UC Merge Bargaining Discussion

Description

ssigned Readings:
Chapter 16. Bargaining.
Chapter 17. Making Decisions with Uncertainty.
Chapter 18. Auctions.
Initial Postings: Read and reflect on the assigned readings for the week. Then post what you thought was the most important concept(s), method(s), term(s), and/or any other thing that you felt was worthy of your understanding in each assigned textbook chapter.Your initial post should be based upon the assigned reading for the week, so the textbook should be a source listed in your reference section and cited within the body of the text. Other sources are not required but feel free to use them if they aid in your discussion.
Also, provide a graduate-level response to each of the following questions:
American Airlines and British Airways are proposing to merge. If British pilots and American pilots are represented by different unions, how would this merger affect airline costs?

3 attachmentsSlide 1 of 3attachment_1attachment_1attachment_2attachment_2attachment_3attachment_3.slider-slide > img { width: 100%; display: block; }
.slider-slide > img:focus { margin: auto; }

Unformatted Attachment Preview

CHAPTER
PowerPoint Slides
© Luke M. Froeb,
Vanderbilt 2014
16
Bargaining
? Strategic view of bargaining: model as either a
simultaneous-move or sequential-move game.
? A player can gain bigger share of the “pie” by
• changing a simultaneous-move game into a sequentialmove game with a ?rst-mover advantage;
or by
• committing to a position.
? Credible commitments (threats) are difficult to make
because they require players to commit to a course of
action against their self-interest. Thus, the best threat
is one you never have to use.
©2018 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part, except for use as permitted in a license distributed with a certain product or service or
otherwise on a password-protected website for classroom use. ©Kamira/Shutterstock Images
2
• continued
? The strategic view of bargaining focuses on how the
outcome of bargaining games depends on who moves
?rst and who can commit to a bargaining position, as
well as whether the other player can make a counteroffer.
? The non-strategic view of bargaining focuses on the
gains and alternatives to agreement to determine the
outcome of barganing.
• Main insight: The gains from agreement relative to the
alternatives to agreement determine the terms of any
agreement.
• Anything you can do to increase your opponent’s relative
gains from reaching agreement or to decrease your own
will improve your bargaining position.
©2018 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part, except for use as permitted in a license distributed with a certain product or service or
otherwise on a password-protected website for classroom use. ©Kamira/Shutterstock Images
3
1: NBA
• In summer 2011, National Basketball Assc. owners were negotiating
with the players’ union over how to split revenues
• Union wanted 57%, owners only offered 50%
• Owners locked out the players, cancelling the start of the season
• After months of legal threats and lost revenue, players finally
accepted owners’ initial offer
4
2: Texaco
• In 1985, Texaco was found guilty by a Texas jury for interfering with
Pennzoil’s attempt to buy Getty Oil.
• Texaco was fined $10.5 billion, but appealed the verdict and began negotiating
with Pennzoil.
• In 1987, Texaco filed for bankruptcy. Pennzoil was then unable to seize
control of Texaco’s assets.
• Texaco was also freed from the responsibility to pay interest and dividends.
• One year later Texaco and Pennzoil settled the case, with Texaco having
to pay only $3 billion. Texaco successfully used bankruptcy to reduce its
liability by over 70%
• This chapter examines bargaining, and strategies to improve your
bargaining position, like those used by Bear Stearns and Texaco .
5
Introduction: Bargaining
• There are two complementary ways to look at bargaining:
• the strategic view analyzes bargaining using the tools of game theory (ch 15).
Bargaining can be viewed as either a simultaneous-move game with two
equilibria or a sequential-move game, where one player gains an advantage by
committing to a position.
• the non-strategic view acknowledges that real life negotiations don’t have fixed
rules as formal games do. This view postulates that the alternatives to
agreement determine the terms of agreement, regardless of the rules of the
negotiating game.
• If you can increase your opponent’s relative gain, or decrease your own, you can gain a
bigger share of the pie.
• By declaring (or threatening) bankruptcy, Bear Stearns and Texaco were able to improve their
bargaining “position”, i.e., by changing the alternatives to agreement, they changed the
terms of agreement.
6
Bargaining: a simultaneous-move game
• Example: Wage negotiations
• Management and labor are bargaining over a fixed sum of $200 million
• Two possible strategies are available to each player: “bargain hard” or
“accommodate.”
• If both bargain hard, no deal is reached. Neither side gains.
• If both accommodate, they split the gains from trade.
