Microeconomics
Leopold S¨ogner c
Department of Economics and Finance
Institute for Advanced Studies
Josefst¨adter Straße 39
1080 Vienna
Masaryk University; October, 2017
Course Outline (1)
Microeconomics
Learning Objectives:
• This course covers key concepts of microeconomic theory. The
main goal of this course is to provide students with both, a basic
understanding and analytical traceability of these concepts.
• The main concepts are discussed in detail during the lectures. In
addition students have to work through the textbooks and have
to solve problems to improve their understanding and to acquire
skills to apply these tools to related problems.
1
Course Outline (2)
Microeconomics
Literature:
• Andreu Mas-Colell, A., Whinston, M.D., Green, J.R., Microeconomic Theory,
Oxford University Press, 1995. (MWG in the following)
Supplementary Literature:
• Gilboa, I., Theory of Decision under Uncertainty, Cambridge University Press,
2009.
• Gollier C., The Economics of Risk and Time, Mit Press, 2004.
• Gravelle, H. and Rees, R., Microeconomics, 3rd
edition, Prentice Hall, 2004.
(GR in the following)
• Jehle G.A. and P. J. Reny, Advanced Microeconomic Theory, Addison-Wesley
Series in Economics, Longman, Amsterdam, 2000.
• Ritzberger, K., Foundations of Non-Cooperative Game Theory, Oxford
University Press, 2002.
2
c ˇS´arka Hor´akov´a
3
Course Outline (3)
Microeconomics
• The slides you find in the university information system contain
(most of) the topics covered in MWG Chapters 1-6, 10 and 15
and 16.
• For the slides I’m grateful to comments by my colleges Egbert
Dierker and Martin Meier as well as to all former students in
Bruno and Vienna.
• For the course in Brno I’m grateful for comments from and
discussions with Rostislav Stanˇek, Josef Menˇsik, by Ondˇrej Krˇc´al,
and many others.
4
Course Outline (4)
Microeconomics
1. Decision theory and the theory of the consumer:
– Rationality, preference primitives and axioms, preference
representations and utility (MWG 1-3, GR 2A-B).
– Utility maximization, Walrasian demand and comparative
statics (MWG, 2, 3D, GR 2C-D).
– Indirect utility, expenditure function, Hicksian demand (MWG
3E,G, GR 3A).
– Slutsky equation, substitution and wealth effect (MWG 3 G,
GR 3B).
5
Course Outline (5)
Microeconomics
2. Production and cost:
– Production functions, returns to scale (GR 5).
– Production set (MWG 5 B).
– Cost minimization, conditional factor demands, cost function
(MWG 5 C, GR 6.A,B,E).
– Profit maximization, input demands, profit function, objectives
of the firm (MWG 5 C,G, GR 7.A,C,D).
6
Course Outline (6)
Microeconomics
3. General Equilibrium:
– Introduction, Walrasian equilibrium (MWG 15, GR 12.A-D).
– The Edgeworth box (MWG 15B, GR 12.E).
– Welfare theorems (MWG16 A-D, GR 13).
4. Decisions under uncertainty:
– Expected utility theorem, risk aversion (MWG 6A-C,GR 17
A-D).
7
Course Outline (7)
Microeconomics
• Winter Term 2015 (see university information system)
– First block: October 19-20, 2017.
– Second block: November 23-24, 2017.
– Third block: December 7-8, 2017.
• Contact hours (per semester): 12 units a 90 minutes.
• Practice session will be organized by Rostislav Stanˇek.
8
Course Outline (8)
Microeconomics
Some more comments on homework and grading:
• Final test (80%).
• Homework and practice session (20%).
• Final test and retakes: tba
9
Consumer Theory (1)
Rationality and Preferences
Microeconomics
• Consumption set X
• Rationality
• Preference relations and utility
• Choice correspondences and the weak axiom of revealed
preference
• Relationship between the axiomatic approach and the revealed
preference approach
MWG, Chapter 1.
10
Consumer Theory (1)
Rationality (1)
Microeconomics
• We consider agents/individuals and goods that are available for
purchase in the market.
• Definition: The set X of all possible mutually exclusive
alternatives (complete consumption plans) is called
consumption set or choice set.
• ”Simplest form of a consumption set”: We assume that each
good, xl ∈ X, l = 1, . . . , L can be consumed in infinitely divisible
units, i.e. xl ∈ R+. With L goods we get the commodity vector
x in the commodity space RL
+.
11
Consumer Theory (1)
Rationality (2)
Microeconomics
• Approach I: describe behavior by means of preference relations;
preference relation is the primitive characteristic of the individual.
• Approach II: the choice behavior is the primitive behavior of an
individual.
12
Consumer Theory 1
Rationality (3)
Microeconomics
• Consider the binary relation “at least good as“, abbreviated by
the symbol .
• For x, y ∈ X, x y implies that from a particular consumer’s
point of view x is preferred to y or that he/she is indifferent
between consuming x and y.
• From we derive the strict preference relation : x y if
x y but not y x and the indifference relation ∼ where
x y and y x.
13
Consumer Theory 1
Rationality (4)
Microeconomics
• Often we require that pair-wise comparisons of consumption
bundles are possible for all elements of X.
• Completeness: For all x, y ∈ X either x y, y x or both.
• Transitivity: For the elements x, y, z ∈ X: If x y and y z,
then x z
• Definition [D 1.B.1]: The preference relation is called rational
if it is complete and transitive.
• Remark: Reflexive x x follows from completeness.
14
Consumer Theory 1
Rationality (5)
Microeconomics
• Based on these remarks it follows that:
Proposition [P 1.B.1]: If is rational then,
– is transitive and irreflexive.
– ∼ is transitive, reflexive and symmetric.
– If x y z then x z.
15
Consumer Theory 1
Utility (1)
Microeconomics
• Definition : A function X → R is a utility function representing
if for all x, y ∈ X x y ⇔ u(x) ≥ u(y). [D 1.B.2]
• Does the assumption of a rational consumer imply that the
preferences can be represented by means of a utility function and
vice versa?
• Theorem: If there is a utility function representing , then
must be complete and transitive. [P 1.B.2]
• The other direction requires more assumptions on the preferences
- this comes later.
16
Consumer Theory 1
Choice Rules (1)
Microeconomics
• Now we follow the behavioral approach (vs. axiomatic approach),
where choice behavior is represented by means of a choice
structure.
• Definition - Choice structure: A choice structure (B, C(.))
consists: (i) of a family of nonempty subsets of X; with B ⊂ X.
The elements of B are the budget sets B. and (ii) a choice rule
C(.) that assigns a nonempty set of chosen elements C(B) ∈ B
for every B ∈ B.
• B is an exhaustive listing of all choice experiments that a
restricted situation can pose on the decision maker. It need not
include all subsets of X.
17
Consumer Theory 1
Choice Rules (2)
Microeconomics
• The choice rule assigns a set C(B) to every element B, i.e. it is
a correspondence.
• Example [1.C.1,(i)]: X = {x, y, z}. B = {B1, B2} with
B1 = {x, y} and B2 = {x, y, z}. C(B1) = {x} and
C(B2) = {x}.
• Example [1.C.1,(ii)]: X = {x, y, z}. B = {B1, B2} with
B1 = {x, y} and B2 = {x, y, z}. C(B1) = {x} and
C(B2) = {x, y}.
• Put structure on choice rule - weak axiom of revealed preference.
18
Consumer Theory 1
Choice Rules (3)
Microeconomics
• Definition - Weak Axiom of Revealed Preference: A choice
structure (B, C(.)) satisfies the weak axiom of revealed
preference if for some B ∈ B with x, y ∈ B we have x ∈ C(B),
then for any B ∈ B with x, y ∈ B and y ∈ C(B ) , we must
also have x ∈ C(B ). [D 1.C.1]
• I.e. if we have a budget set where both x and y are available and
x is chosen, then there cannot be a budget set containing both
bundles, for which y is chosen and x is not.
• Example: C({x, y}) = x, then C({x, y, z}) = y is not possible.
19
Consumer Theory 1
Choice Rules (4)
Microeconomics
• Example [1.C.1,(i)] satisfies the weak axiom, Example [1.C.1,(ii)]
does not.
• Definition - Revealed Preference Relation ∗
based on the
Weak Axiom: Given a choice structure (B, C(.)), ∗
is defined
by:
x ∗
y ⇔ there is some B ∈ B such that x, y ∈ B and
x ∈ C(B). [D 1.C.2]
• In other words: If x is revealed at least as good as y, then y
cannot be revealed preferred to x.
20
Consumer Theory 1
Choice Rules (5)
Microeconomics
• If a decision maker has a rational preference ordering , does the
choice structure satisfy the weak axiom of reveal preferences?
• Given a choice structure satisfying the weak axiom, does this
result in a rational preference relation?
• Definition: Choice structure implied by a rational preference
relation : C∗
(B, ) = {x ∈ B|x y for every y ∈ B}. Assume
that C∗
(B, ) is nonempty.
• Proposition: If is a rational preference relation, then the
choice structure C∗
(B, ) satisfies the weak axiom. [P 1.D.1]
21
Consumer Theory 1
Choice Rules (6)
Microeconomics
Proof:
• Suppose that x, y in some B ∈ B and x ∈ C∗
(B, ), then (i)
x y.
• We have to show that if x, y are in some B and y ∈ C∗
(B , ),
then also x ∈ C∗
(B , ).
• Suppose that x, y in B ∈ B and y ∈ C∗
(B , ), then (ii) y z
for all z ∈ B .
• The rational preference relation is transitive. Hence (i), (ii)
and transitivity imply that x y z. Therefore x ∈ C∗
(B, )
by the definition of this set.
22
Consumer Theory 1
Choice Rules (7)
Microeconomics
• Other direction. We want to know whether the choice rule C(.)
satisfying the weak axiom is equal to C∗
(., ) for some rational
preference relation .
• Definition: Given a choice rule C(.). rationalizes the choice
rule if C(B) = C∗
(B, ) for all B ∈ B. [D 1.D.1]
• Proposition: If a choice structure (B, C(.)) satisfies the weak
axiom of revealed preference and B includes all subsets of X of
up to three elements, then there is a rational preference relation
that rationalizes C(.) relative to B. That is,
C(B) = C∗
(B, ) for all B ∈ B. The rational preference relation
is the unique preference relation that does so. [P 1.D.2]
23
Consumer Theory 1
Choice Rules (8)
Microeconomics
• Example [1.D.1]: Counterexample that choice structure cannot
be rationalized if B does not contain all subsets up to three
elements. X = {x, y, z}. B = {B1, B2, B3} with B1 = {x, y},
B2 = {y, z} and B3 = {x, z}. C(B1) = {x}, C(B2) = {y} and
C(B3) = {z}. Here the weak axiom is satisfied. To rationalize
this choice structure it would be necessary that x y, y z and
z x, which does not satisfy the transitivity property.
24
Consumer Theory 1
Choice Rules (9)
Microeconomics
• Remark: The more subsets of X we consider, the stronger the
restrictions implied by the weak axiom.
• Remark: If the choice structure is defined for all subsets of X,
then the approaches based on the preference relation and on the
weak axiom are equivalent. To consider all budget sets is a strong
requirement. Another way to get equivalence, is to replace the
weak axiom by the strong axiom of revealed preference.
25
Consumer Theory 2
Walrasian Demand (1)
Microeconomics
• Competitive budget sets.
• Walrasian/Marshallian Demand
• Walras’ law, Cournot and Engel aggregation.
MWG: Chapter 2
26
Consumer Theory 2
Consumption Set (1)
Microeconomics
• We have already defined the consumption set: The set of all
alternatives (complete consumption plans). We assumed
X = RL
+.
• Each x represents a different consumption plan.
• Physical restrictions: divisibility, time constraints, survival needs,
etc. might lead to a strict subset of RL
+ as consumption set X.
27
Consumer Theory 2
Budget Set (1)
Microeconomics
Definition - Budget Set: B
• Due to constraints (e.g. income) we cannot afford all elements in
X, problem of scarcity.
• The budget set B is defined by the elements of X, which are
achievable given the economic realities.
• B ⊂ X.
28
Consumer Theory 2
Budget Set (2)
Microeconomics
• By the consumption set and the budget set we can describe a
consumer’s alternatives of choice.
• These sets do not tell us what x is going to be chosen by the
consumer.
• To describe the choice of the consumer we need a theory to
model or describe the preferences of a consumer (or the choice
structure).
29
Consumer Theory 2
Competitive Budgets (1)
Microeconomics
• Assumption: All L goods are traded in the market (principle of completeness),
the prices are given by the price vector p, pl > 0 for all l = 1, . . . , L.
Notation: p 0. Assumption - the prices are constant and not affected by
the consumer.
• Given a wealth level w ≥ 0, the set of affordable bundles is described by
p · x = p1x1 + · · · + pLxL ≤ w.
• Definition - Walrasian Budget Set: The set Bp,w = {x ∈ RL
+|p · x ≤ w}
is called Walrasian or competitive budget set. [D 2.D.1]
• Definition - Consumer’s problem: Given p and w choose the optimal
bundle x from Bp,w.
30
Consumer Theory 2
Competitive Budgets (2)
Microeconomics
• Definition - Relative Price: The ratios of prices pj/pi are called
Relative Prices.
• Here the price of good j is expressed in terms of good i. In other
words: The price of good xi is expressed in the units of good xj.
• On the market we receive for one unit of xj, pj/pi · 1 units of xi.
• Example: pj = 4, pi = 2. Then pj/pi = 2 and we get two units of
xi for one unit of xj.
31
Consumer Theory 2
Competitive Budgets (3)
Microeconomics
• The budget set B describes the goods a consumer is able to buy
given wealth level w.
• Definition - Numeraire Good: If all prices pj are expressed in
the prices of good n, then this good is called numeraire. pj/pn,
j = 1, . . . , L. The relative price of the numeraire is 1.
• There are L − 1 relative prices.
32
Consumer Theory 2
Competitive Budgets (4)
Microeconomics
• The set {x ∈ RL
+|p · x = w} is called budget hyperplane, for
L = 2 it is called budget line.
• Given x and x in the budget hyperplane, p · x = p · x = w
holds. This results in p(x − x ) = 0, i.e. p and (x − x ) are
orthogonal - see Figure 2.D.3 page 22.
• The budget hyperplane is a convex set. In addition it is closed
and bounded ⇒ compact. 0 ∈ Bp,w (given the assumption that
p 0).
33
Consumer Theory 2
Demand Functions (1)
Microeconomics
• Definition - Walrasian demand correspondence: The
correspondence assigning to a pair (p, w) a set of consumption
bundles is called Walrasian demand correspondence x(p, w);
i.e. (p, w) x(w, p). If x(p, w) is single valued for all p, w,
x(., .) is called demand function.
• Definition - Homogeneity of degree zero: x(., .) is
homogeneous of degree zero if x(αp, αw) = x(p, w) for all p, w
and α > 0. [D 2.E.1]
• Definition - Walras law, budget balancedness: x(., .) satisfies
Walras law if for every p 0 and w > 0, we get p · x = w for all
x ∈ x(p, w). That is, the consumer spends all income w with
her/his optimal consumption decision. [D 2.E.2]
34
Consumer Theory 2
Demand Functions (2)
Microeconomics
• Assume that x(., .) is a function:
• With p fixed at ¯p, the function x(¯p, .) is called Engel function.
• If the demand function is differentiable we can derive the gradient
vector: Dwx(p, w) = (∂x1(p, w)/∂w, . . . , ∂xL(p, w)/∂w). If
∂xl(p, w)/∂w ≥ 0, xl is called normal or superior, otherwise it is
inferior.
• See Figure 2.E.1, page 25
• Notation: Dwx(p, w) results in a 1 × L row matrix,
Dwx(p, w) = ( wx(p, w)) .
35
Consumer Theory 2
Demand Functions (3)
Microeconomics
• With w fixed, we can derive the L × L matrix of partial
derivatives with respect to the prices: Dpx(p, w).
• ∂xl(p, w)/∂pk = [Dpx(p, w)]l,k is called the price effect.
• A Giffen good is a good where the own price effect is positive,
i.e. ∂xl(p, w)/∂pl > 0
• See Figure 2.E.2-2.E.4, page 26.
36
Consumer Theory 2
Demand Functions (4)
Microeconomics
• Proposition: If a Walrasian demand function x(., .) is
homogeneous of degree zero and differentiable, then for all p and
w:
L
k=1
∂xl(p, w)
∂pk
pk +
∂xl(p, w)
∂w
w = 0
for l = 1, . . . , L;, or in matrix notation
Dpx(p, w)p + Dwx(p, w) = 0. [P 2.E.1]
• Proof: By the Euler theorem (if g(.) is homogeneous of degree r,
then ∂g(x)/∂x · x = rg(x), [MWG, Theorem M.B.2, p. 929]),
the result follows directly when using the stacked vector
x = (p , w) . Apply this to x1(p, w), . . . , xL(p, w).
37
Consumer Theory 2
Demand Functions (5)
Microeconomics
• Definition - Price Elasticity of Demand: ηij = ∂xi(p,w)
∂pj
pj
xi(p,w) .
• Definition - Income Elasticity: ηiw = ∂xi(p,w)
∂w
w
xi(p,w) .
• Definition - Income Share:
si =
pixi(p, w)
w
,
where si ≥ 0 and
n
i=1 si = 1.
38
Consumer Theory 3
The Axiomatic Approach (1)
Microeconomics
• Axioms on preferences
• Preference relations, behavioral assumptions and utility (axioms,
utility functions)
• The consumer’s problem
• Walrasian/Marshallian Demand
MWG, Chapter 3.A-3.D
39
Consumer Theory 3
The Axiomatic Approach (2)
Microeconomics
• Axiom 1 - Completeness: For all x, y ∈ X either x y, y x
or both.
• Axiom 2 - Transitivity: For the elements x, y, z ∈ X: If x y
and y z, then x z.
• We have already defined a rational preference relation by
completeness and transitivity [D 3.B.1].
• If the number of elements is finite it is easy to see that one can
describe a preference relation by means of a function.
40
Consumer Theory 3
The Axiomatic Approach (3)
Microeconomics
Sets arising from the preference relations:
• (x) := {y|y ∈ X, y x} - at least as good (sub)set
• (x) := {y|y ∈ X, y x} - the no better set
• (x) := {y|y ∈ X, y x} - at preferred to set
• (x) := {y|y ∈ X, y x} - worse than set
• ∼ (x) := {y|y ∈ X, y ∼ x} - indifference set
41
Consumer Theory 3
The Axiomatic Approach (4)
Microeconomics
• Axiom 3.A - Local Nonsatiation: For all x ∈ X and for all
ε > 0 there exists some y ∈ X such that ||x − y|| ≤ ε and y x.
[D 3.B.3]
• This assumptions implies that for every small distance ε there
must exist at least one y, which is preferred to x.
• Indifference “zones“ are excluded by this assumption. See
Figure 3.B.1 on page 43.
42
Consumer Theory 3
The Axiomatic Approach (5)
Microeconomics
• Axiom 3.B - Monotonicity: For all x, y ∈ RL
+: If x ≥ y then
x y while if x y then x y (weakly monotone). It is
strongly/strict monotone if x ≥ y and x = y imply x y.
[D 3.B.2]
• Here ≥ means that at least one element of x is larger than the
elements of y, while x y implies that all elements of x are
larger than the elements of y.
• Remark: Local nonsatiation vs. monotonicity: The latter implies
that more is always better, while Axiom 3.A only implies that in
a set described by ||x − y|| ≤ ε there has to exist a preferred
alternative.
43
Consumer Theory 3
The Axiomatic Approach (6)
Microeconomics
• Discuss the differences of Axioms 3.A and 3.B (what are their
impacts on indifference sets?), e.g. by means Figures 3.B.1 and
3.B.2, page 43.
44
Consumer Theory 3
The Axiomatic Approach (7)
Microeconomics
• Last assumption on taste - “mixtures are preferred to extreme
realizations“
• See Figure 3.B.3, page 44.
• Axiom 4.A - Convexity: For every x ∈ X, if y x and z x
then νy + (1 − ν)z x for ν ∈ [0, 1]. [D 3.B.4]
• Axiom 4.B - Strict Convexity: For every x ∈ X, y x, z x
and y = z then νy + (1 − ν)z x for ν ∈ (0, 1). [D 3.B.5]
• Given these assumptions, indifference curves become (strict)
convex.
45
Consumer Theory 3
The Axiomatic Approach (8)
Microeconomics
• Definition - Homothetic Preferences: A monotone preference
on X is homothetic if all indifference sets are related by
proportional expansions along rays. I.e. x ∼ y then αx ∼ αy.
[D 3.B.6]
• Definition - Quasilinear Preferences: A monotone preference
on X = (−∞, ∞) × RL−1
is quasilinear with respect to
commodity 1 if : (i) all indifference sets are parallel displacements
of each other along the axis of commodity 1. That is, if x ∼ y
then x + αe1 ∼ y + αe1 and e1 = (1, 0, . . . ). (ii) Good one is
desirable: x + αe1 x for all α > 0. [D 3.B.7]
46
Consumer Theory 3
The Axiomatic Approach (9)
Microeconomics
• With the next axiom we regularize our preference order by
making it continuous:
• Axiom 5 - Continuity: A preference order is continuous if it
is preserved under limits. For any sequence (x(n)
, y(n)
) with
x(n)
y(n)
for all n, and limits x, y (x = limn→∞ x(n)
and
y = limn→∞ y(n)
) we get x y. [D 3.C.1]
• Equivalently: For all x ∈ X the set “at least as good as“
( (x)) and the set “no better than“ ( (x)) are closed in X.
47
Consumer Theory 3
The Axiomatic Approach (10)
Microeconomics
• Topological property of the preference relation (important
assumption in the existence proof of a utility function).
• By this axiom the set (x) and (x) are open sets (the
complement of a closed set is open ...). (x) is the complement
of X\ (x).
• The intersection of (x)∩ (x) is closed (intersection of closed
sets). Hence, indifference sets are closed.
• Consider a sequence of bundles y(n)
fulfilling y(n)
x, for all n.
For y(n)
converging to y, Axiom 5 imposes that y x.
48
Consumer Theory 3
The Axiomatic Approach (11)
Microeconomics
Lexicographic order/dictionary order:
• Given two partially order sets X1 and X2, an order is called
lexicographical on X1 × X2 if (x1, x2) (x1, x2) if and only if
x1 < x1 (or x1 = x1, x2 < x2). That is, “good 1 is infinitely
more desired than good 2”.
• Example in R2
+ (Example of Debreu): x = (0, 1) and
y(n)
= (1/n, 0), y = (0, 0). For all n, y(n)
x, while for n → ∞:
y(n)
→ y= (0, 0) (0, 1) = x.
• The lexicographic ordering is a rational (strict) preference
relation (we have to show completeness and transitivity).
49
Consumer Theory 3
The Axiomatic Approach (12)
Microeconomics
• Axioms 1 and 2 guarantee that an agent is able to make
consistent comparisons among all alternatives.
• Axiom 5 imposes the restriction that preferences do not exhibit
“discontinuous behavior”; mathematically important.
• Axioms 3 and 4 make assumptions on a consumer’s taste
(satiation, mixtures).
50
Consumer Theory 3
Utility Function (1)
Microeconomics
• Definition: Utility Function: A real-valued function
u : RL
+ → R is called utility function representing the preference
relation if for all x, y ∈ RL
+ u(x) ≥ u(y) if and only if x y.
• I.e. a utility function is a mathematical device to describe the
preferences of a consumer.
• Pair-wise comparisons are replaced by comparing real valued
functions evaluated for different consumption bundles.
• Function is of no economic substance (for its own).
51
Consumer Theory 3
Utility Function (2)
Microeconomics
• First of all we want to know if such a function exists.
• Theorem: Existence of a Utility Function: If a binary relation
is complete, transitive and continuous, then there exists a
continuous real valued function u(x) representing the preference
ordering .
• Proof: by assuming monotonicity see page 47; Debreu’s (1959)
proof is more advanced.[P 3.C.1]
52
Consumer Theory 3
Utility Function (3)
Microeconomics
• Consider y = u(x) and the transformations v = g(u(x));
v = log y, v = y2
, v = a + by, v = −a − by (see MWG, page
49). Do these transformations fulfill the properties of a utility
function?
• Theorem: Invariance to Positive Monotonic
Transformations: Consider the preference relation and the
utility function u(x) representing this relation. Then also v(x)
represents if and only if v(x) = g(u(x)) is strictly increasing
on the set of values taken by u(x).
53
Consumer Theory 3
Utility Function (4)
Microeconomics
Proof:
• ⇒ Assume that x y with u(x) ≥ u(y): A strictly monotone
transformation g(.) then results in g(u(x)) ≥ g(u(y)). I.e. v(x) is
a utility function describing the preference ordering of a
consumer.
54
Consumer Theory 3
Utility Function (5)
Microeconomics
Proof:
• ⇐ Now assume that g(u(x)) is a utility representation, but g is
not strictly positive monotonic on the range of u(.): Then
g(u(x)) is not > to g(u(y)) for some pair x, y where
u(x) > u(y). Hence, for the pair x, y we have x y since
u(x) > u(y), but g(u(x)) ≤ g(u(y)).
