= −1C 1 ( 1+y( 10,T ) )2 − 2C 2 ( 1+y( 10,T ) )3 − 3C 3 ( 1+y( 10,T ) )4 − ... − nC n 1
( 1+y ( 0,Tn ) )n +1
dy ( 0,Tn ) n n n

dBC ( 0,Tn ) 1 + y ( 0,Tn ) 1C 1 (1+y ( 0,T ) )1 + 2C 2 ( 1+y( 0,T ) )2 + 3C 3 (1+y( 0,T ) )3 + ... + nC n ( 1+y( 0,T ) )n
1 1 1 1
= − n n n n

dy ( 0,Tn ) BC ( 0,Tn ) C
( 0,Tn )
B 
 
<0

• What happens for less infinitesimal change in rates?

183
Prof. Dr. Bernd Scherer

Duration

• Linear approximation to a nonlinear relation

PV function (exact bond price)

dBC ( 0,Tn ) Approximation


dy ( 0,Tn )
B ( 0,Tn )
C

y ( 0, Tn )

184
Prof. Dr. Bernd Scherer

Duration Types

• Dollar duration: dollar value of 1 basis point (DV01)

dBC
∆B$C = ∆y = D$C ∆y
dy y =y

• Modified duration: percentage change in bond price

dBC 1 DC
DCmod = ⋅ C = $
dy y=y B BC

185
Prof. Dr. Bernd Scherer

3. Swap Market

Main Issues

• Definition of swaps
• Applications of swaps
• Fixed for fixed currency swaps
• Fixed for floating interest rate swaps

186
Prof. Dr. Bernd Scherer

Definition

• A swap is a transaction where at the time of the contract’s initiation the two
parties agree to exchange two cash-flow streams of equal present value. The
deal is structured as a single contract, with a right of offset.

• Example: Fixed for Floating Swap (receive a fixed rate and pay a floating rate) is
equivalent to a long position in a fixed bond and a short position in a floating
rate note. Swap spread creates equal present value.

187
Prof. Dr. Bernd Scherer

Application of Swaps

• Changing the exposure of a firm’s cash-flows to exchange rates, interest rates,


and commodity prices.

• Match interest sensitivity of asset and liability sides of balance sheet;

• Reduce financing costs or increasing debt capacity

• Taking advantage of tax regulations that


 differ across countries;
 differ in treatment of capital gains and income.

• Reducing transactions costs because several transactions are bundled together.

188
Prof. Dr. Bernd Scherer

Fixed for Fixed Currency Swap

• Suppose you want to swap a 7 year 18 million HC loan into a FC loan. The current
exchange rate is S = HC/FC = 1.8

• Why would a company want to do this? Get better loan in HC (better known by
domestic banks, less monitoring and screening costs).

• What is the economic position? This is a long/short portfolio of being long a HC


par bond (to pay off your HC loan) and short a FC par bond (which is the FC loan
you swapped the HC loan into)

189
Prof. Dr. Bernd Scherer

Fixed for Fixed Currency Swap

• Suppose FC par bond yield is 7% while HC par bond yield is 8%

1 1
∑ i =1 1.44 ( 1 + 0.08 )i
7
PVDC = + 18 = 18mDC
( 1 + 0.08 )7

1 1 FC
⋅ = [ ]
S 1.8 DC
1 1
∑ i =1 0.7 ( 1 + 0.07 )i
7
PVFC = + 10 = 10mFC
( 1 + 0.07 )7

190
Prof. Dr. Bernd Scherer

How do we value the swap after initialization?

• The swap value is driven by


– the domestic yield curve
– the foreign yield curve
– and the exchange rate

1
V = PVHC ( r1, r2 ,... ) − PVFC ( r1* , r2* ,... )
S

191
Prof. Dr. Bernd Scherer

Floating Rate Note

• Little interest rate risk, much credit risk


• Valuation: Floating-rate note cash-flows consist of a series of one-year forward
rates, derived from the current term structure. However discounting these
payments at the current spot rates will yield to a value of 1

Time t1 t2 t3 t4 … tn
Cash Flow 1
B ( 0,t1 )
−1 1
B ( t1 ,t2 )
−1 1
B ( t2 ,t3 )
−1 1
B ( t3 ,t4 )
−1 … 1
B ( tn −1 ,tn )
−1
Discount
B ( 0, t1 ) B ( 0, t2 ) B ( 0, t3 ) B ( 0, t4 ) … B ( 0, t3 )
Factor

Floating Coupon paid Discount factor


for year 3 Forward (zero bond price):
at the end of year 2
B ( 0, t2 ) B ( t2 , t3 ) = B ( 0, t3 )
B ( 0, t3 )
B ( t2 , t 3 ) =
B ( 0, t2 )

• Can you proof the discounted cash flows add up to one?

192
Prof. Dr. Bernd Scherer

Floating Rate Note - Value after Initialization

• Valuation
 
 
 
  1 
FRN ( t0 + dt ) = B ( t0 + dt , t1 )  1 +  − 1  
 value 
B ( t0 , t1 )  
 after
 known 
 reset Coupon 

• Suppose a floating rate note starts to contain some credit risk. It has been issued
with Libor + 0 bps, but the credit risk rises to Libor plus 50bps.
– How is the FRN affected?
– Can you use the formula above?
– Is the impact large or small?

193
Prof. Dr. Bernd Scherer

Fixed for Floating Swap

• Receive fixed pay floating

• Given this position is equivalent to a long position in a fixed bond (receive


coupon) and a short position in a floating rate note (pay floating) the value of
this portfolio is

V = BTC − FRN T

1

• The swap rate is a par bond rate!

194
Prof. Dr. Bernd Scherer

7. Fixed Income Options

• Learn how to price fixed income options / fixed income contingent claims
• Which capital market products contain fixed income options?
• What can we use fixed income options for?

195
Prof. Dr. Bernd Scherer

Simple Binomial Model

• Suppose interest rates for 6 and 12 month maturity are 10%. Suppose
further rates fall to 8% or increase to 12% during the next 6 month. Whats
the value of a call options (strike 95.5) on a zerobond with 6 month
maturity in?

1
2 ր 12% 1
2 ր 94
10% 1 ⇒ 95 1
2 ց 2 ց
8% 96

196
Prof. Dr. Bernd Scherer

Simple Binomial Model

• The call payoffs are given by max[B-X,0], i.e the payoff is either 0 (if B=94)
oder 0.5 (if B=96). The investor has two traded instrumentds to replicate
the payoff from this call:
– One year zerobond that costs 90 and pays off either 94 or 96 in 6 month
– Six month zero bond that costs 95 and pays off 100 for sure

• Find a portfolio of these two bonds that replicates the payoff from our call
option.
φ1100 + φ2 94 = 0
φ1100 + φ2 96 = 0.5

 100 94   φ1   0 
    =   ⇒ φ2 = 0.25, φ1 = −0.235
 100 96  φ2   0.5 

197
Prof. Dr. Bernd Scherer

Simple Binomial Model

• The value of our 6 month call on a 6 month zero bond is given by

C = 0.25 ⋅ 90 − 0.235 ⋅ 95 = 0.175

• Note: We determined the cal value by pure arbitrage. No assumption about


preferences, risk aversion, etc. was made

198
Prof. Dr. Bernd Scherer

Multi-period valuation

• The real worl is not binomial and sometimes we need the whole interest rate
path to evaluate options (early exccise).

• Dynamic programming method, i.e. Backward induction (solution process starts


from the last period).

• We asssume we already calibrated an interest rate tree (i.e. A tree that correctly
prices a set of reference options in a risk neutral world)

199
Prof. Dr. Bernd Scherer

Three Period Example

• Interest rate and price trees – calculate value of a 3 period zero bond from
belows interest treee

• First we calculate the values of a 3 year zero bond in period 2 (trivially its 100 in
period 3)

100
1
ր 14% 1
ր = 87.72
2
12% 1
2
12% 1 1.14
1 1
2 ր 2 ց 2 ր
ց 100 2
10% 1 10% ⇒ 10% 1 = 90.01
2 ց 1
2 ր 2 ց 1
2 ր 1.1
8% 1 8% 1 100
2 ց 6% 2 ց = 94.34
1.06

200
Prof. Dr. Bernd Scherer

Three Period Example (continued)

• In the „up“ state of year 1the risk free rate is 12% (see previous slide). Given
that the binomal tree represents the risk neutral world (all calims can be
replicated arbitrage free), the expected return also eaquals 12%.

0.5 ( 87.72 + 90.01 )


1.12 =
x
0.5 ( 87.72 + 90.01 )
x = = 79.95
1.12

• The same also applies to the down state


0.5 ( 90.91 + 94.34 )
y = = 85.76
1.08

201
Prof. Dr. Bernd Scherer

Three Period Example (continued)

• New price-tree

1
2 ր 87.72
1
79.95 1
2 ր 2 ց
10% 1 1
90.01
2 ց 2 ր
85.76 1
2 ց 94.34

• We can now calculate todays value for the three year zero bond

0.5 ( 79.95 + 85.76 )


z = = 75.23
1.1

202
Prof. Dr. Bernd Scherer

Exercise

• We want to value the option to buy a one year zero bond at a strike of 90 in two
years from now. Assume the previous interest rate tree.

1
2 ր 0
1
?1
2 ր 2 ց
?1 1
0.01
2 ց 2 ր
?1
2 ց 4.34

203
Prof. Dr. Bernd Scherer

HO/LEE - Model

• Recombining interest rate tree for the die „short rate“, Allows us to price
long bonds, opttions,etc.

• (2,2) denotes the state of the world where rates rose twice in a row.

1
2 ր ...
( 3, 3 )
1 1
2 ր 2 ց
( 2, 2 )
1
2 ր 1
2 ց 1
2 ր ...
( 1, 1 ) ( 3, 1 )
1 1 1 1
2 ր 2 ց 2 ր 2 ց
( 0, 0 ) ( 2, 0 ) ...
1 1 1 1
2 ց 2 ր 2 ց 2 ր
( 1, -1 ) ( 3, -1 )
1 1 1
2 ց 2 ր 2 ց ...
( 2, -2 )
1 1
2 ց 2 ր
( 3, -3 ) ...
1
2 ց

s ( ti , j ) = s ( 0, 0 ) + ( ∑ tt =1 at )dt + j σ
i
dt

204
Prof. Dr. Bernd Scherer

Binomial Tree in HO/LEE model

• Calibration means to fit the free parameters of a tree to a set of bond


prices assumed to be correct (option valuation is relative valuation)

1
2 ր ...
1
s + (a1 + a 2 + a 3 ) + 3σ 1
2 ր 2 ց
s + ( a1 + a 2 ) + 2σ
1
2 ր 1
2 ց 1
2 ր ...
1
s + a1 + 1σ 1 1
s + ( a 1 + a 2 + a 3 ) + 1σ 1
2 ր 2 ց 2 ր 2 ց
s s + ( a1 + a 2 ) + 0σ ...
1 1 1 1
2 ց 2 ր 2 ց 2 ր
s + a 1 − 1σ 1 1
s + ( a 1 + a 2 + a 3 ) − 1σ 1
2 ց 2 ր 2 ց ...
s + ( a 1 + a 2 ) − 2σ 1 1
2 ց 2 ր
s + (a1 + a 2 + a 3 ) − 3σ 1 ...
2 ց

205
Prof. Dr. Bernd Scherer

Example tree

• BRITTEN/JONES, p. 92, assumes zero drift and one percent volatility


1
ր8
5 + 0 + jσ
2
1
7 1
1
1
= 5 + 2 ⋅1 = 7
2 ր 2 ց
6
1
2 ր 1
2 ց 1
2 ր 6
5 1 1
5 1
ց ր ց
2
4
2 2 4 5 + 0 + jσ 1
1
= 5 −1⋅1⋅ 1
1
=4
1 1
2 ց 2 ր
3 1
2 ց 3

• We can now introduce state price securities -

206
Prof. Dr. Bernd Scherer

Valuation of 4 year zero

0.5 ⋅ 100 + 0.5 ⋅ 100


1 + 8 /100

100
92.5925926
87.3515024 100
83.9918293 94.3396226
82.3449307 90.7111756 100
88.9325251 96.1538462
94.2684766 100
98.0392157
100
0.5 ⋅ 94.33 + 0.5 ⋅ 96.15
t=0 t=1 t=2 t=3 t=4 1 + 5 /100

207
Prof. Dr. Bernd Scherer

Valuation of a Coupon Bond

0.5 ⋅ 100 + 0.5 ⋅ 100 + 4


1 + 8 /100

100
96.2963
94.5838803 100
92.8884448 98.1132
94.8054713 94.3396226 100
98.2030449 100.0000
101.922711 100
101.9608
100

t=0 t=1 t=2 t=3 t=4

208
Prof. Dr. Bernd Scherer

Option Valuation

• Call option on 1 year zero bond in 3 years with strike 95

0.5 ⋅ 1,15 + 0.5 ⋅ 3, 04


1 + 3 /100

0
0
0.25917479 0
0.7152017 0.54945055
1.24274878 1.15384615
2.03546691 max ( 98.03 − 95, 0 )
3.03921569

t=0 t=1 t=2 t=3

209
Prof. Dr. Bernd Scherer

Problems

• Short rate can become negative


• No mean reversion
• Constant interest rate volatility
• From here a very rich literature of term structur models developed (not the
objective of our course)

210
Prof. Dr. Bernd Scherer

Use of Fixed Income Options – Callable Bonds

• Callable bonds: gives the issuer the right to buy back the bond at a given
price (for example at par, i.e. at 100)

• Callable bond = Straight Bond – Value of Call Option


– If interest rates fall, bond prices rise and the value of the call rises
– Callable bond has the same downside but a lower downside (duration
compression) as a straight bond.
– Higher coupon (collect premium from short call)

• Who issues callable bonds?


