Market Failure
Market Failure
FREE RESOURCES
EXAM RESOURCES
PREMIUM
Market failure guide: all six types, diagrams and model answers
A market can be busy every day and still produce the wrong amount of a good.
Imora is a small town on the coast, and its cement market is busy. Builders buy cement every day at the market price. Yet the dust and noise from making cement fall on the households near the works, and those households are not part of any sale.
Imora's market for computer courses is busy too. A worker who takes a course also helps the people who work beside them, but none of those people pays towards the course.
Is either market making the right amount? To answer that, you need a quantity to compare it with. Economists use the socially optimal level of output, or the social optimum for short. Output is the quantity produced. It is the quantity there would be if every cost and every benefit to society were counted, not only the costs the buyer and the seller bear and the benefits they receive.
The dust is a cost to society that nobody in the cement market pays. Builders and the works leave it out of their choices, so they keep trading some cement that costs society more than it is worth to builders. So Imora makes more cement than the social optimum. The help to other workers is a benefit that nobody pays for. Each trainee counts only their own benefit, so some people decide a course is not worth its price. Added up across Imora, some courses that are worth more to society than they cost to provide are never taken. So Imora trains fewer workers than the social optimum.
Market failure: Too much or too little of a good being produced and/or consumed compared with the socially optimal level of output.
Put simply, a market fails when it makes the wrong amount, even if it keeps trading.
So market failure has a direction: the quantity can be too high or too low. When you use the term, say which way the quantity is wrong, because the two directions have different causes: the dust gave too much cement, and the unpaid help gave too few courses.
When a price leaves out a cost or a benefit, buyers and sellers who choose sensibly still trade the wrong quantity.
Why does a busy market make the wrong amount? A buyer compares the price with the benefit the good gives them, and a seller compares it with their own cost. Neither counts anything the price leaves out. In Imora's cement market, each step causes the next:
Builders use every bag of cement, so why is the cement made past the social optimum a loss? Imora has only so much land, labour and capital: the resources used to make goods. Land includes the raw materials it gives, labour is the work people do, and capital is the machines and buildings. The workers and machines that make the extra tonnes could make other goods instead. For those tonnes, the cost to society, the works' own cost plus the dust, is more than the benefit to builders. So society gives up more to make them than it gains from them, and it would be better off if that land, labour and capital made something else. This is a misallocation of resources: resources used where they add less to society than they could.
When the price leaves out a benefit, the quantity is wrong in the other direction. Nobody pays for the help a trained worker gives the people beside them, so the price of a course does not include it. Each trainee weighs the price against only their own benefit, so some courses are never taken, and Imora trains fewer workers than the social optimum. Those courses would be worth more to society than they cost to provide, yet the land, labour and capital that could provide them stay in other uses. A course's cost is what its resources would add elsewhere, so those other uses are worth less to society.
In the second step nobody behaves badly: the wrong quantity comes from what the price leaves out.
Exam Matters: When you explain a market failure, complete the chain: say which way the quantity is wrong against the social optimum, and what that means for resources. An answer that stops at "the firm pollutes" has not yet reached the failure.
Market failure has five different sources, and one question spots each of them in a case.
There are five sources of market failure, each a different reason why the quantity settles away from the social optimum. The first needs a term of its own.
Externality: A cost or benefit to a third party that is not reflected in the price.
Put simply, an externality is a cost or benefit that falls on somebody outside the deal. That somebody is a third party: neither the buyer nor the seller. The households near Imora's cement works are third parties.
For each source, one question helps you spot it in a case:
Exam Matters: In a case question, name the source the case gives evidence for (a case can show more than one), quote the fact that shows it, and say which way it moves the quantity.
A good can be worth more to society than to the person who buys it.
A worker in Imora pays for a computer course. The course helps her: she can do more skilled work and earn more. It also helps the people she works with, because she can show them how to use new software. They pay nothing towards her course.
Economists name three kinds of benefit in a case like this.
Private benefit: The benefit to the individual or firm directly involved in consuming or producing a good or service.
External benefit: The benefit to third parties, who are not directly involved in consuming or producing a good or service.
Social benefit: The total benefit to society from consuming or producing a good or service.
Put simply, private benefit is what the user gets, external benefit is what other people get without paying, and social benefit is everything society gets. So social benefit = private benefit + external benefit.
The key difference between private and external benefit is who receives it. The worker, who is now a trainee, receives the private benefit, and she pays for the course. Her colleagues receive the external benefit, but they pay nothing. So when the trainee decides whether a course is worth its price, she counts only her own benefit.
So the 20th course is worth $72 to Imora but only $60 to the trainee, and the trainee is the one who decides whether to take it. Where a good has no external benefit, its private benefit and its social benefit are the same.
Making a good can cost society more than it costs the producer.