This contradicts the assumption that v(.) = g(u(.)) is a utility
representation of .
55
Consumer Theory 3
Utility Function (6)
Microeconomics
• By Axioms 1,2,5 the existence of a utility function is guaranteed.
By the further Axioms the utility function exhibits the following
properties.
• Theorem: Preferences and Properties of the Utility
Function:
– u(x) is strictly increasing if and only if is strictly monotonic.
– u(x) is quasiconcave if and only if is convex:
u(xν
) ≥ min{u(x), u(y)}, where xν
= νx + (1 − ν)y.
– u(x) is strictly quasiconcave if and only if is strictly convex.
That is, u(xν
) > min{u(x), u(y)} for xν
= νx + (1 − ν)y,
x = y and ν ∈ (0, 1).
56
Consumer Theory 3
Utility Function (7)
Microeconomics
• Definition: Indifference Curve: Bundles where utility is constant (in R2
).
• Marginal rate of substitution and utility: Assume that u(x) is differentiable,
then
du(x1, x2) =
∂u(x1, x2)
∂x1
dx1 +
∂u(x1, x2)
∂x2
dx2 = 0
dx2
dx1
= −
∂u(x1, x2)/∂x1
∂u(x1, x2)/∂x2
MRS12 =
∂u(x1, x2)/∂x1
∂u(x1, x2)/∂x2
.
• The marginal rate of substitution describes the trade-off between goods 1 and
2 that marginally keep the consumer indifferent at a given consumption
bundle (x1, x2). That is, the “amount of good 2” the consumer has to obtain
for giving up “one unit of good 1” while staying at the same utility level.
57
Consumer Theory 3
Utility Function (8)
Microeconomics
• If u(x) is differentiable and the preferences are strictly
monotonic, then marginal utility is strictly positive.
• With strictly convex preferences the marginal rate of substitution
is a strictly decreasing function (i.e. in R2
the slope of the
indifference curve becomes flatter).
• For a quasiconcave utility function (i.e.
u(xν
) ≥ min{u(x1
), u(x2
)}, with xν
= νx1
+ (1 − ν)x2
) and its
Hessian H(u(x))D2
(u(x)) we get: yH(u(x))y ≤ 0 for all
vectors y, where grad(u(x)) · y = 0. That is, when moving from
x to y that is tangent to the indifference surface at x utility does
not increase (decreases if the equality is strict).
58
Consumer Theory 3
Consumer’s Problem (1)
Microeconomics
• The consumer is looking for a bundle x∗
such that x∗
∈ B and
x∗
x for all x in the feasible set B.
• Assume that the preferences are complete, transitive, continuous,
strictly monotonic and strictly convex. Then can be
represented by a continuous, strictly increasing and strictly
quasiconcave utility function. Moreover we can assume that we
can take first and second partial derivatives of u(x). These are
usual assumptions, we can also solve the utility maximization
problem (UMP) with less stringent assumptions.
59
Consumer Theory 3
Consumer’s Problem (1)
Microeconomics
• We assume prices pi > 0, p = (p1, . . . , pL) is the vector of prices.
We assume that the prices are fixed from the consumer’s point of
view. (Notation: p 0 means that all coordinates of p are
strictly larger than zero.)
• The consumer is endowed with wealth w.
60
Consumer Theory 3
Consumer’s Problem (2)
Microeconomics
• Budget set induced by w: Bp,w = {x|x ∈ RL
+ ∧ p · x ≤ w}.
• With the constant w and the consumption of the other goods
constant, we get:
dw = p1dx1 + p2dx2 = 0
dx2
dx1
= −
p1
p2
with other prices constant .
• Budget line with two goods; slope −p1/p2. See Figure 2.D.1,
page 21, MWG.
61
Consumer Theory 3
Consumer’s Problem (3)
Microeconomics
• Definition - Utility Maximization Problem [UMP]: Find the
optimal solution for:
max
x
u(x) s.t. xi ≥ 0 , p · x ≤ w.
The solution x(p, w) is called Walrasian demand .
• Remark: Some textbooks call the UMP also Consumer’s
Problem.
• Remark: Some textbooks call x(p, w) Marshallian demand.
62
Correspondences (1)
Microeconomics
• Generalized concept of a function.
• Definition - Correspondence: Given a set A ∈ Rn
, a
correspondence f : A Rk
is a rule that assigns a set
f(x) ⊆ Y ⊂ Rk
to every x ∈ A.
• If f(x) contains exactly one element for every x ∈ A, then (up to
abuse of notation) f is a function.
• A ⊆ Rn
and Y ⊆ Rk
are the domain and the codomain.
• Literture: e.g. MWG, chapter M.H, page 949.
63
Correspondences (2)
Microeconomics
• The set {(x, y)|x ∈ A , y ∈ Rk
, y ∈ f(x)} is called graph of
the correspondence.
• Definition - Closed Graph: A correspondence has a closed
graph if for any pair of sequences x(m)
→ x ∈ A, with x(m)
∈ A
and y(m)
→ y, with y(m)
∈ f(x(m)
), we have y ∈ f(x).
64
Correspondences (3)
Microeconomics
• Regarding continuity there are two concepts with
correspondences.
• Definition - Upper Hemicontinuous: A correspondence is UHC
if the graph is closed and the images of compact sets are
bounded. That is, for every compact set B ⊆ A, the set
f(B) = {y ∈ Rk
: y ∈ f(x) for some x ∈ B} is bounded.
• Definition - Lower Hemicontinuous: Given A ⊆ Rn
and a
compact set Y ⊆ Rk
, the correspondence is LHC if for every
sequence x(m)
→ x, x(m)
, x ∈ A for all m, and every y ∈ f(x),
we can find a sequence y(m)
→ y and an integer M such that
y(m)
∈ f(x(m)
) for m > M.
65
Consumer Theory 3
Consumer’s Problem (4)
Microeconomics
• Proposition - Existence: If p 0, w > 0 and u(x) is
continuous, then the utility maximization problem has a solution.
[P 3.D.1]
• Proof: By the assumptions Bp,w is compact. u(x) is a continuous
function. By the Weierstraß theorem (Theorem M.F.2(ii),
p. 945, MWG; maximum value theorem in calculus), there exists
an x ∈ Bp,w maximizing u(x).
66
Consumer Theory 3
Consumer’s Problem (5)
Microeconomics
• Theorem - Properties of x(p, w): Consider a continuous utility
function u(x) representing a rational locally nonsatiated
preference relation defined on the consumption set X = RL
+.
Then x(p, w) has the following properties: [P 3.D.2]
– Homogeneity of degree zero in (p, w).
– Walras’ law: p · x = w for all x ∈ x(p, w).
– Convexity/uniqueness: If is convex, so that u(x) is
quasiconcave, then x(p, w) is a convex set. If is strictly
convex, where u(x) is strictly quasiconcave, then x(p, w)
consists of a single element.
67
Consumer Theory 3
Consumer’s Problem (6)
Microeconomics
Proof:
• Property 1 - Homogeneity in p, w: We have to show that
x(µp, µw) = µ0
x(p, w). Plug in µp and µw in the optimization
problem ⇒ Bp,w = Bµp,µw. The result follows immediately.
68
Consumer Theory 3
Consumer’s Problem (7)
Microeconomics
Proof:
• Property 2- Walras’ law: If x ∈ x(p, w) and p · x < w, then there
exists a y in the neighborhood of x, with y x and p · y < w by
local nonsatiation. Therefore x cannot be an optimal bundle.
This argument holds for all interior points of Bp,w.
69
Consumer Theory 3
Consumer’s Problem (8)
Microeconomics
Proof:
• Property 3 - x(p, w) is a convex set: If preferences are convex
then u(xν
) ≥ min{u(x), u(y)}, where xν
= νx + (1 − ν)y;
replace ≥ by > if is strictly convex. I.e. u(x) is quasiconcave.
We have to show that if x, y ∈ x(p, w), then xν
∈ x(p, w). From
the above property x, y and xν
have to be elements of the
budget hyperplane {x|x ∈ X and p · x = w}.
Since x and y solve the UMP we get u(x) = u(y), therefore
u(xν
) ≤ u(x) = u(y). By quasiconcavity of u(x) we get
u(xν
) ≥ u(x) = u(y), such that u(xν
) = u(x) = u(y) holds for
arbitrary x, y ∈ x(p, w). I.e. the set x(p, w) has to be convex.
70
Consumer Theory 3
Consumer’s Problem (9)
Microeconomics
Proof:
• Property 3 - x(p, w) is single valued if preferences are strictly
convex: Assume, like above, the x and y solve the UMP; x = y.
Then u(x) = u(y) ≥ u(z) for all z ∈ Bp,w. By the above result
x, y are elements of the budget hyperplane.
• Since preferences are strictly convex, u(x) is strictly quasiconcave
⇒ u(xν
) > min{u(x ), u(y )}. xν
= νx + (1 − ν)y and x , y
are some arbitrary elements of the budget hyperplane; (a
contradiction to strict convexity).
• Now u(xν
) > min{u(x), u(y)}, also for x, y. Therefore the pair
x, y, where x = y, cannot solve the UMP (contradiction to x, y
solve UMP). Therefore, x(p, w) has to be single valued.
71
Consumer Theory 3
Consumer’s Problem (10)
Microeconomics
• By altering the price vector p and income w, the consumer’s
maximization provides us with the correspondence x(p, w), which
is called Walrasian demand correspondence. If preferences are
strictly convex we get Walrasian demand functions x(p, w).
• What happens to the function if w or pj changes? See MWG,
Figure 3.D.1 - 3.D.4
72
Consumer Theory 3
Consumer’s Problem (11)
Microeconomics
• In a general setting demand need not be a smooth function.
• Theorem - Differentiable Walrasian Demand Function: Let
x∗
0 solve the consumers maximization problem at price
p0 0 and w0 > 0. If u(x) is twice continuously differentiable,
∂u(x)/∂xi > 0 for some i = 1, . . . , n and the bordered Hessian
of u(x),
D2
u(x) u(x)
u(x) 0
,
has a non-zero determinant at x∗
, then x(p, w) is differentiable
at p0, w0.
• More details are provided in MWG, p. 94-95.
73
Consumer Theory 4
Duality
Microeconomics
• Instead of looking at u(x), we’ll have an alternative look on
utility via prices, income and the utility maximization problem ⇒
indirect utility
• Expenditure function, the dual problem and Hicksian demand
• Income- and substitution effects, Slutsky equation
MWG, Chapter 3.D-3.H
74
Consumer Theory 4
Indirect Utility (1)
Microeconomics
• We have already considered the direct utility function u(x) in the
former parts.
• Start with the utility maximization problem
max
x
u(x) s.t. p · x ≤ w
x∗
∈ x(p, w) solves this problem for (p, w) 0.
• Definition - Indirect Utility: By the highest levels of utility
attainable with p, w, we define a maximal value function. This
function is called indirect utility function v(p, w). It is the
maximum value function corresponding to the consumer’s
optimization problem (utility maximization problem).
75
Consumer Theory 4
Indirect Utility (2)
Microeconomics
• v(p, w) is a function, by Berge’s theorem of the maximum
x(p, w) is upper hemicontinuous and v(p, w) is continuous (see
MWG, page 963, [M.K.6]).
• If u(x) is strictly quasiconcave such that maximum x∗
is unique,
we derive the demand function x∗
= x(p, w).
• In this case the indirect utility function is the composition of the
direct utility function and the demand function x(p, w), i.e.
v(p, w) = u(x∗
) = u(x(p, w)).
76
Consumer Theory 4
Indirect Utility (3)
Microeconomics
• Theorem: Properties of the Indirect Utility Function
v(p, w): [P 3.D.3] Suppose that u(x) is a continuous utility
function representing a locally nonsatiated preference relation
on the consumption set X = RL
+. Then the indirect utility
function v(p, w) is
– Continuous in p and w.
– Homogeneous of degree zero in p, w.
– Strictly increasing in w.
– Nonincreasing in pl, l = 1, . . . , L.
– Quasiconvex in (p, w).
77
Consumer Theory 4
Indirect Utility (4)
Microeconomics
Proof:
• Property 1 - Continuity: follows from Berge’s theorem of the
maximum.
• Property 2 - Homogeneous in (p, w): We have to show that
v(µp, µw) = µ0
v(p, w) = v(p, w); µ > 0. Plug in µp and µw in
the optimization problem ⇒
v(µp, µw) = {maxx u(x) s.t. µp · x ≤ µw} ⇔
{maxx u(x) s.t. p · x ≤ w} = v(p, w).
78
Consumer Theory 4
Indirect Utility (5)
Microeconomics
Proof:
• Property 3 - increasing in w: Given the solutions of the UMP
with p and w, w , where w > w: x(p, w) and x(p, w ).
• The corresponding budget sets are Bp,w and Bp,w , by
assumption Bp,w ⊂ Bp,w (here we have a proper subset).
• Define Sp,w = {x ∈ X|p · x = w} (Walrasian budget
hyperplane). Then Bp,w is still contained in Bp,w \ Sp,w .
• Therefore also Sp,w ∈ (Bp,w \ Sp,w ). From the above
consideration we know that for any y ∈ Sp,w, we have p · y < w .
By local nonsatiation there are better bundles in Bp,w .
79
Consumer Theory 4
Indirect Utility (6)
Microeconomics
Proof:
• Since v(p, w) is a maximal value function, it has to increase if w
increases.
• In other words: By local nonsatiation Walras law has to hold, i.e.
x(p, w) and x(p, w ) are subsets of the budget hyperplanes
{x|x ∈ X and p · x = w}, {x|x ∈ X and p · x = w },
respectively. We know where we find the optimal bundles. The
hyperplane for w is a subset of Bp,w and Bp,w (while the
hyperplane for w is not contained in Bp,w). Interior points
cannot be an optimum under local nonsatiation.
• If v(p, w) is differentiable this result can be obtained by means of
the envelope theorem.
80
Consumer Theory 4
Indirect Utility (7)
Microeconomics
Proof:
• Property 4 - non-increasing in pl: W.l.g. pl > pl, then we get
Bp,w and Bp ,w, where Bp ,w ⊆ Bp,w. But Sp ,w is not fully
contained in Bp,w \ Sp,w. (Observe the ”common point” in R2
.)
The rest is similar to Property 3.
81
Consumer Theory 4
Indirect Utility (8)
Microeconomics
Proof:
• Property 5 - Quasiconvex: Consider two arbitrary pairs p1
, x1
and
p2
, x2
and the convex combinations pν
= νp1
+ (1 − ν)p2
and
wν = νw1 + (1 − ν)w2; ν ∈ [0, 1].
• v(p, w) would be quasiconvex if
v(pν
, wν) ≤ max{v(p1
, w1), v(p2
, w2)}.
• Define the consumption sets: Bj = {x|p(j)
· x ≤ wj} for
j = 1, 2, ν.
82
Consumer Theory 4
Indirect Utility (9)
Microeconomics
Proof:
• First we show: If x ∈ Bν, then x ∈ B1 or x ∈ B2.
This statement trivially holds for ν equal to 0 or 1.
For ν ∈ (0, 1) we get: Suppose that x ∈ Bν but x ∈ B1 or
x ∈ B2 is not true (then x ∈ B1 and x ∈ B2), i.e.
p1
· x > w1 ∧ p2
· x > w2
Multiplying the first term with ν and the second with 1 − ν
results in
νp1
· x > νw1 ∧ (1 − ν)p2
· x > (1 − ν)w2 83
Consumer Theory 4
Indirect Utility (10)
Microeconomics
Proof:
• Summing up both terms results in:
(νp1
+ (1 − ν)p2
) · x = pν
· x > νw1 + (1 − ν)w2 = wν
which contradicts our assumption that x ∈ Bν.
• From the fact that xν ∈ x(pν
, wν) is either ∈ B1 or ∈ B2, it
follows that v(pν
, wν) ≤ max{v(p1
, w1), v(p2
, w2)}. The last
expression corresponds to the definition of a quasiconvex
function.
84
Consumer Theory 4
Expenditure Function (1)
Microeconomics
• With indirect utility we looked at maximized utility levels given
prices and income.
• Now we raise the question a little bit different: what expenditures
e are necessary to attain an utility level u given prices p.
• Expenditures e can be described by the function e = p · x.
85
Consumer Theory 4
Expenditure Function (2)
Microeconomics
• Definition - Expenditure Minimization Problem [EMP]:
minx p · x s.t. u(x) ≥ u, x ∈ X = RL
+, p 0. (We only look at
u ≥ u(0), such that U = {u|u ≥ u(0) ∧ u ∈ Range(u(x))}.)
• It is the dual problem of the utility maximization problem. The
solution of the EMP h(p, u) will be called Hicksian demand
correspondence.
• Definition - Expenditure Function: The minimum value
function e(p, u) solving the expenditure minimization problem
minx p · x s.t. u(x) ≥ u, p 0, is called expenditure function.
• Existence: The Weierstraß theorem guarantees the existence of
an x∗
s.t. p · x∗
are the minimal expenditures necessary to attain
an utility level u.
86
Consumer Theory 4
Expenditure Function (3)
Microeconomics
• Theorem: Properties of the Expenditure Function e(p, u):
[P 3.E.2]
If u(x) is continuous utility function representing a locally
nonsatiated preference relation. Then the expenditure function
e(p, u) is
– Continuous in p, u domain Rn
++ × U.
– ∀p 0 strictly increasing in u.
– Non-decreasing in pl for all l = 1, . . . , L.
– Concave in p.
– Homogeneous of degree one in p.
87
Consumer Theory 4
Expenditure Function (5)
Microeconomics
Proof:
• Property 1 - continuous: Apply the theorem of the maximum.
88
Consumer Theory 4
Expenditure Function (6)
Microeconomics
Proof:
• Property 2 - increasing in u: We have to show that if u2 > u1
then e(p, u2) > e(p, u1).
• Suppose that h1 ∈ h(p, u1) and h2 ∈ h(p, u2) solve the EMP for
u2 and u1, but e(p, u2) ≤ e(p, u1). We show that this results in a
contradiction (i.e. u2 > u1 but 0 ≤ p · h2 ≤ p · h1).
• Then by continuity of u(x) and local nonsatiation we can find an
α ∈ (0, 1) such that αh2 is preferred to h1 (remember u2 > u1 is
assumed) with expenditures αp · h2 < p · h2 ≤ p · h1. This
contradicts that h1 solves the EMP for p, u1.
89
Consumer Theory 4
Expenditure Function (7)
Microeconomics
Proof:
• Property 2 - with calculus: From
minx p · x s.t. u(x) ≥ u , x ≥ 0 we derive the Lagrangian:
L(x, λ) = p · x + λ(u − u(x)) .
• From this Kuhn-Tucker problem we get:
∂L
∂xi
= pi − λ
∂u(x)
∂xi
≥ 0 ,
∂L
∂xi
xi = 0
∂L
∂λ
= u − u(x) ≤ 0 ,
∂L
∂λ
λ = 0
90
Consumer Theory 4
Expenditure Function (8)
Microeconomics
Proof:
• λ = 0 would imply that utility could be increased without
increasing the expenditures (in an optimum) ⇒ u = u(x) and
λ > 0.
• Good xi is demanded if the price does not exceed λ∂u(x)
∂xi
for all
xi > 0.
• The envelope theorem tells us that
∂e(p, u)
∂u
=
∂L(x, u)
∂u
= λ > 0
• Since u(x) is continuous and increasing the expenditure function
has to be unbounded.
91
Consumer Theory 4
Expenditure Function (9)
Microeconomics
Proof:
• Property 3 - non-decreasing in pl: similar to property 2.
92
Consumer Theory 4
Expenditure Function (10)
Microeconomics
Proof:
• Property 4 - concave in p: Consider an arbitrary pair p1
and p2
and the convex combination pν = νp1
+ (1 − ν)p2
. The
expenditure function is concave if
e(pν, u) ≥ νe(p1
, u) + (1 − ν)e(p2
, u).
• For minimized expenditures it has to hold that p1
x1
≤ p1
x and
p2
x2
≤ p2
x for all x fulfilling u(x) ≥ u.
• x∗
ν minimizes expenditure at a convex combination of p1
and p2
.
• Then p1
x1
≤ p1
x∗
ν and p2
x2
≤ p2
x∗
ν have to hold.
93
Consumer Theory 4
Expenditure Function (11)
Microeconomics
Proof:
• Multiplying the first term with ν and the second with 1 − ν and
taking the sum results in νp1
x1
+ (1 − ν)p2
x2
≤ pνx∗
ν.
• Therefore the expenditure function is concave in p.
94
Consumer Theory 4
Expenditure Function (12)
Microeconomics
Proof:
• Property 5 - homogeneous of degree one in p: We have to show
that e(µp, u) = µ1
e(p, u); µ > 0. Plug in µp in the optimization
problem ⇒ e(µp, u) = {minx µp · x s.t. u(x) ≥ u}. Objective
function is linear in µ, the constraint is not affected by µ. With
calculus we immediately see the µ cancels out in the first order
conditions ⇒ xh
remains the same ⇒
µ{minx p · x s.t. u(x) ≥ u} = µe(p, u).
95
Consumer Theory 4
Hicksian Demand (1)
Microeconomics
• Theorem: Hicksian demand: [P 3.E.3] Let u(x) be a
continuous utility function representing a locally nonsatiatated
preference order; p 0. Then the Hicksian demand
correspondence has the following properties:
– Homogeneous of degree zero.
– No excess utility u(x) = u.
– Convexity/uniqueness: If is convex, then h(p, u) is a convex
set. If is strictly convex, then h(p, u) is single valued.
96
Consumer Theory 4
Hicksian Demand (2)
Microeconomics
Proof:
• Homogeneity follows directly from the EMP.
min{p · x s.t. u(x) ≥ u} ⇔ α min{p · x s.t. u(x) ≥ u} ⇔
min{αp · x s.t. u(x) ≥ u} for α > 0.
• Suppose that there is an x ∈ h(p, u) with u(x) > u. By the
continuity of u we find an α ∈ (0, 1) such that x = αx and
u(x ) > u. But with x we get p · x < p · x. A contradiction that
x solves the EMP.
• For the last property see the theorem on Walrasian demand or
apply the forthcoming theorem.
97
Consumer Theory 4
Expenditure vs. Indirect Utility (1)
Microeconomics
• With (p, w) the indirect utility function provides us with the
maximum of utility u. Suppose w = e(p, u). By this definition
v(p, e(p, u)) ≥ u.
• Given p, u and an the expenditure function, we must derive
e(p, v(p, w)) ≤ w.
• Given an x∗
solving the utility maximization problem, i.e.
x∗
∈ x(p, w). Does x∗
solve the EMP if u = v(p, w)?
• Given an h∗
solving the EMP, i.e. h∗
∈ h(p, u). Does h∗
solve
the UMP if w = e(p, u)?
98
Consumer Theory 4
Expenditure vs. Indirect Utility (2)
Microeconomics
• Theorem: Equivalence between Indirect Utility and
Expenditure Function: [P 3.E.1] Let u(x) be continuous utility
function representing a locally nonsatiatated preference order;
p 0.
– If x∗
is optimal in the UMP with w > 0, then x∗
is optimal in
the EMP when u = u(x∗
). e(p, u(x∗
)) = w.
– If h∗
is optimal in the EMP with u > u(0), then h∗
is optimal
in the UMP when w = e(p, u). v(p, e(p, u)) = u.
99
Consumer Theory 4
Expenditure vs. Indirect Utility (3)
Microeconomics
Proof:
• We prove e(p, v(p, w)) = w by means of a contradiction. p, w
∈ Rn
++ × R++. By the definition of the expenditure function we get
e(p, v(p, w)) ≤ w. In addition h∗
∈ h(p, u).
To show equality assume that e(p, u) < w, where u = v(p, w) and x∗
solves the UMP: e(p, u) is continuous in u. Choose ε such that
e(p, u + ε) < w and e(p, u + ε) =: wε.
The properties of the indirect utility function imply v(p, wε) ≥ u + ε. Since
wε < w and v(p, w) is strictly increasing in w (by local nonstatiation) we
get: v(p, w) > v(p, wε) ≥ u + ε but u = v(p, w), which is a
contradiction. Therefore e(p, v(p, w)) = w and x∗
also solves the EMP,
such that x∗
∈ h(p, u) when u = v(p, u).
100
Consumer Theory 4
Expenditure vs. Indirect Utility (4)
Microeconomics
Proof:
• Next we prove v(p, e(p, u)) = u in the same way. p, u
∈ Rn
++ × U. By the definition of the indirect utility function we
get v(p, e(p, u)) ≥ u.
Assume that v(p, w) > u, where w = e(p, u) and h∗
solves the
EMP: v(p, w) is continuous in w. Choose ε such that
v(p, w − ε) > u and v(p, w − ε) =: uε.
The properties of the expenditure function imply
e(p, uε) ≤ w − ε. Since uε > u and e(p, u) is strictly increasing
in u we get: e(p, u) < e(p, uε) ≤ w − ε but w = e(p, u), which is
a contradiction. Therefore v(p, e(p, u)) = u. In addition h∗
also
solves the UMP.
