UNIVERSITY OF BERGEN
Lecture 5 Consumer Theory
UNIVERSITY OF BERGEN
Course Schedule
Lecture Topic Book Chapters
1Mathematics
2Mathematics
3Mathematics
4Mathematics
5Consumer Theory 3&4
6Consumer Theory 4&5
7Consumer Theory and Producer Theory 5&9
8Producer Theory and Partial Equilibrium 10 & 11
9Partial Equilibrium and Market power 11 & 14
10General Equilibrium 12
11General Equilibrium 12
12Labor Markets 13
13Risk and Uncertainty 7
14Game Theory 8
15Game Theory 8
16Game Theory 8
17Game Theory 8
18Externalities and Preference Aggregation 17
19Asymmetric Information 8 & 16
20Asymmetric Information 16
21Asymmetric Information 16
22Asymmetric Information 16
23Summary Lecture
UNIVERSITY OF BERGEN
Today
• Consumer theory: Utility Maximisation and Choice
– The consumer environment
– Basic Preference Axioms
– Connection between the axioms and Utility
representation
– Utility functions and indifference curves
– Optimal choice using utility functions
PAGE 3
UNIVERSITY OF BERGEN
The consumer environment
• We assume that for the consumer in the economy there is a set of
possible goods 𝑥1 , 𝑥2 , … , 𝑥𝑛
• These could represent many types of things
• Apples
• Bread
• Cheap brand shoes
• Expensive brand shoes
• Free time
• Fresh air
• …
• The main choice that consumers will face is to decide what quantity
(possibly 0) of each good 𝑥𝑖 to consume
PAGE 4
UNIVERSITY OF BERGEN
The consumer enviroment
• The consumption of each commodity should be non-
negative
• We will typically assume commodities are divisible, so
a consumer can have half an apple or half a glas of
milk
• Of course there are restrictions to this choice.
Depending on the budget constraints a consumer
faces, she will in the end choose from a limited choice-
set.
PAGE 5
Basic Preference Axioms
• To model the consumer behavior side of the economy, neoclassical economists
decided to assume the consumer population consists of ‘rational and internally
consistent’ decisionmakers
• This is done by assuming the preferences of each consumer satisfies the
following three properties, or axioms:
– Completeness
– Transitivity
– Continuity
• These and other ‘rationality’ assumptions have been criticised, and alternative
theories (e.g behavioral economics) are developing
• Nevertheless this is a framework allowing for models that give a lot of insight,
and that are widely used in e.g. public policy, taxation and financial market
analyses
PAGE 6
UNIVERSITY OF BERGEN
Basic Preference Axioms
• Completeness: For any two options, say 𝑎 and 𝑏, a
consumer can choose, she either:
– Prefers 𝑎 over 𝑏
– Prefers 𝑏 over 𝑎
– Finds 𝑎 and 𝑏 exactly equally attractive
• So a consumer is here expected to be aware of all
possible consumption options and have a clear
preference over any possible pair of options.
PAGE 7
UNIVERSITY OF BERGEN
Basic Preference Axioms
• Transitivity: If a consumer prefers option 𝑎 over
option 𝑏, and simultaneously prefers option 𝑏 over
option 𝑐..
• ..then it should follow that she also prefers option 𝑎
over option 𝑐.
• This axom requires consumer preferences to be
‘consistent’ and it gives us that when a consumer is
given a menu of options, we know there should exist a
‘favorite’ option
PAGE 8
UNIVERSITY OF BERGEN
Basic Preference Axioms
• Continuity: If a consumer prefers option 𝑎 over option
𝑏, then she will also prefer situations that ‘close
enough’ resemble 𝑎 over option 𝑏.
• This is a slightly more ‘mathematically driven’
assumption and is usually given in a bit more formal
language
• But it means essentially that the consumer does not
exibit erratic jumps in preferences, based on minimal
changes to the given options.
PAGE 9
Connection between the axioms
and Utility representation
• One key result is that if a consumer’s preferences
satisfy all three mentioned axioms.
