EC202 Lecture 1: Introduction and Recap of

fundamentals

The outline of this lecture is:

1. Administrative stuff: An overview of the structure and what support

you can expect.

2. Mathematical recap: Mainly on logic.

3. Consumer theory recap: Most of which will be familiar from year 1.

1 Administrative stuff
This outlines the rules of the game. This is a two term module. This term
with me; next term with Christian Soegaard. The below refers to the half of
the module taught by me.

1.1 Assessments
There are two forms of assessment.

• Midterm test worth 10% of the module. Multiple choice questions.

• Final exam worth 80% of the module, examining material from both
terms in the summer.

1
1.2 Support we give you
You have the following resources at your disposal:

• Lecture notes on moodle. These are quite comprehensive and written

mathematically with focus on prescision instead of ease of comprehen-
sion.

• Pre-recorded videos: These will be less formal and explain the material
covered in lecture notes in a more accessible way. I won’t cover all
parts of the lecture notes, especially where there are large blocks of text
where I have nothing to add besides what is already written, but will
instead focus on parts where I can hopefully add to your understanding.
These are a complement not a substitute to reading the lecture
notes. There will be one video for each Section of each lecture notes.
For example, Lecture 1 has 3 Sections which will be covered by Videos
1.1, 1.2, 1.3.

• Live lectures via MS teams: These happen every Thursday 2-3pm dur-
ing term . These are to help fill the gaps in student understanding,
where students would like me to expand on either what was presented
in video or written in lecture notes. The material covered here is flexible
to reflect student wishes.

– You should read through lecture notes and listen to pre-recorded

lecture at least 2 days in advance of the live lecture. This way if
there is anything you would especially like me to cover, you can
drop me an email in advance.

• 8 Seminars, one per week in weeks 3 to 9, with an additional seminar

in term 3.

• Problem sets, for which my solutions will be posted on moodle.

2
• Past exam papers:

– This is my 2nd year teaching this module and you have last year’s
exam plus mock exam questions to help you revise.

• Advice on how to set your own practice questions and check solutions.

– This is key. While you have only finitely many of my questions

with solutions, this is a potentially infinite resource.
– Tip: Pay attention to methods I will teach you to check your
answers.

• Each other: a great way to learn is to discuss concepts, work on problem

sets and attempt practice questions with your classmates.

– Tip: Write practice questions for each other. That way when you
solve them, you can compare solutions to each other.

• Staff office hours (advice and feedback hours). Myself and your seminar
tutors all have office hours every week during term time (posted on
webpage). These might be in-person or online due to coronavirus but
the principle is the same.

– Tip: Discuss material with friends as they might have similar

questions to you and feel free to attend office hours in groups of
two or three.
– Tip: Attend office hours as problems come up. Don’t wait until
just before the exam when office hours will be crowded.
– Tip: If staff office hours are at a time you have lectures then ask
for an alternate time. There is no guarantee of being able to find
an alternate time, but normally staff will try to be flexible.

3
 • Email: My email address is James.Massey@warwick.ac.uk

– If you have short admin questions that can be answered by email,

then this is best form of communication.
– If you have mathematical questions, then come to my office hour
(or email me for a meeting time). Explaining technical concepts by
email is not realistic as it requires a back and forth conversation
to check you understand what I am saying. As such, if anyone
asks me such questions, my response will be “come to my office
hour”. If you would like me to explain something but can’t make
my office hour then email me for an alternative time.
– Tip: Do not waste time sending long emails about stuff you don’t
understand. Instead, make a list and come to see me or a seminar
tutor in person during an office hour, (or arrnge a video call if
office hours are online).

1.3 Advice:
Keep up to date with the material as we go along. Don’t
be lazy throughout term and assume you can learn
everything just before the exam.

Here’s why this is so key:

• Often material later on in the term builds on material covered earlier,

so if you don’t understand the earlier stuff you will be confused later.

• The key to success in this module is understanding the concepts, not

memorising stuff.

• All the face-to-face help we give you (lectures, seminars, office hours) is
during term time, not when you are revising during the Easter vacation.

4
To keep up to date, I recommend:

• Read through lecture notes on moodle before lecture, so you know

where you need extra help during lecture.

• After lecture discuss anything you didn’t understand with friends.

• Attempt problem sets before seminars and compare your solutions to

those of your seminar tutor.

• If having done the above, you still have points of query then attend
staff office hours.

1.4 The Syllabus

1.4.1 The 10 lecture topics

We have one lecture topic per week and eight seminars covering a problem
set.
1. Introduction and Recap of fundamentals. (Problem set 1.)
2. General Equilibrium in Pure Exchange Economies. (Problem set 2.)
3. Pareto optimality and Welfare Theorems in Pure Exchange Economies.
(Problem set 3.)
4. Adding in production: Robinson Crusoe example. (Problem set 4.)
5. Walrasian Equilibrium and Welfare Theorems in Economies with Pro-
duction. (Problem set 4.)
6. Market failure due to Externalities. (Problem set 5.)
7. Public goods games. (Problem set 6.)
8. Social Choice Theory. (Problem set 7.)
9. Bargaining. (Problem set 8.)
10. Co-operative Game Theory. (Problem set 8.)

5
1.4.2 Overall vision

We can split the overall message into 3 parts:

Part 1: Efficiency of the free market (Weeks 2-5)

Weeks 2 and 3 show that under our basic model and appropriate assumptions
on preferences, the free market works well. Allowing consumers to trade with
one another results in a Pareto efficient outcome (First Welfare theorem)
and any Pareto efficient outcome can be achieved as an Equilibrium with
appropriate transfers (Second Welfare Theorem). Weeks 4 and 5 show that
this remains true when we add production to our model too.

Part 2: Market Failure (Weeks 6-8)

Week 6 shows that when we have externalities, either positive or negative

then Walrasian Equilibria are no longer Pareto efficient. Week 7 extends this
further focusing on the issue of Public goods which would be under-provided
if left to the free market and hence the need for government to step in. Week 8
discusses the problems that arise when trying to aggregate voter preferences.
Arrows’ Theorem and the Gibbard-Satterthwaite Theorem show that there
are no voting systems that satisfy desirable axioms when we have at least 3
alternatives to choose between.

