Download as pdf or txt
Download as pdf or txt
You are on page 1of 6

Microeconomics-I: Section 2

Micro-1 TAs∗
September 13, 2023

1.C.1 Consider the choice structure (B, C(·)) with B = ({x, y}, {x, y, z}) and C({x, y}) =
x. Show that if (B, C(·)) satisfies the weak axiom, then we must have C({x, y, z}) =
{x}, = {z}, or = {x, z}.

Solution:

Firstly, by definition, C({x, y, z}) ̸= ∅. Then it is sufficient to show that y ̸∈ C({x, y, z}).
Proof by contradiction. Assume C({x, y, z}) ̸= ∅ and use WARP for B = {x, y, z} and
B ′ = {x, y}:

[x ∈ B, y ∈ C(B)] & [y ∈ B ′ , x ∈ C(B ′ )] ====⇒ y ∈ C(B ′ )


W ARP

, but we have C(B ′ ) = {x}, so we obtain contradiction, so y ̸∈ C({x, y, z}).

Notice, that all options {x}, {z}, {x, z} are possible. Indeed, the only pair of alternatives
that belongs to both sets in B is x, y. However, y has not been chosen in any of them.
Hence, the contradiction can not be built in terms of WARP. □

1) Let X = Rk for some k ≥ 2, with a generic element x = (x1 , . . . , xk ). The pointwise


order ≥ on Rk is defined as

x ≥ y ⇐⇒ ∀ i = 1, . . . , k, xi ≥ yi .

Determine if ≥ is complete or transitive. (Explain your answer.)

Solution:

Not complete. Counter-example: x = (1, 0, 0, . . . , 0) , y = (0, 1, 0, . . . , 0). Indeed,


x2 < y2 =⇒ ¬(x ≥ y) and y1 < x1 =⇒ = ̸ (y ≥ x).
Transitive. Proof : x ≥ y ≥ z =⇒ ∀ i = 1, . . . , k, xi ≥ yi ≥ zi =⇒ x ≥ z □


Section notes are prepared by several generations of Micro-1 TAs: Olga Rospuskova (2022), Alexey
Belyaev (2022), Alena Gorbuntsova (2023), Anastas Akopian (2023). Please contact your TA if a
mistake is found!

1
2) Let X = R3 , and define a binary relation R on X by the following rule:
3
X
∀ x, y ∈ X, x R y ⇐⇒ Q (x, y) ≥ 0, where Q (x, y) =
p
sign (xi − yi ) |xi − yi |
i=1

Here, sign α is the sign of a real number. Note that Q (x, y) = −Q (y, x).
(i) Determine if R is complete of not.

(ii) Determine if R is transitive or not.


Solution:

(i) Complete. We need to show that ∀x, y ∈ X, xRy or yRx. Indeed, Q is defined
for any pair x, y ∈ R3 and maps it into R, on which natural ≥ is complete, so it is
always true that either Q (x, y) ≥ 0 or 0 ≥ Q (x, y). So, provided with Q (x, y) =
−Q (y, x) we immediately get that it is always true that either Q (x, y) ≥ 0 or
Q (y, x) ≥ 0, which completes the proof.

(ii) Not transitive. We need to find a counter-example: x, y, z ∈ R3 : x ≥ y, y ≥


z, but ¬ (x ≥ z). Let x = (1, 1, 0), y = (0, 1, 1), z = (α, 0, 0), α > 0. One could
check that x ≥ y (actually, x ∼ y). Let’s also write conditions on y ≥ z and
¬ (x ≥ z):  √
− α + 1 + p
 1≥0
sign (1 − α) |1 − α| + 1 < 0 ⇐⇒ 2 < α ≤ 4

α>0

So, e.g. α = 3 delivers counter-example.


A binary relation R on a set X is said to be


• asymmetric: x R y =⇒ ¬(y R x)

• symmetric: x R y =⇒ (y R x)

• negatively transitive: ¬(x R y) and ¬(y R z) ⇒ ¬(x R z)


3) Let ≿ be a complete and transitive preference relation on a set X. Consider the
associated strict preference relation ≻ and the indifference relation ∼. Show that:
(i) ≻ is transitive and asymmetric.

(ii) ∼ is transitive and symmetric (that is, x ∼ y ⇒ y ∼ x).

(iii) x ≻ y and y ≿ z imply x ≻ z.


