ECO 503 Problem Set 2 Solutions

1. We have u (x1 , x2 ) = g (x1 ) + x2 , where g is a continuous increasing function. Suppose g is concave. We have to show that

the preferences ↭ represented by u are convex.

Informally, the reasoning is this: u is the sum of a concave and linear function, so u is concave. Recall from Problem Set

1 that u concave =→ u quasi-concave ↑→ ↭ convex.

But here is a formal proof:

Proof. Let x = (x1 , x2 ) and y = (y1 , y2 ) be arbitrary consumption bundles, and suppose, WLOG1 , that x ↭ y, so

u (x) ↓ u (y). Let ω ↔ [0, 1] be arbitrary. Notice that:

u (ωx + (1 ↗ ω) y) = g (ωx1 + (1 ↗ ω) y1 ) + [ωx2 + (1 ↗ ω) y2 ]

↓ ωg (x1 ) + (1 ↗ ω) g (y1 ) + [ωx2 + (1 ↗ ω) y2 ]

= ω [g (x1 ) + x2 ] + (1 ↗ ω) [g (y1 ) + y2 ]

= ωu (x1 , x2 ) + (1 ↗ ω) u (y1 , y2 )

= ωu (x) + (1 ↗ ω) u (y)

↓ ωu (y) + (1 ↗ ω) u (y)

= u (y)

↫ u (ωx + (1 ↗ ω) y) ↓ u (y)

↫ ωx + (1 ↗ ω) y ↭ y

↘
2. Here, U (x, y) = 2 x + y

ωU (x,y) 1 ωU (x,y)
(a) M Ux = = 2 2→ = →1 ; M Uy = = 1.
ωx x x ωy

(b) Yes, monotonicity is satisfied.

(c) You can easily see that →1 decreases as x increases. If you can’t, you can take the derivative of →1 and see that it is
x x

negative for any x > 0, so M Ux is diminishing.

1 [Link]

1
 →1
(d) Here, M RSx,y = M Ux
= x
= →1 . This is decreasing in x, however, does not change as y changes. This is a property
M Uy 1 x

of quasilinear preferences, where M U and M RS is constant in quantity of the numeraire good (the good that enters

the utility function linearly).

↘ ! "2
(e) For an interior solution, M RSx,y = →1 = py , so x= and optimal x↑ = . We can see that her demand
px py py
x px px

does not depend on income.

(f) Now, as monotonicity is satisfied, the budget constraint must hold with equality, so

px x↑ + py y ↑ = I
# $2
py
px + py y ↑ = I
px
2
(py )
+ py y ↑ = I
px
2
(py )
py y ↑ = I↗
px
I py
↫ y↑ = ↗
py px

(py )2
Her demand for y changes linearly as I changes. Basically, for a fixed price ratio, she spends a constant amount px

on x, and spends the rest of her income on y. This is true for any interior solution with quasilinear preferences.

3. We have to show that the preferences represented by the utility function u (x1 , x2 ) = min {x1 , x2 } are convex.

Let x = (x1 , x2 ) and y = (y1 , y2 ) be arbitrary consumption bundles, and suppose, WLOG, that x ↭ y, so u (x) ↓ u (y).

Let ω ↔ [0, 1] be arbitrary. We have to show that ωx + (1 ↗ ω) y ↭ y.

Notice that u (x) ↓ u (y) means:

min {x1 , x2 } ↓ min {y1 , y2 }

ω min {x1 , x2 } ↓ ω min {y1 , y2 }

ω min {x1 , x2 } + (1 ↗ ω) min {y1 , y2 } ↓ ω min {y1 , y2 } + (1 ↗ ω) min {y1 , y2 }

ω min {x1 , x2 } + (1 ↗ ω) min {y1 , y2 } ↓ min {y1 , y2 }

Now, u (ωx + (1 ↗ ω) y) = min {ωx1 + (1 ↗ ω) y1 , ωx2 + (1 ↗ ω) y2 }, we want to compare this with ω min {x1 , x2 }+(1 ↗ ω) min {y1 ,

If you are familiar with concave/convex functions, you will notice that this comparison is asking whether the min function

is concave or convex. This problem leads to you learning a very useful result: the min {x1 , x2 } function is concave.

Lemma. The function f (x1 , x2 ) = min {x1 , x2 } is concave.

Proof. Let x1 , x2 , y1 , y2 ↔ R be arbitrary, and let ω ↔ [0, 1] be arbitrary. Notice that x1 ↓ min {x1 , x2 } and x2 ↓

min {x1 , x2 }. Similarly, y1 ↓ min {y1 , y2 } and y2 ↓ min {y1 , y2 }. Now, either ωx1 + (1 ↗ ω) y1 ↓ ωx2 + (1 ↗ ω) y2 or

2
 ωx1 + (1 ↗ ω) y1 < ωx2 + (1 ↗ ω) y2 .

