Section C Managerial and individual decision problems Asymmetric

Section C – Managerial and individual decision problems • Asymmetric information and risk – Main problems: adverse selection and moral hazard • Applications and case studies: second hand car markets, the supply of health care, managerial incentives, hiring and contracting with employees

Suggested background reading • Allen et al. 2009. Managerial Economics. Norton. Part 7: Chapters 13 -15 • Kreps, D. M. 2004. Microeconomics for Managers. Norton Chapters 18 -19 (16 -17 provide more background) • Frank, R. H. 2008. Microeconomics and behaviour. Mc. Graw Hill. Chapter 6 • Wall, S. , Minocha, S. and Rees, B. 2010. International Business, Pearson. Chapter 6 • Grimes, P, Register, C. and Sharp, A. 2009. Economics of Social Issues, Mc. Graw Hill. Chapter 15 • Rasmusen, E. 2007. Games and Information, Blackwell. Chapters 7 -9

Objectives By the end of this section you should be able to: • – – – Characterise decision making under risk Use detailed examples to explain what is implied by the concept of adverse selection. In the context of your example, discuss how the problems associated with adverse selection can be resolved. Use detailed examples to explain what is implied by the concept of moral hazard. In the context of your example, discuss how the problems associated with moral hazard can be resolved.

Risk • Risk implies a chance of loss – E. g. some exogenous chance of an investment failing as well as some chance it will succeed • Need to incorporate probabilities into the decision problem – Probabilities may be known or only defined subjectively (uncertainty) » How do people decide what to do?

Decision trees Probability =0. 1 Invest in R&D succeeds M+w Chance R&D fails Probability = 0. 9 Dr Punter Don’t invest M-c M Should Dr Punter invest? What would be a sensible decision rule?

Dr Punter’s decision 0. 1 Invest in R&D succeeds M+w Chance R&D fails 0. 9 Dr Punter Don’t invest M-c M Expected value of payoff from investing = 0. 1(M+w) + 0. 9(M-c) Payoff from not investing = M (sure thing) So perhaps should invest if: 0. 1(M+w) + 0. 9(M-c) > M or: w > 9 c …………a simple rule…………. but ignores attitudes to risk

Do attitudes to risk matter? • Which gamble do you prefer: • Gamble A: – Win $100, 000 probability 0. 01 – Win nothing probability 0. 99 • Gamble B: – Win $2000 probability 0. 5 – Win nothing probability 0. 5

Do attitudes to risk matter? • Gamble A: – Win $100, 000 probability 0. 01 – Win nothing probability 0. 99 • Expected value = $1, 000 • Gamble B: – Win $2000 probability 0. 5 – Win nothing probability 0. 5 • Expected value = $1, 000 • If attitudes to risk don’t matter you should be indifferent but are you? – If indifferent you are risk neutral but if you have a preference you are risk adverse or a risk lover • In either case you shouldn’t use the expected value rule • Maybe maximise expected utility instead

Asymmetric or hidden information • For managers and consumers information is important – E. g. information about rivals’ products, information about sellers and buyers of their product – There is a lot of information available (e. g. internet) but if information is imperfect this can lead to problems e. g. if information is asymmetric • one side to an exchange knows more about some important detail than the other – they have an informational advantage • Analytical framework is the principal agent model (agency theory) – The side with the information is known as the agent - informed – The side with limited information is known as the principle – uninformed • they are at an informational disadvantage

Adverse selection • Adverse selection arises when an agent and a principle are involved in a transaction but there is asymmetric information about a fixed condition or characteristic/type – E. g. The quality of a product or an innate characteristic such as ability, intelligence, reliability , attitude to risk • the agent knows more than the principle about a characteristic (i. e. quality) of the agent and; • the information about the agent is relevant to the principle’s evaluation of the transaction between them

Adverse selection • Examples – The riskiness of an investment for a venture capitalist (the principle) due to uncertainty about the effectiveness of new technology employed by an entrepreneur (the agent) – The riskiness of employing a new worker because uncertainty about their innate ability (productivity) – The state of health of someone buying health insurance • If principle offers a contract that is based on expected quality this may only be acceptable to low quality agents ADVERSE SELECTION; bad drives out good