– Government sponsored entities, mortgage financing.
– Hedge against free option given to borrower (take out new loan and pay back the
old loan)

• Suppose you are a pension fund with 30 year duration liabilities. You buy a
30 year duration mortgage bond. Is this a good hedge?

211
Prof. Dr. Bernd Scherer

Use of Fixed Income Options – Cap

• Caps are interest rate call options on interest rate (hedges against rising
rates). Payoff:
Cap = max [ r − cap, 0 ]

• Can you value a cap with cap = 5 for the below tree?
1
2 ր8
1
71
2 ր 2 ց
61
1
2 ր 2 ց 1
2 ր6
51 1
51
2 ց 2 ր 2 ց
41 4
1
2 ց 2 ր
31
2 ց3

• Caps help limit exposure to rising rates yet allow the borrower to benefit
when rates fall;

212
Prof. Dr. Bernd Scherer

Session 6: Credit Markets and Their Products

1. Defining Credit Risk

2. Credit Derivative Products

3. Applications of Credit Derivatives

4. Pricing of Credit Derivatives

213
Prof. Dr. Bernd Scherer

Literature

• MEISSNER G. (2006), Credit Derivatives, Blackwell

• HULL J. (2009), Option, Futures and Other Derivatives, Prentice Hall

• SKINNER F. (2005), Pricing and hedging interest and credit sensitive


instruments, Elsevier

214
Prof. Dr. Bernd Scherer

1. Credit Risk

• Rating based approach


• Equity based approach
• Credit correlation

215
Prof. Dr. Bernd Scherer

1. Defining Credit Risk

• Default risk: risk an obligor does not repay parts of his financial obligation

• Credit deterioration risk: risk that the credit quality of the debtor
decreases.
– Downgrade to a lower rating category
– Spread widening leads to mark to market losses
– Forced selling due to rating constraints; can no longer be used as collateral

• Credit Derivatives: financial instruments to transfer credit risk from one


counterparty to another counterparty.

216
Prof. Dr. Bernd Scherer

Credit Risk

• Is it in the shareholders interest to obtain a high rating?

• Shoud shareholders care about bankruptcy risk?

• Discuss the following statement: „Each company chooses its own


rating“.

217
Prof. Dr. Bernd Scherer

Credit Rating Approach

• Rating agencies (the biggest are MOODYS, S&P, FITCH) create ratings on bonds in
accordance with their default probability.

• AAA represents the best rating (smallest likelihood to default in a given period), AA
represents the second best rating, …

• One way to measure credit risk is to model the transition between rating caegories.
This is usually done with a transition matrix where it is assumed that transition
probabilities follow a MARKOV process (see next slides).

• Problems with MARKOV approach


– Rating changes are autocorrelated (Markov process assumes no memory)
– Rating changes will be regime dependent (transition matrix changes over time)
– Transition matrices are derived from observable bonds. Might be misleading to
apply to loans, given that covenqats, collateral etc. might induce different
behaviour in the loans market.
218
Prof. Dr. Bernd Scherer

Transition (migration) matrix

• The probability for a AAA bond stay AAA in one year time is 90.82%. It will
migrate to BB (within a year) with a 0.07% probability.

AAA AA A BBB BB B CCC D


AAA 90.82% 8.27% 0.74% 0.07% 0.07% 0.01% 0.01% 0.01%
AA 0.66% 90.89% 7.70% 0.58% 0.06% 0.07% 0.30% 0.01%
A 0.08% 2.42% 91.31% 5.23% 0.68% 0.23% 0.01% 0.04%
BBB 0.04% 0.32% 5.88% 87.46% 4.96% 1.08% 0.11% 0.15%
BB 0.04% 0.13% 0.64% 7.71% 81.16% 8.40% 0.97% 0.95%
B 0.02% 0.11% 0.26% 0.52% 6.86% 83.49% 3.89% 4.80%
CCC 0.13% 0.12% 0.42% 1.20% 2.69% 11.71% 64.48% 19.25%
D 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% 100.00%

219
Prof. Dr. Bernd Scherer

MARKOV chain – Multiple Periods

• What is the likelihood for a


AA bond to be single B after
10 years? There are multiple ∑ j
pij = 1
paths to arrive there. How
do we calculate all of them?
 pAAA, AAA pAAA, AA … pAAA,D 
p p … p 
P = P0,1 =  AA , AAA AA, AA AA, D 
 ⋮ ⋮ ⋱ ⋮ 
 
 0 0 1 

 
T

P0,n = P ⋅ P ⋅… ⋅ P ⋅ P = PT

220
Prof. Dr. Bernd Scherer

Risk Neutral Transition Probabilities

• Assume we know the price of two zero bonds. One without credit risk B(0,T)
and one with credit risk BAA(0,T). In a risk neutral world we would expect

BAA ( 0,T ) = (1− qAA ) ⋅1⋅ B( 0,T ) + qAAδ B( 0,T )


B( 0,T ) − BAA ( 0,T ) 1
qAA =
B( 0,T ) 1−δ

• Here, qAA denotes the risk neutral probability to default any time between
time 0 and time T in Konkurs

• In this example qAA denotes the second element of the last column in Q0,T (T-
Period transition matrix with risik neutral probabilities)

221
Prof. Dr. Bernd Scherer

Risk neutral transition matrix

• We need to transform the real world transition matrix P via a choice of Π to Q. The
matrix Π contains risk premia to adjust real world to risk neutral probabilities
Qt,t+1 =I+Πt ( P-I)
Π(1) 0
1   
 1 0   Π( 2) 0 0
I=   ,Π =  0 ⋱ 0
 0 ⋱  t  
   Π( k ) 0
 1
 0
 0 0 0 1

• We choose the diagonal elements in P such that each period satisfies


BAA ( 0,1) = (1− qAA ) ⋅1⋅ B( 0,1) + qAAδ B( 0,1)

– For period one we fit k risk premia to k rating categories according to the above
– As in bootstrapping we continue to work ourselve through the term structure for all
further periods
– With these estimates we can then price credit derivatives (see later section)

222
Prof. Dr. Bernd Scherer

(Real World) Credit RIsk - Zero bonds for different ratings

• Discount factors for different credit risks

B(0,1) B(0,2) B(0,3) B(0,4) B(0,5)


AAA 96.48% 92.71% 88.13% 83.44% 78.43%
AA 96.43% 92.63% 88.03% 83.28% 78.24%
A 96.37% 92.49% 87.81% 83.00% 77.91%
BBB 96.26% 92.24% 87.43% 82.52% 77.32%
BB 95.86% 91.33% 85.97% 80.48% 78.35%
B 95.30% 90.12% 84.01% 77.85% 71.43%
CCC 94.48% 88.27% 81.04% 73.94% 66.69%

• Zero bond valuation


C
BBB ( 0,5) = 4.4⋅ 0.9586+ 4.4⋅ 0.9133+…+104.4⋅ 0.7835

223
Prof. Dr. Bernd Scherer

Forward price for different ratings

B( 0,2)
B( 0,1)
= 0.9648
0.9271
= 0.9609

B(1,2) B(1,3) B(1,4) B(1,5)


AAA 96,09% 91,35% 86,48% 81,80%
AA 96,06% 91,29% 86,36% 81,65%
A 95,98% 91,12% 86,11% 81,37%
BBB 95,83% 90,82% 85,72% 80,84%
BB 95,27% 89,68% 83,96% 78,58%
B 94,56% 88,15% 81,70% 75,69%
CCC 93,42% 85,77% 78,25% 71,52%

• What is the price for a 5 year BB corporate bond in one year, i.e. with 4 years
remaining live?
C
BBB (1,5) = 4.4 + 0.9527 ⋅ 4.4 + … + 104.4 ⋅ 0.7858 = 100.79

224
Prof. Dr. Bernd Scherer

BBB – bond distribution in 1 year

• 4,4% Coupon, 5 year maturity


• Distribution of credt risk is asymmetric. There are much more potentially bad than
good news for a AAA bond (market risk is symmetric, credit risk is not)
• Calculated from previous tables (please replicate)

Rating Probability Price


AAA 0.04% 101.85
AA 0.13% 101.68
A 0.64% 101.37
BBB 7.71% 100.79
BB 81.16% 98.27
B 8.40% 95.06
CCC 0.97% 90.40
D 0.95% 52.00

225
Prof. Dr. Bernd Scherer

Diversification: 8 uncorrelated BBB bonds

• Still asymmetric

ExtValue (0,0042281; 0,0072182)

-40 -30 -20 -10 0 10 20 30


Portfoliorendite ( x 1000)

226
Prof. Dr. Bernd Scherer

Diversification: 30 uncorrelated BBB bonds

Normal(0,0073082; 0,0030345)

-10 -5 0 5 10 15 20
Portfoliorendite (mal 1000)

227
Prof. Dr. Bernd Scherer

Option Pricing (MERTON / KMV ) approach

• Equity as long call

228
Prof. Dr. Bernd Scherer

Option Pricing (MERTON / KMV ) approach

• Firm value = value of equity plus value of debt

• We can apply the BLACK/SCHOLES model to price the value of firm equity, where
V represents the firms assets, D its outstanding debt, r the risk free rate, T
time to maturity, and σV the volatility of assets. The value of equity is given by

E 0 = V0N ( d1 ) − De −rT N ( d2 )

d1 =
ln (VD ) + ( r + 1 σ2
2 V )T
, d2 = d1 − σV T
σV T

V0 dV
σe = σ
E 0 dE V

• The default probability is given by N ( −d2 )


• The credit spread is given by the excess yield (over risk free) from the risky
corporate bond
• We can determine V0 and σV from the two equations above.

229
Prof. Dr. Bernd Scherer

Credit Correlation

• Credit (default) correlation describes the tendency for two companies to default at
the same time.
– Similar industry, leverage
– Similar business risk

• Non-zero default correlation is the reason why credit risk does not completely
diversify away (and we real world and risk neutral default probabilities are not the
same)

230
Prof. Dr. Bernd Scherer

Market Model For Credit Correlation

• Define t1 and t2 as the time to default for company 1 and company 2. Both
can (and will) be highly non-normal with cumulative distribution Q1 ( t1 ) and
Q2 ( t2 ) . These could be taken from a migration matrix.

• Example: Assume the cumulative probability to default in 1,2,3,4, and 5


years of time to be 1%, 3%, 6%, 10% and 15% for both companies. Use a
percentile to percentile distribution, i.e. use

−1
xi = N (Qi ( ti ) ) .

inverse of
cummulative
standard
normal

231
Prof. Dr. Bernd Scherer

Market Model For Credit Correlation

This allows us to first draw x i and then work out ti . For example if we draw

x i < −2.33 [ = N −1 ( 0.01 ) ]

from the standard normal, default happened in the first year. If

−2.33 [ = N −1 ( 0.01 ) ] < x i < −1.88 [ = N −1 ( 0.03 ) ]

Default happened in the 2nd year …

We can assume a correlation between x1 and x 2 to model the correlation in highly


non-normal default times.

232
Prof. Dr. Bernd Scherer

Market Model For Credit Correlation

Assume a “market model” for x i , where M (common default factor) and


zi (individual default factor) are standard normal, while ai is the correlation to the
common default factor (could use stock market correlation)

x i = ai M + 1 − ai2 zi

Default happens if

x i = ai M + 1 − ai2 zi < N −1 (Qi (T ) )


N −1 (Qi (T ) ) − ai M
zi <
1 − ai2

The probability of default (conditional on M ) for time T is therefore

 N −1 (Qi (T ) ) − ai M 
Qi (T | M ) = N  
 1 − ai2 

233
Prof. Dr. Bernd Scherer

Market Model For Credit Correlation

• Assume 100 bonds with ai = a = 0.1 and Qi (T ) = Q (T ) , i.e. a homogenous


0.06
0.04
0.02
0.0
universe

0 10 20 30 40

number of bonds defaulting after 4 years


234
Prof. Dr. Bernd Scherer

2. Credit Derivative Products

• Credit Default Swaps


• Equity Default Swaps
• Total Return Swaps
• Credit Spread Products
• Collateralized Debt Obligation

235
Prof. Dr. Bernd Scherer

Credit Default Swaps (CDS)

• CDS buyer makes a periodic payment to the seller of the default swap. The
default swap seller promise to make a payment in the event of default of a
reference bond or loan.
• This is more like a put option rather than a swap.

