1

Preface
This is a supplement to Mathematical Methods for Economics, 2nd edition, by Michael W. Klein. This manual provides answers to all the end-of-section exercises in the text. Solving the end-of-section exercises is an integral part of the learning process for a course using this textbook. The exercises in Mathematical Methods for Economics are designed to develop students' economic and mathematical intuition as well as to hone their skills at practical problem solving. The end-of-section exercises also oer students some additional economic applications. Deirdre M. Savarese developed most of the end-ofsection exercises in Mathematical Methods for Economics, and her assistance is gratefully acknowledged. This manual also oers concordances to aid instructors who would like to supplement their microeconomic or macroeconomic courses with Mathematical Methods for Economics. The concordances list, for seven leading microeconomic and macroeconomic textbooks, the mathematical tools required in specic chapters of those books and the place where those tools are presented in Mathematical Methods for Economics.1 The concordances also list applications in Mathematical Methods for Economics that instructors can use when teaching material from specic chapters in these leading macroeconomic and microeconomic textbooks. A conversion chart is also presented to aid instructors who want to switch from another mathematics for economics textbook to Mathematical Methods for Economics. Deirdre M. Savarese put together the concordances and the conversion chart for the 1rst edition of Mathematical Methods for Economics and Kieran Brenner updated this material for the 2nd edition. This manual also oers reproductions of gures from Mathematical Methods for Economics and these reproductions can serve as transparency masters. Graphs are important pedagogic tools in a course like this one. Instructors may nd transparencies more accurate, easier to use, and less timeconsuming to present than gures drawn on a chalkboard in the midst of class.
There are concordances for microeconomic texts by Perlo, Pindyck and Rubinfeld, and Varian, and for macroeconomic texts by Abel and Bernanke, Blanchard, Gordon, and Mankiw.
1

Chapter 2
Section 2.1 1. The intervals are (a) (5; 0) (c) (1; 100) (e) (0; 1)

(b) [5; 0)

(d) (1; 100] (f) (1; 1)

2. Do the following represent a function? (a) Yes (b) No (c) Yes (d) Yes (e) No (f) The function y is not dened at x = 3: (g) No 3. Can the function be dened according to the mapping? (a) Yes (b) Yes (c) No (d) No 4. It would not be a function since there would be a mapping between one argument and more than one value.

3 5. With this cost function (a) When Q = 10; T C = 125: When Q = 25; T C = 200: When there is no production (Q = 0) ; T C = 75: (b) Linear graph. (c) Domain [0:50] Range [75; 325] 6. Four representative ordered pairs are (a) (2; 140) ; (1; 120) ; (0; 100) ; (1; 80) (c) (100; 101) ; (0; 1) ; (1; 2) ; (100; 101)

(b) (1; 2) ; (0; 0) ; (1; 2) ; (2; 10) 7. The limits are (a) lim = 2 (b) lim =1 + =7 (c) lim + (d) lim = 0
x!1 x!7 x!7 x!1

8. Are the functions continuous? (a) No (b) Yes (c) Yes (d) Yes 9. The function presented in 8(c) y = 3 + 1 x+7

is not continuous over the domain (1; 0] since the function is undened at the point x = 7:

4 Section 2.2 1. The respective functions are (a) strictly monotonic (b) nonmonotonic (c) strictly monotonic (d) monotonic 2. Are these one-to-one? (a) no. An individual may be a citizen of more than one country (b) no. While a street address has one zip code, a given zip code applies to more than one street address. (c) yes (d) no. While one identication number is linked with one course grade, a given course grade may correspond to numerous students' identication numbers. 3. The inverse functions, if they exist, are (a) x = f 1 (y ) =
y14 7

(b) This function does not have an inverse unless we restrict the dop 1 main to y 6 so that the inverse is x = f (y ) = y 6 (c) This function does not have an inverse unless we restrict the domain to y 0 so that the inverse is x = f 1 (y ) = y 2
1

(d) x = f 1 (y ) = y 3 4. The inverse is x =

y 10

+1 . To check this, note that 2 y 1 10x 5 1 + = x and 10 + 5 = y: 10 2 10 2

5. For continuous functions with extreme points, the answers are (a) No

5 (b) Yes (c) Either two minima and one maximum or two maxima and one minimum 6. This is a linear function so the global minimum and maximums are the respective endpoints. The global minimum is (0; 50) and the global maximum is (100; 100) 7. The average rates of return are (a) Average rate of return = 12 (b) Average rate of return= 16 (d) Average rate of return = 0 (c) Average rate of return = 12

8. For a strictly increasing function, f (xB ) > f (xA ) and xB > xA , so the average rate of change will always be positive since both the numerator and the denominator yield a positive answer. For a strictly decreasing function, the numerator will be negative and the denominator will be positive. 9. The answers are (a) y 0 = 5
B )f (xA ) (b) The slope of the secant line f (xx = 4: The value of the B xA slope of the secant line represents the average rate of change of a function over the interval dened by the two endpoints of the secant line.

(c) The function is strictly convex over the given interval since the secant line lies wholly above the function. 10. The answers are (a) x0 = 2:8; f (x0 ) = 28:16 (b) y 0 = 26 (c) The function is strictly concave since f (x0 ) > y 0 ; 28:16 > 26

6 11. The answers are (a) A ( B (c) A ) B

(b) A ( B (d) A , B This set satises the necessary and sucient condition

Section 2.3 1. These expressions can be written as (a) x1 (b) (xy )4 (c) x20 (d) x3 y 2 (no further simplication possible) 6 1 (e) xy (a) x(a+b+cd) (b) x 3 (c) x 12 (d) x2 y 2 3. Simplifying the expressions, we get (a) 8 (b) 212 = 4096 (c) (d)
1 2 x+3 x+2
31 5

2. These expressions can be condensed as

(e) (x + 1)25

7 4. The quadrants through in which the graphs of the functions appear are (a) Quadrants I and II (b) Quadrants I and III (c) Quadrants I and II (d) Quadrants I and III 5. The roots are (a) x =
6 5

(b) x = 1; 6

(d) f (x) = (x + 1) (x2 3x + 2) = (x + 1)(x 1)(x 2) = 0 so x = (1; 1; 2) p q q2 pr 6. The root is x = p 7. Some points of these functions are x 2x2 2x 0 0 1 1 2 1.19 1.41 1.68 2
1 4 1 8 1 2 1 2 3 4 9 8

(c) x = 3 (two equal roots)

The two functions do not share a y-intercept but do have a common value when x = 1: 8. The average rate of change for the function y = 2x2 is 10. The average : Both functions are strictly rate of change for the function y = 2x is 14 3 0 0 convex such that f (x ) < y . 9. Some points of these functions are x 1 x 2 2x 1 4 2 0
1 2 1 2

1 2 1 2 2 8

3 4 4 8 32 128

The two curves share a y-intercept since in both cases when x = 0, y = 1 . When A = 1 , the curves no longer share a common y value 2 4 since the graph of the rst function shifts down.

8 10. When the domain of x is restricted, some points of these functions are x 1 x 2 2x 1 4 2 -2
1 8 1 32

-1 0
1 4 1 8 1 2 1 2

The functions are graphed in Quadrant II. The y-intercept is still at 1 2 but the curve slopes down and to the left, asymptotically approaching the x-axis. 11. Matching the functions to the economic relations, we have (a) a = iii (b) b = v (c) c = i (d) d = ii (e) e = iv 12. a. There are two tax rates consistent with raising $60 billion, t = 0:3, t = 0:4. These are found by using the quadratic formula. q 350 (350)2 4 (60) (500) t1 ; t2 = 2 (500) b. There is one tax rate consistent with raising $61:25 billion, t = 0:35, since, using quadratic formula. q 350 (350)2 4 (61:25) (500) 350 = = 0:35 t= 2 (500) 2 (500) This is a case of a single root to a quadratic equation.

Chapter 3
Section 3.1 1. The values are (a) Xt = 100 (b) Xt+5 = 115:93 (c) Xt+1 = 200 (d) Xt+50 = 710:67 2. The values are (a) Xt4 = 36:75 (b) Xt4 = 39:6 (c) Xt1 = 47:17 (d) Xt10 = 41:01 3. The values are (a) Xt = 22:22 (b) Xt = 20:4 (c) Xt = 17:82 (d) Xt = 25 (e) Xt = 27:6 4. The value of Xt+n approaches zero as n approaches innity but never equals zero. Rather the function is a horizontal asymptote, with n plotted on the horizontal axis and Xt+n on the vertical. 5. Your salary will be (a) $52; 500 (b) $56; 784

10 (c) $80; 549 _. 6. Let Pt stand for prots in year t growth rate of 6%, we have With P2000 = $50 million, and a

P2001 = $50 million (1:06) = $53 million P2003 = $50 million (1:06)3 = $59:55 million P1998 = $50 million (1:06)2 = $44:50 million 7. The levels of national income in 2010, Y2010 , under dierent assumptions about the rate of growth beginning in 1994, and given Y1994 , the level of income in 1994, are Growth Rate of 2:5% : Growth Rate of 3:0% : Growth Rate of 3:5% : Y2010 = e160:025 Y1994 = 1:49Y1994 Y2010 = e160:030 Y1994 = 1:62Y1994 Y2010 = e160:035 Y1994 = 1:75Y1994

The ratios of national income in 2010 under dierent scenarios concerning the rate of growth are Growth Rate of 3:0% compared to 2:5% : Growth Rate of 3:5% compared to 2:5% : 1:62Y1994 = 1:087 1:49Y1994 1:75Y1994 = 1:174 1:49Y1994

8. For Indonesia, 207 (1:0175)11 = 250:5 and for China 1; 250 (1:0104)11 = 1; 400:7 9. 100(1:02)5 = 110:41; 100(1:03)5 = 86:26

11 Section 3.2 1. The values are (a) Xt+1 = 21:65 (b) Xt+1 = 20:1 (c) Xt+1 = 22:1 2. The value of Xt is impacted more by a change in r than by a change in k . Doubling the frequency of compounding, holding r constant, reduced the value of Xt by 0.03%. Doubling the rate of decay, r, increased the value of Xt by 5.2%. (a) Xt = 105:19 (b) Xt = 105:16 (c) Xt = 110:66 3. Choose Bank B which will oer a 0.25% better return ($3,916). 4. rE = erA 1 so rE = 8:3% 5. The values are (a) Xt+3 = 98:25 (b) Xt+0:5 = 75:94 (c) Xt2 = 60:19 (d) Xt+0:25 = 76:13 (e) Xt+0:75 = 73:33 6. If there is no capital appreciation, the return on the stock is simply the value of the dividend, or 3%. With capital appreciation and a dividend payment, the following returns apply: Q1 Q2 Q3 Stock 46.14 47.31 48.50 Dividend 00.35 00.35 00.36 Total Value 46.49 47.66 48.86 Q4 49.72 00.37 50.09

The total return on the original investment is thus $5.09, or 11.31%.

12 7. Using the data given, we nd, for Indonesia 207 e110:0175 = 250:9

and for China

The ratio of population in 2010 is 5:59, whereas the ratio in 1999 is 6:04. The percentage dierence between the continuous compounding calculation and the discrete compounding calculation is 0:16% for Indonesia and 0:06% for China. 8. At 7%, PV= 13; 986. At 5%, PV= 14; 268. At 9.5%, P V = 13; 641 9. P V = 5; 578 10. Yt = Y0 exp(:01 + (0:03 0:7) + (0:02 0:3)) = Y0 exp(0:037) where Y0 = A0 L0 K0 11. The answers are (a) L (0) = 1 (b) lim L (t) = 2
t!1

1; 250 e110:0104 = 1; 401:5

(c) L (2) L (1) = 1:181 1:095 = 0:086: L (11) L (10) = 1:667 1:632 = 0:035: This represents diminishing marginal advances in productivity over time. 12. The answers are (a) $1320.00 (b) $1322.50 (c) In each case, with continuously compounded interest, the bond pays $1349.86.

13 Section 3.3 1. The answers are (a) 100 (b) 0 (d) log2 (a + b) (e) a + bx + cz (f) 3 ln 4x = 3(ln 4 + ln x) (g) 5 + 2( ln x ln y ) 2. The answers are (a) log2 256 = 8 (b) log3 27 = 3 (c) log3 1 = 0 3. Assuming that x > 1 and since e > 2, ln 32 < log2 32: Note that ln 32 = 3:47 and log2 32 = 5. 4. In the stock market you would get $2,585.71. Keeping the collection you'll get $3,000 but you will have to pay storage fees equivalent to $329.75, which leaves you with $2,670.25 so you should still keep the collection. 5. U =
n P

(c) 5 log10 x

i=1

qi i

6. The various means are (a) Arithmetic mean =

420+1070+3470 3

= 1653:33
1

(b) Geometric mean= (420 1070 3470) 3 = 1159:6

(c) Log Transformation= 1 (ln 420 + ln 1; 070 + ln 3; 470) = 7:056 so 3 Xg = exp(ln Xg ) = exp(7:056) = 1; 159:6

14 7. The ratio of health care in 1999 (H1999 ) to GDP in 1999 (Y1999 ) is H1999 =Y1999 = 0:10. With Health Care growing at a continuously compounded annual rate of 4:8% and GDP growing at a continuously compounded rate of 2%, we would have, after 30 years, H2029 H1999 e300:048 = = 0:10e300:028 = 0:23 Y2029 Y1999 e300:02 and health care would command almost a quarter of GDP. It is likely, however, that economic forces rein in the rise in health care costs before they reach this proportion of national income. 8. Bank Betterrates eective rate of 8.25% is equivalent to an annual interest rate of 7.9% through the equation rE = erA 1 and taking the natural logarithm of both sides. Bank Highrates is oering its investors a better deal. 9. When n is bigger, the Rule of 70 appears to hold better. (a) n = 69:66 Rule of 70= 70 (b) n = 14:21 Rule of 70 = 14 (c) n = 7:27 Rule of 70 = 7 (d) n = 3:10 Rule of 70 = 2.8 (e) n = 1 Rule of 70 = 0.7 ln L + 1 ln K . When L = 10 and K = 5, then ln Q = 10. ln Q = ln 15 + 4 5 5 4:872 and Q = exp(ln Q) = 130:6 11. n =
ln 1:5 0:05

= 8:1 years

12. The average ination rate is calculated by solving for in 144:8 e9 = 192:9 Rearranging this, we have = ln (192:9=144:8) 100% = 3:2%: 9

15

Chapter 4
Section 4.1 1. The equilibrium values are (a) (w; x; y ) = 1 ; 4; 4 4

(b) (w; x; y ) = (19; 232; 116)

2. For this system of equations

(c) (w; x; y ) = (3; 6; 14) b2h ; (8 5 g b 4 h ) ; (8 5 g b 4 h ) (d) (w; x; y ) = 64g 5 (a) (x; y; z ) = [(8a 9h 2) ; (8a 9h 2) ; (2a 2h 2)] (c) (x; y; z ) = (16; 16; 4)

(b) (x; y; z ) = (8a; 8a; 2a) I

3. With h = 3, the answers are (a) (x; y; z ) = [(8a 9h 2) ; (8a 9h 2) ; (2a 2h 2)] (c) (x; y; z ) = (27; 27; 6)

(b) (x; y; z ) = (9h; 9h; 2h)

+c bc ; Q = ad The parameters b and d are slope parameters. 4. P = a b+d b+d The linear demand curve must have a negative slope reecting the inverse relationship between the quantity demanded and price. The linear supply curve has a positive slope since producers are willing to supply more at a higher price. The parameters a and c are y-intercept parameters, and are positive to reect the notion that prices are always positive.

5. In this Keynesian model, the equilibrium values are (a) Y = 12; 000; C = 11; 000 (b) Y = 10; 500; C = 10; 000

16 6. Y = 3I 7. In this IS/LM model, the answers are (a) Y = 20; 500 R = 0:05 (b) For M = 400, Y = 5 400 = 2000 and R = 0:0005 400 = 0:2. 8. The equilibrium quantity in this model is Q= + N T G : +

Therefore the change in the equilibrium quantity with a given increase T in the tax is Q = T: + A given increase in the tax has a bigger eect on the equilibrium quantity with an increase in the sensitivity either demand or supply to price since the fraction increases with an increase in either or . + 9. In solving the model, we found the common wage to be W = and therefore B Y LB

B Y LB : Using the solution for LB , we obtain, after some manipulation, W = W = A + B Y: 2

10. Solving these equations, we nd LA LB

B A = Y A + B A B = Y A + B

Firm A experiences job creation and Firm B experiences job destruction in the face of Y > 0 if B > A , that is, if labor demand is less responsive to the increase in wages in Firm A than in Firm B .