Case 1: ωx1 +(1 ↗ ω) y1 ↓ ωx2 +(1 ↗ ω) y2 . So, min {ωx1 + (1 ↗ ω) y1 , ωx2 + (1 ↗ ω) y2 } = ωx2 +(1 ↗ ω) y2 ↓ ω min {x1 , x2 }+

(1 ↗ ω) min {y1 , y2 }.

Case 2: ωx1 +(1 ↗ ω) y1 < ωx2 +(1 ↗ ω) y2 . So, min {ωx1 + (1 ↗ ω) y1 , ωx2 + (1 ↗ ω) y2 } = ωx1 +(1 ↗ ω) y1 ↓ ω min {x1 , x2 }+

(1 ↗ ω) min {y1 , y2 }.

So in either case, we have min {ωx1 + (1 ↗ ω) y1 , ωx2 + (1 ↗ ω) y2 } ↓ ω min {x1 , x2 } + (1 ↗ ω) min {y1 , y2 }.

So now we have u (ωx + (1 ↗ ω) y) ↓ u (y); hence, ωx + (1 ↗ ω) y ↭ y, and so the preferences that u represents are convex.

4. Here, U (S, B) = SB + 10(S + B). Notice that here marginal utilities are linear (so not diminishing marginal utilities), and

here we may not have an interior solution, so the M RS = p1

p2 method may not work. But let’s try it anyway to see what

goes wrong.
3
Here, M US = B + 10, M UB = S + 10. So M RSS,B = M US
M UB = S+10 .
B+10
Setting this equal to pS
pB = 9 = 13 , we get the equation

B + 10 1
=
S + 10 3
3B + 30 = S + 10

S = 3B + 20

Plugging this condition into the budget constraint, which will hold with equality because of monotonicity of preferences,

we get

pS S + pB B = 30

3S + 9B = 30

3 (3B + 20) + 9B = 30

9B + 60 + 9B = 30

18B = ↗30
30
↫B = ↗
18

Now, eating ↗ 30
18 burgers is not physically possible (except in really gross fast food joints), so we must have a corner solution

with B = 0, and spend all the money on salad, so 3S = 30 and S = 10. What is the reason for this corner solution?

Notice that the marginal utility of spending one more dollar on salad is M US
pS = B+10
3 and for a burger it is M UB
pB = 9 .
S+10

10 20
Even when B = 0 and S = 10, the marginal utility of a dollar spent on salad is 3 ≃ 3.33, which is greater than 9 ≃ 2.22,

the marginal utility of a dollar spent on a burger, and if you increase B from 0 (and decrease S from 10), M US
pS gets even

bigger and M UB
pB gets even smaller, so at the current income level, spending one more dollar on Salad always adds more

3
 utility than spending one more dollar on Burger, hence Roy should spend all his money on Salad.

5. Here, U (x, y) = x2 y. Given the utility function, it’s easy to see we won’t have a corner solution, because choosing either

x = 0 or y = 0 leads to a utility level of 0, whereas positive amounts of both goods gives a higher utility. So we can use

the M RS = px
py trick. So,

M Ux px
=
M Uy py
2xy 8
=
x2 2
4y = 8x

↫ y↑ = 2x↑

Plugging this in his budget constraint, we get

px x↑ + py y ↑ = I

8x↑ + 2 (2x↑ ) = 240

12x↑ = 240

↫ x↑ = 20

↫ y↑ = 40

6. Here, U (x, y) = x + y, and I = 4.

1
(a) Here, px = 1 and py = 1, so px
py = 1. Also, M RSx,y = 1 = 1, so M RSx,y = px
py at all levels of x, y, even at the corners.

So any combination of (x, y) in which she exhausts her budget is optimal, so her Marshallian demands are the set

{(x↑ , y ↑ ) : 0 ⇐ x↑ ⇐ 4, y ↑ = 4 ↗ x↑ }.

1
(b) When px = 1 and py = 2, M RSx,y = 1 just as before but px
py = 2 < 1, so we cannot have an interior solution. This
1
is because = 1 and = so spending an extra dollar on good X is always better than spending an extra
M Ux M Uy
px py 2

dollar in Y , so the only solution is (x↑ , y ↑ ) = (4, 0).

(c) When px = py = 1, her demand for Y is any y ↑ ↔ [0, 4]. When px = 1 and py > 1, she chooses only good X so her
4
demand for Y is y ↑ = 0. When px = 1 and py < 1, she chooses only good Y so her demand for Y is y ↑ = I
py = py .