101
Consumer Theory 4
Hicksian Demand (3)
Microeconomics
• Theorem: Hicksian/ Compensated law of demand: [P 3.E.4]
Let u(x) be continuous utility function representing a locally
nonsatiatated preference order and h(p, u) consists of a single
element for all p 0. Then the Hicksian demand function
satisfies the compensated law of demand: For all p and p :
(p − p )[h(p , u) − h(p , u)] ≤ 0.
102
Consumer Theory 4
Hicksian Demand (4)
Microeconomics
Proof:
• By the EMP: p · h(p , u) − p · h(p , u) ≤ 0 and
p · h(p , u) − p · h(p , u) ≤ 0 have to hold.
• Adding up the inequalities yields the result.
103
Consumer Theory 4
Shephard’s Lemma (1)
Microeconomics
• Investigate the relationship between a Hicksian demand function
and the expenditure function.
• Theorem - Shephard’s Lemma: [P 3.G.1] Let u(x) be
continuous utility function representing a locally nonsatiatated
preference order and h(p, u) consists of a single element. Then
for all p and u, the gradient vector of the expenditure function
with respect to p gives Hicksian demand, i.e.
pe(p, u) = h(p, u).
104
Consumer Theory 4
Shephard’s Lemma (2)
Microeconomics
Proof by means of calculus:
• Suppose that the envelope theorem can be applied (see e.g.
MWG [M.L.1], page 965):
• Then the Lagrangian is given by: L(x, λ) = p · x + λ(u − u(x)).
• λ > 0 follows from u = u(x).
105
Consumer Theory 4
Shephard’s Lemma (3)
Microeconomics
Proof with calculus:
• The Kuhn-Tucker conditions are:
∂L
∂xi
= pi − λ
∂u(x)
∂xi
≥ 0
∂L
∂xi
xi = 0
∂L
∂λ
= u − u(x) ≥ 0
∂L
∂λ
λ = 0
106
Consumer Theory 4
Shephard’s Lemma (4)
Microeconomics
Proof with calculus:
• Good xi is demanded if the price does not exceed λ∂u(x)
∂xi
for all
xi > 0.
• The envelope theorem tells us that
∂e(p, u)
∂pl
=
∂L(x, u)
∂pl
= hl(p, u)
for l = 1, . . . , L.
107
Consumer Theory 4
Shephard’s Lemma (5)
Microeconomics
Proof:
• The expenditure function is the support function µk of the
non-empty and closed set K = {x|u(x) ≥ u}. Since the solution
is unique by assumption, µK(p) = pe(p, u) = h(p, u) has to
hold by the Duality theorem.
• Alternatively: Assume differentiability and apply the envelope
theorem.
108
Consumer Theory 4
Expenditure F. and Hicksian Demand (1)
Microeconomics
• Furthermore, investigate the relationship between a Hicksian
demand function and the expenditure function.
• Theorem:: [P 3.G.1] Let u(x) be continuous utility function
representing a locally nonsatiatated and strictly convex
preference relation on X = RL
+. Suppose that h(p, u) is
continuously differentiable, then
– Dph(p, u) = D2
pe(p, u)
– Dph(p, u) is negative semidefinite
– Dph(p, u) is symmetric.
– Dph(p, u)p = 0.
109
Consumer Theory 4
Expenditure F. and Hicksian Demand (2)
Microeconomics
Proof:
• To show Dph(p, u)p = 0, we can use the fact that h(p, u) is
homogeneous of degree zero in prices (r = 0).
• By the Euler theorem [MWG, Theorem M.B.2, p. 929] we get
L
l=1
∂h(p, u)
∂pl
pl = rh(p, u).
110
Consumer Theory 4
Walrasian vs. Hicksian Demand (1)
Microeconomics
• Here we want to analyze what happens if income w changes:
normal vs. inferior good.
• How is demand effected by prices changes: change in relative
prices - substitution effect, change in real income - income effect
• Properties of the demand and the law of demand.
• How does a price change of good i affect demand of good j.
• Although utility is continuous and strictly increasing, there might
be goods where demand declines while the price falls.
111
Consumer Theory 4
Walrasian vs. Hicksian Demand (2)
Microeconomics
• Definition - Substitution Effect, Income Effect: We split up
the total effect of a price change into
– an effect accounting for the change in the relative prices pi/pj
(with constant utility or real income) ⇒ substitution effect.
Here the consumer will substitute the relatively more expensive
good by the cheaper one.
– an effect induced by a change in real income (with constant
relative prices) ⇒ income/wealth effect.
112
Consumer Theory 4
Walrasian vs. Hicksian Demand (3)
Microeconomics
• Hicksian decomposition - keeps utility level constant to identify
the substitution effect.
• The residual between the total effect and the substitution effect
is the income effect.
• See Figures in MWG, Chapter 2 and
113
Consumer Theory 4
Walrasian vs. Hicksian Demand (4)
Microeconomics
• Here we observe that the Hicksian demand function exactly
accounts for the substitution effect.
• The difference between the change in Walrasian (total effect)
demand induced by a price change and the change in Hicksian
demand (substitution effect) results in the income effect.
• Note that the income effect need not be positive.
114
Consumer Theory 4
Walrasian vs. Hicksian Demand (5)
Microeconomics
• Formal description of these effects is given by the Slutsky
equation.
• Theorem - Slutsky Equation: [P 3.G.3] Assume that the
consumer’s preference relation is complete, transitive,
continuous, locally nonsatiated and strictly convex defined on
X = RL
+. Then for all (p, w) and u = v(p, w) we have
∂xl(p, w)
∂pj
T E
=
∂hl(p, u)
∂pj
SE
−xj(p, w)
∂xl(p, w)
∂w
IE
l, j = 1, . . . , L.
115
Consumer Theory 4
Walrasian vs. Hicksian Demand (6)
Microeconomics
• Equivalently:
Dph(p, u) = Dpx(p, w) + Dwx(p, w)x(p, w)
• Remark: In the following proof we shall assume that h(p, u) and
x(p, w) are differentiable. (Differentiability of h(p, u) follows
from duality theory presented in Section 3.F.)
116
Consumer Theory 4
Walrasian vs. Hicksian Demand (7)
Microeconomics
Proof:
• First, we use the Duality result on demand:
hl(p, u) = xl(p, e(p, u)) and take partial derivatives with respect
to pj:
∂hl(p, u)
∂pj
=
∂xl(p, e(p, u))
∂pj
+
∂xl(p, e(p, u))
∂w
∂e(p, u)
∂pj
.
• Second: By the relationship between the expenditure function
and the indirect utility it follows that u = v(p, w) and
e(p, u) = e(p, v(p, w)) = w.
117
Consumer Theory 4
Walrasian vs. Hicksian Demand (8)
Microeconomics
Proof:
• Third: Shephard’s Lemma tells us that ∂e(p,u)
∂pj
= hj(p, u), this
gives
∂hl(p, u)
∂pj
=
∂xl(p, w)
∂pj
+
∂xl(p, w)
∂w
hj(p, u)
118
Consumer Theory 4
Walrasian vs. Hicksian Demand (9)
Microeconomics
Proof:
• Forth: Duality between Hicksian and Walrasian demand implies
that h(p, v(p, w)) = x(p, w) with v(p, w) = u. Thus
∂e(p,u)
∂pj
= xj(p, w).
• Arranging terms yields:
∂xl(p, w)
∂pj
=
∂hl(p, u)
∂pj
− xj(p, w)
∂xl(p, w)
∂w
.
119
Consumer Theory 4
Walrasian vs. Hicksian Demand (10)
Microeconomics
• From the Sultsky equation we can construct the following matrix:
Definition - Slutsky Matrix:
S(p, w) :=





∂x1(p,w)
∂p1
+ x1(p, w)
∂x1(p,w)
∂w · · ·
∂x1(p,w)
∂pL
+ xL(p, w)
∂x1(p,w)
∂w
. . . ... . . .
∂xL(p,w)
∂p1
+ x1(p, w)
∂xL(p,w)
∂w · · ·
∂xL(p,w)
∂pL
+ xL(p, w)
∂xL(p,w)
∂w





120
Consumer Theory 4
Walrasian vs. Hicksian Demand (11)
Microeconomics
• Theorem Suppose that e(p, u) is twice continuously
differentiable. Then the Slutsky Matrix S(p, w) is negative
semidefinite, symmetric and satisfies S(p, w)p = 0.
121
Consumer Theory 4
Walrasian vs. Hicksian Demand (12)
Microeconomics
Proof:
• Negative semidefiniteness follows from the negative
semidefiniteness of Dph(p, u) which followed from the concavity
of the expenditure function.
• Symmetry follows from the existence of the expenditure function
and Young’s theorem.
• S(p, w) · p = 0 follows from an Euler theorem (see [MWG,
Theorem M.B.2, p. 929]) argument already used in [P 3.G.2]
122
Consumer Theory 4
Roy’s Identity (1)
Microeconomics
• Goal is to connect Walrasian demand with the indirect utility
function.
• Theorem - Roy’s Identity: [P 3.G.4] Let u(x) be continuous
utility function representing a locally nonsatiatated and strictly
convex preference relation defined on X = RL
+. Suppose that
the indirect utility function v(p, w) is differentiable for any
p, w 0, then
x(p, w) = −
1
wv(p, w)
pv(p, w)
i.e.
xl(p, w) = −
∂v(p, w)/∂pl
∂v(p, w)/∂w
.
123
Consumer Theory 4
Roy’s Identity (1)
Microeconomics
Proof:
• Roy’s Identity: Assume that the envelope theorem can be applied
to v(p, w).
• Let (x∗
, λ∗
) maximize {maxx u(x) s.t. p · x ≤ w} then the
partial derivatives of the Lagrangian L(x, λ) with respect to pl
and w provide us with:
∂v(p, w)
∂pl
=
∂L(x∗
, λ∗
)
∂pl
= −λ∗
x∗
l , l = 1, . . . , L.
∂v(p, w)
∂w
=
∂L(x∗
, λ∗
)
∂w
= λ∗
.
124
Consumer Theory 2
Indirect Utility (11)
Microeconomics
Proof:
• Plug in −λ from the second equation results in
∂v(p, w)
∂pl
= −
∂v(p, w)
∂w
x∗
l
such that
−
∂v(p, w)/∂pl
∂v(p, w)/∂w
= xl(p, w).
• Note that ∂v(p, w)/∂w by our properties on the indirect utility
function.
125
Theorem of the Maximum (1)
Microeconomics
• Consider a constrained optimization problem:
max f(x) s.t. g(x, q) = 0
where q ∈ Q is a vector of parameters. Q ∈ RS
and x ∈ RN
.
f(x) is assumed to be continuous. C(q) is the constraint set
implied by g.
126
Theorem of the Maximum (1)
Microeconomics
• Definition: x(q) is the set of solutions of the problem, such that
x(q) ⊂ C(q) and v(q) is the maximum value function, i.e. f(x)
evaluated at an optimal x ∈ x(q).
• Theorem of the Maximum: Suppose that the constraint
correspondence is continuous and f is continuous. Then the
maximizer correspondence x : Q → RN
is upper hemicontinuous
and the value function v : Q → R is continuous. [T M.K.6], page
963.
127
Duality Theorem (1)
Microeconomics
• Until now we have not shown that c(w, q) or e(p, u) is
differentiable when u(x) is strictly quasiconcave.
• This property follows from the Duality Theorem.
• MWG, Chapter 3.F, page 63.
128
Duality Theorem (2)
Microeconomics
• A set is K ∈ Rn
is convex if αx + (1 − α)y ∈ K for all x, y ∈ K
and α ∈ [0, 1].
• A half space is a set of the form {x ∈ Rn
|p · x ≥ c}.
• p = 0 is called the normal vector: if x and x fulfill
p · x = p · x = c, then p · (x − x ) = 0.
• The boundary set {x ∈ Rn
|p · x = c} is called hyperplane. The
half-space and the hyperplane are convex.
129
Duality Theorem (3)
Microeconomics
• Assume that K is convex and closed. Consider ¯x /∈ K. Then
there exists a half-space containing K and excluding ¯x. There is
a p and a c such that p · ¯x < c ≤ p · x for all x ∈ K (separating
hyperplane theorem).
• Basic idea of duality theory: A closed convex set can be
equivalently (dually) described by the intersection of half-spaces
containing this set.
• MWG, figure 3.F.1 and 3.F.2 page 64.
130
Duality Theorem (4)
Microeconomics
• If K is not convex the intersection of the half-spaces that
contain K is the smallest, convex set containing K. (closed
convex hull of K, abbreviated by ¯K).
• For any closed (but not necessarily convex) set K we can define
the support function of K:
µK(p) = inf{p · x|x ∈ K}
• When K is convex the support function provides us with the dual
description of K.
• µK(p) is homogeneous of degree one and concave in p.
131
Duality Theorem (5)
Microeconomics
• Theorem - Duality Theorem: Let K be a nonempty closed set
and let µK(p) be its support function. Then there is a unique
¯x ∈ K such that ¯p · ¯x = µK(¯p) if and only if µK(p) is
differentiable at ¯p. In this case pµK(¯p) = ¯x.
• Proof see literature. E.g. see section 25 in R.T. Rockafellar,
Convex Analysis, Princeton University Press, New York 1970.
132
Consumer Theory 4
Revealed Preference Theory (1)
Microeconomics
• Weak Axiom of Revealed Preference.
• Strong Axiom of Revealed Preference.
• Revealed preferences and utility maximization.
MWG, Chapter 1.C and 2.F., 3.J.
133
Consumer Theory 4
Revealed Preference Theory (2)
Microeconomics
• Samuelson’s idea: Cannot we start with observed behavior
instead of assumptions on preferences.
• Idea: if a consumer buys a bundle x0
instead of an other
affordable bundle x1
, then the first bundle is called revealed
preferred to x1
(see Consumer Theory 1).
• Definition - Weak Axiom on Revealed Preference: [D 2.F.1]
A Walrasian demand function x(p, w) satisfies the weak axiom of
revealed preference if for any two wealth price situations (p,w)
and (p’,w’) the following relationship holds: If p · x(p , w ) ≤ w
and x(p , w ) = x(p, w) then p · x(p, w) > w .
134
Consumer Theory 4
Revealed Preference Theory (3)
Microeconomics
• Interpret the weak axiom by means for Figure 2.F.1, page 30.
• We assume that x(p, w) is a function, which is homogeneous of
degree zero and Walras’ law holds.
135
Consumer Theory 4
Revealed Preference Theory (4)
Microeconomics
• From the former parts we already know:
Theorem - Weak Revealed Preference and Utility
maximization: If x(p, w) solves the utility maximization problem
with strictly increasing and strictly quasiconcave utility function,
then the weak axiom of revealed preference has to hold.
• See also P 1.D.1
136
Consumer Theory 4
Revealed Preference Theory (5)
Microeconomics
Proof:
• Consider a pair x0
and x1
where x0
= x(p0
, w) solves the utility
maximization problem for p0
, x1
for p1
.
• Assume u(x0
) > u(x1
): w = p0
· x0
≥ p0
· x1
. Then
p1
· x0
> p1
· x1
= w. Otherwise a consumer would have chosen
x0
if it were affordable in the second maximization problem.
• I.e. p1
· x0
> p1
· x1
has to be fulfilled. Since x0
and x1
are
arbitrary pairs, the weak axiom of revealed preference has to hold.
137
Consumer Theory 4
Slutsky Compensation (1)
Microeconomics
• Definition - Slutsky compensation: Given a bundle
x0
= x(p, w) and income is compensated such that the consumer
can always buy the bundle x0
, i.e. w = p · x(p, w). Then
demand is called Slutsky compensated demand xS
(p, w(x0
)).
• Discuss this concept by means of Figure 2.F.2, page 31.
138
Consumer Theory 4
Slutsky Compensation (2)
Microeconomics
• Proposition: Suppose that the Walrasian demand function
x(p, w) is homogeneous of degree zero and satisfies Walras’ law.
Then x(p, w) satisfies the weak axiom if and only if the following
property holds:
For any compensated price change form the initial situation
(p, w) to a new pair (p , w ), where w = p · x(p, w), we have
(p − p) · [x(p , w ) − x(p, w)] ≤ 0
with strict inequality whenever x(p, w) = x(p , w ). [P 2.F.1]
• Remark: x(p , w ) = xS
(p , w(x0
)).
139
Consumer Theory 4
Slutsky Compensation (3)
Microeconomics
Proof:
• (i) The weak axiom implies (p − p) · [x(p , w ) − x(p, w)] ≤ 0
with strict inequality for different demands: If x(p , w ) = x(p, w)
then [x(p , w ) − x(p, w)] = 0.
• Suppose x(p , w ) = x(p, w) and expand
(p − p) · [x(p , w ) − x(p, w)] to
p · [x(p , w ) − x(p, w)] − p · [x(p , w ) − x(p, w)]. By Walras’ law
and the construction of compensated demand the first term is 0.
• By compensated demand we get p · x(p, w) = w . I.e.
x0
= x(p, w) can be bought with p , w . By the weak axiom
x(p , w ) /∈ Bp,w, such that p · x(p , w ) > w. Walras’ law implies
p · x(p, w) = w. This yields p · [x(p , w ) − x(p, w)] > 0, such
that ... holds. 140
Consumer Theory 4
Slutsky Compensation (4)
Microeconomics
Proof:
• (ii) (p − p) · [x(p , w ) − x(p, w)] ≤ 0 implies the weak law if
x(p , w ) = x(p, w):
• If we consider compensated demand, then the weak axiom has to
hold (replace Walrasian demand by compensated demand in the
Theorem - Weak Revealed Preference and Utility maximization.
• It is necessary that the weak axiom holds for all compensated
demand changes: Assume u(x0
) > u(x1
): w = p0
· x0
≥ p0
· x1
.
Suppose that p1
· x0
≤ p1
· x1
= w. Then x0
cannot be an
optimum by local non-satiation.
• By these arguments the weak law holds if p · x(p , w ) > w
whenever p · x(p, w) = w and x(p , w ) = x(p, w). 141
Consumer Theory 4
Slutsky Compensation (5)
Microeconomics
Proof:
• By this argument we can test for the weak axiom by looking at
compensated price changes. (We show that ¬H ⇒ ¬C.) If the
weak law does not hold, there is a compensated price change
such that p x(p, w ) ≤ w ⇒ p x(p, w) ≤ w , < for different x
(p · x(p, w) = w). By Walras law we get
p · [x(p , w ) − x(p, w)] ≤ 0
and
p · [x(p , w ) − x(p, w)] ≥ 0.
• This results in (p − p) · [x(p , w ) − x(p, w)] ≥ 0; > for
x(p , w ) = x(p, w). This contradicts that
(p − p) · [x(p , w ) − x(p, w)] ≤ 0 holds. 142
Consumer Theory 4
Revealed Preference Theory (6)
Microeconomics
• Definition - Strong Axiom of Revealed Preference: [3.J.1]
The market demand satisfies the strong axiom of revealed
preference if for any list
(p1
, w1
), . . . , (pN
, wN
)
with x(pn+1
, wn+1
) = x(pn
, wn
) for all n ≤ N − 1, we have
pN
x(p1
, w1
) > wN
whenever pn
· x(pn+1
, wn+1
) ≤ wn
for all
n ≤ N − 1.
• I.e. if x(p1
, w1
) is directly or indirectly revealed preferred to
x(pN
, wN
), then x(pN
, wN
) cannot be directly or indirectly be
revealed preferred to x(p1
, w1
). Or for different bundles
x1
, x2
, . . . , xk
: If xq
is revealed preferred to x2
and x2
is preferred
to x3
, then x1
is revealed preferred to x3
. 143
Consumer Theory 4
Revealed Preference Theory (7)
Microeconomics
• Theorem - Revealed Preference and Demand (II): If the
Walrasian demand function x(p, w) satisfies the strong axiom of
revealed preference then there is a rational preference relation
that rationalizes x(p, w). I.e. for all (p, w), x(p, w) y for every
y = x(p, w) with y ∈ Bp,w. [P 3.J.1].
• Proof - see page 92.
144
Consumer Theory 5
Welfare Analysis (1)
Microeconomics
• Measurement of Welfare
• Concept of the Equivalent Variation, the Compensating Variation
and the Consumer Surplus.
• Pareto improvement and Pareto efficient
Literature: MWG, Chapter 3.I, page 80-90.
145
Consumer Theory 5
Welfare Analysis (2)
Microeconomics
• From a social point of view - can we judge that some market
outcomes are better or worse?
• Positive question: How will a proposed policy affect the welfare
of an individual?
• Normative question: How should we weight different effects on
different individuals?
146
Consumer Theory 5
Welfare Analysis (3)
Microeconomics
• Definition - Pareto Improvement: When we can make
someone better off and no one worse off, then a Pareto
improvement can be made.
• Definition - Pareto Efficient: A situation where there is no way
to make somebody better off without making someone else worse
off is called Pareto efficient. I.e. there is no way for Pareto
improvements.
• Strong criterion.
147
Consumer Theory 5
Consumer Welfare Analysis (1)
Microeconomics
• Preference based consumer theory investigates demand from a
descriptive perspective.
• Welfare Analysis can be used to perform a normative analysis.
• E.g. how do changes of prices or income affect the well being of
a consumer.
148
Consumer Theory 5
Consumer Welfare Analysis (2)
Microeconomics
• Given a preference relation and Walrasian demand x(p, w), a
price change from p0
to p1
increases the well-being of a
consumer if indirect utility increases. I.e. v(p1
, w) > v(p0
, w).
• Here we are interested in so called money metric indirect
utility functions. E.g. expressing indirect utility in terms of
monetary units.
149
Consumer Theory 5
Consumer Welfare Analysis (3)
Microeconomics
• Suppose u1 > u0, u1 = v(p1
, w) arises from p1
, w and
u0 = v(p0
, w) from p0
, w.
• With p fixed at ¯p, the property of the expenditure function that
e(p, u) is increasing in u yields:
e(¯p, u1)) = e(¯p, v(¯p, ˜w1
)) = ˜w1
> e(¯p, v(¯p, ˜w0
)) = e(¯p, u0) = ˜w0
- i.e. it is an indirect utility function which measures the degree
of well-being in money terms.
• See Figure 3.I.1, page 81.
150
Consumer Theory 5
Consumer Welfare Analysis (4)
Microeconomics
• Based on these considerations we set ¯p = p0
or p1
and
w = e(p0
, u0
) = e(p1
, u1
); we define:
– Definition - Equivalent Variation: “old prices”
EV (p0
, p1
, w) = e(p0
, u1
)−e(p0
, u0
) = e(p0
, u1
)−e(p1
, u1
) = e(p0
, u1
)−w
– Definition - Compensating Variation: “new prices”
CV (p0
, p1
, w) = e(p1
, u1
)−e(p1
, u0
) = e(p0
, u0
)−e(p1
, u0
) = w−e(p1
, u0
)
151
Consumer Theory 5
Consumer Welfare Analysis (5)
Microeconomics
• EV measures the money amount that a consumer is indifferent
between accepting this amount and the status after the price
change (i.e. to attain a utility level u1
).
• CV measures the money amount a consumer is willing to pay to
induce the price change from p0
to p1
(i.e. to obtain utility level
u0
at the new price p1
). This money amount can be negative as
well.
• Discuss Figure 3.1.2, page 82; if p1 falls then the consumer is
prepared to pay the amount CV , i.e. CV > 0.
152
Consumer Theory 5
Consumer Welfare Analysis (6)
Microeconomics
• Both measures are associated with Hicksian demand.
• Suppose the only p1 changes, then p0
1 = p1
1 and p0
l = p1
l for
l ≥ 2. With w = e(p0
, u0
) = e(p1
, u1
) and
h1(p, u) = ∂e(p, u)/∂p1 we get
EV (p0
, p1
, w) =
p0
1
p1
1
h1((p1, p−), u1
)dp1
CV (p0
, p1
, w) =
p0
1
p1
1
h1((p1, p−), u0
)dp1
153
Consumer Theory 5
Consumer Welfare Analysis (7)
Microeconomics
• Discuss these integrals - MWG, Figure 3.1.3, page 83. Here the following case
is considered. p0
and p1
are L dimensional price vectors. Only the first
component p1 is changed. The other prices p− := (p2, . . . , pL) are kept
constant. w is constant as well.
• EV, CV increase if utility increases and vice versa.
• If x1 is a normal good, then the slope of the Walrasian demand function
x1(p, w) is smaller than the slopes of h1(p, .) (in absolute terms).
• We get EV (p0
, p1
, w) > CV (p0
, p1
, w) if the good is normal (in absolute
value), the converse is true for inferior goods.
• EV (p0
, p1
, w) = CV (p0
, p1
, w) with zero income effect for good 1. This
is the case with quasilinear preferences for good two (see [D 3.B.7]).
154
Consumer Theory 5
Consumer Welfare Analysis (8)
Microeconomics
• EV (p0
, p1
, w) = CV (p0
, p1
, w) with zero income effect for good
1.