17

Section 4.2 1. The answers are (a) (b) (c)

5 P n P

axi = ax1 + ax2 + ax3 + ax4 + ax5 = bi xi = b1 x1 + b2 x2 + :::bn xn xi+1

3 P

i=1

i=1

i=2 x4 y2

(d) (e)

i=1 4 P

3 P

3 P

i=2

yi1

= (x3 + x4 ) (y1 + y2 ) = x3 y1 + x3 y2 + x4 y1 +

ixi (xi + 2) = x1 (x1 + 2) + 2x2 (x2 + 2) + 3x3 (x3 + 2) xi =

1 x

i=1

1 x2

1 x3

1 x4

2. The dimensions are (a) 2 x 2 square matrix (b) 1 x 3 row vector (c) 3 x 1 column vector (d) 2 x 3 matrix 3. The values of the elements are (a) a22 = 100 (c) c21 = 1

(b) b13 = m (d) d23 =

4. For this matrix (a) Diagonal: z; b; 3; 1: O-diagonal: y; x; c; w; d; 4; a; 1; 4; 2; 3; 2

18 z 6 a (b) B 0 = 6 4 1 4 2 2 y b 2 3 x c 3 2 a b c d 3 w d 7 7 4 5 1

(d) b21 = y (in B 0 it becomes b12 ); b32 = c (in B 0 it becomes b23 ); b44 =(in B 0 it remains b44 ). The diagonal and o-diagonal elements in matrix B are the same diagonal and o-diagonal elements in matrix B 0 : 5. These matrices are (a) not conformable (b) conformable; solution dimension is 1 1 (c) not conformable (d) conformable; solution dimension is 1 l 6. For this system of equations (a) y1 = b11 cj + b12 cj + b13 cj ; y2 = b21 cj + b22 cj + b23 cj 2 3 c1 y1 b11 b12 b13 (b) y = B= c = 4 c2 5 y2 b21 b22 b23 c3 (c) The system is conformable
j =1 3 P

z 6 y (c) (B 0 )0 = 6 4 x w

1 2 3 4

3 4 3 7 7 2 5 1

(d) yi =

bij cj

7. The products of the matrices are 3 7a 7b 7c 7d (a) S = 4 9a 9b 9c 9d 5 4a 4b 4c 4d 2

19 (b) S = (c) S = 2 1 2 4 3 7 16

3 a g + 7h 4g + 2h (d) S = 4 x w + 7p 4w + 2p 5 b c + 7d 4c + 2d 8. For the matrices A and B (a) AB is dened. BA is not. (b) AB is 3 1 2 1 7 0 (c) A = 4 5 6 (d) Not conformable 9. The four equations which comprise the model are AD = C + I 3 C = 2000 + Y 4 I = 500 1000r AD = Y 10. For this model 3 32 3 2 1 0 20 X 1000 + 0:20YF (a) The matrix system is 4 0 1 10 54 M 5 = 4 450 + 0:15YD 5 0 1 1 0 E 2 (c) E =
0:2 YF 30 :15 and E = 030 YD

7 5 4 2

15YD +550 (b) The real exchange rate E = 0:2YF 0:30 . The equilibrium level of imports (and exports) is M = X = 633:3 + 0:67YF + 0:1YD

(d) Given the changes in YF and YD , E = 0:18

20 Section 4.3 1. The answers are 1 0 (a) 0 1 2a (b) 1b (c) 3 9 25 5

2. For the matrices A and B 2

3 5 8 14 0 5 (a) A + B = B + A = 4 1 2 3 7 8 0 2 3 25 8 10 2 5 (b) A B = 4 13 2 3 3 2 8 2 3 10 0 2 25 8 10 (c) B = 6 2 1 B + A = 4 13 2 2 5=AB 3 3 5 5 4 3 2 8 (a) M L K

3. For the matrices K , L, and M

(b) (M L) K = M (L K ) 4. For A as given (a) Identity matrix in 2 2 (b) AI = IA = 4 0 5 (c) See (b). 5. The transposes are AI is 4 4: Identity matrix in IA is 3 3: 3 4 5 7 0 1 6 5 8 9 2

21 3 3 7 6 5 4 1 2 7 7 (a) X 0 = Y 0=6 4 0 4 5 6 2 1 3 2 3 30 11 6 32 9 7 7 (b) (XY )0 = Y 0 X 0 = 6 4 24 8 5 22 6 2

Section 4.4 1. The respective matrices are (a) nonsingular (b) singular (c) nonsingular (d) nonsingular 2. The matrix is nonsingular if (a) 6= (b) 6=
cb ad ad cd bc . d

(c) The matrix is singular if a =

3. The only system which can be solved is 2(a). 4. The determinants are (a) jAj = 44 (c) jAj = 0

(b) jAj = 1

(d) jAj = 2 nonsingular as long as 6= 0 (e) jAj = 0

22 (f) jAj = a(d ) (d ) cb 5. The adjoints are 7 3 (a) adjA = 4 2 4 1 (b) adjA = 9 3 1 2 (c) adjA = 2 0 4 6 (d) adjA = 3 1 2 6. A A =
1

5 36 3 36

2 36 6 36

6 2 1 0 = =I 3 5 0 1

7. The inverses (if they exist) are 10 4 1 44 44 (a) A = 6 2 44 44 0 1 1 (b) A = 1 1 (d) A (c) Singular so no inverse # "
1

2 1 2

2 1 2

(e) Singular so no inverse " (f) A1 =

1

(g) A

= =

d a(d)(d)cb c a(d)(d)cb 7 2

(h) A

4 23

3 2 2 1

(i) A1 =

3 23 1 9 23 23 1 1 4 2 1 0 2

b a(d)(d)cb a a(d)(d)cb

23 (j) A
1

4 20 3 20

6 20 1 40

8. The system, with its solution, is 1 1 Y I +G (a) A = x= y= b 1 C a I +G+a 1 (b) x = A1 y = 1 b b(I + G)a 9. The answer is 1 1 A F 2 = (a) 1 1 B G 2 (b) A = B = 37:50 (c) A = 0; B = 112:50 10. The system is 12 4 w PS (a) = 2 6 r PW

(b) jAj = 64 3 1 P P w S W 32 16 (c) = 1 3 32 PS + 16 PW r (d) w = 9; r = 5

3 1 (e) r = 16 PW = 3 w = 16 PW = 1. The return to capital rises because wheat production uses capital more intensively.

24

Chapter 5
Section 5.1 1. The determinants are (a) jB j = 54 (c) jB j =

(b) jB j = 0

2. The minors and cofactors are (a) jM12 j = 6 jM22 j = 12 jM32 j = 6 (b) jM12 j = 7 jM22 j = 28 jM32 j = 28 (c) jM12 j = b2 ac jM22 j = ab c2 jM32 j = a2 bc jC12 j = b2 ac jC22 j = ab c2 jC32 j = a2 bc jC12 j = 7 jC22 j = 28 jC32 j = 28 jC12 j = 6 jC22 j = 12 jC32 j = 6

3. The determinants are as in question 1, namely (a) jB j = 54 (c) jB j =

(b) jB j = 0

25 4. jAj = 6 5. The determinants are (a) jAj = 10:50 (c) jAj = 4

(b) jAj = 441

6. jH j = kd j (1 b) 7. jT j = t11 t22 t33 t44 t55 8. jAj = 0. This matrix is redundant in that each element of the third row consists of k times each element in the second row plus the corresponding element in the rst row.

Section 5.2 1. The adjoints are 2 (a) (b)

(c)

(d)

25 35 4 adj (A) = 20 15 40 30 2 bc a2 adj (A) = 4 (c2 ab) ac b2 2 6 92 6 30 82 adj (A) = 6 4 78 64 192 298 2 18 4 6 18 0 adj (A) = 6 4 216 56 162 42

3 (bc ac) ab bc ac bc (a2 c2 ) 5 (a2 b2 ) ab bc 3 4 16 20 80 7 7 11 44 5 128 134 3 32 8 36 0 7 7 412 40 5 300 30

3 17 5 5 33

26 2. The inverses of the above matrices are 3 2 25 35 17 215 215 215 15 5 20 5 (a) A1 = 4 215 215 215 33 30 40 215 215 215 3 2 bc a2 (bc ac) ab bc 1 4 (c2 ab) ac bc (a2 c2 ) 5 (b) A1 = abca3 bc2 + ab2 +ac2 cb2 2 2 2 (a b ) ab bc ac b 2 3 6 92 4 16 6 30 82 20 80 7 1 6 7 (c) A1 = 422 4 78 64 11 44 5 192 298 128 134 2 3 18 4 32 8 6 0 36 0 7 1 6 18 7 (d) A1 = 236 4 216 56 412 40 5 162 42 300 30 3. For this macroeconomic model 2 3 2 3 2 3 1 1 1 Y G 5 0 54 C 5=4 200 (a) 4 0:8 1 0 0 1 I 1000 2000R 2 3 5 5 5 (b) A1 = 4 4 5 4 5 0 0 1

(c) Y = 250. The change in income due to a change in government spending is more in this scenario than in Chapter 4.1 since money demand is not included in this model and there is no crowding out.

4. For this export/import model 3 2 3 2 3 2 1 0 20 X 1000 + 0:2YF (a) 4 0 1 10 5 4 M 5 = 4 450 + 0:15YD 5 . The inverse of 1 1 0 E 2 10 2 3 2 the matrix of parameters, A1 = 4
3 1 3 1 30 1 30 3 2 3

5. Given the 1 3 1 30

27 solution x = A1 y , the following are the values of the endogenous :3YD :3YD variables: X = 650 + 0:2YF +0 ; M = 650 + 0:2YF +0 ; and 3 3 0:2YF 0:15YD 550 : E = 30 + 30 (b) The impact on exports of a $100 increase in foreign income is X = 6 2 : 3 5. The solutions are (a) (b) (c) (d) (x; y ) = (4; 0) 28 (x; y ) = 320 ; 17 17 44 152 (x; y; z ) = 53 ; ; 9 9 9 38 72 (x; y; z ) = 28 ; 96 ; 28 28

6. The Keynesian style model is set up as follows: 2 3 2 3 2 3 1 1 1 Y G (a) 4 0:8 1 0 5 4 C 5 = 4 200 1000R 5 0 0 1 I 1000 2000R (b) Y = 5G 5000R + 4000 7. If the matrix A is nonsingular, a homogeneous equation system yields the solution that x1 = x2 = = xn = 0: Using Cramer's rule, which, in this case, replaces the rst column of matrix A with a column of zeroes suggests that jAj j = 0. The value of jAj must therefore be non-zero in jA j 0 order for there to be a dened solution since xj = jAjj = jA = 0. If j matrix A were singular then its determinant would be zero making the solution of xj undened. 8. Using Cramer's rule, we have the change in the output of manufactured goods as 0 aKF L aLF aKF M = = L < 0 jAj aKM aLF aLM aKF and the change in the output of food as aKM 0 aLM L aKM L > 0: F = = jAj aKM aLF aLM aKF

28

Section 5.3 1. The characteristic equations and characteristic roots are as follows; (a) 2 6 + 5 = 0;
2 2

1 ; 2 = 1; 5;

(b) + 4 + 3 = 0;

(c) 3 4 = 0; 1 ; 2 = 1; 4; jAj = 10 (6) = 4 4 1 2. For the matrix with characteristic roots 1 = 5 and 2 = 1. 3 2 The characteristic vector associated with 1 solves 4 1 p1 5p1 = 3 2 p2 5p2 which gives us p1 = p2 and the characteristic vector associated with 2 solves 4 1 p1 p1 = 3 2 p2 p2 which gives us 3p1 = p2 . Thus we have 1 1 1 P = and P = 1 3
1 3 4 1 4

1 ; 2 = 1; 3; jAj = 3 0 = 3

jAj = 8 3 = 5

3 4
1 4

5 0 0 1

and P AP =

1 0 For the matrix with characteristic roots 1 = 1 and 2 3 2 = 3. The characteristic vector associated with 1 solves 1 0 p1 p1 = 2 3 p2 p2 which gives us p1 = p2 and the characteristic vector associated with 2 solves 3p1 1 0 p1 = p2 3p2 2 3

3 4 1 4

3 4
1 4

4 1 3 2

1 1 1 3

29 which gives us p1 = 0. Thus we have 1 0 1 0 1 P = and P = 1 1 1 1 and P

1

AP =

5 6 For the matrix with characteristic roots 1 = 1 and 1 2 2 = 4. The characteristic vector associated with 1 solves p1 5 6 p1 = p2 p2 1 2 which gives us p1 = p2 and the characteristic vector associated with 2 solves 5 6 p1 4p1 = 1 2 p2 4p2 which gives us p1 = 6p2 . Thus we have 1 6 1 6 1 5 5 P = and P = 1 1 1 1 5 5 and P
1

1 0 1 1

1 0 2 3

1 0 1 1

1 0 0 3

AP =

1 1 5 5

1 5

6 5

5 6 1 2

1 6 1 1

1 0 0 4

3. For this system (a) In matrix format, 2 6 1 3 x1 x2 = 20 4 1 ; 2 = 3; 4

(b) The characteristic equation is 2 + 12 = 0;

30 (c) Associated with the characteristic root 3, 2 6 p1 3p1 = 1 3 p2 3p2 which gives us p1 = 6p2 . Associated with the characteristic root 4, 2 6 p1 4p1 = 1 3 p2 4p2 which gives us p1 = p2 . So the matrix P is (up to a multiplicative constant) 6 1 P = 1 1 P
1

(d) We have =

and by matrix multiplication we nd P 1 AP = where is a diagonal matrix with the elements 3 and 4. 4. For this system, with P and P 1 as dened in exercise 3; (a) The vectors u = P 1 x and v = P 1 y are 1 x+ 1 y 7 7 u= v= 1 x+ 6 y 7 7 (b) We have AP u = 2x + 6y x 3y Pv = 20 4

1 7 1 7

6 7

1 7

24 7 4 7

which is just the original system. (c) We have P and therefore u =

1

AP = =

3 0 0 4 =v=

3 0 0 4
8 7

u1 u2

24 7 4 7

Solving this we nd u1 =

and u2 = 1 : 7

31 (d) Transforming back to the original system using P u = x we have x1 = 7 and x2 = 1: (e) These values solve the original system as well, as they must.

5. The characteristic equations and characteristic roots (found by using the quadratic formula) are as follows;

(a) 2 0:5 + 0:0225 = 0; (b) 2 0:3125 = 0;

1 ; 2 = 0:05; 0:45. This is stable.

1 ; 2 = 0:25; 1:25. This is not stable. 1 ; 2 = 0:25; 1:25. This is not stable.

(c) 2 1:5 + 0:3125 = 0;

The characteristic polynomial is found by evaluating the determinant a 0 0 b

and setting it equal to zero. This gives us 2 (a + b) +(ab) = (a ) (b ) = 0. The two roots of this equation are a and b. More generally, as discussed in Section 5.1, the determinant of a diagonal matrix D with diagonal elements d11 ; d22 ; :::dnn is (d11 ) (d22 ) ::: (dnn ) and, therefore, the n values of for which this determinant equals zero are d11 ; d22 ; :::dnn .