Periodic
Premium
CDS
CDS
Buyer
Seller
Contingent
payment

• ISDA documentation will specify the credit event


• Cash settlement (CDS buyer receives the difference between the price of
defaulted bond and 100)
• Physical settlement (CDS buyer delivers bond and receives 100))

236
Prof. Dr. Bernd Scherer

CDO Structure

• Cash flows can be structured to cater risk aversion with the claim to create
value (departure from neoclassical finance and wrong)
• Offers high upside limited liability exposure to low rated bonds
• Success of CDO depends on selling the risky equity tranche
– Long call on underlying asset pool

AAA,3%

AA, 4%
A, 5%
BBB, 9%

Equity, 15%

237
Prof. Dr. Bernd Scherer

Rating Transformation – Rating Arbitrage

• Mostly B rated assets have been transformed into higher rated assets
under the critical assumption of low correlation (i.e. zero systemic risk)

238
Prof. Dr. Bernd Scherer

Credit Indices

• Credit markets developed indices to track default swap spreads

1. CDX NA IG (5 year / 10 year) tracking the spread of 125 North


American investment grade companies.

2. iTraxx Europe (5 year / 10 year) tracking the spread of 125 North


American investment grade companies.

• Investors can buy credit protection on these portfolios

239
Prof. Dr. Bernd Scherer

Single Tranching

• Use Credit indices to define standard CDO tranches.


• Each tranche can be traded on its own.

240
Prof. Dr. Bernd Scherer

3. Application of Credit Derivatives

• Regulatory Arbitrage
• Hedging
• Yield Enhancement

241
Prof. Dr. Bernd Scherer

Principles of Regulatory Arbitrage

• Regulatory capital refers to the amount of capital (shareholders equity) a


financial institution must hold because of regulatory requirements.

• One function of capital – the function that regulators care about – is to


insulate banks from losses.

• Regulators impose capital requirements in order to help ensure the safety


and soundness of banks. Why?

• The question is how much capital should be required, and the key concept
is risk-based capital.

242
Prof. Dr. Bernd Scherer

Principles of Regulatory Arbitrage – Risk Based Capital

• The theory is simple: for any asset hold an amount of capital proportional
to its risk. The tricky thing is putting this into practice:
– The potential complexity of financial transactions far exceeds the
ability of regulators to specify rules for every one.

• Simple Solution: 1988 Basel Accord (now known as Basel I) - introduced


international standards for risk-based capital requirements.
– Banks have to hold capital equivalent to 8% of their risk-weighted
assets. Each type of asset has a risk weight that reflects its riskiness.
– OECD government bonds have a zero risk weight – theoretically, they
have zero risk, and hence require zero capital.
– Problem: Singapore (AAA) but non OECD has 100% risk weighting while
MEXICO (A) OECD member only has 0% risk weighting, Capital is
misallocated to OECD countries

243
Prof. Dr. Bernd Scherer

Basel I: example

• Home mortgages have a 50% risk weight; and uncollateralized commercial


loans have a 100% risk weight.
• If a bank held $100 in Treasuries (0 weight), $100 in home mortgages (50%
weight , and $100 in commercial loans (100% weight) it would have $300 in
assets, but only $150 in risk-weighted assets (0% * $100 + 50% * $100 +
100% * $100); therefore would have to hold $12 in capital (8% * $150).

244
Prof. Dr. Bernd Scherer

Regulatory Arbitrage – Returns From Leverage

• Regulatory capital arbitrage happens because, all other things being equal,
banks would like to hold less rather than more capital. The reason is that, in
general, bank profits are proportional to the amount of assets that they
hold.
– One main source of banking profits is interest margin: the spread
between the interest charged on loans and the interest paid on
deposits and other sources of funding. For any given interest margin,
profits will be strictly proportional to loan volume (assets).
– The same logic applies to banks’ principal investment and trading
businesses; for any given strategy, doubling the size of the position will
double the expected profit. So to increase profits, you have to increase
assets. If a bank wants to increase its assets, it can do so either by
increasing its leverage (lowering its capital as a percentage of assets) or
increasing its capital; the former is preferable, because the latter
requires issuing new shares, which dilutes current shareholders. (Also,
issuing new shares results in lower earnings per share, lowering the
stock price.)

245
Prof. Dr. Bernd Scherer

Strategies for Regulatory Arbitrage - Securitization

• Our bank earlier had $100 in mortgages, for which it had to hold $4 in
capital.

• Assume it creates a simple collateralized debt obligation (CDO) out of these


mortgages.
– It sells them to a special-purpose vehicle (SPV) that issues bonds to investors;
these bonds are backed by the cash flows from the monthly mortgage payments.

• The bonds are divided into a set of tranches ordered by seniority (priority),
so the incoming cash flows first pay off the most senior tranche, then the
next most senior tranche, and so on.
• If these are high-quality mortgages, all the credit risk (at least according to
the rating agencies) can be concentrated in the bottom few tranches
(because it’s unlikely that more than a few percent of borrowers will
default), so you end up with a few risky bonds and a lot of “very safe” ones.

246
Prof. Dr. Bernd Scherer

Strategies for Regulatory Arbitrage - Securitization

• Getting sufficiently high credit ratings for the senior tranches, the bank can
lower the risk weights on those assets, thereby lowering the amount of
capital it has to hold for those tranches.
– The risky tranches will require more capital, but it is possible to do the math so that the lower
capital requirements on the senior tranches more than outweigh the higher requirements on
the junior tranches.

• The passive side of the CDO will have a lower risk weighting than the
active side of a CDO (up to 50% in some cases)

• An extension of this strategy is to selectively sell some tranches and hold


onto others. In this way, a bank can end up with assets that have a high
degree of economic risk but a low risk weight for capital purposes.

• All this is possible because the rules setting capital requirements are lumpy
while there is an infinite range of actual financial assets.

247
Prof. Dr. Bernd Scherer

Regulatory Arbitrage – Did Basel II Help?

• Recall from session 2: Basel II introduces VaR based methods as well as a finer grid
on credit risk (recall Basel I was very crude). Below are two reasons Basel II failed

1. The shift from fixed-percentage capital requirements (Basel I) to value-at-risk (VaR)


methodologies (Basel II) increased the potential for regulatory arbitrage.

– In VaR, the riskiness of any asset is determined by a model based on the historical
attributes of the asset. In theory, this is an improvement, because it gets around the
problem of lumpy fixed percentages, and tailors the risk weight to the unique
characteristics of the asset itself.

– In practice, however, it made it possible to assess riskiness based on small amounts of


historical data from periods during which, for example, subprime loans rarely defaulted
because rising housing prices always made it possible to refinance. By underestimating
the risk of certain assets, these models underestimated the capital required to support
these assets.

2. Securitization got around finer rating grid. This will continue!

248
Prof. Dr. Bernd Scherer

Hedging - CDS and Regulatory Capital Relief

• Assume 1million USD BBB rated non OECD debt

• Basel I: 100% risk weight for non OECD, i.e. regulatory capital amounts to

1000000 ⋅ 100% ⋅ 8% = 80000

• Basel II: BBB risk weighting is 50%, AAA banks have 20% risk weight, Residual risk
from CDS hedge (contractual risks, enforceability, …) is 15%
 
 
  = 19600
1000000 ⋅  15%  ⋅ 50% + (
 1-15% ) ×20%
 
 BBB risk weight
 CDS risk weight 
 after hedge 

249
Prof. Dr. Bernd Scherer

4. Pricing of Credit Derivatives

• Credit Default Swaps


• Collateralized Debt Obligation

250
Prof. Dr. Bernd Scherer

Credit Default Swaps – Asset Swap Based Pricing

• Risk-free Bond + (Short Default Swap) = Risky Bond

• Example: Government bond yields 4%, CDS premium is 2%, Corporate bond
trades at 7%. Is there arbitrage?

• What are the limits to this approach?

– Only works for par bonds, what happens when interest rates change.
What happens to interest rate sensitivity close to default?

– Are the positions comparable in liquidity?

– Counterparty risk?

251
Prof. Dr. Bernd Scherer

Credit Default Swaps – Using Default Probabilities

• Start with risk neutral default probabilities (for example from transition
matrix)

Prob to default Probability to


Time (T)
in year T survive year T
1 0.0200 0.9800
2 0.0196 0.9604

3 0.0192 0.9412 0.983 = 0.9412


4 0.0188 0.9224
5 0.0184 0.9039

 ⋅ 0.02
0.9604  = 0.0192
survive default
to T −1 in T

252
Prof. Dr. Bernd Scherer

Credit Default Swaps – Using Default Probabilities

• Calculate PV of premium leg (what insurance buyer pays)


• Assume continuous compounding with a risk free rate of 5%
• Assume 40% recovery rate
• Unknown premium (spread) s

Probability to PV of e−0.05⋅2
Time Expected Discout
survive year expected
(T) Payment Factor
T Payment
1 0.9800 0.9800s 0.9512 0.9322s
2 0.9604 0.9604s 0.9048 0.8690s 0.9604 ⋅ s ⋅ 0.9084 = 0.8690s
3 0.9412 0.9412s 0.8607 0.8101s
4 0.9224 0.9224s 0.8187 0.7552s
5 0.9039 0.9039s 0.7788 0.7040s

0.9322s + 0.8690s + ... + 0.7040s = 4.0704s


4.0704s

253
Prof. Dr. Bernd Scherer

Credit Default Swaps – Using Default Probabilities

• PV of insurance leg (payout of 1-recovery rate conditional on default)

PV of
Prob to default in Recovery Expected Discount
Time (T) Expected
year T Rate Payoff Factor
Payoff
0.5 0.0200 0.4 0.0120 0.9753 0.0117
1.5 0.0196 0.4 0.0118 0.9277 0.0109
2.5 0.0192 0.4 0.0115 x 0.8825 = 0.0102
3.5 0.0188 0.4 0.0113 0.8395 0.0095
4.5 0.0184 0.4 0.0111 0.7985 0.0088

0.0192 ⋅ ( 1 − 0.4 ) ⋅ 1
0.0511

254
Prof. Dr. Bernd Scherer

Credit Default Swaps – Using Default Probabilities

• PV of accrual payment ( we assumed that default happened mid year so


end of year you have to pay the premium for the first half)

PV of
Prob to default Expected Discount
Time (T) expected
in year T Accrual Factor
accrual
0.5 0.0200 0.0100s 0.9753 0.0098s
1.5 0.0196 0.0098s 0.9277 0.0091s
2.5 0.0192 0.0096s 0.8825 0.0085s
3.5 0.0188 0.0094s 0.8395 0.0079s
4.5 0.0184 0.0092s 0.7985 0.0074s

0.0426s

• The “fair” CDS spread value is given by

4.0704s + 0.0426s = 0.0511


s = 0.0124

255
Prof. Dr. Bernd Scherer

Credit Default Swaps – Is the Market a Fair Game?


• Equities, Fixed Income, Currencies – no reason to assume counterparties
are better informed

• Credit market – asymmetric information (see lemon problem), ie some are


better informed

• Banks buy protection against default (is this decision done by risk manager
or speculator)

• Do insurance companies (seller of CDS) have the same information? What


happened to AIG?

256
Prof. Dr. Bernd Scherer

CDO Pricing
• The value of the CDO must be the sum of the bond values (asset side of the
balance sheet)!

• CDO pricing is about pricing the relative value of different stakes (tranches).
Each tranch holds options against other tranches. Industry practice of rating
based discount rates is wrong!

• Credit correlation is the most crucial input into CDO pricing

• Example: 5 years n-th to default basket. Assume 100 bonds with 2% default
probability over 5 years.
– With zero default correlation the likelihood that one or more bonds default is
86.74% (1-0.98^100), while the likelihood that 10 or more default is 0.0034%
– First to default is valuable, while 10th to default is not
– If correlation increases the likelihood of one or more declines while the
likelihood of 10 or more increases
– For perfect correlation there will be no difference between 1st and nth to default

257
Prof. Dr. Bernd Scherer

Session 7: Currency Markets

1. Spot Market

2. Forward Market

3. International Parity Conditions

4. International Diversification

258
Prof. Dr. Bernd Scherer

Literature

• EUN/RESNICK (EU), 2009, International Financial Management, McGrawHill

• MOFFEIT/STONEHILL/EITEMAN (MSE), 2006, Foundations of Multinational


Finance, Pearson

259
Prof. Dr. Bernd Scherer

1. The Spot Market

• Key Questions
– How is the market for trading spot exchange organized?
– How are spot exchange rate quoted?

• Motivating Problem
– You work for a large multinational firm that has its headquarters in Paris.
The multinational firm has subsidiaries in 2 countries—Japan and Canada.
All foreign sales are made through these subsidiaries. Each month, the
subsidiaries remit to the Paris office their income over the month. This
income is in the local currency of the country in which the subsidiary is
located (JPY and CAD). The exchange rates that are quoted to you are in
terms of the USD; that is, USD/JPY and USD/CAD.

• How do you find the rates in terms of EUR. If the inflows are JPY 100 m and
CAD 10 m, what is their total value in terms of EUR?

260
Prof. Dr. Bernd Scherer

The Spot Rate - Definition

Definition
• The spot rate is the amount of home currency one pays/receives in
exchange for one unit of foreign currency today.

Spot contract signed at t = 0


Exchange of HC and FC at t = 0

St ST

Example
– If one needs USD 100 to buy a copy of the textbook, then the exchange rate
is USD 100/textbook.
– If one receives USD 2000 when selling a computer, then the exchange rate
USD 2000/computer.