Solution:

2
(i) Firstly, let me write definition of these preferences:
∀x, y ∈ X, x ≻ y ⇐⇒ x ≿ y & ¬[y ≿ x]
Transitivity: Let’s consider any x, y, z from X such that [x ≻ y & y ≻ z].
Firstly, [x ≻ y & y ≻ z] ⇐⇒
⇐⇒ [x ≿ y & ¬(y ≿ x) & y ≿ z & ¬(z ≿ y)] ============⇒ [x ≿ z].
by transitivity of ≿
Now let z ≿ x. Then by transitivity of ≿:
z ≿ x & x ≿ y =⇒ z ≿ y, but it contradicts one of the conditions above:
¬(z ≿ y), so ¬(z ≿ x) and we obtained second part of definition of z ≻ x, so ≻ is
indeed transitive.
Asymmetry: We should prove that ∀x, y ∈ X, x ≻ y =⇒ ¬(y ≻ x). Let’s
consider any x, y ∈ X : x ≻ y. Proof by contradiction: let’s also suggest y ≻ x.
From transitivity, which is already proven, we get y ≻ y ⇐=======⇒ y ≿ y &
by def inition
¬(y ≿ y), so we have contradiction, which proves asymmetry of ≻.

(ii) Firstly, let me write definition of these preferences:


∀x, y ∈ X, x ∼ y ⇐⇒ x ≿ y & y ≿ x
Transitivity: proof is straightforward. ∀x, y, z ∈ X, [x ∼ y & y ∼ z] ⇐=======⇒
by def inition
[x ≿ y & y ≿ x & y ≿ z & z ≿ y] ⇐==============================⇒
by transitivity of ≿: group (1,3) relations and (2,4)
[x ≿ z & z ≿ x] ⇐=======⇒ x ∼ z.
by def inition
Symmetry: obvious: ∀x, y ∈ X, x ∼ y ⇐===⇒ x ≿ y & y ≿ x ⇐==========⇒
by def. A & B ⇐⇒ B & A
y ≿ x & x ≿ y ⇐===⇒ y ∼ x.
by def.

(iii) Firstly, x ≻ y & y ≿ z ========⇒ x ≿ y & y ≿ z =========⇒ x ≿ z.


by def. of ≻ by transitivity
Secondly, let z ≿ x. Therefore: y ≿ z & z ≿ x =====⇒ y ≿ x. But it contradicts
by trans.
x ≻ y, so we obtain ¬(z ≿ x).
Together these two statements complete the proof. □

It is important to get familiar with this notation, so reading this proof is


exercise itself (in case one is new to this notation). But if anything was not
clear in it, here is text version (some proofs are a little bit different).

(i) Pick any x, y, z ∈ X such that x ≻ y and y ≻ z. to prove transitivity we need


to show x ≻ z. By definition of strict preference relation x ≻ y ⇒ x ≿ y and
y ≻ z ⇒ y ≿ z. Then, by transitivity of weak preference relation, x ≿ z. Finally,
we need to show that ¬(z ≿ x). Assume the opposite, that is z ≿ x. Since we also
have x ≿ y, by transitivity we obtain z ≿ y, however it contradicts to the fact
y ≻ z. Thus, ¬z ≿ x. The two facts x ≿ z and ¬z ≿ x imply that x ≻ z.
To prove asymmetry, pick any x, y ∈ X with x ≻ y. By definition of strict
preference relation we have x ≿ y and ¬(y ≿ x). The last fact implies ¬y ≻ x. So,
x ≻ y ⇒ ¬(y ≻ x) as we seek.

(ii) Pick any x, y, z ∈ X such that x ∼ y and y ∼ z. to prove transitivity we need


to show x ∼ z. By definition of strict preference relation x ∼ y ⇒ x ≿ y and
y ∼ z ⇒ y ≿ z. Thus, by transitivity of weak preference relation x ≿ z. At the
same time, x ∼ y ⇒ y ≿ x and y ∼ z ⇒ z ≿ y. Again, by transitivity, z ≿ x.

3
Since x ≿ z and z ≿ x, x ∼ z as we seek.
To prove symmetry, pick any x, y ∈ X with x ∼ y. By definition of indifference
preference relation, we have x ≿ y and y ≿ x. The last two facts imply y ∼ x.
Thus, x ∼ y ⇒ y ∼ x as we seek.

(iii) Firstly, x ≻ y implies x ≿ y. By transitivity of weak preference relations from


x ≿ y and y ≿ z it follows x ≿ z. Finally, we need to show that ¬(z ≿ x). Suppose
the opposite, that is z ≿ x. Then, y ≿ z and z ≿ x together imply y ≿ x by
transitivity, but it contradicts to x ≻ y. Thus, x ≿ z and ¬(z ≿ x), that is x ≻ z
as we seek.

4) Consider a binary relation P on a set X that is


asymmetric: xP y ⇒ ¬(yP x), and
negatively transitive: ¬(xP y) and ¬(yP z) ⇒ ¬(xP z).
Define a further binary relation ≿P on X as follows:

x ≿P y ⇐⇒ ¬(yP x).

Show that ≿P is a complete and transitive binary relation. Moreover, its strict part ≻P
equals P . (Conclusion: We can as well take a strict preference P as our primitive, and
derive from that a weak preference ≿P ).