32

Chapter 6
Section 6.2 1. The dierence quotients are (a) (b) (c) (d)
y x y x y x y x

=5 = 15 = 2 + 12x0 + 6x = 2x0 x

2. When x = 2, the dierence quotients for questions a and b do not change. The dierence quotients for questions c and d are 38 and -6, respectively, where for each the variance is equal to cx: (a) The average rate of change = 5: (b) The average rate of change = 15. (c) The average rate of change = 50: (d) The average rate of change = 8: 3. The dierence quotients are (a) (b)
z x w x 2 = 3gx2 0 + 3gx0 x + g x 2 = b + 2cx0 + cx + 3gx2 0 + 3gx0 x + g x

4. The dierence quotients are (a) (b)

y x h x

= b + 2cx0 + cx

2 3 2 = b + 2cx0 + cx + 3gx2 0 + 3gx0 x + g x + 4hx0 + 6hx0 x + 4hx0 x2 + hx3

5. With the given functions y = 10x 4 y = 3x2 + 6x 5 y = x3 + 4x2 6x + 12 x0 = 1; x = 1 y = 10 x y = 15 x y = 13 x x0 = 2; x = 1 10 21 33 x0 = 2; x = 2 10 24 46

33 Changing the values of x0 and x has no impact on (a) since it is linear. Doubling the value of x0 has a greater impact on the quadratic and cubic functions than does doubling the value of x: 6. The change in wages are: w(20) = 49:75; w(30) = 44:75; w (40) = 39:75; w (50) = 34:75. As this employee ages, his wages increase at a decreasing rate as indicated by the negative coecient on the squared term. 7. The revenue function and its properties are as follows (a) T R = 10Q 0:5Q2
T R Q

(b)

= 10 Q 0:5Q. When Q = 1 and Q0 = 5;the impact on total revenue of a one unit change in Q is 4.5.

(c) The impact on total revenue is 6.5 when Q0 = 3 and Q = 1: (d) The impact on total revenue is 6 when Q0 = 3 and Q = 2: 8. When x0 = 3 and x = 3,
y x

= 34

y y = 28: When x = 0:5; = 24: (a) When x = 1:5; x x

(b) In the limit, as x approaches zero,

y x

= 22:

Section 6.3 1. The derivatives are (a) f 0 (x) = 30 (b) f 0 (x) = 16x 6 (c) f 0 (x) = 0 (d) f 0 (x) = 4x 2. The derivatives evaluated at 3 and at 6 are (a) f 0 (3) = 30; f 0 (6) = 30

34 (b) f 0 (3) = 42; f 0 (6) = 90 (c) f 0 (3) = 0; f 0 (6) = 0 (d) f 0 (3) = 12; f 0 (6) = 24 3. R0 (t) = 25 150t; tm = 0:167: An increase in the tax rate beyond tm would decrease tax revenue since the eect of lower sales on tax revenue more than osets the eect of the higher tax rate so higher taxes would reduce consumption. 4. Are the functions dierentiable? (a) Yes (b) Yes (c) No; Not smooth at x = 2 (d) Yes (e) No; Not continuous at x = 1 5. Equal average tax rates: at a given level of income, the tax systems should have secant lines drawn from the origin with the same slope. Equal marginal tax rates: the line tangent to each of the three tax functions at a given level of income should have the same slope. 6. In all cases, the average rate of change of a function is equal to the average of the function's marginal changes over a given interval. The answers for these functions are y y y y = x + 2x 3 =1 x2 4 = 4 3x x2 1 = x2
2

Average Change Marginal Change 7 f 0 (2) = 6; f 0 (3) = 8 2 f 0 (2) = 1; f 0 (6) = 3 3 f 0 (1) = 1; f 0 (5) = 7 1 f 0 (4) = 1 ; f 0 (9) = 1 5 4 6

7. C 0 (q ) = 4q ; AC (q ) = 10 + 2q . The total cost curve, C (q ) = 10 + 2q 2 ; q where q 0; is a convex function. Therefore, the marginal cost, which is measured as the slope of a line tangent to the total cost function at a given point, is greater than the slope of the secant line from the origin to that same point. An increase in output will increase the slope of the secant line, but the slope of the marginal cost curve will always be greater than that of the secant line.

35 8. For this cost function (a) The total revenue function is a concave function with a y-intercept of (0,0). It is increasing for Q = 0 to Q = 8. (b) The derivative of the function, which is marginal revenue, is 8Q (c) Average revenue is
T R(Q) Q dT R dQ

=8 1 Q. 2

(d) In the range where the function is increasing, the slope of the marginal revenue function is less than that of the average revenue function. (e) The slope of the AR curve decreases with increasing values of Q: ; 1 , which means that 9. The roots of the equation y = 4x2 8x + 3 are 12 8 2 the value of the function at these values of x equals zero: The following ordered pairs can be plotted for this function: (0; 3) ; (1; 1) ; (2; 3) (a) f 0 (x) = 8x 8. When f 0 (x) = 0; x = 1: The extreme value is a global minimum. (b) The graph is a convex curve with y-intercept (0; 3). It reaches a minimum 1 global 12 at (1; 1) and passes through the x-axis at ; 0 and ;0 : 2 8

10. The roots of this function are (3; 3; 1), which are all values where y = 0: Other points on the function are (0; 9) ; (2; 5) ;and (1; 16) (b) When f 0 (x) = 0; x = 2:08 and 1:4 which are the two roots of the derivative equation and indicate the values of x where the derivative function passes through the x- axis. (c) The derivative function is convex. 11. The derivative of this function is
dP d ln Y

(a) f 0 (x) = 3x2 2x 9

= 58:2 + 7:8 ln Y:

(a) This derivative equals zero for ln Y = 58:2=7:8 = 7:46, which corresponds to a level of income of e7:46 = $1737.

36 (b) Evaluating this derivative at ln Y = 7:04 we obtain 3:29. Evaluating this derivative at ln Y = 8:60 we obtain 8:88. A sketch of this function, then, would show that it is \atter" than the function presented in the application (and, because the value of ln Y where the derivative is equal to zero is smaller in this function, the minimum point of the function is to the left of the minimum point of the function in the application. The association of the minimum point of a function with the point where the derivative equals zero will be discussed in detail in Chapter 9).

Section 6.4 1. The dierentials are (a) dy = (14x 3) dx (b) dy = 10 1 x dx 2 (d) dy = (3x2 + 3)dx 2. Using the dierential to approximate the changes in the value of the function, we get (a) y = 13:70 (b) y = 0:5 (d) y = 30:3 3. The dierentials and approximate changes are (a) dy = (6x + 1 )dx When x = 0:5; y = 6 1 . When x = 2; y = 3 6 2 24 3 (b) dy = (x) dx When x = 1 ; y = 1 : When x = 10; y = 20 4 2 (c) dy = (3x2 4) dx When x = 8; y = 64: When x = 0:2; y = 1:6 (c) y = 2 (c) dy = (2x)dx

37 4. The percentage dierence results should be read as how much smaller the estimate is than the actual result. (a) x = 0:5 x = 2 (b) x = x = 10 (c) x = 8 x = 0:2
1 4

Actual Estimated % Dierence 15.58 14.833 4.8% 45.333 33.333 26.5% Actual Estimated % Dierence -11.47 -11.50 0.26% 58 8 86% Actual Estimated % Dierence 976 80 92% 17.84 17.6 1.4%

5. With this consumption function

dC (a) The derivative of the consumption function is dY = 0:8, which in economic terms is the marginal propensity to consume, or the portion of current income which is consumed in a given period.

(b) Using the dierential approximation, C = f 0 (y )y = 800 which in this case equals the actual dierence in consumption due to the Y since the consumption function is linear. 6. The answers are in the following table r I (actual) 0.02 0.0 0.025 -0.66 0.030 -1.30 I (estimated) 0.0 -0.67 -1.34 % dierence 0% 1.5% 3.1%

7. The roots of the function are x1 ; x2 = 1; 11. The dierential equation is dy = (10 2x) dx Actual x = 0:5 22.75 x = 1 27 x = 1:75 30.94 x = 3 35 Estimate 24 28 34 44 % Dierence 5.5% 3.7% 9.9% 26%

38 The percentage dierence should be read as how much bigger the differential estimate is than the actual value of y . The dierential approximation becomes a less accurate estimate for the actual value of y as the x gets larger. 8. The dierential can be expressed as dP = (58:2 + 7:8 ln Y0 ) d ln Y where ln Y0 is the initial level of income. (a) ln Y0 = 6:64 and d ln Y = 0:1 so dP = 0:64. (c) ln Y0 = 9:08 and d ln Y = 0:1 so dP = 1:26.

(b) ln Y0 = 6:64 and d ln Y = 0:25 so dP = 1:60. (d) ln Y0 = 9:08 and d ln Y = 0:25 so dP = 3:16.

39

Chapter 7
Section 7.1 1. The derivatives are x2 (a) f 0 (x) = 3 2 (b) f 0 (x) = 0 (d) f 0 (x) = 16x + (c) f 0 (x) = 7x5
3 p 2 x
1

(e) f 0 (x) = 2 x3 2 3 2. Evaluated at 1 and 4, the derivatives are (a) g 0 (1) = 24; g 0 (4) = 96
a (b) g 0 (1) = 2a; g 0 (4) = 32

(c) g 0 (1) = 2; g 0 (4) = 1 (d) g 0 (1) = 2e2 = 14:8; g 0 (4) = 2e8 = 5; 962

3. f 0 (x) = 0 is the derivative of a constant function such as f (x) = 5 for any x. f 0 (x0 ) = 0 is the derivative of a function evaluated at a particular x, which we may call x0 , equals 0. For example, for f (x) = 2x2 + 4x with the derivative f 0 (x) = 4x + 4, we have at x0 = 1; f 0 (x0 ) = f 0 (1) = 0. 4. The derivatives are (a) f 0 (x) = 24x 1

(b) f 0 (x) = 45x2 + 100x + 58 (c) f 0 (x) = 20x3 + 66x2 + 16x (d) f 0 (x) = 5x4 20x (e) f 0 (x) = 6x2 24

40 5. T R = AR Q = f (Q) Q so the derivative of T R, marginal revenue (MR) is MR = Qf 0 (Q) + f (Q) 6. The derivatives are (a) f 0 (x) = ex + xex (b) f 0 (x) = 2e( (d) f 0 (x) = 2e(x
x ) 5

(c) f 0 (x) = a(e(bx) + xbe(bx) )

2 +4)

+ 4x2 e(x

2 +4) 3 x2 +5x6)

(e) f 0 (x) = (9x2 2x + 5)e(3x

7. The derivatives evaluated at particular points are as follows (a) f 0 (x) = 20 8x; f 0 (1) = 28; f 0 (1) = 12 so the derivative is decreasing (b) f 0 (x) = 2e2x ; f 0 (0) = 2; f 0 (2) = 109 so the derivative is increasing
3 ; f 0 (4) = 3 ; f 0 (6) = (c) f 0 (x) = x 4 1 2

so the derivative is decreasing

(d) f 0 (x) = 4(x + 1); f 0 (5) = 16; f 0 (1) = 8 so the derivative is increasing 8. MR(Q) = 70 20q ; M C (Q) = 40q 50; Q = 2 9. = e(i) ;When the interest rate rises, people want to hold fewer real balances and more of other assets which earn interest so M falls as P i rises.
dM P di

10. The instantaneous growth rates are (a) 5 percent (b) 4.5 percent (c) 100 percent

41 Section 7.2 1. The derivatives are (a) f 0 (x) = 4x3 9x2 + 14x + 3 (c) f 0 (x) = (200x + 200)(5x2 + 10x + 3)19 (d) f 0 (x) = abeabx a b1 a1 xa (e) f 0 (x) = b ex ax e 10 2 (f) f 0 (x) = ea+bx+cx 10 (b + 2cx) (a) f 0 (x) = (b) f 0 (x) = (c) f 0 (x) = (d) f 0 (x) =
2x2 14x2 (x2 1)2 2bx3 +cx2 +4 x2 2xe(2x) 2e(2x) x3 9x2 4 x2

(b) f 0 (x) = 198(2x + 4)98

2. The derivatives are

3. For this cost function (a) The derivative of average cost is A0 (x) =
1 x (f 0 (x)x)(1(f (x)) x2

1 x

[f 0 (x) A(x)] where the last term in brackets is marginal cost 1 minus average cost times x :

h f 0 (x)

f (x) x

(b) When the derivative of average cost is less than zero then it is decreasing. Marginal cost is less than average cost. (c) When the derivative of average cost is more than zero then it is increasing. Marginal cost is greater than average cost. (d) If marginal cost equals average cost, then A0 (x) = 0, which is a point of horizontal tangency 4. The derivatives are (a) f 0 (x) = 4x5 + x1

42 (b) f 0 (x) = 8x2 + 24x2 (ln x) (c) f 0 (x) = (d) f 0 (x) = (e) f 0 (x) = 5.
dYf dDef (+) 1 x(1+x) 1 x 1 x(ln 2)

dYf dT BB (+)

dT BA ed (+)

ded dKf (+)

dKf did (+)

did . dDef (+)

A (+) indicates a positive

relationship between the two variables and a () indicates a negative B relationship between the variables. In the case of dT , where the exed change rate is measured in units of domestic currency, a strengthening currency is measured as a decrease in the exchange rate since it takes fewer units of domestic currency to purchase one unit of foreign currency. A stronger exchange rate has a dampening eect on country A's trade balance, since it encourages imports, but it has a strengthening eect on country's B's trade balance by stimulating that country's exports. The ultimate eect of an increase in country A's budget decit, therefore, is to increase country B's income. 6.
d dt

logb f (x) =

f 0 (x) f (x)

1 ln b

7. The derivatives are (a) f 0 (x) = (b) f 0 (x) =

2 ln 2(2x+3) 2 x ln 4 x2 ln 2

(c) f 0 (x) = 3x2 log2 x + (d) f 0 (x) = 8.

d(ln Q) dt
1 log3 ln 3 2x2

= c + f + (1 ) g

9. These elasticities are (a) " = 4 and therefore elastic. Elasticity will change because linear function. (b) " = 1:2 and therefore elastic. Elasticity will change with changes in x and y: (c) " = 0:5 which is inelastic and constant.

43 (d) " = 2 which is elastic and constant 10. The elasticities are Quadratic Sri Lanka 2.0017 Venezuela 2.0000 Sweden 2.0000 11. The answers are (a) QM = 10; PM = 20. If QM = 8; PM = 24 and demand is elastic. If QM rises to 11, this part of the demand curve is inelastic. (b) "q;p = 0:5 which is inelastic. Because it is a constant elasticity, it will not change when quantity demanded changes. 12. The answers are (a) P = 1; Q = 21 (b) "D = 1 7 (c) "S =
2 7

Linear 1.037 1.007 1.001

(d) Response of PP to a change in T = 1 . Response of PC to change 3 2 in T = 3 (e) The price facing consumers rises by more than the price facing producers for a given increase in the tax becasue the elasticity of demand is less than the elasticity of supply in absolute value terms. With T = 0:3, the new PC = 1:20 and the new PP = 0:9: The new Q = 20:40: 13. The answers are (a) P = 7:389; Q = 403:4 (b) "D = 2 (c) "S = 0:5 (d) Response of PP to a change in T = 0:8. Response of PC to change in T = 0:2

44 (e) The price facing consumers rises by more than the price facing producers for a given increase in the tax because the elasticity of demand is less than the elasticity of supply in absolute value terms. With T = 1, the new PC = 7:589 and the new PP = 6:589: The new Q is approximately 381: 14. The marginal change in the eective labor input with respect to time is dL = e0:1t dt which is positive. Thus, this specication shows assumes that workers become more eective the longer their tenure in a position.

Section 7.3 1. The second derivatives are (a) f 00 (x) = 14 6x (c) f 00 (x) = x2

(b) f 00 (x) = 10x3

(d) f 00 (x) = 2ex + 4xex + x2 ex (e) f 00 (x) = 600x26 2. f 0 (x) = 8x 3 ; f 0 (0) = 0; f 00 (x) = is undened
5

40 2 x 3 ; f 00 (0) 3

= 0; f 000 (x) =

80 1 x 3 ; f 000 (0) 9

3. Determining the second derivatives we nd (a) f 00 (x) = 2 Convex. The second derivative is constant. (b) f 00 (x) =
3 p 2 x3

Concave. The second derivative is always negative.

(c) f 00 (x) = 6x 12 Concave and convex portions. The second derivative changes sign. (d) f 00 (x) = 3x2 Concave. The second derivative is always negative.