In this case EV (p0
, p1
, w) = CV (p0
, p1
, w) is also equal to the
change in Marshallian Consumer Surplus.
• Definition - Marshallian Consumer Surplus:
MCSl(p, w) =
∞
p
xl((pl, p−), w)dpl
• Definition - Area Variation:
AV (p0
, p1
, w) =
p0
l
p1
l
x(pl, p−, w)dpl.
155
Consumer Theory 5
Area Variation Measure (1)
Microeconomics
• Definition - Area Variation:
AV (p0
, p1
, w) =
p0
1
p1
1
x(p1, p−, w)dp1.
• It measures the change in Marshallian consumer surplus.
• If the income effect is zero this measure corresponds to EV and
CV . (see Marshallian Consumer Surplus)
• The argument that AV provides are good approximation of EV
or CV can but need not hold. See MWG, Figure 3.1.8, page 90.
Jehle/Reny, 1st edition, Theorem 6.3.2, page 278: Willing’s
upper and lower bounds on the difference between EV and CV.
156
Consumer Theory 5
Partial Information (1)
Microeconomics
• Consider a bundle x0
, price vectors p0
, p1
and wealth w. Often a
complete Walrasian demand function cannot be observed,
however:
• Theorem - Welfare and Partial Information I: Consider a
consumer with complete, transitive, continuous, and locally
non-satiated preferences. If (p1
− p0
) · x0
< 0, then the consumer
is strictly better of with (p1
, w) compared to (p0
, w). [P 3.I.1]
157
Consumer Theory 5
Partial Information (2)
Microeconomics
Proof:
• With non-satiation the consumer chooses a set on the boundary
of the budget set, such that p0
· x = w. Then p1
· x < w.
• ⇒ x is affordable within the budget set under p1
. By the
assumption of local non-satiation, there exists a closed set with
distance ≤ ε including a better bundle which remains within the
budget set. Then the consumer is strictly better off with p1
.
158
Consumer Theory 5
Partial Information (3)
Microeconomics
• What happens if (p1
− p0
) · x0
> 0 ? This implies
(αp1
+ (1 − α)p0
− p0
) · x0
> 0 for α > 0.
• Theorem - Welfare and Partial Information II: Consider a
consumer with a twice differentiable expenditure function. If
(p1
− p0
) · x0
> 0, then there exists an ¯α ∈ (0, 1) such that for
all 0 < α ≤ ¯α, we have e((1 − α)p0
+ αp1
), u0
) > w, and so the
consumer is strictly better off under p0
, w than under
(1 − α)p0
+ αp1
, w. [P 3.I.2]
159
Consumer Theory 5
Partial Information (4)
Microeconomics
Proof:
• We want to show that CV is negative, if we move from p0
to p1
.
Let pα
= (1 − α)p0
+ αp1
. We want to show that
CV = e(p0
, u0
) − e(pα
, u0
) < 0 for some ¯α ≥ α > 0. In other
words e(pα
, u0
) − e(p0
, u0
) > 0.
• Taylor expand e(p, u) at p0
, u0
:
e(pα
, u0
) = e(p0
, u0
) + (pα
− p0
) pe(p0
, u0
) + R(p0
, pα
)
where R(p0
, pα
)/||pα
− p0
|| → 0 if pα
→ p0
. e(., .) has to be at
least C1
. (fulfilled since second derivatives are assumed to exist).
160
Consumer Theory 5
Partial Information (5)
Microeconomics
Proof:
• By the properties of this approximation, there has to exist an ¯α,
where the Lagrange residual can be neglected. Then
sgn(e(pα
, u0
) − e(p0
, u0
)) = sgn (pα
− p0
) pe(p0
, u0
) for all
α ∈ [0, ¯α].
• This results in e(pα
, u0
) − e(p0
, u0
) > 0 by the assumption that
(pα
− p0
) pe(p0
, u0
) > 0 and the fact that
pe(p0
, u0
) = h(p0
, u0
) = x(p0
, e(p0
, u0
)).
161
Consumer Theory 5
Partial Information (6)
Microeconomics
• Remark: Note that with a differentiable expenditure function the
second order term is non-positive, since the expenditure function
is concave.
• Remark: We can show the former theorem also in this way
(differentiability assumptions have to hold in addition). There the
non-positive second order term does not cause a problem, since
there we wanted to show that e(p1
, u0
) − e(p0
, u0
) < 0 if
(p1
− p0
) · x0
< 0.
162
Production 1
Motivation
Microeconomics
• Production
• Production possibility sets and the production function
• Marginal product, marginal rate of substitution and returns to
scale.
MasColell, Chapter 5
163
Production 1
Firms (1)
Microeconomics
• In this section we treat the firm as a black box. We abstract from
ownership, management, organization, etc.
• Assumption: A firm maximizes its profit.
• How can we justify this assumption?
164
Production 1
Production Possibility Set (1)
Microeconomics
• Definition - Production: The process of transforming inputs to
outputs is called production.
• The state of technology restricts what is possible in combining
inputs to produce output (technological feasibility).
• Definition - Production Possibility Set: A set Y ∈ RL
describing possible production plans is called production
possibility set, Y = {y ∈ RL
| y is a feasible production plan}.
yi < 0 are called inputs, yi > 0 outputs.
165
Production 1
Production Possibility Set (2)
Microeconomics
• Often the production possibility set is described by a function
F(.) called transformation function. This function has the
property Y = {y ∈ RL
|F(y) ≤ 0} and F(y) = 0 if and only if we
are on the boundary of the set Y . {y ∈ RL
|F(y) = 0} is called
transformation frontier.
• Definition - Marginal Rate of Transformation: If F(.) is
differentiable and F(¯y) = 0, then for commodities k and l the
ration
−
dyk
dyl
= MRTlk(¯y) =
∂F(¯y)/∂yl
∂F(¯y)/∂yk
is called marginal rate of transformation of good l for good k.
166
Production 1
Production Possibility Set (3)
Microeconomics
• If l and k are outputs we observe how output of k increases if l
decreases by one marginal unit.
• With inputs .... In this case the marginal rate of transformation is
called marginal rate of technical substitution.
• With a single output q, production is often described by means of
a production function q = f(z1, . . . , zm), where the inputs
zi ≥ 0, i = 1, . . . , m. In this case Y =
{(−z1, . . . , −zm, q) |q − f(z1, . . . , zm) ≤ 0 and z1, . . . , zm ≥ 0}.
167
Production 1
Production Possibility Set (4)
Microeconomics
• Assumption and Properties of production possibility sets
P1 Y is non-empty.
P2 Y is closed. I.e. Y includes its boundary, if yn ∈ Y converges
to y then y ∈ Y .
P3 No free lunch. If yl ≥ 0 for l = 1, . . . , L, then y = 0. It is not
possible to produce something from nothing. Therefore
Y ∩ RL
+ = 0 ∈ Y (note that 0 ∈ Y has to be assumed here).
See Figure 5.B.2, page 131.
168
Production 1
Production Possibility Set (5)
Microeconomics
P4 Possibility of inaction: 0 ∈ Y . This assumption holds at least
ex-ante, before the setup of the firm. If we have entered into
some irrevocable contracts, then a sunk cost might arise.
P5 Free Disposal: New inputs can be acquired without any reduction
of output. If y ∈ Y and y ≤ y then y ∈ Y . For any y ∈ Y and
x ∈ RL
+, we get y − x ∈ Y . See Figure 5.B.4, page 132.
P6 Irreversibility: If y ∈ Y and y = 0, then −y /∈ Y . It is impossible
to reverse a possible production vector. We do not come from
output to input.
169
Production 1
Production Possibility Set (6)
Microeconomics
P7 Nonincreasing returns to scale: If y ∈ Y , then αy ∈ Y for all
α ∈ [0, 1]. I.e. any feasible input-output vector y can be scaled
down. See Figure 5.B.5.
P8 Nondecreasing returns to scale: If y ∈ Y , then αy ∈ Y for any
scale α ≥ 1. I.e. any feasible input-output vector y can be scaled
up. See Figure 5.B.6.
P9 Constant returns to scale: If y ∈ Y , then αy ∈ Y for any scale
α ≥ 0. I.e. any feasible input-output vector y can be scaled up
and down.
170
Production 1
Production Possibility Set (7)
Microeconomics
P10 Additivity - free entry: If y ∈ Y and y ∈ Y , then y + y ∈ Y .
This implies that ky ∈ Y for any positive integer k.
• Example: Output is an integer. If y and y are possible, additivity
means that y + y is still possible and the production of y has no
impact on y and vice versa. E.g. we have two independent
plants.
• As regards free-entry: If the aggregate production set Y is
additive, then unrestricted entry is possible. To see this, if y ∈ Y
is produced by firm A and y ∈ Y by firm B, then y + y ∈ Y if
additivity holds. That is, the production plans of firm A do not
interfere with the production plans of firm B (and vice versa). In
other words, the aggregate production set has to satisfy
additivity whenever unrestricted entry is possible.
171
Production 1
Production Possibility Set (8)
Microeconomics
P11 Convexity: Y is a convex set. I.e. if y ∈ Y and y ∈ Y , then
αy + (1 − α)y ∈ Y , for any α ∈ [0, 1].
• Convexity implies nonincreasing returns to scale.
• We do not increase productivity by using unbalanced input
combinations. If y and y produce the same output, then a
convex combination of the correspond inputs must at least
produce an output larger or equal to the output with y and y .
172
Production 1
Production Possibility Set (9)
Microeconomics
P12 Y is convex cone: Y is a convex cone if for any y, y ∈ Y and
constants α, β ≥ 0, αy + βy ∈ Y . Conjunction between
convexity and constant returns to scale property.
173
Production 1
Production Possibility Set (10)
Microeconomics
• Proposition [P 5.B.1]: The production set Y is additive and
satisfies the nonincreasing returns to scale property if and only if
it is convex cone.
174
Production 1
Production Possibility Set (11)
Microeconomics
Proof:
• If Y is a convex cone then Y is additive and satisfies the
nonincreasing returns to scale by the definition of a convex cone.
• We have to show that with additivity and nonincreasing returns
to scale we get αy + βy ∈ Y for any y, y and α, β > 0 (note
that with α ≥ 0, β = 0, α = 0, β ≥ 0 the relation αy + βy ∈ Y
follows from RTS and additivity): Let γ be a natural number
where γ ≥ max{α, β} > 0. By additivity γy ∈ Y and γy ∈ Y .
• α/γ and β/γ are ≤ 1. Due to nonincreasing returns to scale
αy = (α/γ)γy and βy = (β/γ)γy ∈ Y . By additivity
αy + βy ∈ Y .
175
Production 1
Production Possibility Set (12)
Microeconomics
• Proposition [P 5.B.2]: For any convex production set Y ⊂ RL
with 0 ∈ Y , there is a constant returns to scale convex
production set Y ∈ RL+1
such that Y = {y ∈ RL
|(y, −1) ∈ Y }.
Y is called extended production set.
• Proof: Let Y = {y ∈ RL+1
|y = α(y, −1), y ∈ Y, α ≥ 0}. If
y ∈ Y , then the first L components are in Y by construction.
Since βy1
+ (1 − β)y2
∈ Y we get
β(y1
, −1) + (1 − β)(y2
, −1) ∈ Y . αy ∈ Y by construction.
176
Production 1
Production Function (1)
Microeconomics
• Often it is sufficient to work with an output q ≥ 0 and inputs
z = (z1, . . . , zm) where zi ≥ 0.
• Definition - Production Function: A function describing the
the relationship between q and z is called production function f.
• Remark: The production functions assigns the maximum of
output q that can be attained to an input vector z.
f(z) = max{q ≥ 0|z ∈ Rm
+ }; (output efficient production).
177
Production 1
Production Function (2)
Microeconomics
• Assumption PF on Production Function: The production
function f : Rm
+ → R+ is continuous, strictly increasing and
strictly quasiconcave on Rm
+ ; f(0) = 0.
• Assumption PF’ - Production Function: The production
function f : Rm
+ → R+ is continuous, increasing and
quasiconcave on Rm
+ ; f(0) = 0.
• How can we motivate these assumptions?
178
Production 1
Production Function (3)
Microeconomics
• Considering production functions two approaches are common:
(i) variation one factor, (ii) variation all factors in the same
proportion.
• Definition - Marginal Product: If f is differentiable then
∂f(z)
∂zi
= MPi(z) is called marginal product of the input factor zi.
• By Assumption PF all marginal products are strictly larger than
zero, with PF’ MPi(z) ≥ 0.
• Definition - Average Product: The fraction f(z)/zi = APi(z)
is called average product of the input factor zi.
179
Production 1
Production Function (4)
Microeconomics
• Definition - Isoquant: The set Q(q) where output is constant is
called q-level isoquant. I.e. Q(q) = {z ≥ 0|f(z) = q}.
• In addition to Q(q) we can define the the contour set
¯S(q) = {z ≥ 0|f(z) ≥ q}. Since f is quasiconcave, this set is
convex ⇒ isoquants are convex curves.
180
Production 1
Production Function (5)
Microeconomics
• In addition, by means of the isoquant we can observe how input
factors can be substituted to remain on the same level of output.
• Definition - Marginal Rate of Technical Substitution:
MRTSij(z) =
MPi
MPj
• The slope of the isoquant is given by −
dzj
dzi
= MPi
MPj
• Discuss: MPi
MPj
> 0 (≥ 0) and the concept of technical efficiency:
To remain on the same level of output at least one input has to
be increased if one input factor has been decreased.
181
Production 1
Production Function (6)
Microeconomics
• In general the MRTS of two input depends on all other inputs
(note that the MPi depends on z).
• In applied work it is often assumed that inputs can be classified,
such that the MRTS within a class is not affected by inputs
outside this class.
182
Production 1
Production Function (7)
Microeconomics
• Since MRTSij is sensitive to the dimension of the
measurements of zi and zj an elasticity can be used.
• Definition - Elasticity of Substitution: For a differentiable
production function the elasticity of substitution between inputs
zi and zj is defined by
σij :=
d(zj/zi)
d(MPi/MPj)
(MPi/MPj)
(zj/zi)
=
d log(zj/zi)
d log(MPi/MPj)
.
• With a quasiconcave production function σij ≥ 0
183
Production 1
Production Function (8)
Microeconomics
• Another way to look at the properties of production is to alter
inputs proportionally. I.e. zi/zj remains constant.
• Discuss: This analysis is of interest especially for the long run
behavior of a firm.
184
Production 1
Production Function (9)
Microeconomics
• Definition - Returns to Scale. A production function f(z)
exhibits
– Constant returns to scale if f(µz) = µf(z) for µ > 0 and all z.
– Increasing returns to scale if f(µz) > µf(z) for µ > 1 and all
z.
– Decreasing returns to scale if f(µz) < µf(z) for µ > 1 and all
z.
185
Production 1
Production Function (10)
Microeconomics
• With constant returns the scale the production function has to
be homogeneous of degree one.
• Homogeneity larger than one is sufficient for increasing returns to
scale but not necessary.
• Most production function/technologies often exhibit regions with
constant, increasing and decreasing returns to scale.
186
Production 2
Profits and Cost (1)
Microeconomics
• Profit Maximization
• Cost minimization
• Price taking
• Cost, profit and supply function
MasColell, Chapter 5.C
187
Production 2
Profits - Single Output Case (1)
Microeconomics
• Suppose the there is only one output q ≥ 0 and input z ≥ 0. The
relationship between q and z is described by a differentiable
production function. The price of q is p > 0. Input factor prices
are w 0. We assume price taking behavior and that the second
order conditions are met.
• The profit maximization problem now reads as follows:
π(p, w) := {max
z,q≥0
pf(z) − w · z s.t. f(z) ≥ q}
• The input factor demand arising from this problem z = z(w, q) is
called input factor demand.
188
Production 2
Profits - Single Output Case (2)
Microeconomics
• Is the profit function well defined?
• Suppose that pf(z) − w · z ≥ 0. What happens if f(z) exhibits
increasing returns to scale?
• Here pf(µz) − w · µz > pµf(z) − w · µz for all µ > 1.
• That is, the profit can always be increased when increasing µ.
• With constant returns to scale no problem arises when
π(w, p) = 0. Then pf(µz) − w · µz = pµf(z) − w · µz = 0 for all
µ.
189
Production 2
Profits - Single Output Case (3)
Microeconomics
• From these remarks we get the (long run) problem:
max{pq − w · z} s.t f(z) ≥ q
• The Lagrangian is now given by:
L(q, z, λ) = pq − w · z + λ(f(z) − q)
• The marginal product will be abbreviated by MPi = ∂f(z)
∂zi
.
190
Production 2
Profits - Single Output Case (4)
Microeconomics
• Then the Kuhn-Tucker conditions are given by:
∂L
∂q
= p − λ ≤ 0 ,
∂L
∂q
q = 0
∂L
∂zi
= −wi + λMPi ≤ 0 ,
∂L
∂zi
zi = 0
∂L
∂λ
= f(z) − q ≥ 0 ,
∂L
∂λ
λ = 0
191
Production 2
Profits - Single Output Case (5)
Microeconomics
• This yields:
wi = p
∂f(z)
∂zi
, ∀zi > 0
• Definition - Marginal Revenue Product: p∂f(z)
∂zi
.
• For inputs i and j (both larger zero) we derive:
∂f(z)/∂zi
∂f(z)/∂zj
=
wi
wj
192
Production 2
Profits (1)
Microeconomics
• Assume that p = (p1, . . . , pL) are larger than zero and fixed
(price taking assumption).
• We assume that firms maximize profits.
• Given an Input-Output vector y, the profit generated by a firm is
p · y.
• We assume that Y is non-empty, closed and free disposal holds.
193
Production 2
Profits (2)
Microeconomics
• Definition: Given the production possibility set Y , we get the
profit maximization problem
max
y
p · y s.t. y ∈ Y.
• If Y can be described by a transformation function F, this
problem reads as follows:
max
y
p · y s.t. F(y) ≤ 0.
• Define π(p) = supy p · y s.t. y ∈ Y .
194
Production 2
Profits (3)
Microeconomics
• Definition - Profit function π(p): The maximum value function
associated with the profit maximization problem is called profit
function. The firm’s supply correspondence y(p) is the set of
profit maximizing vectors {y ∈ Y |p · y = π(p)}.
• The value function π(p) is defined on extended real numbers
(¯R = R ∪ {−∞, +∞}). The set Sp = {p · y|y ∈ Y } is a subset of R.
{p · y|y ∈ Y } has an upper bound in ¯R. For p where Sp is unbounded (from
above) in R we set π(p) = ∞.
• If Y is compact a solution (and also the max) for the profit maximization
problem exits. If this is not the case π(p) = ∞ is still possible. The profit
function exists by Bergs theorem of the maximum if the constraint
correspondence is continuous.
• We follow MWG and write maxy p · y s.t. y ∈ Y , although ....; Jehle/Reny
call π(p, w) well defined if π(p, w) < ∞. 195
Production 2
Profits (4)
Microeconomics
• Suppose that F (.) is differentiable, then we can formulate the profit
maximization problem as a Kuhn-Tucker problem:
• The Lagrangian is given by: L(y, λ) = p · y − λF (y)
• Then the Kuhn-Tucker conditions are given by:
∂L
∂yl
= pl − λ
∂F (y)
∂yl
= 0 ,
∂L
∂λ
= −F (y) ≥ 0
∂L
∂λ
λ = 0
• We assume that the second order conditions are met.
196
Production 2
Profits (5)
Microeconomics
• Remark: In the above maximization problem each yl ∈ R. That
is, it can be used as an output or an input.
• If however yl ≥ 0 for l ∈ “set of outputs” and yk ≤ 0 for k ∈
“set of input” (and both sets are non-empty and disjoint), we
obtain the first order conditions ∂L
∂yl
= pl − λ∂F (y)
∂yl
≤ 0 and
∂L
∂yl
yl = 0 for an output, while for an input we get
∂L
∂yl
= pk − λ∂F (y)
∂yk
≥ 0 and ∂L
∂yk
yk = 0.
197
Production 2
Profits (6)
Microeconomics
• For those inputs and output different from zero we get:
p = λ yF(y)
This implies that
pl
pk
=
∂F/∂yl
∂F/∂yk
= MRTlk.
• Since the left hand side is positive by assumption, the fraction of
the right hand side and λ have to be positive.
198
Production 2
Profits (7)
Microeconomics
• If yl, yk > 0, i.e. both goods are outputs, then yl, yk have to be
chosen such that the fraction of marginal rates of transformation
is equal to the ratio of prices.
• If yl, yk < 0, i.e. both goods are inputs, then yl, yk have to be
chosen such that the fraction of marginal rates of transformation
(= marginal rate of technical substitution) is equal to the
ratio of prices.
• If yl > 0, yk < 0, i.e. yl is an output and yk is an input, then
pl = ∂F/∂yl
∂F/∂yk
pk. Later on we shall observe that ∂F/∂yl
∂F/∂yk
pk is the
marginal cost of good l. See Figure 5.C.1. page 136.
199
Production 2
Profit Function (1)
Microeconomics
• By means of π(p) we can reconstruct −Y , if −Y is a convex set.
• That is to say: π(p) follows from {maxy p · y s.t. y ∈ Y }, which
is equivalent to {miny − p · y s.t. y ∈ Y } and
{min−y p · (−y) s.t. (−y) ∈ −Y }.
• Remember the concept of a support function: By means of the
support function µX(p) we get by means of {x|p · x ≥ µX(p)} a
dual representation of the closed and convex set X.
• Here −π(p) = µ−Y (p) where µ−Y (p) = miny{p · (−y)|y ∈ Y }
such that −π(p) is a support function of −Y .
200
Production 2
Profit Function (2)
Microeconomics
• Proposition: [5.C.1] Suppose that π(p) is the profit function of
the production set Y and y(p) is the associated supply
correspondence. Assume that Y is closed and satisfies the free
disposal property. Then
1. π(p) is homogeneous of degree one.
2. π(p) is convex.
3. If Y is convex, then Y = {y ∈ RL
|p · y ≤ π(p) , ∀p 0}
4. y(p) is homogeneous of degree zero.
5. If Y is convex, then y(p) is convex for all p. If Y is strictly
convex, then y(p) is single valued.
6. Hotelling’s Lemma: If y(¯p) consists of a single point, then
π(p) is differentiable at ¯p and π(¯p) = y(¯p).
7. If y is differentiable at ¯p, then Dy(¯p) = D2
π(¯p) is a symmetric
and positive semidefinite matrix with Dy(¯p)¯p = 0.
201
Production 2
Profit Function (3)
Microeconomics
Proof:
• π(p) is homogeneous of degree one and y(p) is homogeneous of
degree zero follow from the structure of the optimization
problem. If y ∈ y(p) solves {max p · y s.t. F(y) ≤ 0} then it also
solves α{max p · y s.t. F(y) ≤ 0} and
{max αp · y s.t. F(y) ≤ 0}, such that y ∈ y(αp) for any α > 0.
• This hold for every y ∈ y(p) ⇒ y(p) is homogeneous of degree
zero and π(p) is homogeneous of degree one by the structure of
the profit equation.
202
Production 2
Profit Function (4)
Microeconomics
Proof:
• π(p) is convex: Consider p1
and p2
and the convex combination
pν
. y1
, y2
and yν
are arbitrary elements of the optimal supply
correspondences.
• We get p1
y1
≥ p1
yν
and p2
y2
≥ p2
yν
• Multiplying the first term with ν and the second with 1 − ν,
where ν ∈ [0, 1] results in
νp1
y1
+ (1 − ν)p2
y2
≥ νp1
yν
+ (1 − ν)p2
yν
= pν
yν
which implies
νπ(p1
) + (1 − ν)π(p2
) ≥ π(pν
)
203
Production 2
Profit Function (5)
Microeconomics
Proof:
• If Y is convex then Y = {y ∈ RL
|p · y ≤ π(p) , for all p 0}: If
Y is convex, closed and free disposal holds, then π(p) provides a
dual description of the production possibility set.
204
Production 2
Profit Function (6)
Microeconomics
Proof:
• If Y is convex then y(p) is a convex set, with strict convexity
y(p) is a function: If Y is convex then yν
= νy1
+ (1 − ν)y2
∈ Y .
• If y1
and y2
solve the PMP for p, then π(p) = p · y1
= p · y2
. A
rescaling of the production vectors has to result in
yν
= νy1
+ (1 − ν)y2
where p · yν
= π(p) has to hold.
This follows from p · y1
= p · y2
= π(p)= νπ(p) + (1 − ν)π(p)=
νp·y1
+(1−ν)p·y2
= pν ·y1
+p(1−ν)·y2
= p(ν ·y1
+(1−ν)·y2
).
205
Production 2
Profit Function (7)
Microeconomics
Proof:
• Suppose that yα
solves the PMP and Y is strictly convex (every
point on the boundary is an extreme point, i.e. this point is not a
convex combination of other points in Y ). yα
is an element of
Y ∩ H(p, π(p)). H(p, π(p)) stands for an isoprofit hyperplane.
