UGBA 101A Microeconomic Analysis for Business

Decisions
Section 3: Uncertainty, Producer Theory, Cost Functions

Tianyu Han

Haas School of Business, UC Berkeley

June 10, 2021

Tianyu Han (Berkeley Haas) UGBA 101A (Summer 2021) Section 3 June 10, 2021 1 / 60
Outline
1 Choice under Uncertainty
Setup
Maximax, Maximin, and Minimax
Expected Money Value
Expected Utility Hypothesis
2 Production
Technology
Short-Run Production
Long-Run Production
Returns to Scale
3 Costs
Different Types of Costs
Cost in the Short-Run
Cost in the Long-Run
Relationship between Short-Run and Long-Run
Efficiencies in the Long-Run

Tianyu Han (Berkeley Haas) UGBA 101A (Summer 2021) Section 3 June 10, 2021 2 / 60
Outline
1 Choice under Uncertainty
Setup
Maximax, Maximin, and Minimax
Expected Money Value
Expected Utility Hypothesis
2 Production
Technology
Short-Run Production
Long-Run Production
Returns to Scale
3 Costs
Different Types of Costs
Cost in the Short-Run
Cost in the Long-Run
Relationship between Short-Run and Long-Run
Efficiencies in the Long-Run

Tianyu Han (Berkeley Haas) UGBA 101A (Summer 2021) Section 3 June 10, 2021 3 / 60
Uncertainty: Setup
So far, we have assumed that prices, incomes, and other variables are
known with certainty.
However, many of the economic choices involve uncertainty (or risk).
In this class, we use a discrete random variable to define a risky object.
Specifically, for a random variable X with n possible realizations
(outcomes) x1 , x2 , . . . , xn , it has a probability distribution given by
X Pr(X = x)
x1 p1
x2 p2
.. ..
. .
xn pn
where p1 + p2 + · · · + pn = 1.
Mean: E [X ] = p1 x1 + p2 x2 + · · · + pn xn
Variance: Var (X ) ≡ 𝜎X2 = p1 (x1 − E [X ])2 + · · · + pn (xn − E [X ])2
q
Standard deviation: 𝜎X = 𝜎X2
Tianyu Han (Berkeley Haas) UGBA 101A (Summer 2021) Section 3 June 10, 2021 4 / 60
Exercise (Homework 2-B Q1-2)
Mobon Oil Company must now decide whether to drill for oil at the site or to
sell the lease to Exxil Oil Company, which has offered Mobon $50,000.
Mobon estimates that it would cost $100,000 to drill at the site.
If the well were dry, all this cost would be lost. If the well were successful,
its value to Mobon would depend on the extent of the oil discovered: a
minor success or a major success. In excess of the drilling cost,
I A minor success would result in revenues of $200,000
I A major success would result in revenues of $600,000
Mobon has assessed the following probabilities:

Type of Well Dry Minor Success Major Success

Probability 0.7 0.2 0.1

(a) Construct an appropriate payoff matrix. Back

Dry ( ) Minor ( ) Major ( )

Drill
Sell

Tianyu Han (Berkeley Haas) UGBA 101A (Summer 2021) Section 3 June 10, 2021 5 / 60
Choice Rules

There is no uniform choice rule – it depends on the tasks and goals in the
setting
In this class, we consider five choice rules:
1 Maximax: choose decision having best possible outcome
2 Maximin: choose the decision having the best of the worst possible
outcomes
3 Minimax regret: choose the decision having the least potential regret
4 Expected money value: choose the decision having the best mean value
5 Expected utility value: choose decision with the best mean utility value

Tianyu Han (Berkeley Haas) UGBA 101A (Summer 2021) Section 3 June 10, 2021 6 / 60
Outline
1 Choice under Uncertainty
Setup
Maximax, Maximin, and Minimax
Expected Money Value
Expected Utility Hypothesis
2 Production
Technology
Short-Run Production
Long-Run Production
Returns to Scale
3 Costs
Different Types of Costs
Cost in the Short-Run
Cost in the Long-Run
Relationship between Short-Run and Long-Run
Efficiencies in the Long-Run

Tianyu Han (Berkeley Haas) UGBA 101A (Summer 2021) Section 3 June 10, 2021 7 / 60
Maximax

Maximax: choose decision having best possible outcome

Steps:
1 Find the best outcome value for each action
2 Take the action with the largest best outcome value
(b) Using the Maximax criterion, identify the optimal decision.

Dry (.70) Minor (.20) Major (.10)

Drill -100 200 600
Sell 50 50 50

Tianyu Han (Berkeley Haas) UGBA 101A (Summer 2021) Section 3 June 10, 2021 8 / 60
Maximin

Maximin: choose the decision having the best of the worst possible
outcomes
Steps:
1 Find the worst outcome value for each action
2 Take the action with the largest worst outcome value
(c) Using the Maximin criterion, identify the optimal decision.

Dry (.70) Minor (.20) Major (.10)

Drill -100 200 600
Sell 50 50 50

Tianyu Han (Berkeley Haas) UGBA 101A (Summer 2021) Section 3 June 10, 2021 9 / 60
Minimax Regret

Minimax Regret: choose the decision with the least worst regret
Steps:
1 Configure a regret table:
F Find the regret of each action under each state (each cell)
F Regret is defined as the difference between what the best outcome would have
been and what the decision maker received given their decision
2 and find the maximum regret of the each action
3 Take the action with the minimum max regret
(d) Using the Minimax Regret criterion, identify the optimal decision.

Dry (.70) Minor (.20) Major (.10)

Drill -100 200 600
Sell 50 50 50

Tianyu Han (Berkeley Haas) UGBA 101A (Summer 2021) Section 3 June 10, 2021 10 / 60
Outline
1 Choice under Uncertainty
Setup
Maximax, Maximin, and Minimax
Expected Money Value
Expected Utility Hypothesis
2 Production
Technology
Short-Run Production
Long-Run Production
Returns to Scale
3 Costs
Different Types of Costs
Cost in the Short-Run
Cost in the Long-Run
Relationship between Short-Run and Long-Run
Efficiencies in the Long-Run

Tianyu Han (Berkeley Haas) UGBA 101A (Summer 2021) Section 3 June 10, 2021 11 / 60
Expected Money Value (EMV)
Expected money value (EMV): choose the decision having the best mean
value
Steps to construct a decision tree:
1 Draw the decision node ()
2 For each action from the decision node, draw the state node (◦)
3 For each state from the state node, compute the EMV of each action
4 Prune the actions with lower EMVs
(e) Using the Expected Money Value (EMV) criterion, identify the optimal
decision.
(g) Construct a decision tree for this problem, prune the decision tree
using the EMV criterion, and verify that you have reached the same
optimal decision as in part (e).