45 4. The second derivative is f 00 (x) = 2x 8 . The function is concave over the interval [0; 4] since f "(x) 0: Using the denition for concavity which appears in Chapter 2 2 f (2:4) 0:4(8) + 0:6(10 ) 3 5. f 00 (x) = 30x4 so the second derivative is unambiguously positive, and the function is strictly convex since a secant line connecting two points lies wholly below the function itself. When evaluated at f (x0 ), the second derivative equals zero but the function remains strictly convex.
1 ; T P 00 (Q) = 100 L3 . 6. T P (Q) = 150L 3 ; T P 0 (Q) = MP (Q) = 100L 3 3 The negative second derivative indicates diminishing marginal returns.
2 4

7.

i = qi > 0 A positive marginal utility implies that the total util i ity is a monotonically increasing function, as consumer demand the2 ln U i ory requires. d dq = (q 2 2 < 0 The function exhibits diminishing i i ) i marginal utility and is concave.

d ln U dqi

8. The necessary restrictions are found by considering the derivatives and the second deivatives. (a) u0 (c) = aec > 0 if a > 0: u00 (c) = a 2 ec < 0 if a < 0 therefore <0 (b) u0 (c) = c1 if > 0: u00 (c) = ( 1) c2 < 0 if 0 < < 1 9. Coecient of Relative Risk Aversion is x. Coecient of Absolute Risk Aversion is 10. The coecient of absolute risk aversion for the square root utility func1 tion is = 2 . It is = 1 for the logarithmic function. c c 11. The relationship between the actual function and its approximations are x=2 Actual f (x) = 3 (a) Linear h(x) = 3 + 7(x 2) Quadratic j (x) = 3 + 7(x 2) + 3(x 2)2 x = 2:1 f (x) = 3:73 h(x) = 3:7 j (x) = 3:73

46 x=2 Actual f (x) = 1:3863 (b) Linear h(x) = ln 4 + 1 (x 2) 2 1 Quadratic j (x) = ln 4 + 2 (x 2) 1 (x 2)2 4 x=2 Actual f (x) = 403:4 (c) Linear h(x) = e6 + 3e6 (x 2) Quadratic j (x) = e6 + 3e6 (x 2) + 9 e6 (x 2)2 2 x = 2:1 f (x) = 1:4351 h(x) = 1:4363 j (x) = 1:4338 x = 2:1 f (x) = 544:6 h(x) = 566:8 j (x) = 591:3

12. f (x0 ) = e0 = 1: The derivative of ex is simply ex for each successively higher order derivative. Thus the Taylor series expansion for ex evalu2 3 n1 ated around the point x = 0 is ex = 1 + x + x +x + ::: (x where 2! 3! n1)! each x in the Taylor series term represents a value around the original point x = 0:

47

Chapter 8
Section 8.2 1. The partial derivatives are (a) (b) (c) (d) (e) (f)
@y @x1 @y @x1 @y @x1 @y @x1 @y @x1 @y @x1

= 48x3 1 12x1 x2 ; =
2x2 2 14 ; @y (x1 2)2 @x2

@y @x2

2 = 6x2 1 + 12x2 @y @x2

= 6x1 x2 + 24x1 + 5x2 + 20; =

2x1 x2 x1 2

= 3x2 1 + 5x1 + 1

(3x1 ) = 6x2 ; 2e

@y @x2

= 4x2 e(3x1 )

4 = x2 p @y = 2x1 + 2 x2 ; @x = 2

2 @y ; x1 @x2

x1 p x2

2. The second partial derivatives and cross partial derivates are (a) (b) (c) (d) (e) (f)
@2y @x2 1 @2y @x2 1 @2y @x2 1 @2y @x2 1 @2y @x2 1 @2y @x2 1

= 144x2 1 12x2 ; = 6x2 + 24; =

@2y @x2 2

@2y @x2 2

2 = 6x2 1 + 12x2 ; @2y @x1 @x2

@2y @x1 @x2

= 12x1
4x2 (x1 2)2

= 0;

= 6x1 + 5
@ 2y 4x2 2 ; @x @x (x1 2) 1 2

2 4(x1 x2 @2y 2 2x2 7x1 +14) ; 2 2 @x ((x1 2) ) 2

3x1 = 18x2 ; 2e

@2y @x2 2

= 4e3x1 ;

@2y @x1 @x2

= 12e3x1 x2

2 @2y ; @x2 x2 1 2 @2y @x2 2

@2y 4 ; @x x2 1 @x2 2 3

=0
@2y @x1 @x2

= 2;

= 1 x x 2; 2 1 2

= x2 2

3. The partial derivatives at x1 = 1 and x2 = 4 are (a) (b) (c) (d)

@y @x1 @y @x1 @y @x1 @y @x1

= 0;

@y @x2

= 186 =9 =8 = 321

= 88; = 18;

@y @x2 @y @x2

= 1928;

@y @x2

48 (e) (f)
@y @x1 @y @x1

= 2; = 6;

@y = 1 @x2 @y = 7 @x2 2

4. The partial derivatives are

@u @v @u @x @u @y @u @z

= = = =

4v3 4xyz v 4 +x4 +y4 +z 4 4vxyz 4x3 4vyz v 4 +x4 +y4 +z 4 4vxyz 4y3 4vxz v 4 +x4 +y4 +z 4 4vxyz 4z3 4vxy v 4 +x4 +y4 +z 4 4vxyz

Adding together the above partial derivatives and simplifying results 4(v4 +x4 +y4 +z4 4vxyz ) in v4 +x4 +y4 +z4 4vxyz = 4: 5. The solution to this problem is as follows fw (w; x; y; z ) = w 1 ; fww (w; x; y; z ) = ( 1) w 2 fx (w; x; y; z ) = y ln(z ); fxx (w; x; y; z ) = 0 (a) x x fy (w; x; y; z ) = (y ln (z ); fyy (w; x; y; z ) = 22 (y ln (z ) )2 )3 x 1 x 1 fz (w; x; y; z ) = y z ; fzz (w; x; y; z ) = y z2 (b) fxz (w; x; y; z ) =
y

(c) According to Young's Theorem, this function has six cross partial derivatives. 6. In this demand and supply model (a)
@Q @Y

1 ; fwy (w; x; y; z ) = 0 z

= :05;

@Q @PS

(b) "QD ;Y = 1:12: (d) "QD ;PC = 0:16

= 10Ps ;

@Q @PC

= 10

(c) "QD ;PS = 0:56

7. The partial derivative is fx (x; y; z ) = 1 y+x z + x (z + x)2 zy = (z + x)2

which is positive if z > y , negative if y > z , and 0 if y = z .

49 8. The properties of the earnings relationship are as follows (a)

@ ln E @x

= 0:004S + 0:15 0:004X

(b) 5 percent (c) Earnings increase by $495 (d) Earnings increase by 0.2 percent. (e) This specication exhibits diminishing marginal returns with re2 ln E is spect to experience since the second partial derivative @@X 2 negative. 9. The relative return relationship has the following properties. (a) (b) 2:5 rIndia = 2 rU S 3 r @ r I US = 1 A 1 @ A I
US

1 1:5
15 AI AU S

1 1 Q L

21:08

10. For this production function (a) When w = 10, L = 16: When the wage falls to 8, the demand for labor, L, increases to 25. The labor demand schedule is downward sloping reecting the inverse relationship between labor demand and real wages. (b) When capital is allowed to vary and the real wage, w = 8, the labor demanded will be L = 39:06: An increase in the amount of capital is represented by an outward shift in the labor demand schedule, which reects an increase in labor demand at any real wage due to an increase in the productivity of labor.
@ y (c) The cross partial derivative, @L@K = 2p5 : The cross partial is LK positive since the marginal product of either input rises as more of the other input is used.
2

11. The demand for hamburgers

50 (a) varies with price in according to the relationship @QD = 0:25e4 P 1:25 in September @P @QD = 0:25e3:5 P 1:25 in January @P @QD = 0:25e3 P 1:25 in May @P (b) varies with the number of months since September according to the relationship dQD = 0:125e4 P 0:25 e0:125t dt 12. The partial derivative of the weighted arithmetic mean with respect to the iith argument is @xA = !i @xi and the second derivative is zero. The partial derivative of the weighted arithmetic mean with respect to the iith argument is @xG ! i 1 = ! i x! i6=j ! j xi j i @xi

and the second partial derivative is negative (since ! i 1 < 0) and equal to @xG ! i 2 = (! i 1) ! i x! i6=j ! j xi j : i @xi These results are consistent since they show that the eect of an outlier diminshes for the geometric mean, but not the arithmetic mean (which has a second partial derivative of zero).

Section 8.3 1. The derivatives are (a)

dy dz

= 40x 8w

51 (b) (c) (d)

dy dz dy dz dy dz

= (12x2 + 1 z 2 ) (2z 3 ) + 4 = =
@y @v 4xt2 +10t z

dv dz

@y @w

2x2 t+5x p z

dw dz

@y @z

1
2

xz 2

2. The partial derivatives are (a) (b)

@z @u @z @u

= 48xu + 12yu + 2x + 2y ;

@z @v

(d)

@z = (3ax2 2bxy ) + (bx2 + c); @v = (3ax2 2bxy ) + (bx2 + c) (2v ) @z 1 2 4 @z (c) @u 2u; @v = (2ex + xy ) 1 + x = 3x2 24 4 2 y y 1 = (6x2 3y 2 ) 2pu + 0:75y 10u; +v (6xy + 0:75u) 2v @z @u @z @v 1 = (6x2 3y 2 ) 2pu + +v

= 4x 4y

3. The partial derivatives of the trade balance (a) (b)

dT B dM dE dE US = fE dM gE dM gYU S dY dM

dT B dM

are

dE dE dYU S dT B = fE gE gYU S R0 dM dM dM dM dT B = fYJ > 0 dYJ The total impact on the trade balance due to dM is ambiguous while the eect due to dYJ is positive. 4. The homogeneity properties are as follows (a) homogeneous of degree 0 (b) homogeneous of degree 1 (c) homogeneous of degree 3 (d) homogeneous of degree 1 (e) not homogeneous (f) not homogeneous

52 5. The homogeneity properties are (a) Partial derivatives are homogeneous of degree -1. For example, fx (sx; sy; sw) = sw1 = s1 (w 1 ) (b) Partial derivatives are homogeneous of degree 0. For example, fx (sx; sy; sw ) = 2sx1 sw1 = s0 (2x1 w 1 ) (c) Partial derivatives are homogeneous of degree 2. For example, fx (sx; sy; sw) = 3s2 x2 s1 y 1 s1 w1 +2s1 y 1 s1 w1 = s2 (3x2 yw1 )+s2 (2yw ) (d) Partial derivatives are homogeneous of degree 0. For example, s1 s1 x1 p fx(sx; sy ) = p = s2 x2 + s2 y 2 s1 x2 + y 2

6. The production function properties are as follows

(a) Consider the case of a doubling of inputs: A (2K ) (2L) = A 2 K 2 L = AK L 2(+ ) = 2+ Q. Since + = 1, 2Q results which represents a doubling of output given a doubling of inputs (constant returns to scale). (b) For increasing returns to scale, Q increases more than proportionately so 2+ > 2. This requires that + > 1: (c) + < 1 7. The homogeneity properties are (a) The function is homogeneous of degree k =
1 1 7

7 12

since

f (sx1 ; sx2 ) = (sx1 ) 4 (sx2 ) 3 = s 12 f (x1 ; x2 ) (b) The partial derivatives of the production function are @y 1 3 1 @y 1 1 2 3 4 = x1 4 x2 and = x1 x2 3 @x1 4 @x2 3

53
5 where the fi are homogeneous of degree k 1 = 12 as follows:

f1 (sx1 ; sx2 ) =

3 1 5 1 (sx1 ) 4 (sx2 ) 3 = s 12 f1 (x1 ; x2 ) 4 1 2 5 1 f2 (sx1 ; sx2 ) = (sx1 ) 4 (sx2 ) 3 = s 12 f1 (x1 ; x2 ) 3

(c) Euler's Theorem shows that

3 1 1 2 5 1 1 7 x1 ( (sx1 ) 4 (sx2 ) 3 ) + x2 ( (sx1 ) 4 (sx2 ) 3 ) = s 12 f (x1 ; x2 ) 4 3 12 1 1

4 3 x2 where f (x1 ; x2 ) = y = x1

8. wL = 30K 0:4 L0:6 ; rK = 20K 0:4 L0:6 The sum of wL + rK = 50K 0:4 L0:6 9. The answers are (a) Homogeneous of degree 2 (b) Homothetic but not homogeneous (c) Homothetic and homogeneous of degree 2 (d) Homothetic and homogeneous of degree 4 (e) Homothetic but not homogeneous 10. The answers are (a) ey = xz (b) ey = L0:3 K 0:7 (c) ey =
x2 y w

(d) ln (y ) = xz Exercise 8.4 1. The total dierentials are 1 2 (a) dw = 4x + 1 y dx + x 9 y dy 2 2 1 1 (b) dy = (12x2 1 x1 )dx1 + 6 x2 dx2

54 (c) dz =
2xy 3 +x2 y (y3 +xy )2

(d) dy = 4x1 e(3x2

2. The actual and approximate changes in the values of the functions are (a) y = x2 + 4x z 2 2xz
2 +3z

dy (3x2 ) dx1 + 6x2 dx2 1e

dx

3x2 y2 +x3 (y 3 +xy)2

y =24 dy =16

(b) y = ex

y =7007 dy =-2194 y =19.3 dy =6 = 41 e = 2:64 while the actual change, 2

1 z=2 e dz 2

(c) y = ln x3 4z + 2xz (d) dy = 2xdx + y = 3:93.

3. The function y = 3x2 2x + z 3 1:5z 2 xz , evaluated at x; z = (2; 1) = 5:5 Actual Change (y ) x; z = 1 11.5 x; z = 0:5 4.5 x; z = 0:1 0.74 4. dI = (af
a1

Dierential Approx. (dy ) 7 3.5 0.7

Percent Dierence 40% 22% 5.4%

) df +

salary increases by 5. The answers are (a) .05 (b) -50 (c) 5 (d) -2.50

dp p . 100p

+ ( ) dc. If protability increases by 1%,

6. The deriviatives are (a) 6x+x 2 y 1 6y+y 3 x 3 y 2x+y +1 (b) x 2y+x+1

xy (c) 3x2x2+2 y 4
1 2 2 3 1 3

55
3x +wy (d) 3w 3 y 2 +wx y (e) x(2+ y)
2

7. With this production function

K (a) 7 3 L

(b) Slope = 7; Slope = 1 2

(c) MRTS is 7 and 1 , respectively. 2

8. With this production function (a) Slope of isoquant = degree 1. 1 3 x2 . The function is homogeneous of x1

(b) The partial derivatives are homogeneous of degree 0 so f1 (sx1 ; sx2 ) dx2 = dx1 f2 (sx1 ; sx2 ) Therefore, sk1 f1 (x1 ; x2 ) f1 (x1 ; x2 ) = k 1 s f2 (x1 ; x2 ) f2 (x1 ; x2 ) Regardless of the scaling factor, s, the slopes of the isoquants are equal and unchanged. 9. The marginal rates of substitution are (a) MRSA;B = (b) MRSA;B =
B A B 3A

10. This utility function (a) Is homogeneous of degree 2 since (sL) (sT ) = s2 LT . The slope of dL L the indierence curve is dT = T .

(b) The function z = ln L + ln T is homothetic because ln U is strictly increasing and U = LT is homogeneous.

56 (c) The total dierential is dz = and therefore

dL dT

s s 1 1 dL + dT = dL + dT sL sT L T

L = T and s does not appear.