Part 3: Dividing the Pie (Weeks 9-10)

Week 9 discusses some bargaining models which discuss how, in the absence
of competitive markets, players might split the gains from trade. We see
models both from a game theory perspective and an axiomatic perspective.
Week 10 extends this discussion of how to split gains from trade from an
axiomatic perspective to situations where there are more than two players.

6
1.5 What is examinable?
• For the midterm test, The first 7 topics and first 6 problem sets are
examinable.

• For the Summer Exam, everything is examinable.

1.6 Further Readings

These lecture notes and associated problem sets are designed to be sufficient.
However I will reference further readings for those who would like them.
Often in the notes, I will refer to places where students can read further.

1.6.1 Main texts

Unfortunately, different textbooks use different notation and terminology.

As such the notation and terminology I will use will not follow any single
textbook but will be consistent with itself and where texts disagree, I will
try to use the more common terminology.

• “Microeconomics: An intuitive approach with Calculus”, by Thomas J

Nechyba. (Lectures 1-7)

– This offers a relatively accessible coverage of the subject. Al-

though in parts, where I feel more rigourous approach is needed I
will take definitions and Theorems from Mas-Colell:

• “Microeconomic Theory” by Mas-Colell, Whinston and Green. (all

weeks)

– This book is advanced and as such is more suitable for those who
have mastered my lecture notes and want to learn more, rather
than those who are struggling. It is a good book to read for anyone
looking to study Microeconomics at postgraduate level.

7
For Lectures 8-10, I recommend:

• “Game Theory: A Multi-Leveled Approach”, by Hans Peters (available

as an e-book from Library)

1.6.2 Supplementary texts

In addition to Nechyba’s book, the following texts and many other under-
graduate Microeconomics textbooks will also be good on Lectures 1-7:

• “Microeconomics,” by Jeffrey M Perloff

• “Intermediate Economics with Calculus” by Hal R Varian

• “Microeconomics” by Morgan, Katz and Rozen

• “Microeconomic Theory concepts and connections” by Michael E Wet-

zstein

Additionally the following may be useful for further reading in some parts of
the course:

• “Essential Mathematics for Economic Analysis” by Sydsaeter and Ham-

mond.

• “Bargaining Theory with Applications” by Abhinay Muthoo.

• “Game Theory: A multi-leveled approach” by Hans Peters

2 Mathematical Recap
For further reading see “Essential Mathematics for Economic Analysis” by
Sydsaeter and Hammond.
The main mathematial skills you need for EC202 this term are:

• Algebra (Chapters 1 and 2)

8
 A

• Calculus (Chapter 6)

• Logic (Section 3.4)

B A
• Optimisation (Chapters 8, 13, 14)

From last year, you should be well-practiced on the first two already, but
perhaps not so much the third and fourth. So I will give you a recap:

2.1 Necessary and sufficient conditions

The foundation of all logical thinking is the statement “ A implies B”. In this
section we consider what it means and give examples of when it does and
does not hold.

Definition 2.1. Consider two statements A and B. Then A implies B,

written A =∆ B or B ≈= A, means that the set of states where A is true is
a subset of those states where B is true.
Put another way, the statement “A implies B” is logically correct unless
it is possible to simultaneously have A is true and B is false.

Exercise 2.1. Represent the statement A =∆ B using a Venn-diagram.

Remark 2.1. If the statement A is always false then “A implies B” holds

trivially since the empty set is a subset of every set.1
Typically though, we like to write more useful statements where both A
and B might be true or might be false.

Example 2.1. x = 3 =∆ x2 = 9. Here both statements could be true or

false, depending on the value of x. But every time we have x = 3, we also
have x2 = 9, thus the statement x = 3 =∆ x2 = 9 is correct.
1
As an example where this could be helpful to know: in EC228, you could be asked
questions of the form: “Show that if a voting system satisfies certain axioms and elects
x under preferences P Õ , then it must elect y under preferences P ÕÕ . If you can show that
it is impossible for a voting system to satisfy the axioms and elect x under P1, then you
have proved the statement, without even needing to look at P2.

9
 However x = 3 ≈= x2 = 9 is incorrect. To express this, we could write
x = 3 : x2 = 9. The reason is that x2 = 9 also happens when x = ≠3.

There are also many other ways of expressing the idea of implication.
Here are a few applied to the example above:

• x = 3 implies x2 = 9.

• x2 = 9 is implied by x = 3 .

• If x = 3, then x2 = 9.

• x2 = 9 if/whenever x = 3.

• x = 3 is a sufficient condition for x2 = 9.

• x2 = 9 is a necessary condition for x = 3.

• x = 3 only if x2 = 9.

Remark 2.2. If the statement “A implies B” is correct, then note that this
only tells you something about the veracity of B when A is true. If A is false,
then B could be either true or false.

Example 2.2. x > 1 =∆ x > 0. This implication is correct. Suppose we are

working out an equation and we can show that x > 1, then this implication
means we can conclude that x > 0. If we are unable to show that x > 1 then
this implication gives us no help determining whether x > 0. It is possible
x > 0 might still hold (eg x = 12 ) but also possible x > 0 will not hold (eg
x = ≠1).

Remark 2.3. To show a statement of the form A =∆ B is false, you need to

find an example where A is true but B is false. If no such example exists,
then the statement is true.

10
 Example 2.3. Let y = x + 5. Then we can verify:
1. x > 4 =∆ y > 4 is correct because there is no (x, y) satisfying all of
y = x + 5, x > 4 and y Æ 4.
2. The statements x2 > 4 and y 2 > 4 are unrelated, because for each
possible implication we can construct a counter-example to disprove it:
To disprove x2 > 4 =∆ y 2 > 4, we can use the example (x, y) = (≠5, 0).
To disprove x2 > 4 ≈= y 2 > 4 we can use the example (x, y) = (1, 6).