261
Prof. Dr. Bernd Scherer

Quotation

• We quote exchange rates in terms of HC/FC.


– This quote tells us the price of the FC.
– If the so quoted currency goes up, what does this mean?
– Thus, statements about buying or selling will always refer to the currency in the
denominator (the “foreign” currency).
– This convention, standard in continental Europe, is called the direct quote.
– To keep things clear, you should remember that the object of interest (the currency
that we are buying or selling) is the one that is in the denominator.

• Example
– 30.000 USD/car—is the price for a car with the ”item” being bought or sold, in this case
a car, in the denominator
– 2.03 USD/GBP is the price for one British Pound, i.e. you need to pay 2.03 USD for one
GBP. Where is the exchange rate today?

262
Prof. Dr. Bernd Scherer

The Spot Rate - Example

TABLE 1: Foreign exchange spot rates as on 11 September 2007

USD as FC USD as HC
263
Prof. Dr. Bernd Scherer

Market Organization

• Currency trading does not occur on an organized exchange


– In contrast to stock markets and futures markets, trading is not limited to a
particular time, there is no centralized clearing mechanism, and contracts are not
standardized.
• The currency market consists of
– a wholesale tier, (an informal network of about 500 banks and currency brokers
that deal with each other and with large corporations), and
– a retail tier.
• Most interbank dealing is done electronically
• Many players in the wholesale market act as market makers.
• Market maker makes a two-way quote (that is, bid and ask quotes), you
don’t have to reveal whether you intends to buy or sell. Limits to the market
maker’s commitment to this quote.
– Time, i.e. immediate execution (Why?)
– Size, quote is only good for a limited amount (Why)
• Deal via a broker (tries to shop around for a commission, about half the
market)

264
Prof. Dr. Bernd Scherer

Market Statistics

• Size:
– The daily volume of trading on the exchange market (including the currency
futures, options, and swaps markets) is more than USD 1800 billion.
– It is about five to ten times the daily volume of international trade in goods and
services.

• Location
– The major markets are, in order of importance, London, and New York, with less
important markets being Singapore, Zurich, Hong Kong, and Frankfurt.

• Currencies
– The most important markets, per currency, have been USD/EUR and the USD/JPY;
together they represent over half of the world trading volume.

265
Prof. Dr. Bernd Scherer

Currency Markets by Delivery Date

• Spot Market
– the exchange market for payment (of home currency) and delivery (of foreign
currency) ”today”.
– In practice, ”today” means the same day only when you buy or sell notes or coins
(and this part of the market is very small).
– For ’electronic’ deposits, delivery is within two working days for most currencies,
and one day between Canada and the US
• Forward Market: exchange market for contracts signed today but payment
and delivery take place at some future date.
– The forward market consists of many sub-segments corresponding to different
delivery dates, with each sub-segment having its own price.
– The most active forward markets are for 30, 90, 180, 270, and 360 days, but
bankers nowadays quote rates up to ten years forward.
 Note that months are indicated as 30 days.
 A 30-day contract is settled one month later than a spot contract, and a 180-day
forward contract is settled six months later than a spot contract—each time
including the two-day delay convention.
 You can always obtain a price for non-standard maturities, too, for instance 64 days
(two months and four days), or for a specific date.

266
Prof. Dr. Bernd Scherer

Conventions For Quoting Spot Rates

• We first look at the simple (but unrealistic) case where there are no transactions
costs; then, we consider the case with bid-ask spreads.

• See TABLE 1:
– One needs USD 0.83 to buy one AUD; thus, the exchange rate is USD/AUD 0.83.
– The fourth column treats the USD as the HC and the other currency as the FC.
– The third column treats the USD as the FC and the other currency as the HC. Thus, the
numbers in the third column are the inverse of the numbers in the fourth column.

• Example: Inverting exchange rates


– USD/AUD = 0.83 which implies that AUD/USD = 1/0.83 =1.20
– Given an exchange rate HC/FC (which is a quote for FC since it is in the denominator),
we can get the quote for the HC by inverting this exchange rate.

267
Prof. Dr. Bernd Scherer

Cross Rates

• All the quotes in the third and fourth columns of TABLE 1 are in terms of the USD.
But one may also wish to make spot transactions between other currencies
without going through the USD. The rate at which one can exchange two non-
USD currencies directly is called the cross rate.

HC

• TABLE 2: Currency cross spot rates on 11 September 2007


• The cross exchange rate for CAD/EUR is 1.4416 and the cross exchange rate for
EUR/JPY is 0.0063.

268
Prof. Dr. Bernd Scherer

Calculating Cross Rates

• To calculate the cross rate between two non-USD currencies simply note:
FC 1 FC 1 USD
= ⋅
FC 2 USD FC 2

• Example: The cross rate for CAD/EUR can be calculated from

CAD CAD USD


= ⋅ = 1.0388 ⋅ 1.388 = 1.4416
EUR USD EUR

• This is exactly what we find in TABLE 2

269
Prof. Dr. Bernd Scherer

Spot Rates With Bid Ask Spreads

• So far, we have ignored transactions costs. But when we buy an object we


typically have to pay more for it than the price we would get for selling it.
– Ask price: the price paid to buy the object (this is the seller’s asking price).
– Bid price: the price received when selling an object (this is the buyer’s bid price).
– The difference between the ask price and the bid price is called the bid-ask spread?

– The spread is the ”market-maker’s” commission for executing the trade. This
spread compensates the market-maker for taking a position that she may not
desire.
– Markets where the trading volume is small will typically have larger spreads; this is
because in these markets it is more difficult for the market-maker to get out of an
undesirable position.
– Bid ask spreads are rising when volatility is rising. Why?

270
Prof. Dr. Bernd Scherer

Spot Rates With Bid Ask Spreads (cont.)

• The spot ask exchange rate is the amount of HC one requires to buy one unit of
the FC today, and the spot bid exchange rate is the amount of HC that one
receives when selling one unit of the FC today.
• TABLE 3: FX Spot bid and ask

– If you live in the U.K. and are planning a holiday to the U.S and wish to buy USD, then
the you have to pay GBP/USD 0.4927. This is the ask rate for USD (the currency in the
denominator).
– If you have some dollars left over after your holiday, when you sell them you will
received the bid rate, GBP/USD 0.4917, which is less than what you paid to purchase
dollars.

271
Prof. Dr. Bernd Scherer

Inverting Spot Rates With Bid Ask Spreads

• The spot bid-ask quotes in TABLE 3 are in terms of FC. To get the inverse quotes,
that is the quotes in terms of USD/FC, we need to invert the bid-ask quotes. To
get the inverse bid quote we invert the FC/USD ask quote; to get the inverse ask
quote, we invert the FC/USD bid quote.

• The reason why we invert the USD ask quote to get the inverse bid quote is that,
to preclude arbitrage, the bid quote must always be smaller than the ask quote.

• Thus, to get the inverse bid quote to be the smaller number we need invert the
(larger) direct ask quote; similarly, the inverse ask quote will be larger than the
inverse bid quote only if it obtained by inverting the (smaller) direct bid quote.

272
Prof. Dr. Bernd Scherer

Example

• Table 4: Foreign exchange spot bid and ask rates as of 11 September 2007

• In Table 4, the sixth column (bid per US$) is obtained by inverting the ask quotes in
the fifth column; the eighth column is obtained by inverting the quotes in the third
column. Consequently, the bid per quotes in column 6 are smaller than the
corresponding ask quotes in column 8.

• This is easily testable in any exam!

273
Prof. Dr. Bernd Scherer

Summary

• Currencies are not traded on an organized exchange; instead trading takes place
via banks (market-makers) and brokers.

• The market for currencies is very deep—over USD 1.8 trillion is traded daily,
– with most of the trade being in the spot and forward markets, and
– a majority of this being unrelated to trade in goods and services.

• Quotes for spot exchange rate are typically against the USD (direct).
– Indirect quotes can be obtained by inverting the appropriate direct quote.
– Cross rates can be obtained by computing the effective rate if one transacted via the
USD.

274
Prof. Dr. Bernd Scherer

2. The Forward Market

• Key Questions
– Understanding forward exchange contracts.
– Relation between spot and forward exchange rates.
– Valuation of forward contracts

• Motivating Problems
1. Suppose you work for Sony (Japan) in the division that exports computer games to
Britain. You receive GBP 100 m for these exports at the end of each quarter. You are
worried about the possibility of a decrease in the JPY value of the GBP. How can you
hedge your position against such a decrease?
2. You work in the treasury office of BC Gas in Vancouver, Canada. It is 6 AM, and you
have yet to see your first cup of coffee. The Chief Financial Officer has told you, Marie,
that there is a cash shortfall, and that to buy gas you need to borrow USD 10 m, for 1
month. How do you decide where to finance this loan—in the US or the Canadian
money market.

275
Prof. Dr. Bernd Scherer

The Forward Rate - Definition

• Definition: The forward exchange rate between two currencies, is the price
agreed upon today (t) at which one currency can be exchanged for the
other currency at a certain future date (T).

Forward contract signed at t Exchange of FC for HC at


Ft ,T previously agreed price

t T

• Some Observations
– The rate at which the transaction will take place is determined today;
thus, there is no uncertainty about the price (it is not random).
– The actual transaction (settlement) takes place at a future date, T, in
contrast to a spot transaction, which is settled on the same date (t).
– There is no exchange of cashflows today.

276
Prof. Dr. Bernd Scherer

Example

• TABLE 1: Foreign exchange spot and forward rates as on 11 September 2007

– The spot and forward rate will typically not be the same.
– If the forward rate for the FC (the currency in the denominator) is larger than the spot
rate, then the FC is said to be trading at a premium otherwise, it is at a discount. From
the last column, we see that the Euro is at a premium relative to the USD.
– If the FC is at a forward premium, then the HC is at a discount, and vice versa.

277
Prof. Dr. Bernd Scherer

Example (continued)

• Forward rate for goods


– You agree today to sell your house after 3 months for USD 1 m. Thus, the 3-month
forward exchange rate for your house is USD/house 1 m. This
– implies that: no exchange takes place today; After 3 months you will receive USD 1 m
and will transfer ownership of your house.

• Forward rate for currencies


– Suppose you are a resident of the U.S. You are planning a vacation to Japan in 6 months
time. Your budget for the holiday is JPY 1 m. To buy this amount forward, the amount
of USD you need in 6 months time is given by the current USD/JPY 6-month forward
rate in TABLE 1:

JPY [ 1m ] ⋅ USD / JPY [ 0.00895 ] = USD [ 0.00895m ]

278
Prof. Dr. Bernd Scherer

Hedging FC Cash-Flows

• FC inflows at a future date can be hedged by selling this FC forward.


• FC outflows at a future date can be hedged by buying FC forward.
• Hedging allows us to translate
– a future outflow of FC, whose HC value would depend on the future random spot rate
– into an outflow of HC whose value is the current forward rate which is known today

279
Prof. Dr. Bernd Scherer

Example

• Suppose that you live in the US and intend to vacation in London (UK) twelve
months from today. You expect to spend GBP 1 M (monetary unit ☺ ) on your
vacation. You are worried, however, that if there is an increase in the value of the
GBP, you will either have to cut back on what you spend, or you will need
additional USD to keep the same expenditure in GBP. How can you use forward
contracts to hedge your position?

• Your position is short forward GBP. You can hedge this by going long forward GBP
(buying GBP forward). From the data below, we see that the price for buying 12-
month forward GBP would be: USD/GBP = 2.0080.

• What is the cost of hedging: spread, bid ask or both?

280
Prof. Dr. Bernd Scherer

Covered Interest Rate Parity (CIP)

• Consider the following strategy

Invest in HC

( 1 + rt,T )
1 1 ⋅ ( 1 + rt ,T )

Ft,T
Exchange in FC Change into HC

( 1 + rt*,T )
1
st 1
St ⋅ ( 1 + rt*,T )

Invest in FC

281
Prof. Dr. Bernd Scherer

CIP (cont.)

• In frictionless markets and in the absence of arbitrage we have covered Interest


Rate Parity (CIP):

( 1 + rt,T )
Ft,T = St ⋅
( 1 + rt*,T )

– Hedging transforms foreign cash into home cash (taxman might view capital gains from
a forward position different than interest rates)
– This is a pure arbitrage argument. It does not imply that the forward rate is
determined by the spot rate.
– Holds extremely close in empirical data (do not confuse with UIP! See later)

• Memory tip

EUR ( 1 + rtEUR
,T )
Ft,T = St USD

( 1 + rtUSD
,T )

282
Prof. Dr. Bernd Scherer

What are the Corporate Finance Implications?

• From the above, in frictionless markets (no transactions costs, no taxes etc.): Can
you save money by borrowing in one currency rather than another?

• No, according to CIP, once you hedge exchange rate risk, borrowing cost will be
the same across currencies.

1
S
( 1 + rt*,T ) Ft,T = (
1 + rt,T )
t
  borrowing at hom e
borrowing abroad

• Of course, this is true only in frictionless markets


– Example: Credit risk is assessed differently in different markets

• If you are not allowed to use forward contract you can replicate them via the
local money markets

283
Prof. Dr. Bernd Scherer

What is the Value of a Forward Contract after Initialization?