Tianyu Han (Berkeley Haas) UGBA 101A (Summer 2021) Section 3 June 10, 2021 12 / 60
Expected Money Value (EMV)
Expected money value (EMV): choose the decision having the best mean
value
Steps to construct a decision tree:
1 Draw the decision node ()
2 For each action from the decision node, draw the state node (◦)
3 For each state from the state node, compute the EMV of each action
4 Prune the actions with lower EMVs
(e) Using the Expected Money Value (EMV) criterion, identify the optimal
decision.
(g) Construct a decision tree for this problem, prune the decision tree
using the EMV criterion, and verify that you have reached the same
optimal decision as in part (e).
(f) Using the Expected Regret criterion, identify the optimal decision.

Dry (.70) Minor (.20) Major (.10)

Drill -100 200 600
Sell 50 50 50

Tianyu Han (Berkeley Haas) UGBA 101A (Summer 2021) Section 3 June 10, 2021 12 / 60
Value of Information

Value of information: the difference between the expected value of a

choice when additional information is present and absent
Value of Perfect Information (EVPI): Suppose the decision maker
knows the random state ex ante (perfect information), what is the value
added to the EMV?
Steps to construct a decision tree:
1 Draw the state node (◦) first, because with perfect information, we will learn
the state ex-ante
2 For each state from the state node, draw the decision node ()
3 For each action from the decision node, pick the decision with the highest
payoff
4 Compute the EMV of the best decision for each state

Tianyu Han (Berkeley Haas) UGBA 101A (Summer 2021) Section 3 June 10, 2021 13 / 60
Value of Information (cont.)
(h) Using the Regret table from part (f) and a decision tree analysis in
part (e), compute the Expected Value of Perfect Information EVPI for this
problem and verify you obtain the same result.

Dry (.70) Minor (.20) Major (.10)

Drill -100 200 600
Sell 50 50 50

Tianyu Han (Berkeley Haas) UGBA 101A (Summer 2021) Section 3 June 10, 2021 14 / 60
Value of Information (cont.)

Value of Imperfect Information (EVII): The expected value of imperfect

information is the value added to the EMV by making the best decision(s)
using the information over the EMV achieved by making the best decision
without any information.
Steps to construct a decision tree:
1 Add an action of acquiring the information to the decision node and
complete the structure of the information’s “sub-tree”
2 Use the Bayes rule to find the probability associated with each state in the
state notes
3 Compute the EMV of each action; eliminate the actions with lower EMVs

Tianyu Han (Berkeley Haas) UGBA 101A (Summer 2021) Section 3 June 10, 2021 15 / 60
Value of Information (cont.)
4. Now assume that for a fee of $25,000, Mobon can hire a consulting firm to
conduct a test known as a seismic survey. Using seismic soundings, the
survey can determine whether a site’s underlying rock formations are bowed
up into a “dome.” If a dome structure exists, the chances of finding oil are
substantially better than if no dome structure exists. The consulting firm has
furnished Mobon with data on how well its seismic surveys have performed in
the past. Based on this data, Mobon has assessed the following probabilities:
If the well is dry, the seismic survey will indicate “dome” structure with
probability 0.2.
If the well is a minor success, the seismic survey will indicate a “dome”
structure with probability 0.6.
If the well is a major success, the seismic survey will indicate a “dome”
structure with probability 0.9.

Tianyu Han (Berkeley Haas) UGBA 101A (Summer 2021) Section 3 June 10, 2021 16 / 60
Value of Information (cont.)

Tianyu Han (Berkeley Haas) UGBA 101A (Summer 2021) Section 3 June 10, 2021 17 / 60
Value of Information (cont.)
(a) By constructing and pruning an appropriate (multistage) decision tree,
determine Mobon’s optimal strategy (including whether or not Mobon
should purchase the seismic survey) and the net EMV.

Tianyu Han (Berkeley Haas) UGBA 101A (Summer 2021) Section 3 June 10, 2021 17 / 60
Value of Information (cont.)
(a) By constructing and pruning an appropriate (multistage) decision tree,
determine Mobon’s optimal strategy (including whether or not Mobon
should purchase the seismic survey) and the net EMV.
(b) Regardless of whether Mobon should purchase the survey for
$25,000, what is the maximum Mobon should pay the consulting firm for
its seismic survey?

Tianyu Han (Berkeley Haas) UGBA 101A (Summer 2021) Section 3 June 10, 2021 17 / 60
Sensitivity Analysis

Sensitivity analysis: investigate the robustness of an assumption in the

decision problem; determine over what range will maintain the current
optimal decision and what decisions become optimal beyond that range.
In this class, we focus on two problems:
1 Sensitivity to a change in a payoff: allow a single payoff to change, holding
everything else fixed
2 Sensitivity to a change in a probability of a state: allow a probability (and
other probabilities that are involved) to change, holding everything else fixed

Tianyu Han (Berkeley Haas) UGBA 101A (Summer 2021) Section 3 June 10, 2021 18 / 60
Sensitivity Analysis (cont.)
2. Reconsider Mobon’s decision problem in Q1 above. Question

Dry ( ) Minor ( ) Major ( )

Drill
Sell

(a) Assuming all other data remain unchanged, perform a sensitivity

analysis on the amount Exxil offers to buy the lease.

Tianyu Han (Berkeley Haas) UGBA 101A (Summer 2021) Section 3 June 10, 2021 19 / 60
Sensitivity Analysis (cont.)
2. Reconsider Mobon’s decision problem in Q1 above. Question

Dry ( ) Minor ( ) Major ( )

Drill
Sell

(a) Assuming all other data remain unchanged, perform a sensitivity

analysis on the amount Exxil offers to buy the lease.
(b) Assuming all other data remain unchanged, perform a sensitivity
analysis on Mobon’s estimate of the cost to drill for oil at the site. Use a
graphical analysis within your answer.

Tianyu Han (Berkeley Haas) UGBA 101A (Summer 2021) Section 3 June 10, 2021 19 / 60
Sensitivity Analysis (cont.)
2. Reconsider Mobon’s decision problem in Q1 above. Question

Dry ( ) Minor ( ) Major ( )

Drill
Sell

(a) Assuming all other data remain unchanged, perform a sensitivity

analysis on the amount Exxil offers to buy the lease.
(b) Assuming all other data remain unchanged, perform a sensitivity
analysis on Mobon’s estimate of the cost to drill for oil at the site. Use a
graphical analysis within your answer.
(c) Assuming all other data remain unchanged, perform a sensitivity
analysis on Mobon’s estimate of the probability of a dry well (currently
0.70). Use a graphical analysis within your answer.