Definition 2.2. When we have both A =∆ B and A ≈= B, we write

They represent the

same set on the Venn A ≈∆ B and say that A and B are equivalent to each other.
diagram .

An alternative way of expressing this is A happens if and only if B

happens.

Example 2.4. Again let y = x + 5. Then y Ø 0 ≈∆ x Ø ≠5 because the

set of states where y Ø 0 is exactly the same set of states as where x Ø ≠5.

In conclusion, between any two statements A and B we have 4 possibilities:

• A ∆ B (but A : B).

– This happens if the set of states where A happens is a subset of

the set of states where B happens. And there exists at least one
instance of B happening without A.

• A ≈ B (but A ; B).

– Same idea as above with A and B switched.

• A…B

– The set of states where A happens is exactly equal to the set of

states where B happens.

• A and B are unrelated.

11
 – This happens if neither implies the other. That is, there are times
when A happens without B; and times when B happens without
A.

2.2 The Contrapositive

Every statement of the form A =∆ B has a logical twin, known as the
contrapositive. To see this, lets represent what A =∆ B means in a Venn-
diagram.

Graphical representation of “A implies B”. In

every state where A holds true, B also holds Set of all states of
the world, X.
true, since 𝐴 ⊆ 𝐵.
X
Set of states of the
world where B holds
true.
B
Set of states of the
world where A holds
true.
A
𝐴𝑐 = ˺𝐴 = 𝑋 ∖ 𝐴.
𝐵𝑐 = ˺𝐵 = 𝑋 ∖ 𝐵.

Notice that 𝐵𝑐 ⊆ 𝐴𝑐 .

What is the significance of

this?

Definition 2.3. The contrapositive of the statement “A =∆ B” is “not B =∆

not A”.
These two statements are equivalent to each other. I.e. if one is true,
then the other must be true; and if one is false then the other must be false
too.

Remark 2.4. If we take the contrapositive of the contrapositive, we get back

to the original statement itself.

:ThmmdomkBm%Ba
12

÷:÷÷::÷÷÷÷÷÷÷:÷:÷÷
 The contrapositive of “not B =∆ not A” is “not not A =∆ not not B”
which is “A =∆ B” once we cancel the double negative.
By definition, if we take the contrapositive of any true statement we get
another true statement.

Example 2.5. x Ø 4 =∆ x Ø 3 has contrapositive x < 3 =∆ x < 4

Similarly if we take the contrapositive of any false statement, we get

another false statement.

Example 2.6. x2 > 0 =∆ x > 0 has contrapositive x Æ 0 =∆ x2 Æ 0.

Here, taking the contrapositive has made it far more obvious why the
initial statement was false: If we take any negative number, its square is
positive instead of negative.

Remark 2.5. When trying to prove or disprove a statement of the form

“A =∆ B”, we could equally well try to prove or disprove the contrapos-
itive “not B =∆ not A”.
Another way to see this relationship is to consider what we need to do to
disprove “A =∆ B”:
Recall that “A =∆ B” is true unless there is a state of the world where
A happens and B does not. While “not B =∆ not A” is true unless there
is a state of the world where not B happens and not A does not. This is
equivalent to saying B doesn’t happen and A does happen, which is exactly
what was needed to falsify “A =∆ B”. We have already seen an example
of this in Example 2.6, where the same set of examples falsified both the
original statement and its contrapositive.

Exercise 2.2. For each of the following statements, find it’s contrapositive:
1. x > 0 =∆ y > 0.
2. x = 3 =∆ x2 = 9

13
2.3 ’, ÷ and negation.
Definition 2.4. Mathematicians often use the symbol “’” to mean “for all”
or “for every”.
Ô
For example “’x > 0, we can define y = x” meaning “for all x greater
than 0, we can define y to be the square root of x”.
Sometimes this formulation can mask a hidden implication:
Example 2.7. ’x > 0 we have y > 3 translates as “for all x > 0, we have
y > 3.
This could alternatively be written as x > 0 =∆ y > 3. This statement
holds as long as there doesn’t exist some x > 0 such that y Æ 3.
Definition 2.5. Mathematicians often use the symbol “÷” to mean “there
exists”. And they use the symbol @ to mean “there does not exist”.
Example 2.8. Consider the statement “’a, b œ R, with a < b, ÷c œ (a, b)”.
This translates as: for any two Real numbers a and b, there exists a third
Real number c which lies between a and b. This is indeed a true statement,
for example c = a+b
2
.
Next we ask how to negate statements:
Remark 2.6. Take the statement

’ (this), we have (that)

its negation is

÷ an instance of (this) where (that) does not happen.

If we negate this statement again but in a different way, we get2

@ an instance of (this) where (that) does not happen.

2
Often the “we have” in the first statement is dropped. While the “where” in 2nd and
3rd statements is sometimes replaced by s.t. which stands for “such that”.

14
 Thus the first and third statements are equivalent to each other, since
they are both the negation of the second.

Example 2.9. Take the statement

’x œ A, (we have) x œ B

An equivalent way of expressing this is:

x œ A =∆ x œ B (2.1)

This statement is true unless there is a counterexample to it. A coun-

terexample would be where

x œ A is true but x œ B is false

Thus the negation of our statement is3

÷x œ A, s.t. x œ
/B

Negating this statement in a different way we get

@x œ A, s.t. x œ
/B

To translate this into English:

There doesn’t exist an x which is in A but not in B

Thus if you consider an x not in B then it must not be in A. Translating

back to mathspeak:

/ B =∆ x œ
xœ /A (2.2)
3
“s.t.” stands for “such that” or “satisfying”

15
 ¥÷÷÷ .

> C E 7 A V > B

Note: these statements should be familiar: the initial statement was a

wordy definition of A ™ B. Equation (2.1) translates this into mathematical
symbols, while equation (2.2) is the contrapositive of (2.1).