Costs come in the same three kinds. Imora's cement works makes cement by the tonne (1,000 kilograms). It pays for limestone, the rock that cement is made from, and for fuel, wages and machines. Making cement also spreads dust and noise over the households nearby. They must clean dust from their homes and sleep through the noise, but nobody pays them for it.
Private cost: The cost to the individual or firm directly involved in producing or consuming a good or service.
External cost: The cost to third parties, who are not directly involved in producing or consuming a good or service.
Social cost: The total cost to society of producing or consuming a good or service.
Put simply, private cost is what the producer pays, external cost is the harm that falls on other people without any payment, and social cost is the full cost to society. So social cost = private cost + external cost.
These have the same shape as the three benefits: a private part, an external part, and a social total. The key difference is again who bears the cost. The works pays its private cost, so it counts that cost when it decides how much to make. The works never has to pay for the dust, so it leaves the external cost out when it decides how much to make.
So the 50th tonne costs Imora $10 more than it costs the works.
An externality comes either from making a good or from using it.
Every externality, a cost or benefit to a third party that the price leaves out, comes from an activity. Ask which activity causes the effect on the third party. If it is the making of the good, it is an externality of production. If it is the using of the good, it is an externality of consumption.
In Imora, the dust comes from making cement: an external cost of production. The trainee's help to her colleagues comes from using her course: an external benefit of consumption.
The answer also tells you where the externality counts. An external cost of production adds to the cost of making the good. Each tonne of cement costs society the works' own cost plus the dust, so social cost = private cost + external cost.
An external cost of consumption is different, because nothing extra is spent on making the good. Think of a radio played loudly late at night. The music is worth something to its owner, but the neighbours lose sleep. The lost sleep is a cost that comes with using the radio, so society gains less from the music than its owner does: social benefit = private benefit − external cost. An external cost of consumption lowers the social benefit of a good, not the cost of making it.
The same good can carry both. A car causes fumes at the factory when it is made: an externality of production. It causes fumes and delays on the road when it is driven: an externality of consumption. In an answer, say which one you are analysing.
A cost the price misses leads to too much of a good; a benefit it misses leads to too little.
Marginal: The cost or benefit of one more unit of a good or service.
Put simply, a marginal figure says what the next unit adds. In Imora, the marginal private cost of the 50th tonne of cement is $35: what the works pays to make that one tonne.
Each height on these curves is a per-unit amount: dollars for one more unit. Some of these amounts, such as the external cost, are not a price that anyone pays, so the vertical axis is labelled "Costs, benefits", not "Price".
Two questions place any externality. Making or using decides which pair of curves separates. Cost or benefit decides which way the social curve moves.
Why does the quantity move this way? Buyers and sellers trade until MPB equals MPC, because those are the benefit and cost they themselves receive and bear. When the price misses a cost, they keep trading units that cost society more than they are worth. So the market produces or uses too much. This holds whichever pair of curves has separated: with a cost of consumption, MSB lies below MPB, and the units past the social optimum are worth less to society than they cost.
When the price misses a benefit, third parties gain from each unit, so some units are worth more to society than they cost. Buyers and sellers count only their own benefit and cost, so to them those units are not worth trading. So the market produces or uses too little.
An external benefit of production is easy to miss. Suppose a factory pays guards to watch its building at night. The guards also make the street safer for the shops and homes nearby, which pay nothing for them. So running the factory costs society less than it costs the firm: MSC lies below MPC. The firm counts only its own cost, so it produces less than the social optimum.
With an external cost of production, the market makes more than the social optimum.
In Imora, the dust and noise from making cement cost the households nearby $10 a tonne, at every quantity. So MSC = MPC + external cost, $10 higher at every quantity. Using cement harms nobody else, so MPB and MSB are the same curve.
The market position is where MPB equals MPC, because buyers and sellers count only their own benefit and cost. In Imora, that is 50 tonnes a day at $35. The social optimum position is where MSB equals MSC, because there every cost and benefit is counted. Here MSB is MPB, so that is 40 tonnes, where MPB and MSC are both $40.
Why does the market go past 40 tonnes? Builders are willing to pay $35 for the 50th tonne, and it costs the works the same to make, so both sides trade it. The $10 of dust is in nobody's calculation. Yet that tonne costs Imora $45 and is worth only $35 to its buyer. Each tonne beyond 40 costs society more than it is worth to its buyer. Each tonne up to 40 is worth at least as much to its buyer as its full cost to society.
On the diagram, Qm is the point on the quantity axis at the market quantity, 50 tonnes, and Pm is the point on the vertical axis at the market price, $35. Q is the point on the quantity axis at the social optimum quantity, 40 tonnes, and P is the point on the vertical axis at $40, where MPB and MSC meet.
Welfare loss: The net loss to society from producing or consuming units whose marginal social cost is greater than their marginal social benefit.
