m3f33chII

Chapter II: ECONOMIC AND FINANCIAL BACKGROUND

1. Time Value of Money; discounting

Recall first the definition of simple interest, at 100x% p.a. With one unit
(a pound, say) invested for one year, at the end of the year we have:
1 + x with interest calculated yearly;
(1 + x2 )2 with interest calculated twice-yearly;
(1 + nx )n with interest calculated (or compounded) n times per year (12 for
monthly, 365 for daily, etc.).
We would like the compounding to be done as often as possible (so as to
be able to exploit the essence of compound interest – getting interest on the
interest). To do this, we use
x n
(1 + ) ↑ ex (n → ∞) (x > 0) :
n
exponential growth is the limit of compound interest as the interest is com-
pounded continuously.
Note. 1. Exponential growth (or decay) occurs widely in nature, as well as
in economics/finance! – e.g.
Growth of biological populations. A species expanding into a new environ-
ment with no limit on resources can and does grow exponentially (rabbits in
Australia, bacteria overwhelming a weakened organism, etc.).
Radioactive decay. The amount of material decays exponentially (the rate is
measured by the half-life – the time it takes for half of what is left to decay).
2. Exponential growth is extremely rapid, and hence is destabilizing. This
is one of the underlying causes behind financial crises (such as the Credit
Crunch, or Crash of 2008). The wonder is that the financial world is as sta-
ble as it is, most of the time.
3. Possibly for this reason, there were religious prohibitions on lending money
at interest – ‘usury’, as the (Christian, Catholic) Church called it in the Mid-
dle Ages (there are still such reservations in parts of the Islamic world). But:
4. Without the inducement of interest, it is difficult to justify taking the risk
of lending money. Without lending to finance investment and business ac-
tivity, credit dries up, business activity slows down and the economy shrinks
(as recently).

1
Discounting.
With economic activity driven by money lent and borrowed at interest,
one must distinguish between prices in nominal terms and prices in real
terms, whenever any length of time has elapsed. Otherwise, one is not com-
paring like with like. (This gives a good way to distinguish ‘good’ from ‘bad’
politicians, with an election pending!) We do this by discounting. With r
the current short-term interest rate (short rate), one discounts over a time-
interval of length t by a factor e−rt . In much of the mathematics below, we
will discount everything, and if St denotes a price at the present time t, its real
value at a later time T in present (time t) prices is e−r(T −t) St . This is done
for accounting purposes in Net Present Value (NPV) calculations. These are
also used as a tool in assessing whether or not an investment should be made.
We will not pursue this in detail here: part of the message of Real Options
(VII.5) is that the NPV approach to investment decisions is misleading.
For most purposes, we use as interest rate the rate for riskless lend-
ing/borrowing. Of course, this overlooks a lot of complications! For example:
1. Borrowing v. lending. Traditionally, banks made their money on the dif-
ference between the interest they paid to depositors on their savings accounts,
and the higher rate they charged borrowers, whether private customers on
overdrafts or businesses on borrowing for investment. However, banks not
only lend, they borrow, from each other (LIBOR = London Inter-Bank Of-
fer Rate, etc.).1 The difference between such borrowing and lending rates
introduces friction into the markets. One often (e.g., in the standard Black-
Scholes theory below) neglects this, and works with an idealized market.
2. Risky v. riskless. Major government debt (Treasury bonds in the US, gilts
in the UK) is traditionally regarded as riskless. However, governments do
occasionally default on their debts (Mexico and Russia, within fairly recent
memory). Banks also occasionally default (though government institutions
– the Fed in US, the Bank of England in UK – may intervene as ‘lender of
last resort’ to rescue a troubled institution – e.g. Northern Rock, UK, 2008).
3. Fixed rate or variable. Interest rates vary, and one may need to reflect
this by using a function r(t, T ) rather than a constant r.
4. Stochastic or deterministic. It may be that the variability r(t, T ) is better
1
In the recent LIBOR scandal – known inevitably as the Lie-bor scandal – it emerged
that individual traders in the participating banks (which only included the large and
reputable ones) were systematically distorting the quotes at which they purported to lend
or borrow. This is a blatant form of market manipulation; it has led to large fines already,
and to heavy (and expensive) reputational damage to the banks.

2
modelled as random, or stochastic – see Ch. IV below, and VII.5.
5. Term structure of interest rates. The mathematics of this – the term
structure of interest rates – is very interesting and important. But it is (a
lot) harder than most of the mathematics we shall be doing here.
6. Stock markets v. money markets (bond markets). There are two princi-
pal kinds of financial markets, stock markets and money markets. In stock
markets, what are traded – bought and sold – are stocks and shares – part-
ownership in a company, both its physical assets, its intangible assets (good-
will, brand name etc.) and its future earnings, which may be released to
shareholders as dividends. Investors may buy for capital appreciation, div-
idends or both. The prototype of the relevant mathematics is the Black-
Scholes theory, of which more below in Ch. V and VII. On the other hand,
in money markets, money is being borrowed (usually by governments, or
large companies) over time, to be repaid later at specified rates of interest
and/or with specified payments or coupons.
The borrower binds himself to repay with interest, and the agreement to
do so is called a bond, hence bond markets (Shylock, Merchant of Venice: I
will have my bond!). Bonds issued by leading governments are regarded as
safe, or ‘gilt-edged’, hence the term gilts in the UK. Their US equivalents are
called Treasury bonds, or T-bonds.