Suppose that there is another solution y solving the profit
maximization problem (PMP). So y, y are elements of this
hyperplane. Since y, y ∈ Y this implies that Y cannot be strictly
convex.
• Remark by Proposition P 5.F.1, page 150, y(p) cannot be an
interior point of y. Suppose that an interior point y solves the
PMP, then π(p) = p · y . For any interior point, there is an y
such that y ≥ y and y = y . Since p 0 this implies
p · y > p · y such that an interior point cannot be optimal. 206
Production 2
Profit Function (8)
Microeconomics
Proof:
• Hotellings lemma: Follows directly from the duality theorem:
pπ(¯p) = y(¯p); (see [P 3.F.1], page 66).
207
Production 2
Profit Function (9)
Microeconomics
Proof:
• Property 7: If y(p) and π are differentiable, then
Dpy(¯p) = D2
pπ(p). By Young’s theorem this matrix is symmetric,
since π(p) is convex in p the matrix has to be positive
semidefinite (see Theorem M.C.2, p.. 932).
• Dpy(p)p = 0L×1 follows from the Euler theorem (see MWG,
Theorem M.B.2, p. 929).
208
Production 2
Profit Function (10)
Microeconomics
• By Hotellings lemma inputs and outputs react in the same
direction as the price change: Output increases if output prices
increase, while inputs decrease if its prices increase (law of
supply), i.e.:
(p − p )[y(p) − y(p )] ≥ 0
• This law holds for any price change (there is no budget
constraint, therefore any form of compensation is not necessary.
We have no wealth effect but only substitution effects).
• We can also show that the law of supply holds also for the
non-differentiable case. (We know that p1
y1
≥ p1
y for any
y1
∈ y(p1
) and p2
y2
≥ p2
y for any y2
∈ y(p1
), sum up ....)
209
Production 2
Cost Function (1)
Microeconomics
• Profit maximization implies cost minimization!
• Production does not tell us anything about the minimal cost to
get output.
• On the other hand side - if the firm is not a price taker in the
output market, we cannot use the profit function, however the
results on the cost function are still valid.
• With increasing returns to scale where the profit function can
only take the values 0 or +∞, the cost function is better
behaved since the output is kept fixed there.
210
Production 2
Cost Function (2)
Microeconomics
• Assume that the input factor prices w 0 are constant. In
addition we assume that the production function is at least
continuous.
• Definition - Cost: Expenditures to acquire input factors z to
produce output q; i.e. w · z.
• Definition - Cost Minimization Problem (CMP): minz w · z
s.t. f(z) ≥ q. The minimal value function C(w, q) is called cost
function. The optimal input factor choices are called
conditional factor demand correspondence z(w, q).
211
Production 2
Cost Function (3)
Microeconomics
• Existence: Construct the set {z|f(z) ≥ q}. Under the usual
assumptions on the production function the set is closed. By
compactifying this set by means of {z|f(z) ≥ q, zi ≤ w · ¯z/wi}
for some ¯z with f(¯z) = q we can apply the Weierstraß theorem.
• By Berge’s theorem of the maximum we get a continuous cost
function C(w, q), if the constraint correspondence is continuous.
212
Production 2
Cost Function (4)
Microeconomics
• Definition - Marginal Cost: LMC(q) = ∂c(w,q)
∂q is called
marginal cost.
• Definition - Average Cost: LAC(q) = c(w,q)
q is called average
cost.
213
Production 2
Cost Function (5)
Microeconomics
• Theorem: Properties of the Cost Function c(w, q): [P 5.C.2]
Suppose that c(w, q) is a cost function of a single output
technology Y with production function f(z) and z(w, q) is the
associated conditional factor demand correspondence. Assume
that Y is closed and satisfies the free disposal property. Then
(i) c(w, q) is homogeneous of degree one in w and nondecreasing
in q.
(ii) Concave in w.
(iii) If the set {z ≥ 0|f(z) ≥ q} is convex for every q, then
Y = {(−z, q)|w · z ≥ c(w, q) for all w 0}.
(iv) z(w, q) is homogeneous of degree zero in w.
(v) If the set {z ≥ 0|f(z) ≥ q} is convex then z(w, q) is a convex
set, with strict convexity z(w, q) is a function.
214
Production 2
Cost Function (6)
Microeconomics
• Theorem: Properties of the Cost Function c(w, q): [P 5.C.2]
Suppose that c(w, q) is a cost function of a single output
technology Y with production function f(z) and z(w, q) is the
associated conditional factor demand correspondence. Assume
that Y is closed and satisfies the free disposal property. Then
(vi) Shepard’s lemma: If z( ¯w, q) consists of a single point, then
c(.) is differentiable with respect to w at ¯w and
wc( ¯w, q) = z( ¯w, q).
(vii) If z(.) is differentiable at ¯w then Dwz( ¯w, q) = D2
c( ¯w, q) is
symmetric and negative semidefinite with Dwz( ¯w, q) ¯w = 0.
(viii) If f(.) is homogeneous of degree one, then c(.) and z(.) are
homogeneous of degree one in q.
(ix) If f(.) is concave, then c(.) is a convex function of q (marginal
costs are nondecreasing in q).
215
Production 2
Cost Function (7)
Microeconomics
• By means of the cost function we can restate the profit
maximization problem (PMP for the single output case):
max
q≥0
pq − C(w, q)
• The first order condition becomes:
p −
∂C(w, q)
∂q
≤ 0
with (p − ∂C(w,q)
∂q ) = 0 if q > 0.
216
Production 3
Aggregate Supply and Efficiency
Microeconomics
• Aggregate Supply
• Joint profit maximization is a result of individual profit
maximization
• Efficient Production
Mas-Colell Chapters 5.D, 5.E
217
Production 3
Aggregate Supply (1)
Microeconomics
• Consider J units (firms, plants) with production sets Y1, . . . , YJ
equipped with profit functions πj(p) and supply correspondences
yj(p), j = 1, . . . , J.
• Definition - Aggregate Supply Correspondence: The sum of
the yj(p) is called aggregate supply correspondence:
y(p) :=
J
j
yj(p) = {y ∈ R
L
|y =
J
j
yj for some yj ∈ yj(p)}, j = 1, . . . , J}
• Definition - Aggregate Production Set: The sum of the
individual Yj is called aggregate production set:
Y =
J
j
Yj = {y ∈ R
L
|y =
J
j
yj for some yj ∈ Yj, j = 1, . . . , J}
218
Production 3
Aggregate Supply (2)
Microeconomics
• Proposition The law of supply also holds for the aggregate
supply function.
• Proof: Since (p − p )[yj(p) − yj(p )] ≥ 0 for all j = 1, . . . , J it
has also to hold for the sum.
• Definition: π∗
(p) and y∗
(p) are the profit function and the supply
correspondence of the aggregate production set Y .
219
Production 3
Aggregate Supply (3)
Microeconomics
• Proposition[5.E.1] For all p 0 we have
– π∗
(p) =
J
j πj(p)
– y∗
(p) =
J
j yj(p) (= {
J
j yj|yj ∈ yj(p)})
• Suppose that prices are fixed, this proposition implies that the
aggregate profit obtained by production of each unit separately is
the same as if we maximize the joint profit.
220
Production 3
Aggregate Supply (4)
Microeconomics
Proof:
• π∗
(p) =
J
j πj(p): Since π∗
is the maximum value function
obtained from the aggregate maximization problem, we have
π∗
(p) ≥ p · ( j yj) = j p · yj which implies π∗
(p) ≥ j πj(p).
• To show equality, note that there are yj in Yj such that
y = j yj. Then p · y = j p · yj ≤ j πj(p) for all y ∈ Y .
221
Production 3
Aggregate Supply (5)
Microeconomics
Proof:
• y∗
(p) =
J
j yj(p): Here we have to show that j yj(p) ⊂ y∗
(p)
and y∗
(p) ⊂ j yj(p). Consider yj ∈ yj(p), then
p · ( j yj) = j πj(p) = j pyj = j πj(p) = π∗
(p) (the last
step by the first part of [5.E.1]).
• From this argument is follows that yj(p) ⊂ y∗
(p).
• To get the second direction we start with y ∈ y∗
(p). Then
y = j yj with yj ∈ Yj. Since p · y = p · ( j ·yj) = π∗
(p) and
π∗
(p) =
J
j πj(p), it must be that p · yj = πj(p) (because
yj ∈ Yj implies p · yj ≤ πj(p)). Hence, we get y∗
(p) ⊂ j yj(p).
222
Production 3
Aggregate Supply (6)
Microeconomics
• The same aggregation procedure can also be applied to derive
aggregate cost.
223
Production 3
Efficiency (1)
Microeconomics
• We want to check whether or what production plans are wasteful.
• Definition: [D 5.F.1] A production vector is efficient, if there is
no y ∈ Y such that y ≥ y and y = y.
• There is no way to increase output with given inputs or to
decrease input with given output (sometimes called technical
efficiency).
• Discuss MWG, Figure 5.F.1, p. 150.
224
Production 3
Efficiency (2)
Microeconomics
• Proposition [P 5.F.1] If y ∈ Y is profit maximizing for some
p 0, then y is efficient.
• Version of the first fundamental theorem of welfare economics.
See MWG, Chapter 16 C, p. 549.
• It also tells us that a profit maximizing firm does not choose
interior points in the production set.
225
Production 3
Efficiency (3)
Microeconomics
Proof:
• We show this by means of a contradiction: Suppose that there is
a y ∈ Y such that y = y and y ≥ y. Because p 0 we get
p · y > p · y, contradicting the assumption that y solves the PMP.
• For interior points suppose that y is the interior. By the same
argument we see that this is neither efficient nor optimal.
226
Production 3
Efficiency (4)
Microeconomics
• This result implies that a firm chooses y in the convex part of Y
(with a differentiable transfer function F(.) this follows
immediately from the first order conditions; otherwise we choose
0 or ∞).
• The result also holds for nonconvex production sets - see Figure
5.F.2, page 150.
• Generally it is not true that every efficient production vector is
profit maximizing for some p ≥ 0, this only works with convex Y .
227
Production 3
Efficiency (6)
Microeconomics
• Proposition [P 5.F.2] Suppose that Y is convex. Then every
efficient production y ∈ Y is profit maximizing for some p ≥ 0
and p = 0.
228
Production 3
Efficiency (7)
Microeconomics
Proof:
• Suppose that y is efficient. Construct the set
Py = {y ∈ RL
|y y}. This set has to be convex. Since y is
efficient the intersection of Y and Py has to be empty.
• This implies that we can use the separating hyperplane theorem
[T M.G.2], page 948: There is some p = 0 such that
p · y ≥ p · y for every y ∈ Py and y ∈ Y . This implies
p · y ≥ p · y for every y y. Therefore, we also must have
p ≥ 0. If some pl < 0 then we could have p · y < p · y for some
y y with yl − yl sufficiently large. This procedure works for
each arbitrary y. p = 0.
229
Production 3
Efficiency (8)
Microeconomics
Proof:
• It remains to show that y maximizes the profit: Take an arbitrary
y ∈ Y , y was fixed, p has been derived by the separating
hyperplane theorem. Then p · y ≥ p · y for every y ∈ Py.
y ∈ Py can be chosen arbitrary close to y, such that
p · y ≥ p · y still has to hold. I.e. y maximizes the profit given p.
• Regarding this proof see also MWG, Figure 5.F.3, p. 151.
230
Production 4
Objectives of the Firm (1)
Microeconomics
• Until now we have assumed that the firm maximizes its profit.
• The price vector p was assumed to be fixed.
• We shall see that although preference maximization makes sense
when we consider consumers, this need not hold with profit
maximization with firms.
• Only if p is fixed we can rationalize profit maximization.
231
Production 4
Objectives of the Firm (2)
Microeconomics
• The objectives of a firm should be a result of the objectives of the
owners controlling the firm. That is to say, firm owners are also
consumers who look at their preferences. So profit maximization
need not be clear even if the firm is owned by one individual.
• MWG argue (”optimistically”) that the problem of profit
maximization is resolved, when the prices are fixed. This arises
with firms with no market power.
232
Production 4
Objectives of the Firm (3)
Microeconomics
• Consider a production possibility set Y owned by consumers
i = 1, . . . , I. The consumers own the shares θi, with
I
i=1 θi = 1.
y ∈ Y is a production decision. wi is non-profit wealth.
• Consumer i maximizes utility maxxi≥0 u(xi), s.t.
p · xi ≤ wi + θip · y.
• With fixed prices the budget set described by p · xi ≤ wi + θip · y
increases if p · y increases.
• With higher p · y each consumer i is better off. Here maximizing
profits p · y makes sense.
233
Production 4
Objectives of the Firm (4)
Microeconomics
• Problems arise (e.g.) if
– Prices depend on the action taken by the firm.
– Profits are uncertain (risk attitude plays a role).
– Firms are not controlled by its owners.
234
Production 4
Objectives of the Firm (5)
Microeconomics
• Suppose that the output of a firm is uncertain. It is important to
know whether output is sold before or after uncertainty is
resolved.
• If the goods are sold on a spot market (i.e. after uncertainty is
resolved), then also the owner’s attitude towards risk will play a
role in the output decision. Maybe less risky production plans are
preferred (although the expected profit is lower).
• If there is a futures market the firm can sell the good before
uncertainty is resolved and the consumers bear the risk. Profit
maximization can still be optimal.
235
Production 4
Objectives of the Firm (6)
Microeconomics
• Consider a two good economy with goods x1 and x2; L = 2,
non-profit wealth wi = 0. Suppose that the firm can influence
the price of good 1, p1 = p1(x1). We normalize the price of good
2, such that p2 = 1. z units of x2 are used to produce x1 with
production function x1 = f(z). The cost is given by p2z = z.
• We consider the maximization problem maxxi≥0 u(xi1, xi2), s.t.
p · xi ≤ wi + θip · y.
Given the above notation p = (p1(x1), 1), y = (f(z), −z).
wi = 0 by assumption. The profit is
p · y = p1(x1)x1 − p2z = p1(f(z))f(z) − z.
236
Production 4
Objectives of the Firm (7)
Microeconomics
• Assume that the preferences of the owners are such that they are
only interested in good 2.
• The aggregate amount of x2 the consumers can buy is
1
p2
(p1(f(z))f(z) − p2z) = p1(f(z))f(z) − z (since p2 = 1).
• Hence, maxxi≥0 u(xi2), s.t. p · xi ≤ wi + θip · y results in
max p1(f(z))f(z) − z.
237
Production 4
Objectives of the Firm (8)
Microeconomics
• Assume that the preferences of the owners are such that they
only look at good 1.
• The aggregate amount of x1 the consumers can buy is
1
p1(f(z)) (p1(f(z))f(z) − z) = f(z) − z/p1(f(z)).
• Then maxxi≥0 u(xi1), s.t. p · xi ≤ wi + θip · y results in
max f(z) − z/p1(f(z)).
• We have two different optimization problems - solutions are
different.
238
Production 4
Objectives of the Firm (9)
Microeconomics
• Example: Let p1(f(z)) =
√
z, then the first order conditions are
different, i.e. 1
2
√
z
f(z) +
√
zf (z) − 1 = 0 and f (z) − 1
2
√
z
= 0.
• We have considered two extreme cases: all owners prefer (i) good
2, (ii) good 1. There is no unique output decision based on
max p · y.
• If the preferences become heterogeneous things do not become
better.
239
Expected Utility
Uncertainty (1)
Microeconomics
• Preferences and Lotteries.
• Von Neumann-Morgenstern Expected Utility Theorem.
• Attitudes towards risk.
• State Dependent Utility, Subjective Utility
MWG, Chapter 6.
240
Expected Utility
Lotteries (1)
Microeconomics
• A risky alternative results in one of a number of different states
of the world, ωi.
• The states are associated with consequences or outcomes, zn.
Each zn involves no uncertainty.
• Outcomes can be money prices, wealth levels, consumption
bundles, etc.
• Assume that the set of outcomes is finite. Then
Z = {z1, . . . , zN}.
• E.g. flip a coin: States {H, T} and outcomes Z = {−1, 1}, with
head H or tail T.
241
Expected Utility
Lotteries (2)
Microeconomics
• Definition - Simple Gamble/Simple Lottery: [D 6.B.1] With the
consequences {z1, . . . , zN} ⊆ Z and N finite. A simple gamble assigns a
probability pn to each outcome zn. pn ≥ 0 and
N
n=1 pn = 1.
• Notation: L = (p1, . . . , pN). pi ≥ 0 is the probability of consequence zi,
for i = 1, . . . , N.
• Let us fix the set of outcomes Z: Different lotteries correspond to a different
set of probabilities.
• Definition - Set of Simple Gambles: The set of simple gambles on Z is
given by
LS = {(p1, . . . , pN)|pn ≥ 0 ,
N
n=1
pn = 1} = {L|pn ≥ 0 ,
N
N=1
pn = 1}
242
Expected Utility
Lotteries (3)
Microeconomics
• Definition - Degenerated Lottery:
˜Ln
= (0, . . . , 1, . . . , 0) = en.
• ’Z ⊆ LS’, since one can identify zn with ˜Ln
.
243
Expected Utility
Lotteries (4)
Microeconomics
• With N consequences, every simple lottery can be represented by
a point in a N − 1 dimensional simplex
∆(N−1)
= {p ∈ RN
+ | pn = 1} .
• At each corner n we have the degenerated case that pn = 1.
• With interior points pn > 0 for all i.
• See Ritzberger, p. 36,37, Figures 2.1 and 2.2 or MWG, Figure
6.B.1, page 169.
• Equivalent to Machina’s triangle; with N = 3;
{(p1, p3) ∈ [0, 1]2
|0 ≤ 1 − p1 − p3 ≤ 1}.
244
Expected Utility
Lotteries (5)
Microeconomics
• The consequences of a lottery need not be a z ∈ Z but can also
be a further lottery.
• Definition - Compound Lottery: [D 6.B.2] Given K simple
lotteries Lk and probabilities αk ≥ 0 and αk = 1, the
compound lottery
LC = (L1, . . . , Lk, . . . , LK; α1, . . . , αk, . . . , αK). It is the risky
alternative that yields the simple lottery Lk with probability αk.
• The support of the compound lottery (the set of consequences
with positive probability) is the union of the supports generating
this lotteries.
245
Expected Utility
Lotteries (6)
Microeconomics
• Definition - Reduced Lottery: For any compound lottery LC
we can construct a reduced lottery/simple gamble L ∈ LS.
With the probabilities pk
for each Lk
we get p = αkpk
, such
that probabilities for each zn ∈ Z are pn =
K
k=1 αkpk
n.
• Examples: Example 2.5, Ritzberger p. 37
• A reduced lottery can be expressed by a convex combination of
elements of compound lotteries (see Ritzberger, Figure 2.3, page
38). I.e. αpl1
+ (1 − α)pl2
= plreduced
.
246
Expected Utility
von Neumann-Morgenstern Utility (1)
Microeconomics
• Here we assume that any decision problem can be expressed by
means of a lottery (simple gamble).
• Only the outcomes matter.
• Consumers are able to perform calculations like in probability
theory, gambles with the same probability distribution on Z are
equivalent.
247
Expected Utility
von Neumann-Morgenstern Utility (2)
Microeconomics
• Axiom vNM1 - Completeness: For two gambles L1 and L2 in
LS either L1 L2, L2 L1 or both.
• Here we assume that a consumer is able to rank lotteries (risky
alternatives). I.e. Axiom vNM1 is stronger than Axiom 1 under
certainty.
• Axiom vNM2 - Transitivity: For three gambles L1, L2 and L3:
L1 L2 and L2 L3 implies L1 L3.
248
Expected Utility
von Neumann-Morgenstern Utility (3)
Microeconomics
• Axiom vNM3 - Continuity: [D 6.B.3] The preference relation
on the space of simple lotteries is continuous if for any L1, L2, L3
the sets {α ∈ [0, 1]|αL1 + (1 − α)L2 L3} ⊂ [0, 1] and
{α ∈ [0, 1]|L3 αL1 + (1 − α)L2} ⊂ [0, 1] are closed.
• Later we show: for any gambles L ∈ LS, there exists some
probability α such that L ∼ α¯L + (1 − α)L, where ¯L is the most
preferred and L the least preferred lottery.
• This assumption rules out a lexicographical ordering of
preferences (safety first preferences).
249
Expected Utility
von Neumann-Morgenstern Utility (4)
Microeconomics
• Consider the outcomes Z = {1000, 10, death}, where
1000 10 death. L1 gives 10 with certainty.
• If vNM3 holds then L1 can be expressed by means of a linear
combination of 1000 and death. If there is no α ∈ [0, 1] fulfilling
this requirement vNM3 does not hold.
250
Expected Utility
von Neumann-Morgenstern Utility (5)
Microeconomics
• Monotonicity: For all probabilities α, β ∈ [0, 1],
α¯L + (1 − α)L β ¯L + (1 − β)L
if and only if α ≥ β.
• Monotonicity is implied by the axioms vNM1-vNM4.
251
Expected Utility
von Neumann-Morgenstern Utility (6)
Microeconomics
• Axiom vNM4 - Independence, Substitution: For all
probabilities L1, L2 and L3 in LS and α ∈ (0, 1):
L1 L2 ⇔ αL1 + (1 − α)L3 αL2 + (1 − α)L3 .
• This axiom implies that the preference orderings of the mixtures
are independent of the third lottery.
• This axiom has no parallel in consumer theory under certainty.
252
Expected Utility
von Neumann-Morgenstern Utility (7)
Microeconomics
• Example: consider a bundle x1
consisting of 1 cake and 1 bottle
of wine x1
= (1, 1), x2
= (3, 0); x3
= (3, 3). Assume that
x1
x2
.
Axiom vNM4 requires that αx1
+ (1 − α)x3
αx2
+ (1 − α)x3
;
here α > 0.
253
Expected Utility
von Neumann-Morgenstern Utility (8)
Microeconomics
• Lemma - vNM1-4 imply monotonicity: Moreover, if L1 L2
then αL1 + (1 − α)L2 βL1 + (1 − β)L2 for arbitrary
α, β ∈ [0, 1] where α ≥ β. For every L1 L L2, there is
unique γ ∈ [0, 1] such that γL1 + (1 − γ)L2 ∼ L.
• See steps 2-3 of the vNM existence proof.
254
Expected Utility
von Neumann-Morgenstern Utility (9)
Microeconomics
• Definition - von Neumann Morgenstern Expected Utility
Function: [D 6.B.5] A real valued function U : LS → R has
expected utility form if there is an assignment of numbers
(u1, . . . , uN) (with un = u(zn)) such that for every lottery
L ∈ LS we have U(L) = zn∈Z p(zn)u(zn). A function of this
structure is said to satisfy the expected utility property - it is
called von Neumann-Morgenstern (expected) utility function.
• Note that this function is linear in the probabilities pn.
• u(zn) is called Bernoulli utility function.
255
Expected Utility
von Neumann-Morgenstern Utility (10)
Microeconomics
• Proposition - Linearity of the von Neumann Morgenstern
Expect Utility Function: [P 6.B.1] A utility function has
expected utility form if and only if it is linear. That is to say:
U
K
k=1
αkLk =
K
k=1
αkU(Lk)
256
Expected Utility
von Neumann-Morgenstern Utility (11)
Microeconomics
Proof:
• Suppose that U(
K
k=1 αkLk) =
K
k=1 αkU(Lk) holds. We have
to show that U has expected utility form, i.e. if
U( k αkLk) = k αkU(Lk) then U(L) = pnu(zn).
• If U is linear then we can express any lottery L by means of a
compound lottery with probabilities αn = pn and degenerated
lotteries ˜Ln
. I.e. L = pn
˜Ln
. By linearity we get
U(L) = U( pn
˜Ln
) = pnU(˜Ln
).
• Define u(zn) = U(˜Ln
). Then
U(L) = U( pn
˜Ln
) = pnU(˜Ln
) = pnu(zn). Therefore
U(.) has expected utility form.
257
Expected Utility
von Neumann-Morgenstern Utility (12)
Microeconomics
Proof:
• Suppose that U(L) =
N
n=1 pnu(zn) holds. We have to show
that utility is linear, i.e. if U(L) = pnu(zn) then
U( k αkLk) = k αkU(Lk)
• Consider a compound lottery (L1, . . . , LK, α1, . . . , αK). Its
reduced lottery is L = k αkLk.
• Then U( k αkLk) = n k αkpk
n u(zn) =
k αk n pk
nu(zn) = k αkU(Lk).
258
Expected Utility
von Neumann-Morgenstern Utility (13)
Microeconomics
• Proposition - Existence of a von Neumann Morgenstern
Expect Utility Function: [P 6.B.3] If the Axioms vNM 1-4 are
satisfied for a preference ordering on LS. Then admits an
expected utility representation. I.e. there exists a real valued
function u(.) on Z which assigns a real number to each outcome
zn, n = 1, . . . , N, such that for any pair of lotteries
L1 = (p1, . . . , pN) and L2 = (p1, . . . , pN) we get
L1 L2 if and only if
U(L1) :=
N
n=1
pnu(zn) ≥ U(L2) :=
N
n=1
pnu(zn) .