Tianyu Han (Berkeley Haas) UGBA 101A (Summer 2021) Section 3 June 10, 2021 19 / 60
Outline
1 Choice under Uncertainty
Setup
Maximax, Maximin, and Minimax
Expected Money Value
Expected Utility Hypothesis
2 Production
Technology
Short-Run Production
Long-Run Production
Returns to Scale
3 Costs
Different Types of Costs
Cost in the Short-Run
Cost in the Long-Run
Relationship between Short-Run and Long-Run
Efficiencies in the Long-Run

Tianyu Han (Berkeley Haas) UGBA 101A (Summer 2021) Section 3 June 10, 2021 20 / 60
Expected Utility Value (EUV)
Similar to the utility representation we developed for choices under
certainty, we can have something similar – the expected utility
representation.
This is also known as the expected utility hypothesis: an agent chooses
between risky prospects by comparing expected utility values.

Tianyu Han (Berkeley Haas) UGBA 101A (Summer 2021) Section 3 June 10, 2021 21 / 60
Expected Utility Value (EUV)
Similar to the utility representation we developed for choices under
certainty, we can have something similar – the expected utility
representation.
This is also known as the expected utility hypothesis: an agent chooses
between risky prospects by comparing expected utility values.
Specifically, for a uncertain choice set X with n possible realizations
(outcomes) x1 , x2 , . . . , xn , it has a probability distribution given by
X Pr(X = x)
x1 p1
x2 p2
.. ..
. .
xn pn
where p1 + p2 + · · · + pn = 1.
I u(xi ) denotes the utility of a realization (outcome) xi
I U(X ) denotes the von Neumann Morgenstern (vNM) expected utility of the
uncertain choice set X , where
U(X ) = p1 u(x1 ) + p2 u(x2 ) + · · · + pn u(xn ).

Tianyu Han (Berkeley Haas) UGBA 101A (Summer 2021) Section 3 June 10, 2021 21 / 60
Different Preferences toward Risk

People’s preferences over risk are also heterogeneous.

Specifically,
I Risk averse (U(X ) < u(E [X ])): one prefers a certain income to a risky
income with the same expected value
F Risk premium: the maximum amount of money that a risk-averse person will pay
to avoid taking a risk

Tianyu Han (Berkeley Haas) UGBA 101A (Summer 2021) Section 3 June 10, 2021 22 / 60
Different Preferences toward Risk

People’s preferences over risk are also heterogeneous.

Specifically,
I Risk averse (U(X ) < u(E [X ])): one prefers a certain income to a risky
income with the same expected value
F Risk premium: the maximum amount of money that a risk-averse person will pay
to avoid taking a risk
I Risk neutral (U(X ) = u(E [X ])): one is indifferent between a certain income
and a risky income with the same expected value

Tianyu Han (Berkeley Haas) UGBA 101A (Summer 2021) Section 3 June 10, 2021 22 / 60
Different Preferences toward Risk

People’s preferences over risk are also heterogeneous.

Specifically,
I Risk averse (U(X ) < u(E [X ])): one prefers a certain income to a risky
income with the same expected value
F Risk premium: the maximum amount of money that a risk-averse person will pay
to avoid taking a risk
I Risk neutral (U(X ) = u(E [X ])): one is indifferent between a certain income
and a risky income with the same expected value
I Risk loving (U(X ) > u(E [X ])): one prefers a risky income to a certain income
with the same expected value

Tianyu Han (Berkeley Haas) UGBA 101A (Summer 2021) Section 3 June 10, 2021 22 / 60
 job with an expected income of $20,000. This outcome is shown graphically by
drawing a horizontal line to the vertical axis from point F, which bisects straight

Different Preferences toward Risk: Visualization

Utility
E
18
D
16
C
14
B F
13.5
A
10

0 10 15 16 20 30
Income ($1000)
(a)

Utility Utility
E E
18 18

C
12

C
8
A
6
A
3

0 10 20 30 0 10 20 30
Income ($1000) Income ($1000)
(b) (c)

FIGURE 5.3
RISK AVERSE, RISK LOVING, AND RISK NEUTRAL
People differ in their preferences toward risk. In (a), a consumer’s marginal utility diminishes as income in-
creases. The consumer is risk averse because she would prefer a certain income of $20,000 (with a utility of 16)
to a gamble with a .5 probability of $10,000 and a .5 probability of $30,000 (and expected utility of 14). In (b),
the consumer is risk loving: She would prefer the same gamble (with expected utility of 10.5) to the certain
income (with a utility of 8). Finally, the consumer in (c) is risk neutral and indifferent between certain and uncer-
tain events with the same expected income.

Tianyu Han (Berkeley Haas) Figure 1: Different Preferences

UGBA 101A (Summer 2021) Section toward
3 Risk June 10, 2021 23 / 60
 188 PART 2 Producers, Consumers, and Competitive Markets
Risk Premium: Visualization
Utility
G
20

18 E

C
14 F

A Risk Premium
10

10 16 20 30 40
Income ($1000)

FIGURE 5.4
RISK PREMIUM
The risk premium, CF, measures the amount of income that an individual would give up
to leave her indifferent between a risky choice and a certain one. Here, the risk premium is
$4000 because a certain income of $16,000 (at point C) gives her the same expected utility
(14) as the uncertain income (a .5 probability of being at point A and a .5 probability of being
at point E) that has an expected value of $20,000.

Figure
line AE (thus representing an 2: RiskofPremium
average $10,000 and $30,000). But the utility
level of 14 can also be achieved if the woman has a certain income of $16,000,
as shown by dropping a vertical line from point C. Thus, the risk premium of
Tianyu Han $4000,
(Berkeley Haas) given by line segment CF,(Summer
UGBA 101A is the amount of expected
2021) Section 3 income ($20,000 mi-June 10, 2021 24 / 60
Outline
1 Choice under Uncertainty
Setup
Maximax, Maximin, and Minimax
Expected Money Value
Expected Utility Hypothesis
2 Production
Technology
Short-Run Production
Long-Run Production
Returns to Scale
3 Costs
Different Types of Costs
Cost in the Short-Run
Cost in the Long-Run
Relationship between Short-Run and Long-Run
Efficiencies in the Long-Run

Tianyu Han (Berkeley Haas) UGBA 101A (Summer 2021) Section 3 June 10, 2021 25 / 60
Production and Technology

We are switching gears to look at the supply side of the market.

Going back to the big picture of Economics, to satisfy unlimited human
needs, firms take inputs and transform them into products
The primitive of the firms is the

Tianyu Han (Berkeley Haas) UGBA 101A (Summer 2021) Section 3 June 10, 2021 26 / 60
Production and Technology

We are switching gears to look at the supply side of the market.