• Long a forward contract is equivalent to


– a long position in a foreign zero bond and
– a short position in a domestic zero bond

1 1
Vt +1 = St +1 ⋅ − Ft ,T ⋅
1 + rt*+1,T 1 + rt*+1,T

• Alternatively we can rewrite this


1 + rt*+1,T
St +1 = Ft +1,T
1 + rt +1,T
 1 + rt*+1,T  1 1
Vt +1 
=  Ft +1,T  ⋅ − Ft ,T ⋅
 1 + rt +1,T  1 + rt +1,T
* 1 + rt +1,T
1 1
= Ft +1,T ⋅ − Ft ,T ⋅
1 + rt +1,T 1 + rt +1,T
1
= ∆F
1 + rt +1,T

284
Prof. Dr. Bernd Scherer

Summary

• Forward currency is traded just like spot currency—not on an organized exchange


but via banks and brokers.

• Forward contracts can be used for hedging FC-denominated Cash-Flows.

• The value of a forward contract is the discounted value of the difference


between the current forward rate and the rate at which the forward contract
was initiated.

• The value of a just-issued forward contract is zero. After this it moves with the
discounted P&L.

• Covered Interest Parity implies that one cannot profit by borrowing in one
currency and lending in another. With frictions, one can always save money: it
will always be cheaper to borrow in one currency rather than another, even
after hedging against exchange rate risk.

285
Prof. Dr. Bernd Scherer

3. Exchange Rate Behavior

• Some Questions we will answer

1. In theory, should one be able to forecast the future spot exchange


rate (say one-month or three-months in the future) given the
current and historical values of the spot exchange rate? In
practice, do past exchange rates help us to forecast future
exchange rates?
2. In theory, should the forward exchange rate today for a contract
maturing in 90 days (or 360 days) tell us what the future value of
the spot exchange rate will be in 90 days (or 360 days)? In practice,
does the forward exchange rate today tell us what the future spot
exchange rate is going to be?
3. How should one identify the present value in home-currency terms
of a foreign-currency cash flow that is to be received one year
from now?

286
Prof. Dr. Bernd Scherer

Main Issues

Question Answer

1. Do prices change to offset exactly changes in the nominal exchange rate; No


that is, is the real exchange rate constant?

2. Can one forecast accurately the future spot exchange rate using past No
values of the exchange rate (weak form of predictability).

3. Is it possible to forecast accurately the future spot exchange rate using No


the current forward rate (Unbiased Expectations Hypothesis; semi-strong
form of predictability).
4. Can one forecast the exchange rate accurately using other variables No
suggested by macroeconomic theories of exchange rate determination
(semi-strong form of predictability).
5. Can one forecast the future spot exchange rate using private information Maybe
(strong form of predictability).

287
Prof. Dr. Bernd Scherer

Motivating Problems

• Real exchange rate risk


– You work in the treasury office of International Harvester, a manufacturer of heavy
equipment. Its main competitor is a Japanese firm, Komatsu, which has only a small
market share. When you talk to your boss about hedging against changes in the
USD/JPY exchange rate, your boss replies that, “Hedging is not important because in
the long run changes in the exchange rate are offset by changes in relative prices; thus,
it all averages out in the long run.”? Do you agree with the opinion of your boss?

• Forecasting Exchange Rates


– It is your first day of work at Fletcher Challenge, a forestry firm in Canada. As you walk
into the office, your boss introduces you to the other members of the Fletcher
Challenge treasury team. You are beaming brightly in your new business attire,
dreaming about how you are going to spend your signing bonus.
– Two minutes later, you hear a voice say that, ”Our new recruit will now tell us what the
value of CAD/NZD will be a year from now. We will use this estimate for capital
budgeting decisions.” Your boss looks at you expectantly. You are seeking help (and
cover), but the only thing you can hide yourself under is today’s Financial Times. What
is your response?

288
Prof. Dr. Bernd Scherer

Real Exchange Rate Risk

• The real exchange rate is defined as the nominal exchange rate, deflated by the
domestic and foreign price level.

Pt*  HC FC 
et = St  
Pt  FC HC 

• The real exchange rate is an index number. It has no dimension


• If commodity prices offset perfectly the change in the exchange rate, then the
real exchange rate will be equal to 1
• If below (above) one a currency is said to be under (over) valued.
• Main issue: do price changes (at home and abroad) and exchange rate changes
offset each other.
– If yes, then do not have to worry about fluctuations in the nominal exchange rate, for
they will be offset by changes in prices implying that the real effect is zero.
– If not, then changes in the exchange rate will have real effects—exchange rate
fluctuations will change the relative prices of goods at home and abroad; that is, there
is real exchange rate risk.

289
Prof. Dr. Bernd Scherer

Example: Changes in exchange rates and prices

• Suppose that the current price of gold is USD 330 and the Pound exchange rate is
USD/GBP is 1.65. Then, to preclude arbitrage opportunities the gold price in Sterling
must be:
[USD ] 330
= [GBP ] 200
[ GBP ]
USD 1.65

• That is, the price of the gold in the US must be the same as Britain, after translation
into USD.

• Another way of stating this is that the relative price of gold in the US and in Britain
must equal the exchange rate.

290
Prof. Dr. Bernd Scherer

Commodity Price Parity

• In frictionless commodity markets, in the absence of arbitrage opportunities, the


price of commodity k at time t in one country, must equal the price of the same
commodity in another country, expressed in terms of the same currency:

Pkt = St ⋅ Pkt*

• Clearly, if CPP holds for all goods produced in an economy then we need not care
about fluctuations in nominal exchange rates.

• Question: Should we expect CPP to hold for all goods?

• Answer: While CPP may hold for easily traded and homogenous commodities such as
gold and silver, it is unlikely to hold for other commodities such as cars or houses
which are difficult to trade across countries.

291
Prof. Dr. Bernd Scherer

Commodity Price Parity (cont.)

• Reasons why CPP may be violated


– Transactions costs;
– Tariffs and non-tariff barriers to trade, which constrain arbitrage;
– Imperfect competition arising from exclusive dealership contracts, costs for entering new
markets, etc.

• Example (CPP violations)


– Would you expect CPP to hold for the Big Mac hamburger? See The Economist magazine
for comparison of Big Mac prices in different countries.
– Would you expect CPP to hold for The Economist magazine itself? Compare the price of The
Economist magazine in different countries (given on the front cover) with the exchange
rates (published on the last few pages of The Economist).

292
Prof. Dr. Bernd Scherer

Absolute Purchasing Power Parity

• Question
– Even if CPP does not hold, is the general price level in one country related to the price
level in another country, after being translated into a common currency? That is, is one
country more expensive than another?

• Definition
– The absolute version of the Purchasing Power Parity (PPP) hypothesis states that the
price of a representative basket of goods at time t in a particular country, should equal
the price of a representative basket of goods in another country, translated into a
common currency:

Pt = St ⋅ Pt*
• Empirical evidence
– shows that there are large violations of Absolute PPP because the prices of identical
goods in the two baskets may vary across countries (violations of CPP); The
composition of the representative baskets might be different across countries.

293
Prof. Dr. Bernd Scherer

Long Run Evidence on Purchasing Power Parity

• Neither absolute nor relative


PPP appear to hold closely in
the short run.

• Both appear to hold well as a


long-run average and when
there are large movements in
relative prices.

• Both appear to hold better


between producer price than
between consumer price
indices.

294
Prof. Dr. Bernd Scherer

Empirical evidence on PPP

• Tests of relative PPP seem to do better when inflation is high (when economic
agents are more responsive to price changes). But even in these studies, there
are substantial period-by-period deviations from relative PPP.

• Relative PPP is also more likely to hold in countries where wages are indexed (for
example, Israel in the 1970s).

• Conclusion:
– Relative PPP may have some power in the long run (several years rather than several
months), and in times of high inflation,
– but in the short run there are large deviations that tend to persist for aslong as 3-6
years.

295
Prof. Dr. Bernd Scherer

Big Mac Index (4th of Janary 2010)

• Big Mac reflects a basket of


good ingredients
• Distorted by costs for local
(nontraded)) assets like rental
or labour costs.
• In lower income countries
non-traded goods tend to be
cheaper due to lower wages.
• Brand value of MCDonalds the
same in US and China?
• Long run concept and in the
long run prices could simply
take the burden of adjstment
(in particular for a fixed
exchange rate)

296
Prof. Dr. Bernd Scherer

Corporate Finance - Exchange Rate Pass Through

• Will Porsche pass through exchange rate variations?

 USD 
 = PPorsche
USD EUR
PPorsche [USD ] St 
 EUR 
• If Euro appreciates by 10% (exchange rate pass through) versus USD will Porsche
be 10% more expensive?
– Accept smaller profit margins?
– Cost reduction as imports become less expensive?
– Price elasticity of demand < 1?
– Cost reactions and efficiency gains

• Less than 100% exchange rate pass through exposes a firm to economic currency
exposure (risk)

297
Prof. Dr. Bernd Scherer

Managing Risk from Imperfect Exchange Rate Pass Through

• Local Production
• Flexible Sourcing
• Diversification across markets (currencies)
• Lowering price elasticity (product differentiation, create monopolistic positions)
• Hedging with financial instruments

298
Prof. Dr. Bernd Scherer

Relative Purchasing Power Parity

• Question
– Even if the level of prices is not the same across two countries, are changes in the price
level in one country related to changes in the exchange-rate adjusted price level in
another country?

• Definition
– According to the relative PPP hypothesis, changes in the price of a representative
basket of goods at time t in a particular country, should equal the changes in the
exchange-rate adjusted price of a representative basket of goods in another country:

Pt +1 St +1 Pt*+1
= ⋅ *
Pt St Pt

299
Prof. Dr. Bernd Scherer

Relative Purchasing Power Parity - Restatements

• Take logs to arrive at simple differences

 Pt +1   St +1 Pt*+1   St +1   Pt*+1 
ln   = ln  ⋅  = ln   + ln  
 Pt   St * 
Pt    S   * 
 Pt 
t

S 
ln  t +1  = πt +1 − π *
+1
 St 

 t
percentage change percentage change
percentage change in domestic prices in foreign prices
in exchange rate

300
Prof. Dr. Bernd Scherer

Bottom Line – Real Exchange Rate Risk

• There are large, frequent, and persistent deviations from PPP, implying that there
is real exchange rate risk;
– that is, changes in the nominal spot rate are not completely offset by changes in prices
at home and abroad.

• Hence, changes in the nominal spot rate will have an impact on the competitive
position of firms.

• Also, changes in the nominal spot rate will have an impact on the returns realized
from investing in foreign-currency denominated securities.

• Thus, we cannot ignore fluctuations in the nominal exchange rate

301
Prof. Dr. Bernd Scherer

Forecasting exchange rates

• Is it possible to forecast future value of nominal exchange rate?


• Weak form tests - those that forecast the future exchange rate on the basis of
the past exchange rate; these are tests of technical models.
• Semi-strong form tests - those that predict the exchange rate on the basis of
other available information, in addition to realized exchange rates. This is called
fundamental analysis.
– The fundamental variables may include the inflation rate, money supply, industrial
production and the forward exchange rate.
– The forecasts based on these fundamentals may be made using an econometric model,
or on a judgmental basis.
• Strong form tests - forecasts made by
– professional forecasting agencies who may be better than individuals at interpreting
information, and
– central banks, who may have more information than that available to the typical
investor.

302
Prof. Dr. Bernd Scherer

Autocorrelation Models

• Autocorrelation models are similar to regressions where one regresses the left-
hand side variable on its lagged values.

S   S 
ln  t +1  = a + b ln  t  + εt +1
 St   St −1 

• Then, one examines the autocorrelation coefficients, b, for significance.


• Autocorrelation tests on the exchange market generally find
– small, and typically significantly positive autocorrelations;
– autocorrelation coefficients are frequently larger than for common stocks.
• Conclusion: For floating exchange rates the information content of past exchange
rates is low: the R2 statistics from such tests rarely exceed 5%, so the
predictability of the change in the exchange rate, based on past changes, is not
economically significant

303
Prof. Dr. Bernd Scherer

Filter Rules

• If increases tend to be followed by increases, then a strategy of buying after


observing an x% rise from a “low” would, on average, generate profits;
• if price-drops tend to be followed by more down-ticks, then selling after
observing an x% fall from a “high” would again pay, on average.
• The percentage x used is called the size of the filter. The filter is meant to detect
significant changes as opposed to meaningless changes generated by temporary
fluctuations in demand and supply.

• Conclusion: While there is some persistence in exchange rates, of the more than
580 filters tested
– less than 10% of the rules produced profits that are significant at the 10% level before
transaction costs; and,
– only 0.3% of the rules were profitable after transaction costs.

304
Prof. Dr. Bernd Scherer

Forecasting the Exchange Rate Using the Forward Exchange Rate

• The conjecture that the future spot rate is equal to the current forward rate is
called the Unbiased Expectations Hypothesis (UEH)

E ( Sɶt +1 | Ωt ) = Ft ,t +1

• This hypothesis would also be true if


– all investors were risk neutral, or
– if all exchange risk was completely diversifiable (so that investors were indifferent
between the certain forward rate and the random future spot rate)

• Question: Is UEH true in the real world?