– Their preferences over the available options
can be represented as a list where for each
option, an option higher on the list is (weakly)
preferred to options lower in the list
– Weakly as two neighboring options on the list
could be preferred equally
PAGE 10
Connection between the axioms
and Utility representation
• Moreover the preferences of such an individual can be described by a
numerical ‘utility function’ 𝑈 𝑥 where 𝑈 𝑎 > 𝑈 𝑏 if option 𝑎 is
preferred to 𝑏 and 𝑈 𝑎 = 𝑈 𝑏 if the options are equally preferred
• A possible (but debatable) interpretation here is that a consumer will
likely experience more ‘utility’ from consuming a more preferred bundle
• However given the assumptions made, such a utility representation
does not specify ‘how much’ one option is preferred to another
• For example, multiplying all the utility amounts by 2, or adding ‘10’ to
each option’s utility score would still give a utility function describing
the same preferences
PAGE 11
Utility functions and indifference
curves
• Typically we will use the utility function to describe
the preference 𝑈 𝑥1 , 𝑥2 , … , 𝑥𝑛 over a bundle of
goods, where each 𝑥𝑖 denotes the amount of good
𝑖 that the consumer can consume with this option
• An additional assumption we typically make is that
all goods are economic goods. Meaning that the
consumer prefers, ceteris paribus, a bundle with
more of good 𝑖 in it.
PAGE 12
Utility functions and indifference
curves
• To gain insight into the properties of different utility
functions, we usually do not draw the utility function
directly, but we look at so called ‘indifference curves’
• An indifference curve is a set of points in the ‘space’ of
the set of possible consumption bundles, which gives
the consumer equal utility
• The continuity axiom guarantees that this is indeed a
connected set or a ‘curve’
• We can draw indifference curves for examples with
two goods
PAGE 13
Utility functions and indifference curves
Quantity of y
U3
U2
U1
Increasing utility
Curve of bundles that are
equally preffered
Quantity of x
PAGE 14
Utility functions and indifference curves
Quantity of y
U1 Slope of the indifference
curve at bundle (x1,y1) is
equal to minus the MRS
(marginal rate of
substitution) of good y for
good x
y1
Curve of bundles that are
equally preffered
Quantity of x
x1
PAGE 15
UNIVERSITY OF BERGEN
Utility functions and indifference curves
• The Marginal Rate of Substitution (of good y for good x) is
𝑑𝑦
defined as: 𝑀𝑅𝑆 = − ቚ
𝑑𝑥 𝑈=𝑈1
– Translated this means the MRS is equal to minus the
slope of the indifference curve corresponding to the
utility level 𝑈1
• Intuitively this corresponds (approximately) to how many
items of good y a consumer is willing to give up, to aquire
one more item of good x, while staying at the same utility
level (so being indifferent between the old and the new
bundle)
PAGE 16
UNIVERSITY OF BERGEN
Utility functions and indifference curves
• One common assumption made is that a
consumer’s preference satisfies ‘diminishing
marginal rates of substitution’ for each good
• For the two goods case, this means that as the
relative amount of good 𝑥 in a bundle with utility
level 𝑈1 goes up, the amount of good 𝑦 the
consumer is willing to give up for an extra unit of
good 𝑥 is decreasing
PAGE 17
UNIVERSITY OF BERGEN
Utility functions and indifference curves
• A direct consequence of the assumption of diminishing
marginal rates of substitution is that for two goods,
indifference curves will be ‘convex’
• Convexity here means that if you take two bundles 𝑥
and 𝑥 ′ that both lie on the same indifference curve,
than any bundle 𝑥 ′′ = 𝛼 ∙ 𝑥 + 1 − 𝛼 ∙ 𝑥 ′ with 0 < 𝛼 <
1 will be preferred over 𝑥 and 𝑥 ′
• So the diminishing MRS assumption will lead to
‘balanced’ bundles being typically more preferred than
bundles with mostly one good in them
PAGE 18
UNIVERSITY OF BERGEN
Calculating the MRS
𝑑𝑦
• So we defined the MRS as 𝑀𝑅𝑆 = − 𝑑𝑥 ቚ , but how does one
𝑈=𝑘
practically calculate this for a given utility function 𝑈 𝑥, 𝑦 ?