Going back to the big picture of Economics, to satisfy unlimited human
needs, firms take inputs and transform them into products
The primitive of the firms is the technology (cf. preferences in the
consumer theory) represented by a production function (cf. utility in the
consumer theory)

Tianyu Han (Berkeley Haas) UGBA 101A (Summer 2021) Section 3 June 10, 2021 26 / 60
Production and Technology

We are switching gears to look at the supply side of the market.

Going back to the big picture of Economics, to satisfy unlimited human
needs, firms take inputs and transform them into products
The primitive of the firms is the technology (cf. preferences in the
consumer theory) represented by a production function (cf. utility in the
consumer theory)
In this class, we focus on the technology of two inputs (Labor L and
Capital K ) and one output
Together, we have the production function: q = F(K , L )
Interpretation: what is the most output q the firm could get from a given
combination of Labor L and Capital K ?

Tianyu Han (Berkeley Haas) UGBA 101A (Summer 2021) Section 3 June 10, 2021 26 / 60
Outline
1 Choice under Uncertainty
Setup
Maximax, Maximin, and Minimax
Expected Money Value
Expected Utility Hypothesis
2 Production
Technology
Short-Run Production
Long-Run Production
Returns to Scale
3 Costs
Different Types of Costs
Cost in the Short-Run
Cost in the Long-Run
Relationship between Short-Run and Long-Run
Efficiencies in the Long-Run

Tianyu Han (Berkeley Haas) UGBA 101A (Summer 2021) Section 3 June 10, 2021 27 / 60
Short-Run: Production with One Variable Input (Labor)

Short-run: a period of time in which quantities of one or more production

inputs cannot be changed
Usually, in the short-run, we assume that labor is variable but capital is
fixed. In other words, in the production function, we take K as a given
parameter, and the production becomes a function of one variable, L

Tianyu Han (Berkeley Haas) UGBA 101A (Summer 2021) Section 3 June 10, 2021 28 / 60
Short-Run: Production with One Variable Input (Labor)

Short-run: a period of time in which quantities of one or more production

inputs cannot be changed
Usually, in the short-run, we assume that labor is variable but capital is
fixed. In other words, in the production function, we take K as a given
parameter, and the production becomes a function of one variable, L
Short-run production function: q

Tianyu Han (Berkeley Haas) UGBA 101A (Summer 2021) Section 3 June 10, 2021 28 / 60
Short-Run: Production with One Variable Input (Labor)

Short-run: a period of time in which quantities of one or more production

inputs cannot be changed
Usually, in the short-run, we assume that labor is variable but capital is
fixed. In other words, in the production function, we take K as a given
parameter, and the production becomes a function of one variable, L
Short-run production function: q = F(K̄ , L )

Tianyu Han (Berkeley Haas) UGBA 101A (Summer 2021) Section 3 June 10, 2021 28 / 60
Short-Run: Production with One Variable Input (Labor)

Short-run: a period of time in which quantities of one or more production

inputs cannot be changed
Usually, in the short-run, we assume that labor is variable but capital is
fixed. In other words, in the production function, we take K as a given
parameter, and the production becomes a function of one variable, L
Short-run production function: q = F(K̄ , L ) = F(L )

Tianyu Han (Berkeley Haas) UGBA 101A (Summer 2021) Section 3 June 10, 2021 28 / 60
Short-Run: Production with One Variable Input (Labor)

Short-run: a period of time in which quantities of one or more production

inputs cannot be changed
Usually, in the short-run, we assume that labor is variable but capital is
fixed. In other words, in the production function, we take K as a given
parameter, and the production becomes a function of one variable, L
Short-run production function: q = F(K̄ , L ) = F(L )
To study how L contributes to the production:
I Total product (TP): total output produced q
I Average product (AP): output per unit of a particular input
output q
APL = =
labor input L
I Marginal product (MP): additional output produced as a input is increased
by one unit
change in output Δq
MPL = =
change in labor input ΔL

Tianyu Han (Berkeley Haas) UGBA 101A (Summer 2021) Section 3 June 10, 2021 28 / 60
The Average-Marginal Relationship
If MPL > APL , then APL increases in L
If MPL < APL , then APL6 .decreases U N CL
2 P R O D U C T I O N Fin TIONS WITH A SINGLE INPUT 211
If MPL = APL , then APL is at a maximum (point A in Figure 3)

10 A
APL, MPL, chips per man-hour

5
APL is increasing APL is decreasing
so MPL > APL so MPL < APL
APL
0 6 12 18 24 30 36
L, thousands of man-hours per day

−5

−10
MPL

FIGURE 6.3 Average and Marginal Product Functions Figure 3: AP and MP

APL is the average product function. MPL is the marginal product function. The marginal product
function rises in the region of increasing marginal returns (L ⬍ 12) and falls in the region of
diminishing marginal returns (12 ⬍ L ⬍ 24). It becomes negative in the region of diminishing
total returns (L ⬎ 24). At point A, where APL is at a maximum, APL ⫽ MPL.

The other notion of productivity is the marginal product of labor, which we marginal product of
Tianyu Han (Berkeley MPL. The
write as Haas) marginal product
UGBA of labor
101A is the2021)
(Summer rate Section
at which3 total output changes laborJune
The10,
rate at which29 / 60
2021
The Average-Marginal Relationship
If MPL > APL , then APL increases in L
If MPL < APL , then APL6 .decreases U N CL
2 P R O D U C T I O N Fin TIONS WITH A SINGLE INPUT 211
If MPL = APL , then APL is at a maximum (point A in Figure 3)

10 A
APL, MPL, chips per man-hour

5
APL is increasing APL is decreasing
so MPL > APL so MPL < APL
APL
0 6 12 18 24 30 36
L, thousands of man-hours per day

−5

−10
MPL

FIGURE 6.3 Average and Marginal Product Functions Figure 3: AP and MP

APL is the average product function. MPL is the marginal product function. The marginal product
function rises in the region of increasing marginal returns (L ⬍ 12) and falls in the region of
As the use of an input increases with other inputs fixed, the resulting
diminishing marginal returns (12 ⬍ L ⬍ 24). It becomes negative in the region of diminishing
total returns (L ⬎ 24). At point A, where APL is at a maximum, APL ⫽ MPL.
additions to output will eventually decrease. This principle is known as
the law of diminishing marginal returns.
The other notion of productivity is the marginal product of labor, which we marginal product of
Tianyu Han (Berkeley MPL. The
write as Haas) marginal product
UGBA of labor
101A is the2021)
(Summer rate Section
at which3 total output changes laborJune
The10,
rate at which29 / 60
2021
 212 CHAPTER 6 INPUTS AND PRODUCTION FUNCTIONS
Relationship between TP, MP, and AP
C

Marginal product at
L1 equals slope of

Q, thousands of chips per day

line BC
B
Average product at
L0 equals slope of
ray 0A

Total
A product
Q0 function

0 L0 18 L1 24
L, thousands of man-hours per day
APL, MPL , chips per man-hour

FIGURE 6.4 Relationship APL

among Total, Average, and
Marginal Product Functions
The marginal product of labor 0 L0 18 L1 24
at any point equals the slope of L, thousands of man-hours per day
the total product curve at that
point. The average product at
any point is equal to the slope
of the ray from the origin to
the total product curve at that MPL
point.