259
Expected Utility
von Neumann-Morgenstern Utility (14)
Microeconomics
Proof:
• Suppose that there is a best and a worst lottery. With a finite set
of outcomes this can be easily shown by means of the
independence axiom. In addition ¯L L.
• By the definition of ¯L and L we get: ¯L Lc L, ¯L L1 L
and ¯L L2 L.
• We have to show that (i) U(L) exists and (ii) that for any
compound lottery Lc = βL1 + (1 − β)L2 we have
U(βL1 + (1 − β)L2) = βU(L1) + (1 − β)U(L2) (expected utility
structure).
260
Expected Utility
von Neumann-Morgenstern Utility (15)
Microeconomics
Proof:
• Step 1: By the independence Axiom vNM4 we get if L1 L2
and α ∈ (0, 1) then L1 αL1 + (1 − α)L2 L2.
• This follows directly from the independence axiom.
L1 ∼ αL1+(1−α)L1 αL1+(1−α)L2 αL2+(1−α)L2 = L2
261
Expected Utility
von Neumann-Morgenstern Utility (16)
Microeconomics
Proof:
• Step 2: Want to show that β > α , if and only if
β ¯L + (1 − β)L α¯L + (1 − α)L (monotonicity):
• Define γ = (β − α)/(1 − α); the assumptions imply γ ∈ (0, 1].
262
Expected Utility
von Neumann-Morgenstern Utility (17)
Microeconomics
Proof:
• Then
β ¯L + (1 − β)L = γ ¯L + (1 − γ)(α¯L + (1 − α)L)
γ(α¯L + (1 − α)L) + (1 − γ)(α¯L + (1 − α)L)
∼ α¯L + (1 − α)L
263
Expected Utility
von Neumann-Morgenstern Utility (18)
Microeconomics
Proof:
• Step 2: For the converse we have to show that
β ¯L + (1 − β)L α¯L + (1 − α)L results in β > α. We show this
by means of the contrapositive: If β > α then
β ¯L + (1 − β)L α¯L + (1 − α)L.
• Thus assume β ≤ α, then α¯L + (1 − α)L β ¯L + (1 − β)L
follows in the same way as above. If α = β indifference follows.
264
Expected Utility
von Neumann-Morgenstern Utility (19)
Microeconomics
Proof:
• Step 3: There is a unique αL such that L ∼ αL
¯L + (1 − αL)L.
• Existence follows from ¯L L and the continuity axiom:
• Ad existence: define the sets {α ∈ [0, 1]|α¯L + (1 − α)L L}
and {α ∈ [0, 1]|L α¯L + (1 − α)L}. Both sets are closed. Any α
belongs to at least one of these two sets. Both sets are nonempty.
Their complements are open and disjoint. The set [0, 1] is
connected ⇒ there is at least one α belonging to both sets.
• Uniqueness follows directly from step 2.
265
Expected Utility
Excursion: Connected Sets
Microeconomics
• Definition: Let Ω = ∅ be an arbitrary set. A class τ ⊂ 2Ω
of
subsets of Ω is called a topology on Ω if it has the three
properties:
– ∅, Ω ∈ τ
– A ∩ B ∈ τ for any two sets A, B ∈ τ.
– A∈F A ∈ τ for any F ⊂ τ.
• The pair (Ω, τ) is called a topological space. The sets A ∈ τ
are called open sets, and the sets A ⊂ Ω with Ac
∈ τ are called
closed sets; Ac
stands for complementary set.
266
Expected Utility
Excursion: Connected Sets
Microeconomics
• Consider the family τR of subsets of R: O ∈ τR if and only if for
each x ∈ O, there is an ε > 0 such that (x − ε, x + ε) ⊂ O. That
is, elements of O are arbitrary unions of open intervals.
• Fact from Math: τR forms a topology on R. It is called
Euclidean topology.
• We consider the closed interval [0, 1] with the following topology:
A ⊂ [0, 1] is open if and only if there is an O ∈ τR such that
A = O ∩ [0, 1]. This topology is induced by τR.
267
Expected Utility
Excursion: Connected Sets
Microeconomics
• Definition: Let (X, τ) be a topological space. The space is said
to be connected, if for any two non-empty closed subsets
A, B ⊂ X, A ∪ B = X implies A ∩ B = ∅.
• Fact from Math: [0, 1] with by τR the induced topology is
connected.
268
Expected Utility
von Neumann-Morgenstern Utility (20)
Microeconomics
Proof:
• Step 4: The function U(L) = αL represents the preference
relations .
• Consider L1, L2 ∈ LS: If L1 L2 then α1 ≥ α2. If α1 ≥ α2 then
L1 L2 by steps 2-3.
• It remains to show that this utility function has expected utility
form.
269
Expected Utility
von Neumann-Morgenstern Utility (21)
Microeconomics
Proof:
• Step 5: U(L) is has expected utility form.
• We show that the linear structure also holds for the compound
lottery Lc = βL1 + (1 − β)L2.
• By using the independence we get:
βL1 + (1 − β)L2 ∼ β(α1
¯L + (1 − α1)L) + (1 − β)L2
∼ β(α1
¯L + (1 − α1)L) + (1 − β)(α2
¯L + (1 − α2)L)
∼ (βα1 + (1 − β)α2)¯L + (β(1 − α1) + (1 − β)(1 − α2))L
• By the rule developed in step 4, this shows that
U(Lc) = U(βL1 + (1 − β)L2) = βU(L1) + (1 − β)U(L2).
270
Expected Utility
von Neumann-Morgenstern Utility (22)
Microeconomics
• Proposition - von Neumann Morgenstern Expect Utility
Function are unique up to Positive Affine Transformations:
[P 6.B.2] If U(.) represents the preference ordering , then V
represents the same preference ordering if and only if
V = α + βU, where β > 0.
271
Expected Utility
von Neumann-Morgenstern Utility (23)
Microeconomics
Proof:
• Note that if V (L) = α + βU(L), V (L) fulfills the expected
utility property (see also MWG p. 174).
• We have to show that if U and V represent preferences, then V
has to be an affine linear transformation of U.
• If U is constant on LS, then V has to be constant. Both
functions can only differ by a constant α.
272
Expected Utility
von Neumann-Morgenstern Utility (24)
Microeconomics
Proof:
• Alternatively, for any L ∈ LS and ¯L L, we get
f1 :=
U(L) − U(L)
U(¯L) − U(L)
and
f2 :=
V (L) − V (L)
V (¯L) − V (L)
.
• f1 and f2 are linear transformations of U and V that satisfy the
expected utility property.
• fi(L) = 0 and fi(¯L) = 1, for i = 1, 2.
273
Expected Utility
von Neumann-Morgenstern Utility (25)
Microeconomics
Proof:
• L ∼ L then f1 = f2 = 0; if L ∼ ¯L then f1 = f2 = 1.
• By expected utility U(L) = γU(¯L) + (1 − γ)U(L) and
V (L) = γV (¯L) + (1 − γ)V (L).
• If ¯L L L then there has to exist a unique γ, such that
L L ∼ γ ¯L + (1 − γ)L ¯L. Therefore
γ =
U(L) − U(L)
U(¯L) − U(L)
=
V (L) − V (L)
V (¯L) − V (L)
274
Expected Utility
von Neumann-Morgenstern Utility (26)
Microeconomics
Proof:
• Then V (L) = α + βU(L) where
α = V (L) − U(L)
V (¯L) − V (L)
U(¯L) − U(L)
and
β =
V (¯L) − V (L)
U(¯L) − U(L)
.
275
Expected Utility
von Neumann-Morgenstern Utility (27)
Microeconomics
• The concept of expected utility can be extended to a set of
distributions F(x) where the expectation of u(x) exists, i.e.
A
u(x)dF(x) < ∞, z ∈ R and A ⊂ R.
• For technical details see e.g. Robert (1994), The Bayesian Choice
and DeGroot, Optimal Statistical Decisions.
• Note that expected utility is a probability weighted combination
of Bernoulli utility functions. I.e. the properties of the random
variable z, described by the lottery l(z), are separated from the
attitudes towards risk.
276
Expected Utility
VNM Indifference Curves (1)
Microeconomics
• Indifferences curves are straight lines; see Ritzberger, Figure 2.4,
page 41.
• Consider a VNM utility function and two indifferent lotteries L1
and L2. It has to hold that U(L1) = U(L2).
• By the expected utility theorem
U(αL1 + (1 − α)L2) = αU(L1) + (1 − α)U(L2).
• If U(L1) = U(L2) then U(αL1 + (1 − α)L2) = U(L1) = U(L2)
has to hold and the indifferent lotteries is linear combinations of
L1 and L2.
277
Expected Utility
VNM Indifference Curves (2)
Microeconomics
• Indifference curves are parallel; see Ritzberger, Figure 2.5, 2.6,
page 42.
• Consider L1 ∼ L2 and a further lottery L3 L1 (w.l.g.).
• From βL1 + (1 − β)L3 and βL2 + (1 − β)L3 we have received
two compound lotteries.
• By construction these lotteries are on a line parallel to the line
connecting L1 and L2.
278
Expected Utility
VNM Indifference Curves (3)
Microeconomics
• The independence axiom vNM4 implies that
βL1 + (1 − β)L3 ∼ βL2 + (1 − β)L3 for β ∈ [0, 1].
• Therefore the line connecting the points βL1 + (1 − β)L3 and
βL2 + (1 − β)L3 is an indifference curve.
• The new indifference curve is a parallel shift of the old curve; by
the linear structure of the expected utility function no other
indifference curves are possible.
279
Expected Utility
Allais Paradoxon (1)
Microeconomics
Lottery 0 1-10 11-99
pz 1/100 10/100 89/100
La 500,000 500,000 500,000
Lb 0 2,500,000 500,000
Ma 500,000 500,000 0
Mb 0 2,500,000 0
280
Expected Utility
Allais Paradoxon (2)
Microeconomics
• Most people prefer La to Lb and Mb to Ma.
• This is a contradiction to the independence axiom.
• Allais paradoxon in the Machina triangle, Gollier, Figure 1.2,
page 8.
281
Expected Utility
Allais Paradoxon (3)
Microeconomics
• Expected utility theory avoids problems of time inconsistency.
• Agents violating the independence axiom are subject to Dutch
book outcomes (violate no money pump assumption).
282
Expected Utility
Allais Paradoxon (4)
Microeconomics
• Three lotteries: La Lb and La Lc.
• But Ld = 0.5Lb + 0.5Lc La.
• Gambler is willing to pay some fee to replace La by Ld.
283
Expected Utility
Allais Paradoxon (5)
Microeconomics
• After nature moves: Lb or Lc with Ld.
• Now the agents is once again willing to pay a positive amount for
receiving La
• Gambler starting with La and holding at the end La has paid two
fees!
• Dynamically inconsistent/Time inconsistent.
• Discuss Figure 1.3, Gollier, page 12.
284
Expected Utility
Risk Attitude (1)
Microeconomics
• For the proof of the vNM-utility function we did not place any
assumptions on the Bernoulli utility function u(z).
• For applications often a Bernoulli utility function has to be
specified.
• In the following we consider z ∈ R and u (z) > 0; abbreviate
lotteries with money amounts l ∈ LS.
• There are interesting interdependences between the Bernoulli
utility function and an agent’s attitude towards risk.
285
Expected Utility
Risk Attitude (2)
Microeconomics
• Consider a nondegenerated lottery l ∈ LS and a degenerated lottery ˜l.
Assume that El(z) = ˜zl holds. I.e. the degenerated lottery ˜l pays the
expectation ˜zl of l for sure.
• Definition - Risk Aversion: A consumer is risk averse if for any lottery l, ˜zl
is at least as good as l. A consumer is strictly risk averse if for any lottery l, ˜z
is strictly preferred to l, whenever l is non-degenerate.
• Definition - Risk Neutrality: A consumer is risk neutral if ˜zl ∼ l for all l.
• Definition - Risk Loving: A consumer is risk loving if for any lottery l, ˜zl is
at most as good as l. A consumer is strictly risk loving if for any lottery l, l is
strictly preferred to ˜zl, whenever l is non-degenerate.
286
Expected Utility
Risk Attitude (3)
Microeconomics
• By the definition of risk aversion, we see that the utility function u(.) has to
satisfy for any non-degenerate distribution F ,
u(E(z)) = u( zdF (z)) ≥ E(u(z)) = u(z)dF (z).
• If u(z) is a concave function and z is distributed according to F (z) (such
that the expectations exist), then
u(z)dF (z) ≤ u( zdF (z))
Jensen’s inequality. In addition, if u(z)dF (z) ≤ u( zdF (z)) holds for
any distribution F , then u(z) is concave.
• For sums this implies:
pzu(z) ≤ u( pzz) .
For strictly concave function, < has to hold whenever F is nondegenerate, for
convex functions we get ≥; for strictly convex functions > whenever F is
non-degenerate. 287
Expected Utility
Risk Attitude (4)
Microeconomics
• For a lottery l where E(u(z)) < ∞ and E(z) < ∞ we can
calculate the amount C where a consumer is indifferent between
receiving C for sure and the lottery l. I.e. l ∼ C and
E(u(z)) = u(C) hold.
• In addition we are able to calculate the maximum amount π an
agent is willing to pay for receiving the fixed amount E(z) for
sure instead of the lottery l. I.e. l ∼ E(z) − π or
E(u(z)) = u(E(z) − π).
288
Expected Utility
Risk Attitude (5)
Microeconomics
• Definition - Certainty Equivalent [D 6.C.2]: The fixed amount
C where a consumer is indifferent between C an a gamble l is
called certainty equivalent.
• Definition - Risk Premium: The maximum amount π a
consumer is willing to pay to exchange the gamble l for a sure
state with outcome E(z) is called risk premium.
• Note that C and π depend on the properties of the random
variable (described by l) and the attitude towards risk (described
by u).
289
Expected Utility
Risk Attitude (6)
Microeconomics
• Remark: the same analysis can also be performed with risk
neutral and risk loving agents.
• Remark: MWG defines a probability premium, which is
abbreviated by π in the textbook. Given a degenerated lottery
and some ε > 0. The probability-premium πR
is defined as
u(˜lz) = (1
2 + πR
)u(z + ε) + (1
2 − πR
)u(z − ε). I.e.
mean-preserving spreads are considered here.
290
Expected Utility
Risk Attitude (7)
Microeconomics
• Proposition - Risk Aversion and Bernoulli Utility: Consider
an expected utility maximizer with Bernoulli utility function u(.).
The following statements are equivalent:
– The agent is risk averse.
– u(.) is a (strictly) concave function.
– C ≤ E(z). (< with strict version)
– π ≥ 0. (> with strict version)
291
Expected Utility
Risk Attitude (8)
Microeconomics
Proof: (sketch)
• By the definition of risk aversion: for a lottery l where E(z) = z˜l,
a risk avers agent ˜l l.
• I.e. E(u(z)) ≤ u(z˜l) = u(E(z)) for a VNM utility maximizer.
• (ii) follows from Jensen’s inequality.
• (iii) If u(.) is (strictly) concave then E(u(z)) = u(C) ≤ u(E(z))
can only be matched with C ≤ E(z).
• (iv) With a strictly concave u(.),
E(u(z)) = u(E(z) − π) ≤ u(E(z)) can only be matched with
π ≥ 0.
292
Expected Utility
Arrow Pratt Coefficients (1)
Microeconomics
• Using simply the second derivative u (z) of the Bernoulli utility
function, causes problems with affine linear transformations.
• Definition - Arrow-Pratt Coefficient of Absolute Risk
Aversion: [D 6.C.3] Given a twice differentiable Bernoulli utility
function u(.), the coefficient of absolute risk aversion is defined
by A(z) = −u (z)/u (z).
• Definition - Arrow-Pratt Coefficient of Relative Risk
Aversion: [D 6.C.5] Given a twice differentiable Bernoulli utility
function u(.), the coefficient of relative risk aversion is defined by
R(z) = −zu (z)/u (z).
293
Expected Utility
Comparative Analysis (1)
Microeconomics
• Consider two agents with Bernoulli utility functions u1 and u2.
We want to compare their attitudes towards risk.
Definition - More Risk Averse: Agent 1 is more risk averse
than agent 2: Whenever agent 1 finds a lottery F at least good
as a riskless outcome ˜x, then agent 2 finds F at least good as ˜x.
I.e. if F 1
˜L˜x then F 2
˜L˜x.
In terms of a VNM-ultility maximizer: If
EF (u1(z)) = u1(z)dF(z) ≥ u1(˜x) then
EF (u2(z)) = u2(z)dF(z) ≥ u2(˜x) for any F(.) and ˜x.
294
Expected Utility
Comparative Analysis (2)
Microeconomics
• Define a function φ(x) = u1(u−1
2 (x)). Since u2(.) is an
increasing function this expression is well defined. We, in
addition, assume that the first and the second derivatives exist.
• By construction with x = u2(z) we get:
φ(x) = u1(u−1
2 (x)) = u1(u−1
2 (u2(z))) = u1(z). I.e. φ(x)
transforms u2 into u1, such that u1(z) = φ(u2(z)).
• In the following we assume that ui and φ are differentiable. In the
following theorem we shall observe that φ > 0 for u1 and u2 > 0.
295
Expected Utility
Comparative Analysis (3)
Microeconomics
• Proposition - More Risk Averse Agents [P 6.C.2]: Assume
that the first and second derivatives of the Bernoulli utility
functions u1 and u2 exist (u > 0 and u < 0). Then the
following statements are equivalent:
– Agent 1 is (strictly) more risk averse than agent 2.
– u1 is a (strictly) concave transformation of u2 (that is, there
exists a (strictly) concave φ such that u1(.) = φ(u2(.)))
– A1(z) ≥ A2(z) (> for strict) for all z.
– C1 ≤ C2 and π1 ≥ π2; (<> for strict).
296
Expected Utility
Comparative Analysis (4)
Microeconomics
Proof:
• Step 1: (i) follows from (ii): We have to show that if φ is concave, then if
EF (u1(z)) = u1(z)dF (z) ≥ u1(˜x) ⇒
EF (u2(z)) = u2(z)dF (z) ≥ u2(˜x) has to follow.
• Suppose that for some lottery F the inequality
EF (u1(z)) = u1(z)dF (z) ≥ u1(˜x) holds. This implies
EF (u1(z)) = u1(z)dF (z) ≥ u1(˜x) = φ(u2(˜x)).
• By means of Jensen’s inequality we get for a concave φ(.); (with strict
concave we get <) E(u1(z)) = E(φ(u2(z)) ≤ φ(E(u2(z))).
• Then φ(E(u2(z))) ≥ E(u1(z)) and E(u1(z)) ≥ u1(˜x) = φ(u2(˜x))
implies φ(E(u2(z))) ≥ φ(u2(˜x)).
• Since φ is increasing this implies E(u2(z)) ≥ u2(˜x).
297
Expected Utility
Comparative Analysis (5)
Microeconomics
Proof:
• (ii) follows from (i): Suppose that
EF (u1(z)) = u1(z)dF(z) ≥ u1(˜x) ⇒
EF (u2(z)) = u2(z)dF(z) ≥ u2(˜x) for any F(.) and ˜x holds
and φ is not concave.
• Then EF (u1(z)) = u1(CF 1) has to hold as well with ˜x = CF 1.
This implies EF (u1(z)) = EF (φ(u2(z))) = φ(u2(CF 1)) for
lottery F.
• Since φ is not concave, there exits a lottery where
φ(EF (u2(z))) < EF (φ(u2(z))) = φ(u2(CF 1)). This yields
EF (u2(z)) < u2(CF 1). Contradiction!
298
Expected Utility
Comparative Analysis (6)
Microeconomics
Proof:
• Step 2 (iii)∼ (ii): By the definition of φ and our assumptions we
get
u1(z) =
dφ((u2(z)))
dz
= φ (u2(z))u2(z) .
(since u1, u2 > 0 ⇒ φ > 0) and
u1(z) = φ (u2(z))u2(z) + φ (u2(z))(u2(z))2
.
299
Expected Utility
Comparative Analysis (7)
Microeconomics
Proof:
• Divide both sides by −u1(z) < 0 and using u1(z) = ... yields:
−
u1(z)
u1(z)
= A1(z) = A2(z) −
φ (u2(z))
φ (u2(z))
u2(z) .
• Since A1, A2 > 0 due to risk aversion, φ > 0 and φ ≤ 0 (<)
due to its concave shape we get A1(z) ≥ A2(z) (>) for all z.
300
Expected Utility
Comparative Analysis (8)
Microeconomics
Proof:
• Step 3 (iv)∼ (ii): Jensen’s inequality yields (with strictly concave
φ)
u1(C1) = E(u1(z)) = E(φ(u2(z)) < φ(E(u2(z))) = φ(u2(C2)) = u1(C2)
• Since u1 > 0 we get C1 < C2.
• π1 > π2 works in the same way.
• The above considerations also work in both directions, therefore
(ii) and (iv) are equivalent.
301
Expected Utility
Comparative Analysis (9)
Microeconomics
Proof:
• Step 4 (vi)∼ (ii): Jensen’s inequality yields (with strictly concave
φ)
u1(E(z)−π1) = E(u1(z)) = E(φ(u2(z)) < φ(E(u2(z))) = φ(u2(E(z)−π2)) = u1(E(z)−π2)
• Since u1 > 0 we get π1 > π2.
302
Expected Utility
Stochastic Dominance (1)
Microeconomics
• In an application, do we have to specify the Bernoulli utility
function?
• Are there some lotteries (distributions) such that F(z) is
(strictly) preferred to G(z)?
• E.g. if X(ω) > Y (ω) a.s.?
• YES ⇒ Concept of stochastic dominance.
• MWG, Figure 6.D.1., page 196.
303
Expected Utility
Stochastic Dominance (2)
Microeconomics
• Definition - First Order Stochastic Dominance: [D 6.D.1] A
distribution F(z) first order dominates the distribution G(z) if
for every nondecreasing function u : R → R we have
∞
−∞
u(z)dF(z) ≥
∞
−∞
u(z)dG(z).
• Definition - Second Order Stochastic Dominance: [D 6.D.2]
A distribution F(z) second order dominates the distribution G(z)
if EF (z) = EG(z) and for every nondecreasing concave function
u : R+ → R the inequality
∞
0
u(z)dF(z) ≥
∞
0
u(z)dG(z)
holds.
304
Expected Utility
Stochastic Dominance (3)
Microeconomics
• Proposition - First Order Stochastic Dominance: [P 6.D.1]
F(z) first order dominates the distribution G(z) if and only if
F(z) ≤ G(z).
• Proposition - Second Order Stochastic Dominance: [D
6.D.2] F(z) second order dominates the distribution G(z) if and
only if
¯z
0
F(z)dz ≤
¯z
0
G(z)dz for all ¯z in R+ .
• Remark: I.e. if we can show stochastic dominance we do not
have to specify any Bernoulli utility function!
305
Expected Utility
Stochastic Dominance (4)
Microeconomics
Proof:
• Assume that u is differentiable and u ≥ 0
• Step 1: First order, if part: If F(z) ≤ G(z) integration by parts
yields:
∞
−∞
u(z)dF (z) −
∞
−∞
u(z)dG(z) =
∞
−∞
u(z)F (z)dz −
∞
−∞
u(z)G (z)dz
= u(z)(F (z) − G(z))|
∞
−∞ −
∞
−∞
u (z)(F (z) − G(z))dz
= −
∞
−∞
u (z)(F (z) − G(z))dz ≥ 0 .
• The above inequality holds since the terms inside the integral
(F(z) − G(z)) ≤ 0. In addition, limt→∞ F(t) = 1 and
limt→−∞ F(t) = 0 and likewise for G(.).
306
Expected Utility
Stochastic Dominance (5)
Microeconomics
Proof:
• Step 2: First order, only if part: If FOSD then F (z) ≤ G(z) holds.
• Assume there is a ¯z ∈ R such that F (¯z) > G(¯z). ¯z > −∞ by
construction. Set u(z) = 0 for z ≤ ¯z and u(z) = 1 for z > ¯z. Here we get
0 ≤
∞
−∞
u(z)dF (z) −
∞
−∞
u(z)dG(z)
=
∞
−∞
u(z)F (z)dz −
∞
−∞
u(z)G (z)dz
=
∞
¯z
F (z)dz −
∞
¯z
G (z)dz
= (1 − F (¯z)) − (1 − G(¯z)) = −F (¯z) + G(¯z) .
307
Expected Utility
Stochastic Dominance (6)
Microeconomics
Proof:
• Second Order SD: Assume that u is twice continuously differentiable, such
that u (z) ≤ 0, w.l.g. u(0) = 0.
• Remark: The equality of means implies:
0 =
∞
0
zdF (z) −
∞
0
zdG(z)
=
∞
0
zF (z)dz −
∞
0
zG (z)dz
= z(F (z) − G(z))|
∞
0 −
∞
0
(F (z) − G(z))dz
= −
∞
0
(F (z) − G(z))dz .