Put simply, a welfare loss is how much worse off society is because units were made that cost more than they are worth.
So making tonnes 41 to 50 leaves Imora $50 a day worse off than making 40. On the diagram, the welfare loss is the triangle between MSC and MPB, from Q to Qm. Naming an area by the curves and quantities that bound it shows exactly which area you mean.
Exam Matters: Label the vertical axis "Costs, benefits" and every curve, mark Q* and Qm on the quantity axis, and shade only the triangle between them.
The welfare loss counts only the units past the social optimum, so it is smaller than the total external cost.
An area on a marginal diagram is a total: a height, the amount for one unit, times a number of units. Two areas on Imora's cement diagram are easy to mix up.
The total external cost is the external cost of every tonne made. At the market position, 50 tonnes a day are made, and each causes $10 of dust. So the total is $10 × 50 = $500 a day. This is what the households near the works bear. On the diagram it is the band, the strip of even height, between MSC and MPC, from zero to Qm, the point on the quantity axis at 50 tonnes. Q is the point on the quantity axis at 40 tonnes, the social optimum quantity.
Most of that cost comes from the first 40 tonnes, which are worth making: they carry $10 × 40 = $400 of it. Yet each of those tonnes is worth at least as much to builders as its full cost to society, the works' own cost plus the dust. So making them does not leave Imora worse off.
The welfare loss, $50 a day, comes only from tonnes 41 to 50, past the social optimum of 40. Each of these costs society more than it is worth to builders.
Why does the difference matter? If Imora made only 40 tonnes, the welfare loss would be gone, but the dust would not. The households would still bear $400 a day, because those 40 tonnes would still be made, and they are worth making. A question about the harm to the households asks for the total external cost. A question about making the wrong quantity asks for the welfare loss.
With an external benefit of consumption, the market makes less than the social optimum.
In Imora, a trainee's colleagues gain $12 from each course, at every quantity. This is an external benefit of consumption, so MSB = MPB + external benefit: MSB is $12 above MPB. Providing courses harms nobody, so MPC and MSC are the same curve.
The market position is where MPB equals MPC: 20 courses a week at $60. At 23 courses, a course costs $66 to provide but is worth only $54 to the trainee, so trainees do not buy it. The social optimum position is where MSB equals MSC: 23 courses at $66. Courses 21 to 23 are worth more to Imora than they cost, because each adds $12 for colleagues, but nobody pays for the colleagues' gain, so they are not taken. Beyond 23, each course costs more than it is worth to Imora.
Welfare gain: The net gain to society from producing or consuming units whose marginal social benefit is greater than their marginal social cost.
Put simply, a welfare gain is how much better off society would be if the missing units were bought.
The $240 is a benefit Imora already gets from the 20 courses taken. The $18 is the extra benefit from courses 21 to 23. On the diagram, Qm is the point on the quantity axis at 20 courses, the market quantity, and Q is the point on the quantity axis at 23 courses, the social optimum quantity. The welfare gain is the triangle between MSB and MSC, from Qm to Q.
One rule covers both: the social optimum is where the social curve that has moved meets the curve that has not. In the cement market, MSC lies above MPC, so it meets MPB at a smaller quantity than the market's. In the courses market, MSB lies above MPB, so it meets MPC at a larger quantity.
Each extra car on a busy road slows other drivers, a cost the driver does not pay.
A car journey has a private benefit: getting where you need to go. It has private costs too: fuel, wear on the car and the driver's own time. The driver bears these, so they are part of the decision to travel.
Two costs fall on other people. On a busy road, each extra car slows the cars around it a little. This is congestion: a cost paid in other people's time. The fumes and noise fall on people along the route, and many of them are not travelling at all.
These costs come from using the car, not from making it, so they are external costs of consumption. MSB lies below MPB, because each journey is worth less to society than to the driver. Drivers count only their own benefit and cost, so more journeys are made than the social optimum.
Transport has one feature that can seem odd. Many of the people delayed by congestion are other drivers. They add to the delay themselves. It is still an external cost. Each driver counts their own delay when deciding whether to travel. They do not count the extra delay their car adds for the drivers behind them, because nobody charges them for it.
The size of the external cost depends on the road. On an empty road, one more car slows nobody, so the congestion cost is close to zero. On a crowded road at rush hour, one more car slows many others.
Real Example: A study by the World Bank, an international organisation that lends to developing countries, estimated in 2014 that traffic congestion in Greater Cairo costs Egypt about 47 billion Egyptian pounds a year, almost 4% of the value of everything the country produces. Half of the cost is time lost in traffic, and about a fifth is harm to people's health. The study counts everyone's losses together. Each driver counts only their own delay, so the delay their car adds for other people is an external cost of using the roads.
Health care and education help people beyond those who use them, so the market provides too little of both.