• If one player bargains hard and the other accommodates, then the player
who bargains hard takes 75% of the “pie”
7
Bargaining: a simultaneous game (cont.)
• There are two equilibria for this game
• Management prefers the lower-left equilibrium
• Labor prefers the upper-right.
• This bargaining game has the same structure as a game of “chicken”
• Each party can gain by committing to a position, which turns it into a sequential
game
8
Bargaining: a sequential-move game
• In sequential-move bargaining the first “player” makes an
offer that the second “player” can accept or refuse.
• Again to analyze a sequential-move game look ahead and
reason back.
• The first-mover “looks ahead and reasons back” to determine the
how her rival will react to each possible move. Then the first-mover
can determine the consequences of each possible move.
• In this case, the sequential-move games present a “firstmover advantage,” i.e., by moving first a player can gain an
advantage.
• Using the same wage negotiation example, we can look at
sequential-move bargaining and first-mover advantage.
9
Bargaining game: first-mover advantage
• Management “wins” by moving first and making a low offer
Management
low offer
generous offer
Union
strike
accept
0,0
150 , 50
strike
0,0
accept
50 , 150
10
Bargaining game: first-mover advantage
• Union can change the outcome by credibly committing to strike if a
low offer is made
Management
low offer
generous offer
Union
strike
0,0
strike
0,0
accept
50 , 150
11
Sequential-move bargaining (cont.)
• Because the management has the first-mover
advantage, it is in their best interest to make a low offer,
and it is in the union’s best interest to accept that offer.
• However, if the union can effectively threaten to strike
(in such a way that the management believes them)
they can change the outcome of the game despite
management’s first-mover advantage.
• Credible threats are hard to make because they require the
union act against its self interest.
• If management doesn’t believe the threat, the union might
actually have to follow through on the threat.
• So, again, the best threat is one you never have to use.
12
Non-strategic View of Bargaining
• The outcome in strategic bargaining “games” is dependent
on the rules of the game, but in real life, the rules are not
always clear.
• John Nash proved that any reasonable outcome to a bargain
would maximize the product of the bargainers’ surplus.
• This is known as an “axiomatic” or “non-strategic” view of
bargaining.
• In this view, the gains from bargaining relative to the alternatives
to bargaining, determine the terms of any bargain.
• This view also teaches that to increase your bargaining power,
• you can increase your opponent’s gain from reaching agreement or decrease
your own.
• If your rival has more to gain by agreeing, he becomes more eager to reach
agreement, and accepts a smaller share of the surplus.
13
Non-strategic view (cont.)
• Nash’s axiomatic approach:
•
•
•
•
[ S1(z) – D1 ] x [ (S2(z) – D2 ] , where:
z is the agreement
S1(z) is the value of the agreement to player 1 (sub 2 for player two)
D1 is “disagreement value,” or pay-off if no agreement is reached, for player 1
(sub 2 for player two)
• So player 1’s gain from agreement is (S1(z) – D1)
14
Non-strategic view (cont.)
• For example, two brothers are bargaining over a dollar.
• If no agreement is reached, neither participant gains.
• If they reach an agreement (z)
• Player one, the older brother, has a surplus of z
• Player two, the younger brother, has a surplus of 1 – z
• Nash’s solution is for them to “split” the gains from trade, i.e., {½, ½} is the axiomatic
solution.
• But, now the older brother receives a $0.50 bonus for “sharing nicely,” and the
total gain rises from $1.00 to $1.50
• The Nash bargaining outcome is for the brothers to split to total gains – each receiving
$0.75, meaning the older brother effectively shares half of his bonus.
• By increasing the first player’s gain to reaching agreement, he becomes more eager to reach
agreement, and “shares” his gain with his brother.
15
Bonuses for agreement
• Giving a bonus for reaching agreement is similar to incentive
compensation schemes used by many companies.
• When salespeople are offered bonuses it increases their eagerness to
reach agreement and this induces them to accept “weaker”
agreements.
• So giving salespeople such a bonus driven incentive will lead to lower prices
when they negotiate with customers.
• (This concept will be further addressed in chapter 20)
16
Alternatives to agreement
• Nash’s bargaining solution incorporates the effect of alternatives to
agreement on the agreement itself. This creates some sound
bargaining advice:
• To improve your own bargaining position, increase your opponent’s gain
from reaching agreement, S2(z) – D2, or reduce your own gain from reaching
agreement, S1(z) – D1.