2. Economics and Finance; Utility

Economics is largely concerned with questions such as supply and demand
(for everything: commodities, manufactures, capital, labour, etc.). Much of
economics deals with how prices are determined.
Finance is concerned with the borrowing and lending of money needed to
engage in economic activity. We will be dealing here with the more math-
ematical side of this. In (mathematical) finance, one takes prices largely
as given, and deals with questions such as how to price options (real or
otherwise). Thus Finance is a part of Economics – and a fairly small and
specialized part at that.
Small economic agents are price takers. They have no power to influence
prices, which they can either take or leave – but equally, do have the power
to enter the market without thereby moving the market against them. By
contrast, large economic agents are price makers. They do have the power to
influence prices – but against this, are visible, and so are vulnerable, when
forced to enter the market through weakness (example: the financial author-
ities of a major country, defending the value of its currency by buying it on

3
the market).
Trading.
The price of common everyday items is accurately known at any given
time. Anyone trying to sell at a higher price than the ‘going rate’ would tend
to lose market share to cheaper competitors, and eventually have to reduce
towards the going rate or go out of business. At the other extreme, items
never bought and sold do not have a price – are literally priceless (Bucking-
ham Palace, Westminster Abbey, the Houses of Parliament, ...).
In between the two, prices are known but not accurately – to within some
interval. This is the bid-ask spread – the gap between the price at which a
market participant will buy, and the (higher) price at which he will sell.
When large trades are made, prices jump. This is because the large trade
affects the current balance between supply and demand, and the price is the
level at which markets clear – that is, at which supply and demand balance.
One can model such a market by means of a stochastic process (Ch. IV)
with jumps (prototype: Poisson process).
With small trades, one can look at things at two different levels of detail.
‘From a distance’, prices seem to move continuously – so can be modelled by
a stochastic process which is continuous (prototype: Brownian motion, VI.3).
But imagine a trader spending a trading day tracking the price movement
of a heavily traded stock under normal market conditions. From a distance,
price movement looks continuous, but close up, prices move by lots of little
jumps – the effects of the individual small trades, and how they briefly affect
the current balance between supply and demand. Such movement of prices
by ‘lots of little jumps’ is called jitter.
Utility.
A pound is worth much more to a poor man than to a rich man. For
ordinary people, a 10% increase in income might well give an extra 10% of
satisfaction. So for small amounts x of money, we can think of utility as
being the same as money. But to a billionaire, it would be hardly noticeable.
To model this, one uses a utility function, U (.), which measures how much
utility – genuine use – money is to the economic agent in question: income x
gives utility U (x). The effect above is called the Law of diminishing returns
(or diminishing utility – U is strictly increasing, but its graph bends below
the line y = x, and indeed is typically bounded above. Often used here are
the Inada conditions (Ken-Ichi Inada in 1963):
U (0) = 0; U ∈ C 1 ; U ↑; U 00 < 0 (so U concave); U 0 (0+) = ∞; U 0 (∞−) = 0.
A guiding principle that is often used here is that each economic agent

4
should seek to maximize his expected utility. This approach goes back to John
Von Neumann and Oscar Morgenstern in 1947 (in their classic book Theory
of games and economic behaviour, one of ‘the books of the last century’),
and earlier to F. P. Ramsey (1906-1930) in 1931 (posthumously).
Loss.
This is often looked at the other way round. One uses a loss function –
which can usually be thought of as a negative of utility. One then seeks to
minimize one’s expected loss.
Arbitrage.
An arbitrage opportunity (see II.6) is the possibility of extracting riskless
profit from the market. In an orderly market, this should not be possible
– at least, to a first approximation. For, an arbitrage opportunity is ‘free
money’; arbitrageurs will take this, in unlimited quantities – until the per-
son or institution being so exploited is driven from the market (bankrupt or
otherwise). In view of this, we make the assumption that the market is free
of arbitrage – is arbitrage-free, or has no arbitrage, NA.
Idealized markets.
Various assumptions are commonly made, in order to bring to bear the
tools of mathematics on the broad field of economic/financial activity. All
are useful, but valid to a first approximation only.
1. No arbitrage (NA).
2. No transaction costs or transaction taxes.
3. Same interest rates for borrowing and lending.
4. Unlimited liquidity (the ability to turn goods into money, and vice versa,
at the currently quoted prices).
5. No limitations of scale.
Markets satisfying such assumptions will be called perfect, or frictionless –
unrealistic in detail, but a useful first approximation in practice.

3. Brief history of Mathematical Finance

Mathematical Finance I: Markowitz and CAPM.
We deal with the history of put-call parity (II.7) below. It has ancient
roots, but entered the textbooks around 1904. Louis Bachelier (1870-1946)
first put mathematics to work on finance in his 1900 thesis Théorie de la
spéculation.2 Bachelier’s thesis is also remarkable as he used Brownian mo-
2
Mark Davis and Alison Etheridge: Louis Bachelier’s Theory of speculation: The ori-
gins of modern finance, translated and with a commentary; foreword by Paul A. Samuelson.

5
tion as a model for the driving noise in the price of a risky asset. This was
remarkable, as the relevant mathematics did not exist until 1923 (Wiener),
and later (Itô, stochastic calculus, 1944).
Until 1952, finance was more an art than a science. This changed with the
1952 thesis of Harry Markowitz (1927–), which introduced modern portfolio
theory. Markowitz gave us two key insights, both so ‘obvious’ that they are
all around us now. There is no point in investing in the stock market, which
is risky, when one can instead invest risklessly by putting money in the bank,
unless one expects the (rate of) return on the stock, µ, to be higher than the
riskless return r. The riskiness of the stock is measured by a parameter,
the volatility σ, which corresponds to the standard deviation (square root of
the variance) in a model of the risky stock price as a stochastic process (Ch.
IV), while µ, r correspond to means, for risky and riskless assets respectively.
Markowitz’s first key insight is: think of risk and return together, not sepa-
rately. This leads to mean-variance analysis.
Next, the investor is free to choose which sector of the economy to invest
in. He is investing in the face of uncertainty (or risk), and in each sector
he chooses, prices may move against him. He should insure against this by
holding a balanced portfolio, of assets from a number of different sectors,
chosen so that they will tend to ‘move against each other’. Then, ‘what he
loses on the swings he will gain on the roundabouts’. This tendency to move
against each other is measured by negative correlation (the term comes from
Statistics). Markowitz’s second key insight is:
diversify, by holding a balanced portfolio with lots of negative correlation.
Markowitz’s theory was developed during the 1960s, in the capital asset
pricing model (CAPM – ‘cap-emm’), of Sharpe, Lintner and Mossin (William
Sharpe (1964), John Lintner (1965), Jan Mossin (1966); Jack Treynor (1961,
1962)). In CAPM, one looks at the excess of a particular stock over that
of the market overall, and the risk (as measured by volatility), and seeks to
obtain the maximum return for a given risk (or minimum risk for a given
return), which will hold on the efficient frontier. The relevant mathematics
involves Linear Regression in Statistics, and Linear Programming in Opera-
tional Research (OR).
Mathematical Finance II: Black, Scholes and Merton.
If one is contemplating buying a particular stock, intending to hold it
for a year say, what one would love to know is the price in a year’s time,
Princeton UP, 2006.