308
Expected Utility
Stochastic Dominance (7)
Microeconomics
Proof:
• Step 3: Second order, if part: Integration by parts yields:
∞
0
u(z)dF (z) −
∞
0
u(z)dG(z)
= u(z)(F (z) − G(z))|
∞
0 −
∞
0
u (z)(F (z) − G(z))dz
= −
∞
0
u (z)(F (z) − G(z))dz
= −u (z)
z
0
(F (x) − G(x))dx|
∞
0 −
∞
0
−u (z)
z
0
(F (x) − G(x))dx dz
=
∞
0
u (z)
z
0
(F (x) − G(x))dx dz ≥ 0
• Note that u ≤ 0 by assumption.
309
Expected Utility
Stochastic Dominance (8)
Microeconomics
Proof:
• Step 4: Second order, only if part: Consider a ¯z such that
u(z) = ¯z for all z > ¯z and u(z) = z for all z ≤ ¯z. This yields:
∞
0
u(z)dF(z) −
∞
0
u(z)dG(z)
=
¯z
0
zdF(z) −
¯z
0
zdG(z) + ¯z ((1 − F(¯z)) − (1 − G(¯z)))
= z (F(z) − G(z)) |¯z
0 −
¯z
0
(F(z) − G(z)) dz − ¯z (F(¯z) − G(¯z))
= −
¯z
0
(F(z) − G(z)) dz ≥ 0 .
310
Expected Utility
Stochastic Dominance (9)
Microeconomics
• Definiton - Monotone Likelihood Ratio Property: The
distributions F(z) and G(z) fulfill, the monotone likelihood rate
property if G(z)/F(z) is non-increasing in z.
• For x → ∞ G(z)/F(z) = 1 has to hold. This and the fact that
G(z)/F(z) is non-increasing implies G(z)/F(z) ≥ 1 for all z.
• Proposition - First Order Stochastic Dominance follows
from MLP: MLP results in F(z) ≤ G(z).
311
Expected Utility
Arrow-Pratt Approximation (1)
Microeconomics
• By means of the Arrow-Pratt approximation we can express the
risk premium π in terms of the Arrow-Pratt measures of risk.
• Assume that z = w + kx, where w is a fixed constant (e.g.
wealth), x is a mean zero random variable and k ≥ 0. By this
assumption the variance of z is given by
V(z) = k2
V(x) = k2
E(x2
).
• Proposition - Arrow-Pratt Risk Premium with respect to
Additive risk: If risk is additive, i.e. z = w + kx, then the risk
premium π is approximately equal to 0.5A(w)V(z).
312
Expected Utility
Arrow-Pratt Approximation (2)
Microeconomics
Proof:
• By the definition of the risk premium we have
E(u(z)) = E(u(w + kx)) = u(w − π(k)).
• For k = 0 we get π(k) = 0. For risk averse agents dπ(k)/dk ≥ 0.
• Use the definition of the risk premium and take the first derivate
with respect to k on both sides:
E(xu (w + kx)) = −π (k)u (w − π(k)) .
313
Expected Utility
Arrow-Pratt Approximation (3)
Microeconomics
Proof:
• For the left hand side we get at k = 0:
E(xu (w + 0x)) = u (w)E(x) = 0 since E(x) = 0 by assumption.
• Matching LHS with RHS results in π (k) = 0 at k = 0, while
u (.) > 0 by assumption.
314
Expected Utility
Arrow-Pratt Approximation (4)
Microeconomics
Proof:
• Taking the second derivative with respect to k yields:
E(x2
u (w + kx)) = (π (k))2
u (w − π(k)) − π (k)u (w − π(k))
• At k = 0 this results in (note that π (0) = 0):
π (0) = −
u (w)
u (w)
E(x2
) = A(w)E(x2
)
315
Expected Utility
Arrow-Pratt Approximation (5)
Microeconomics
• A second order Taylor expansion of π(k) around k = 0 results in
π(k) ≈ π(0) + π (0)k +
π (0)
2
k2
• Thus
π(k) ≈ 0.5A(w)E(x2
)k2
• Since E(x) = 0 by assumption, the risk premium is proportional
to the variance of x, that is V(z) = k2
E(x2
).
316
Expected Utility
Arrow-Pratt Approximation (6)
Microeconomics
• For multiplicative risk we can proceed as follows: z = w(1 + kx)
where E(x) = 0.
• Proceeding the same way results in:
π(k)
w
≈ −
wu (w)
u (w)
k2
E(x2
) = 0.5R(w)E(x2
)k2
• Proposition - Arrow-Pratt Relative Risk Premium with
respect to Multiplicative risk: If risk is multiplicative, i.e.
z = w(1 + kx), then the relative risk premium π/w is
approximately equal to 0.5R(w)k2
V(x).
• Interpretation: Risk premium per monetary unit of wealth.
317
Expected Utility
Decreasing Absolute Risk Aversion (1)
Microeconomics
• It is widely believed that the more wealthy an agent, the smaller
his/her willingness to pay to escape a given additive risk.
• Definition - Decreasing Absolute Risk Aversion[D 6.C.4]:
The Bernoulli utility function for money exhibits decreasing
absolute risk aversion if the Arrow-Pratt coefficient of absolute
risk aversion −u (.)
u (.) is a decreasing function of wealth w.
318
Expected Utility
Decreasing Absolute Risk Aversion (2)
Microeconomics
• Proposition - Decreasing Absolute Risk Aversion: [P 6.C.3]
The following statements are equivalent
– The risk premium is a decreasing function in wealth w.
– Absolute risk aversion A(w) is decreasing in wealth.
– −u (z) is a concave transformation of u. I.e. u is sufficiently
convex.
319
Expected Utility
Decreasing Absolute Risk Aversion (3)
Microeconomics
Proof: (sketch)
• Step 1, (i) ∼ (iii): Consider additive risk and the definition of
the risk premium. Treat π as a function of wealth:
E(u(w + kx)) = u(w − π(w)) .
• Taking the first derivative yields:
E(1u (w + kx)) = (1 − π (w))u (w − π(w)) .
320
Expected Utility
Decreasing Absolute Risk Aversion (4)
Microeconomics
Proof: (sketch)
• This yields:
π (w) = −
E(u (w + kx)) − u (w − π(w))
u (w − π(w))
.
• π (w) decreases if E(u (w + kx)) − u (w − π(w)) ≥ 0.
• This is equivalent to E(−u (w + kx)) ≤ −u (w − π(w)).
• Note that we have proven that if E(u2(z)) = u2(z − π2) then
E(u1(z)) ≤ u1(z − π2) if agent 1 were more risk averse.
321
Expected Utility
Decreasing Absolute Risk Aversion (5)
Microeconomics
Proof: (sketch)
• Here we have the same mathematical structure (see slides on
Comparative Analysis): set z = w + kx, u1 = −u and u2 = u.
• ⇒ −u is more concave than u such that −u is a concave
transformation of u.
322
Expected Utility
Decreasing Absolute Risk Aversion (6)
Microeconomics
Proof: (sketch)
• Step 2, (iii) ∼ (ii): Next define P(w) := −u
u which is often
called degree of absolute prudence.
• From our former theorems we get: P(w) ≥ A(w) has to be
fulfilled (see A1 and A2).
• Take the first derivative of the Arrow-Pratt measure yields:
A (w) = −
1
(u (w))2
(u (w)u (w) − (u (w))
2
)
= −
u (w)
(u (w))
(u (w)/u (w) − u (w)/u (w))
=
u (w)
(u (w))
(P (w) − A(w))
323
Expected Utility
Decreasing Absolute Risk Aversion (7)
Microeconomics
Proof: (sketch)
• A decreases in wealth if A (w) ≤ 0.
• We get A (w) ≤ 0 if P(w) ≥ A(w).
324
Expected Utility
HARA Utility (1)
Microeconomics
• Definition - Harmonic Absolute Risk Aversion: A Bernoulli
utility function exhibits HARA if its absolute risk tolerance (=
inverse of absolute risk aversion) T(z) := 1/A(z) is linear in
wealth z.
• I.e. T(z) = −u (z)/u (z) is linear in z
• These functions have the form u(z) = ζ (η + z/γ)
1−γ
.
• Given the domain of z, η + z/γ > 0 has to hold.
325
Expected Utility
HARA Utility (2)
Microeconomics
• Taking derivatives results in:
u (z) = ζ
1 − γ
γ
(η + z/γ)
−γ
u (z) = −ζ
1 − γ
γ
(η + z/γ)
−γ−1
u (z) = ζ
(1 − γ)(γ + 1)
γ2
(η + z/γ)
−γ−2
326
Expected Utility
HARA Utility (3)
Microeconomics
• Risk aversion: A(z) = (η + z/γ)
−1
• Risk Tolerance is linear in z: T(z) = η + z/γ
• Absolute Prudence: P(z) = γ+1
γ (η + z/γ)
−1
• Relative Risk Aversion: R(z) = z (η + z/γ)
−1
327
Expected Utility
HARA Utility (4)
Microeconomics
• With η = 0, R(z) = γ: Constant Relative Risk Aversion
Utility Function: u(z) = log(z) for γ = 1 and u(z) = z1−γ
1−γ for
γ = 1.
• This function exhibits DARA; A (z) = −γ2
/z2
< 0.
328
Expected Utility
HARA Utility (5)
Microeconomics
• With γ → ∞: Constant Absolute Risk Aversion Utility
Function: A(z) = 1/η.
• Since u (z) = Au (z) we get u(z) = − exp(−Az)/A.
• This function exhibits increasing relative risk aversion.
329
Expected Utility
HARA Utility (6)
Microeconomics
• With γ = −1: Quadratic Utility Function:
• This functions requires z < η, since it is decreasing over η.
• Increasing absolute risk aversion.
330
Expected Utility
State Dependent Utility (1)
Microeconomics
• With von Neumann Morgenstern utility theory only the
consequences and their corresponding probabilities matter.
• I.e. the underlying cause of the consequence does not play any
role.
• If the cause is one’s state of health this assumption is unlikely to
be fulfilled.
• Example car insurance: Consider fair full cover insurance. Under
VNM utility U(l) = pu(w − P) + (1 − p)u(w − P), etc. If
however it plays a role whether we have a wealth of w − P in the
case of no accident or getting compensated by the insurance
company such the wealth is w − P, the agent’s preferences
depend on the states accident and no accident.
331
Expected Utility
State Dependent Utility (2)
Microeconomics
• With VNM utility theory we have considered the set of simple
lotteries LS over the set of consequences Z. Each lottery li
corresponds to a probability distribution on Z.
• Assume that Ω has finite states. Define a random variable f
mapping from Ω into LS. Then f(ω) = lω for all ω of Ω. I.e. f
assigns a simple lottery to each state ω.
• If the probabilities of the states are given by π(ω), we arrive at
the compound lotteries lSDU = π(ω)lω.
• I.e. we have calculated probabilities of compound lotteries.
332
Expected Utility
State Dependent Utility (3)
Microeconomics
• The set of lSDU will be called LSDU. Such lotteries are also
called horse lotteries.
• Note that also convex combinations of lSDU are ∈ LSDU.
• Definition - Extended Independence Axiom: The preference
relation satisfies extended independence if for all
l1
SDU, l2
SDU, lSDU ∈ LSDU and α ∈ (0, 1) we have l1
SDU lSDU
if and only if αl1
SDU + (1 − α)l2
SDU αlSDU + (1 − α)l2
SDU.
333
Expected Utility
State Dependent Utility (4)
Microeconomics
• Proposition - Extended Expected Utility/State Dependent
Utility: Suppose that Ω is finite and the preference relation
satisfies continuity and in independence on LSDU. Then there
exists a real valued function u : Z × Ω → R such that
l1
SDU l2
SDU
if and only if
ω∈Ω
π(ω)
z∈supp(l1
SDU
(ω))
pl1(z|ω)u(z, ω) ≥
ω∈Ω
π(ω)
z∈supp(l2
SDU
(ω))
pl2(z|ω)u(z, ω) .
334
Expected Utility
State Dependent Utility (4)
Microeconomics
• u is unique up to positive linear transformations.
• Proof: see Ritzberger, page 73.
• If only consequences matter such that u(z, ω) = u(z) then state
dependent utility is equal to VNM utility.
335
General Equilibrium
Outline
Microeconomics
• Motivation and main questions to be investigated:
– Does a competitive economy result in a Pareto efficient
allocation?
– Can any Pareto efficient allocation be obtained by means of a
price system in a competitive economy?
• Edgeworth Box
• Robinson Crusoe economies
• Small open economies and trade
• General vs. partial equilibrium
MWG, Chapter 15
336
General Equilibrium
Motivation (1)
Microeconomics
• Consider the economy as a closed and interrelated system.
• With the partial equilibrium approach these interrelations are
mainly ignored.
• The exogenous variables in general equilibrium theory are
reduced to a small number of physical realities (number of
agents, technologies available, preferences of the agents,
endowments of various agents).
337
General Equilibrium
Motivation (2)
Microeconomics
• First we consider:
– A pure exchange economy: no production is possible,
commodities are ultimately consumed, the individuals are
permitted to trade the commodities among themselves. With
two consumers and two goods this can be represented in the
Edgeworth box.
– One consumer - one firm economy, to get a first impression
on the impacts of production.
338
General Equilibrium
Motivation (3)
Microeconomics
• Consider I consumers, indexed i = 1, . . . , I. Xi ⊂ RL
are the
consumption sets. Each consumer chooses a consumption bundle
xi, the utility is given by ui(xi). The preferences are i.
• J firms, indexed j = 1, . . . , J. The production possibility sets are
Yj ⊂ RL
. The production vectors are yj.
• L goods, indexed = 1, . . . , L.
339
General Equilibrium
Motivation (4)
Microeconomics
• Total endowments of good is ω ≥ 0. The total net amount of
good available is ω + j y j, = 1, . . . , L.
• We assume that the initial endowments and technological
possibilities (i.e. the firms) are owned by consumers. Shares θij,
where
I
i=1 θij = 1 for all j = 1, . . . , J.
• The wealth of consumer i is wi(p) = p · ωi.
• Remark: often the endowments are abbreviated by e . Here we
stick to MWG and use ω .
340
General Equilibrium
Motivation (5)
Microeconomics
• Definition - Economic Allocation [D 10.B.1]: An economic
allocation (x, y) = (x1, . . . , xI, y1, . . . , yJ) is a specification of a
consumption vector xi ∈ Xi for each consumer i = 1, . . . , I and
a production vector yj ∈ Yj for each firm j = 1, . . . , J. The
allocation is feasible if
I
i=1
x i ≤ ω +
J
j=1
y j for = 1, . . . , L.
341
General Equilibrium
Motivation (6)
Microeconomics
• Definition - Competitive Economy
– Suppose that consumer i initially owns ω i, where
ω =
I
i=1 ω i for = 1, . . . , L, ωi = (ωi1, . . . , ωiL).
– Consumer i owns the shares θi = (θi1, . . . , θij, . . . , θiJ), where
I
i=1 θij = 1 for j = 1, . . . , J.
– Markets exist for all L goods and all firms are price takers; the
prices are p = (p1, . . . , pL).
342
General Equilibrium
Motivation (7)
Microeconomics
• Definition - Walrasian/Competitive Equilibrium [D 10.B.3] The allocation
(x, y) and the price vector p ∈ RL
constitute a competitive (Walrasian)
equilibirium if the following conditions are met:
– Profit maximization: each firm j solves maxyj∈Yj
p · yj where yj ∈ Yj.
– Utility maximization: each consumer i solves
max
xi∈Xi
ui(xi) s.t. p · xi ≤ p · ωi +
J
j=1
θij(p · yj).
– Market clearing: For each good = 1, . . . , L:
I
i=1
x i = ω +
J
j=1
y j.
343
General Equilibrium
Motivation (8)
Microeconomics
• Definition - Pareto Optimality [D 10.B.2]: A feasible allocation
(x, y) = (x1, . . . , xI, y1, . . . , yJ) is Pareto optimal (efficient) if
there is no other feasible allocation (x1, . . . , xI, y1, . . . , yJ) such
that ui(xi) ≥ ui(xi) for all i = 1, . . . , I and ui(xi) > ui(xi) for
some i.
• Definition - Utility Possibility Set: ”The set of attainable
utility levels”.
U = {(u1, . . . , uI) ∈ R
I
|∃ feasible allocation (x, y): ui ≤ ui(xi) for i = 1, . . . , I}
• Pareto efficient allocations are on the north-east boundary of this
set. See MWG, Figure 10.B.1.
344
Edgeworth Box (1)
Microeconomics
• We consider a pure exchange economy.
• Consumers posses initial endowments of commodities. Economic
activity consists of trading and consumption.
• Now we restrict to a two good - two consumer exchange
economy. Then, L = 2, X1 = X2 = R2
+, Y1 = Y2 = −R2
+ (the
free disposal technology). i is the index of the consumer, the
index of our goods.
• xi = (x1i, x2i) ∈ Xi. i are the preferences of consumer i.
• The initial endowments are ω i ≥ 0. The endowment vector of
consumer i is ωi = (ω1i, ω2i). The total endowments of good
are ¯ω = ω 1 + ω 2. We assume that ¯ω > 0 for = 1, 2.
345
Edgeworth Box (2)
Microeconomics
• From the above Definition [D 10.B.1] it follows that an economic
allocation (x, y) = (x1, . . . , xI, y1, . . . , yJ) is a specification of a
consumption vector xi ∈ Xi for each consumer i = 1, . . . , I and
a production vector yj ∈ Yj for each firm j = 1, . . . , J. It is
feasible if
I
i=1
x i ≤ ¯ω +
J
j=1
y j for = 1, . . . , L.
• For the Edgeworth Box an allocation is some consumption vector
x = (x11, x21, x21, x22) ∈ R4
+.
• An allocation is feasible if x 1 + x 2 ≤ ¯ω for = 1, 2.
346
Edgeworth Box (3)
Microeconomics
• Definition - Nonwasteful allocation: If x 1 + x 2 = ¯ω for
= 1, 2, then the allocation is called nonwasteful.
• Nonwasteful allocations can be described by means of an
Edgeworth box.
• See MWG, Figure 15.B.1.
• For a given price vector p = (p1, p2) the budget line intersects
the initial endowment point ωi = (ω1i, ω2i). The slope is −p1
p2
.
Note that only the relative price −p1
p2
matters, with −λp1
λp2
,
λ ∈ R++, we get the same Edgeworth box with the same budget
sets.
• See MWG, Figure 15.B.2.
347
Edgeworth Box (4)
Microeconomics
• Next we assume that the preferences of both consumers are
strongly monotone and strictly convex.
• For each price p consumer i obtains the budget set Bi(p). By
solving the utility maximization problem
maxx1i,x2i
ui(xi) s.t. p · xi ≤ wi(p)
we obtain the optimal quantities x1i(p), x2i(p). By collecting
x1i(p), x2i(p) for different p, we obtain the mapping
OCi : R2
+ → R2
+, p → (x1i(p), x2i(p)). This mapping is called
offer curve.
• By the assumptions on the preferences the solution of the UMP
is unique, hence here we obtain a function.
348
Edgeworth Box (5)
Microeconomics
• The consumer’s offer curve lies within the upper contour set of
ωi.
• See MWG, Figures 15.B.3.-15.B.5.
349
Edgeworth Box (6)
Microeconomics
• Definition [D 15.B.1] A Walrasian/Competivie Equilibrium for
an Edgeworth box economy is a price vector p∗
and a feasible
allocation x∗
= (x∗
1, x∗
2) in the Edgeworth box such that for
i = 1, 2,
x∗
i i xi for all xi ∈ Bi(p∗
).
• At any equilibrium the offer curves intersect.
• Consumer’s demand is homogeneous of degree zero in p, i.e. only
the relative price matters.
• See MWG, Figures 15.B.7 and 15.B.8.
350
Edgeworth Box (7)
Microeconomics
• A Walrasian equilibrium need not be unique.
• See MWG, Figure 15.B.9.
• This could already happen with quasilinear preferences, where
the preferences are such that different numeraire goods are used.
• MWG, Chapter 10 constructs a model where all agents have
quasilinear preferences with respect to the same numeraire good.
351
Edgeworth Box (8)
Microeconomics
• Recall: Definition - Quasilinear Preferences: A monotone
preference relation on X = (−∞, ∞) × RL−1
is quasilinear
with respect to commodity one (the numeraire good) if : (i) all
indifference sets are parallel displacements of each other along
the axis of commodity one. I.e. x ∼ y then x + αe1 ∼ y + αe1
and e1 = (1, 0, . . . ). (ii) Good one is desirable: x + αe1 x for
all α > 0. [D 3.B.7]
• A Walrasian equilibrium need not exist.
• This happens e.g. if (i) one consumer only desires only one good
or (ii) preferences are non-convex.
• See MWG, Figure 15.B.10.
352
Edgeworth Box (9)
Microeconomics
• Definition - Pareto Optimality [D 15.B.2]: A feasible allocation
x in the Edgeworth box is Pareto optimal (or Pareto efficient) if
there is no other allocation x in the Edgeworth box with
xi i xi for i = 1, 2 and xi i xi for some i. The set of all
Pareto optimal allocations is called Pareto set. The contract
curve is the part of the Pareto set where both consumers do at
least as well as at their initial endowments.
• See MWG, Figures 15.B.11 and 15.B.12.
353
Edgeworth Box (10)
Microeconomics
• We observe in the Edgeworth box that ”every Walrasian
equilibrium allocation x∗
belongs to the Pareto set”. This
corresponds to the first theorem of welfare economics.
Figure 15.B.7(a) and 15.B.8, MWG, page 520.
• Regarding the second theorem: a planner can (under convexity
assumptions, see MWG, Chapter 16) achieve any desired Pareto
efficient allocation.
• Hence we define:
354
Edgeworth Box (11)
Microeconomics
• Definition - Equilibrium with Transfers [D 15.B.3]: An
allocation x in the Edgeworth box is supportable as an
equilibrium with transfers, if there is a price system p∗
and
wealth transfers T1 and T2 satisfying T1 + T2 = 0, such that for
each consumer i we have
x∗
i i xi for all xi ∈ R2
+ such that p∗
· xi ≤ p∗
· ωi + Ti.
• In the Edworth box we observe that with continuous, strongly
monotone and strictly convex preferences any Pareto optimal
allocation is supportable.
• See MWG, Figure 15.B.13.
• See MWG, Figure 15.B.14 - to observe how the second theorem
fails with non-convex preferences. 355
One-Consumer, One-Producer (1)
Microeconomics
• We introduce production in the most simple way.
• There are two price taking agents, a single consumer and a single
firm.
• There are two goods, labor (or leisure) of the consumer and the
consumption good produced by the firm.
• The preferences defined over leisure x1 and the consumption
good x2 are continuous, strongly monotone and strictly convex.
The initial endowment consists of ¯L units of leisure and no
endowment of the consumption good.
356
One-Consumer, One-Producer (2)
Microeconomics
• The firm uses labor to produce the consumption good under the
increasing and strictly concave production function q = f(z),
where z is labor input and q the amount of x2 produced.
• The firm maximizes its profit:
max
z≥0
pf(z) − wz
given the prices (p, w). This optimization problem results in the
optimal labor demand z(p, w) and output q(p, w). The profit is
π(p, w).
357
One-Consumer, One-Producer (3)
Microeconomics
• The consumer maximizes the utility function u(x1, x2):
max
x1,x2≥0
u(x1, x2) s.t. px2 ≤ w(¯L − x1) + π(p, w).
This results in the Walrasian demand x1(p, w) and x2(p, w).
Labor supply corresponds to ¯L − x1(p, w).
• See MWG, Figure 15.C.1 on these optimization problems.
358
One-Consumer, One-Producer (4)
Microeconomics
• Walrasian equilibrium is attained at a pair (p∗
, w∗
) where
x2(p∗
, w∗
) = q(p∗
, w∗
) and z(p∗
, w∗
) = ¯L − x1(p∗
, w∗
).
• See MWG, Figure 15.C.1 on these optimization problems. See
MWG, Figure 15.C.2 for an equilibrium.
359
One-Consumer, One-Producer (5)
Microeconomics
• Remark: A particular consumption-leisure combination can arise
in a competitive equilibrium if and only if it maximizes the
consumer’s utility subject to the technological and endowment
constraints.
• ⇒ A Walrasian equilibrium allocation is the same as if a social
planner would maximize the consumer’s utility given the
technological constraints of the economy. A Walrasian
equilibrium is Pareto optimal.
360
One-Consumer, One-Producer (6)
Microeconomics
• Remark on Non-convexity: Suppose the the production set is not
convex, then we can construct examples where the price system
does not support the allocation x∗
.
• See MWG, Figure 15.C.3 (a).
361
General vs. Partial Equilibrium (1)
Microeconomics
• Bradford’s (1978) example on taxation:
• Consider an economy with N large towns. Each town has a
single price taking firm producing a consumption good by means
of a strictly concave production function f(z). The consumption
good is identical.
• The overall economy has M units of labor, inelastically supplied.
Utility is derived from consuming the output.
• Workers are free to move to another town. Hence the equilibrium
wage must be the same, i.e. w1, . . . , wN = ¯w.