Solution:

Completeness: Assume that x ̸≿p y, then yP x (Indeed, if we assume the opposite,


¬(yP x), then by the definition of binary relation ≿p we get x ≿p y, that contradicts
to our assumption). Then, due to asymmetry ¬(xP y). By the definition of ≿p , we get
y ≿p x.
Conclusion: if one alternative, x, is not preferred to another alternative, y, that is x ̸≿p y,
then with necessity y is preferred to x: y ≿p x, which means that ≿p is complete.

Transitivity: Assume x ≿p y and y ≿p z ⇒ ¬(yP x) and ¬(zP y) by definition. Then,


due to negative transitivity ¬(zP x). Then, by the definition of binary relation ≿p we
get x ≿p z.
So, x ≿p y & y ≿p z ⇒ x ≿p z, which means that ≿p is transitive.

Equivalence of strict part ≻p to P:

i) Assume x ≻p y, then x ≿p y & y ̸≿p x. Suppose ¬(xP y), then by definition y ≿p x,


that contradicts to y ̸≿p x. Contradiction ⇒ xP y

ii) Assume xP y. The, due to asymmetry xP y ⇒ ¬(yP x). Then, by definition x ≿p y.


Finally, notice that y ̸≿p x. Suppose the opposite: y ≿p x, then ¬(xP y). This
contradicts to the initial assumption.
As x ≿p y and y ̸≿p x, we obtain x ≻p y.

4

5) Consider a choice structure (B, C(·)) on a set X. Show that if C is single valued,
that is, if a unique alternative is chosen from each B ∈ B, then WARP is equivalent to
the following statement:
For any x, y ∈ X with x ̸= y, if x ∈ C(B) and y ∈ B for some B ∈ B, then for any
B ′ ∈ B with y ∈ C(B ′ ), we have x ∈ / B′.
Interpret this statement in terms of the revealed preferences of the agent.

Solution:

Definition. The choice structure (B, C(·)) satisfies the weak axiom of revealed
preference (WARP) if the following property holds:

If for some A ∈ B with x, y ∈ A we have x ∈ C(A), then for any B ∈ X with


x, y ∈ B and y ∈ C(B), we must also have x ∈ C(B) .

Firstly, let’s prove that WARP implies the statement . Assume that statement does not
hold. That is, for some x, y ∈ X(x ̸= y) and for some B ∈ B : x ∈ C(B) and y ∈ B,
and there exists B ′ ∈ B such that y ∈ C(B ′ ) and x ∈ B ′ .
That is, x, y ∈ B ∩ B ′ , x ∈ C(B), y ∈ C(B ′ ). Then, by WARP x ∈ C(B ′ ). But since
C(·) is single-valued, it cannot be the case. Contradiction.

Secondly, let’s prove that statement implies WARP. Let’s formulate WARP in terms
of 2 implications: [∃B ∈ B : x, y ∈ B & x ∈ C(B)] =⇒ [∀B ′ : (x ∈ B ′ &
y ∈ C(B ′ )) =⇒ x ∈ C(B ′ )]. If x = y in WARP then WARP is always true (it
becomes tautological), so this case does not violate implication. Further I consider
x ̸= y. To understand, why provided with statement we obtain WARP, one should
know that if premise of implication is false, than implication itself is true. Actually,
given statement, [x ∈ B ′ & y ∈ C(B ′ )] is always false (because statement states that
y ∈ C(B ′ ) =⇒ x ̸∈ B ′ ), so premise of inner implication of WARP is always false, so
inner implication is always true, so outer implication is always true. □

Definition Given a choice structure (B, C(·)) the revealed preference relation ≿∗
is defined by

x ≿∗ y ⇐⇒ there is some B ∈ B such that x, y ∈ B and x ∈ C(B)

Thus, for a single-valued C(·), the statement is equivalent to the following:

If x is revealed preferred to y, then y cannot be revealed preferred to x.

5
6) One more equivalence to WARP for a single-valued choice
function.
Consider a choice structure (B, C(·)) on a set X. If C is single valued, then WARP is
equivalent to the following statement:

For any pair of sets B and B ′ , if C(B ′ ) belongs to B, and C(B ′ ) is not equal to C(B),
then C(B) does not belong to B ′ .

Proof:
In other words, we need to show:

W ARP ⇐⇒ ∀B, B ′ ∈ B : C(B ′ ) ∈ B & C(B ′ ) ̸= C(B) ⇒ C(B) ∈


/ B′
Consider any pair B, B ′ ∈ B: C(B) = x, C(B ′ ) = y, x ̸= y. So, for these x, y we have
y ∈ B, x ∈ C(B) and y ∈ C(B ′ ). Then, x ∈ / B ′ . But this is equivalent to statement (b)
from task 4) which, in turn, is equivalent to WARP.

7) Proposition 2.F.1 from MWG (pp. 30-32)

You might also like