0% found this document useful (0 votes)
157 views26 pages

Problem Set I: Preferences, W.A.R.P., Consumer Choice: Paolo Crosetto Paolo - Crosetto@unimi - It

This document provides an overview and solutions to problem set 1 on preferences, weak axiom of revealed preference (WARP), and consumer choice. It begins with definitions of strict preference, indifference, and rational preference relations. It then proves various properties of these relations, including that strict preference is irreflexive and transitive, and indifference is reflexive, transitive, and symmetric. It shows that a strictly increasing transformation of a utility function yields another valid utility function. It proves existence of a utility function representing a rational preference on a finite set. Finally, it recaps the definitions of WARP and revealed preference relation. The document provides proofs and intuition for the concepts and properties discussed.

Uploaded by

damla87
Copyright
© Attribution Non-Commercial (BY-NC)
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
0% found this document useful (0 votes)
157 views26 pages

Problem Set I: Preferences, W.A.R.P., Consumer Choice: Paolo Crosetto Paolo - Crosetto@unimi - It

This document provides an overview and solutions to problem set 1 on preferences, weak axiom of revealed preference (WARP), and consumer choice. It begins with definitions of strict preference, indifference, and rational preference relations. It then proves various properties of these relations, including that strict preference is irreflexive and transitive, and indifference is reflexive, transitive, and symmetric. It shows that a strictly increasing transformation of a utility function yields another valid utility function. It proves existence of a utility function representing a rational preference on a finite set. Finally, it recaps the definitions of WARP and revealed preference relation. The document provides proofs and intuition for the concepts and properties discussed.

Uploaded by

damla87
Copyright
© Attribution Non-Commercial (BY-NC)
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 26

Problem Set I: Preferences, W.A.R.P.

, consumer choice
Paolo Crosetto
[email protected]
Exercises solved in class on 18th January 2010
Recap: , ~,
Denition
The strict preference relation ~ is
x ~ y x y but not y x
Denition
The indierence relation is
x y x y and y x
2 / 26
Recap: rationality assumptions
is rational if it is

Complete: x, y X, we have x y or y x or both;

Transitive: x, y, z X, if x y and y z, then x z.


3 / 26
Recap: and utility function u()
Denition
A function u : X R is a utility function representing if
x, y X : x y u(x) u(y)
4 / 26
1. MWG, Exercise 1.B.1 + 1.B.2: properties of
Prove that if is rational (complete and transitive), then
1. ~ is both irreexive (x ~ x never holds) and transitive (if x ~ y and y ~ z,
then x ~ z);
2. is reexive (x x, x), transitive (if x y and y z, then x z) and
symmetric (if x y then y x);
3. if x ~ y z then x ~ z.
5 / 26
Solution: property 3 rst
Proof.
Property 3: if x ~ y z then x ~ z
1. By denition, x ~ y means that x y but not y x;
2. then, x ~ y z means x y z;
3. for transitivity (assumed), this means that x z.
4. Now, lets suppose that z x. Since y z, by transitivity wed have y x
5. but this is a contradiction, since we had in the beginning that x ~ y.
6. So, we have x z but we cannot have z x: this means that x ~ z
6 / 26
Solution: property 1
Proof.
Property 1: ~ is irreexive and transitive
1. Irreexivity. Use completeness: x y, x, y X:
2. hence, it must hold also for x x, x X;
3. this means that in no case there can be x ~ x.
4. Transitivity. Suppose x ~ y and y ~ z:
5. this means that at least x ~ y z.
6. But we have proved before that this means x ~ z.
7 / 26
Solution: property 2
Proof.
Property 2: is reexive, transitive and symmetric
1. Reexivity. By completeness, x x, x X:
2. this implies also that x x, x X, by denition of .
3. Transitivity. Suppose x y and y z:
4. by the denition of , this means that all of these hold:
5. x y, y x, y z, z y.
6. By transitivity of , this implies both x z and z x: hence x z.
7. Symmetry. Suppose x y: by denition, then x y and y x.
8. But the latter is also the denition of y x, if you look it the other way
around.
9. hence, x y implies y x.
8 / 26
2. MWG 1.B.3 + 1.B.4.: and u()

Show that if f : R R is a strictly increasing function and u : X R is a


utility function representing the preference relation , then the function
v : X R dened by v(x) = f (u(x)) is also a utility function representing ;

Consider a preference relation and a function u : X R. Show that if


u(x) = u(y) implies x y and if u(x) > u(y) implies x ~ y then u() is a
utility function representing .
9 / 26
Solution: strictly increasing function, intuition

We will prove that a utility function associated with is ordinal and not
cardinal in nature.

