1 General Equilibrium in the Robinson Crusoe

Economy.
Robinson Crusoe is a fictional shipwrecked person from a novel by Daniel De-
foe.1 In the novel he lived alone on an Island for 28 years, producing everything
he consumed by himself.2 This term has been adopted by economists to describe
a very simple economy, where there is one consumer and one production possi-
bilities frontier (PPF). One can assume that the PPF is created by the various
ways the one consumer can divide his labor and other resources. In this handout
I will assume preferences are Cobb-Douglass U (F, C) = F α C β and that the pro-
duction possibilities frontier has the following simple form: af F γ f + ac C γ c = T ,
where af , ac , T > 0 and γ f , γ c > 1.3 The following graph shows us such an
economy.

10
C
9

0
0 1 2 3 4 5 6 7 8 9 10
F

The dark line is the production possibilities frontier, and the light lines are
various indifference curves. Now obviously maximizing the utility over this set
will be very similar to maximizing it over a budget set, and so the solution
will be at a tangency. The slope of the PPF we call the Marginal Rate of
Transformation, or the MRT. We find both the MRS and the MRT by taking
1 http://en.wikipedia.org/wiki/Robinson_Crusoe
2 He did have the assistance of "Friday," a local man. But his man Friday was a native

American in a novel written by a European, so he didn’t count.

3 In most applications γ = γ = γ.
f c

1
derivatives of an indifference curve and the PPF respectively.

U (F, C) = Ū
∂C
M Uf + M Uc = 0
∂F
∂C MUf
M RS = − =
∂F M Uc
αU
αU βU F Cα
MUf = , MUc = , M RS = βU
=
F C C
Fβ

likewise we can find the MRT:

∂C
af γ f F γ f −1 + ac γ c C γ c −1 = 0
∂F
∂C af γ f F γ f −1
M RT = − =
∂F ac γ c C γ c −1

and the tangency condition is:

αC af γ f F γ f −1
=
βF ac γ c C γ c −1
C γ c αγ c ac = F γ f βγ f af

1
βγ f af γ γ c
C = F f .
αγ c ac

To find out how much clothing this consumer should consume we plug this back
into the PPF:

af F γ f + ac C γ c = T

 1 γ c
βγ f af γ γ c
af F γ f + ac F f = T
αγ c ac
βγ f af γ
af F γ f + ac F f = T
αγ c ac


βγ f
1+ af F γ f = T
αγ c

 γ1
αγ c T f
F =
αγ c + βγ f af

you should recognize the close similarity between this solution and the standard
β I
Cobb-Douglass demands, In the standard demand F = α+β pf here αγ c replaces
β, βγ f replaces α and the entire thing is raised to a power. Overall it is not
really that different. I will let you solve for the general value for C on your
own.

2
 The important thing about this equilibrium–to an economist–is that Robin-
son’s decision process can be separated into Utility maximizing and Revenue
maximizing. The revenue maximizer will solve the following problem:

R = max pf F + pc C − λ (af F γ f + ac C γ c − T )
F,C

the utility maximizer will solve the problem

max F α C β − λ (pf F + pc C − I)
F,C

and the model will be closed by having I = R, or the revenue the firm makes.
I should note that in order to solve such a problem we really need γ f = γ c , we
also need this if the utility function is not a Cobb-Douglass. In general we will
assume this.
To give an example assume that U (F, C) = F C 4 and F 2 + C 2 = 125. Then
the Pareto efficient outcome is:
 
max min F C 4 − λ C 2 + F 2 − 125
F,C λ

C 4 − 2F λ = 0
4C 3 F − 2Cλ = 0
C 2 + F 2 − 125 = 0

1 C4
λ = = 2C 2 F
2 F
1
F = C
2

2
1
C2 + C = 125
2
5 2
C = 125
4
C = 10
1
F = (10) = 5
2
and we can decentralize this by finding the appropriate prices.