6
compared with the price today (one should discount this, as above). If the
(discounted) price goes up, one will be glad in a year’s time that one bought;
if it goes down, one will be sorry.
Suppose one’s Fairy Godmother appeared, and gave one a piece of paper,
which said that if one bought now, then in a year’s time if one was glad one
had done so one did buy, but if one was sorry, one didn’t. Such pieces of
paper do exist, and are called options – see Ch. V, VII. Clearly such options
are valuable: they may lead to a profit, but cannot lead to a loss.
Question: What is an option worth?
Note that unless one can price options, they will not be traded (at least
in any quantity) – as with anything else.
Before 1973, the conventional wisdom was that this question had no an-
swer: it could have no answer, because the answer would necessarily depend
on the economic agent’s attitude to risk (that is, on his utility function, or
loss function – see above). It turns out that this view is incorrect. Subject
to the above assumptions of an idealized market (NA, etc.), one can price
options, according to the famous Black-Scholes formula of 1973 (Ch. V,
VII – Fischer Black (1938-1995) and Myron Scholes (1941-)). They derived
their formula by showing that the option price satisfied a partial differential
equation (PDE), of parabolic type (a variant of the heat equation). In 1973
Robert Merton (1944-) gave a more direct approach. Meanwhile, 1973 was
also the year when the first exchange for buying and selling options opened,
the Chicago Board Options Exchange (CBOE).
To see why options can be priced, one only needs to know that the stan-
dard options are (under our idealized assumptions) redundant financial as-
sets: an option is equivalent to an appropriate combination of cash and stock.
Knowing how much cash, how much stock and the current stock price, one
can thus calculate the current option price by simple arithmetic.
In 1981, it was shown (by J. M. Harrison and S. R. Pliska) that the
right mathematical machinery to use in this area involves a particular type
of stochastic process – martingales – and a particular type of calculus, for
stochastic processes – Itô calculus (Kiyosi Itô (1915-2008)); see Ch. VII.
The subject of Mathematical Finance is by now well-established, and
rapidly growing in popularity in universities, in UK, US and elsewhere. This
is because of its relevance to the needs of the financial sector (or financial
services industry) in the City of London (also Edinburgh) within UK, New
York in USA, Tokyo in Japan, Frankfurt in Germany, etc. This sector needs
technical people with good skills in mathematics, statistics, numerics etc.,

7
as well as economic insight and financial awareness, problem-solving skills
and ability to work in a team, etc. Such people are variously called financial
engineers, quantitative analysts (‘quants’) or ‘rocket scientists’.
Academically, the subject falls broadly in the interface between Eco-
nomics on the one hand and Mathematics on the other. In Economics, much
of the subject, again broadly speaking, relates to how prices are determined
– by the interplay between supply and demand, etc. By contrast, here in
this course we will usually take prices as given. Our task is to study how,
starting from the given prices, we can price other things related to them (op-
tions, and other financial derivatives – see below), and guard our operations
against unpredictable hazards (hedge – again, see below).
In this sense, Finance as a subject appears as a small – specialised, highly
mathematical – part of Economics (note that Finance here is not used quite
in the traditional non-technical sense). Risk is the key danger – the key con-
cept even – in finance; risk reflects uncertainty; uncertainty reflects chance
or probability. So it was clear that Probability Theory, a branch of Mathe-
matics related to Statistics, had to be relevant here. Quite how was shown
in 1981 by J. M. (Michael) Harrison (a probabilist) and David Kreps (an
economist), who simplified and generalized the Black-Scholes-Merton theory
by using the language of Probability Theory and Stochastic Processes – in
particular, martingales (and Itô calculus, again). These developments – and
what followed – constituted the ‘second revolution in mathematical finance’.
This is the subject-matter of this course. (We can cover the mathematics of
the developments outlined above. More recent developments are very impor-
tant, but go beyond a first undergraduate course – see e.g. our MSc in MF.)
On the mathematical side: you will learn a lot about stochastic processes,
martingales and Itô calculus, and see them put to use on financial problems.
On the practical side: the best proof of the relevance and usefulness of these
ideas is the explosive growth in volumes of trades in financial derivatives over
the last forty-odd years, and the corresponding explosive growth in employ-
ment opportunities (and salaries!) for those who understand what is going
on.
Black, Scholes and Merton
As everywhere, triumph and disaster can always happen, and one has to
use common sense. Triumph: Scholes and Merton were awarded the Nobel
Prize for Economics in 1997 (Black died in 1995, and the prize cannot be
awarded posthumously). Disaster: Scholes and Merton were on the board of
the hedge fund Long Term Capital Management, which ignominiously col-

8
lapsed with enormous losses in 1998. Pushing a good theory too far – beyond
all sensible limits – is asking for trouble, even if one invented the theory and
got the Nobel Prize for it, and if one asks for trouble, one can expect to get it.