In health, a person vaccinated against an infectious disease, one that passes from person to person, gains a private benefit: they are unlikely to catch it. Other people gain too. A person who does not catch the disease cannot pass it to their family, classmates or colleagues. These people did not pay for the vaccination, and nothing in its price counts their gain. This is an external benefit.
The more people are vaccinated, the harder it is for the disease to spread. Once enough people are protected, even people who cannot be vaccinated, such as very young babies, are safer.
In education, a student gains knowledge, a qualification and, often, higher pay later. The external benefits are spread widely and come later. A better-trained worker helps colleagues work better. A firm can use new machines sooner when its workers can learn to run them. A parent who can read can help their children learn to read.
Both are external benefits of consumption, because the gain comes from people using the vaccination or the education. So MSB lies above MPB. Each buyer counts only their own benefit, so the quantity bought is below the social optimum.
Both external benefits are hard to measure. In health, much of the gain is an illness that other people never catch, and nobody can count an illness that did not happen. In education, much of it arrives years after the student chooses. So their exact size is uncertain, and so is the distance between the market and the social optimum.
Pollution is usually an external cost of production, and the people who bear it may be far away.
Much environmental damage is an external cost of production: waste poured into a river, smoke from a chimney, land left unusable after mining. It works like Imora's cement market: MSC lies above MPC, so the market produces more than the social optimum, and there is a welfare loss.
What is special about the environment is who bears the cost. Three features make it harder to deal with than dust on a nearby street:
So each feature also makes the external cost hard to measure. That matters, because the size of the welfare loss depends on how far MSC lies above MPC.
Not all environmental damage comes from production. Heating a home with a coal stove harms the air through using a good, so it is an external cost of consumption. Ask, as with any externality, whether the damage comes from making the good or using it.
When a bank fails, people who took no part in its decisions lose too.
Banks take deposits, the money savers keep at a bank, and lend money. When a bank decides how much risk to take, it weighs the profit it could make against the losses it could suffer. These are its private benefits and private costs.
What it does not weigh is the harm its failure would do to others. If a bank fails, depositors may lose savings, and firms that borrowed from it may lose the credit they relied on. Other banks that lent to it may make losses. Even people who never dealt with it may find borrowing harder, because other lenders become more careful. None of these people took part in the bank's decision.
These are external costs of producing the bank's service. So MSC lies above MPC, and the quantity of risky lending is greater than the social optimum.
How far does MSC lie above MPC? It depends on how connected the bank is. A bank that has borrowed from many other banks cannot repay them all if it fails, so the losses reach all of them. When those banks lend less, firms and households lose credit too. So the failure of a well-connected bank costs others more than the failure of a bank of the same size that deals with few others.
Real Example: In July 1997, Thailand stopped holding its currency (its money), the baht, at a fixed value. It later closed 56 finance companies that could not pay their debts. Investors then lost confidence in other countries in the region, such as Indonesia and South Korea, which they believed had similar weaknesses. So the effects of a crisis in Thailand's finance sector reached people in other countries who took no part in it.
Most things you buy are private goods: one person's use leaves less for others, and you must pay to get one.
A builder in Imora buys a sack of cement. Once the cement is in the builder's wall, nobody else can use it. The builder also cannot take the sack from the shop without paying. Economists name these two features rival and excludable, and most goods and services you buy have both.
Private good: A good that is both rival and excludable.
Put simply, a private good is one where one person's use leaves less for others, and where nobody can have it without paying for it.
Rival means that one person's use reduces the amount left for others. A seat on an Imora bus is rival: while you sit in it, nobody else can. The cement is rival too, because the cement in one wall cannot go into another.
Excludable means that a person who has not paid can be kept from the good. The bus fare keeps people without a ticket off the bus. A lock, a ticket or a password makes a good excludable in the same way.
Why do these two features matter for selling the good? Because the good is excludable, a seller can charge for it: nobody gets it without paying. Because it is rival, the cement in one builder's sack cannot also go into another builder's wall. To have cement for their own wall, each builder must buy their own. So each buyer pays, and the money from buyers covers the seller's costs. The seller can then make a profit by making the good, so the market supplies it.
A good must be both rival and excludable to be a private good. If either feature is missing, it is not one.
A public good fails both tests of a private good: nobody can be kept from it, and one person's use leaves others no less.
Storms push the sea towards the houses on Imora's shore. A sea wall protects every house behind it. Apply the two tests of a private good to the wall, and it fails both.
Non-rival means that one person's use does not reduce the amount left for others. When the wall protects one house, the house next door is protected just as well. One more house behind the wall leaves every other house with no less protection.
Non-excludable means that nobody can be kept from the benefit, whether they pay or not. The wall cannot protect only the households that paid for it. It holds back the sea for every house behind it.
Public good: A good that is both non-rival and non-excludable.