This is important: among other things, it implies that it is impossible to make


interpersonal utility comparisons directly.

Note the denition of strictly increasing function:


Denition
A function f (x) is said to be strictly increasing over an interval I if f (b) > f (a)
for all b > a, when a, b I .
Example
Functions that are strictly increasing over their whole domain are among others all
positive straight lines (y = ax, a > 0) and positive exponentials (y = a
x
, a > 0);
other functions can be increasing over a part of their domain, as parabola (y = x
2
,
for x > 0).

Tip: a strictly increasing function on interval I has its derivative positive on I .


10 / 26
Solution: strictly increasing functions, plots
11 / 26
Solution: strictly increasing function, proof
Proof.
A strictly increasing transformation of a utility function is still a utility function
1. Lets take x, y X. Since u() represents , by denition:
2. if x y then u(x) u(y).
3. since f () is strictly increasing, applying f () to u() does not change order,
but only magnitude;
4. hence, f (u(x)) f (u(y)), i.e. v(x) v(y) when x y:
5. hence, v() is a utility function representing .
12 / 26
Solution: u(), ~ and .
Proof.
if x y, then u(x) u(y)
1. Suppose x y.
2. if at this we add y x, then x y and u(x) = u(y).
3. if instead we dont have y x, then x ~ y and u(x) > u(y).
4. hence, if x y, then u(x) u(y)
Proof.
if u(x) u(y), then x y
1. Suppose u(x) u(y).
2. if at this we add u(x) = u(y), then x y.
3. if instead we add u(x) > u(y), then x ~ y.
4. hence, if u(x) u(y), then x y.
13 / 26
3. MWG 1.B.5: and u(), II
Show that if X is nite and is a rational preference relation on X, then there is
a utility function u : X R that represents .
14 / 26
Solution: intuition

Since X is nite, the set of pairwise combination of elements of X is nite too;

Since is rational (hence complete and transitive):

it denes a preference over all of the nite set of pairs;

it excludes contradictory cycles of preferences.

Hence, intuitively it is possible to rank all x, y X according to ;

it must then be possible to build a utility function with such a complete


ranking using .
15 / 26
Solution: proving by induction

A proof by induction is done by showing that something is true for n = 1 and


then for n + 1;

it then follows that it must be true for all n up to N.

Proof by induction is used in the set of natural numbers N.

More formally, for any proposition P(n) about positive integers:

Prove that P(1) is true (base case);

Prove that for each k 1, if P(k) is true, then P(k + 1) is true (inductive
step).
Example
Consider a set of domino tiles. If domino tile n falls, tile n + 1 will fall. If we prove
that tile 1 has fallen, then we can conclude that all tiles will fall.
16 / 26
Solution: proof, x y
Proof.
if X is nite, then there exists a u() representing : no indierence
1. Start considering that no two items are indierent, i.e. x y, x, y X;
2. Lets prove by induction that in such a setting there exists a u() representing
.
3. Base case: if N = 1 there is nothing to prove.
4. Inductive step: Lets suppose the claim is true for N 1, and lets prove it is
still ture for N.
1. Lets take X = x
1
, x
2
, . . . , x
N1
, x
N
.
2. By hypothesis, there exists a u() on dened up to x
N1
.
3. Lets order the x: lets assume u(x
1
) > u(x
2
) > > u(x
N1
).
4. Since we have assumed no indierence, the above ranking means exclusively:

i < N, x
N
~ x
i
i < N, x
i
~ x
N
i < N and j < N s.t. x
i
~ x
N
~ x
j
17 / 26
Solution: proof, x y continued
Proof.
...continued
In all the three cases above we can nd a value of u() that is consistent:
1. In Case 1, we can take u(N) > u(x
1
);
2. In Case 2, we can take u(N) < u(x
N
1);
3. in Case 3:

Dene two intervals I = i (1 . . . N) : x


i
~ x
N
and J = j (1 . . . N) : x
N
~ x
j
;

I and J are disjoint intervals on N by our hypotheses;

then if i

= max I , i

+1 = min J.

We can then take u(x


N
) to lie in the interval (u(i

), u(i

+1)).
Hence, in all of three cases an utility function can be built.
18 / 26
Recap: W.A.R.P.,

Denition (WARP)
A choice structure (B, C()) satises the weak axiom if for some B B with
x, y B we have x C(B), then for any B
/
B with x, y B
/
, if we have
y C(B
/
) we must also have x C(B
/
)

Which is indeed a minimal consistency requirement. Note that completeness


and transitivity are not required.
Denition (Revealed preference relation

)
Given a choice structure (B, C()), the revealed preference relation

is dened
as:
x

y there is some B B such that x, y B and x C(B).