Figure 4: TP, MP, and AP

A P P L I C A T I O N 6.2

Tianyu
TheHan (Berkeley Haas)
Resurgence of Labor UGBA 101A (Summer 2021) Section 3 June 10, 2021 30 / 60
Outline
1 Choice under Uncertainty
Setup
Maximax, Maximin, and Minimax
Expected Money Value
Expected Utility Hypothesis
2 Production
Technology
Short-Run Production
Long-Run Production
Returns to Scale
3 Costs
Different Types of Costs
Cost in the Short-Run
Cost in the Long-Run
Relationship between Short-Run and Long-Run
Efficiencies in the Long-Run

Tianyu Han (Berkeley Haas) UGBA 101A (Summer 2021) Section 3 June 10, 2021 31 / 60
Long-Run: Production with Two Variable Inputs
Long-run: amount of time needed to make all production inputs variable
Long-run production function: q = F(K , L )

Tianyu Han (Berkeley Haas) UGBA 101A (Summer 2021) Section 3 June 10, 2021 32 / 60
Long-Run: Production with Two Variable Inputs
Long-run: amount of time needed to make all production inputs variable
Long-run production function: q = F(K , L )
Isoquant: a curve that shows all of the combinations of labor and capital
that can produce a given level of output. See Figure 5 as an example.

q3
q2
q1
L
Figure 5: An Isoquant Example (q1 < q2 < q3 )

Tianyu Han (Berkeley Haas) UGBA 101A (Summer 2021) Section 3 June 10, 2021 32 / 60
Marginal Rate of Technical Substitution

Marginal rate of technical substitution (of labor for capital) (MRTSL ,K ):

the rate at which the quantity of capital can be reduced for every one-unit
increase in the quantity of labor, holding the quantity of output constant.
The marginal rate of technical substitution is analogous to the marginal
rate of substitution from consumer theory.
Just as the marginal rate of substitution of good 1 for good 2 is the slope
of an indifference curve drawn with x1 on the horizontal axis and x2 on
the vertical axis, the marginal rate of technical substitution of labor for
capital is the slope of an isoquant drawn with L on the horizontal axis and
K on the vertical axis.

Tianyu Han (Berkeley Haas) UGBA 101A (Summer 2021) Section 3 June 10, 2021 33 / 60
Marginal Rate of Technical Substitution: Derivation

To find MRTSL ,K , we compute the total derivative of q,

=MPL =MPK
=0
𝜕F 𝜕F
dq = ·dL + ·dK
𝜕L 𝜕K
=⇒ MPL · dL = −MPK · dK
dK MPL
=⇒ MRTSL ,K ≡ − = .
dL MPK

Tianyu Han (Berkeley Haas) UGBA 101A (Summer 2021) Section 3 June 10, 2021 34 / 60
Outline
1 Choice under Uncertainty
Setup
Maximax, Maximin, and Minimax
Expected Money Value
Expected Utility Hypothesis
2 Production
Technology
Short-Run Production
Long-Run Production
Returns to Scale
3 Costs
Different Types of Costs
Cost in the Short-Run
Cost in the Long-Run
Relationship between Short-Run and Long-Run
Efficiencies in the Long-Run

Tianyu Han (Berkeley Haas) UGBA 101A (Summer 2021) Section 3 June 10, 2021 35 / 60
Returns to Scale

Returns to scale: The percentage by which output will increase when all
inputs are increased by a given percentage.

%Δ (quantity of output)
RTS = .
%Δ (quantity of all inputs)

This is to say, suppose that all inputs are scaled up by the same proportionate
amount 𝛼, where 𝛼 > 1, then:
Increasing returns to scale if F(𝛼K , 𝛼L ) > 𝛼F(K , L );
Constant returns to scale if F(𝛼K , 𝛼L ) = 𝛼F(K , L );
Decreasing returns to scale if F(𝛼K , 𝛼L ) < 𝛼F(K , L ).

Tianyu Han (Berkeley Haas) UGBA 101A (Summer 2021) Section 3 June 10, 2021 36 / 60
 smaller firms. This cost advantage of large-scale operation has been the traditional
justification for allowing firms to operate as regulated monopolists in markets such as
Returns to Scale (cont.)
electric power and oil pipeline transportation.

Returns to scale is related to the spacing of isoquants as shown in Figure 6:

K

K
Q=3
2 2 2
Q=3
1 Q=3 1 1 Q=2
Q=2
Q=2
Q=1
Q=1 Q=1
0 1 2 0 1 2 0 1 2
L L L
(a) Increasing Returns to Scale (b) Constant Returns to Scale (c) Decreasing Returns to Scale

FIGURE 6.18 Increasing, Constant,Figure 6: Returns

and Decreasing Returns toto Scale
Scale
In panel (a), doubling the quantities of capital and labor more than doubles output. In panel (b),
doubling the quantities of capital and labor exactly doubles output. In panel (c), doubling the
quantities of capital and labor less than doubles output.
Difference between returns to scales and diminishing marginal returns:
15
Returns to scale pertains to the impact of an increase in all input
Therefore, the percentage change in all input quantities is (  1)  100 percent.
quantities simultaneously;
Marginal returns pertains to the impact of an increase in the quantity of a
single input (such as labor) holding the quantities of all other inputs fixed.

Tianyu Han (Berkeley Haas) UGBA 101A (Summer 2021) Section 3 June 10, 2021 37 / 60
Exercise (Ch.7 Appendix Q1)
Of the following production functions, which exhibit increasing, constant, or
decreasing returns to scale?
1 F(K , L ) = K 2 L
2 F(K , L ) = 10K + 5L
1
3 F(K , L ) = (KL ) 2

Tianyu Han (Berkeley Haas) UGBA 101A (Summer 2021) Section 3 June 10, 2021 38 / 60
Outline
1 Choice under Uncertainty
Setup
Maximax, Maximin, and Minimax
Expected Money Value
Expected Utility Hypothesis
2 Production
Technology
Short-Run Production
Long-Run Production
Returns to Scale
3 Costs
Different Types of Costs
Cost in the Short-Run
Cost in the Long-Run
Relationship between Short-Run and Long-Run
Efficiencies in the Long-Run