11. For a given level of M , dY Y = 0:4 dH H 12. To maintain a constant value of M , (that is, dM = 0), for a given level of r, we have 1r dH = > 0: dc (c + r) (1 + c) To maintain a constant value of M for a given level of H , we have dc 1+c = < 0: dr 1r

57

Chapter 9
Section 9.1 1. The stationary points are (a) x = 5 which is a minimum (b) x = (1; 2) which are maximum and minimum, respectively (c) x = 0 which is a maximum (d) x =
1 2

which is a maximum

(e) x = (1:80; 0:46) which are maximum and minimum, respectively (f) x = exp = 2:718 which is a maximum 2. Using the second-order conditions we nd (a) f 00 (x) = 2 which is sucient for a minimum (b) f 00 (x) = 2x 3 At x = 1, sucient for a maximum. At x = 2, sucient for minimum. (d) f 00 (x) =
00

(c) f 00 (x) = 15x4 which is sucient but not necessary for maximum
2 x2

which is sucient for a maximum

(e) f (x) = 12x + 8 At x = 1:80, sucient for a maximum. At x = 0:46, sucient for minimum. (f) f 00 (x) =
2 ln(x)3 x3

which is sucient for a maximum.

3. The maximum number of extreme points for the function is n 1: 4. The maximum number of inection points for a polynomial of degree n is n 2. 5. The general formula for determining the critical points of a quadratic function is: + 2x = 0. The general formula for the critical points of a cubic function is: q 2 (2 )2 12 = 0: 6

58 6. A quadratic function has only one bend and is either concave or convex. The second derivative is constant because f 0 (x) is linear, therefore, there is no inection point since the second derivative never changes sign. 7. Substituting the functions, we get 2 R2 R V (R) = 9R 2 = 9R : 4 2 The vote-maximizing use of resources is found by setting the derivative equal to zero, V 0 (R) = 9 R = 0

so R = 9 and then V (9) = 81 81 = 40:5 percent of the vote; she 2 loses anyway. This is a maximum since V 00 (R) = 1.

8. The critical points are Y1 Y2 = 13:72 (local minimum), 3:95 (local maximum) 9. The prot function is also a cubic function with two bends. The rst curvature is convex and crosses the x-axis at the quantity where the concave portion of the total cost curve intersects the total revenue curve. The prot function reaches a maximum at Q the point where there is a maximum positive dierence between the total revenue and total cost curves. The marginal revenue curve is a horizontal line. The marginal cost curve is a u-shaped convex curve, and the marginal prot curve is a concave curve. q C iY iY 10. FOC: dT = c +( 2) = 0 so solving for n = . Using the SOC dn 2c (2n)2 to check that this is a minimum gives us
Y 2n

that average cash balances = M = Section 9.2 1.

= 8( (2iY 3 ) > 0. Assume q n) then M = cY . 2i

d2 T C (dn)2

= (2 + 2 ) exp(( + r) + ay ) = 0 Since the exponential term cannot equal zero, the rst set of terms on the left must equal zero for 2 the product to equal zero. The inection point is = :

d2 S d 2

2. For the rm described in this question

59 (a) The rm's prot function is = 12Q (Q3 4:5Q2 + 18Q 7). The extreme values of this function are Q = 1 and Q = 2: To obtain maximum prots, the rm should produce 2 units. The second order condition is: 9 6Q < 0 so Q > 3 2

(b) Marginal cost = 3Q2 9Q + 18 = 12, which is price. The two values which obtain are Q = 1 and Q = 2: The slope of the marginal revenue curve is zero, and at Q = 2, the slope of the marginal cost curve = 3, which is greater than the slope of the marginal revenue curve. At Q = 1; the slope of the marginal cost curve is -3 which is less than the slope of the marginal revenue curve. 3. In order to maximize prots, the monopolist should produce at Q = 5 which corresponds to a price of P = 2: 4. For the monopolist Babydrink (a) Q = (b) (c)
ac c ; P = a a 2b 2 ct ct Q = a ; P = a a2 2b c t = a 2

5. For these two duopolists, rms A and B , (a) A = (pA c) QA = (pA c) (a pA + bpB ) ; B = (pB c) QB = (pB c) (a pB + bpA ) (c) p A =
a+c b +2 pB 2 a+c pB = 2b

(b) pA =

= RA ; pB =

a+c 2

b +2 pA = RB

6. In the Solow growth model (a) Long-run consumption per capita = c = y i = k a ( + n)k: The dc level of k that maximizes consumption is dk = ak a1 ( + n)

(b) The Golden Rule level of capital accumulation sets the marginal product of capital equal to the sum of depreciation and population growth. The economy with the higher population growth, holding everything else equal, will have the higher marginal product of capital.

60 (c) An increase in the number of workers causes capital per worker to decline because capital is spread over a larger population since in the steady state k = 0: The Solow model also predicts that countries with higher population growth will have lower income/output per capita. 7. The intensive production function is y = k a which incorporates laboraugmenting technological progress and i = ( + n + g ) k . According dc to the Solow growth model, c = y i = k a ( + n + g ) k so dk = 1 p 0:145 = 0 and k = 11:91 which is the Golden Rule level of capital 2 k accumulation that maximizes consumption per eciency unit of labor. Labor-augmenting technological progress and population growth, or an increase in the labor force, have the same directional impact on k: 8. For this monopsonist who is the only purchaser of labor (a) The First order condition is F OC = 9L amount of labor to employ is L = 9 8
2 3

8L 3 = 0: The optimal
5 2

(b) The second order condition is SOC = 6L 3 8 L 3 < 0. The 3 negative coecient suggests diminishing marginal returns to labor. 9. The cost-minimizing quantity is Q = 14:47. The original function has a y-intercept at (0; 500), a local maximum at Q = 5:53, a local minimum at Q = 14:47, and an inection point at Q = 10. The prot maximizing quantity is Q = 12, however, and if the rm produced the cost-minimizing quantity it would be earning less than maximum prot. 10. a. The derivative of the prot function is d = pL 1 (w + b) : dL Setting this equal to zero we have the optimal amount of labor demanded, L , equal to 1=( 1) 1=(1 ) w+b p L = = : p w+b This shows that L decreases as b rises.

61 b. The second derivative is d2 = p ( 1) L 2 < 0 dL2 which shows that the critical point derived in part (a) represents a maximum.

62

Chapter 10
Section 10.1 1. The stationary points are (a) (x; z ) = 2; 1 4 (b) (x; z ) = (4; 1) 1 ;9 (c) (x; z ) = 64 (d) (x; z ) = (0; 0)

2. The stationary points are (a) (u; v ) = (2; 6) (b) (u; v ) = 11 ;2 8

(d) (u; v ) = (128; 256) 3. The stationary points are (a) (a; b) = (5; 3)

(c) u = v = (1; 2)

(b) (a; b) = (0; 0) 1 9 ; (c) (a; b) = 4 4 (d) (a; b) = (1; 1)

4. The function in (d) should be f (x; z ) = 6x +2z 2 x2 +4z . The answers are Original Function a. (3; 2) = 25 b. (18; 8) = 196 c. 1 ; 2 = 12 2 d. (3; 1) = 7 5. For this government x + a 24 195:5 4 6 z + a 22 194 11 9 x a 24 195:5 4 6 z a 21 194 8 9 Maximum Maximum Minimum Saddle Point

63 (a) The level of ination that maximizes seignorage revenue is =

b(X ct) 2(abg )

(b) The tax rate that maximizes income tax revenues is t =

X +g 2c

ct)+gt (c) They choose dierent outcome. = b(X which is 2(agt 2(abg ) bg ) bc b bigger than answer 4a. t = X +g which is smaller than 2c 2 answer 4b. R is higher in the cooperation case.

6. The partial derivatives are @C = 4E + 78 2X @E @C = 4X + 66 2E: @X Setting these equal to zero, we nd the stationary point is (E; X ) = (15; 9) : 7. (E; U ) = (70; 30) 8. (L ; K ) = (2; 4); fLL = 4; fKK = 1 and fLL fKK > (fLK )2 9. For these two cigarette rms
(a) A C = 500 and AM = 250; C = 187; 500 and M = 62; 500; C + M = 250; 000

(b) A C =

3000 7

and A M =

1000 ; C 7

+ M = 306; 122:45

(c) A = 250 ; C + M = 250; 000 10. For this publishing rm c c ; . Both of the second partial derivatives (a) (QT ; QM ) = a2 2 are negative so these stationary points represent a maximum. 2cQM 2cQT (b) (QT ; QM ) = a2( ; +c) 2(+c) (c) With linear costs, Q =
+ c( +) : 2

11. The critical points of this function are (QB ; QS ) = ( 185 ; 40 ). 26 13

64 12. For this multi-plant monopolist

2 (a) = 40(Q1 +Q2 +Q3 )(Q1 + Q2 + Q3 )2 (Q1 + Q2 1 )(Q2 2Q2 ) (2Q2 3 Q3 )

(b) (Q1 ; Q2 ; Q3 ) = (1; 16:5; 1)

Section 10.2 1. The stationary points are (a) Saddle Point (b) Maximum (c) Minimum (d) Maximum 2. The stationary points are (a) Maximum (b) Minimum (c) Maximum (d) Maximum 3. The stationary points are (a) Saddle Point (b) Saddle Point (c) Saddle Point (d) Minimum

65 4. This leads to maximum government revenue under the following condition. We have @ 2R = 2a 2bg < 0 @ 2 @2R = 2c < 0 @t2 @ 2R = g bc @@t so the condition needed for a maximum is 4ac + 6bcg > g 2 + b2 c2 : 5. For this problem @ 2C = 4 < 0 @E 2 @ 2C = 4 < 0 @X 2 @ 2C = 2 @E@X and (4) (4) > (2)2 . 6. 1 4 > 12 , so maximum 7. Conrming maximum prots in part (a) 2 2 > 02 Conrming maximum prots in part (b) (2B 2C ) (2 2C ) = 4 + 4C + 4C + 4C 2 > (2C )2 8. Conrming maximum prots: fQB QB = 6 fQS QS = 7 fQS QS > (fQB QS )2

fQB QB

66 9. The stationary point is r = s = 0. The second derivatives are 2z 2 and @ = 2: @s2 (a) > 0, > 0 so minimum (b) < 0, < 0 so maximum (c) and are dierent, so saddle point 10. The second order conditions are S @2 2bQ N = 2 b 1 e N = <0 2 @P q 1 S 1 S @ 2 = e N aP bP 2 C (a bP ) = e N q (P C ) < 0 2 @S N N 2 S S @ = e N [(a 2bP ) + bC ] = e N [q b (P C )] @P @S where the signs assigned to the second partial derivatives assume that P C > 0, S > 0 and Q > 0. Some algebra will show that the condition 2 2 @2 @2 @ > @P 2 @S 2 @P @S is also met. 11. The rst order conditions are 12. The rst order conditions are @y = 4x 4 + 4z = 0 @x @y = z + 4 + 4x = 0: @z (b) The second partial derivatives and the cross partial derivative are @ 2y = 4 @x2 @ 2y = 1 @z 2 @ 2y = 4: @x@z (a) The stationary point is x = 1 and z = 0.
@2z @r2

67
@ y @ y Thus, we have @x 2 < 0 and @z 2 < 0, two of the conditions needed for the stationary point to be but the other condition a maximum, @ y@ y @ y is not met since @x and, therefore, this stationary 2 @z 2 < @x@z point is neither a maximum nor a minimum.
2 2 2 2 2

Exercises 10.3
@ y 1. For this function, all three second partial derivatives @x 2 = 2 for i = i 1; 2; 3 and all cross partial derivatives are zero. Therefore the stationary point, (0; 0; 0), is a minimum.
2

2. If all cross partial derivatives are zero, then the stationary point is a minimum if fii > 0 for all i and it is a maximum if fii < 0 for all i: 3. For a function with all cross partial derivatives equal to zero, a stationary point is a minimum if all second derivatives are positive when evaluated at the stationary point and a stationary point is a maximum if all second partial derivatives are negative when evaluated at the stationary point. If some second partial derivatives are zero, or if they are of dierent signs, then the stationary point is a saddle point.
4. (x 1 ; x2 ; x3 ) = (12; 9; 9) The sucient condition shows that these represent a local minimum

f11 f22 f33 + 2f12 f23 f13

f11 = 4 > 0 f11 f22 = 4 6 > (f12 )2 = 9 > f11 (f23 )2 + f22 (f13 )2 + f33 (f12 )2

5. (x 1 ; x2 ; x3 ) = (4; 2; 4). The sucient condition shows that these represent a local maximum

f11 f22 f33 + 2f12 f23 f13

f11 = 1 < 0 f11 f22 = 6 > (f12 )2 = 4 < f11 (f23 )2 + f22 (f13 )2 + f33 (f12 )2

68 6. Conrming point of maximum prots f11 = 4 < 0 f11 f22 > (f12 )2 = 4 2 > 22 f11 f22 f33 + 2f12 f23 f13 < f11 (f23 )2 + f22 (f13 )2 + f33 (f12 )2 = 64 < 48 7. In the matrix A= a11 a12 a21 a22

the two 1rst order principal minors are a11 and a22 and the single 2nd order principal minor is the determinant of the matrix, jAj = a11 a22 a12 a21 . This matrix is positive denite if the two leading principle minors, a11 and jAj, are positive. This matrix is positive semidenite if a11 0, a22 0, and jAj 0. This matrix is negative denite if a11 < 0 and jAj > 0. This matrix is negative semidenite if a11 0, a22 0, and jAj 0. Therefore, using these criteria we nd that the matrix in (a) is negative denite, the matrix in (b) is positive semidenite, the matrix in (c) is negative semidenite, and the matrix in (d) is positive denite. 8. A 4 4 matrix has 4 leading principal minors and 15 principal minors.

69

Chapter 11
Section 11.1 1. The extreme points are (a) (x; z ) = (8; 0) Minimum 2 Minimum (b) (x; z ) = 1 ; 3 3

2. (x; z ) = (8; 4) :

(c) (x; z ) = (4; 1)Minimum 32 (d) (x; z ) = 4 ; 3 Maximum 3 (e) (x; z ) = 1 ; 1 Maximum 2 2

3. R = 16; L = 8 (R = 0 yields minimum) 4. U (S ) =

1 2(24S )2

ln(S ) + ln(12 S ); 2 < 0:

1 2

@U @S

= 0 , S = 12;

@ 2U @S 2

2 = (2S )2

5. For the consumption problem with three items (a)

S 4

V 2

J 12

=6
1 3

(b) U (V; J ) = 6V = J

ln(24 2V J )+ 3

1 3

ln(V ) + 1 ln(J ). The ratio is 3

(d) The optimal amounts of soup, salad, and juice are S = 8, V = 4, and J = 24: (e) The solution provides a maximum level of utility since each of the second order conditions are negative. 6. This problem requires the minimization of 50C + 100P subject to the constraint 5 = C 1=3 P 2=3 . Solving for C from the constraint we get C = 125P 2

ln(S )+ 1 ln(12 J S )+ 1 ln(J ). The ratio is 3S = J (c) U (S; J ) = 1 3 3 6 2 3

70 which then gives us the internalized function f (P ) = The rst order condition is f 0 (P ) = 12500 + 100 = 0 P3 6250 + 100P: P2

and the solution is P 3 = 125, that is, P = 5. Substituting this into the constraint, we have C = 5 as the optimal amount of cappuccino. The second derivative, evaluated at P = 5, is f 00 (P ) = 37500 37500 = = 60 > 0 4 P 625

which shows that the stationary point is a minimum. 7. (x; y; z ) = (1; 4; 9) It is a saddle point. 8. V = 6; S = 12; J = 36 Optimal relative proportion is same as the case of $6 budget. 9. The constraint is T + P = 12. Expressing this as T = 12 P , substituting it into the revenue function and taking the deriviative, we nd the rst order condition to be
1 1 1 R (P ) = (12 P ) 2 + 2 4

1 2 P = 0: 2

Solving, we obtain the optimal level P = 4 and, using the constraint, T = 8. The second derivative of the internalized objective function is
3 1 1 U (M ) = (12 P ) 2 4 16

00

3 2 P <0 2

for P = 4. Therefore, this choice of P and T represents a maximum increase in revenues.

71 Section 11.2 1. The solutions are (a) (x; z; ) = (12; 4; 15). (b) (a; b; c; ) = (12; 6; 1; 16)
(136) (c) (x; z; ) = ( 5100 ; 136 ; 4(5100) ). 7 7 76 1 (d) (x; y; ) = (4; 4; 32 ).
3 3

2. For this school lunch problem (a) L = 2C 2 + A 2 (C + 1 A 24); (A ; C ) = (16; 16) 2

1 1

; This is the marginal utility of money available for cookies (b) = 1 4 0 1 and apples. = 4p Marginal utility is diminishing according to 2 the increase in budget.
ln 2 3. X = ln 2+4 ; M = 83 , is the marginal utility regarding the increase 3 in time you devote to these activities.