Compound statements

We often want to build longer statements and we should know what they
mean and how to negate them.

Example 2.10. Take the statement “for all natural numbers n, if xn > yn
then xn < 10.”
’n œ N, xn > yn =∆ xn < 10 (2.3)

Taking it’s negation we get “there exists some natural number n satisfying
xn > yn but not xn < 10. So we get:

÷n œ N, s.t. xn > yn and xn Ø 10

Negating this, we get that an equivalent way of expressing equation (2.3)

is “there does not exist a natural number n satisfying both xn > yn and
xn Ø 10.
@n œ N, s.t. xn > yn and xn Ø 10

To say this another way, “for any natural number n, at least one of xn > yn
and xn Ø 10 must be violated”. Thus if one holds, the other cannot. Equation
(2.3) can be seen as saying if the first holds, then the second does not. But
alternatively we could have if xn Ø 10 holds then xn > yn does not. So we
get

’n œ N, xn Ø 10 =∆ xn Æ yn (2.4)

Note that equations (2.3) and (2.4) are contrapositives of each other.

Often we want to express things like “if both A and B happen then C

16
happens” and know what they mean and what their contrapositives are.

Example 2.11. Take the following statement (which we know to be true):

If x satisfies both x > 0 and x2 = 9 then x = 3.
1 2
x > 0 and x2 = 9 =∆ x = 3 (2.5)

Forming its contrapositive

1 2
not (x = 3) =∆ not x > 0 and x2 = 9

Which simplifies to

x ”= 3 =∆ x Æ 0 or x2 ”= 9 (2.6)

To see why the term to the right of the implication simplifies in this way,
put it into words: not (x > 0 and x2 = 9) means that at least one of x > 0
and x2 = 9 doesn’t happen. Ie x Æ 0 or x2 ”= 9.4
Note that since the original statement in equation (2.5) was true, its con-
trapositive (2.6) is also a statement we know to be true. While its negation,
below, will be a statement we know to always be false:

÷x s.t. x > 0 and x2 = 9 but x ”= 3

The implication above actually holds in both directions so we could write

1 2
x > 0 and x2 = 9 ≈∆ x = 3

Lets explore the contrapositive and negation of the implication in the

opposite:
4
Recall from the section on sets, when taking intersection and union, that “and” means
both things happen; while “or” means at least one happens.

17
Example 2.12. Take the following statement (which we know to be true):
1 2
x > 0 and x2 = 9 ≈= x = 3 (2.7)

Forming its contrapositive

1 2
not x = 3 ≈= not x > 0 and x2 = 9

Which simplifies to

x ”= 3 ≈= x Æ 0 or x2 ”= 9 (2.8)

To express what this means in words: if either of x Æ 0 or x2 ”= 9 happen

then x ”= 3.
Note, as expected, (Remark 2.4) if we take the contrapositive of the
contrapositive (2.8) then we get back to the original statement (2.7).
Again, both the original statement and the contrapositive were statements
we know to always be true. so taking the negation of (2.7) we should get a
statement we know to be false:
1 2
÷x st x = 3 but not x > 0 and x2 = 9

Which simplifies to

÷x st x = 3 and either x Æ 0 or x2 ”= 9

This is the statement we get more directly from negating (2.8) and is a
statement we know to be false.

Remark 2.7. As seen from our examples above, negating a statement often
turns “and” into “or” and vice-versa.

18
 This is because

not (X and Y) = not (X)or not (Y )

not (X or Y) = not (X)and not (Y )

Give yourself a few examples and translate them into words if you need
further convincing. Using logical reasoning such as the above, we can show

Lemma 2.1. Let A ,B andC be any three statements. (Using “¬” to mean
“not”)
1. A and B =∆ C has contrapositive ¬C =∆ ¬A or ¬B .
2. C =∆ A or B has contrapositive ¬A and ¬B =∆ ¬C.
3. C =∆ A and B has contrapositive ¬A or ¬B =∆ ¬C.
4. A or B =∆ C has contrapositive ¬C =∆ ¬A and ¬B

Exercise 2.3. Translate the implications in Lemma 2.1 into set notation
and use this to verify Lemma 2.1. The diagrams below should give you much
help.

19
 A A

B B

Note the
C C
switch between
N for the 𝐴 ∩ 𝐵 ⊆ 𝐶. So “A and B” implies C” 𝐴 ∪ 𝐵 ⊇ 𝐶. So “C” implies “A or B”
original statement Also note ¬𝐶 ⊆ ¬𝐴 ∪ ¬𝐵 for contrapositive. Also note ¬𝐶 ⊇ ¬𝐴 ∩ ¬𝐵 for contrapositive.
& u for the
contra A
positive
B
A B

C
𝐴 ∪ 𝐵 ⊆ 𝐶. So “A or B” implies “C”
𝐴 ∩ 𝐵 ⊇ 𝐶. So “C” implies “A and B”.
Also note ¬𝐶 ⊆ ¬𝐴 ∩ ¬𝐵 for contrapositive.
Also note ¬𝐶 ⊇ ¬𝐴 ∪ ¬𝐵 for contrapositive.

2.4 Proving and disproving statements

• To prove a statement of the form A =∆ B:

– It is not sufficient to find an example where A is true and show

that B also holds.
– You need to show that B holds in every example where A is true.
– Alternatively by the contrapositive, you can show that if B is
violated then A must be too.
– This is closely linked to “proof by contradiction” where you as-
sume A is true and B is false and show that it generates a logical
contradiction.

Exercise 2.4. Recall the Game Theoretic definitions of a strategy weakly

dominating and strictly dominating another strategy. Show that strict dom-
ination implies weak domination.

20
 • To disprove a statement of the form A =∆ B:

– You do not need to show that whenever A is true, B must be false.

This would be to prove A =∆ ¬B and is a far stricter condition
than simply A =∆ B not holding.
– All you need to do is find an example where A is true and B.

Exercise 2.5. Recall the Game Theoretic definitions of a strategy weakly

dominating and strictly dominating another strategy. Disprove that weak
domination implies strict domination.