• When you increase your opponent’s gain in agreement, you make him more
willing to agree.
• Reducing your own gain makes you less willing to compromise and helps to
improve your position.
17
How Nash’s view differs from strategic
• The strategic view of bargaining places a greater emphasis on timing
and commitment in determining the outcome of the game.
• With the labor/management example, the union’s commitment to strike,
or management making the first move, changes the equilibrium of the
game.
• But neither action changes the gains of the agreement so neither would
affect the Nash bargaining outcome.
• The Nash bargaining outcome incorporates the idea that if you
decrease your own gain to agreement you become a better
bargainer.
• EXAMPLE: the best time to ask for a raise is when you have another
attractive offer waiting for you, you have less to gain by reaching
agreement. Your bargaining position improves.
• This is similar to the idea of “opportunity cost.” The opportunity cost of
staying at your current job is giving up the new offer; if the new job pays
more, you’re costs (bottom line) go up.
18
Improving a Bargaining Position
• Discussion Question: When is the best time to buy a car?
• Hint: Remember, car salesmen are generally paid a commission for the sales
they make.
• Discussion Question: How can mergers or acquisitions improve
bargaining power?
19
Merger bargaining example
• A Managed Care Organization (MCO) markets its network to
an employer
• Network value is $100 if it contains either one of two local hospitals
• But the value rises to $120 if it contains both
• And there is no value without at least one of the hospitals
• The gain to the MCO from adding either of the hospitals to
its network when it already has the other is $20
• Nash bargaining solution predicts this is evenly split
• So, each hospital gets $10 for joining the MCO
• But if the hospitals merge and bargain together,
• The MCO can no longer drop one of the hospitals, so the gain from striking a
bargain with the merged hospital is the full $120
• The gain is evenly split in the Nash bargaining solution
• The merged hospitals thus receive $60, a post-merger gain of $40
20
Health care mergers
• In Rhode Island in 2003, Blue Cross Blue Shield (BCBS, the health
insurance company covering state employees) hired PharmaCare to
provide pharmaceutical services.
• PharmaCare created a network of retail pharmacies willing to sell drugs
to state employees at discounted rates.
• The previous contract had allowed employees to buy from any pharmacy but
was considerably more expensive.
• In the new PharmaCare contract, 4 retail pharmacies were excluded from the
plan. These 4 firms lobbied RI legislature to include them in the new plan and
offered to provide the same discounted price but PharmCare declined their
request to join.
• Pharmacare maintained that allowing the other stores to join would
eliminate the savings generated by having a restricted network.
PharmaCare’s bargaining position would deteriorate.
• Many politicians, though, like “freedom-of-choice” bills that would open any
pharmacy willing to meet the negotiated prices.
21
Title?
• Under the 2002 CHAOS (Create Havoc Around Our System) plan, flight
attendants threatened to either stage a mass walkout for several days
or to strike individual flights of Midwest Express, with no advance
warning to either customers or management.
• Midwest Express reacted by cancelling all flight attendant vacation,
and threatened to lock out any employee who participated in the
strike
• Flight attendant union promised funding from its strike fund to
support any attendant who ended up locked out.
• The biggest strength of the union’s threat was that it could be
effective without full implementation.
• The threat of random strikes was enough to push passengers to other
airlines.
• After 30 days of CHAOS, the union successfully negotiated a new
contract.
22
CHAPTER
17 Making Decisions
with Uncertainty
©2018 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part, except for use as permitted in a license distributed with a certain product or service or otherwise on
a password-protected website for classroom use. ©Kamira/Shutterstock Images
? When you’re uncertain about the costs or bene?ts of
a decision, replace numbers with random variables
and compute expected costs and benefits.
? Uncertainty in pricing: When customers have
unknown values, you face a familiar trade-off: Price
high and sell only to high-value customers, or price
low and sell to all customers.
? If you can identify high-value and low-value
customers, you can price discriminate and avoid the
trade-off. To avoid being discriminated against, highvalue customers will try to mimic the behavior and
appearance of low-value customers.
©2018 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part, except for use as permitted in a license distributed with a certain product or service or
otherwise on a password-protected website for classroom use. ©Kamira/Shutterstock Images
2
• continued
? Difference-in-difference estimators are a good way
to gather information about the bene?ts and costs of
a decision. The ?rst difference is before versus after
the decision or event. The second difference is the
difference between a control and an experimental
group.