𝑑𝑦 𝑈𝑥
• The rule to use here is 𝑀𝑅𝑆 = − ቚ =
𝑑𝑥 𝑈=𝑘 𝑈𝑦
• The MRS can be calculated by looking at the ratio of the partial
derivatives of utility to goods 𝑥 and 𝑦
– And notice that the partial derivative of 𝑈 to good 𝑥 is simply
the marginal utility of consuming an extra unit of 𝑥
• In the case of a utility function with more than two goods
𝑈 𝑥1 , 𝑥2 , … , 𝑥𝑛 we can calculate the 𝑀𝑅𝑆 of two goods 𝑖 and 𝑗 as
𝑑𝑥𝑗 𝑈𝑖 𝑥1 ,𝑥2 ,…,𝑥𝑛
𝑀𝑅𝑆 = − ቚ =
𝑑𝑥𝑖 𝑈 𝑥 ,𝑥 ,…,𝑥 =𝑘 𝑈𝑗 𝑥1 ,𝑥2 ,…,𝑥𝑛
1 2 𝑛
UNIVERSITY OF BERGEN
Some typically used Utility functions
• Cobb-Douglas Utility: 𝑈 𝑥, 𝑦 = 𝑥 𝛼 ∙ 𝑦 𝛽
– Often ‘normalized’ with 𝛼 + 𝛽 = 1
– 𝛼 and 𝛽 could here be interpreted as the relative importance that a
consumer with these preferences attaches to goods 𝑥 and 𝑦
• Perfect substitutes: 𝑈 𝑥, 𝑦 = 𝛼 ∙ 𝑥 + 𝛽 ∙ 𝑦
– Indifference curves become straight lines
• Perfect complements: 𝑈 𝑥, 𝑦 = min 𝛼 ∙ 𝑥, 𝛽 ∙ 𝑦
– This models preferences over goods that can only be enjoyed together in a
certain ratio
𝑥𝛿 𝑦𝛿
• CES Utility: 𝑈 𝑥, 𝑦 = + if 𝛿 ≤ 1 and 𝑈 𝑥, 𝑦 = 𝑙𝑜𝑔 𝑥 + 𝑙𝑜𝑔 𝑦 if 𝛿 = 0
𝛿 𝛿
– This function has versions of the other three uitility specifications as special
cases for certain levels of 𝛿
PAGE 20
UNIVERSITY OF BERGEN
Homothetic preferences
• These are preferences where the marginal rate of
substitution only depends on the ratio of the
amounts of two goods.
• This means that the indifference curves at
different utility levels look similar and do not
change shape, for those type of preferences
• Here on a straight line from the origin, all points
on the line should correspond to a similar MRS
PAGE 21
Class question
UNIVERSITY OF BERGEN
Class question
• Show that the preferences described by the utility
function 𝑈 𝐴 = 𝑐1 ∙ 𝑐2 are homothetic, but those
described by 𝑈 𝐵 = 𝑐1 ∙ 𝑐2 + 𝑐1 are not.
PAGE 23
UNIVERSITY OF BERGEN
Class question + answer
• Show that the preferences described by the utility function 𝑈 𝐴 𝑐1 , 𝑐2 =
𝑐1 ∙ 𝑐2 are homothetic, but those described by 𝑈 𝐵 𝑐1 , 𝑐2 = 𝑐1 ∙ 𝑐2 + 𝑐1
are not.
𝑈1𝐴 𝑐 ,𝑐 𝑐
• 𝑀𝑅𝑆 𝐴 = 1 2
= 2
. Clearly here the marginal rate of substitution
𝑈2𝐴
𝑐1 ,𝑐2 𝑐1
only depends on the ratio of the amounts consumed of the two goods.
𝑈1𝐵 𝑐1 ,𝑐2 𝑐 +1
• 𝑀𝑅𝑆 𝐵 = = 2 . Here for example, consuming equal amounts
𝑈2𝐵 𝑐1 ,𝑐2 𝑐1
of goods one and two, would give different levels of 𝑀𝑅𝑆 depending on
the total consumption amount (compare for example the bundles 1,1
and 2,2 , so here the preferences are not homothetic.
PAGE 24
Optimal choice using utility
functions
• So what sort of consumption bundles would
consumers choose if they
– Hold preferences represented by a (mathematically
well behaved) utility function
– Tend to choose the bundle that will maximise their
utility level
– Faced constraints by having a limited budget and
prices connected to the consumption of each good
PAGE 25
UNIVERSITY OF BERGEN
Optimisation principle
• To start with the main insight that will come out of most
optimisation exercises:
• A consumer (as modeled by our previous assumptions) will
buy those quantities of goods that will
– Exhaust her budget
– Equate the MRS for the goods at the chosen bundle with
the price ratio of prices she faces for the different goods
• With exeptions of course such as ‘corner solutions’
PAGE 26
UNIVERSITY OF BERGEN
Two good graphical optimisation
𝐼 = 𝑝𝑥 ∙ 𝑥 + 𝑝𝑦 ∙ 𝑦
Quantity of y
𝐼 U3
𝑝𝑦 U2
U1 Optimum where slope of
𝑝
budget constraint, − 𝑝𝑥
𝑦
Is equal to the slope of indifference curve
𝑑𝑦
− 𝑑𝑥 ቚ = 𝑀𝑅𝑆 (of x for y)
𝑈=𝑘
PAGE 27 𝐼 Quantity of x
𝑝𝑥
UNIVERSITY OF BERGEN
Second order conditions
• This tangency condition is ‘neccesary’ for an optimum, but
not ‘sufficient’. Quantity of y
• A ‘real’ optimal
consumption bundle
also satisfies the
‘second order conditions’
• One important sufficient Second order conditions
condition would be if each not satisfied!