3 A Recap of Consumer Theory

Most or all of this will have been covered last year, hence we will go through
it pretty quick.

3.1 Consumption Set and Budget Constraint

For those looking to read in textbooks, see Chapter 2 of Nechyba or Chapter
2 of Mas-Colell.

Definition 3.1. The Consumption Set is defined as the set of commodities

*
(or goods) consumers can consume. Let there be J goods then we define
This is 70 because
Ø0 . A typical member of this set is x =
the consumption set by X = R0*
J
consumers can only
Consume (x1 , x2 , . . . , xJ ) where xj is the consumer’s consumption of good j, for j œ
a positive
no of
. the
good {1, 2, . . . , J}.

µ Generally consuming a negative amount of a good doesn’t make sense,

which is why we take X = RJØ0 , meaning that xj Ø 0 for all j œ {1, 2, . . . , J}.5 .
If we have two goods then the consumption set is the positive orthant.
I will henceforth refer to a member of the consumption set as a bundle of
goods or bundle for short.
5
Although in financial markets, investors are often allowed to short sell a commodity.

21
Remark 3.1. Notice the consumption set is unbounded. That is there is no
upper bound to how much of any one good can be consumed. Also note that
the consumption set is infinitely divisible: that is any Real number of any
good can be consumed, not just integer amounts. These assumptions are
made for technical convenience to allow us to apply our mathematical tools.
In some cases these may be fair assumptions and in other cases less so.
Consider these assumptions with the following goods:

• Litres of water

• Number of armchairs

• Tonnes of manure

• Number of diamonds

• Hours of leisure time

• Houses

Often, even when a commodity is very discrete like houses, we can still think
of it in continuous terms by considering a more abstract commodity like
housing quality. For example, you are a student who has many things you
can spend your money on, from beer to pizza to clothing etc. But one of
your most important decisions as a consumer will be finding somewhere to
live. In general, the more money you spend on rent, the greater quality of
house you will receive, in terms of size, facilities, location etc. So while the
number of houses you will rent at any one time is fixed at one, you do have
an important decision on how much of your budget you wish to spend on
housing compared to other goods.
This brings us to the budget constraint. This gives the set of bundles
consumers can afford. For j œ {1, 2, . . . , J}, let price of good j be pj Ø 0 (we
assume non-negative prices) and suppose the consumer has wealth M .

22
 Amount spent E Income
Definition 3.2. The budget constraint equation is: p1 x1 + . . . + pJ xJ Æ M .

This leads us onto the next definition:

Definition 3.3. The Walrasian budget set is the set of bundles our consumer
can choose between
Consumer BC
Ó
(x1 , . . . xJ ) œ RJØ0 | p1 x1 + . . . + pJ xJ Æ M
Ô µ for all goods

Exercise 3.1. Let J = 2. Draw this on a diagram and show what happens
to the budget constraint as prices change.

We can also define budget constraints when consumers have an initial

allocation instead of some exogeneous wealth, M . If a consumer has initial
allocation Ê = (Ê1 , . . . ÊJ ) then the value of this initial allocation is p1 Ê1 +
. . . + pJ ÊJ and this can be substituted as M in the above budget constraint
equation.

Exercise 3.2. Let J = 2 and the consumer’s initial allocation be Ê = (1, 2)

and let p1 = p2 . Draw this on a diagram and show what happens to the
budget constraint as prices change.

Note that the budget constraint here is always linear. This is something
that I will assume throughout the course and it is a simplification of the
real world. This assumption says that, no matter what quantity you wish to
purchase of the commodity you will be charged the same price per unit. When
we go to a supermarket and see “buy one get one free” offers or the like we
see an immediate contradiction to this. This kind of thing is unavoidable and
nothing to be perturbed about. Think of an economic model as being like a
drvier’s roadmap. To be useful it doesn’t have to list all the details like where
every tree is. It simply has to capture the most important aspects of reality
so as to be useful to us. Indeed with a roadmap, if it did tell you every little
detail like the location of every tree, it would obscure the more important like

23
 PRICE CHANGES
Xin '

If P , t to P, the
pivots inwards as consumer can afford less of P
•
curve ,

m -

The slope BC will

of
always be equal to the price ratio (as proven below)
-
.

%
slope =
ffs
, =Tf%- If = = Price Ratio

I $ X,
I
MP
,
'
P,

ENDOWMENTS
•
A consumer starts with a certain bundle (an endowment) and the value of this endowment determines
income for that
period (hence the BC will always pass through the endowment point e (shown below)

Xz

A
Pz
I -

PL

e
a.

r XI
MP7
M
p
,
where the main roads go. Most textbooks including Mas-Colell and Nechyba
will briefly discuss non-linear budget constraints for those interested but I
won’t delve into this issue. The aim in this course as you will see in future
weeks is to present a model for how the whole economy works and in forming
this, it is convenient to assume linear budget constraints. These represent a
reasonable approximation of reality and in our judgement this simplification
doesn’t impact the results of our model too much. Those willing to think
about this issue can attempt this exercise:

Exercise 3.3. Consider two goods: apples and bananas. Our consumer has
income M = 200 and faces prices (pa , pb ) = (10, 5). Draw the budget set.
How will your answer change if:
i) The consumer gets “buy one get one free on apples”
ii) The shop selling bananas at pb = 5 only has 10 units to sell. So if
the consumer wants more they will need to buy off another shop which sells
them at pb = 10.

3.2 Preferences
For those looking to read in textbooks, see Chapters 4 and 5 of Nechyba or
Chapters 1 and 3 of Mas-Colell.
Next we want to define a consumer’s preferences over all bundles of goods
in the consumption set. I introduce the following notation:

Definition 3.4. The preference relation ≤ defines weak preference over

bundles. For any two bundles of goods in X, x̂ = (x̂1 , . . . x̂J ) and x̄ =
(x̄1 , . . . , x̄J ), x̂ ≤ x̄ translates as x̂ is weakly preferred to x̄.