4. Markets and Options.

Markets.
This course is about the mathematics of financial markets. Types:
Stock markets [New York, London, ...], dealing in stocks/shares/equities, etc.,
Bond markets, dealing in government bonds (gilts, ...),
Currency or foreign exchange (‘forex’) markets,
Futures and options markets, dealing in financial instruments derived from
the above - financial derivatives such as options of various types.
Options.
Economic activity, and trading, involves risk. One may have to, or choose
to, make a judgement involving committing funds (‘taking a position’) based
on prediction of the future in the presence of uncertainty. With hindsight,
one might or might not regret taking that position. An option is a financial
instrument giving one the right but not the obligation to make a specified
transaction at (or by) a specified date at a specified price. Whether or not
the option will be exercised depends on (is contingent on) the uncertain
future, so is also known as a contingent claim.
Types of option.
Call options give one the right (but not – without further comment now –
the obligation) to buy. [Remember: to buy something, one calls for it.]
Put options give one the right to sell. [Remember: to sell something, one
puts it on the market.]
European options give one the right to buy/sell on the specified date, the
expiry date, when the option expires or matures.
American options give one the right to buy/sell at any time prior to or at
expiry. Thus:
European options: exercise at expiry,
American options: exercise by expiry.
Note. The terms European, American (Asian, etc.) refer only to the type
of option, and no longer bear any relation to the area in the name. Most
options traded worldwide these days are American.
History. As discussed in §1, over-the-counter (OTC) options were long ago
negotiated by a broker between a buyer and a seller. Then in 1973 (the year
of the Black-Scholes formula, the central result of the course), the Chicago

9
Board Options Exchange (CBOE) began trading in options on some stocks.
Since then, the growth of options has been explosive. Options are now traded
on all the major world exchanges, in enormous volumes. Often, the market
in derivatives is much larger than the market in the underlying assets – an
important source of instability in financial markets.
The simplest call and put options are now so standard they are called
vanilla options. Many kinds of options now exist, including so-called exotic
options (Asian, barrier, etc.) – on which you need a separate course. For
real options (also called investment options) – see L30.
Terminology.
The asset to which the option refers is called the underlying asset or the
underlying. The price at which the transaction to buy/sell the underlying,
on/by the expiry date (if exercised), is called the exercise price or strike price.
We shall usually use K for the strike price, time t = 0 for the initial time
(when the contract between the buyer and the seller of the option is struck),
time t = T for the expiry or final time.
Calls.
Consider a European call option, with strike price K and underlying
worth St at time t. If ST > K, the option is in the money: the holder
will/should exercise the option, obtaining an asset worth ST (> K) for K.
He can immediately sell the asset for ST , making a profit of ST − K (> 0).
If ST = K, the option is said to be at the money.
If ST < K, the option is out of the money, and should not be exercised. It is
worthless, and is thrown away.
The pay-off from the option is thus
ST − K if ST > K, 0 otherwise,
which may be written more briefly as
max(ST − K, 0) or (ST − K)+
(x+ := max(x, 0), x− := −min(x, 0); x = x+ − x− , |x| = x+ + x− ; (−x)+ =
max(−x, 0) = −min(x, 0) = x− ).
Puts.
Similarly, the payoff from a put option is
K − ST if ST ≤ K, 0 if ST > K, or (K − ST )+ .
Option pricing.
The fundamental problem in the mathematics of options is that of option

10
pricing. The modern theory began with the Black-Scholes formula for pricing
European options in 1973. We shall deal with the Black-Scholes theory, and
cover the pricing of European options in full. We also discuss American op-
tions: these are harder, and lack explicit formulae such as the Black-Scholes
formula; consequently, one needs to evaluate them numerically. The pricing
of Asian options is even harder.
Perfect Markets. For simplicity, we shall confine ourselves to option pricing
in the simplest (idealised) case, of a perfect, or frictionless, market. First,
there are no transaction costs (one can include transaction costs in the the-
ory, but this is considerably harder). Similarly, we assume that interest rates
for borrowing and for lending are the same (which is unrealistic, as banks
make their money on the difference), and also that all traders have access to
the same – perfect – information about the past history of price movements,
but have no foreknowledge of price-sensitive information (i.e. no insider trad-
ing). We shall assume no restrictions on liquidity – that is, one can buy or
sell unlimited quantities of stock at the currently quoted price. That is, our
economic agents are price takers and not price makers. (This comes back
to §1 on the relationship between Economics and Finance. In practice, big
trades do move markets. Also, in a crisis, no-one wants to trade, and liquid-
ity dries up – basically, this is what did for LTCM.) In practice, very small
trades are not economic (the stockbroker may only deal in units of reasonable
size, etc.). We shall ignore all these complications for the sake of simplicity.

5. Portfolios and Hedging.

Portfolios.
We consider an investor with capital to invest. The simplest model is
that in which he has two (or more) choices: to invest in
(i) a bank account – assumed riskless, and yielding interest. For simplicity,
we assume the interest rate is a constant r > 0 (usually called the short rate
of interest: interest rates may be different outside [0, T ]); thus $B invested
at time t grows to $Ber(T −t) by time t;
(ii) one (or more) risky assets or stocks, whose value at time t is St (scalar
or vector).
A portfolio is a division (Bt , St ) of the investor’s capital between bank ac-
count and stock holdings at time t.
A trading strategy is a rule (suitably restricted – see Chapters V and VII)
chosen by the investor to update his portfolio over time as new price infor-
mation on the risky stock(s) comes in.

11
Hedging. Hedging is an attempt to reduce exposure to risk by adopting
opposite positions – e.g., in holding both call and put options in the same
underlying, or by adjusting a portfolio as above to cover possible losses.
Why buy options? Options have two main uses: speculation and hedging. In
speculation, available funds (‘hot money’) are invested opportunistically in
the hope of making a profit: the underlying itself is irrelevant to the investor
(speculator), who is only interested in the potential for possible profit that
trade involving it may present. Hedging, by contrast, is typically engaged in
by companies who have to deal habitually in intrinsically risky assets such as
foreign exchange next year, wheat/copper/oil next year, etc. This protects
their economic base (trade in wheat/copper/oil, or manufacture of products
using these as raw materials), and lets them focus their effort in their chosen
area of trade or manufacture. But for speculators, it is the market (forex,
commodities or whatever) itself which is their main focus.
Because the value of an option at expiry is so sensitive to price – it reflects
movements in the price of the underlying in exaggerated form – the holding
(or trading) of options and other derivatives presents greater opportunities
for profit (and indeed, for loss) than trade in the underlying (this is why
speculators buy options!). They are correspondingly more risky.
One of the main insights of the fundamental work of Black and Scholes
was that one can (at least in the most basic model) hedge against meeting
a contingent claim by replicating it: constructing a portfolio, adjusted or re-
balanced as time unfolds and new price information comes in, whose pay-off
is the amount of the contingent claim.