of the indifference curves at
all bundles satisfy the
condition of diminishing MRS
PAGE 28
Quantity of x
UNIVERSITY OF BERGEN
Corner solutions
• In certain cases, the Quantity of y
internal optimisation
solution where
𝑝
𝑀𝑅𝑆 = − 𝑥 does not exist.
𝑝𝑦
• In this case the optimal 𝐼
𝑝𝑦
consumption bundle will
be a ‘corner solution’ with
the consumption amounts
for some of the goods set
to zero
PAGE 29 𝐼 Quantity of x
𝑝𝑥
UNIVERSITY OF BERGEN
N-good optimisation using the Lagrange method
• The general optimisation problem for the consumer will
be to optimise 𝑈 𝑥1 , 𝑥2 , … , 𝑥𝑛 with all 𝑥𝑖 ≥ 0
• Subject to the budget constraint
𝐼 = 𝑝1 ∙ 𝑥1 + 𝑝2 ∙ 𝑥2 + ⋯ + 𝑝𝑛 ∙ 𝑥𝑛
• And for this general case the main solution method is
to optimise the Lagrangian expression:
ℒ = 𝑈 𝑥1 , 𝑥2 , … , 𝑥𝑛 + λ ∙ 𝐼 − 𝑝1 ∙ 𝑥1 − 𝑝2 ∙ 𝑥2 − ⋯ − 𝑝𝑛 ∙ 𝑥𝑛
PAGE 30
UNIVERSITY OF BERGEN
N-good optimisation using the Lagrange method
• ℒ = 𝑈 𝑥1 , 𝑥2 , … , 𝑥𝑛 + λ ∙ 𝐼 − 𝑝1 ∙ 𝑥1 − 𝑝2 ∙ 𝑥2 − ⋯ − 𝑝𝑛 ∙ 𝑥𝑛
• The first order conditions become:
𝜕ℒ 𝜕𝑈
– = 𝜕𝑥 − λ ∙ 𝑝1 = 0
𝜕𝑥1 1
𝜕ℒ 𝜕𝑈
– = − λ ∙ 𝑝2 = 0
𝜕𝑥2 𝜕𝑥2
⋮
𝜕ℒ 𝜕𝑈
– = − λ ∙ 𝑝𝑛 = 0
𝜕𝑥𝑛 𝜕𝑥𝑛
𝜕ℒ
– 𝜕λ
= 𝐼 − 𝑝1 ∙ 𝑥1 − 𝑝2 ∙ 𝑥2 − ⋯ − 𝑝𝑛 ∙ 𝑥𝑛 = 0
PAGE 31
UNIVERSITY OF BERGEN
N-good optimisation using the Lagrange method
• Now given that for each two goods i and j we have :
𝜕ℒ 𝜕𝑈 𝜕𝑈
– = − λ ∙ 𝑝𝑖 = 0 so = λ ∙ 𝑝𝑖
𝜕𝑥𝑖 𝜕𝑥𝑖 𝜕𝑥𝑖
𝜕ℒ 𝜕𝑈 𝜕𝑈
– 𝜕𝑥𝑗
= 𝜕𝑥 − λ ∙ 𝑝𝑗 = 0 so 𝜕𝑥𝑗
= λ ∙ 𝑝𝑗
𝑗
𝜕𝑈
ൗ𝜕𝑥
• Dividing both first order condition equations will give us: 𝜕𝑈 𝑖
=
ൗ𝜕𝑥𝑗
𝑝𝑖
𝑝𝑗
which is exactly the optimality condition
𝑑𝑥𝑗 𝑈 𝑥1 ,𝑥2 ,…,𝑥𝑛 𝑝
𝑀𝑅𝑆 = − 𝑑𝑥 ቚ = 𝑈𝑖 = 𝑝 𝑖 described
𝑖 𝑈 𝑥1 ,𝑥2 ,…,𝑥𝑛 =𝑘 𝑗 𝑥1 ,𝑥2 ,…,𝑥𝑛 𝑗
before.
