Problem Set 15 (a) Show that there exists some x∗ ∈ RK + such that

Mathematical Methods for Economists1 y ∗ (w, B) = f (x∗ ) and w · x∗ = B, i.e., the budget
constraint is binding.
1. For each constrained optimization problem, (1) define the
Lagrangian function, (2) derive necessary conditions (in- (b) Show that y ∗ (w, B) is quasi-convex in (w, B) (note
cluding FOCs) for the optimal solution, and (3) find the that we have not assumed that f is concave).
optimal solution
5. Consider the production function f (x1 , x2 ) = δ ln(x1 +
(a) max 3x + 2y s.t. x, y ≥ 0, x + y = 1 1) + (1 − δ) ln(x2 + 1), if (p, w1 , w2 )  0 denotes the vec-
x,y
tor of commodity prices, we can define the profit func-
(b) max −(x − y)2 s.t. −1 ≤ x ≤ 2 tion π(p, w1 , w2 ) = max{py − w1 x1 − w2 x2 |y ≤ f (x1 , x2 )}
x
(c) max x4 s.t. −1 ≤ x ≤ 2 which represents the maximum profits that can be earned
x from any feasible production activity. For what values of

in
2. For each (unconstrained) optimization problem, derive (p, w1 , w2 ) is π(p, w1 , w2 ) > 0? When does π(p, w1 , w2 ) =
first order conditions, find critical points (points to solve 0? [Hint: You may use Kuhn-Tucker]
FOC), and discuss whether the point you find is a (local)

e.
6. The cost of producing qj units at plant j is Cj (qj ) = αj qj +
maximizer or (local) minimizer or neither of them
qj2
, j = 1, 2, . . . , n. The product price is p.

nc
(a) max −(x − 7)2 2βj
x∈R
(b) max x3 − x (a) Initially assume that α = (α1 , . . . , αn ) is zero and
x∈R

ra
solve for aggregate supply as a function of the output
(c) max −ex − e−x
x∈R price.
3. A is selfish with utility function (b) Suppose instead that α1 < α2 < · · · < αn . For any

nt
price p, which plants will the firm operate? Again
ua (xa , ya ) = u(xa , ya ) solve for aggregate supply as a function of the output
se
where x and y are two commodities. B cares about her price.
own consumption and A’s happiness. Her utility function
7. A monopolist sells his output in two markets (with no
ic
is
resale.) The demand price for product 1 is p1 = 10 − q1
ub (xa , ya , xb , yb ) = u(xb , yb ) + αu(xa , ya )
and for product 2 is p2 = b−q2 . The cost of producing the
om

A has an income of Ia and B has an income of Ib . total output q1 + q2 is C(q) = (q1 + q2 )2 . For what values
of b will the profit maximizing firm only sell in market 2.
(a) If u(x, y) = ln x + ln y, assuming given market prices
of px , py , under what conditions will B give part of 8. Consider the following problem: max(x1 ,x2 ) 3x1 + x2 sub-
her income to A?
on

ject to (x1 + 1)2 + (x2 + 1)2 ≤ 5, xi ≥ 0 for i ∈ {1, 2}

(b) Assuming the condition you derived in (a) holds,
what gift will she make? (a) Suppose that (x∗1 , x∗2 ) solves this problem. Is there
necessarily a value of λ∗ such that (x∗1 , x∗2 , λ∗ ) satis-
ec

(c) Provide a necessary and sufficient condition for gift

fies the Kuhn-Tucker conditions?
giving for a general concave utility function u(x, y).
(d) Consider an economy with equal numbers of 2 types (b) Now suppose that (x∗1 , x∗2 , λ∗ ) satisfies the Kuhn-
of consumers (so now prices are endogenous). Type Tucker conditions. Does (x∗1 , x∗2 ) necessarily solve the
A are selfish but each type A has one type B who problem?
cares. Assuming price taking behaviour, and ignoring (c) Given the information in your answers to (a) and (b),
bargaining, define an equilibrium for this model. Is use the Kuhn-Tucker method to solve the problem.
it Pareto efficient?
9. A firm produces in each of n periods. In any period i
4. Consider a competitive firm with a single output produc- its output is xi and the price of output is pi (which is
tion function f (x), where x ∈ RK + is the amount of K > 1 fixed, but varies between periods). If it chooses the ca-
goods used as inputs. The price of output is one, and the pacity k, then its output in each period is at most k. The
price vector for the inputs is w ∈ RK ++ , so the total cost cost of maintaining the capacity k is D(k), where D is
of input is w · x. The firm is under a budget constraint so convex, and the cost of producing the vector of outputs
that there is a maximum amount B on the total cost of x = (x1 , . . . , xn ) is C(x),
Pnwhere C is convex. Thus the
inputs. Suppose f is continuous and satisfies the following firm’s problem is maxx,k i=1 pi xi − C(x1 , . . . , xn ) − D(k)
“local productiveness condition”: For all x = (x1 , . . . , xK ) subject to k ≥ 0, 0 ≤ xi ≤ k for i = 1, . . . , n.
and  > 0, there is some x0 = (x01 , . . . , x0K ), such that
maxk=1,...,K |x0k − xk | <  and f (x0 ) > f (x). Assume the (a) Write down the Kuhn-Tucker conditions for this
firm maximizes output, let y ∗ (w, B) denote this output, problem.
i.e., (b) What is the relation between the solutions of the
y ∗ (w, B) = sup {f (x) : w · x ≤ B} Kuhn-Tucker conditions and the solutions of the
1 Topics: Linear Algebra Contact: [email protected] problem?