6. Arbitrage.
Economic agents go to the market for various reasons. One the one hand,
companies may wish to insure, or hedge, against adverse price movements
that might affect their core business. On the other hand, speculators may
be uninterested in the specific economic background, but only interested in
making a profit from some financial transaction. The relation between hedg-
ing (‘good’) and speculation (‘bad’) is to some extent symbiotic (one cannot
lay off a risk unless someone else is prepared to take it on, and why should he
unless he expects to make money by doing so). Nevertheless, one feels that it
should not be possible to extract money from the market without genuinely
engaging in it, by taking risk: all business activity is risky. Indeed, were it
possible to do so, people would do so – in unlimited quantities, thus sucking
money parasitically out of the market, using it as a ‘money-pump’. This

12
would undermine the stability and viability of the market in the long run –
and in particular, make it impossible for the market to be in equilibrium.
The view we take of modelling markets as NA is not that arbitrage op-
portunities do not exist, but that if they do exist in any quantity, people will
rush to exploit them, and thereby dissipate them – ‘arbitrage them away’.
Financial markets involve both riskless (bank account) and risky (stocks,
etc.) assets. For investors, the only point of taking risk is the chance of a
greater profit than the riskless procedure of putting one’s money in the bank
(the mathematics of which – compound interest – does not require a degree
or MSc course!). In general, the greater the risk, the greater the return re-
quired to make investment an attractive enough prospect to attract funds.
It is usually better to work, not in face-value or nominal terms, but in
discounted terms, allowing for the exponential growth-rate ert of risklessly
invested money. So, profit and loss are generally reckoned against this dis-
counted benchmark. So a market with arbitrage opportunities would be a
disorderly market – too disorderly to model. The remarkable thing is the
converse. It turns out that the minimal requirement of absence of arbitrage
opportunities is enough to allow one to build a model of a financial mar-
ket which – while admittedly idealised (frictionless market – no transaction
costs, etc.) – is realistic enough both to provide real insight and to handle the
mathematics necessary to price standard options (Black-Scholes theory). We
shall see that arbitrage arguments suffice to determine prices – the arbitrage
pricing technique (APT – S. A. Ross, 1976, 1978).
Short-selling.
Just as we can borrow money from the bank, in many markets, risky as-
sets such as stocks may be treated in the same way. We may have a positive
or zero holding – or a negative holding (notionally borrowing stock, which we
will be obliged to repay – or repay its current value). In particular, we may
be allowed to sell stock we do not own. This is called short-selling, and is
perfectly legal (subject to appropriate regulation) in many markets. Think
of short-selling as borrowing.
Not only is short-selling both routine and necessary in some contexts, such
as foreign exchange and commodities futures, it simplifies the mathematics.
So we assume, unless otherwise specified, no restriction on short-selling. By
extension, we call a portfolio, or position, short in an asset if the holding
of the asset is negative, long if the holding of the asset is positive. It turns
out that in some important contexts – such as the Black-Scholes theory of
European and American calls – short-selling can be avoided. In such cases,

13
it is natural and sensible to do so: see Ch. VII.

7. Put-Call Parity.
Just as long and short positions are diametrical opposites, so are call and
put options. We now use arbitrage to show how they are linked.
Suppose there is a risky asset, value S (or St at time t), with European
call and put options on it, value C, P (or Ct , Pt ), with expiry time T and
strike-price K. Consider a portfolio which is long one asset, long one put
and short one call; write Π (or Πt ) for the value of this portfolio. So

Π=S+P −C (S: long asset; P: long put; -C: short call).

Recall that the payoffs at expiry are:


max(S − K, 0) or (S − K)+ for a call, C,
max(K − S, 0) or (K − S)+ = (S − K)− for a put, P.

So the value of the above portfolio at expiry is K: for, it is

S + 0 − (S − K) = K if S ≥ K, S + (K − S) − 0 = K if K ≥ S.

Alternatively, use x = x+ − x− and (−x)+ = x− with x = S − K.

This portfolio thus guarantees a payoff K at time T . How much is it worth
at time t?
Short answer (correct, and complete): Ke−r(T −t) , because it is financially
equivalent to cash K, so has the same time-t value as cash K.
Longer answer (included as an example of arbitrage arguments). The
riskless way to guarantee a payoff K at time T is to deposit Ke−r(T −t) in the
bank at time t and do nothing. If the portfolio is offered for sale at time t
too cheaply – at a price Π < Ke−r(T −t) – I can buy it, borrow Ke−r(T −t) from
the bank, and pocket a positive profit Ke−r(T −t) − Π > 0. At time T my
portfolio yields K (above), while my bank debt has grown to K. I clear my
cash account – use the one to pay off the other – thus locking in my earlier
profit, which is riskless. If on the other hand the portfolio is offered for sale
at time t at too high a price – at price Π > Ke−r(T −t) – I can do the exact
opposite. I sell the portfolio short – that is, I buy its negative, long one call,
short one put, short one asset, for −Π, and invest Ke−r(T −t) in the bank,
pocketing a positive profit −(−Π) − Ke−r(T −t) = Π − Ke−r(T −t) > 0. At time
T , my bank deposit has grown to K, and I again clear my cash account –