? If you are facing a decision in which one of your
alternatives would work well in one state of the
world, and you are uncertain about which state of the
world you are in, think about how to minimize
expected error costs.
©2018 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part, except for use as permitted in a license distributed with a certain product or service or
otherwise on a password-protected website for classroom use. ©Kamira/Shutterstock Images
3
TeleSwitch
? A large telecom supplier, TeleSwitch, sold its product
only through distributors.
? In 2000, their largest clients wanted to deal directly with
TeleSwitch – and avoid the middle man distributor.
TeleSwitch was unsure what to do.
• They might lose large customers if they didn’t switch.
• But, they might lose distributors (and their small
customers) if they did.
• There is a lower probability of losing dealers (because they
would have to incur costs to change suppliers)
• But this would have a much larger impact on profit.
? How should we analyze decisions like this??
©2018 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part, except for use as permitted in a license distributed with a certain product or service or
otherwise on a password-protected website for classroom use. ©Kamira/Shutterstock Images
4
Introduction: Undertainty
? This problem illustrates the type of uncertainty that
exist in most business decisions.
? This chapter looks at ways to help deal with
uncertainty and arrive at decisions that will best
profit your firm.
? By modeling uncertainty, you can:
• Learn to make better decisions
• Identify the source(s) of risk in a decisions
• Compute the value of collecting more information.
©2018 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part, except for use as permitted in a license distributed with a certain product or service or
otherwise on a password-protected website for classroom use. ©Kamira/Shutterstock Images
5
Modeling Uncertainty
? To model uncertainty we use random variables to
compute the expected costs and benefits of a decision.
? Definition: a random variable is simply a way of
representing numerical outcomes that occur with
different probabilities.
? To represent values that are uncertain,
• list the possible values the variable could take,
• assign a probability to each value, and
• compute the expected values (average outcomes) by
calculating a weighted average using the probabilities
as the weights.
©2018 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part, except for use as permitted in a license distributed with a certain product or service or
otherwise on a password-protected website for classroom use. ©Kamira/Shutterstock Images
6
Random variables
? Definition: a binomial random variable, X, can
have two values, x1 or x2, with probabilities, p and 1p. The expected value (mean) for a binomial random
variable is:
E[X]=p*x1+(1-p)x2
? Definition: a trinomial random variable, X, can
have three values, x1, x2, or x3, with probabilities p1,
p2, and 1-p1–p2. The mean for a trinomial random
variable is:
E[X]= p1*x1+ p2*x2+(1- p1-p2) x3
©2018 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part, except for use as permitted in a license distributed with a certain product or service or
otherwise on a password-protected website for classroom use. ©Kamira/Shutterstock Images
7
How to model uncertainty
? “Wheel of Cash” example:
• The carnival game wheel is divided like a pie into thirds, with
values of $100, $75, and $5 painted on each of the slices
• The cost to play is $50.00
• Should you play the game?
• Three possible outcomes: $100, $75, and $5 with equal
probability of occurring (assuming the wheel is “fair”)
• Expected value of playing the game is
1/3 ($100) + 1/3 ($75) + 1/3 ($5) = $60
• But, if the wheel is biased toward the $5 outcome, the
expected value is
1/6 ($100) + 1/6 ($75) + 2/3 ($5) = $32.50
©2018 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part, except for use as permitted in a license distributed with a certain product or service or
otherwise on a password-protected website for classroom use. ©Kamira/Shutterstock Images
8
TeleSwitch’s Decision Tree
? The probability of losing customers is 0.6
? The probability of losing distributors is 0.2
Telecom Firm
Sell directly to large customers
Sell only through dealers
(.20) × $30 + (.80) × $130 = $110
(.60) × $100 + (.40) × $130 = $112
Distributors leave
Distributors stay
Large customers leave
Large customers stay
(probability = .20)
Firm profit = $30
(probability = .80)
Firm profit = $130
(probability = .60)
Firm profit = $100
(probability = .40)
Firm profit = $130
©2018 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part, except for use as permitted in a license distributed with a certain product or service or
otherwise on a password-protected website for classroom use. ©Kamira/Shutterstock Images
9
Entry Decision with Uncertainty
? The probability of retaliation (no accommodation) to an entry
decision (as modeled in ch 15) is 0.5
Entrant
Enter
Stay Out
(.50) × $60 + (.50) × $-40 = $10
(.50) × $0 + (.50) × $0 = $0
Incumbent prices high
Incumbent prices low
Incumbent prices high
Incumbent prices low
(probability = .50)
Entrant profit = $60
(probability = .50)
Entrant profit = $-40
(probability = .50)
Entrant profit = $0
(probability = .50)
Entrant profit = $0
©2018 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part, except for use as permitted in a license distributed with a certain product or service or
otherwise on a password-protected website for classroom use. ©Kamira/Shutterstock Images
10
Dealing with uncertainty
? Discussion: How do you respond to an invitation from a
friend to invest in a real estate venture that depends on
uncertain future demand and interest rates?