Put simply, a public good is one that everybody nearby receives once it exists, and one more user leaves nobody with less. Street lighting works in the same way: everybody on the street can see by it, paid or not, and one more walker leaves the others with no less light.
A good must have both features to be a public good. Some goods have one feature but not the other. A crowded road with no toll is non-excludable, because any driver can use it. But it is rival: each extra car takes space that another driver could have used. A film on a subscription service, which you pay for each month, is non-rival, because your watching leaves the film there for every other viewer. But it is excludable: without paying, you cannot watch it. So neither is a public good.
Why does this matter for a market? A seller can only charge for a good if it can keep people who do not pay from having it. A public good is non-excludable, so it is hard to sell, even though the households behind the wall gain from it.
Exam Matters: To explain why a good is a public good, show each feature with a fact about that good: why nobody can be kept from it, and why one more user leaves the others with no less.
Because nobody can be kept from a public good, each person can wait for others to pay, so too little is paid.
A stronger sea wall would protect every house on Imora's shore. Suppose a firm offers to build it if the households pay enough between them. Look at the choice from one household's side.
If the wall is built, it protects this house whether or not this household pays, because the wall is non-excludable. If enough others pay, the household is protected without paying. If too few pay, one household's payment is too small to make the difference, and the wall is not built: paying would lose the household its money and buy it no wall. Either way, the household is better off not paying. Every other household faces the same choice, so many of them wait for others to pay.
Free-rider problem: The problem that arises when people can consume a good without paying for it, so the market provides too little of it or none.
Put simply, a free rider is a person who takes the benefit and leaves others to pay. When many people do this, too little money comes in to pay for the good.
The chain starts with non-excludability, not with non-rivalry. Non-rivalry only means that one more house behind the wall leaves the others no less protected. It gives nobody a reason to avoid paying. If the firm could leave a household that did not pay unprotected, waiting would no longer work, and the household would have a reason to pay.
Exam Matters: When you explain free-riding, give both steps: the good is non-excludable, so each person's best choice is to wait for others to pay. Then say what that does to the money collected and the quantity provided.
The free-rider problem means a market provides too little of a public good, but not always none of it.
Because of the free-rider problem, the households behind a sea wall pay less between them than the wall is worth to them. So a market is likely to provide less of a public good than the social optimum. This is under-provision. But the argument does not show that nothing is provided. The private sector may not provide a public good, but sometimes it does.
When does private provision happen? There are three common cases.
So how much is provided depends on the case. Private provision becomes less likely when more people benefit, and when it is harder to keep people who do not pay from the good.
Real Example: From the 1600s to the 1800s, many English lighthouses were built and run by private owners, with permission from the king or queen. A passing ship could not be kept from the light, so the owners could not charge for it at sea. Instead they charged a fee called "light dues", and the state made sure ships paid it in port. So the owners were paid through something ships could be made to pay for: the second case, with the state's help.
Exam Matters: When you weigh the free-rider argument, use facts from the case: how many people benefit, and whether one of them gains enough to pay alone. Then judge how much is likely to be provided.
In some sales the buyer and the seller know the same things; in others, one side knows more.
Every sale, or transaction, needs some knowledge. The buyer needs to know what the good will do for them, and the seller needs to know what it costs and how good it is. Sometimes the two sides know the same things. Sometimes one side knows more.
Symmetric information: When both parties to a transaction have the same information.
Put simply, the buyer and the seller know the same things, so neither can use knowledge the other lacks. At a fruit stall where buyers can see and touch every mango, the seller knows little that a buyer cannot check.
Asymmetric information: When one party to a transaction has more or better information than the other.
Put simply, one side knows something important that the other side does not. The seller of a used car knows how it was driven and what was repaired. The buyer sees only what a short test drive shows. It can also be the buyer who knows more.
The key difference is whether one side knows more than the other. With asymmetric information, the better-informed side can gain at the other's expense: the car's seller can ask a good car's price for a car that will soon need repairs, because the buyer cannot tell the difference. With symmetric information, neither side has that chance.
Symmetric does not mean complete. Information is symmetric when both sides know the same, even if both know very little. Suppose a builder in Imora sells a homeowner a new kind of roof tile, and neither knows how long it will last. The information is symmetric, because neither knows more. Yet both may still choose badly, because neither knows enough.
When buyers do not know what a good is really worth to them, they may buy the wrong quantity of it.
An information gap is knowledge that is missing when a choice is made. It can be on one side, which is asymmetric information, or on both.
Imperfect market information: When buyers and/or sellers do not have all the information they need to make a rational decision.
Put simply, somebody is choosing without the facts they need to weigh the real costs and benefits of the choice. Most often it is the buyer, who cannot see how good a product is.