305
Prof. Dr. Bernd Scherer

Regression Tests of the UEH

• Typical regression test of UIP

 S t +1 − S t   Ft ,t +1 − St 
 = + 
 St  a b  St  + εt +1

• Conclusion: Over 75 empirical studies test this regression, and reject it.
– The average value of b in these studies is -0.88. What regression coefficient would you
expect. What does this tell you?
– The R2 value is also very small, so there is little evidence of predictability in forward
exchange rates.
– What do theses results tell you about the likely success of carry strategies? How would
you optimally design those funds?

306
Prof. Dr. Bernd Scherer

Regression Tests of the UEH

• There are several possible interpretations of the result that b < 1.


1. Investors are not risk neutral and that the bias in the forward rate’s prediction of the
spot rate reflects a risk premium.
2. Investors make errors in forming expectations.
3. The transactions costs may be sufficiently large to obscure the link between the
forward rate and the future spot exchange rate.

• None of these explanations get much support in the data—and this is an ongoing
area of investigation.
– Recent work finds that sign of slope coefficient depends on whether USD at
discount/premium
– slope coefficient is not negative in emerging countries.

307
Prof. Dr. Bernd Scherer

Two More Parities

• Fisher Closed: Nominal rates should react one to one to an increase in expected
inflation to protect the real rate investors require (no money illusion)

rt +1 = ρ + E ( πt ,t +1 )

• Fisher Open: The real rate of return is the same across the world. If not
investments continue. In other words, the differential in nominal interest equals
the differential in inflation rates

rt +1 − rt*+1 = E ( πt ,t +1 ) − E ( πt*+1 )

308
Prof. Dr. Bernd Scherer

Bringing it All Together

• Example: Nominal interest rates in Europe are 1% higher than in the US. What
does this imply for inflation differential, forward contracts and expected
exchange rate movements?

St +1 −St
UEH St PPP

Ft ,t +1 −St
E ( πt ,t +1 ) − E ( πt*+1 )
St

CIP rt +1 − rt*+1 Fisher open

309
Prof. Dr. Bernd Scherer

FX Forecasting with Macroeconomic Variables

• Right-hand side includes many of the fundamental variables proposed by models


of the exchange rate.
– Interest rate differential
– Inflation differential
– Monetary growth differential
– GDP growth differential

 St + 1 − St 
  = a + b ( rt,t +1 − rt ,t +1 ) + c ( πt ,+1 − πt ,t +1 )
* *
 St
+ d ( mt ,+1 − mt*,t +1 ) + e ( gt ,+1 − gt*,t +1 ) + εt +1

• For the case of US versus Japan, the R2 statistic is 0.0098; and for the case of US
versus Germany, the R2 is 0.0118;
• None of the slope coefficients in either regression is significant even at the 10%
level.

310
Prof. Dr. Bernd Scherer

FX Forecasting with Macroeconomic Variables

• Correlation between exchange rates and explanatory variables is small and not
significant

311
Prof. Dr. Bernd Scherer

The Forecasting Record of Forecasting Services

• Why do we ask this question?

• Would you sell a forecast or apply it a set of assets? How do the economics
compare? What is your prior on forecasting services?

• Forecasting services using technical models to predict the future spot rate have a
superior record compared to those analyzing fundamental variables.

• However, the record of technical forecasting services with a superior record


initially appeared to deteriorate over time; that is, it was not the same services
that had the superior record over time.

312
Prof. Dr. Bernd Scherer

The Forecasting Record of Central Banks

• Central banks claim that they intervene in currency markets to maintain an


orderly market and to smooth out excessive swings in exchange rates;
– they also claim that they do not try and move the exchange rate away from its
fundamental value.
– If this is true, it means that central banks must be quite good a predicting exchange
rates.

• When the profits from intervention are measured for eight central banks, it was
found that seven central banks actually made substantial losses from currency
trading.
• Studies undertaken by the Dutch central bank and the Canadian central bank find
that intervention has been modestly profitable, though there have been long
periods during which the banks incurred substantial losses.
• Conclusion: Thus, the evidence is mixed and one cannot conclude that central
banks can predict the future spot rate, even though they have access to private
information about monetary and exchange-rate policy.

313
Prof. Dr. Bernd Scherer

Summary

• There is strong evidence of frequent, significant and persistent deviations from


PPP, indicating the presence of real exchange rate risk.

• The overall evidence suggests that it is quite difficult to predict future exchange
rates accurately.

– Predictions based on fundamental variables do not seem to be very accurate.


– Technical forecasts seem to do better, though even their record is not impressive.
– While economic theory may be useful for explaining the past, and for making broad
predictions about the long-run effects of government policy, it has only limited
successful in making precise predictions about the exchange rate in the short run

• Question: Do firm managers and investors need to worry about fluctuations in


exchange rates?
– Do fluctuations in the exchange rate influence the value of firms?
– Do changes in the exchange rate affect the payoff from investing in international
assets?

314
Prof. Dr. Bernd Scherer

4. International Portfolio Theory

• In an international environment, one is not limited to holding assets traded


on the domestic stock exchanges. One can also hold international assets.

– What are the theoretical gains from investing in international assets?


– What are the disadvantages of investing in international stocks and bonds?
– What does the data suggest: are the gains from international diversification
large or small?

315
Prof. Dr. Bernd Scherer

Return and risk from investing in foreign stock

• Return on a domestic stock

Pt +1 − Pt + dt +1 P + dt +1
rt,t +1 = = t +1t −1
Pt Pt

• Return on a foreign stock measured in home currency


( PtFC
+1 +1 ) ⋅ St +1 − Pt
+ DtFC FC
⋅ St
rt,t +1 =
PtFC ⋅ St
= ( 1 + rtFC
,t +1 ) ⋅ ( 1 + st ,t +1 ) − 1

= rtFC
,t +1 + st ,t +1 + rt ,t +1 ⋅ st ,t +1
FC

≈ rtFC
,t +1 + st ,t +1

– The approximation is good when the returns are small, and is exact for
continuously compounded returns
– Currency risk is always on TOP OF asset risk

316
Prof. Dr. Bernd Scherer

Risk of foreign stocks

• The riskiness of foreign stocks can be derived from

rt ,t + 1 = rtFC
,t + 1 + ( 1 − h ) ⋅ st ,t + 1

var ( rt,t + 1 ) = var ( rtFC


,t + 1 ) + ( 1 − h ) var ( st ,t + 1 ) + 2 ( 1 − h ) cov ( rt ,t + 1 , st ,t + 1 )
2 FC

• Will currency hedging always decrease risks, i.e. what is the optimal hedge ratio?

d var ( rt ,t + 1 )
= −2 ( 1 − h ) var ( st ,t + 1 ) − 2 cov ( rtFC
,t + 1 , st ,t + 1 ) = 0
dh

cov ( rt ,t + 1 , st ,t + 1 )
FC
h = 1+
var ( st ,t + 1 )

317
Prof. Dr. Bernd Scherer

Developed equity markets

• Many international equity markets have outperformed the US market in local


currency terms, and in USD terms (that is, after adjusting for exchange rates).

• But, the volatility of non-US equity markets has also been higher than that of the
US market. The variance of returns on non-US stocks in local currency terms are
only slightly lower than returns measured in USD because even though the
exchange rate has high volatility, its correlation with stock returns is quite low.

• Because the correlation among international stock markets is less than one
there are substantial gains from diversifying internationally. But, investors
typically invest only in domestic stocks and do not diversify internationally.

318
Prof. Dr. Bernd Scherer

Emerging equity markets

• Average returns in merging markets have been high and very variable relative to
returns in developed markets.

– Emerging markets have more currency risk than developed markets.


– The correlations among emerging equity markets are lower than those among
developed markets (about 0.1 instead of 0.4).
– Thus, there seem to be substantial diversification gains by adding emerging market
equities to ones portfolio.

319
Prof. Dr. Bernd Scherer

Optimal Portfolio Choice - Value of International Diversification

• MSCI Country Performance in UK Sterling

320
Prof. Dr. Bernd Scherer

Optimal Portfolio Choice - Value of International Diversification

• Correlation between equity indices

321
Prof. Dr. Bernd Scherer

Optimal Portfolio Choice - Value of International Diversification

• Domestic Diversification in the UK

322
Prof. Dr. Bernd Scherer

Optimal Portfolio Choice - Value of International Diversification

• Global Diversification

323
Prof. Dr. Bernd Scherer

Optimal Portfolio Choice - Value of International Diversification

• Global Diversification plus Emerging Markets

324
Prof. Dr. Bernd Scherer

Summary

• The data seems to suggest that there are large gains from diversifying one’s
portfolio internationally.

• However note that observed portfolios are heavily weighted toward domestic
assets. It is a puzzle why investors do not diversify internationally – If its so great
why don’t investors do it?

• Designing optimal portfolios with international assets is very similar to choosing


portfolios of only domestic assets; the only difference is that now one needs to
translate returns on foreign assets into HC terms.

325
Prof. Dr. Bernd Scherer

Session 9: Variance Products

1. Volatility Returns

2. Volatilities as Asset Class – Portfolio Context of Variance Products

326
Prof. Dr. Bernd Scherer

Literature

• NEFTCI (2008), Principles of Financial Engineering, 2nd edition,


Academic Press

• ROUAH/WEINBERG (2007), Option Pricing Models and Volatility,


Wiley

• HAFNER/WALLMEIER (2005), An Investors Perspective on Volatility


as an Asset Class, SSRN

327
Prof. Dr. Bernd Scherer

1. Volatility (Positions) Returns

• Delta Hedged Option Positions


• Straddle Returns
• Variance Swaps

328
Prof. Dr. Bernd Scherer

Delta Hedged Option Positions

• Example: Delta hedged long call (20% volatility), i.e. long call with delta
equivalent short position in the underlying

• If implied volatility jumps (to 30%), is the holder of the above strategy
guaranteed to make a profit?

• Note a delta hedged position leaves the investors still with three exposures
– Theta: Time decay, i.e. option sensitivity to the passage of time
– Gamma: Sensitivity of the option delta to changes in underlying
– Vega: Option sensitivity to changes in implied (expected future)
volatility

• In the next few slides we will show this is a grossly imperfect volatility
position

329
Prof. Dr. Bernd Scherer

Delta Hedging and Vega Exposure

• Example: 100 Calls with 0.1 years maturity, Strike 100 and implied volatility
of 20%
100 ⋅ C = 100 ⋅ 2.52 = 252
delta = 100 ⋅ dC
dS = 51.2
2
gamma = 100 ⋅ ddSC2 = 6.3
vega = 100 ⋅ ddC
σ2
= 3152

• What is the expected change in option value if variance jumps?

100dC ≅ 3152 ( - 0.04 ) = 157.6


0.09
change in
variance

330
Prof. Dr. Bernd Scherer

Vega Exposure and P&L

• Vega is a function of the underlying stock price (deep ITM and OTM) “cease” to
have option value and hence have almost no vega.
35
100 ⋅ 31.52 ⋅ ( 0.09 - 0.04 ) = 157.6
30

25

100 ⋅ 0.056 ⋅ ( 0.09 - 0.04 ) = 0.28


20
Variance Vega

15

10

0
70.00 80.00 90.00 100.00 110.00 120.00 130.00 140.00
-5
Strike/Underlying
331
Prof. Dr. Bernd Scherer

Delta Hedging and Gamma Exposure

• Sell 110% out of the money call at 30% implied volatility. Realized volatility is 27%
(prima facie good) but rises towards the end when option was at the money (and
hence had large gamma). Position lost money. Why?

332
Prof. Dr. Bernd Scherer

Long Straddle (Static Volatility Position)

500%
• A dynamically hedged position
might loose its exposure to
400%
volatility and hence works no
longer as a hedge. How will a
300% static long straddle (long call
Straddle Return

and put at the same strike) do?


200%

• Extend previous example with


100%
long call and put (strike 100),
0.1 days to maturity.
0%
75 80 85 90 95 100 105 110 115 120 125

-100%
Spot Price (S) at Maturity T

333
Prof. Dr. Bernd Scherer

Long Straddle (Static Volatility Position)

• Expected payoff is twice as large! Why?

100dC ≅ 2 ⋅ 3152 (


0.09 - 0.04 ) = 315.2
change in
variance
• All other problems remain, i.e. vega becomes very small if the underlying
moves away from the strike. Again long straddle does not isolate volatility
as source of return.

• What does change?


a) If increased volatility results in large realized movements (underlying
ends at 70 or 130) the long straddle holder will benefit
b) Conversely even if volatility rises a long straddle might pay nothing if
the stock finishes after violent moves at 100

334
Prof. Dr. Bernd Scherer

Trading Pure Volatility – Vega Neutral Positions

• We want to trade pure volatility.

• How can we construct a portfolio whose Vega is insensitive to the price of


the underlying?

• Intuition: Vega of out of the money puts became increasingly smaller, i.e.
we need to increase the weight on OTM Calls and Puts to avoid Vega decay.
But by how much?

