Chapter 4: Non-renewable Resources:

Optimal Extraction

Instructor: Aschalew Sh.

Department of Economics
PhD Candidate

1
 Non renewable resources
• Non-renewable resources include fossil-fuel energy supplies – oil, gas and coal –
and minerals – copper and nickel, for example.
• They are formed by geological processes over millions of years and so, in effect,
exist as fixed stocks which, once extracted, cannot be renewed.
• One question is of central importance: what is the optimal extraction path over
time for any particular non-renewable resource stock?

2
 Theory of optimal depletion of non renewable
resources
• Assumption:
– utility comes directly from consuming the extracted resource.
– it is assumed that there exists a known, finite stock of each kind of non-renewable
resource.
– No adverse external effects arising from the extraction or consumption of the resource.
• Not surprisingly, the optimal extraction path will be different if adverse
externalities are present causing environmental damage.

3
 A non-renewable resource two-period model
• The planning horizon that consists of two periods, period 0 and period 1.
• There is a fixed stock of known size of one type of a non-renewable resource.
The initial stock of the resource (at the start of period 0) is denoted S.
• Rt is the quantity extracted in period t
• Assume that an inverse demand function exists for this resource at each time,
given by
Pt  a  bR t
– where Pt is the price in period t, with a and b being positive constant
numbers. So, the demand functions for the two periods will be:
P0  a  bR 0
P1  a  bR 1 4
 The shaded area (the integral of P with respect to R over the interval R = 0 to R = Rt)
shows the total benefit consumers obtain from consuming the quantity Rt in period t.
From a social point of view, this area represents the gross social benefit, B, derived from
P
the extraction and consumption of quantity Rt of the resource.
a
Rt

B Rt  
 a  bR dR
0

b 2
 aRt  Rt
2

a - bR

0
Rt a/b R

Figure 15.1 The non-renewable resource demand function for the two-period model
5
A linear and negatively sloped demand function such as this one has the property that
demand goes to zero at some price, in this case the price a.

Hence, either this resource is non-essential or it possesses a substitute which at the

price a becomes economically more attractive.

The assumption of linearity of demand is arbitrary and so you should bear in mind that
the particular results derived here are conditional upon the assumption that the demand
curve is of this form.

6
 Gross and net benefit
• The gross benefit obtained by consumers is not identical to the net social benefit
of the resource, as resource extraction involves costs.
• In this chapter, we assume that these costs are fully borne by the resource-
extracting firms, and so private and social costs are identical.
• We also assume that benefits represented in the resource demand function are
the only benefits to society, so there are no beneficial externalities.

7
 Cont...
• Assume that there are constant marginal costs of extraction, c with c ≥ 0.
• Then total extraction costs, Ct, for the extracted quantity Rt units will be Ct = cRt
• The total net social benefit from extracting the quantity Rt is NSBt = Bt – Ct
where NSB denotes the total net social benefit and B is the gross social benefit
of resource extraction and use. Hence
Rt
b 2
NSB Rt  
 a  bR dR  cR
0
t
 aRt 
2
Rt  cRt (15.1)

8
 A socially optimal extraction policy

• We develop a socially optimal extraction program.

– This will serve as a benchmark in terms of which any particular extraction
program can be assessed.
• In order to find the socially optimal extraction program, two things are required.
1. A social welfare function that embodies society’s objectives.
2. A statement of the technical possibilities and constraints available at any
point in time.
• We deal first with the social welfare function, using a discounted utilitarian in
form.

9
10
11
12
 Interpretation
• This is the Hotelling rule of an inter-temporal efficiency condition which must be
satisfied by any efficient process of resource extraction.
• What does this result tell us? The left-hand side of equation 15.5, , is the social utility
discount rate, which embodies some view about how future utility should be valued in
terms of present utility. The right-hand side is the proportionate rate of growth of the
resource’s net price.
• So if, for example, society chooses a discount rate of 0.1 (or 10%), Hotelling’s rule
states that an efficient extraction program requires the net price of the resource to grow
at a proportionate rate of 0.1 (or 10%) over time.

