Week 3

Session 7: Economics
Today’s session gives answers to the question: how do firms behave under imperfect competition?

Imperfect competition

Under perfect competition and monopoly, we could ignore how firms interact. In an oligopoly we can
no longer ignore such interactions.

Interaction among firms create interesting issues.

Characteristics of an oligopoly:

• High barriers of entry

Example:

2 firms, A and B, produce the same product:

1. If firm A changes its output, the optimal output of firm B will change. This may in fact change
what firm A should do.
2. Through which variables do firms interact (price or quantities)?

There will be not a single answer to these questions → there are several approaches. In a practical
application, you have to decide which is the most applicable.

Nash equilibrium

Nash equilibrium: A simple and robust way to think about outcomes in situations with small number
of players.

The Nash equilibrium is the outcome that is obtained when each firm does its best, taken as given the
behavior of other firms.

Once equilibrium is reached, no one wants to change it.

In game theory, (we talked about in philosophy of science), we also have the concept of collectively
optimal. A collectively optimal situation is a strategy that provides the best payoff for a player in a
game.

The Nash equilibrium can be worse than the collective optimum, but not better.
In this case player B will choose to defect as player B wants 0 years in jail. Player A will also defect as
this person also wants 0 years in jail. The Nash equilibrium will be both defect, so both 5 years in jail.

Collectively optimum is both 1 year in jail.

The Nash equilibrium considers uncoordinated choices among firms. As we have seen, these choices
don’t necessarily result in the best outcomes for firms.

Firms can do better by colluding: agree on their strategies.

Optimal collusions: Agree on monopolistic outcome (to maximize join surplus) and split the
revenues.

In contrast to the Nash equilibrium, collusion is unstable. Each firm has an incentive to cheat and
deviate from strategies.

Firms have 2 ways to compete with each other:

- They can compete with each other by setting prices (which results then in quantity
demanded)
- Or by setting quantities they want to produce (which results in a market clearing price)

Bertrand-competition

Firms competing through price is called Bertrand-competition. In this case competition is very
intense and prices will be driven down to equal each firm’s marginal cost.

Firms want to undercut their opponents. They do this by decreasing their prices until they reach their
marginal costs. Thus firms with the lowest marginal cost, will have an advantage.

Cournot competition

Cournot competition: Firms competing through setting quantities. When there are capacity
constraints, firms are more likely to use Cournot competition.

This type of competition is less aggressive and prices will still be above marginal costs.

Competitiveness of an industry depends usually on whether firms are competing via price or
quantities.
Stackelberg competition

Stackelberg competition: Is a form of sequential competition where firms make choices before or
after their competitor.

First-mover advantage: a firm's ability to be better off than its competitors as a result of being first to
market in a new product category.
Session 8: Economics
Today’s session is about information incompleteness.

So far in the course we have assumed that everyone knows everything. For instance, all buyers know
exactly the quality of the good they are buying. In reality, there’s incomplete and asymmetric
information.

For instance, a firm doesn’t know the productivity of a potential hired employee. Or buyers don’t
know the quality of a product they bought.

Asymmetric information: One party knows more than the other party during a transaction.

Asymmetric information

In real life, transactions often occur under incomplete information. One or more parties to a market
cannot determine with certainty all of the important attributes of a market.

The consequence of this is asymmetric information: an imbalance of information across participants

Asymmetric information creates 2 types of problems:

- Moral hazard
- Adverse selection

Adverse selection

Adverse selection is a market process where buyers or sellers are able to use their private knowledge
of the risk factors involved in the transaction to maximize their outcomes (at the expense of other
parties).

For example: In the insurance industry, one person gains health insurance at a cost that is below their
true level of risk. We have “good” (plums) and “bad” risks (lemons). Good risks are people who don’t
often use their insurance and “bad” risks are people who often need the insurance.

One person knows he is going to get sick so he will get a health insurance. The person that knows he
is going to get sick is a “bad” risk. The insurance will lose a lot of money because of this person.

People who never get sick will have to make up for these high costs. The monthly payment goes up
for all people (whether they’re good or bad risks).

The “good” risks (people who never get sick) will cancel the insurance as it costs them a lot of money.
This means that only the “bad” risks (people who often get sick) pay for the insurance and this is not
doable for the insurance to finance all people.

Adverse selection is a situation where there are stronger incentives for “bad” types of a product to
be involved in a transaction than “good” types of the product.

The information asymmetry also hurt those with more information; because the agents with less
information are aware of the problem and adjust their behavior to it.

Example of adverse selection and information asymmetry in the car industry

Asymmetric information arises because the buyer of the insurance has more information (he knows
how he drives).

This creates adverse selection: Bad drivers have higher incentives to insure. This drives up the cost of
the car insurance, which in turn makes it undesirable for good drivers to insurance (since they’ll have
to pay a lot of insurance to cover the bad drivers).

To counter this issue, there is a system called the bonus-Malus system. You can build up claim-free
years which results in a discount on your insurance (so you pay less). Now people want to drive safe
in order for them to pay less insurance.