Why does a gap change the quantity? Demand shows how much buyers are willing to pay. That willingness rests on what buyers believe a good will do for them. If buyers know the good's real benefit and nobody else is affected, what they are willing to pay for one more unit equals its benefit to society. So demand is then the same as the marginal social benefit curve. In Imora, many households do not know how much a water filter protects them from stomach illness:
A gap can also lead to too much. Some households in Imora believe a vitamin pill stops colds, and it does not. Their willingness to pay is above the pill's real benefit to them, so demand sits above marginal social benefit, and Imora buys more pills than the socially optimal level.
Three things decide how much a gap matters: its size, which is how far buyers' beliefs are from the real benefit; the cost of a mistake, small for cough sweets and large for heart surgery; and how hard the gap is to close. A drink's label shows its sugar, so that gap is easy to close. A patient cannot easily test a surgeon before an operation, so that gap is hard to close.
Exam Matters: When you explain an information gap, follow it to the quantity: say what buyers wrongly believe, which way that puts demand against marginal social benefit, and whether too much or too little is bought.
In healthcare the seller usually knows more, while in insurance the buyer usually knows more.
Healthcare and insurance both have asymmetric information, but the gap runs in opposite directions.
In healthcare, the seller knows more. A patient at Imora's clinic cannot judge whether a treatment is needed, whether a cheaper one would work as well, or whether the diagnosis, the doctor's finding of what is wrong, is right. The doctor can judge, and the doctor is often also paid for the treatment. So the patient's willingness to pay rests on the doctor's advice, not on her own knowledge.
This can move the quantity either way. Where a doctor is paid for each treatment, the doctor has a reason to advise more. A patient who cannot check the advice may accept it. Her willingness to pay is then above the treatment's real benefit to her, so more treatment is bought than the socially optimal level.
The gap can also lead to too little treatment. A patient who cannot judge how serious a symptom is may wait instead of seeing a doctor. A patient who does not trust advice she cannot check may refuse a treatment she needs. Her willingness to pay is then below the treatment's real benefit to her, so less treatment is bought than the socially optimal level.
In insurance, the buyer knows more. A driver knows how carefully she drives, and a household in Imora knows how well its house is locked. The insurer asks questions, and it may ask for a medical test. Even so, it cannot see everything about each buyer's risk. So it sets the premium, the price of the cover (the insurer's promise to pay for a loss), for the average buyer.
For a careful buyer, that premium is more than her own risk is worth. For a careless buyer, it is less. So careful buyers buy less cover than they would at a price that matched their own risk. For careless buyers the premium is low for their risk, so they keep buying cover. A larger share of the insurer's buyers are then careless than if each buyer paid a price that matched their own risk.
Real Example: In India's national family health survey published in 2021, 47.4% of births in private health facilities were by caesarean section, an operation to deliver the baby, against 14.3% in public health facilities. A woman giving birth can rarely judge for herself whether the operation is needed. The figures do not prove that the extra operations were unneeded, because women who use private facilities may differ. But a gap this large fits what the seller's advantage could produce: where the patient relies on the doctor's advice, more operations can be bought than are needed.
Exam Matters: On healthcare, give both directions, too much and too little treatment, each with its own cause. Then weigh which is more likely in the case, for example where doctors are paid for each treatment.
In education and pensions, the choice is made now, but its result arrives years later and cannot be known in advance.
In healthcare and insurance, one side knows more than the other. In education and pensions, the main gap lies between now, when the choice is made, and later, when its result arrives. Often neither the buyer nor the seller can know that result in advance.
A student in Imora chooses a course at seventeen and pays for it now. The benefit, a higher income, arrives over the next forty years. To choose well, the student must guess what the course will be worth in a job market that may change. Nor can the teachers and parents who advise the student see that future.
This leads to two failures. First, too little education is bought. The cost is certain and paid today, but the benefit is uncertain and far away. People tend to give a benefit less weight when it is uncertain and far in the future, so many students and families underestimate it. Their willingness to pay is below the real benefit, so fewer people study than the socially optimal level. Second, some of the education bought is the wrong kind, because students choose courses for jobs that may change before they finish.
A pension, savings that pay an income after you stop working, has the same gap, over an even longer time. A worker of twenty-five must decide how much of her pay to save for an income after she stops working. She cannot know how long she will live or what her savings will earn. The cost of saving is felt now, and having too little in forty years is hard to imagine. So the real benefit of saving is above many workers' willingness to save, and they save less than the socially optimal amount.
Pensions also give no feedback in time to correct the choice. A bad pair of shoes is clear within weeks, and you buy better ones next time. Too little pension saving shows only when you stop working, when it is too late to save more.
Once somebody else pays for a loss, people take less care to prevent it.
A shopkeeper in Imora has no insurance on her shop. If a fire destroys it, the whole cost falls on her. So she takes care: she has the wiring checked and keeps a fire extinguisher by the door.