• Without loss of generality the price of the output is normalized,
i.e. p = 1.
362
General vs. Partial Equilibrium (2)
Microeconomics
• By the symmetric construction of the model we get: each firm
hires M/N workers, the output of each firm is f(M/N).
• Due to price taking we get ¯w = f = ∂f(M/N)
∂(M/N) .
• The equilibrium profits are: f(M/N) − ∂f(M/N)
∂(M/N) (M/N).
363
General vs. Partial Equilibrium (3)
Microeconomics
• Suppose that town 1 levies a tax on labor, the tax rate is t > 0.
• Given the wage w1 and the tax rate t we arrive at a labor
demand z1, which is implicitly given by f (z1) = t + w1.
364
General vs. Partial Equilibrium (4)
Microeconomics
• Partial equilibrium argument: N is large, in impact on the other
wage rates can be neglected. Hence ¯w remains the same.
• Since labor moves freely, we get w1 = ¯w. The supply
correspondence is 0 at w1 < ¯w and ∞ at w1 > ¯w. It is [0, ∞] at
w1 = ¯w.
• Then f (z1) = t + ¯w. z1 falls by our assumptions on f(.), labor
moves to other towns.
• The incomes of the workers and the profits in towns 2, . . . , N
remain the same. The profit of firm 1 decreases, the firms
completely bear the tax burden.
365
General vs. Partial Equilibrium (5)
Microeconomics
• General equilibrium argument: Since labor moves freely,
w1, . . . , wN = w still has to hold. All M units of labor are
employed by the structure of f(.).
• w(t) denotes the equilibrium wage rate when the tax rate is t. By
symmetry z2(t) = · · · = zN(t) = z(t). z1(t) is the labor demand
in town 1.
• Then equilibrium demands for:
z1(t)+(N −1)z(t) = M , f (z(t)) = w(t) , f (z1(t)) = w(t)+t.
366
General vs. Partial Equilibrium (6)
Microeconomics
• Next, f (z1(t)) = w(t) + t = f (M − (N − 1)z(t)) = w(t) + t.
By taking the first derivative w.r.t. to t and evaluating at t = 0
(where z1(0) = z(0) = M/N) we obtain
f (M/N)[−(N − 1)]z (0) = w (0) + 1.
• The derivative of f (z(t)) = w(t) (the FOC of the firms) w.r.t.
to t and evaluated at t = 0 yields f (M/N)z (0) = w (0) such
that [−(N − 1)]w (0) = w (0) + 1 and
w (0) = −
1
N
.
367
General vs. Partial Equilibrium (7)
Microeconomics
• Hence, the wage rates in all towns decrease due to the tax in
town 1. Only if N goes to infinity this effect becomes zero.
• For N large the partial equilibrium approximation regarding
prices and quantities is correct. However, the distributional
effects remain wrong.
• Hence, the wage rates in all towns decrease due to the tax in
town 1. Only if N goes to infinity this effect becomes zero.
368
General vs. Partial Equilibrium (8)
Microeconomics
• In addition, when we consider the profits of the firms, we observe:
π ( ¯w)(w (0) + 1) + (N − 1)π ( ¯w)w (0) = π ( ¯w) −
N − 1
N
+
N − 1
N
= 0.
• Hence, aggregate profit remains constant. The complete burden
is attributed to the workers.
• For N large the partial equilibrium approximation regarding
prices and quantities is correct. However, the distributional
effects remain wrong.
369
General Equilibrium
Microeconomics
• First Fundamental Theorem of Welfare Economics
• Second Fundamental Theorem of Welfare Economics
MWG, Chapter 16
370
Notation (1)
Microeconomics
• Consider I consumers, indexed i = 1, . . . , I. Xi ⊂ RL
are the
consumption sets. The preferences are i. i is complete and
transitive (rationale consumers).
• J firms, indexed j = 1, . . . , J. The production possibility sets are
Yj ⊂ RL
. Yj is non-empty and closed. The production vectors are
yj.
• L goods, indexed = 1, . . . , L.
371
Notation (2)
Microeconomics
• The initial endowment of good is ¯ω ∈ RL
. The total
endowments are ¯ω = (¯ω1, . . . , ¯ωL) ∈ RL
.
• Basis data of the economy: ([Xi, i]I
i=1, [Yj]J
j=1, ¯ω).
• The wealth of consumer i is wi(p) = p · ωi.
372
Notation (3)
Microeconomics
• Definition - Economic Allocation [D 16.B.1]: An economic
allocation (x, y) = (x1, . . . , xI, y1, . . . , yJ) is a specification of a
consumption vector xi ∈ Xi for each consumer i = 1, . . . , I and
a production vector yj ∈ Yj for each firm j = 1, . . . , J. The
allocation is feasible if
I
i=1
x i = ¯ω +
J
j=1
y j for = 1, . . . , L.
This is
I
i=1 xi = ¯ω +
J
j=1 yj. We denote the set of feasible
allocations by
A := {(x, y) ∈ X1×· · ·×XI×Y1×· · ·×YJ :
I
i=1
xi = ¯ω +
J
j=1
yj} ⊂ R
L(I+J)
.
373
Notation (4)
Microeconomics
• Definition - Pareto Optimality [D 16.B.2]: A feasible allocation
(x, y) = (x1, . . . , xI, y1, . . . , yJ) is Pareto optimal (efficient) if
there is no other feasible allocation (x , y ) ∈ A that Pareto
dominates it. This is, if there is no feasible allocation (x , y ) such
that xi i xi for all i = 1, . . . , I and and xi i xi for some i.
374
Notation (5)
Microeconomics
• Suppose that consumer i initially owns ω i, where ¯ω =
I
i=1 ω i
for = 1, . . . , L, ωi = (ωi1, . . . , ωiL).
• Consumer i owns the shares θi = (θi1, . . . , θij, . . . , θiJ), where
I
i=1 θij = 1 for j = 1, . . . , J.
• Markets exist for all L goods and all firms are price takers; the
prices are p = (p1, . . . , pL).
375
Notation (6)
Microeconomics
• Definition [D 16.B.3] (Walrasian/Competitive Equilibrium)
– Given a private ownership economy by ([Xi, i]I
i=1, [Yj]J
j=1, ¯ω, θ). An
allocation (x∗
, y∗
) and the price vector p ∈ RL
constitute a competitive
(Walrasian) equilibrium if the following conditions are met:
∗ Profit maximization: For each firm j, y∗
j solves the profit maximization
problem, i.e.
p · yj ≤ p · y
∗
j for all yj ∈ Yj.
∗ Preference maximization: For each consumer i, x∗
i is maximal for i in
the budget set
{xi ∈ Xi : p · xi ≤ p · ωi +
J
j=1
θijp · y
∗
j }.
∗ Market clearing: For each good = 1, . . . , L:
I
i=1
x
∗
i = ¯ω +
J
j=1
y
∗
i or
I
i=1
x
∗
i = ¯ω +
J
j=1
y
∗
j .
376
Notation (7)
Microeconomics
• Definition [D 16.B.4] (Price Equilibrium with Transfers)
– Given a private ownership economy by ([Xi, i]I
i=1, [Yj]J
j=1, ¯ω, θ). An
allocation (x∗
, y∗
) and the price vector p ∈ RL
constitute a price
equilibrium with transfers if there is an assignment of wealth levels
(w1, . . . , wI) with
I
i=1 wi = p · ¯ω + j p · y∗
j such that
∗ For each firm j, y∗
j solves the profit maximization problem, i.e.
p · yj ≤ p · y
∗
j for all yj ∈ Yj.
∗ For each consumer i, x∗
i is maximal for i in the budget set
{xi ∈ Xi : p · xi ≤ p · wi}.
∗ Market clearing:
I
i=1 x∗
i = ¯ω +
J
j=1 y∗
j .
377
First Fundamental Theorem of Welfare
Economics (1)
Microeconomics
• Proposition [16.C.1] (First Fundamental Theorem of Welfare
Economics)
– If the preference relations i are locally nonsatiated and if
(x∗
, y∗
, p) is a price equilibrium with transfers, then the
allocation x∗
, y∗
is Pareto optimal. In particular, any Walrasian
equilibrium is Pareto optimal.
• Proof: See MWG page 549.
378
First Fundamental Theorem of Welfare
Economics (2)
Microeconomics
• The First Fundamental Theorem of Welfare Economics is on
Pareto optimality.
• Recall - Local Nonsatiation: For all x ∈ X and for all ε > 0 there
exists some y ∈ X such that ||x − y|| ≤ ε and y x. [D 3.B.3]
• Note that markets are complete and price taking is assumed.
379
Second Fundamental Theorem of Welfare
Economics (1)
Microeconomics
• First theorem: Given some assumptions and a price equilibrium with transfers
⇒ Pareto.
• Consider a competitive economy with transfers. Given some Pareto efficient
allocation (x, y). Does there exist a price system p which supports this
Pareto efficient allocation?
• Problem I: Convexity - see MWG, Figure 15.C.3 (a).
• Problem II: Minimum wealth problem - see MWG, Figure 15.B.10 (a). Let
xi ∈ R2
and consider the sequence pn = (1, 1/n), n = 1, 2, . . . . Let
w1 = (0, w21) = (0, ¯ω2). Then pn · w1 → 0. This is the minimum wealth
you can have on X = R2
. In the limit p = (1, 0) and the budget set is the
vertical axis intersected with R2
+. So everything of x2 costs nothing. The
budget correspondence is not continuous so Berge’s maximum theorem does
not apply.
• First investigate convexity. To do this we consider the concept of a
quasi-equilibrium.
380
Second Fundamental Theorem of Welfare
Economics (2)
Microeconomics
• Definition [16.D.1] (Price Quasi-equilibrium with Transfers)
– Given a private ownership economy by
([Xi, i]I
i=1, [Yj]J
j=1, ¯ω). An allocation (x∗
, y∗
) and the price
vector p = 0 constitute a price quasi-equilibrium with
transfers if there is an assignment of wealth levels
(w1, . . . , wI) with wi = p · ¯ω + j p · y∗
j such that
∗ For each firm j, y∗
j solves the profit maximization problem,
i.e.
p · yj ≤ p · y∗
j for all yj ∈ Yj.
∗ For each consumer i: If xi i x∗
i , then p · xi ≥ wi.
∗ Market clearing:
I
i=1 x∗
i = ¯ω +
J
j=1 y∗
j .
381
Second Fundamental Theorem of Welfare
Economics (3)
Microeconomics
• With local nonsatiation the second condition becomes: If xi x∗
i
then p · xi ≥ p · x∗
i .
• I.e. with local non-satiation, x∗
i minimizes the expenditures given
{xi : xi i x∗
i }.
382
Second Fundamental Theorem of Welfare
Economics (4)
Microeconomics
• Proposition [16.D.1] (Second Fundamental Theorem of Welfare
Economics)
– Consider an economy specified by ([Xi, i]I
i=1, [Yj]J
j=1, ¯ω), and
suppose that every Yj is convex and every preference relation
i is convex (the set {xi ∈ Xi : xi i xi} is convex for every
xi ∈ Xi) and locally non-satiated.
Then for every Pareto optimal allocation (x∗
, y∗
) there exists a
price vector p = 0 such that (x∗
, y∗
, p) is a price
quasi-equilibrium with transfers.
• Proof: See MWG, page 553.
383
Second Fundamental Theorem of Welfare
Economics (5)
Microeconomics
• When is a price quasi-equilibrium with transfers a price
equilibrium with transfers?
• The example considered in MWG, Figure 15.B.10 (a) and on
page 554, is a quasi-equilibrium but not an equilibrium.
• In this example the wealth of consumer 1 is zero (hence, zero
wealth problem).
• We need a sufficient condition under which which
”xi i x∗
i ⇒ p · xi ≥ wi” implies ”xi i x∗
i ⇒ p · xi>wi”.
384
Second Fundamental Theorem of Welfare
Economics (6)
Microeconomics
• Proposition [16.D.2]
– Assume that Xi is convex and i is continuous. Suppose also
that the consumption vector x∗
i ∈ Xi, the price vector p and
the wealth level wi are such that xi i x∗
i implies p · xi≥wi.
Then, if there is a consumption vector xi ∈ Xi such that
p · xi < wi [a cheaper consumption for (p, wi)], it follows that
xi i x∗
i implies p · xi>wi.
• Proof: See MWG page 555. See also MWG, Figure 16.D.3
(right).
385
Second Fundamental Theorem of Welfare
Economics (7)
Microeconomics
• Proposition [16.D.3]
– Suppose that for every i = 1, . . . , L, Xi is convex and i is
continuous. Then, any price quasi-equilibrium with transfers
that has (w1, . . . , wL) 0 is a price equilibrium with transfers.
• Proof: See MWG page 556.
386
Quasiconcave Functions
Motivation (1)
Microeconomics
• Motivation: Sufficient conditions for maximum for Kuhn Tucker
problem: Suppose that there are no nonlinear equality constraints
and each inequality constraint is given by a quasiconvex function.
Suppose that the objective function satisfies f(x)(x − x) > 0
for any x and x with f(x ) > f(x). If x∗
satisfied the
Kuhn-Tucker conditions, then x∗
is a global maximizer. (see
Mas-Colell, Theorem [M.K.3]).
• Jehle, Reny: Chapter A 1.4, 1.4.
• Mas-Colell, Chapter M.C
387
Quasiconcave Functions
Concave Functions (1)
Microeconomics
• Consider a convex subset A of Rn
.
• Definition - Concave Function: A function f : A → R is
concave if
f(νx + (1 − ν)x) ≥ νf(x ) + (1 − ν)f(x) , ν ∈ [0, 1].
If strict > holds then f is strictly concave; ν ∈ (0, 1) and x = x .
This last equation can be rewritten with z = x − x and α = ν:
f(x + αz) ≥ αf(x ) + (1 − α)f(x) .
• If f is (strictly) concave then −f is (strictly) convex. 388
Quasiconcave Functions
Concave Functions (2)
Microeconomics
• Theorem - Tangents and Concave Functions: If f is
continuously differentiable and concave, then
f(x ) ≤ f(x) + f(x)) · (x − x) (and vice versa). < holds if f is
strict concave for all x = x . [Theorem M.C.1]
• For the univariate case this implies that the tangent line is above
the function graph of f(x); strictly for x = x with strict concave
functions.
389
Quasiconcave Functions
Concave Functions (3)
Microeconomics
Proof:
• ⇒: For α ∈ (0, 1] the definition of a concave function implies:
f(x ) = f(x + z) ≤ f(x) +
f(x + αz) − f(x)
α
If f is differentiable the limit of the last term exists such that
f(x + z) ≤ f(x) + f(x) · z
390
Quasiconcave Functions
Concave Functions (4)
Microeconomics
Proof:
• ⇐: Suppose that f(x + z) − f(x) ≤ f(x) · z for any
non-concave function. Since f(.) is not concave
f(x + z) − f(x) >
f(x + αz) − f(x)
α
for some x, z and α ∈ (0, 1].
• Taking the limit results in f(x + z) − f(x) > f(x) · z, i.e. we
arrive at a contradiction.
391
Quasiconcave Functions
Concave Functions (5)
Microeconomics
• Theorem - Hessian and Concave Functions: If f is twice
continuously differentiable and concave, then the Hessian matrix
D2
f(x) is negative semidefinite; negative definite for strict
concave functions (and vice versa). [Theorem M.C.2]
392
Quasiconcave Functions
Concave Functions (6)
Microeconomics
Proof:
• ⇒: A Taylor expansion of f(x ) around the point α = 0 results in
f(x + αz) = f(x) + f(x) · (αz) +
α2
2
(z · D2
(f(x + β(α)z))z)
By the former theorem we know that
f(x + αz) − f(x) − f(x) · (αz) ≤ 0 for concave functions ⇒
z D2
(f(x + β(α)z))z ≤ 0. For arbitrary small α we get
z D2
(f(x))z ≤ 0.
393
Quasiconcave Functions
Concave Functions (7)
Microeconomics
Proof:
• ⇐: If the right hand side of
f(x +αz)−f(x)− f(x)·(αz) = 0.5α2
(z D2
(f(x +β(α)z))z)
is ≤ 0 then the left hand side. By the former theorem f is
concave.
394
Quasiconcave Functions
Quasiconcave Functions (1)
Microeconomics
• Definition - Quasiconcave Function: A function f : A → R is
quasiconcave if
f(νx + (1 − ν)x) ≥ min{f(x ), f(x)} , ν ∈ [0, 1].
If > holds it is said to be strict quasiconcave; ν ∈ (0, 1) and
x = x .
• Quasiconvex is defined by f(νx + (1 − ν)x) ≤ max{f(x ), f(x)}.
If f is quasiconcave than −f is quasiconvex.
• If f is concave then f is quasiconcave but not vice versa. E.g.
f(x) =
√
x for x > 0 is concave and also quasiconcave. x3
is
quasiconcave but not concave.
395
Quasiconcave Functions
Quasiconcave Functions (2)
Microeconomics
• Transformation property: Positive monotone transformations of
quasiconcave functions result in a quasiconcave function.
• Definition - Superior Set: S(x) := {x ∈ A|f(x ) ≥ f(x)} is
called superior set of x (upper contour set of x).
• Note that if f(xν
) ≥ min{f(x ), f(x )}, then if f(x ) ≥ t and
f(x ) ≥ t this implies that f(xν
) ≥ t; where t = f(x).
396
Quasiconcave Functions
Quasiconcave Functions (3)
Microeconomics
• Theorem - Quasiconcave Function and Convex Sets: The
function f is quasiconcave if and only if S(x) is convex for all
x ∈ A.
397
Quasiconcave Functions
Quasiconcave Functions (4)
Microeconomics
Proof:
• Sufficient condition ⇒: If f is quasiconcave then S(x) is convex.
Consider x1
and x2
in S(x). We need to show that f(xν
) in
S(x); f(x) = t.
• Since f(x1
) ≥ t and f(x2
) ≥ t, the quasiconcave f implies
f(xν
) ≥ min{f(x1
), f(x2
)} ≥ t.
• Therefore f(xν
) ∈ S(x); i.e. the set S(x) is convex.
398
Quasiconcave Functions
Quasiconcave Functions (5)
Microeconomics
Proof:
• Necessary condition ⇐: If S(x) is convex then f(x) has to be
quasiconcave. W.l.g. assume that f(x1
) ≥ f(x2
), x1
and x2
in A.
• By assumption S(x) is convex, such that S(x2
) is convex. Since
f(x1
) ≥ f(x2
), we get x1
∈ S(x2
) and xν
∈ S(x2
).
• From the definition of S(x2
) we conclude that
f(xν
) ≥ f(x2
) = min{f(x1
), f(x2
)}.
• Therefore f(x) has to be quasiconcave.
399
Quasiconcave Functions
Quasiconcave Functions (6)
Microeconomics
• Theorem - Gradients and Quasiconcave Functions: If f is
continuously differentiable and quasiconcave, then
f(x) · (x − x) ≥ 0 whenever f(x ) ≥ f(x) (and vice versa).
[Theorem M.C.3]
• If f(x) · (x − x) > 0 whenever f(x ) ≥ f(x) and x = x then
f(x) is strictly quasiconcave. If f(x) is strictly quasiconcave and
if f(x) = 0 for all x ∈ A, then f(x) · (x − x) > 0 whenever
f(x ) ≥ f(x) and x = x .
400
Quasiconcave Functions
Quasiconcave Functions (7)
Microeconomics
Proof:
• ⇒: For f(x ) ≥ f(x) and α ∈ (0, 1] the definition of a
quasiconcave function implies:
f(x + α(x − x)) − f(x)
α
≥ 0
If f is differentiable, then the limit exists such that
f(x) · z ≥ 0
401
Quasiconcave Functions
Quasiconcave Functions (8)
Microeconomics
Proof:
• ⇐: Suppose that f(x) · z ≥ 0 holds but f is not quasiconcave.
Then f(x + αz) − f(x) < 0 for some x, z and α ∈ (0, 1]. Such
that (f(x + αz) − f(x))/α < 0. Taking the limit results in a
contradiction.
402
Quasiconcave Functions
Quasiconcave Functions (9)
Microeconomics
• Theorem - Hessian Matrix and Quasiconcave Functions:
Suppose f is twice continuously differentiable. f(x) is
quasiconcave if and only if D2
(f(x)) is negative semidefinite in
the subspace {z| f(x) · z = 0}. I.e. z D2
(f(x))z ≤ 0 whenever
f(x) · z = 0. [Theorem M.C.4]
• If the Hessian D2
(f(x)) is negative definite in the subspace
{z| f(x) · z = 0} for every x ∈ A then f(x) is strictly
quasiconcave.
403
Quasiconcave Functions
Quasiconcave Functions (10)
Microeconomics
Proof:
• ⇒: If f is quasiconcave then whenever f(xν
) ≥ f(x), so
f(x) · (αz) ≥ 0 has to hold.
• Thus f(x1
) − f(x) ≤ 0 and the above theorem imply:
f(x) · (z) ≤ 0, where z = x1
− x.
• A first order Taylor series expansion of f in α (at α = 0) results
in
f(x + αz) = f(x) + f(x)αz +
α2
2
· z D2
f(x + β(α)z)z .
404
Quasiconcave Functions
Quasiconcave Functions (11)
Microeconomics
Proof:
• Apply this to x1
, x with f(x1
) ≤ f(x):
f(x + αz) − f(x) − f(x)αz =
α2
2
· z D2
f(x + β(α)z)z.
• If z = x1
− x fulfills f(x)(x1
− x) = 0 the above inequality still
has to hold.
• This implies α2
/2z D2
f(x + β(α)z)z ≤ 0.
405
Quasiconcave Functions
Quasiconcave Functions (12)
Microeconomics
Proof:
• To fulfill this requirement on the subspace {z| f(x) · z = 0},
where f(x)αz = 0, this requires a negative definite Hessian of
f(x).
• ⇐: In the above equation a negative semidefinite Hessian implies
that . . . .
406
Envelope Theorem (1)
Microeconomics
• Consider f(x; q), x are variables in RN
and q are parameters in
RS
.
• We look at the constrained maximization problem
max
x
f(x; q) s.t. gm(x; q) ≤ bm
m = 1, . . . , M.
• Assume that the solution of this optimization problem x∗
= x(q)
is at least locally differentiable function (in a neighborhood of a ¯q
considered).
• v(q) = f(x(q); q) is the maximum value function associated with
this problem.
407
Envelope Theorem (2)
Microeconomics
• With no constraints (M = 0) and S, N = 1 the chain rule yields:
d
dq
v(¯q) =
∂f(x(¯q); ¯q)
∂x
∂x(¯q)
∂q
+
∂f(x(¯q); ¯q)
∂q
.
• With an unconstrained maximization problem the first order
condition ∂f(x(¯q);¯q)
∂x = 0 results in
d
dq
v(¯q) =
∂f(x(¯q); ¯q)
∂q
.
408
Envelope Theorem (3)
Microeconomics
[T. M.L.1] Consider the value function v(q) for the above
constrained maximization problem. Assume that v(q) is
differentiable at ¯q ∈ RS
and (λ1, . . . , λM) are the Lagrange
multipliers associated with the maximizing solution x(q) at ¯q. In
addition the inequality constraints remain unaltered in a
neighborhood of ¯q. Then
∂v(¯q)
∂qs
=
∂f(x(¯q); ¯q)
∂qs
−
M
m=1
λm
∂gm(x(¯q); ¯q)
∂qs
.
For s = 1, . . . , S.
409
Envelope Theorem (4)
Microeconomics
Proof:
• Let x(.) stand for the maximizer of the function f(.) and
v(q) = f(x(q), q) for all q. The chain rule yields:
∂v(¯q)
∂qs
=
N
n=1
∂f(x(¯q); ¯q)
∂xn
∂xn(¯q)
∂qs
+
∂f(x(¯q); ¯q)
∂qs
.
• The first order conditions tell us
∂f(x(¯q); ¯q)
∂xn
=
M
m=1
λm
∂gm(x(¯q); ¯q)
∂xn
.
410
Envelope Theorem (5)
Microeconomics
Proof:
• In addition we observe
N
n=1
∂gm(x(¯q); ¯q)
∂xn
∂xn(¯q)
∂qs
+
∂gm(¯q)
∂qs
= 0.
if a constraint is binding; if not the multiplier λm is zero.
411
Envelope Theorem (6)
Microeconomics
Proof:
• Plugging in and changing the order of summation results in :
∂v(¯q)
dqs
=
M
m=1
λm
N
n=1
∂gm(x(¯q); ¯q)
∂xn
∂xn(¯q)
∂qs
+
∂f(x(¯q); ¯q)
∂qs
.
• and
∂v(¯q)
dqs
= −
M
m=1
λm
∂gm(x(¯q); ¯q)
∂qs
+
∂f(x(¯q); ¯q)
∂qs
.
• Remark: remember that the Lagrangian of the problem is
L(x, λ; q) = f(x; q) − m λmgm(x; q). Hence we get ∂v(¯q)
dqs
by
means of the partial derivative of the Lagrangian with respect to
ql, evaluated at ¯q.
412