14
use this to meet my obligation K on the portfolio I sold short, again locking
in my earlier riskless profit. So the rational price for the portfolio at time t
is exactly Ke−r(T −t) . Any other price presents arbitrageurs with an arbitrage
opportunity (to make a riskless profit) – which they will take! Thus
(i) The price (or value) of the portfolio at time t is Ke−r(T −t) , that is,
S + P − C = Ke−r(T −t) .
This link between the prices of the underlying asset S and call and put op-
tions on it is called put-call parity.
(ii) The value of the portfolio S + P − C is the discounted value of the riskless
equivalent. This is a first glimpse at the central principle, or insight, of the
entire subject of option pricing. But in general, we will have ‘risk-neutral’ in
place of ‘riskless’; see II.8 below, Ch. V and Ch. VII.
(iii) Arbitrage arguments, although apparently qualitative, have quantita-
tive conclusions, and allow one to calculate precisely the rational price – or
arbitrage price – of a portfolio. The put-call parity argument above is the
simplest example – though typical – of the arbitrage pricing technique (APT).
(iv) The APT is due to S. A. Ross in 1976-78 (details in [BK], Preface). Put-
call parity has a long history (see Wikipedia).
Note. 1. History shows both that arbitrage opportunities exist (or are
sought) in the real world and that the exploiting of them is a delicate matter.
The collapse of Baring’s Bank in 1995 (the UK’s oldest bank, and bankers to
HMQ) was triggered by unauthorised dealings by one individual, who tried
and failed to exploit a fine margin between the Singapore and Osaka Stock
Exchanges. The leadership of Baring’s Bank at that time thought that the
trader involved had discovered a clever way to exploit price movements in
either direction between Singapore and Osaka. This is obviously impossible
on theoretical grounds, to anyone who knows any Physics. See Problems 2
Q1 (key phrases: perpetual motion machine; Maxwell’s demon; Second Law
of Thermodynamics; entropy).
2. Major finance houses have an arbitrage desk, where their arbs work.

8. An Example: Single-Period Binary Model.

We consider the following simple example, taken from
[CRR] COX, J. C., ROSS, S. A. & RUBINSTEIN, M. (1979): Option pricing:
a simplified approach. J. Financial Economics 7, 229-263.
For definiteness, we use the language of foreign exchange. Our risky asset
will be the current price in Swiss francs (SFR) of (say) 100 US $, supposed

15
X0 = 150 at time 0. Consider a call option with strike price K = 150 at time
T . The simplest case is the binary model, with two outcomes: suppose the
price XT of 100 $ at time T is (in SFR)

180 with probability p
XT =
90 with probability 1 − p.

The payoff H of the option will be 30 = 180 − 150 with probability p, 0 with
probability 1 − p, so has expectation EH = 30p. This would seem to be
the fair price for the option at t = 0, or allowing for an interest-rate r and
discounting, we get the value
H
30p
V0 = E = .
1+r 1+r
Take for simplicity p = 12 and r = 0 (no interest): the naive, or expectation,
value of the option at time 0 is

V0 = 15.

The Black-Scholes value of the option, however, is different. To derive it, we

follow the Black-Scholes prescription (Ch. IV, VI):
(i) First replace p by p∗ so that the price, properly discounted, behaves like
a fair game:
XT

X0 = E ∗ .
1+r
That is,
1
150 = (p∗ .180 + (1 − p∗ ).90);
1+r
for r = 0 this gives 60 = 90p∗ or p∗ = 2/3.
(ii) Now compute the fair price of the expected value in this new model:

∗ H
30p∗
V0 = E = ;
1+r 1+r
for r = 0 this gives the Black-Scholes value as V0 = 20.
Justification: it works! – as the arbitrage constructed below shows. For sim-
plicity, take r = 0.
We sell the option at time 0, for a price π(H), say. We then prepare for the
resulting contingent claim on us at time T by the option holder by using the

16
following strategy:

Sell the option for π(H) +π(H)

Buy $33.33 at the present exchange rate of 1.50 −50
Borrow SFR 30 +30
Balance π(H) − 20.
So our balance at time 0 is π(H) − 20. At time T , two cases are possible:
(i) The dollar has risen:

Option is exercised (against us) −30

Sell dollars at 1.80 +60
Repay loan −30
Balance 0.
(ii) The dollar has fallen:

Option is worthless 0.00

Sell dollars at 0.90 +30
Repay loan −30
Balance 0.
So the balance at time T is zero in both cases. The balance π(H) − 20 at
time 0 should thus also be zero, giving the Black-Scholes price π(H) = 20 as
above. For, any other price gives an arbitrage opportunity. Argue as in put-
call parity in §4: if the option is offered too cheaply, buy it; if it is offered
too dearly, write it (the equivalent for options to ‘sell it short’ for stock).
Thus any other price would offer an arbitrageur the opportunity to extract a
riskless profit, by appropriately buying and selling (Swiss francs, US dollars
and options) so as to exploit your mis-pricing.
The same argument with interest-rate r also applies: divide everything
through by 1 + r.
Note. This argument, and result, are independent of p, the ‘real’ probabil-
ity, and depend instead only on this ‘fictitious’ new probability, p∗ (which is
called the risk-neutral or risk-adjusted probability.
The example above is highly instructive. First, it clearly represents the
simplest possible non-trivial case: only two time-points (with one time-period
between them, hence the ‘single-period’ of the title), and only two possible
outcomes (hence the ‘binary’ of the title). Secondly, it shows that there is a
theory hidden here, which gives us a definite prescription to follow (and some