1
 (c) Consider the case in which C(x1 , x2 ) = x21 + 2. x∗2 solves maxx2 ∈B(p∗ ,p∗ ·e2 ) u2 (x2 , x∗11 ) and
x22 , D(k) = k 2 , p1 = 1, and p2 = 3. Is there a so- 3. x∗1 + x∗2 ≤ r
lution in which x1 = x2 = k > 0? Is there a solution
Use the characterization in (b) to show that no
in which 0 < x1 < x2 = k?
strictly positive competitive equilibrium allocation
10. Unbinding Constraints in Programming Problem. Con- x∗  0 in this economy is Pareto efficient (in partic-
sider the programming problem maxx∈RL f (x) s.t. x ∈ X, ular, the first welfare theorem fails in this economy).
where X is a closed and convex set in RL . Suppose that 2 2
you added another constraint g(x) ≥ 0 to the problem, 12. Consider the following problem: max(x1 ,x2 ) −x1 −x1 x2 −x2
derived an optimal solution x∗ , and found that this addi- subject to x1 −2x2 ≤ −1 and 2x1 +x2 ≤ 2. Use the Kuhn-
∗
tional constraint is not really binding at x . Tucker conditions to find the solution(s) of the problem.

(a) Can you conclude that x∗ is also a solution for 13. Consider  the following
 problem:
2 2
maxx∈RL f (x) s.t. x ∈ X? If not, find a counterex- max (x1 ,x2 ) −(x 1 − 4) − (x 2 − 4) subject to
ample (a graphical example would suffice). x 1 + 2x 2 ≤ 4 and x 1 + 3x 2 ≤ 9. Use the Kuhn-Tucker
conditions to find the solution(s) of the problem.
(b) Suppose that f is strictly quasi-concave and g is con-
tinuous. Is x∗ a solution for maxx∈RL f (x) s.t. x ∈ X 14. A consumer has income I > 0, and faces a price vector
in this case? p ∈ R3++ for the three commodities she consumes. All
(c) Consider the consumer utility maximization problem commodities mush be consumed in nonnegative amounts.
with truncated consumption set (maxxi ∈Xi ui (xi ) s.t. Moreover, she must consume at least two units of com-
p · xi ≤ p · ei , xi ≤ r̄). Let x̄i (p) be a solution for this modity 2, and cannot consume more than one unit of com-
problem. Apply the above result to show that x̄i (p) is modity 1. Assuming I = 4 and p = (1, 1, 1), calculate the
also a solution for the standard utility maximization optimal consumption bundle if the utility function is given
problem if x̄i (p)  r̄. by u(x 1 x2 , x3 ) = x1 x2 x3 . What if I = 6 and p = (1, 2, 3)?
,

11. Consider a pure exchange economy with two consumers, 15. A firm manufactures two outputs y1 and y2 using two in-
who are endowed with strictly positive initial endow- puts x1 and x2 . The production function f : R2+ → R2+
ments ei ∈ RL ++ , i = 1, 2. Consumer 1’s utility func- 1/2 1/2 1/3
is given by (y1 , y2 ) = f (x1 , x2 ) = (x1 , x1 x2 ). Let pi
tion u1 : RL + → R is differentiable, concave and satisfies denote the unit price of yi , and wj that of xj . Describe the
Du1 (x1 )  0 for any x1 ∈ RL + . Assume that consumer 2 firm’s optimization problem and derive the equations that
cares about consumer 1’s consumption of good 1 as well define the critical points of the Lagrangian L. Calculate
as his or her own consumption. More specifically, con- the solution when p1 = p2 = 1 and w1 = w2 = 2.
sumer 2’s utility function u2 : RL+1 + → R is given by
u2 (x2 , x11 ) = v2 (x2 ) − x11 , where v2 is differentiable, con- 16. Construct a subset of the xy-plane R2 that is
cave and satisfies Dv2 (x2 )  0 for any x2 ∈ RL +.
Pareto efficiency is defined in the standard way. A fea- (a) closed under vector addition and subtraction, but not
sible allocation (x1 , x2 ) ∈ R2L + (x 1 + x 2 ≤ r = e 1 + e 2 ) scalar multiplication.
is Pareto efficient if and only if there is no other fea- (b) closed under scalar multiplication but not under vec-
sible allocation (x01 , x02 ) such that u1 (x01 ) ≥ u1 (x1 ) and tor addition.
u2 (x02 , x011 ) ≥ u2 (x2 , x11 ) with at least one inequality be-
ing strict. Answer the following questions. 17. Addition (⊕) and scalar multiplication (·) are required to
satisfy these eight rules:
(a) Consider the following programming problem:
1. x ⊕ y = y ⊕ x
(P) max u1 (x1 )
(x1 ,x2 )∈R2L
+ 2. x ⊕ (y ⊕ z) = (x ⊕ y) ⊕ z
s.t. v2 (x2 ) − x11 ≥ ū2 3. There is a unique “zero vector” such that x ⊕ 0 = x
and x1 + x2 ≤ r for all x.
4. For each x there is a unique vector −x such that
Show that a feasible allocation (x∗1 , x∗2 ) ∈ R2L
+ is x ⊕ (−x) = 0.
Pareto efficient if and only it is a solution of prob-
lem (P) with ū2 = v2 (x∗2 ) − x∗11 . 5. 1 · x = x
(b) Write down the Kuhn-Tucker conditions to charac- 6. (c1 c2 ) · x = c1 · (c2 · x)
terize the set of all solutions in R2L
+ of (P). Discuss 7. c · (x ⊕ y) = c · x ⊕ c · y
why they characterize the entire set of (interior) so-
lutions. 8. (c1 + c2 ) · x = c1 · x ⊕ c2 · x
(c) Define competitive equilibrium in this pure exchange (a) Suppose addition in R2 adds an extra 1 to each com-
economy as follows: (x∗ , p∗ ) ∈ R2L L
+ × R+ is a com- ponent, so that (3, 1) ⊕ (5, 0) equals (9, 2) instead of
petitive equilibrium if (8, 1). With scalar multiplication unchanged, which
1. x∗1 solves maxx1 ∈B(p∗ ,p∗ ·e1 ) u1 (x1 ), rules are broken?