We can then derive the strict preference relation and indifference relation
from the weak preference relation:

Definition 3.5. Given ≤ we define:

The strict preference relation: x̂ º x̄ ≈∆ x̂ ≤ x̄ and not x̂ ∞ x̄.

24 ↳ I is strictly
preffeed to
I
 The indifference relation: x̂ ≥ x̄ ≈∆ x̂ ≤ x̄ and x̂ ∞ x̄.
Note : we
Two important axioms on preferences are completeness and transitivity: will
always
COMPLETENESS
µ choose the
Definition 3.6. Completeness: For any two bundles of goods in X, x̂ = preferred bundle
/ (x̂ , . . . x̂ ) and x̄ = (x̄ , . . . , x̄ ), at least one of x̂ ≤ x̄ or x̄ ≤ x̂ must hold.
L
1 J 1 J
because it has a
higher utility
When you have a (If both hold then x̄ ≥ x̂)
for the consumer
Utility function you
already have completeness This says that for any two bundles in our consumption set, we can com-
because the bundle pare them and say which is better. In reality, this is not always easy for
with a
higher utility is consumers to do. But for us economists, wanting to build a model we can
preferred to the other
analyse, this axiom is a must.

TRANSITIVITY Definition 3.7. Transitivity: If we prefer bundle 1 to bundle 2 and bundle

2 to bundle 3, then we should also prefer bundle 1 to bundle 3. If x(1) ≤ x(2)
If Ulx )7UHz)
and x(2) ≤ x(3) then x(1) ≤ x(3) .
,

And U4H 7UK 3)

Then UK17 0HD Typically in Economics, we call a consumer rational if they satisfy Com-
pleteness and Transitivity. Note that a consumer who violated transitivity
would be open to money pumps: If they had preferences x(1) ≤ x(2) ≤ x(3) ≤
x(1) , they would be prepared to pay a small amount of money to move from
x(1) to x(2) and again from x(2) to x(3) and again from x(3) to x(1) . In describ-
ing preferences which satisfy these axioms as rational, we do not pass any
judgement on what is good for a consumer, we simply represent what bundles
consumers prefer. For example it is possible a consumer really likes smoking
lots of cigarrettes despite them giving her lung cancer. Note that assuming
rationality allows us to rank the various alternatives for the consumer from
best to worst.
In all what follows, we will always assume the consumer has rational pref-
erences. There are five further properties on preferences we will sometimes
like to impose. The first two are convexity properties. Basically convexity
says that consumers prefer averages of goods to extremes.

25
 This is why the K Strict Convexity Weak Convexity
Xz Xz
Prefer averages to curve is
convexly •
extremes i. you would shaped because it •

rather have bundle "2

•
a
you take 2. points on •
• IC1 1CI
With a mixture of the curve and took
✗, X,
apples & bananas .
the
average of them
, The weighted average The weighted average
you would be on a of the 2 points is on of the 2 points is at
a higher Illit is strictly least
higher 1C C. higher on the same / (
preferred)
utility)
Definition 3.8. Convex preferences: For every x œ X, the upper contour
CONVEXITY
set {x̂ œ X | x̂ ≤ x} is a convex set. In other words, if x̂ ≤ x and x̄ ≤ x
then –x̂ + (1 ≠ –) x̄ ≤ x for any – œ [0, 1].

To give some rationale for this, suppose you like both apples and bananas
and are indifferent between one bundle that is {no apples and 10 bananas}
and a second bundle which is {10 apples and no bananas}. Most people
would prefer a bit of variety - if you have too much of the same thing you are
likely to get bored of it. Thus our consumer would likely prefer a third bundle
of {5 apples and 5 bananas}. Of course there may be scenarios this is not
true, for example if this consumer is wanting to use the fruit to make either
an apple pie or banoffee pie and with 10 of the same kind they can make a
dessert, while with 5 of each they cannot. Although, if instead you consider
apples and bananas over a course of a month, then convexity again appears
a valid assumption. In much of this course, we will make this assumption.
There is also a stronger version of it:

Definition 3.9. Strictly convex preferences: For every x œ X, if x̂ ≤ x and

x̄ ≤ x, with x̂ ”= x̄ then for any – œ (0, 1), –x̂ + (1 ≠ –) x̄ º x̄.

Next we come to our monotonicity properties. These are central to the

problem at hand. Mahatma Gandhi once said “The world has enough for
everyone’s need, but not for everyone’s greed.” Of course the line between
need and greed can be a grey one6 and as economists we won’t go into such
semantics. But this quote does capture an important truth about human
nature: No matter how much or how little we have, we always want more.
When I was in your shoes many years ago, this notion was taught to me by
a rather large American professor who came out with the immortal phrase
“more hamburgers are always preferred to less”. Whether or not he was
trying to be funny, the point stuck as I hope it will for you too. Of course we
6
Does one “need” a house to live when many people do without and survive. Do one
“need” clean unpolluted air when many people live in big polluted cities and survive?

26
 can question the validity of this claim. If you have 100 hamburgers already
or however many is too much to eat, do you really want one more? You
might argue not, but on the other hand I’m sure there would be some other
commodity, perhaps soda or extra comfortable seating, that you would like
more of. If we look around us at the very rich, they still want more, whether
that means upgrading a Volvo to a Lambroghini or buying a luxury yacht.
Even a billionaire philanthropist, still fits into this framework, if you consider
the charitable giving as a good in itself due to the warm glow effect and
prestige the giver receives.
These notions are encapsulated in the following three definitions.
"

Bbs Definition 3.10. Preferences are Strongly Monotone if for every x, x̂ œ X,

if xj Æ x̂j ’j œ J and ÷j œ J with xj < x̂j then x̂ º x.