Tianyu Han (Berkeley Haas) UGBA 101A (Summer 2021) Section 3 June 10, 2021 39 / 60
Different Types of Costs
Total cost (TC or C): Total economic cost of production, consisting of
fixed and variable costs.
Fixed cost (FC): Cost that does not vary with the level of output and that
can be eliminated only by shutting down.
Variable cost (VC): Cost that varies as output varies.
Marginal cost (MC): Increase in cost resulting from the production of
one extra unit of output.
ΔTC Δ(FC + VC) ΔVC
MC = = =
Δq Δq Δq
Average total cost (ATC or AC): Firm’s total cost divided by its level of
output.
AC = TC/q
Average fixed cost (AFC): Fixed cost divided by the level of output.
AFC = FC/q
Average variable cost (AVC): Variable cost divided by the level of
output.
AVC = VC/q
Tianyu Han (Berkeley Haas) UGBA 101A (Summer 2021) Section 3 June 10, 2021 40 / 60
Outline
1 Choice under Uncertainty
Setup
Maximax, Maximin, and Minimax
Expected Money Value
Expected Utility Hypothesis
2 Production
Technology
Short-Run Production
Long-Run Production
Returns to Scale
3 Costs
Different Types of Costs
Cost in the Short-Run
Cost in the Long-Run
Relationship between Short-Run and Long-Run
Efficiencies in the Long-Run

Tianyu Han (Berkeley Haas) UGBA 101A (Summer 2021) Section 3 June 10, 2021 41 / 60
 CHAPTER 7 THE COST OF PRODUCTION 247
Short-Run Cost Curves
TC
Cost 400 VC
(dollars
per
year)
300

175 A
FIGURE 7.1
100 COST CURVES
FC FOR A FIRM
In (a) total cost TC is the vertical
0
1 2 3 4 5 6 7 8 9 10 11 sum of fixed cost FC and vari-
Output (units per year) able cost VC. In (b) average total
(a) cost ATC is the sum of average
Cost 100 MC
variable cost AVC and average
(dollars
fixed cost AFC. Marginal cost
per
MC crosses the average vari-
unit)
75 able cost and average total cost
curves at their minimum points.

50 ATC
AVC

25

AFC
0
1 2 3 4 5 6 7 8 9 10 11
Output (units per year)
(b)

Figure 7: Short-Run Cost Curves

Observe in Figure 7.1 (a) that fixed cost FC does not vary with output—it is
shown as a horizontal line at $50. Variable cost VC is zero when output is zero and
then
Tianyu increases
Han continuously
(Berkeley Haas) as output increases. The(Summer
UGBA 101A total cost2021)
curve TC is3deter-
Section June 10, 2021 42 / 60
Exercise (Ch.7 Problem 9)
The short-run cost function of a company is given by the equation
TC = 200 + 55q, where TC is the total cost and q is the total quantity of
output, both measured in thousands.
1 What is the company’s fixed cost?
2 If the company produced 100,000 units of goods, what would be its
average variable cost?
3 What would be its marginal cost of production?
4 What would be its average fixed cost?

Tianyu Han (Berkeley Haas) UGBA 101A (Summer 2021) Section 3 June 10, 2021 43 / 60
Exercise (Ch.7 Problem 9)
The short-run cost function of a company is given by the equation
TC = 200 + 55q, where TC is the total cost and q is the total quantity of
output, both measured in thousands.
1 What is the company’s fixed cost?
2 If the company produced 100,000 units of goods, what would be its
average variable cost?
3 What would be its marginal cost of production?
4 What would be its average fixed cost?
5 Suppose the company borrows money and expands its factory. Its fixed
cost rises by $50,000, but its variable cost falls to $45,000 per 1000 units.
The cost of interest (i) also enters into the equation. Each 1-point
increase in the interest rate raises costs by $3000. Write the new cost
equation.

Tianyu Han (Berkeley Haas) UGBA 101A (Summer 2021) Section 3 June 10, 2021 43 / 60
Outline
1 Choice under Uncertainty
Setup
Maximax, Maximin, and Minimax
Expected Money Value
Expected Utility Hypothesis
2 Production
Technology
Short-Run Production
Long-Run Production
Returns to Scale
3 Costs
Different Types of Costs
Cost in the Short-Run
Cost in the Long-Run
Relationship between Short-Run and Long-Run
Efficiencies in the Long-Run

Tianyu Han (Berkeley Haas) UGBA 101A (Summer 2021) Section 3 June 10, 2021 44 / 60
Isocost
Isocost: The set of combinations of labor and capital that yield the same total
cost for the firm. See Figure 8 as an example.

K
TC2
r
TC1
r
TC0
r

TC0 TC1 TC2 L

w w w

Figure 8: Isocost Lines (TC0 < TC1 < TC2 )

More generally, for an arbitrary level of total cost TC, and input prices w and r,
the equation of the isocost line is K = TC/r − (w/r)L , where the slope is −w/r.
Tianyu Han (Berkeley Haas) UGBA 101A (Summer 2021) Section 3 June 10, 2021 45 / 60
Cost Minimization Problem

(Long-Run) Cost minimization problem (CMP): The problem of finding

the input combination that minimizes a firm’s total cost of producing a
particular level of output.
Let C = wL + rK be the cost function and q = F(L , K ) be the production
function. To achieve a production q0 , our problem is

min wL + rK subject to q0 = F(L , K ).

L ,K

The solutions to the CMP L ∗ (w , r , q0 ) and K ∗ (w , r , q0 ) are the firm’s

(long-run) input demand for the production.
I Interpretation: given production function F(L , K ) and input prices w and r, to
most effectivelyproduce q0 goods, we need L ∗ labor and K ∗ capital.

Tianyu Han (Berkeley Haas) UGBA 101A (Summer 2021) Section 3 June 10, 2021 46 / 60
Cost Minimization Problem (cont.)

Similar to a consumer problem, the optimal production set is the

tangency point of the isoquant curve and isocost curve.
Therefore, we have the following optimality condition for a cost
minimization problem:
MPL w
=
MPK r
where MPL /MPK is the (negative) slope of the isoquant and w/r is the
(negative) slope of the isocost.
Standard procedure to solve a CMP:
MPL
I First, we use the optimality condition MP = wr to find the relationship
K
∗
between optimal L and K ∗
I Then, we apply the relationship above to the isoquant, and solve for L ∗ and
K ∗ respectively as a function of w, r, and q0 .

Tianyu Han (Berkeley Haas) UGBA 101A (Summer 2021) Section 3 June 10, 2021 47 / 60
Allocation Rules of Inputs

Similar to the consumer problems , we also have the allocation rules of

inputs (Figure 9):
MPL MPK
If w = r , then L , K : This is optimal (Point A );
MPL MPK
If w> then L ↑ , K ↓ : Keeping output (q0 ) constant, a firm operating
r ,
at Point B could spend an additional dollar on labor (L ) input and save
more than one dollar by reducing its capital (K ) input;
If MPL MPK
w < r , then L , K : Keeping output (q0 ) constant, a firm operating
↓ ↑

at Point C could spend an additional dollar on capital (K ) input and save

more than one dollar by reducing its labor (L ) input.