4. For these two pastries, the optimal amounts are (a) R = 2; T = 4 (b) R = 4; T = 16 (c) Answer: (ii), exactly twice as much 5. (M; L) = (4; 8); GPA = 4:0 6. The optimal amounts are (a) K = (b) K =
C ;L r Q A ) = C (1 w 1 w1 1 r ;L=

Q A

7. (w; x; y ) = (4; 4; 4)

w
r

8. The bordered Hessians are

72 (a) For part (a) 3 0 2 2 H=4 2 1 1 5 2 1 6 2

which has the determinant 24 so the stationary point is a minimum. (b) For part (b) 3 3 1 6 1 1 0 0 7 2 7 H=6 4 3 0 8 0 5 1 0 0 16 2 0
1 2

(c) For part (d), the bordered Hessian is 2 3 0 4 4 1 1 4 4 5 64 64 1 1 4 64 64 and with = minimum.
1 , 32

is 11 and therefore this is a minimum.

The determinant is 184. The determinant of the leading principal minor, which is 2 3 0 1 3 2 4 1 1 0 5 2 3 0 8

the determinant is 1 so the critical point is a

9. The Lagrangean function is L = G 3 U 3 (WG G + WU U B ) where B is the total wage-bill. The rst order conditions are @L 1 2 2 = G 3 U 3 WG = 0 @G 3 @L 2 1 1 = G 3 U 3 WU = 0 @U 3 @L = WG G + WU U B = 0 @
1 2

73 a. Solving the rst order conditions, we get G 1 WU = U 2 WG b. If WU rises then G rises.(c) With B = 1500, WG = 50 and U WU = 25, we have G = 10 and U = 40. Section 11.3 1. The student's utility is maximized with (a) (x1 ; x2 ) = (15; 30) (b) (x1 ; x2 ) = (15; 30) r r 1 1 (c) P ART (a) = = 0:1777; P ART (b) = = 0:00833 32 120 2. Expenditure minimization requires (a) (x1 ; x2 ) = (12:5; 25) (b) The results are the same because the utility constraint in Question 2 was set at exactly the level of utility achieved in Question 1 through the maximization of x1 and x2 (c) If the utility constraint rises to u = 20, x1 rises to 14.14 and x2 increases to 28.28. 3. Solve the Lagrangian function F (K; L; ) = rK + wL (K L1 Q) and get w r = = 1 1 L K (1 ) L K w K = , L 1 r K w , ln = ln + ln L 1 r therefore, the elasticity of substitution is d ln K L = 1 = d ln w r

74
4. Max U (X1 ; X2 ; X3 ) = (X1 +X2 +X3 ) s.t. P1 X1 +P2 X2 +P3 X3 = I . Set Lagrangian function L = (X1 + X2 + X3 ) (P1 X1 + P2 X2 + P3 X3 I ):
1 1

5. The Euler equations are (a) (b)

1 C1

The elasticity of substitution is 2 X3 X3 d log X d log d log X 1 X1 1 2 P2 = P3 X = = = = P3 1 d log P1 d log P2 d log P1 = (1 + R)

C2 1

(c) exp(C1 ) = (1 + R) exp(C2 ) 6. For this problem,

1 1

= (1 + R)

, 1 = (1 + R)

I1 = C1 + S1 (1); I2 + (1 + R) S1 = C2 + S2 (2); I3 C3 I3 + (1 + R) S2 = C3 (3); , (1 + R) I1 + I2 + = (1 + R) C1 + C2 + 1+R 1+R If we put (1 + R) I1 + I2 + = I , the Lagrangian function will be C3 I : L = u(C1 ) + u(C2 ) + u(C3 ) (1 + R) C1 + C2 + 1+R Solve this for . = u0 (C1 ) = u0 (C2 ) = (1 + R) u0 (C3 ): 1+R
I3 1+R

7. The savings problem with period 2 income is as follows (a) Put Y1 ; C1 as income and consumption at the rst period. A = Y1 C1 (1 + r)A + Y2 = C2 + R + (E S ) A consolidated intertemporal budget constraint is (1 + r )Y1 + Y2 = (1 + r )C1 + C2 + R + (1 )E

75 (b) Put (1 + r )Y1 + Y2 = Y : The Lagrangian function is written as p p p p L = 2 C1 + 2 C2 + 2 E + 2 2 R (1 + r )C1 + C2 + R + (1 )E Y Solve the Lagrangian function and we get C1 = C2 =
2

(Hint: Express all variablesh in terms of C2 and plug them i into the 1 = Y ) constraint. You will get C2 2 (1+1r) + 1 + 2 + 1

(1 + r ) 2 (1 ) (Y + ) (1 ) + (1 + 2 ) 2 (1 ) (1 + r ) + 2 (1 + r ) (1 + r ) 4 (1 ) (Y + ) R = (1 ) + (1 + 2 ) 2 (1 ) (1 + r ) + 2 (1 + r ) (1 + r ) 2 (Y + ) E = (1 ) (1 ) + (1 + 2 ) 2 (1 ) (1 + r ) + 2 (1 + r )

(1 ) (Y + ) (1 + r ) (1 ) + (1 + ) 2 (1 ) (1 + r ) + 2 (1 + r )

I2 I2 8. I will be W0 + I1 + (1+ , rather than I1 + (1+ . However, the Euler R) R) 0 equation will be the same as u (C1 ) = (1 + R) u0 (C2 ). A non-zero initial stock of wealth does NOT alter the results of the model except for the fact that the initial income dier. (W0 + I1 , rather than I1 )

9. This time-allocation problem has the following solution. (a) The optimal allocations are A1 = 3:75 A2 = 2:50 w = 10:5 t1 = 7:5 t2 = 6:0 (b) G1 = 7:50; G2 = 15

76 (c) This technological change would reduce n1 thereby increasing the relative amount of activity A1 (d) This improvement in \physical productivity" will reduce n2 thereby increasing the relative amount of activity A2 (e) The increase in wages will decrease the consumption of the more time-intensive activity, A2 . Section 11.4 1. The Lagrangian function is L = 2x2 y 2 1 x2 + y 2 1 + 2 x + 3 y @L = 4 x 2 1 x + 2 = 0 @x @L = 2y 21 y + 3 = 0 @y 1 x2 + y 2 1 = 0 2 x = 0 3 y = 0 2 0; 3 0; x2 + y 2 = 1 The optimal solution is x = 1, y = 0, which means that 2 = 0: 2. The Lagrangian function is L = 2a2 + b2 (2a + b 9) a2 + b2 16 @L = 4a 2 2a = 0 @a @L = 2b 2b = 0 @b (2a + b 9) = 0 2 a + b2 16 = 0 0; 0; 2 16 a + b2 ; 2a + b 9

and the solution that provides the maximum value must satisfy

and the solution that provides the maximum value must satisfy

77 The optimal solution has 6= 0, = 0, with a = b = 3: 3. The Lagrangian function is 1 1 1 S V J L = ln (S )+ ln (V )+ ln (J ) + + 6 (S + J 40) 3 3 3 4 2 12 and the solution that provides the maximum value must satisfy 1 @L = =0 @S 3S 4 @L 1 = =0 @V 3V 2 @L 1 = =0 @J 3 J 12 S V J + + 6 =0 4 2 12 (S + J 40) = 0 0; 0; S V J + + 6; S + J 40 4 2 12 The optimal solution has 6= 0, = 0, with S = 8, V = 4, and J = 24. 4. For this problem; (a) The Lagrangian function is L = AK N 1 (wN + rK c) + 1 N + 2 K where labor is N and capital is K . The rst order conditons are @L = AK 1 N 1 r + 2 = 0 @K @L = (1 ) AK N w + 1 = 0 @N (wN + rK c) = 0 1 N = 0; 2 K = 0 2 0 0; 1 0; wN + rK c

78 The optimal solution involves 1 = 0 and 2 = 0 while wN + rK = c and 6= 0which gives us rK = c and wN = (1 ) c for optimal amounts spent on capital and labor services, respectively. (b) The Lagrangian function is L = (wN + rK ) AK N 1 Q + 1 N + 2 K: @L = r AK 1 N 1 + 2 = 0 @K @L = w (1 ) AK N + 1 = 0 @L AK N 1 Q = 0 1 N = 0; 2 K = 0 0; 1 0; 2 0 AK N 1 Q The optimal solution involves 1 = 0 and 2 = 0 while AK N 1 = Q and 6= 0 which gives us K 1 w = N r (c) These solutions are the same as the respective solutions in exercise 11.2.6 since the optimal solution requires using some of each factor of production and either using the full budget, in the output maximization problem, or producing no more than the required amount, in the cost minimization problem. The optimization functions with inequality constraints thus are the same as those with equality constraints. 5. The Lagrangian function is L = u (x1 ; x2 ) (p1 x1 + p2 x2 m) (t1 x1 + t2 x2 T )

The rst order conditons are

79 and the solution that provides the maximum value must satisfy L1 L2 (p1 x1 + p2 x2 m) (t1 x1 + t2 x2 T ) m = = = = u1 p1 t1 = 0 u2 p2 t2 = 0 0 0 0; 0; p1 x1 + p2 x2 ; T t1 x1 + t2 x2 :

It is will not be the case that p1 x1 + p2 x2 = m and t1 x1 + t2 x2 = T since then more of one or both goods could be consumed, which would raise utility. So it is unlikely that the solution is one where = 0 and = 0. It is also unlikely that both 6= 0 and 6= 0 since this would mean that both constraints would be binding. It is more likely that either 6= 0 and = 0, in which case the budget constraint is binding and the time constraint is not binding, or = 0 and 6= 0, in which case the time constraint is binding and the budget constraint is not binding.

80

Chapter 12
Exercises 12.1 1. For this asset; (a) 1; 000 (b) 941:76 (c) Overestimate equals 985; Underestimate equals 955 (d) Overestimate equals 977:60; Underestimate equals 963 (e) Overestimate equals 973:04; Underestimate equals 968:19 (f) Continuous stream equals 970:61 2. The price of each of the following bonds is (a) 975:41 (b) 1812:69 (c) 3625:39 (d) 2673:76 3. The antiderivatives can be checked by taking the derivatives. 4. The antiderivatives are (a) F (x) = 5x +C 2 x3 (c) H (x) = 1 3
x 2 2 2

x3 + C (b) G(x) = 1 3 (d) J (x) = 2e + C (e) K (x) = ex + C

5x2 2

+C

5. The values of the antiderivatives evaluated over the interval 1 to 5 are (a) 60

81 (b) 41 1 3 (d) 21:0675 (e) 7:210 6. Given your job oer: (a) The company will pay you approximately $506,107 over the course of ten years. (b) The present discounted value = $374; 933 (c) The present discounted value = $446; 389 so it is more attractive to accept the higher starting oer with the lower growth rate. 7. The stream of services is discounted by the interest rate to obtain the present value of services. The present value of a given stream of services decreases as the interest rate increases. 8. The price of a durable will decline over a given length of time given a rate of depreciation, v: 9. Marginal revenue equals
dz dx

(c) 18 2 3

= 4e0:5x +

x2 2

+C

10. For this rm with the given marginal cost function (a) C (Q) = 20e0:2Q +Constant (b) c = 100

Exercises 12.2 1. Integration yields x4 + C (a) F (x) = 1 4 (b) F (x) = x + C (c) F (x) = 2 x2 + C 7
7

82 (d) F (x) = 3x1 + C (e) F (x) = 2e2x + C (f) F (x) = e 4 + C (g) F (x) = ln(6x3 + 2) + C 2. Evaluating the integrals in question 1 over x = 2 to 5 (a) 152:25 (b) 3 (c) 76:63 (d) 0:9 (e) 43; 943:74 (f) 515:29 (g) 2:71 3. The integrals are (a) F (x) =
x4 4
x2

+2 x3 + 4x + C 3
5

(b) F (x) = e2x ln(3x2 1) + 2x 2 (d) F (x) = 2 x2 + 2 x2 2 x 2 5 3 3 4. Using substitution, we nd (a) ln (x2 + 4) + C (b) 2e
p x

(c) F (x) = x2 2x 2 ln(x)

5 3

+C
1 12

(c) F (x) =

(d) F (x) = 2ex + C

(x2 1) + C
2

5. Using integration by parts, we nd (a) F (x) = x ln(x) x + C

x3 3

(b) F (x) =

ln(x)

x3 9

+C

83 (c) This requires repeated use of integration by parts. Each time, let dv = e4x dx and therefore v = 1 e4x . The solution is 4 x3 4x 3x2 4x 3x 4x 3 4x e e + e e + C: 4 16 32 128 (d) Let u =
x2 2

p and dv = 2x 1 + x2 dx. Then we have the solution 3 5 2 x2 1 + x2 2 1 + x2 2 : 3 15

6. Over the interval x = 1 to x = 3; the values of the integrals are, approximately, (a) 0:96 (b) 5:86 (c) 21; 845:33 (d) 16; 200:73 7. The capital stock at the end of year 25 equals 104. 8. The integral describing the capital stock is Z
25

p 121t + 100dt +

25

10e0:04t dt = 96128

9. The change in greenhouse gases is (a) greenhouse gas = 17; 183 (b) greenhouse gas = 19; 651 10. The present value of the land = $500; 000:

84 Exercises 12.3 1. The present values are (a) Continuous payment equals $1; 014: Payment at the end of the year equals $979: (b) $1; 024 (c) $1; 001:67 (d) Parents should invest in the 10 year bond since the value of the longer term bond will increase more rapidly than that of the veyear bond given a change in interest rates. 2. For these stamps which require special storage (a) Total present value equals 2000(1 e0:05t ) (c) t = 18

(b) N P V (t) = 1000(t)ert 2000(1 ert ) = (1000t +2000)ert 2000 3. Producer's surplus is as follows; (a) Producer surplus increases by 65: (b) Producer surplus increases by 158: 4. In the market for bannanas (a) M = 2:5 (b) PD = 3:55 (c) Consumer surplus decreases by 4, while producer surplus increases by 3:1. The dierence, 0:9, is considered a deadweight loss due to the imposition of a tari high enough to eliminate all imports. 5. The benets of reducing ination under dierent scenarios are (a) Present value = 100V , which when V = 0:0028 of national income, equals 28% of a year's national income. (b) If the discount rate equals the rate of growth in the economy, the present value of the benet of reducing ination to zero is zero.