MUf C4 1C 1 pf
M RS = = = = =
M Uc 4F C 3 4F 2 pc

let us just let pc = 1 and pf = 12 for the fun of it. Then we can solve the
revenue maximization problem first:
1  
max min F + C − λ C 2 + F 2 − 125
F,C λ 2

3
 1
− λ2F = 0
2
1 − λ2C = 0
 2 2

C + F − 125 = 0

1 1
λ = =
4F 2C
1
F = C
2
and at this point from above you can easily see that the solution is C = 10 and
F = 5. This means that Robinson’s Income is I = 12 (5) + 10 = 25 2 , and his
utility maximization problem is:


4 25
max min F C − λ pF + C −
F,C λ 2

C 4 − λp = 0
4F C 3 − λ = 0
1 25
F +C − = 0
2 2

λ = 2C 4 = 4F C 3
1
F = C

 2
1 1 25
C +C− = 0
2 2 2
C = 10

and as you can see the solution is the same.

1.1 Robinson Crusoe and International Trade.

Now what would happen if Robinson Crusoe could trade at a different price
vector from p = 12 as above? What would happen if he could trade at p = 2?
Obviously now he would choose a different bundle to produce (as a revenue
maximizer) and consume (as a utility maximizer.) In the example above

4
(U (F, C) = F C 4 and F 2 + C 2 = 125) this would result in the following graph:

25
y
20

15

10

0
0 1 2 3 4 5 6 7 8 9 10 11 12
x

The revenue
 maximizing
 point is (F, C) = (10, 5) the utility maximizing point is
(F, C) = 52 , 20 so obviously this will not work. There is no way that Robinson
can get the extra 15 units of C that he needs and nothing to do with the 15 2
units of leftover food. Or is there? What if he could trade with other islands?
This is what happens in a model of international trade.
Now say that Robinson has just discovered that the next island over has
a very productive economy run by highly civilized Native Americans that do
not run around half naked killing every European they come across.4 Should
Robinson open up trade? Should the Native Americans? After all they can
produce everything Robinson can, and do it more efficiently. The answer is yes,
and there is a very simple way to figure out if both economies will benefit from
trade. Are the (implicit) price vectors different in the two economies? (Up to
a scaling factor of course, we still can only identify m − 1 prices in a General
Equilibrium model.) If yes, then they will strictly benefit from trade. Pretty
amazing isn’t it?
What could cause these different price vectors? Well many different things,
these are a few in the order of relative importance.

1. Difference in Resource Endowments

2. Differences in Technology
3. Differences in Demand
4. Economies of Scale
4 Though, come to think of it, it probably would have been better for all Native Americans

if they had been a little more... blood thirsty... when meeting the first Europeans.

5
 I should point out that the last one could mean that the PPF was not con-
cave, if this happens it could cause problems–but it generally doesn’t. Notice
as well that the terms of trade will change, almost surely input prices will shift
when trade is opened up. This means that some people will gain and some will
loose. When I say that they should open up to trade I mean that the winners
could pay off the losers. Unfortunately in general this does not happen–this
is a serious problem but not one I want to discuss here.

1.2 Robinson Crusoe and the Dead Weight Loss from Taxes.
Now I want to explore the idea of dead weight loss caused by taxes. To be
precise what we are going to do is look at a Robinson Crusoe economy where
first the government taxes each unit of labor a fixed amount t and then returns
the funds to him as a transfer τ = tL∗ , where L∗ is the optimal amount of labor
Robinson supplies. Since we are trying to mimic an economy where there are
millions of laborers we will assume that Robinson does not think about the fact
that τ will change if L∗ does, or we will assume Robinson acts as if he is a
competitive worker. Notice as well that this government is a waste of time. If
they had any brains they wouldn’t collect a tax just to give it back, in general
there are much better things to do with the money–we will discuss this later
in the semester.
But to continue with
 this problem,√ assume the production function for food
and clothing is F = Lf and C = Lc and Robinson decides how much labor
to supply by maximizing his utility function: U (F, C, L) = F C (24 − L) subject
to the budget constraint pf F + pc C ≤ wL.

max F C (24 − L) − λ (pf F + pc C − wL)

U
− λpf = 0
F
U
− λpc = 0
C
U
− + λw = 0
24 − L
pf F − pc C − wL = 0

U
F λpf
U
=
C
λpc
pc
F = C
pf


pc
pf C + pc C − wL = 0
pf

6
 U
24−L λw
U
=
C
λpc
C
L = − pc + 24
w



pc C
pf C + pc C − w − pc + 24 = 0
pf w
w
C = 8
pc
pc w pc w
F = C =8 =8
pf pc pf pf