17
surprises, such as not involving the ‘real’ probability p above). This prescrip-
tion is simple to implement, and can be justified by explicitly constructing
an arbitrage to exploit doing anything else [if the option is offered for sale
too cheaply, buy it, if too dearly, write it]. This theory is the Black-Scholes
theory, which we consider in detail in Chapters V and VII. The technical key
to the Black-Scholes prescription is the introduction of p∗ and its associated
expectation operator E ∗ . In technical language, this is the equivalent mar-
tingale measure. Now each of these three terms needs full introduction. We
shall talk about measures in III.1 below, about equivalent measures in III.4,
and martingales in IV.3 and VI.2. We stress: the Black-Scholes theory –
that is, rational option pricing – cannot be done without all these concepts.
This is why we need Chapter III on the necessary background on measure
theory, and Chapters IV and VI on the necessary background on stochastic
processes.
There are basically three options open to those teaching, and learning,
how to price options etc.
1. One can avoid measure theory altogether (cf. [CR]). This is technically
possible rigorously in the discrete-time setting of Ch. IV – though at greater
length, because the key concepts cannot be addressed explicitly. It is also
possible non-rigorously in continuous time; cf. [WHD], who base their ap-
proach on partial differential equations (PDE).
2. One can learn measure theory first – say, from the excellent book [W].
This, however, puts the subject beyond the reach of most people who need
it and use it in practice – and beyond reach of most of this audience.
3. One can do as we shall do (and as [BK] and a number of other books
do): state what we need from measure theory, and use its language, con-
cepts, viewpoint and results, without proving anything. This makes good
sense: the constructions and proofs of measure theory are quite hard (say,
final year undergraduate or first-year postgraduate level for good mathemat-
ics students with a bent for analysis – quite a select group!). Using measure
theory taking its results for granted, however, is quite easy, as we shall see.

9. Complements
1. Types of risk.
Institutions encounter risks of various types. These include:
Market risk.
This is the risk that one’s current market position (the aggregate of risky
assets one holds) goes down in value (things one is long on get cheaper,

18
and/or things one is short on get dearer).
Credit risk.
This is the risk that counter-parties to one’s financial transactions may
default on their obligations.
When this happens, debts cannot be (or are not) paid in full. Usu-
ally, payment is made in part, by negotiation between the parties (it may
be cheaper to agree a partial repayment than to force the other party into
bankruptcy), or by the administrators or liquidators in the case of compa-
nies. This raises issues of moral hazard, below.
Operational risk.
This is risk arising from the internal procedures of an institution: failure
of computer systems for implementing transactions (the failure of the Taurus
clearing system on the London Stock Exchange was one example); fraudulent
or unauthorised trading made possible by inadequate supervision; etc.
Liquidity risk.
This is the risk that one will be unable to implement a planned or agreed
transaction because of lack of cash-in-hand to trade with, and/or willingness
to trade. The Credit Crunch of 2007/8 on was caused by banks realising
they had piles of toxic debt on their hands (see below), and so did not know
what their balance sheets were worth; that other banks were similarly placed;
hence that banks no longer trusted themselves or each other, and so refused
to lend to each other. So the financial system froze up; so the real economy
froze up.
Model risk.
To handle real-world phenomena of any complexity, one needs to model
them mathematically. To quote Box’s Dictum: All models are wrong; some
models are useful.3 Use of an inappropriate model to set the prices at which
one buys and sells exposes the institution to open-ended losses, to competi-
tors with better models.
2. Risk management. The problems of 2007/8 on have made the importance
of risk management obvious. For an excellent book-length treatment, see e.g.
[MFE] A. J. McNEIL, R. FREY & P. EMBRECHTS: Quantitative risk man-
agement: Concepts, techniques, tools. Princeton UP, 2005.
We know from Markowitz that we should have a balanced portfolio, with
lots of negative correlation. The danger is large losses. These are quantified
by the tails of the distributions – the joint distribution of our portfolio. The
3
George E. Box, 1919-. British statistician

19
point of diversifying is so that what we lose on the swings we gain on the
roundabouts. Two comments:
(a) Whether this works for large losses depends on the tail properties of the
joint distribution. It does not work if this is normal – as it is in the bench-
mark Black-Scholes model.
(b) When the whole market is falling – as in a financial crisis – none of the
risk-management techniques useful under normal market conditions work.
3. Moral hazard. Before the limited liability company, if one defaulted, one
was liable to the whole of the loss incurred by one’s counter-party. This made
trading very dangerous (the early traders were called merchant adventurers)
– all the more as insurance had not developed by then.4
Limited liability was what made ordinary people willing to undertake the
risks of trading, and so paved the way for the development of modern busi-
ness, commerce, capitalism etc.
The moral hazard here is the possibility of gambling with other people’s
money (see John Kay’s excellent new book, Other people’s money, L0). If it
works, fine. If not, walk away (writing off one’s limited liability) and leave
them to bear the loss. ”Playing Russian roulette with someone else’s head”
is a rather brutal description of this.
Bankruptcy law varies from country to country, and is too complicated
to pursue here. But one sees moral hazard where it concerns us in, e.g.:
(a) start-ups of hedge funds (or, dot-com companies);
(b) aggressive traders – who (for the sake of their bonuses) gamble with their
careers – but with other people’s money;
(c) credit rating – where the credit rating agencies had a financial incentive
to pass as AAA some highly questionable financial asset, etc.
4. Securitization. This term covers the drive in recent years to seek out
new financial markets by identifying risks that people might want to cover
themselves against, and creating new financial derivatives that can be sold to
address this perceived need. These derivatives too could be traded, etc. The
upshot was an explosion of trade in increasingly artificial financial products,
developed by the R&D departments of the financial institutions. By 2007/8,
the leaders of these institutions did not understand these products – could
not price them, and could not value their holdings of them (above).
4
Lloyds of London predates limited liability. The Lloyds participants – ”names” – had
unlimited liability. Many were driven into personal bankruptcy in the Lloyds scandals of
the 90s. See Google for the ghastly details.