• Answer: Build a portfolio of weighted calls and puts

∑ K >K * ( K1 )C
2
OTM
(K ) + ∑
K <K * ( K1 )P
2
OTM
(K )

CALL PUT

335
Prof. Dr. Bernd Scherer

Example

• 13 Options with strikes between 70 and 130, 0.0833 years to maturity, 20% volatility
and 5% risk free rate.

16

12
Vega

0
70 80 90 100 110 120 130
Spot Price (S )

336
Prof. Dr. Bernd Scherer

Example (continued)

• Underlying EXCEL sheet


Strike (K ) 70.00 75.00 80.00 85.00 90.00 95.00 100.00 105.00 110.00 115.00 120.00 125.00 130.00
1/K² 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00
1/K² Weight 0.14 0.12 0.11 0.10 0.09 0.08 0.07 0.06 0.06 0.05 0.05 0.04 0.04

Asset Price Option Vegas


70.00 8.02 4.43 0.70 0.04 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00
72.50 6.49 7.42 2.31 0.25 0.01 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00
75.00 3.73 8.59 5.15 1.02 0.08 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00
77.50 1.57 7.14 8.07 2.90 0.40 0.03 0.00 0.00 0.00 0.00 0.00 0.00 0.00
80.00 0.50 4.38 9.17 5.87 1.41 0.15 0.01 0.00 0.00 0.00 0.00 0.00 0.00
82.50 0.12 2.05 7.77 8.71 3.54 0.61 0.05 0.00 0.00 0.00 0.00 0.00 0.00
85.00 0.02 0.75 5.05 9.74 6.59 1.85 0.25 0.02 0.00 0.00 0.00 0.00 0.00
87.50 0.00 0.22 2.57 8.40 9.35 4.20 0.87 0.09 0.01 0.00 0.00 0.00 0.00
90.00 0.00 0.05 1.05 5.72 10.31 7.31 2.35 0.38 0.03 0.00 0.00 0.00 0.00
92.50 0.00 0.01 0.35 3.13 9.03 9.98 4.88 1.19 0.16 0.01 0.00 0.00 0.00
95.00 0.00 0.00 0.10 1.40 6.39 10.88 8.02 2.89 0.56 0.06 0.00 0.00 0.00
97.50 0.00 0.00 0.02 0.52 3.71 9.65 10.61 5.58 1.56 0.25 0.02 0.00 0.00
100.00 0.00 0.00 0.00 0.16 1.80 7.06 11.46 8.73 3.47 0.78 0.11 0.01 0.00
102.50 0.00 0.00 0.00 0.04 0.74 4.32 10.26 11.23 6.29 1.97 0.37 0.05 0.00
105.00 0.00 0.00 0.00 0.01 0.26 2.25 7.73 12.03 9.43 4.08 1.05 0.17 0.02
107.50 0.00 0.00 0.00 0.00 0.08 1.00 4.95 10.88 11.85 7.01 2.43 0.53 0.08
110.00 0.00 0.00 0.00 0.00 0.02 0.39 2.73 8.40 12.60 10.13 4.71 1.36 0.26
112.50 0.00 0.00 0.00 0.00 0.00 0.13 1.31 5.59 11.49 12.46 7.73 2.93 0.72
115.00 0.00 0.00 0.00 0.00 0.00 0.04 0.55 3.25 9.06 13.18 10.82 5.37 1.71
117.50 0.00 0.00 0.00 0.00 0.00 0.01 0.21 1.66 6.25 12.10 13.07 8.45 3.47
120.00 0.00 0.00 0.00 0.00 0.00 0.00 0.07 0.75 3.79 9.73 13.75 11.50 6.05
122.50 0.00 0.00 0.00 0.00 0.00 0.00 0.02 0.30 2.05 6.91 12.70 13.68 9.17
125.00 0.00 0.00 0.00 0.00 0.00 0.00 0.01 0.11 0.99 4.36 10.39 14.32 12.18
127.50 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.04 0.43 2.47 7.57 13.31 14.28
130.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.01 0.17 1.26 4.95 11.04 14.90

337
Prof. Dr. Bernd Scherer

Example (continued)

• The optimal weighting scheme achieves a constant Vega

0.16

2.5
0.14

2.3 0.12

0.10
2.1

Weight
Vega

0.08

1.9
0.06

1.7 0.04

0.02
1.5
70 90 110 130
Asset Price 0.00
70.00 90.00 110.00 130.00
Stri ke

Weak replication 338


Prof. Dr. Bernd Scherer

Variance Swap - Definition

• A variance swap is a forward contract on realized variance. At maturity it pays the


difference between realized variance and a predetermined variance strike.

• Variance swaps can be replicated with a portfolio of vanilla options.

• However replication is not perfect as option as the range of available strikes is


limited (variance swaps are hence often capped)

• Variance Swap Strike (rule of thumb): 90% implied put volatility

• Price of a variance swap depends on the price of out of the money puts. If smile is
steep, variance swaps are very expensive.

339
Prof. Dr. Bernd Scherer

Who Should Use Variance Swaps?

• CPPI investor (performance suffers if volatility rises)

• Investor in quantitative strategies (typically short volatility)

• Convertible arbitrage funds

340
Prof. Dr. Bernd Scherer

2. Variance as Asset Class

• Properties of variance swap returns

• Variance swaps in a portfolio context

341
Prof. Dr. Bernd Scherer

What defines an asset class?

• Assets that are no mean variance spanned by linear combinations of


existing assets

• Low correlation is not enough (coin flipping also has zero correlation but no
risk premium)

• An asset class needs a risk premium unique to its own!

• Only return streams that are yet un-spanned extend the efficient frontier
further to the left.

342
Prof. Dr. Bernd Scherer

Variance Swap Rates versus ATM Implied Variance

• The more pronounced the smile the more expensive the variance swap
(bigger spread below). Why (relate this to the pricing of variance swaps)?

Source: HAFNER/WALLMEIER (2005)

343
Prof. Dr. Bernd Scherer

Variance Swap Returns

Significant negative (positive) risk


premium from long (short) volatility

Short volatility means providing


insurance to the capital markets.
Payoffs
Is this a reasonable position for a
private investor?

Discrete
Returns

Log
Returns

Source: HAFNER/WALLMEIER (2005)

344
Prof. Dr. Bernd Scherer

Variance Swap Returns and Market History

• Variance swaps tend to pay out in market crisis (increase in volatility)

Source: HAFNER/WALLMEIER (2005)

345
Prof. Dr. Bernd Scherer

Variance Swap Returns vs Market Returns

• Long variance swap offers protection in market downturn. Look at two


conditional OLS regressions below.

Source: HAFNER/WALLMEIER (2005)

346
Prof. Dr. Bernd Scherer

Variance Swaps and Capital Market Equilibrium

• Run CAPM regressions to look for risk adjusted out (under) performance.
Are variance swaps priced (spanned) by systematic risk factors?

Source: HAFNER/WALLMEIER (2005)

Significant risk adjusted


underperformance for
long variance swap

347
Prof. Dr. Bernd Scherer

Efficient Frontiers

Equities and cash

Equities, cash and variance


swaps

Source: HAFNER/WALLMEIER (2005)

348
Prof. Dr. Bernd Scherer

Optimal Asset Allocation

• Expected utility maximization with power utility.


• +/- 2 STD change in variance swap risk premium.

Contrary to popular advice (go long variance swapss to protect downside, the reults show that
investors should go short variance swaps (to collect a risk premium) and go short equities to hedge
the risk from sort variance swaps.

Source: HAFNER/WALLMEIER (2005)

349
Prof. Dr. Bernd Scherer

Session 10: Risk Management

1. Why Risk Management?

2. Risk Management and Capital Allocation

350
Prof. Dr. Bernd Scherer

Literature

• HULL, 2007, Financial Institutions and Risk Management, Prentice-Hall

• JORION, 2001, Value at Risk, McGraw-Hill

• PEROLD (2001), Capital Allocation in Financial Firms, HBS working paper

351
Prof. Dr. Bernd Scherer

1. Why Risk Management

Financial Operational
• Instruments: • Instruments:
– insurance contracts – standard procedures
– derivatives – safety installations
– capital structure – etc
• Aim: • Aim:
– structuring cash flows – continued operations
depending on certain and loss prevention
outcomes

Generally, NPV = 0! Possibly, NPV > 0!


i.e., derivatives are „fairly priced“ i.e., real investments may be profitable

352
Prof. Dr. Bernd Scherer

Why financial risk management?

Conventional answer:

„Risk is Bad! Risk management increases firm value!“

Public company
Entrepeneurial firm

Risk management is equivalent Do shareholders care


to the owner-manager buying about projects with NPV = 0?
insurance.

353
Prof. Dr. Bernd Scherer

Irrelevance of risk management

Balance sheet Balance sheet

Buying Assets
Capital „fairly 60 Capital
Assets
100 100 priced“ 100
derivative Hedge
40

Assumptions of perfect capital market:


Conclusion
- perfect competition
- no transactions costs
- no bankruptcy costs Compare the
- no taxes assumptions of In a perfect capital market,
- no regulation Modigliani & Miller (1958) risk management is irrelevant!
- perfect information

354
Prof. Dr. Bernd Scherer

Derivatives use by US firms

Source: Bodnar et al. (1998): Wharton/CIBC World Markets 1998 Financial Risk Management
Survey of 2000 publicly traded US firms. 399 respondents.

355
Prof. Dr. Bernd Scherer

Bankruptcy Costs (1/2)

• Direct costs
– Lawyers, management time
– Firm specific assets

• Indirect costs („costs of financial distress“)


– Difficulties to sell products
– Difficult to tap capital markets

• Reducing operative leverage increases debt capacity („tax shield“)

356
Prof. Dr. Bernd Scherer

Bankruptcy Costs (2/2)

bc

d a

Bankruptcy costs. The costs of corporate bankruptcy (direct and indirect) indicated by bc rise if the asset
value (a) falls below the value of corporate debt (d). Costs are measured at time of bankruptcy. Asset
volatility will increase the value of this option. Source: Scherer (2005)

357
Prof. Dr. Bernd Scherer

Taxes (1/2)

• The second convexity arises from corporate tax schedules. For example:
firms either pay taxes if they make profits (the more the higher profits
are) or they don’t (irrespective of the size of losses).
• Carrying losses backward and forward is typically limited. Hence there is
a natural kink of the tax function creating convexity. If earnings (e) are
100 in one year and -100 next year the tax base is on average zero (the
company left its owner with zero profits after two years), while the
company did pay 35 in taxes after year one.
• Shareholder value is created if we manage to reduce the governments
contingent claim in a company.
• In summary: a rise in earnings volatility induced by corporate leverage
will reduce the expected value of the tax shield, as this can only be used
if interest payments can be offset against profits.

358
Prof. Dr. Bernd Scherer

Taxes (2/2)

0 e

Non linear tax schedule. The value of total tax payments ( τ ) increases with earnings (e), but it does not
become negative when earnings turn negative. Instead it remains zero. This resembles a long call option
on corporate earnings. Source: Scherer (2005)

359
Prof. Dr. Bernd Scherer

Liquidity Costs (1/4)

• Liquidity (cash on hand) is valuable to firms, because raising cash in the future
might be expensive. Reasons are
– direct transaction costs (investment banking and lawyers‘ fees)
– adverse selection (Myers & Majluf, 1984)

• Adverse selection costs arise because two types of firms might raise capital
– firms that need to finance positive-NPV projects (starting the project is good
for existing investors, new investors get required rate of return)
– firms whose claims are overvalued (raising capital transfers wealth from new
investors to existing investors!)

360
Prof. Dr. Bernd Scherer

Liquidity Costs (2/4)

• Existing managers act in the interest of existing shareholders (because of


compensation contracts) and issue claims whenever they are overvalued
or a positive-NPV project emerges
• Outside investors cannot distiniguish between overvalued firms and those
with positive-NPV projects
– New investors would have, on average, a lower than required return
– Potential investors understand this, and offer to pay LESS so that the earn
their required rate of return on average

361
Prof. Dr. Bernd Scherer

Liquidity Costs (3/4)

• Raising capital is costly (even absent investment banking fees)


• Underinvestment (positive-NPV projects are not realized because
necessary capital can only be raised at additional cost)
• Liquidity preference:
– financing projects from retained earnings is least costly
– cash reserves as such are valuable
– firms with future growth opportunities (real options) hold excess cash
• Risk management preserves cash reserves
• Case: Porsche AG hedging USD exposure
• 50% of revenues in USD, 100% costs in EUR
• USD devaluation would drain cash quickly and force Porsche to raise
expensive capital

362
Prof. Dr. Bernd Scherer

Corporate hedging reduces costs of financial distress

• A firm is said to be in financial distress when its income is not sufficient to cover
its fixed expenses.
• The state of financial distress can lead to bankruptcy, which of course involves
liquidation costs and other direct costs.
• Large, uncovered exposures combined with adverse exchange rate movements
may send a firm into insolvency and bankruptcy, or may at least contribute to
such an outcome.

• Important to note: Without frictional bankruptcy costs avoiding bankruptcy is


not a corporate objective (now try to tell this a regulator ☺)
– In the absence of these costs, shareholders would just have lost control of the firm to
the bondholders, who would carry on the business as before (possibly after selling
their ownership rights to others).
– That is, in the absence of bankruptcy costs the event of insolvency would not have
affected the value of the firm as a whole.