Mitigating adverse selection in insurance

• Group policies: tying insurance to employment removes the link between the individual’s
riskiness and the decision to purchase insurance
• Screening: detailed questionnaires, health exams, driving records
• Denying coverage: insurers try to deny coverage to individual with certain risk factors or pre-
existing conditions.

Expected value

In the car industry a seller values plums at 8000 dollars and buyers value them at 10000 dollars. Thus
plums will be sold at 9000 dollars. Lemons have no value (0 dollars). Buyers know 50% of the market
are lemons and 50% are plums. To calculate the expected value:

50% of 10000 dollars + 50% of 0 dollars = 5000 dollars

This is less than the value of a plum to sellers ( 8000 dollars). So owners of plums won’t offer their
cars and only lemons would be offered.

Moral hazard

Moral hazard arises from asymmetric information. It arises when one side in the transaction cannot
observe the behavior of the other side in the transaction.

One person lacks the incentive to guard against financial risk due to being protected from any
potential consequences.

Example

Once being insured, a driver becomes less careful since he’s insured ‘anyways’.

How to mitigate moral hazard

• Deductibles in insurance contracts

Principal-agent problem is a specific type of moral hazard. Principal-agent problem: A principle hires
an agent to perform a task but the principle cannot observe the actions of the agent. Also, incentives
between the agent and the principle differ.

Example

An employer (principle) wants the employee (agent) to work as hard as possible for the company
whereas the employee (agent) just wants to earn its money and doesn’t really care about the success
of the business.

Signaling

Signaling is a solution to the problem of the principal-agent situation. Signaling: An agent with high
quality undertakes action to signal its quality.

For this to work, the cost of the signal has to be lower for high quality agents.

Example

A university degree is a credible signal to employers that a person is productive. For less-productive
students this signal would be too costly (high risk of not passing courses and lots of hard work).
Session 9: Economics
This last session of the course is about the game theory.

Game theory

In many real-life situations, people interact. They choose actions and these actions influence each
other. This session answers the questions:

• What actions should an individual player in a game take?

Game theory provides tools to answer these questions.

A “game” has 3 elements:

• Players: A player must decide his or her actions based on the actions of others.
• Strategy: The action taken by a player. Strategies can be simple or complex. They also depend
on the actions of other players.
• Payoff: The outcome a player receives from playing the game. The payoff to one player
generally depends on the actions of other players. Otherwise, there’s no strategic interaction.

Payoff matrix

Today we look at games from many different viewpoints instead of only analyzing the Nash-
equilibrium.

We want to predict outcomes in games. We thus have to find optimal strategies for the players.

Dominant strategy: Strategy that gives highest payoff regardless of the strategy of the opponent. In
this image above, both firms’ dominant strategy is ‘develop technology’.

Dominated strategy: Strategy that never gives the highest payoff.

A player will always choose a dominant strategy, but never the dominated strategy.

In more complex games, we can use the concept of dominant and dominated strategies to simplify
the game.
Solving for the Nash-equilibrium

In many situations, using dominant and dominated strategies doesn’t help us to solve the game.
Other concept: Nash-equilibrium.

Recall: In Nash-equilibrium each player chooses optimal action taken as given what the other player is
doing.

How do we find the Nash-equilibrium?

1. Find the best responses of player 1 to each action of player 2

Mixed strategies

We have so far considered “pure” strategies: player 1 chooses an action with 100% likelihood given
what the other player is doing. Sometimes, this isn’t optimal.

In such cases, players “mix” their best response. That is, they play a certain response only with a
certain probability.

Minimizing losses

We have so far assumed that both players are rational and know of each other that they’re rational.
However, there may also be situations where you know that your opponent behaves irrational.
in particular, you may face an opponent who behaves erratic (unpredictable behavior).

In this situation a conservative player may want to minimize losses.

Maximin-Strategy: Choose the strategy with the largest minimum payoff across all possible actions of
your opponent.

Sequential games

Sometimes players make decisions one at a time. These types of games are called sequential games.
In sequential games, one player moves first and other players observe action before making their
decisions.
The matrix approach we used for simultaneous-move games doesn’t work for sequential games as it
ignores the timing of action. We thus use an extensive form or decision tree to analyze sequential
games. This representation takes into account the sequence of actions.

We solve sequential games with backward induction: we first solve the last stage, then the second-
but-last and so on until we arrive at the first stage.

Example of such a decision tree

Side payments

A simple strategic move is to offer a side payment: A bribe that influences the outcome of a strategic
game.

A “side” payment because the payment is outside the normal game. For instance: When two parties
involved in a transaction exchange money that doesn’t belong to the transaction, the exchange is a
side payment. The goal is to convince the player to take part in a transaction.

Commitment problems

Side payments may not be credible. Players may have an incentive to back out of the side payment
(no problem in pass or play game, but in other contexts it is).

Commitments that are not enforceable are referred to as noncredible threats: A threat in a
sequential game that a player wouldn’t actually carry out, since it wouldn’t be in the player’s best
interest to do so. Thus this is an empty threat. To avoid this dilemma, a player must make a credible
commitment.