X-qi.kz
In words, starting from any consumption bundle, if you increase the
- - - . -

amount you consume of any one good without decreasing the consumption
of any other good, then you are strictly better off.
'

,
SX ,
I If you find strong monotonicity hard to justify, you might prefer the
weaker property of monotonicity:
↳ Monotonicity
Suggests that Definition 3.11. Preferences are Monotone if for every x, x̂ œ X, if xj <
any bundle x̂j ’j œ J then x̂ º x.
in red
region
is preferred to In words, starting from any consumption bundle, if you increase the
Bundle X whereas amount you consume of every good, then you are strictly better off.
Bundle × is
preferred to In turn, monotonicity can be weakened to local non-satiation.
any bundle in blue
region is
Definition 3.12. Preferences are locally non-satiated if for every x œ X,
Strongly Monotone ’Á > 0, ÷x̂ œ X such that Îx ≠ x̂Î < Á7 and x̂ º x.
•

Includes the Òq
7
Îx ≠ x̂Î = j=1 (xj ≠ x̂j ). This is known as the Euclidean distance between two
J
boundary lines in
preferences vs points. For J = 2 you can draw this and see that it corresponds to the normal notion of
Weakly monotone distance as the crow flies.
doesn't include
boundary lines
27
 In words, starting from any consumption bundle, we can always find
another bundle “close to” it that the consumer strictly prefers. This holds
regardless of how small the distance must be to be considered “close to” (this
is the ’Á > 0 term) .

In the problem set, you are asked to show the alleged implications between
strong monotonicity, monotonicity and local non-satiation. Note that local
non-satiation allows one or more commodities to be “bads” although we can’t
have all commodities be bads as then 0 œ X would be a point of satiation.
Remark 3.2. In Lecture 3 when we give the First Welfare Theorem, we will
state local non-satiation as a sufficient condition for our result to hold. Since
it is weaker than monotonicity and strict monotonicity, it follows immediately
that they would both be sufficient conditions too.

Exercise 3.4. Let J = 2. For each of the five properties of preferences

defined above, draw indifference curves satisfying it and violating it. Draw
diagrams to illustrate how convexity differs from strict convexity and how
local non-satiation differs from monotonicity and how that in turn differs
from strong monotoncity.

Representability

As economists, it makes our analysis much easier if a consumer’s preferences

can be represented by a utility function.

Definition 3.13. The preference relation ≤ can be represented by a utility

function u : X æ R if for every pair of bundles x̄, x̂ œ X, x̂ ≤ x̄ ≈∆ u (x̂) Ø
u (x̄)

It is possible that there does not exist a u : X æ R that represents ≤.

An example of this is lexicographic preferences which are, for example, often
used for ranking countries on the medal table at the Olympics. In ranking
two countries’ medal hauls, they first look at number of gold medals, if tied

28
 look at the number of silver medals and if tied look at number of bronze
medals.
Remark 3.3. If preferences are represented by a utility function u then any
increasing transformation of u will also represent these preferences. Thus
there are infinitely many utility functions representing the same preferences.
For this reason, when we look at utility functions, we only care about whether
one bundle has higher utility than another; not about the exact numerical
amounts and differences in utility. That is utility functions are ordinal, not
cardinal.
Recall the following definition:

Definition 3.14. A function u : X æ R is quasi-concave if its upper level

sets, {x œ X : u(x) Ø c}, are convex for every c œ R.

Remark 3.4. Suppose preferences can be represented by u : X æ R. Then

preferences are convex iff u is quasi-concave.

Exercise 3.5. Let J = 2 and suppose our consumer has the utility function
u (x1 , x2 ) = x1 x2 . Draw it’s indifference curves. How do things change if the
Ô Ô
utility function changes to v (x1 , x2 ) = x1 x2 or w (x1 , x2 ) = 2 x1 x2 ≠ 17.
Which convexity and monotonicity properties are satisfied?
OPTIMAL BUNDLE
This is the

3.3 Consumer optimisation [Utility Max) highest 1C

you can

µ
get onto 4
For further reading see Chapter 6 of Nechyba or Chapter 3 of Mas-Colell.
Our next step is to combine Sections 3.1 and 3.2 to find the best bundle
the consumer can afford. In other words our consumer wants to solve
µg, ,↳emeµ,
won't work it : L
max u (x) subject to p1 x1 + . . . + pJ xJ Æ M But setting

/
xœX 2 A
→
The convexity assumption
isbroken
You should have had plenty of practice last year solving exercises like
%%%%%
→
Corner Solution case

where slopeotk this: Reminder Method -661×121 Reminder Method ↳ Solve

£9
:
Note : This method
]
.

is never equal to slope

works with convex ks
" BC
L=X,×XPtX(M RX Pak)
29 Slope k=Mks=-Mj

.ee#.? .pq&pII sp=IIeRato/:UsingM--P,x,tpzyz

-
-
of
,

Solve
using partial derivatives
slopes , =p

Solve using simultaneous equi,

Exercise 3.6. Let J = 2 and preferences be given by u (x1 , x2 ) = x1 x2 and
suppose the consumer has income of 200 and faces prices of (p1 , p2 ) = (10, 5).
Draw the budget constraint and indifference curves. Argue that the point of
optimisation occurs where simultaneously the consumer exhausts her budget
and the slope of the budget constraint is equal to the slope of the indifference
curve. Solve this pair of simulatneous equations to find the optimal bundle.
Lets disect further what is happening here: The slope of the budget
constraint, which is the price ratio ≠ pp12 is the rate at which our consumer is
allowed to trade one good for the other. The slope of the indifference curve is
the rate at which the consumer is willing to trade one good for the other. If
the two were not equal, lets say the slope of the indifference curve was a little
steeper, then this would mean the consumer could slide down the indifference
curve slightly, while getting to a point in the interior of the budget set. Draw
a diagram of this to convince yourself. But then if we are at a point in the
interior of the budget set, we could increase our consumption of both goods
and so improve utility. Thus we have argued that slope of indifference curve
equalling slope of budget constraint is a necessary condition for optimality
in this example.
At this stage I’ll refresh your memory of some terms:
• Marginal utility of good i is the rate at which utility of consumer in-
creases with good i: M Ui = du
dxi

• Marginal rate of substitution is the rate slope of the indifference curve,

the rate at which the consumer is willing to substiute good 1 for good
2. M RS1,2 = M U1
M U2

• A utility function of the form u (x1 , x2 ) = x–1 x—2 is called Cobb-Douglas.