Which is just attaching a preference relationship to choices

Note again that nor completeness nor transitivity are implied. It is just
descriptive.
19 / 26
Recap: Rationalizability
Rational preferences W.A.R.P. satised always
W.A.R.P. satised Rational preferences not always
Denition (Rationalizability)
Given a choice structure (B, C()), the rational preference relation rationalizes
C() relative to B if C(B) = C

(B, ) for all B B. In other words, generates


the choice structure (B, C()).

the W.A.R.P. is a necessary but not sucient condition for rationalizability.

if B includes all subsets of X of up to three elements, then it is also sucient:

intuitively, the three-members property implies transitivity...


20 / 26
4. Exercise on W.A.R.P.
Consider a choice problem with choice set X = x, y, z. Consider the following
choice structures:

(B
/
, C()), in which B
/
= x, y, y, z, x, z, x, y, z and
C(x, y) = x, C(y, z) = y, C(x, z) = z, C(x) =
x, C(y) = y, C(z) = z.

(B
//
, C()), in which B
//
= x, y, z, x, y, y, z, x, z, x, y, z and
C(x, y, z) = x,
C(x, y) = x, C(y, z) = z, C(x, z) = z, C(x) =
x, C(y) = y, C(z) = z.

(B
///
, C()), in which B
///
= x, y, z, x, y, y, z, x, z, x, y, z
and C(x, y, z) = x,
C(x, y) = x, C(y, z) = y, C(x, z) = x, C(x) =
x, C(y) = y, C(z) = z.
For every choice structure say if the WARP is satised and if it exists a rational
preference relation that rationalizes C() relative to its B. If such a
rationalization is possible, write it down. Comment on your results.
21 / 26
Solution: (B
/
, C())
The choice structure can be summarised in these three relations:

C(x, y) = x reveals x

y;

C(y, z) = y reveals y

C(x, z) = z reveals z

x
1. W.A.R.P. is trivially satised

the same couple never appears more than once in dierent budgets;

moreover, B
/
does not include all budgets up to three elements.

revealed preference relation is not necessarily transitive


2. B
/
is NOT rationalizable:

C(x, y) = x is rationalised by x ~ y;

C(y, z) = y is rationalised by y ~ z;

C(x, z) = z is rationalised by z ~ x.

It is not transitive, hence (B


/
, C()) is not rationalisable.
22 / 26
Solution: (B
//
, C())
The choice structure can be summarised in these relations:

C(x, y, z) = x reveals x

y and x

C(x, y) = x reveals x

y;

C(y, z) = y reveals y

z;

C(x, z) = z reveals z

x.
1. W.A.R.P. is NOT satised

z and z

x hold at the same time;

in this case it exists x, z B : C(B) = x, but there is also...

...a x, z B
/
: z C(B
/
) but not x C(B
/
)
2. B
//
is NOT rationalisable:

since in general if is rational

satises W.A.R.P.;

then, by using the contrapositive, if A B, it must be true that B A

Hence (B
//
, C()) is not rationalisable
23 / 26
Solution: (B
///
, C())
The choice structure can be summarised in these relations:

C(x, y, z) = x reveals x

y and x

C(x, y) = x reveals x

y;

C(y, z) = y reveals y

z;

C(x, z) = x reveals x

z.
1. W.A.R.P. is satised

there are no violations of the type x

y and y

x;

moreover, B
///
includes all budgets up to three elements.
2. B
///
is rationalizable:

C(x, y, z) = x reveals x ~ y and x ~ z

C(x, y) = x reveals x ~ y;

C(y, z) = y reveals y ~ z;

C(x, z) = x reveals x ~ z.

Hence x ~ y ~ z is complete and transitive and rationalises (B


///
, C())
24 / 26
5. MWG 1.D.2: and W.A.R.P.
Show that if X is nite, then any rational preference relation generates a nonempty
choice rule; that is, C(B) ,= for any B X with B ,= .
25 / 26
Solution
Proof.
X nite C(B) ,=
1. We proved earlier that if X is nite, then u() is a utility function representing
a rational . (by induction. Remember??)
2. Since X is nite, for any B X with B ,= there exists x C(B) such that
u(x) u(y) for all y B...
3. ...remember that niteness implied that we could order all alternatives in X,
and assign a value.
4. Then, it means that x C

(B, ), i.e. x is chosen according to preference


relation in B.
5. Hence, C

(B, ) cannot be empty: C

(B, ) ,=
26 / 26

You might also like