Then she insures the shop against fire. She pays a firm, the insurer, a regular price called a premium. In return the insurer gives her cover: a promise to pay for most of the loss if there is a fire. A fire would now cost her much less, because the insurer pays most of it. Checking the wiring still costs her the same money. But if there is a fire, most of the loss now falls on the insurer, so each check saves her less than before. So she has less reason to take care, and in the months that follow she may take less.
Moral hazard: When an individual or firm takes more risks because somebody else will bear some or all of the cost of those risks.
Put simply, people take less care to avoid a loss once the loss no longer costs them the full amount.
Why is this a market failure, and not just a choice? Each step causes the next:
If the insurer could see the shopkeeper's wiring, it would raise her premium. It cannot check every customer, so the change is not priced, and the cover is now too cheap for the risk taken. This is why moral hazard is a market failure: the price of the cover leaves out the cost of the extra risk, which falls on the insurer and its other customers, so more risk is taken than at the social optimum.
The timing matters. Moral hazard is a change in behaviour after the cover starts. Buying cover because you already know your risk is high is a different problem. That is an information gap at the moment of the sale: the buyer knows more about the risk than the insurer does.
Exam Matters: To explain moral hazard, link two steps: the cover moves the cost of a loss to somebody else, so the reason to take care weakens. Then say why the price of the cover does not change.
Insurance moves the cost of a loss away from the person best placed to prevent it.
Insurance is useful because it moves a cost. Many households each pay a small premium, and the few who suffer a loss are paid from that money. A loss that could ruin one household is shared across many.
But the person best placed to prevent a loss is usually the insured person: the driver, the shop owner, the household. Insurance moves the cost of a loss away from exactly that person. So some insured people take less care, and claims, requests for the insurer to pay for a loss, become more frequent than the premium allowed for.
Why do careful people pay too? An insurer cannot tell its careful customers from its careless ones, so it charges them all the same premium. That premium must cover the claims the insurer expects to pay. When claims rise, the premium rises for every customer, including those whose care never changed. One person's weaker care raises the price of cover for other people.
Insurers often pay less than the whole of a loss, because full cover would remove the last reason to take care. They leave part of the cost with the insured:
With each one, a loss still costs the insured something, so some reason for care remains. None removes moral hazard completely. With any cover, the insured still bears less of a loss than without it, so part of the reason for care is still gone.
Exam Matters: When you weigh moral hazard in insurance, set the gain from sharing risk against the cost of weaker care, and say how far an excess or partial cover reduces that cost.
A bank that expects a rescue if it fails has less reason to avoid risky lending.
Banks take risks every day. A bank lends the money its savers have put in it, their deposits, to households and firms, and some loans are never repaid. Without any cover, the bank's owners bear those losses, so they have a reason to lend carefully.
Now suppose a bank is so large that its failure would harm a great many people: its savers, the firms that borrow from it, and other banks it owes money to. Its owners may expect the national government to rescue it rather than let it fail. That expectation works like insurance. But nobody bought it, no premium was charged for it, and nobody priced the risk it covers.
This is moral hazard in banking: the expected rescue changes the bank's choice between safe and risky loans, because part of the cost of a bad loan now falls on somebody else. A risky loan pays a higher rate of interest, the price a borrower pays for a loan. If it is repaid, the extra profit goes to the bank's owners. If it is not repaid and the government rescues the bank, taxpayers bear part of the loss. The owners keep all of the gain from a risky loan, but they bear only part of the loss. So riskier lending becomes the sensible choice for the bank's owners.
Each rescue also makes banks expect the next one more strongly. Once a government has rescued one bank, it has shown that it will pay rather than let a large bank fail, so other banks have better reason to expect the same help in the years that follow. So the cost of a rescue includes the extra risk it encourages later.
One failing bank can show three sources of market failure at once. Its failure harms people outside its own deals, an externality; its savers cannot see how risky its loans are, imperfect market information; and the expected rescue weakens its care, moral hazard.
Real Example: In January 1998, during the Asian financial crisis, Indonesia's government announced that it would guarantee all depositors and most other lenders of the banks based in Indonesia. Before that, it had guaranteed only smaller deposits, and savers had been withdrawing money across the banking sector. The guarantee was cover that nobody bought. With every deposit guaranteed, savers had less reason to check how safely their bank lent.
Moral hazard costs consumers, producers, workers and governments, not only the person who took less care.
Moral hazard raises costs for four groups, in insurance and in banking: consumers, who buy goods and services; producers, the firms that make and sell them; workers; and governments. Each group bears a different cost.
Some of these groups never chose the risk. Workers did not decide how much risk their firm took, and they had no cover against losing their job. Governments did not take the risk either, yet they pay for a rescue with taxpayers' money. Customers who pay a higher premium because other people take less care are in the same position. All of them bear a cost of decisions made by other people. That is why moral hazard is a market failure, not a private matter.