2
 (b) Show that the set of all positive real numbers, with 26. Find a basis for the subspace of R3 spanned by the three
x ⊕ y and c · x redefined to equal the usual xy and vectors      
xc , is a vector space. What is the “zero vector”? 1 −1 0
−1 ,  0  ,  1 
(c) Suppose (x1 , x2 ) ⊕ (y1 , y2 ) is defined to be (x1 +
0 1 −1
y2 , x2 + y1 ). With the usual c · x = (cx1 , cx2 ), which
of the eight conditions are not satisfied? 27. Find a basis for the subspace of R4 spanned by the four
(d) The functions f (x) = x2 and g(x) = 5x are “vectors” vectors        
in the vector space F of all real functions i.e. set of 1 2 1 2
all f : R → R. Addition is defined in the usual way 1 −1  1  1
 , , , 
((f ⊕ g)(x) = f (x) + g(x), ∀x ∈ R). Which rule 2 −5 −4 1
is broken if multiplying f (x) by c gives the function 4 2 0 6
f (cx).
28. For which values of k does
(e) If the sum of the “vectors” f (x) and g(x) in F is
defined to be (f ⊕ g)(x) = f (g(x)) then what is the kx + y = 1
“zero vector”? Keep the usual scalar multiplication
x + ky = 1
c · f (x) = cf (x), and find two rules that are broken.
have no solution, one solution, or infinitely many solu-
18. Which of the following subsets of R3 are actually sub-
tions?
spaces?
29. Starting with a first plane u+2v−w = 6, find the equation
(a) The plane of vectors b = (b1 , b2 , b3 ) with first compo-
for
nent b1 = 0.
(b) The vectors b = (b1 , b2 , b3 ) with b2 b3 = 0. (a) the parallel plane through the origin.
(c) All linear combinations of two given vectors (1, 1, 0) (b) a second plane that also contains the points (6, 0, 0)
and (2, 0, 1). and (2, 2, 0).
(d) The plane of vectors b = (b1 , b2 , b3 ) that satisfy b3 − (c) a third plane that meets the first and second in the
b2 + 3b1 = 0. point (4, 1, 0).

19. Show that the set of nonsingular 2 by 2 matrices is not

a vector space. Show also that the set of singular 2 by 2
matrices is not a vector space.
 
2 −2
20. The matrix A = is a “vector” in the space M
2 −2
of all 2 by 2 matrices. What matrices are in the smallest
subspace containing A?
21. Find the largest possible number of independent vectors
among
           
1 1 1 0 0 0
−1  0   0   1   1   0 
 , , , , , 
 0  −1  0  −1  0   1 
0 0 −1 0 −1 −1

22. If w1 , w2 , w3 are independent vectors, show that the dif-

ferences v1 = w2 − w3 , v2 = w1 − w3 , and v3 = w1 − w2
are dependent. Find a combination of the v’s that gives
zero.
23. If w1 , w2 , w3 are independent vectors, show that the sums
v1 = w2 + w3 , v2 = w1 + w3 , and v3 = w1 + w2 are
independent.
24. Find two independent vectors on the plane x+2y−3z−t =
0 in R4 . Then find three independent vectors. Why not
four?
25. v + w and v − w are combinations of v and w. Write v and
w as combinations of v + w and v − w. Do they span the
same space? When are they a basis for the same space?