20
 One specific trigger of the US crash in 2007 was the explosive growth
in sub-prime mortgages. These were granted to people who would not have
qualified as financially sound enough to get a mortgage previously, but who
wanted to buy their own house. This new and profitable market proved ir-
resistible to US banks – leading to a great house-price bubble, which burst
(as bubbles do) in 2007. The knock-on effects hit the UK in 2008 (Northern
Rock, etc.). The real damage of this failure of the financial sector has been
its devastating and ongoing consequences on the real economy.
5. Macro-prudential issues. As the above illustrates, financial matters are
too important to be left to financiers. Proper regulation is vital.
6. Forwards and futures. Forwards are agreements between buyer and seller
made now, but concerning delivery in the future. They are not traded. Fu-
tures are options on things that will come to market in the future (next year’s
grain crop, for example), and these are traded (extensively). There are good
accounts in Hull’s books, [H1], [H2].
7. OTC and exchange-traded contracts. OTC – ”over-the-counter” – denotes
a transaction made between an individual buyer and an individual seller. As
options on standard transactions develop, these are assets themselves that
can be traded in exchanges (e.g., the CBOE, which opened in 1973: II.3).
8. Marking to market. This is a system whereby the exchanges cover them-
selves and their clients against the risk of large losses. If one party to a trade
is, on current market prices, exposed to a potentially heavy loss, a margin
call will be required by the exchange. Margin calls actually trigger many
financial failures (but limit the losses of the counter-parties).
9. Forex. Forex is an abbreviation for foreign exchange. International trade
involves more than one currency’ currencies move against each other. There
is a vast market in derivatives to cover the risks involved.
10. Swaps. From Hull [H2] Ch. 5: ”Swaps are private agreements between
two companies to exchange cash flows in the future ... The first swap con-
tracts were negotiated in 1981. Since then the market has grown very rapidly.
...” There are even options on swaps – swaptions – etc.

21
10. Postscript to Ch. II: Systemic issues – ”Big-picture stuff” (not
examinable).
Recall (Ch0) the three underlying truths about this course (indeed, this
subject, at this level): Anything important enough becomes political. Politics
is not an exact science. Mathematics is an exact science.
Thus, while this course is self-contained as “F22”, it is not and cannot be
self-contained if we broaden the picture: the big picture is, to some extent,
unmathematisable – even in principle. We illustrate this with a few themes.
(i) Finance and economics. What ultimately counts – in terms of jobs, peo-
ple’s lives, growing food, making things that people want, etc. – is the real
economy. Finance is the means (the supply of money) to an end (the pro-
duction of food, goods, services, creation of jobs, etc.). But, damage to the
financial system (the Slump of 1929 and the 1930s, the Crash of 2007/8/...)
can be longer-lasting than economic damage. Think of the financial system
as the nervous system of the economy. A nervous breakdown can disrupt life
more than a broken leg, etc. – and have longer-lasting effects (cf. PTSD).
(ii) Economics: Keynes, Hayek/Friedman, the postwar consensus, the neo-
liberal consensus. The great crisis of world economics 1900-50 was the Slump
(or Depression), the misery caused by which was a major cause of WWII.
Three factors helped to cure this: F. D. Roosevelt and the New Deal (US,
1930s); J. M. Keynes (books, 1930s); the stimulus to economies of WWII.
Keynesian economics (governments should spend their way out of a slump by
infrastructure projects etc.) formed the post-war consensus (1945-1979/80).
Then came the neo-liberal consensus (politically, Thatcher and Reagan; in-
tellectual underpinning from Hayek and Friedman) – free markets, legal re-
strictions on trade unions, etc. (and the collapse of communism). The Nobel-
prize winning US economist Paul Krugman puts it thus: big business doesn’t
like Keynesian economics, as it diminishes its bargaining power. Do your
own digging here, and form your own view. But note that in science, liking
something or not has zero effect on whether it is right, or whether it works:
a scientist should like what he sees, rather than see what he likes. But in
economics this is less true, because big issues involve politics, and so per-
ception, and so psychology. Ordinary members of the public do not even
know the names of the economists mentioned above, still less what they did;
this limits their ability to make informed decisions about matters of current
political controversy, where they see things through the ‘distorting prism’ of
media coverage, partisan political debate, etc. Three comments:
(a) The above underlines the difference in mission between the Imperial Col-

22
lege of Science, Technology and Medicine and (say) the London School of
Economics and Political Science;
(b) Anyone intending to work in, say, the financial services industry should
know at least as much about such things as a well-informed member of the
public. This cannot be done overnight! One obvious way is to read a decent
newspaper regularly over a period of time (or online equivalent).
(c) The profound importance of the issues here can be seen in, e.g., the
damage to the life prospects of the young. Even discounting Greece here,
youth unemployment has been over 50% in some European countries for
long periods since the Crash (and much higher in disadvantaged parts of the
community, with predictable consequences for social cohesion, crime etc.).
(iii) The Euro. By common consent, the euro was political in aim and in
origin. It is financial and economic in substance and in content. There is a
tension between these two. This is seen dramatically in the events concerning
Greece (Grexit?) and the UK (Brexit). What is your view on all this?
(iv) Asset-price bubbles. The Greenspan years were one long asset-price bub-
ble, which ended in tears. The Chinese economy now has another one ... .
(v) Quantitative easing (QE). Since the Crash, governments moved to rescue
a banking system at risk of collapse by QE. The idea was to ‘create elec-
tronic money’ for banks to lend to businesses, to free up and kick-start the
real economy, so that life could get back to normal. What tended to happen
instead was that banks – also under pressure to rebuild their balance sheets
– did so by holding onto this new money, rather than lending it as intended.
At the same time, interest rates have been at historic lows for long peri-
ods – indeed, real (as distinct from nominal) interest rates have often been
negative, which would have been dismissed as laughable a decade ago. But
keeping interest rates at near-zero for long periods has itself had distorting
effects – such as fuelling a new asset-price bubble, and so sewing the seeds
for the next Crash. How should policy-makers proceed here?
(vi) Banks: too big to fail? separate investment banking from retail banking?
As Keynes said: if you owe the bank a thousand pounds, it’s your problem;
if you own the bank a million pounds, it’s the bank’s problem. Seeming ‘too
big to fail’ (without risking a collapse of confidence as with Lehman Broth-
ers) helped the banks in their dealings with governments post-Crash. The
Glass-Steagall Act was passed in the US in 1932 to separate investment (UK:
merchant; ”casino”) banks from retail (”high-street”) banks, but this was re-
pealed in 1999. What is the right approach to regulation and legislation here?

23