Tianyu Han (Berkeley Haas) UGBA 101A (Summer 2021) Section 3 June 10, 2021 48 / 60
Allocation Rules of Inputs
K
Q0 isoquant
TC1
r

TC0
B MPL MPK
A: = : L +K
r w r
MPL MPK
B: > : L↑ + K↓
w r
MPL MPK
A C: < : L↓ + K↑
w r

TC0 TC1 L
w w

Figure 9: Cost Minimization and the Allocation Rules

Tianyu Han (Berkeley Haas) UGBA 101A (Summer 2021) Section 3 June 10, 2021 49 / 60
Exercise (Ch.7 Appendix Q2)
The production function for a product is given by q = 100KL . If the price of
capital is $120 per day and the price of labor $30 per day, what is the
minimum cost of producing 1000 units of output?

Tianyu Han (Berkeley Haas) UGBA 101A (Summer 2021) Section 3 June 10, 2021 50 / 60
Long-Run Total Cost Curve

Tianyu Han (Berkeley Haas) UGBA 101A (Summer 2021) Section 3 June 10, 2021 51 / 60
Long-Run Total Cost Curve
258 PART 2 Producers, Consumers, and Competitive Markets

Capital
per
year 150 $3000 Isocost Line

Expansion Path
$2000
Isocost Line
100
C
75
B
50 300 Unit Isoquant
A 200 Unit
25 Isoquant

50 100 150 200 300

Labor per year
(a)

Cost
(dollars F Long-Run Total Cost
per 3000
year)

E
2000

D
1000

100 200 300

Output (units per year)
(b)

Figure
FIGURE 7.6 10: Long-Run Total Cost Curve
A FIRM’S EXPANSION PATH AND LONG-RUN
Tianyu Han (Berkeley Haas) TOTAL COST CURVE
UGBA 101A (Summer 2021) Section 3 June 10, 2021 51 / 60
Outline
1 Choice under Uncertainty
Setup
Maximax, Maximin, and Minimax
Expected Money Value
Expected Utility Hypothesis
2 Production
Technology
Short-Run Production
Long-Run Production
Returns to Scale
3 Costs
Different Types of Costs
Cost in the Short-Run
Cost in the Long-Run
Relationship between Short-Run and Long-Run
Efficiencies in the Long-Run

Tianyu Han (Berkeley Haas) UGBA 101A (Summer 2021) Section 3 June 10, 2021 52 / 60
Inflexibility of Short-Run Production

Tianyu Han (Berkeley Haas) UGBA 101A (Summer 2021) Section 3 June 10, 2021 53 / 60
Inflexibility of Short-Run Production
7.4 Long-Run versus Short-Run Cost Curves
Figure 11 illustrates an example: capital is fixed at level K1 in the
We saw earlier (see Figure 7.1—page 247) that short-run average cost curves are
short-run.
U-shaped. We will see that long-run average cost curves can also be U-shaped,
but different economic factors explain the shapes of these curves. In this section,
Therefore, if we increase output level from q1 to q2 , the firm can only
we discuss long-run average and marginal cost curves and highlight the differ-
ences between these curves and their short-run counterparts.
increase input L . So, we have a horizontal short-run expansion path.
The Inflexibility of Short-Run Production
Point P denotes thethe short-run
Recall that we defined long run as occurringinput setto to
when all inputs produce
the firm are q2 of output. It lies on
variable. In the long run, the firm’s planning horizon is long enough to allow
the isocost curve
for a change in plantEF, which
size. This is higher
added flexibility allows the firmthan the
to produce at long-run isocost curve CD.
a lower average cost than in the short run. To see why, we might compare the
The costsituation
of production is arehigher
in which capital and labor both flexiblein the
to the case inshort-run,
which capital because the firm is
is fixed in the short run.
unable to substitute
Figure relatively
7.8 shows the firm’s inexpensive
production isoquants. The firm’s long-run capital
expan-
sion path is the straight line from the origin that corresponds to the expansion
for more costly labor
when it expands production.
Capital
per E
year

C FIGURE 7.8
THE INFLEXIBILITY OF
SHORT-RUN PRODUCTION
Long-Run
Expansion Path When a firm operates in the short run, its
A cost of production may not be minimized
because of inflexibility in the use of capi-
K2 tal inputs. Output is initially at level q1. In
Short-Run
the short run, output q2 can be produced
Expansion Path
P only by increasing labor from L1 to L3 be-
K1
q2 cause capital is fixed at K1. In the long run,
the same output can be produced more
cheaply by increasing labor from L1 to L2
q1 and capital from K1 to K2.

L1 L2 B L3 D F
Labor per year

Figure 11: Inflexibility of Short-Run Production

Tianyu Han (Berkeley Haas) UGBA 101A (Summer 2021) Section 3 June 10, 2021 53 / 60
Short-Run vs. Long-Run Cost
Suppose a firm is choosing the plant size for production: small, medium,
or large. The short-run AC curves are given by SAC1 , SAC2 , SAC3 in
Figure 12.
If the firm expects to produce q0 of output, then it’s optimal to build a
small plant; if the firm expects to produce q2 of output, then it’s optimal to
build a medium plant; if the firm expects to produce q3 of output, then it’s
optimal to build a large plant.
In the long-run, the firm is able to choose the “right” the plant size
according to its expected output. The long-run AC curve is given by the
crosshatched portions of the short-run AC curves because these show
the minimum cost of production for any output level.
If more plant sizes are available, then there would be more short-run
curves in the figure, and the scalloped LAC curve would become
progressively smoother. If K can be any positive number, the LAC
becomes the lower envelope of an infinite number of SAC curves.
Somewhat surprisingly, a SAC curve does not generally reach its
minimum at the output where short-run and long-run average costs are
equal unless both curves reach a minimum (at q2 ).
Finally, each point on the long-run MC curve is the short-run MC
associated with the most cost-efficient plant.
Tianyu Han (Berkeley Haas) UGBA 101A (Summer 2021) Section 3 June 10, 2021 54 / 60
Short-Run vs. Long-Run Cost

Tianyu Han (Berkeley Haas) UGBA 101A (Summer 2021) Section 3 June 10, 2021 55 / 60
 its diseconomies at higher output levels.
To clarify the relationship between short-run and long-run cost curves, con-
Short-Run vs. Long-Run Cost sider a firm that wants to produce output q1. If it builds a small plant, the short-
run average cost curve SAC1 is relevant. The average cost of production (at B on
SAC1) is $8. A small plant is a better choice than a medium-sized plant with an