13
 Optimal Resource Extraction - Discrete Time

U (Ct )

Social welfare function: maxW   t
Ct ,Rt
t 0 (1   )
t 1
Extraction: St  St 1  Rt 1  S0   R
 0
St  Rt 1
Investment: Kt  Kt 1  Qt 1  Ct 1
Kt  Qt 1  Ct 1

Production function: Qt  Q (Kt , Rt )

14
 Cont’d

U (Ct )

L t
t 0 (1   )
 
  t St  St 1  Rt 1    t Kt  Kt 1  Qt 1  Ct 1 
t 0 t 0

L L L L L L
 0;  0;...;  0;...  0;  0;...;  0;...
C 0 C 1 Ct 0 1 t
L L L L L L
 0;  0;...;  0;...  0;  0;...;  0;...
R0 R1 Rt  0 1 t

15
 Optimal Resource Extraction - Continuous Time

Objective function: max J   L ( x , u ,t )e  t

dt
u
0
x
Subject to: : x  f (x , u ,t ); x (t0 )  x 0
t
• J depends on control variables (u), state variables (x), and time
• State variables describe the economy at any time; the equation of motion governs its
evolution over time
• Control variables are time-dependent policy instruments
• To obtain the solution we construct a current value Hamiltonian

16
 The Hamiltonian
• The Hamiltonian only contains the current state and controls; current optimality is a
necessary condition for intertemporal optimality

H (x , u ,t ,  )  L(x , u ,t )  f (x , u ,t )
H  H
FOC:  0 and    
u x
• The co-state variables (λ) secure intertemporal optimality; they are like Lagrange
multipliers, indeed measure the shadow price

17
Cont’d


Social welfare function:
maxW   U (Ct )e  t dt
Ct ,Rt
t0

Equations of motion: St  Rt and Kt  Qt  Ct

Hamiltonian: Ht  U (Ct )  Pt ( Rt )  t (Q (Kt , Rt )  Ct )

Necessary conditions:
Ht Ht
 UCt  t  0  t  t   t  tQKt
Ct Kt
Ht  Ht
 Pt  tQRt  0 Pt   Pt    Pt
Rt St 18
 Static Efficiency Conditions

• The static efficiency conditions: As with any asset, static efficiency requires
that, in each use to which a resource is put, the marginal value of the services
from it should be equal to the marginal value of that resource stock in situ.
• This ensures that the marginal net benefit (or marginal value) to society of the
resource should be the same in all its possible uses.

19
 Cont’d

Ht
 UCt  t  0  UCt  t
Ct

• In each period, the marginal utility of consumption UC,t must be equal to the

shadow price of capital ωt .

• An efficient outcome will be the same in which the marginal net benefit of using
one unit of output for consumption is equal to its marginal net benefit when it is
added to the capital stock.

20
 Cont’d

Ht
 Pt  tQRt  0  Pt  tQRt
Rt
•The value of the marginal product of the natural resource must be equal to the
marginal value (or shadow price) of the natural resource stock, Pt.

•The value of the marginal product of the resource is the marginal product in units of
output (QR,t) multiplied by the value of one unit of output(ωt).

21
 Dynamic Efficiency Conditions
• Dynamic efficiency requires that each asset or resource earns the same rate of return,
and that this rate of return is the same at all points in time, being equal to the social
rate of discount.

H P
Pt   Pt  t
 Pt   Pt   
St P
• The growth rate of the shadow price of the resource equals the discount rate

Ht 
 t  t   t  tQKt   QKt  
Kt 
• The return to capital equals the discount rate
22
Hotelling‘s Rule
• Dynamic efficiency required: P
 
P
The growth rate of the shadow price equals the discount rate
• An alternative interpretation:

Pt   Pt  Pt  P0e t  Pt e  t  Pt *
• Hotelling’s rule states that the discounted value of the resource should be the
same at all dates. But this is merely a special case of a general asset–efficiency
condition; the discounted (or present value) price of any efficiently managed
asset will remain constant over time.