85 (c) If the real discout rate exceeds the rate of growth, there is actually a cost to reducing ination 6. For this probability distribution function (a) The probability of being in this range is Z4
1 4 3

3x2 dx =

13 32

(b) The probability of being in this range is Z1

1 2

3x2 dx =

7 8

(c) The mean is Z1

0

x 3x2 dx =

3 4

(d) Checking over the relevant range Z1

0

3x2 dx = 1

(e) The variance is 2 Z1 3 3 3x2 dx = : x 4 80

0

7. For this probability distribution (a) To nd U , note that we need ZU

1

1 dx = ln (U ) ln (1) = ln (U ) = 1 x

so U = e:

86 (b) Given U = e, Z2
1
e

e 1 dx = ln ln (1) = 0:3 (approx.) x 2

(c) Given U = e, Ze
e 2

e 1 dx = ln (e) ln = 0:7 (approx.) x 2

(d) The mean is found by Given U = e, Ze

1

1 x dx = e 1 = 1:72 (approx.) x

8. For the probability distribution function g (x) = x1 ex (a) With = 2 and = 1 , 2 Z1

0 2 xe 2 x dx = e 2 x j1 = 0:39 (approx.) 0= 1 e
1 2 1 2 1

(b) For this range from 1 to 2 Z2

1 2 xe 2 x dx = e 2 x j2 e2 = 0:47 (approx.) 1= e
1 2 1 2 1

(c) Over the range from 0 to 1 Z1

0

x1 ex dx = ex j1 0 = 1 0 = 1:

87 9. For the uniform distribution with lower limit a and upper limit b, the b mean is a+ and the variance is 2 a+b 2 1 x dx 2 ba 0 a 3 1 x b+ 2 = jb a ba 3 a 3 a 3 b b+ a b+ 2 2 = 3 (b a) ba 3 ab 3 2 = 2 3 (b a) 2 (b a)3 = 24 (b a) (b a)2 = 12 10. The variance is the solution to the integral Z1
0

Z1

e 2 dx (x 2) 2
2
x

To integrate by parts, let u = (x 2)2 and dv = 1 e 2 dx which can be 2 solved to obtain (x 2) e

2 x 2

j1 0

Z1
0

(2x 4)2 e 2 dx:

x

To integrate by parts now, let u = (2x 4) and dv = e 2 dx which can be solved to obtain (x 2) e (x 2) e
2 x 2 2 x 2

(2x 4) 2e
x 2

(2x 4) 2e

x 2

+ 8e

x 2

j j

1 0 1 0

Z1
0

4 e 2 dx =

= 4:

88 11. As shown in the application Measuring Inequality, the Gini coecient associated with this continuous Lorenz curve function L (p) equals G= 12 Z1
0

L (p) dp = 1 2

Z1
0

p [p 0:375 p (1 p)] dp:

Evaluating the integral, we get Z1

0

p p2 1 [p 0:375 p (1 p)] dp = j 2 0

Z1
0

p [0:375 p (1 p)] dp

1 0:375p1:5 1 0:375p2:5 1 j0 + j0 = 2 1:5 2:5 1 0:375 0:375 = + = 0:4 2 1:5 2:5

89

Chapter 13
Section 13.1 1. The rst six values for the dierence equations are: (a) v0 = 200; v1 = 205; v2 = 209:4; v3 = 213:2; v4 = 216:6; v5 = 219:5 (b) w0 = 50; w1 = 4; w2 = 6:8; w3 = 4:6; w4 = 5:1; w5 = 5 (c) x0 = 120; x1 = 90; x2 = 82:5; x3 = 80:6; x4 = 80:2; x5 = 80 (d) y0 = 30; y1 = 10; y2 = 30; y3 = 10; y4 = 30; y5 = 10 (e) y0 = 0; y1 = 5 ; y2 = 2
15 ; y3 4 35 ; y4 8 75 ;y 16 5 155 : 32

2. For the equations in question 1 (a) (a), (c) and (e) are monotonic while (b) and (d) are oscillatory (b) (a),(b),(c) and (e) are stable while (d) is not stable. 3. The steady states are: (a) v1 = 240 (b) w1 = 5 (c) x1 = 80 (d) y1 is not well-dened, (e) y1 = 10: 4. The steady states are: (a) x1 = 60 (b) y1 = 60 (c) z1 = 30 5. The rst values of these equations are (a) x0 = 100; x1 = 80; x2 = 85; x3 = 76:3; x4 = 76:6; x5 = 72:3. Stable (b) y0 = 100; y1 = 90; y2 = 76:7; y3 = 76:6; y4 = 72:3; y5 = 71:3: Stable (c) z0 = 40; z1 = 30; z2 = 50; z3 = 70; z4 = 150; z5 = 350: Unstable 6. The steady state points are (a) 25; (b) 10; (c) 22; (d) 25. See Figures 13.1.6(a), 13.1.6(b), 13.1.6(c) and 13.1.6(d).

90

7. The starting salary, s0 = 40; 000: The equation showing the relationship in subsequent years is st = 0:95st1 + 2500 (1:1)t1 which gives the sequence s0 = 40; 000; s1 = 40; 500; s2 = 41; 225; s3 = 42; 189; s4 = 43; 407; s5 = 44; 897 and s6 = 46; 678. The percentage increases are, respectively, 1:25%; 1:79%; 2:34%; 2:89%; 3:43% and 3:97%. 8. In this model for the market for hogs (a) The steady state values are P1 = Q1 bF bc + a + B 1 + bg c F + ag + Bg = 1 + bg

91 (b) The dierence equations are

Pt = bgPt1 + bF bc + a + B Qt = bgQt1 + c F + ag + Bg

There will be oscillatory and stable behavior of price and quantities if 1 > bg > 0.

9. The dierence equation for the number of cars is

ct+1 =

1 2 10 c + : 11 t 11

(a) The two steady states are found by nding the roots of c2 1 11c1 + 10 = 0. These roots are c1 = 1; c1 = 10: (b) The function in the phase diagram intersects the 45 degree line at the two steady states, c1 = 1 and c1 = 10 with the function below the 45 degree line in the range 1 < c < 10 and above it for c < 1 and c > 10. The steady state c1 = 1 is stable while the steady state c1 = 10 is unstable. See Figure 13.1.9 (c) With c0 = 0 the number of cars goes to the steady state of c1 = 1. With c0 = 4 the number of cars goes to the steady state of c1 = 1. With c0 = 12 the number of cars rises without bound.

92

Section 13.2 1. For the Keynesian model presented in the text:

(a) The steady state level of income is 1000. and it (b) The slope of the linear function in the phase diagram is 2 3 intersects the 45 degree line at Y = 1000. See Figure 13.2.1.

93

2. For the open-economy Keynesian model, after substitution, we nd that the dierence equation is Yt = (1 ) 1 2 Yt1 + ( + It + Qt ) 1 + 1 + 1 1 +

(a) The steady state is Y1 =

3 2

( + It + Qt )

(b) The phase diagram, with Yt1 on the horizontal axis and Yt on the vertical axis, includes a 45 line and a line with the slope 1+2 1 and the intercept 1 ( + It + Qt ). These two lines intersect + at the steady state identied in part (a).
(1)

(c) Assuming the slope of the line is positive and less than 1, there is a monotonic convergence to the steady state.

94

3. Using the the fact that the steady state value, x1 , is x1 = y : 1a

and that the half-life is calculated as H = 0:6931 ln (2) : ln (a) ln (a)

we obtain the following answers: (a) H = 0:6931= ln (0:25) = 0:6931= (1:3863) = 0:50 periods. We have x1 = 75 so the dierence between x0 and x1 is 75. The x1 steady state is 160 = 80 so x0 + = 40. 0:25 2

(b) H = 0:6931= ln (0:75) = 0:6931= (0:2877) = 2:41 periods. We have x2 = 35 so the dierence between x0 and x2 is 35. The x1 steady state is 120 = 80 so x0 + = 40. 0:75 2 (c) H = 0:6931= ln (0:95) = 0:6931= (0:0513) = 13:51 periods. Using (13:8), we have x14 1 0:9514 4 = 40:99 = 0:95 0 + 1 0:95
14

so the dierence between x0 and x14 is 40:99. The steady state is x1 x1 4 = 80 so x0 + = 40. Note that xA x0 is closer to x0 + 10:95 2 2 for larger values of A. 4. The initial deviation between the starting value, x0 ; and the steady state value, x1 , will be cut by one-third at time T if the value of x in that period, which we denote as xT , is xT = x1 + 2x0 : 3

Following the analysis in the application, at time T , we have x1 + 2x0 = aT x0 + 1 aT x1 3

95 which can be rewritten as 2 3aT x0 = 2 3aT x1 : This can only be true, when x0 6= x1, if 3aT = 2 and, solving for T , we nd T = ln (2=3) 0:4055 : ln (a) ln (a)

Using this result for the given equations

2 (a) xt = 0:85xt1 + 60; H = lnln = 4:26, T = 0:85 ln(2=3) ln 0:85

= 2:50.

2 (b) xt = 0:90xt1 ; H = lnln = 6:58, T = 0:90

ln(2=3) ln 0:90

= 3:85. = 40:35.

2 (c) xt = 0:99xt1 + 1; H = lnln = 68:96, T = 0:99

ln(2=3) ln 0:85

5. We can use these equations to obtain a dierence equation in pt , pt = 1 pt1 + : 1+ 1+

(a) The steady state value of p is p1 =

1+ 1 1 1+

= :

(b) The coecient on pt1 is positive and less than 1. Therefore the dynamics are stable and monotonic. 6. The results are that (a) This doesn't aect the number of periods in the half-life. (b) It takes more periods until half the dierence between the initial value and the steady state is reached (i.e. there is slower convergence) when is larger.

96 (c) It takes fewer periods until half the dierence between the initial value and the steady state is reached (i.e. there is faster convergence) when is larger. 7. The steady state value is p1 = 100. The slope of the line representing the dierence equation in the phase diagram is 1 . See Figure 13.2.7. 3

8. With adaptive expectations, pt = If = 1 then pt = 1 1 pt1 + (mt (1 ) mt1 ) : 1 1 1 pt1 + mt 1 1

which is stable if 1 < 1, that is, if < 1 : Thus stability requires that 2 money demand is not \too sensitive" to interest rates when = 1.

9. Initially, with a constant stream of dividends equal to $250 and r = 0:1, p0 = 250 = 2500. 0:1

97 (a) p0 =
300 0:1

= 3000: i 9 X 1 50: 1+r t=1

(b) The easiest way to calculate this is to subtract from $3000

From the discussion in Section 13.1 we nd ! 1 10 i 9 X 1 1+ 1 r 50 = 50 1 = 337:7 1 1 + r 1 1+ r t=1 so p0 = $2662: (c) p0 = 2500 + 150 (d) p0 = 2500 + (e) p0 =
250 0:5

= 500:

1 2 = 2624: 1+r 1 10 150 1+r = 2558:

98

Section 13.3 1. For the second-order dierence equation 8 7 32 xt = xt1 xt2 + 9 81 9 (a) The steady state is found by solving 8 7 32 x1 = x1 x1 + 9 81 9 which gives us x1 = 4 (b) The two roots are 18 k1 ; k2 = 29 s 8 2
9

7 1 7 = ; 81 9 9

the equation is stable since the roots are less than 1 in absolute value. (c) To solve for the constants, note that at time 0 and at time 1, respectively, 3=A+B+4 7 1 5=A +B +4 9 9 which can be solved to give us A = to the equation is 5 xt = 3
5 3

and B = 8 so the solution 3

t t 7 8 1 + 4: 9 3 9

2. For the second-order dierence equation 2 1 xt = xt1 xt2 + 4 3 9

99 (a) The steady state is found by solving 2 1 x1 = x1 x1 + 4 3 9 which gives us x1 = 9: (b) There is a repeated root since 1 k1 ; k2 = 3 r 1 1 1 = 9 9 3

the equation is stable since the root is less than 1 in absolute value. (c) Using the initial values, we nd the value of the constants by solving 5=A+5 1 6 = (A + B ) + 5 3 which gives us A = 0, B = 3 so the solution is t 1 xt = 3t + 5: 3 3. The roots of this equation are k1 ; k2 = 0:10 1 (0:04) 0:65 4 p = 0:10 0:8 1: r

The real parts of the two complex roots are less than 1 so this is a stable, oscillatory equation. 4. For this system of two equations (a) The two characteristic roots are 1 = 0:25 and 2 = 1:25. (b) The characteristic column vector associated with 1 = 0:25 is [1 1]0 or any other column vector where the one element is 1 times the other element. The characteristic column vector associated with 2 = 1:25 is [1 1]0 or any other column vector where the two elements are equal.

100 (c) A matrix of the characteristic vectors, P , and its inverse, P 1 , are 1 1 1 1 1 2 2 : P = P = 1 1 1 1 2 2 Thus if we dene ut vt
1

=P

xt yt

xt 2 xt 2

yt 2 yt 2

and noting that P 1 AP = , the diagonal matrix of the characteristic roots, we have 5 ut 0 ut1 4 = vt 0 1 vt1 4 or ut = 5 u and vt = 1 v : 4 t1 4 t1 (d) The solution to ut , given some u0 which is a linear combination of x0 and y0 , is t 5 ut = u0 4 while the solution to vt , given some v0 which is a linear combination of x0 and y0 , is t 1 v0 : vt = 4 The solution to ut is unstable and the solution to vt is stable. (e) The solution to the original system is found by premultiplying by P to get # " 5 t u xt 1 1 0 4 : = 1 t yt 1 1 v0 4

Writing this out, and expressing u0 and v0 as linear combinations of x0 and y0 , we have " # " t t x0 y0 x0 y0 5 1 5 t 1 t x0 5 t 1 xt = + + = + + 4 2 2 4 2 2 4 4 2 4 4 " # " t t t t t 5 x0 y0 1 x0 y0 5 1 x0 5 1 yt = + = + + 4 2 2 4 2 2 4 4 2 4 4

101 5. For this 2 2 system (a) The characteristic equation is 2 (a11 + a22 ) + (a11 a22 a12 a21 ) = 0: (b) The roots of the characteristic equation are q (a11 + a22 ) (a11 + a22 )2 4 (a11 a22 a12 a21 ) 1 ; 2 = 2 or, in terms of the trace, tr (A) = a11 a22 and determinant, det (A) = (a11 a22 a12 a21 ) q tr (A) tr (A)2 4 det (A) 1 ; 2 = 2 When the roots are real, both roots are positive if tr (A) > 0 and det (A) > 0; both roots are negative if tr (A) < 0 and det (A) > 0. One root is positive and one is negative if det (A) < 0. (see Chapter 14, section 3 for a discussion of a related point in the context of dierential equations).

102

Chapter 14
Section 14.1 1. The functions are characterized as follows; (a) non-homogeneous, unstable (b) homogeneous, stable (c) non-homogeneous, stable (d) non-homogeneous, unstable 2. The steady state values are (a) 1 2 (c) 3 (d) 8 3. The steady state values are (a) 45 (b) 7 4. The solutions are (a) for x (0) = 1, for x (0) = 5, x (t) = (b) for x (0) = 0, x (t) = 0 for x (0) = 3; x (t) = 3e2t : 11 10t 1 e : 2 2 1 1 x (t) = e10t 2 2

(b) 0

103 (c) for x (0) = 0 for x (0) = 3 for x (0) = 5 x (t) = 2e 3 t + 3: (d) for x (0) = 4 for x (0) = 8 for x (0) = 12 x (t) = 4et + 8: 5. The limiting values as t ! 1 are (a) for x (0) = 1; lim x (t) = 1; for x (0) = 5; lim x (t) = 1; t!1 t!1 initial value matters. (b) for x (0) = 0; lim x (t) = 0; for x (0) = 3; lim x (t) = 0; initial t!1 t!1 value does not matter. (c) for x (0) = 0; lim x (t) = 3; for x (0) = 3; lim x (t) = 3; for x (0) = 5; lim x (t) = 3; initial value does not matter. (d) for x (0) = 4; lim x (t) = 1; for x (0) = 8; lim x (t) = 8; for x (0) = 12; lim x (t) = 1; initial value matters.
t!1 t!1 t!1 t!1 t!1 t!1
2

x (t) = 3e 3 t + 3 x (t) = 3

x (t) = 4et + 8 x (t) = 8

6. In part (a), the line representing the dierential equation in the phase diagram has a negative slope and its x-intercept is 4. This is a stable equation and x (t) monotonically approaches its steady state of 4 over time. See Figure 14.1.6(a). In part (b), the line representing the dierential equation in the phase diagram has a positive slope and its x-intercept is 16. This is a stable equation and x (t) increases without bound, if x (0) > x (1) or decreases without bound if x (0) < x (1). See Figure 14.1.6(b). In part (c), the line representing the dierential equation in the phase diagram has a negative slope and its x-intercept is 15. This is a stable equation and x (t) monotonically approaches its steady state of 15 over time. See Figure 14.1.6(c).

104

7. To nd the half-life, which occurs at time H : (a) First note that x _ (t) = ax (t) + b, which is stable since a < 0, has the solution b at b x (t) = x (0) + e a a and at time H b b x (0) + x1 x (H ) = x (0) + eaH = : a a 2
b Note that x1 = a . Solving for H

x (0) b b x (0) + eaH = a a 2

b a

105 therefore b aH x (0) + x (0) + e = a 2

b a

and 1 1 e = implying aH = ln 2 2 1 0:69 since ln 0:69: H a 2

aH

(b) No, H is only a function of the parameter a. 8. The solutions are as follows; (a) = 0:02 and therefore With Y (0) = 2; 000; 000; 000,
d ln Y (t) dt dY (t) 1 dt t

= 0:02 thus

dY (t) dt

= 0:02t.