Cover itself has benefits. Consumers gain from sharing a risk that could ruin one household. Firms can take on useful projects that would be too risky to bear alone. Moral hazard is a cost of these gains. How large that cost is depends on how much of a loss the cover takes.
Exam Matters: In a Discuss question about the four groups, follow each group's cost through its own link, then weigh the gain from sharing risk against those costs.
In a bubble, buyers expect the price to keep rising, and their buying makes it rise.
Most people buy a flat to live in or to rent out. Some buy one for a different reason: they expect to sell it later for more than they paid. A flat is an asset: something owned that keeps value, like land or gold.
Speculation: The buying of an asset in the expectation of selling it later at a higher price, to make a profit.
Put simply, speculation is buying something to sell it again, not to use it. The profit comes from the rise in price.
Speculation is not harmful in itself. The problem starts when it becomes the main reason people buy, as in Imora, a coastal town, after a year of rising flat prices.
Why does a past rise change what buyers do today? If buyers expect flats to cost more next year, waiting means paying more, and a flat bought now brings a likely profit. So at every price, buyers want more flats than before: demand increases, and the whole demand curve shifts to the right. This is not a movement along the demand curve. Only a change in the flat's own price causes a movement along it.
Then each step causes the next:
Each turn of the loop moves the price further above the flat's underlying value: what it is worth as a home, such as the rent it could earn.
Market bubble: A situation in which the price of an asset rises far above its underlying value, driven by the expectation that the price will keep rising.
Put simply, a bubble is a price held up by the hope of a higher price, not by what the asset is worth to the people who use it.
Exam Matters: When you explain how a bubble arises, name the expectation as the cause of the increase in demand, and go through the loop twice, so that the price is shown rising because it has risen.
A bubble ends in a fall, not a pause, and the fall is usually faster than the rise.
Why can a bubble not simply stop rising and stay where it is? Think of Imora's speculators, who hold flats only to resell them. They bought because they expected the price to rise. If the price stops rising, their reason to hold the flats has gone.
So the loop turns the other way, and each step causes the next:
Only the hope of a further rise held the price above the underlying value. Once that hope has gone, nothing holds the price there.
The fall is usually faster than the rise, for two separate reasons:
The price usually falls below the underlying value before it settles. While the price is falling, buyers expect to pay less next month, so demand stays low, even at prices below the underlying value, while owners keep selling. So the price goes past the value in both directions: far above it on the way up, and below it on the way down.
Exam Matters: When you analyse a burst, keep the two reasons for the fast fall apart: forced selling by people who borrowed, and most owners seeing the fall within days. Each one is its own chain.
Housing bubbles do more harm than bubbles in most markets, because homes are bought with borrowed money and take years to build.
Three features of housing make a bubble there more harmful than in most markets.
When the bubble bursts, the price falls fast, because borrowers are forced to sell and most owners see the fall within days. The burst harms four groups:
Banks lose too, when buyers and builders cannot repay their loans, because a flat sold after the fall raises less than the loan. Many of these losses fall on people who never speculated, such as building workers and families who bought a home to live in.
Exam Matters: When you analyse a housing bust, choose one feature of housing, such as borrowed money, and take it through two or three linked steps to one group's loss.
A share-price bubble sends money to the wrong firms while it rises, and cuts the value of savings when it bursts.
A share is a small part of the ownership of a company. Its owner receives part of the company's profit and can sell the share to someone else. Shares are bought and sold on a stock market, and their prices change every minute it is open.
Shares differ from housing in two ways. First, a share can be sold in seconds, while a flat takes months to sell. Little slows a decision to sell, and an owner who waits risks selling at a lower price after others have sold. So when a share-price bubble bursts, many owners sell together, and the fall can be even faster than in housing.
Second, the share price directs capital: it decides how much money a company can raise by selling new shares, and so where land, labour and capital go. Why does that matter to people who own no shares? Land, labour and capital are limited, so what one company uses is not available to make other goods. A company whose shares are highly priced can sell new shares for a lot of money, and use it to build factories and hire workers. In a bubble, the share price is far above what the company's profits justify, because buyers pay for the rise they expect, not for the company's plans. So the company can raise money for factories that its profits do not justify. Land, labour and capital go there instead of to companies whose plans would use them better. This misallocation of resources happens while the bubble is still rising.
When the bubble bursts, each of the four groups loses in its own way:
The producers' loss can last longest: good projects may wait years for money after the fall has stopped.
Exam Matters: When you compare a housing bubble with a share-price bubble, judge which does more harm by who carries the loss and for how long. Housing losses fall mostly on borrowers and construction; share losses fall on savers and on the money firms can raise.
A market can be busy every day and still produce the wrong amount of a good.
When a price leaves out a cost or a benefit, buyers and sellers who choose sensibly still trade the wrong quantity.