This is a common type of utility function and has the property that
at the optimal bundle the consumer will spend –
–+—
of their income on
good 1 and —
–+—
of their income on good 2. You can use the method
above to verify this.

30
Last year you should have also come across the Lagrangean method where we
take L = x1 x2 +⁄ (200 ≠ 10p1 ≠ 5p2 ) and solve by setting dL
dx1
= dL
dx1
= dL
d⁄
= 0.
This is an alternate way of finding a bundle that exhausts the budget while
setting slope of budget line equal to slope of indifference curve (MRS equal
to price ratio). It has the advantage that it can also be easily scaled up for
cases when J > 2.
However note that using the Lagrangean or setting MRS equal to price
ratio will not always work for the following reasons:

• The utility function may not be differentiable.

• The optimal bundle might not be interior, which can happen for two
reasons:

– Preferences are not convex.

– Preferences are convex but the MRS is either always steeper or
always shallower than the budget line.

When these occur we need to use other tactics. An example of non-differentiability:

Exercise 3.7. Perfect complements: Let J = 2 and preferences be given by

u (x1 , x2 ) = min {x1 , 4x2 } and suppose the consumer has income of 200 and
faces prices of (p1 , p2 ) = (10, 5). Draw the budget constraint and indifference
curves. Argue that the point of optimisation occurs where simultaneously
the consumer exhausts her budget and we are at a kink of an indifference
curve. Solve this pair of simulatneous equations to find the optimal bundle.

An example of non-convexity:

Exercise 3.8. Let J = 2 and preferences be given by u (x1 , x2 ) = 2x21 +x22 and
suppose the consumer has income of 200 and faces prices of (p1 , p2 ) = (10, 5).
Draw the budget constraint and indifference curves. Argue that by using
the Lagrangean or an equivalent method we are not maximising utility but

31
are instead finding the bundle that minimises utility subject to spending all
income. Using your diagram, argue that the actual optimal bundle is a corner
solution and compare utilities at the two corners to find which is the optimal
bundle.
An example of slopes being non-equal:
Exercise 3.9. Perfect substitutes: Let J = 2 and preferences be given by
u (x1 , x2 ) = 3x1 + 2x2 and suppose the consumer has income of 200 and
faces prices of (p1 , p2 ) = (10, 5). Draw the budget constraint and indifference
curves. Show that attempting to set MRS equal to price ratio or applying
Lagrangean doesn’t yield a solution. Using your diagram, argue that the
actual optimal bundle is a corner solution and compare utilities at the two
corners to find which is the optimal bundle.
Note that Perfect substitutes are convex preferences but not strictly con-
vex. Although we could obtain a similar result with indifference curves that
had a slight curvature and so are strictly convex. To understand what is
going on, lets look back to our condition with Cobb-Douglas preferences: we
had
M U1 p1 M U1 M U2
M RS1,2 = = ≈∆ =
M U2 p2 p1 p2
This equation has a nice intuitive explanation M Ui
pi
is the “bang per buck”
you get for spending marginally more on good i. So if instead we had M U1
p1
>
M U2
p2
this means that for some small Á > 0, the consumer could improve utility
by transferring Á of her good 2 expenditure good 2 to good 1. Hence why we
need equality here at an interior solution. While in the perfect substitutes
example of Exercise 3.9 we had that for all bundles, M U1
p1
< M U2
p2
which is
why the consumer wants to spend all their income on good 2. With perfect
substitutes, it is possible for M U1
p1
= M U2
p2
to hold. But then it would hold at
every bundle and so the whole budget line would be optimal bundles. An
example of this occurring would be changing preferences in Exercise 3.9 to
u (x1 , x2 ) = 2x1 + x2 .

32
 Here are some useful results to help find optimal bundles:

Lemma 3.1. If preferences satisfy local non-satiation then the consumer

must expend all her budget to maximise utility.

Exercise 3.10. Prove the above Lemma. Hint: prove by contradiction.

qJ
That is, assume we have local non-satiation and x œ X with j=1 pj xj < M
maximises utility. Then generate a contradiction.

Lemma 3.2. If preferences satisfy local non-satiation and are represented by

a differentiable, quasi-concave utility function (so that preferences are convex)
then x œ X is an optimal bundle if and only if satisfies:
1. The consumer expends all her budget.
2. If xi = 0 then ÷j œ J, j ”= i such that M Ui
pi
Æ M Uj
pj
.
3. If xi > 0 then M Ui
pi
Ø M Uj
pj
’j œ J.
If additionally the utility function is strictly quasi-concave (so that pref-
erences are strictly convex) then the optimal bundle is unique.

This result says that at the optimal bundle, all goods consumed in a
positive quantity produce the same bang per buck as each other, while any
goods not consumed must have a weakly lower bang per buck.
Remark 3.5. Note that the solution to the utility maximisation problem
should be:
1. Homogeneous of degree 0 in terms of income and prices since if all
prices and income double the budget set remains unchanged.
2. Invariant to increasing transformations of the utility function, since
they all represent the same preferences, with the same indifference curves.

SUMMARY Lecture Recap

You should now know:

• How this module will be run this term.

33
• What is expected of you as a student.

• The meaning of “implies” or “if.... then....”, “sufficient”, “necessary”.

• When a statement involving these words is true or false.

• How to prove or disprove a statement involving these words.

• What a consumers budget set is and how to depict it graphically when

there are 2 goods.

• How the consumers bidget set changes as income or initial allocation

and prices change.

• How to depict preferences on a diagram.

• Different properties that preferences can have, in particular what “lo-

cally non-satiated” means.

• How to find the consumer’s optimal bundle subject to budget con-

straint.

34