C 8 pwc
L = − pc + 24 = − pc + 24 = 16
w w
Notice the rather shocking fact that the quantity of Labor supplied is indepen-
dent of all prices. To understand this we should think about him consuming
leisure instead of selling Labor. Leisure is the time this guy uses to do anything
besides work (watch TV, sleep, have dinner, etceteras) so l = 24 − L. If we
substitute for L in the budget constraint we get:

pf F + pc C ≤ w (24 − l)
pf F + pc C + wl ≤ w24

so you can see that his total potential income is 24w and the price of a unit of
leisure is also w, since in Cobb-Douglass demands the general form is ΣNαmαn pIm
m=1
and in this case pm = w, I = 24w we can see why the labor supply is constant.
Now we want to solve for the supply of food and clothing, this is easiest to do
1
if we write the profit function in terms of F , in order to do this since F = Lf2
we can see that Lf = F 2

max pf F − wF 2
Lf
pf − 2wF = 0
1
Fs = pf
2w
1 w
Fs = pf = 8 = Fd
2w pf
1
pf = 4w, Fs = Fd = (4w) = 2
2w
and obviously by symmetry pc = 4w, and at this point let us normalize the
wage rate to one. So our equilibrium isF = C = 2, L = 16 and pf = pc = 4 .
Now with the tax our new problem is:

max F C (24 − L) − λ (pf F + pc C − L − τ)

7
obviously the only difference is that the income of the consumer has increased
by τ , so the solution will still be
pc
F =C , L = −Cpc + 24
pf

when we put this into the budget constraint:

pf F + pc C − L − τ = 0


pc
pf C + pc C − (−Cpc + 24) − τ = 0
pf
8 1 τ
C = +
pc 3 pc
8 1 τ
F = +
pf 3 pc


8 1 τ
L = − + pc + 24
pc 3 pc
1
= − τ − 8 + 24
3
1
= 16 − τ
3
but now let’s plug back in the fact that τ = tL, and we get:
1
L = 16 − tL
3
48 (48) 13 1
L = = 1 = 16 t
t+3 (t + 3) 3 1+ 3

and obviously I like this solution because if t = 0, 1+1 t = 1. Plugging this in

3
all over the place we get:

   

8 1 τ 1 1 1 16t + 24 1 + 3t 8 1+t
C = + = t16 + 24 = =
pc 3 pc 3pc 1 + 3t 3pc 1 + 3t pc 1 + 3t


8 1+t
F =
pf 1 + 3t

and to solve the other side of the problem:

max pf F − (1 + t) F 2
Lf

pf − 2 (1 + t) F = 0
pf
Fs =
2 (1 + t)

8


pf 8 1+t
Fs = = = Fd
2 (1 + t) pf 1 + 3t
2 1
p2f = 16 (1 + t)
1 + 3t

1
pf = 4 (1 + t)
1 + 3t

pf 1 1 1
F = = 4 (1 + t) t =2 t
2 (1 + t) 2 (1 + t) 1+ 3 1+ 3

Not elegant, I agree, but heh. The final solution is F = C = 2 1+1 t , L =
 3

16 1+1 t and pf = pc = 4 (1 + t) 1+1 t . So how do we compare the two

3 3
situations? If we wanted to be precise we should use compensating variation
(how much more income would we have to give the consumer to make him
indifferent) but I am not going to require that this year. An easy rule of thumb
is to just look at the two utilities. The utility after the tax is:

U (F (t) , C (t) , L (t)) = F C (24 − L)





1 1 1
= 2 2 24 − 16
1 + 3t 1 + 3t 1 + 3t

   
4 24 1 + 3t − 16
=
1 + 3t 1 + 3t



4 8t + 8
=
1 + 3t 1 + 3t
t+1
= 32  2
1 + 3t

9
The case without a tax is t = 0, and utility is equal to 32 at this level. Now
theroetically this should be monotonically decreasing in t, but it is not, it only
decreases after t ≥ 1, the maximum is actually at t = 1 as you can see in the
graph.

36
U
34

32

30

28

26

24

22

20

0 1 2 3 4 5 6 7 8 9 10
t

I find this rather peculiar, and don’t really understand what is going on. The
best I can understand is that increasing the tax in this general equilibrium model
has two effects. First there is the increase in all prices, second is the increase

10
in income. Apparently in this model the second effect outweighs the first for
low levels of taxes.

11