• Calculate the potential gains and loses based on different
combinations of high and low interest rates and high and
low demand
• Whoever proposed the venture probably presented the best
case scenario (low interest rates and high demand) – and
that is the only combination (of four possible outcomes)
under which you will do well.
• Either don’t invest or find a way that aligns your friend’s
incentives with your own, i.e., he gets a payoff only if the
venture does well.
©2018 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part, except for use as permitted in a license distributed with a certain product or service or
otherwise on a password-protected website for classroom use. ©Kamira/Shutterstock Images
11
Uncertainty in Pricing
? Uncertainty in pricing arises when the demand for a
product is unknown.
? To model this uncertainty, classify the number and type
of potential customers. For example:
• High-value consumers willing to pay $8
• Low-value consumers willing to pay $5
• Suppose there are equal numbers of each consumer group
? Discussion: If MC= $3, what is optimal price?
• By pricing high, you would earn $5 per sale each time a
high-value costumer shops – or %50 of the time
• By pricing low, you would earn $2 per sale but would be
able to sell to both high- and low-value costumers – 100%
of the time.
©2018 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part, except for use as permitted in a license distributed with a certain product or service or
otherwise on a password-protected website for classroom use. ©Kamira/Shutterstock Images
12
Uncertainty in Pricing (cont’d.)
? Answer: Price High
Pricing Decision
Price High
Price Low
(.50) × $5 + (.50) × $0 = $2.50
(.50) × $2 + (.50) × $2 = $2
Get high-value customer
Get low-value customer
Get high-value customer
Get low-value customer
(probability = .50)
Profit = $5
(probability = .50)
Profit = $0
(probability = .50)
Profit = $2
(probability = .50)
Profit = $2
©2018 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part, except for use as permitted in a license distributed with a certain product or service or
otherwise on a password-protected website for classroom use. ©Kamira/Shutterstock Images
13
Price Discrimination Opportunity
? If you can identify the two types of customers, set
different prices to each group, and prevent arbitrage
between them, then you can price discriminate.
• Price of $8 to the high-value customers
• Price of $5 to the low-value customers.
? Discussion: When buying a new car, sales people
discriminate between high- and low-value customers.
How do they do this?
? Discussion: What can you do to defeat this?
©2018 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part, except for use as permitted in a license distributed with a certain product or service or
otherwise on a password-protected website for classroom use. ©Kamira/Shutterstock Images
14
Natural experiments
? To gather information about the benefits and costs of
a decision you can run natural experiments.
? Natural experiment example: A national restaurant
chain
• A regional manager wanted to test the profitability of a
special holiday menu
• To do this, the menu was introduced in half the restaurants
in her region.
• In comparing sales between the new menu locations and
the regular menu locations (the control group) the manager
hoped to isolate the effect of the holiday menu on profit.
©2018 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part, except for use as permitted in a license distributed with a certain product or service or
otherwise on a password-protected website for classroom use. ©Kamira/Shutterstock Images
15
Natural experiments (cont’d.)
? This is a difference-in-difference estimator. The first
difference is before vs. after the introduction of the menu; the
second difference is the experimental vs. control groups
? Difference-in-difference controls for unobserved factors that
can influence changes
? The manager found that sales jumped during the holiday
season – but the increase was seen both in the control and
experimental groups—both increased by the same amount.
? The manager concluded that the holiday menu’s popularity
came at the expense of the regular menu. So the holiday menu
only cannibalized the regular menu’s demand and didn’t
attract new customers to the restaurant.
©2018 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part, except for use as permitted in a license distributed with a certain product or service or
otherwise on a password-protected website for classroom use. ©Kamira/Shutterstock Images
16
Natural experiments (cont’d.)
? Natural experiments can be useful in many different
contexts.
? When the FTC looked back at a 1998 gasoline
merger in Louisville, they used their own version of
a difference-in-difference estimator.