PAGE 32
UNIVERSITY OF BERGEN
Interpretation of the λ
𝜕𝑈
ൗ𝜕𝑥
• λ= 𝑖
for each good 𝑖
𝑝𝑖
• So lambda is unique for every problem, and equal
to the marginal utility the consumer would get
when allowed to spend one more unit of income
on any of the goods.
PAGE 33
UNIVERSITY OF BERGEN
Corner solutions
• More generally, if we want to allow for corner solutions,
the n first order conditions become:
𝜕𝑈
− λ ∙ 𝑝𝑖 ≤ 0
𝜕𝑥𝑖
𝜕𝑈
• Where if we have for some good − λ ∙ 𝑝𝑖 < 0 then
𝜕𝑥𝑖
we will have in the optimum 𝑥𝑖 = 0
• In the two goods case this is exactly the intuition given
by the corner solution graph before
PAGE 34
Class question
UNIVERSITY OF BERGEN
Class question
• Assume a consumer with the utility function 𝑈 =
𝑥 + 2 ∙ 𝑦 having a budget of 𝐼 = 10 and facing
prices 𝑝𝑥 = 𝑝𝑦 = 1.
• Write down the Lagrangean optimisation, derive
the first order conditions and the optimal
consumption bundle 𝑥 ∗ , 𝑦 ∗
PAGE 36
UNIVERSITY OF BERGEN
Answer
• The Lagrangean becomes:
ℒ = 𝑥 + 2 ∙ 𝑦 + λ ∙ 10 − 𝑥 − 𝑦
• The first order conditions become
𝜕ℒ
=1−λ=0 →λ=1
𝜕𝑥
𝜕ℒ 1 1
= −λ=0 →λ=
𝜕𝑦 𝑦 𝑦
𝜕ℒ
= 10 − 𝑥 − 𝑦 = 0 → 𝑥 + 𝑦 = 10
𝜕λ
PAGE 37
UNIVERSITY OF BERGEN
Answer
• Using the first order conditions
λ=1
1
λ=
𝑦
𝑥 + 𝑦 = 10
• We get 𝑥 ∗ = 9 and 𝑦 ∗ = 1
PAGE 38
UNIVERSITY OF BERGEN
Demand functions and indirect utility
• The optimal consumption level you find for a given good 𝑖, 𝑥𝑖∗ will be a
function of the prices and the income available to the consumer
• 𝑥𝑖∗ = 𝑥𝑖∗ 𝑝1 , 𝑝2 , . . , 𝑝𝑛 , 𝐼
• If prices or consumer income change, optimal consumption 𝑥𝑖∗ will alter
along
• 𝑥𝑖∗ = 𝑥𝑖∗ 𝑝1 , 𝑝2 , . . , 𝑝𝑛 , 𝐼 is the demand function of good i by the
consumer
• Note that given that 𝑥𝑖∗ is the utility optimising consumption of good i,
V 𝑝1 , 𝑝2 , . . , 𝑝𝑛 , 𝐼 =
𝑈 𝑥1∗ 𝑝1 , 𝑝2 , . . , 𝑝𝑛 , 𝐼 , 𝑥2∗ 𝑝1 , 𝑝2 , . . , 𝑝𝑛 , 𝐼 , . . , 𝑥𝑛∗ 𝑝1 , 𝑝2 , . . , 𝑝𝑛 , 𝐼 should be
the maximum utility level the consumer could achieve given prices and
income 𝑝1 , 𝑝2 , . . , 𝑝𝑛 , 𝐼.
PAGE 39
UNIVERSITY OF BERGEN
Lump Sum Principle
• Under the preference assumptions we have used, the following
potentially ‘policy-guiding’ principle can be derived a result:
• Taxing an optimizing consumer with a ‘lump sum’ amount will
leave the consumer with a (weakly) higher utility than collecting
a similar amount of taxation by taxing the consumption of
specific goods.
• Alternatively and relatedly, ‘lump sum’ subsidies to a consumer
will more efficiently lead to higher levels of consumer utility than
good-specific subsidies
• Lump sum taxes and subsidies are in this case ‘non-
distortionary’ and leave a consumer free to optimize.
PAGE 40
UNIVERSITY OF BERGEN