FIGURE 7.11
LONG-RUN Cost
(dollars
COST WITH per unit
SAC 1
ECONOMIES AND of output) SMC 1 A
SAC 2 SMC 3 SAC 3
LAC
DISECONOMIES $10

OF SCALE $8
SMC 2
B
The long-run average cost
curve LAC is the envelope
of the short-run average LMC
cost curves SAC1, SAC2,
and SAC3. With econo-
mies and diseconomies of
scale, the minimum points
of the short-run average
cost curves do not lie on the
long-run average cost curve. q0 q1 q2 q3 Output

Figure 12: Short-Run vs. Long-Run Cost

Tianyu Han (Berkeley Haas) UGBA 101A (Summer 2021) Section 3 June 10, 2021 55 / 60
Outline
1 Choice under Uncertainty
Setup
Maximax, Maximin, and Minimax
Expected Money Value
Expected Utility Hypothesis
2 Production
Technology
Short-Run Production
Long-Run Production
Returns to Scale
3 Costs
Different Types of Costs
Cost in the Short-Run
Cost in the Long-Run
Relationship between Short-Run and Long-Run
Efficiencies in the Long-Run

Tianyu Han (Berkeley Haas) UGBA 101A (Summer 2021) Section 3 June 10, 2021 56 / 60
Economies of Scale
Economies of Scale: A characteristic of production in which average
cost decreases as output goes up.

%ΔC ΔC/C ΔC/Δq MC

EOS = = = = .
%Δq Δq/q C/q AC

Tianyu Han (Berkeley Haas) UGBA 101A (Summer 2021) Section 3 June 10, 2021 57 / 60
Economies of Scale
Economies of Scale: A characteristic of production in which average
cost decreases as output goes up.

%ΔC ΔC/C ΔC/Δq MC

EOS = = = = .
%Δq Δq/q C/q AC

Returns to scale and economies of scale are closely related, because the
returns to scale of the production function determine how long-run
average cost varies with output.
Figure 13 summarizes the relationship between returns to scale and
economies of scale:
1 If average cost decreases as output increases (when q < q2 ), we have
economies of scale and increasing returns to scale.
2 If average cost increases as output increases (when q > q2 ), we have
diseconomies of scale and decreasing returns to scale.
3 If average cost stays the same as output increases (when q = q2 ), we have
neither economies nor diseconomies of scale and constant returns to scale.

Tianyu Han (Berkeley Haas) UGBA 101A (Summer 2021) Section 3 June 10, 2021 57 / 60
Production and Cost

Figure 13: Production and Cost Curves

Tianyu Han (Berkeley Haas) UGBA 101A (Summer 2021) Section 3 June 10, 2021 58 / 60
Economies of Scope

Economies of Scope: A production characteristic in which the total cost

of producing given quantities of two goods in the same firm is less than
the total cost of producing those quantities in two single product firms.
Economies of scope happens if C(q1 ) + C(q2 ) > C(q1 , q2 ).
We can also measure the degree of economies of scope:

C(q1 ) + C(q2 ) − C(q1 , q2 )

SC =
C(q1 , q2 )

Tianyu Han (Berkeley Haas) UGBA 101A (Summer 2021) Section 3 June 10, 2021 59 / 60
Exercise (Ch.7 Appendix Q4)
Suppose the process of producing lightweight parkas by Polly’s Parkas is
described by the function q = 10K 0.8 (L − 40)0.2 where q is the number of
parkas produced, K the number of computerized stitching-machine hours,
and L the number of person-hours of labor. In addition to capital and labor,
$10 worth of raw materials is used in the production of each parka.

Tianyu Han (Berkeley Haas) UGBA 101A (Summer 2021) Section 3 June 10, 2021 60 / 60
Exercise (Ch.7 Appendix Q4)
Suppose the process of producing lightweight parkas by Polly’s Parkas is
described by the function q = 10K 0.8 (L − 40)0.2 where q is the number of
parkas produced, K the number of computerized stitching-machine hours,
and L the number of person-hours of labor. In addition to capital and labor,
$10 worth of raw materials is used in the production of each parka.
1 By minimizing cost subject to the production function, derive the
cost-minimizing demands for K and L as a function of output (q), wage
rates (w), and rental rates on machines (r). Use these results to derive
the total cost function: that is, costs as a function of q, r, w, and the
constant $10 per unit materials cost.

Tianyu Han (Berkeley Haas) UGBA 101A (Summer 2021) Section 3 June 10, 2021 60 / 60
Exercise (Ch.7 Appendix Q4)
Suppose the process of producing lightweight parkas by Polly’s Parkas is
described by the function q = 10K 0.8 (L − 40)0.2 where q is the number of
parkas produced, K the number of computerized stitching-machine hours,
and L the number of person-hours of labor. In addition to capital and labor,
$10 worth of raw materials is used in the production of each parka.
1 By minimizing cost subject to the production function, derive the
cost-minimizing demands for K and L as a function of output (q), wage
rates (w), and rental rates on machines (r). Use these results to derive
the total cost function: that is, costs as a function of q, r, w, and the
constant $10 per unit materials cost.
2 This process requires skilled workers, who earn $32 per hour. The rental
rate on the machines used in the process is $64 per hour. At these factor
prices, what are total costs as a function of q? Does this technology
exhibit decreasing, constant, or increasing returns to scale?

Tianyu Han (Berkeley Haas) UGBA 101A (Summer 2021) Section 3 June 10, 2021 60 / 60
Exercise (Ch.7 Appendix Q4)
Suppose the process of producing lightweight parkas by Polly’s Parkas is
described by the function q = 10K 0.8 (L − 40)0.2 where q is the number of
parkas produced, K the number of computerized stitching-machine hours,
and L the number of person-hours of labor. In addition to capital and labor,
$10 worth of raw materials is used in the production of each parka.
1 By minimizing cost subject to the production function, derive the
cost-minimizing demands for K and L as a function of output (q), wage
rates (w), and rental rates on machines (r). Use these results to derive
the total cost function: that is, costs as a function of q, r, w, and the
constant $10 per unit materials cost.
2 This process requires skilled workers, who earn $32 per hour. The rental
rate on the machines used in the process is $64 per hour. At these factor
prices, what are total costs as a function of q? Does this technology
exhibit decreasing, constant, or increasing returns to scale?
3 Polly’s Parkas plans to produce 2000 parkas per week. At the factor
prices given above, how many workers should the firm hire (at 40 hours
per week) and how many machines should it rent (at 40 machine-hours
per week)? What are the marginal and average costs at this level of
production?
Tianyu Han (Berkeley Haas) UGBA 101A (Summer 2021) Section 3 June 10, 2021 60 / 60