• Three control cities (Chicago, Houston, and Arlington)
were used to control for demand and supply shocks that
could affect price.
• The first difference was before vs. after the merger; the
second difference was Louisville prices vs. prices in control
cities– this allowed the FTC to isolate the effects of the
merger and determine its effect
©2018 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part, except for use as permitted in a license distributed with a certain product or service or
otherwise on a password-protected website for classroom use. ©Kamira/Shutterstock Images
17
1998 LouisGasoline Merger
©2018 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part, except for use as permitted in a license distributed with a certain product or service or
otherwise on a password-protected website for classroom use. ©Kamira/Shutterstock Images
18
Minimizing expected error costs
? Sometimes, when faced with a decision, instead of
focusing on maximizing expected profits (benefits minus
costs) it can be useful to think about minimizing
expected “error costs.”
? This approach is helpful when one alternative would
work well only under certain conditions, and you are
uncertain about whether the conditions hold.
• For example, “should we impose a carbon tax?”
• If global warming is caused by human activity then a carbon tax
will help reduce it.
• But if global warming is not caused by human activity, then a
carbon tax would only reduce economic activity and would not
cool the Earth.
©2018 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part, except for use as permitted in a license distributed with a certain product or service or
otherwise on a password-protected website for classroom use. ©Kamira/Shutterstock Images
19
Error costs (cont’d.)
? The two global warming alternatives can be modeled by:
Carbon Tax
No Tax
GW is caused by human activity (p) 0
(p) x (error cost II)
GW is not caused by human
activity (1-p)
0
(1-p) x (error cost I)
• Type I error is the failure to tax when global warming (GW) is caused by human activity.
• The Type II error is the implementation of a carbon tax when global warming (GW) is
not caused by human activity.
• The optimal decision is the one with the smaller expected error costs, i.e. Tax if (1p)*Cost(Type I) < p*Cost(Type II) • This type of analysis is especially useful for balancing the risks associated with pricing errors (over- v. under-), e.g., for airlines, hotels, cruise ships; as well as production errors (over v. under) ©2018 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part, except for use as permitted in a license distributed with a certain product or service or otherwise on a password-protected website for classroom use. ©Kamira/Shutterstock Images 20 Risk versus uncertainty ? Risk is how we characterize uncertainty about values that are variable. • Risk is modeled using random variables. ? Uncertainty is uncertainty about the about the distribution of the random variables. • E.g., which probabilities should be assigned to the various values the random variables can take? ? This difference is critical in financial markets. Risk can be predicted, priced and traded – people are comfortable with risk. Dealing with uncertainty is much more difficult. ? Mistaking risk for uncertainty can be a costly mistake ©2018 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part, except for use as permitted in a license distributed with a certain product or service or otherwise on a password-protected website for classroom use. ©Kamira/Shutterstock Images 21 IndyMac: Risk vs. Uncertainty ? Risk never went away, investors were just ignoring it ? Black Swans & fat tails I have nothing against economists: you should let them entertain each others with their theories and elegant mathematics, [But]…do not give any of them risk-management responsibilities. —Nassim Nicholas Taleb ©2018 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part, except for use as permitted in a license distributed with a certain product or service or otherwise on a password-protected website for classroom use. ©Kamira/Shutterstock Images 22 Dealing with uncertainty ? Uncertainty is unavoidable. So to cope with uncertainty in decision making, gather more or better information. • Best Buy has used dispersed sets of non-experts to predict future variables, such as a holiday sales rate. • Google uses internal prediction markets to generate demand and usage forecasting. ? The US Marines advise: • Because we can never eliminate uncertainty, we must learn to fight effectively despite it. We can do this by developing simple, flexible plans; planning for likely contingencies; developing standing operating procedures; and fostering initiative among subordinates. ©2018 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part, except for use as permitted in a license distributed with a certain product or service or otherwise on a password-protected website for classroom use. ©Kamira/Shutterstock Images 23 Risk versus uncertainty ? Part of the housing crisis can be attributed to an error in translating uncertainty to risk through a mathematical formula created by David Li. • The formula was designed to measure the correlation between returns of various assets that made up collateralized debt obligations (CDOs). • But there was uncertainty about how one asset’s failure would related to that of another asset. There was also a lack of historical data about relationships among the underlying assets. ? Li’s solution was to use past credit default swap (CDS) prices as an indication of correlation returns (clever but imperfect). • CDS data came from a time when housing prices were on the rise, and the correlation changed during a period of decreasing pri