363
Prof. Dr. Bernd Scherer

Product market and reorganization costs

• Many firms sell products for which after-sales service is needed and the firm
typically offers product warranties.
• Thus, a buyer’s decision to purchase such products will depend on her
confidence to continue to receive after-sales service.
• Such firms will sell more, and will therefore be worth more, the lower the
probability of going out of business.
• Thus hedging, by reducing the volatility of cash flows, decreases the probability
of coming uncomfortably close to bankruptcy.

364
Prof. Dr. Bernd Scherer

Labour market and wage costs

• Risk averse employees are likely to demand higher wages if their future job
prospects are very uncertain.

• In the event of bankruptcy, a forced change of job will generally entail monetary
or nonmonetary losses to the employee. Thus, the employee will want to
protect himself, ex ante, by requiring higher wages when working for a firm that
is more likely to be in financial distress.

• Note that this source of wage risk premium seems to hinge on imperfections in
the labour market. However, ultimately, the validity of this rationale for
corporate hedging can still be traced to imperfections in the market for risks. If
uncertainty of personal income were fully diversifiable or hedgeable, there
would not be a risk premia in wages.

365
Prof. Dr. Bernd Scherer

Capital market and refinancing costs

• Loan covenants can trigger repayment if the firm’s income falls below a stated
level. To the extent that refinancing is difficult or costly, it is useful for the firm
to reduce income volatility by hedging.

• Costs associated with refinancing include transaction costs, and especially the
indirect or agency costs of refinancing when a firm is in financial distress.

• Again, these costs represent a ”friction” in the capital market.

366
Prof. Dr. Bernd Scherer

Homemade hedging is not a substitute for corporate hedging

• The Miller-Modigliani proposition that hedging is irrelevant assume that


investors’ homemade financing decisions are perfect substitutes for the
corporate financing decisions the shareholders would have preferred.
• The term “perfect substitutes” should be read in a double sense:
– the home-made financial contracts the same effect, and*
– they are contracted at the same prices, as the corporate financial decisions they
replace.
• In practice, home-made financing decisions fail to meet both criteria because of:
– Imperfect information: In the real world, shareholders have far less information about
the firm’s exposure than the managers. If shareholders have very imprecise knowledge
of the firm’s exposure, home-made hedging will be far less effective than corporate
hedging.
– Costs of financial distress and agency conflicts: In reality, no individual shareholder can
buy a contract that perfectly hedges against the costs of financial distress.

367
Prof. Dr. Bernd Scherer

Hedging reduces agency costs

• Agency costs are the costs that arise from the conflict of interest between
shareholders, bondholders and the managers of the firm.
– These agency costs can affect the firm’s wage bill, its choice of investment projects,
and its borrowing costs.
• Hedging, by reducing the volatility of a firm’s cashflows, can reduce the conflict
of interests between different claimants to the firm’s cashflows and can
– increase its debt capacity, and
– reduce its cost of capital.

368
Prof. Dr. Bernd Scherer

Regulation

Existing shareholders own a call option on the assets of the firm

D ∂c
Debt Vega = >0
∂σ
Equity
Equity holders have an incentive to engage
in high risk investments

D V
Explains the severe regulation of financial institutions,
Incentive increases with leverage
such as banks and insurance, in particular
- risk management
- asset liability management
Taking risk particularly simple with - restrictions on derivative use
financial instruments

369
Prof. Dr. Bernd Scherer

Regulation

• Sometimes, derivatives allow easy implementation of


regulatory requirements
• case: swap overlay of UK insurance assets
• Mostly, financial innovations are used to bypass
regulatory requirements
• case: structured products in German insurance industry
• loans are not marked-to-market under German GAAP („HGB“), implying no
risk of write-off
• „interest rate payments“ may depend on the return of equity portfolio
• result is economic exposure to equity without balance sheet exposure,
inflating the risk capacity of German insurers
• case: ENRON
• derivatives may be „off-balance sheet“
• value of firm‘s assets are not transparent

370
Prof. Dr. Bernd Scherer

Costs of risk management

• Information costs
– risk management has potentially a multitude of objectives
– optimal strategies may be complex

• Transaction fees
– trading costs
– complex products are difficult to price: low transparency

• Governance of risk management


– derivatives are complex (case: Metallgesellschaft)
– derivatives can be used for asset substitution (case: Barings)

371
Prof. Dr. Bernd Scherer

Optimal amount of hedging

Optimum amount of
hedging
Cost
Firm value

PV(bankruptcy costs)

Extent of Extent of
hedging hedging

372
Prof. Dr. Bernd Scherer

2. Capital Allocation - The CEO Dilemma

• Annual budgeting time. The CEO talks to two business managers

• Lending manager: I can make another million if my loan book increases


by 200 millions

• Trading manager: I can also make another million if my position limit


increases by 10 million

• What should the CEO do? How can he compare across business?

• What should the CEO do at bonus time? Whose profits are more
valuable?

373
Prof. Dr. Bernd Scherer

What is Economic Capital?

• How much capital


should a firm hold • „Amount of capital a bank needs to absorb losses over a certain
in order to arrive time horizon with a certain confidence level“ (HULL, 2007, p.366)
at its optimal – Example: Suppose a company decides to target a AA rating (Question: Is this decision
default probability exogenous?). Given the one year rate of default probability for ist peers (other AA) is
0.03%, the confidence level would be 99.7%.
• Allocating risk
across business • Capital is required to cover unexpected loss. (Expected loss is
units incorporated in pricing decisions)
• Performance
evaluation • We need economic capital to cover operational, market and credit
risks. Sketch the distributions. How can we aggregate these risks.

• What is the time horizon for economic capital calculations? Time


to access the markets for equity funding in times of distress.

374
Prof. Dr. Bernd Scherer

The Diversification Benefit


• Aggregating
economic capital • The following table shows the economic capital (main diagonal for two
business units along the categories of market, operational and credit risk.
• Hybrid approach: Off diagonal entries show correlations.
Calculate
economic capital MR-1 CR1 OR-1 MR-2 CR-2 OR-2
MR-1 30 0.5 0.2 0.4 0 0
for business units CR1 0.5 70 0.2 0 0.6 0
using non-normal OR-1 0.2 0.2 30 0 0 0
MR-2 0.4 0 0 40 0.5 0.2
distributions, EVT,
CR-2 0 0.6 0 0.5 80 0.2
… & aggregate OR-2 0 0 0 0.2 0.2 90
the marginal
distributions
• The diversification benefit is the difference between the sum of individual
using correlations
risk capital (at correlation of one) and the aggregated risk capital using
above correlations.

∑ i =1 ∑ j =1 Ei E j ∑ i =1 ∑ j =1 Ei E j ρij = 287.9 − 203.2


n n n n
DB = −

Economic
Capital

375
Prof. Dr. Bernd Scherer

Allocation of the Diversification Benefit

• This should • From the EULER Theorem we know:


be familiar to
you from dEC
∑ i =1 dI i
n
previous EC = Ii
lectures! dEC
dI i = m arg inal risk capital
Ii dEC
dI i = incremental risk capital
I i = amount invested in unit i

• We should hence allocate capital according to incremental risk!


dEC
∑ i =1 EC i
n
EC i = I ⇒ = EC
dI i i

• How can we calculate incremental risk numerically instead of


analytically from a large scenario matrix? Some call this non-normal
risk budgeting.

376
Prof. Dr. Bernd Scherer

RAROC

• RAROC is a version of risk adjusted performance measurement. It


comes in many different definitions (be careful in practice what your
counterpart means). The most often used definition is:

Re venues - cos ts - exp ected losses


RAROC i =
EC i

• Usually derived from accaunting information and marked to model


• Can you game this RAROC measure? How? What should be changed?
• Does RAROC create a procyclical behaviour?

377
Prof. Dr. Bernd Scherer

The PEROLD Model


• Read the A financial firm issues claims (for example options) at a spread µ over fair market
original value (i.e. it makes economic profits!)
paper! Its
worth it! The issued claim is hedged. However, hedging comes with an error that can not be
reduced any further. We assume the hedging error follows

dh = σ ⋅ dz

In order to manage the firm risks that arise from the hedging error, the firm puts a
risk buffer in form of cash, C , aside. The buffer earns the riskless rate on cash r and
the firms assets, S , after one period are given by

S = C ⋅ (1 + r ) + h

The company is bankrupt if Sɶ < 0 . In this case a shortfall of S − = max [ −S , 0 ] needs


to be covered. We also assume the firm insures against this loss. While the fair
insurance premium would amount to E ( S − ) , i.e. the cost of an option against
default, the insurance provider (for example the bondholder) requires some
monitoring costs (to avoid hidden action in the form of pumping up the risks), m .

378
Prof. Dr. Bernd Scherer

The PEROLD Model

As such insurance costs amount to

(1 + m ) E ( S − )

While we can get these insurance costs down to zero if we hold large amounts of
cash, cash itself comes with agency costs (management waste) diluting the
shareholder claim S − = max [ −S , 0 ] to

(1 − d ) E ( S + )

The present value of the firm is equal to the firms operating profit less the value of
deadweight costs (these costs depend on volatility with no beta exposure, i.e. it is
not only systematic risks that matter!!!)

firm − value = µ − [ d ⋅ E ( S + ) + m ⋅ E ( S + ) ]

deadweight costs

379
Prof. Dr. Bernd Scherer

The PEROLD model

We can calculate the deadweight costs by taking their expectation with respect to
the risk neutral distribution.

E (S + ) =
σ
1+r ((
n
C ⋅ (1 + r )
σ )
+
C ⋅ (1 + r )
σ
N (
C ⋅ (1 + r )
σ ))
E (S + ) =
σ
1+r ((
n
C ⋅ (1 + r )
σ )

C ⋅ (1 + r )
σ (
N −
C ⋅ (1 + r )
σ ))
where n ( ) and N ( ) are the standard normal and cumulative standard normal
distributions.

380
Prof. Dr. Bernd Scherer

Maximizing Firm Value

Cash will decrease monitoring costs but increase agency costs. As these two effects
offset we need to find the cash allocation where

∂E ( S + ) ∂E ( S − )
d =m
∂C ∂C

Define the probability of default as q = F ( −C ( 1 + r ) ) . Now we can write

∂E ( S + ) ∂E ( S − )
d −m =0
∂C ∂C
1 − q ) (
d ( 1 + r ) − mq ( 1 + r ) = 0
prob of pay off from
pay off more cash

d
q =
m +d

381
Prof. Dr. Bernd Scherer

Maximizing Firm Value

Now we know q , we can calculate the optimal cash level:

q = F ( −C * ( 1 + r ) )
d
d +m
=N ( −C * ( 1+r )
σ ) | transform into standard normal

1−
d
d +m
=N ( C * ( 1+ r )
σ ) | use symmetry
C * ( 1 +r )
σ = N −1 ( mm+d )
C * = N −1 ( mm+d ) σ 1+1 r

Firms will hold a low cash portion if the agency costs of cash are high relative to the
monitoring costs from insurance and the other way round. If hedging were
complete, σ = 0 , no cash would be needed.

382
Prof. Dr. Bernd Scherer

Maximizing Firm Value

If we substitute the term for C * into the expression for firm − value we get the
following

firm − value = µ − kR
R= σ
2 π ( 1+r )
2
k = (d + m )e −2[ N ( d +m ) ]
1 −1 m

How do we interpret this?

R is the value of a put option to insure against hedging error. In other words: The
risk capital needed equals the price to insure against a loss. How can we see this?
We simply calculate

0 − 1 ( h )2
E ( max [ −h, 0 ] ) = ∫ -∞ h⋅ 1
1
σ( 2 π )2
e 2 σ2 ⋅ dh = σ

The factor k represents the deadweight costs of this risk capital.

383
Prof. Dr. Bernd Scherer

Capital Allocation
Capital allocation should aim at maximizing the net present value of the financial
firm
µ − kR

For m, d constant we can evaluate any single project by evaluating marginal


economic profits versus marginal risks:

∆µ − k ∆R > 0 or
∆µ
∆R
>k
Steps:

1. Calculate incremental profits ∆µ . Note: We take about economic profits,


i.e. after risk adjusted capital costs.
2. Subtract the incremental deadweight costs, i.e. the costs of increasing
risk. Note that in a MM framework theses costs would be zero as they (in
our example) only represent noise. Note: As long as the project is small
∆R = βR , where β is the regression coefficient of project versus firm
revenues. Projects with negative correlation might be attractive even if
∆µ < 0

384
Prof. Dr. Bernd Scherer

Capital Allocation

∆µ
The expression ∆R
> k looks similar to RAROC. However the differences are:

1. ∆µ is the incremental economic profit not just revenues

2. ∆R is the marginal not the stand alone risk capital

Note all this analysis is incremental, i.e. for small projects only!

385
Prof. Dr. Bernd Scherer

Optimal Cash Allocation


m = 100%, d = 10%, σ = 150, r = 10%

70

60

50
Deadweight Costs

40

30

20

10

0
0

0
0
40

80
12

16

20

24

28

32

36

40

44

48

52

56

60
Cash Cushion

AGENCY COSTS INSURANCE COSTS


Total Deadweight Costs

386