Y (t) = 2; 000; 000; 000e0:02t . (b)

(t) 1 = 0:1 and therefore dE = 0:1 thus dt t 0:1t With E (0) = 125, E (t) = 125e . d ln E (t) dt dE (t) dt

= 0:1t.

9. The solutions are (a) For x _ (t) = 2x (t) + b (t) with b (t) = 0 for 0 t 10 and b (t) = 100 for t > 10 and x (0) = 1. For t 10, x (t) = e2t and for t > 10; x (t) = e2t + =e
2t

+ 50 1 e202t :

s=0

Zt

100e2(ts) ds

106 (b) For x _ (t) = 3x (t) + g (t) with g (t) = 6 for 0 t 5, g (t) = 3 for t > 5, x (0) = 2. For 0 t 5, x (t) = 2e3t + and for t > 5; x (t) = 2e3t + Z5 6e3(5s) ds + Zt 3e3(ts) ds Zt 6e3(ts) ds

s=0

s=0

= 2e3t + 2 e15 1 +

s=5

s=5 Zt

3e3(ts) ds:

(c) For x _ (t) = 10x (t) + h (t) with h (t) = e2t + 10 and x (0) = 3, x (t) = 3e
10t

s=0

Zt

= 3e

10t

s=0 8t e = 3e10t + e10t + 1 e10t : 10

s=0

Zt

2t e + 10 e10(ts) ds ds + 10 Zt e10(ts) ds

e e

8t 10s

10. The phase diagrams require nding the roots of the equations and characterizing the value of the functions as positive or negative at points other than the roots. (a) s _ = 4s2 + 8s has two steady state points, s1 = 0 which is unstable, and s1 = 2 which is stable. The line representing the phase diagram crosses the x-axis at these points and is above the axis in the range 0 < s < 2 and elsewhere is below the axis. See Figure 14.1.10(a). (b) m _ = m3 15m2 +36m has three steady state points, m1 = 0 which is unstable, m1 = 3 which is stable, and m1 = 12 which is

107 stable. The line representing the phase diagram crosses the x-axis at these points and is above the axis in the range 12 < m < 3 and m > 0 and elsewhere is below the axis. See Figure 14.1.10(b).

108

Section 14.2 1. The solution for _ (t) = p _ (t) (t) is (t) = p _ (0) e
t

s=0

Zt

p _ (t) e(ts) ds

where p _ (0) is the initial level of actual ination, which equals the initial level of expected ination (0). 2. The money demand equation can be written as p _ (t) = 2p (t) 2m (t). It has the forward solution Z1 p (0) = 2m (s) e2s ds:
s=0

(a) For m (s) = 10, p (0) = e2s m (s) j1 0 = 10: (b) For m (s) = 10 from t = 0 to t = 20 and m (s) = 15 thereafter e2s m (s) j1 p (0) = e2s m (s) j20 20 0 40 = 10 1 e + 15e40 = 10 + 5e40 (c) For m (s) = 10 from t = 0 to t = 20, m (s) = 15 from t = 20 to t = 30,and m (s) = 10 thereafter
2s 2s p (0) = e2s m (s) j20 m (s) j30 m (s) j1 30 0 e 20 e = 10 1 e40 + 15 e40 e60 + 10e60 = 10 + 5e40 5e60 :

3. The solution of the dierential equation describing the behavior of domestic gold holdings is g (t) = (g (0) (" + p )) et + (" + p )

109 with the steady state g1 = " + p and initially g (0) = " + p . The eect of a devaluation is to increase " to "0 . The new solution is g (t) = ((" + p ) ("0 + p )) et + ("0 + p ) = (" "0 ) et + ("0 + p ) with the steady state g1 = "0 + p : 4. The dierential equation is _ H (t) = GF (t) GH (t) = 2GH (t) + G: G

(a) The steady state value of GH is G . The solution to the dierential 2 equation is G 2t G H H e G (t) = G (0) + : 2 2

(b) There is a common increase in both the world gold stock and GH by a common amount, which we call A. The solution is G + A t G + A H H G (t) = 2 G (0) + A e + 2 2 G A t G + A = 2 GH (0) + e + : 2 2 2 If initially in the steady state such that GH (0) = gold discovery G+A A GH (t) = et + : 2 2 5. With m (t) p (t) = 0:8i (t) we have " _ (t) = 1:25" (t) 1:25m (t) : With m (0) = 0 and
dm(t) dt G 2

then after the

= 0:10, we have m (t) = 0:10t and Z1 1:25 (0:10s) e1:25s ds:

" (0) =

s=0

110 Integrate by parts to get 0:125s 1:25s 1 " (0) = j0 e 1:25 Z1 0:125 1:25s ds e 1:25 = 0:08: 6. With the modication of the interest parity relationship, and a money demand equation m (t) p (t) = i (t) and the purchasing power parity relationship " (t) = p (t) (since p = 0) the dierential equation for the exchange rate becomes (with i = 0) 1 m (t) " _ (t) = " (t) + (t ) : (a) With a constant money supply and a constant risk premium, the solution is Z1 s m e " (0) = +e e ds
s=0

s=0

(b) With a constant money supply and a time-varying risk premium, the solution is Z1 s m e + (s) e ds " (0) =
s=0

=m e + e :

(c) With a constant risk premium and a time-varying money supply, the solution is Z1 s m (s) +e e ds " (0) =
s=0

=m e+

s=0

Z1

(s) e ds:

= e +

s=0

Z1

m (s) s e ds:

111

Section 14.3 1. The steady state values are found by setting x _ (t) = 0 and y _ (t) = 0 and solving.

(a) x1 = 4, y1 = 1: (b) x1 = 2, y1 = 4: (c) x1 = 3, y1 = 5:

2. System I: tr = 7 and det = 6 therefore 1 < 0 and 2 < 0 and the system is globally stable. See Fig. 14.3.2(a). 7 p 49 24 = 1; 6: 2

1 ; 2 =

System II: tr = 8 and det = 7 therefore 1 > 0 and 2 > 0 and the system is globally unstable.See Fig. 14.3.2(b). 8 p 64 28 = 1; 7: 2

1 ; 2 =

System III: tr = 2 and det = 3 therefore 1 > 0 and 2 < 0 and the system is saddlepath stable.See Fig. 14.3.2(c). 2 p 4 (12) = 3; 1: 2

1 ; 2 =

112

3. The only stable solution is that the system is at its steady state and therefore x (t) = 2 and y (t) = 4. 4. System III is saddlepath stable. Its initial point is x = 0, y = 8, which is on the saddlepath, since the equation for the saddlepath is y = x + 8. The stable solution to the system, with this initial point, is x (t) = 3et + 3 y (t) = 3et + 5:

5. This system can be expressed as ' 1' " _ (t) " _ (t) + constant terms = p _ (t) p _ (t)

113 (a) The system is saddlepath stable if the determinant of the matrix is negative, that is if 2 ' 1 ' < 0 or, equivalently, if ' + > :

(b) The slope of the saddlepath is 2 ' dp = d" ' where 2 is the negative characteristic root. The saddlepath has a negative slope if > ' and a positive slope if < '. Either of these conditions could be consistent with saddlepath stability. 6. The system can be written in matrix format as u _ (t) p _ (t) = 0 u (t) u m + : u p (t)

(a) The steady state is u1 = u and p1 = m. (b) The trace of the matrix is < 0 and its determinant is > 0 so the two characteristic roots are negative and the system is globally stable. The two roots are p 2 4 : 1 ; 2 = 2 (c) See gure 14.3.6(c) (d) See gure 14.3.6(d). The u _ = 0 line shifts up with the increase in the money supply. During the transition, the unemployment rate falls below its long-run level (which equals its initial level) and prices rise. If the roots were complex then there could be oscillatory behavior.

114

7. The characteristic roots for the second-order linear dierential equation x (t) = ax _ (t) + bx (t) + c are 1 ; 2 = a p a2 4b : 2

(b) The equation is unstable since the two roots are positive, with and 2 = 3 . 1 = 1 2 2 (c) The equation is unstable since the one root is negative and one is positive, with 1 = 1 and 2 = 5 . 2 2

(a) The equation is stable since the two roots are each negative, with 1 = 1 and 2 = 7 . 2 2

115

Chapter 15
Section 15.1 1. The Hamiltonian is t 1 Ht = (F (Kt ; Lt ) wt Lt It ) + t (It Kt ) : 1+r The rst-order conditions are t 1 @Ht = (FL (Kt ; Lt ) wt ) = 0 @Lt 1+r t @Ht 1 = + t = 0 @It 1+r t @Ht 1 = FK (Kt ; Lt ) t = (t t1 ) @Kt 1+r @Ht = (It Kt ) = (Kt+1 Kt ) @t The optimal amount of labor satises the condition FL (Kt ; Lt ) = wt in each period. Making the substitution from the optimality conditions 1 t t = 1+r , the optimal amount of capital satises the condition 1 1+r t FK (Kt ; Lt ) 1 1+r t = 1 1+r t 1 1+r t1 !

or FK (Kt ; Lt ) = + r: 2. With the cost of adjustment, (a) The Hamiltonian is Ht = 1 1+r t It2 F (Kt ; Lt ) wt Lt It + t It : 2Kt

116 (b) The rst-order condition with respect to labor is the same, namely FL (Kt ; Lt ) = wt . (c) The rst-order condition with respect to investment and capital are t @Ht 1 = + t = 0 @It 1+r t @Ht 1 It2 = FK (Kt ; Lt ) + 2 = (t t1 ) @Kt 1+r Kt (d) Dening qt = t (1 + r)t and after some algebra we can show that It2+1 qt (1 + r) = FK (Kt+1 ; Lt+1 ) + 2 + qt+1 : Kt+1 3. In this T period problem, (a) The Hamiltonian is Ht = t U (Ct ) + t (RBt + Yt Ct ) (b) The rst-order conditions are @Ht = t U 0 (Ct ) t = 0 @Ct @Ht = t R = (t t1 ) @Bt Solving, we nd the optimal condition to be (1 + R) U 0 (Ct ) = U 0 (Ct1 ) : (c) This is the same as the condition in Chapter 11 for the two-period problem. 4. With production rather than endowment income we have (a) The Hamiltonian is Ht = t U (Ct ) + t (RBt + F (Kt ) Ct It ) + t It where the two Lagrange multipliers are t and t :

117 (b) The rst-order conditions are @Ht = t U 0 (Ct ) t = 0 @Ct @Ht = t + t = 0 @It @Ht = t R = (t t1 ) @Bt @Ht = t F 0 (Kt ) = t t1 @Kt

Solving, we nd the optimal condition to be the same with respect to consumption as in the previous question, (1 + R) U 0 (Ct ) = U 0 (Ct1 ) while the optimal amount of capital satises the condition F 0 (Kt ) = R: Section 15.2 1. In this problem: (a) The Hamiltonian is y (t)2 + (t) y (t) : H= 2 (b) The maximum principle shows that the optimal conditions include @H = y (t) + (t) = 0 @y @H _ (t) =0= @x @H = y (t) = x _ (t) @ which shows that is constant and therefore y is constant.

118 (c) To solve for y we use the initial and terminal conditions to obtain Z40 Z40 dx dt = x (40) x (0) = 20 dt Z40

t=0

and

dx dt = dt

ydt = 40y = 20

t=0

t=0

so y =

1 . 2

Also, since x _ = y , x (t) = 1 t + 20: 2

2. In this problem: (a) The Hamiltonian is H = 2x (t) y (t) + y (t)2 + (t) y (t) : @H = 2x (t) 2y (t) + (t) = 0 @y @H _ (t) = 2y (t) = @x @H = y (t) = x _ (t) @ Taking the derivative of the Hy condition with respect to time shows that _ (t) = 0 2x _ (t) 2y _ (t) + and substituting in the other two conditions we nd 2y (t) 2y _ (t) + 2y (t) = 0 or y _ (t) = 0, so y is constant. (c) Using the initial and terminal conditions, Z10 dx dt = x (10) x (0) = 90 dt

(b) The maximum principle shows that the optimal conditions include

t=0

119 and since x _ (t) = y , Z10 dx dt = dt Z10 ydt = 10y = 90

t=0

t=0

so y = 9. Also, since x _ = y , x (t) = 9t + 10: 3. In this problem: (a) The Hamiltonian is H = ln (y (t)) 3 (t) y (t) : (b) The maximum principle shows that the optimal conditions include @H 1 = 3 (t) = 0 @y y (t) @H _ (t) = 0 = @x @H = 3y (t) = x _ (t) @ which shows that is constant and, therefore, y is constant since y = 31 . (c) Using the initial and terminal conditions, Z1 dx dt = x (1) x (0) = 30 dt

t=0

and since x _ (t) = 3y , Z1 dx dt = dt Z1 3ydt = 3y = 30

t=0

t=0

so y = 10. Also, since x _ = 3y , x (t) = 30t: 4. In this problem:

120 (a) The Hamiltonian is I (t)2 K (t) H = K (t) K (t) + (t) I (t) : 2 2
2

(b) The solution for this problem uses the optimal conditions @H = I (t) + (t) = 0 @I @H (t) _ (t) = 1 2K (t) = @K 2 @H K (t) _ (t) = I (t) =K @ 2 Using the rst condition, which shows that I (t) = (t) along the optimal path, we can write this as a two-equation dierential equation system in I (t) and K (t). The homogeneous part of this system is 1 _ (t) I 2 I (t) 2 _ (t) = 1 1 K (t) : K 2 The trace of this matrix is 0 and its determinant is 2 1 and 4 therefore its two characteristic roots are 3 and 3 and the system is saddlepath stable. 5. The Hamiltonian in this problem is H = y (t)2 + (t) (x (t) + y (t)) : 2

The solution for this problem uses the optimal conditions @H = y (t) + (t) = 0 @y @H _ (t) = (t) = @x @H = x (t) + y (t) = x _ (t) @
y (t) along the Using the rst condition, which shows that (t) = optimal path, we can write this as a two-equation dierential equation

121 system in y (t) and x (t) x _ (t) x (t) = : y _ (t) 0 y (t) The trace of this matrix is 0 and its determinant is 2 and therefore its two characteristic roots are 2 and the system is saddlepath stable. Section 15.3 1. Let V = B u (c)and note that c (t) = f (k (t)) _ integrand as V c k; k : (a) Note that Vk = Vc Vk _ = Vc
dk(t) . dt

Consider the

@c = u0 (c) f 0 (k ) @k

@c = u0 (c) (1) = u0 (c) _ @k

d Euler's equation shows that the rst-order condition is Vk = dt Vk _ which gives us d u0 (c) f 0 (k ) = u0 (c) dt which can be rewritten to obtain the equation in the text. d 0 (b) In this case, f 0 (k (t)) = k (t)1 , u0 (c (t)) = c(1t) , and dt u (c (t)) = 1 c _(t) d c (t) = c(t)2 . The optimal condition is, therefore, dt

k (t)1 =

c _ (t) c (t)

which is the same as the condition in the text if n = 0, = 0, and = 0. 2. If the Bests receive a wedding gift of $20,000: (a) The rst-order conditions are unaltered but the budget constraint changes because a (0) = 20; 000 and a (50) = 0 and Z50 ert da (t) = a (50) e50r a (0) = 20; 000

t=0

122 and therefore

t=0

Z50

c (t) ert dt =

t=0

Z50

w (t) ert dt + 20; 000:

This shifts up their lifetime consumption prole, but the slope of that prole is unchanged. (b) If the Bests leave a bequest of $20,000 then they will consume more than if they did not receive the wedding present and leave no bequest because

t=0

Z50

c (t) ert dt =

t=0

Z50

w (t) ert dt + 20; 000 20; 000e50r

where 20; 000 20; 000e50r > 0:

3. The entire stock of gold coins can be sold at once so the \extraction time" in this case is an instant. The optimal time for selling the gold coins is the time where the schedule p (T ) e(tT ) is tangent to the price path, p (t), as shown in Figure 15.3.3.

123

124