23
 Cont’d
• This way of interpreting Hotelling’s rule shows that there is nothing special
about natural resources per se when it comes to thinking about efficiency.
• A natural resource is an asset. All efficiently managed assets will satisfy the
condition that their discounted prices should be equal at all points in time.

24
 A non-renewable resource continuous-time multi-period
model
• We now change to a continuous-time framework which deals with rates of extraction
and use at particular points in time over some continuous-time horizon.
• To keep the maths as simple as possible, we will now define P as the net price of the
non-renewable resource, that is, the price after deduction of the cost of extraction.
• Let P(R) denote the inverse demand function for the resource, indicating that the
resource net price is a function of the quantity extracted, R.

25
 Cont’d
• The social utility from consuming a quantity R of the resource may be
defined as
R

U R  
 PR  dR
0
(15.6a)
• By differentiating total utility with respect to R, the rate of resource
extraction and use, we obtain
U
 P R 
R (15.6b)
which states that the marginal social utility of resource use equals the net
price of the resource.

26
 P

U(R) = shaded area

Ke-aR

0 Quantity of resource extracted, R

R
Figure 15.2 A resource demand curve, and the total utility from consuming a particular quantity of the resource
27
 Cont’d
• Assume that the inter-temporal social welfare function is utilitarian, with social utility discount
rate . Then the value of social welfare over an interval of time from period 0 to period T can be
expressed as
T

W 

0
U Rt  e t dt

• Our problem is to make social-welfare-maximizing choices of

1. Rt, for t = 0 to t = T (that is, we wish to choose a quantity of resource to be extracted in each period), and
2. the optimal value for T (the point in time at which depletion of the resource stock ceases), subject to the constraint
that T

 R dt  S
0
t

• That is, the total extraction of the resource is equal to the size of the initial resource stock. Note that in
this problem, the time horizon to exhaustion is being treated as an endogenous variable to be chosen by
the decision maker.
28
29
 Specific solution
• To obtain specific solutions, we need a particular form of the resource demand
function. We suppose that the resource demand function is

P(R) = Ke–aR (15.8)

which is illustrated in Figure 15.2.

• This function exhibits a non-linear relationship between P and R, and is probably more
representative of the form that resource demands are likely to take than the linear function
used in the section on the two-period model.

30
 Cont’d
• However, it is similar to the previous demand function in so far as it exhibits zero
demand at some finite price level. To see this, just note that P(R = 0) = K.
• K is the so-called choke price for this resource, meaning that the demand for the
resource is driven to zero or is ‘choked off’ at this price.
• Whether or not you have succeeded in obtaining a formal solution to this optimisation
problem, intuition should suggest one condition that must be satisfied if W is to be
maximised.

31
 Cont’d
• Rt must be chosen so that the discounted marginal utility is equal at each point
in time, that is,
U  ρt
e  constant
R
• To understand this, let us use the method of contradiction.
– If the discounted marginal utilities from resource extraction were not equal in every
period, then total welfare W could be increased by shifting some extraction from a
period with a relatively low discounted marginal utility to a period with a relatively
high discounted marginal utility.
– Rearranging the path of extraction in this way would raise welfare.

32
 Cont’d
• It must, therefore, be the case that welfare can only be maximized when discounted
marginal utilities are equal. What follows from this result? First note equation 15.6b
again:
U t
 Pt
Rt
• So, the requirement that the discounted marginal utility be constant is equivalent to
the requirement that the discounted net price is constant as well – this result is noted
in Chapter 14. (see Perman 3rd edition page 485).

33
 Cont’d
• That is,

34
End of the chapter

35