Market Failures#

The assumption we previously made was that free markets naturally converge to an optimal outcome. However, this clearly is not always the case: market failures arise when markets fail to converge to an optimal outcome.

Externalities and Public Goods#

Definition 25

Externalities are side effects on individuals other than the buyer or seller.

Definition 26

Public goods are goods that are accessible and usable by everyone but not owned by anyone.

For example:

  • The environment is a public good, and activities that harm the environment (plane tickets) have negative externalities.

  • WiFi bandwidth is a public good, and one person’s usage imposes a negative externality of slower connections on others.

Two possible solutions.

  1. Pigouvian tax: a tax on transactions proportional to the externality they generate. (e.g., carbon, bags, cigarettes)

  2. Coasian bargaining: auction off public goods and rely on the Coarse Theorem to achieve social efficiency.[1]

Transaction Costs#

Also known as market frictions. Sometimes, transactions themselves can be costly to execute leading to mutually beneficial transactions not being made. For example:

  • Agent one has a boat worth \(9,900\) to them and agent two is willing to pay \(10,000\).

  • Agent one can sell the boat to agent two and both would be happier off.

  • However, if there is a \(9\%\) sales tax on boats, this transaction would not occur.

Transaction costs do not necessarily have to be monetary:

  • A hungry student is willing to buy an apple for \(10\) while other students have applies worth \(1\) to them.

  • Other students can sell the apple to the hungry student and both would be better off.

  • However, there are time/information costs for the hungry student to run around and ask other students if they have an apple on them.

Computers, especially the Internet and search have done a lot to reduce transaction cost (e.g., Lyft, Venmo, Tinder).

Market Thickness#

Definition 27

Informally, a thick market has many buyers, a thin market has few buyers.

A market is thick (thin) if there are many (few) buyers and sellers. In thick markets, sellers and buyers have lots of options which makes prices close to competitive equilibrium. In thin markets, there are few buyers and if buyers come they face monopoly prices. Thick markets generate welfare-maximizing outcomes while thin markets do not.

Why are monopolies bad? Monopoly quantity sold is lower than quantity sold in competitive equilibrium; prices are higher and consumer surplus is lower.

First welfare theorem holds with reserve prices to cover costs. However, sellers setting reserve prices is not strategyproof.

Proof. If reserve prices are nonbinding, then sellers can increase reserve prices by \(\epsilon > 0\) and still sell. If prices are binding, then the good is sold at the reserve price which is equal to cost and the seller makes zero profit. Assuming the necessary assumptions, sellers can increase reserve prices by \(\epsilon\) and make strictly positive profit, which is a profitable deviation.

Some ways to get around this:

  • As markets become far from monopoly, it converges to a seller-strategyproof mechanism (strategyproof-in-the-large). However, this requires fungible goods in a thick market.

  • If the seller doesn’t know buyers’ valuations, then it is difficult for sellers to set a reserve price. However, sellers collect lots of data.

How can markets be thickened?

  • Encourage buyers and sellers to join and stay;

  • Merge markets (larger kidney donation markets have typically led to better outcomes);

  • Batch transactions: wait a bit for more buyers and sellers to join before running transactions (e.g., ride-share, kidney exchange).

Timing Issues#

Market Unraveling#

Committing to contracts too early

In certain markets, one side has incentives to make matches earlier and earlier (to get good students before competitions hire them). As a result, before introduction of the centralized matching mechanism for matching doctors and residencies even first year med students would be matched. However, whether or not a first year med student and a residency would be a good match is a bad predictor of whether or not the student and the residency is still a good match four years later.

Exploding Offers#

Job offers that require a very short response time. In markets where one side needs to look for other candidates if turned down, short response times can arise. Amplified when:

  • if there is a limited time window for transactions due to regulatory issues;

  • if one side of the market can’t absorb variance in the number of matches received.

Impatience#

The third timing issue is not waiting for a market to clear. APPIC rules for psychologists in the 1970-1990’s:

  • All transactions must occur on selection day;

  • Exploding offers are not allowed.

In theory, the result should be similar to Deferred Acceptance. In practice, employers didn’t want to propose to candidates who might cancel right before the deadline leading to earlier and earlier “safe” commits.

Solutions#

How to resolve timing issues? Centralized matching systems will help. However, setting timing rules often fail due to incentives to violate rules being too strong (“Would you like to visit my hospital [wink wink]”?) Rules/norms that allow for accepting and then reneging from exploding offers removes the incentives to cheat and make exploding offers.

See also

Notes on market stability from ECON 136: Market Design.

Information Asymmetries#

Definition 28 (Information Asymmetry)

Informally, the seller knows something that the buyer doesn’t.

Sellers know more about the quality of goods than buyers. As such, buyers might buy a good that they thought was worth it, but later find out that the good is worse than expected. For example, the market for lemons:

  • Market for used cars, some good and some bad.

  • Buyers willing to pay more for good cars than bad cars but do not know which type of car is which.

  • Sellers willing to sell bad cars for less and do know which car is which.

  • Let \(g \in [0,1]\) be the fraction of cars in the market that are good.

  • Let \(h \leq g\) be the fraction of cars that are in good condition and also for sale.

One bad equilibrium: only bad cars are sold at a low price and high cars aren’t worth being sold at that low price.

One good equilibrium: if the portion of good cars is high enough, then the expected value of a car and hence price of a car is high enough to sustain good cars being sold.

Other examples where asymmetric information comes up:

  • Health insurance: buyers know private information about their health status, sellers want to sell insurance to healthier individuals.

  • Clickbait: content creators know if a link is clickbait or not while users do not. Creators what to maximize clicks, users want to avoid clickbait.

Definition 29 (Moral Hazard)

A moral hazards occur when there is an incentive for an agent to risks, because they do not bear the full costs of their actions. For example, insurance might encourage risky behavior.

Example

As a fun (potentially apocryphal) example, Stanford used to give out free bike helmets to freshmen; however, they reportedly stopped doing so because the rate of bike crashes increased, presumably because the helmets emboldened new bikers to take more risks.

Definition 30 (Adverse Selection)

Adverse selection: situations where only the low quality goods/lemons stay in the market.

Possible solutions:

  • Provide more information to both sides (mandatory disclosures, reputation systems);

  • Disallow use of information (universal health care, anti-insider trading regulations);

  • Filter out lemons (require warranties).