Basic idea underlying the adverse selection problem Principle contracts agent Agent is a certain type/quality agent knows own type but principle does not Outcome depends on type/quality of agent

Adverse selection • As principle doesn’t know agent’s type (e. g. high or low quality) the contract will reflect this uncertainty payment based on expected (average) value rather than actual value e. g. expected profitability of a new venture, expected ability of an employee – Proposed payment therefore less than value of high quality agents so high quality agents likely to reject contract and only lows will take contract; ADVERSE SELECTION - bad drives out good – Principle’s payoff is low unless revises the contract downwards

A game theoretic illustration of the general adverse selection problem P 1 c H AH Chance L AL accept Payoffs Principle Agent high negative reject 0 0 accept low positive reject 0 0 Principle, P: offers agent (A) a contract, c, based on average quality Chance/Nature determines agent’s type/quality: quality is high (H) or low (L); probability of each = ½ so principles expected payoff = ½ high + ½ low Agent, A: knows own type: accepts contract if payoff is positive i. e. value of contract is > 0, rejects contract otherwise (payoff = zero) What kind of agent will accept the contract? What will the principle’s payoff be?

A game theoretic illustration of the general adverse selection problem accept P 1 c H AH Chance L AL Payoffs Principle Agent high reject 0 accept low reject 0 negative 0 positive 0 What kind of agent will accept the contract? Only low quality agents will accept the contract What will the principle’s payoff be? so the principles payoff will be lower than expected payoff - lower than (½ high + ½ low)

Vicious circle of adverse selection • If principle (buyer) understands that only low quality agents (sellers) accept a contract based on average quality, then offered payment (price) will be lowered to reflect this adverse selection – ……. . if there any intermediate quality agents they may withdraw from the market as well – adverse selection gets worse (more adverse) • Market gets thinner and thinner

Moral hazard • Moral hazard: asymmetric information about the action of someone (the agent) that affects the welfare of another person (the principle) – But the choice of action cannot be specified in a contract – the actions of the agent are not completely controllable or observable • there is uncertainty because of ‘noise’ leading to confusion – There is also some conflict between the interests of the agent and principle • The agent may have an incentive to lie about the action taken to detriment of principle = Moral hazard • Key issue is one of incentives – the incentives faced by the agent may be influenced by the principle via the structure of the transaction (e. g. through a contract)

Examples of scenarios in which moral hazard may arise • Insurance: An insurance company sells health insurance to a firm and is concerned that the firm’s employees may take less care over their health (e. g. drinking and smoking) now they have insurance • More insurance examples: – how carefully an insured person drives – How carefully an insured factory guards against fire or theft

More examples of scenarios in which moral hazard may arise • Employment: A car mechanic is hired by the hour to fix a car, and the owner of the car is concerned that the mechanic will take a lot of long tea breaks but claim that the problem was complicated • More generally: An employee has a contract of employment and is paid a fixed hourly or daily wage. The employer worries about the amount of effort or care the worker will exert since either gives the worker negative utility. – The employer cannot observe effort or care but can observe output; but there is no 1 -to-1 relationship between effort/care and output

More examples of scenarios in which moral hazard may arise • Team work: Two students working on a team project worry that the other team member will do very little work - but that the team member will claim that s/he put in a lot of effort but that what they tried to do proved very difficult and time consuming because of problems finding relevant data • Borrowing: How careful an entrepreneur will be with the money loaned from a bank – the loan manager worries that the entrepreneur will gamble with the funds – take too many risks

Basic idea underlying the moral hazard problem Principle contracts agent Agent performs action which principle does not see – agent prefers a different action to principle Outcome depends on action of agent

A game theoretic model of moral hazard in employment contracts • Worker: Action = Level of effort/care. More effort gives negative utility to the worker; Utility depends on wage(+) and effort/care(-) – U = F(wage, effort/care) – Wage is constant e. g. $100 a week • Employer: Output = Q; employer prefers higher Q

A game theoretic model of moral hazard in employment contracts • Uncertainty about the environment and asymmetric information about the worker’s action (choice of effort level) – Environment or state of the world can vary; may be good or bad (e. g. bad if machinery breaks down, mistakes made, accidents happen) - some probability of either • In any given state of the world, more effort/care always leads to at least as much Q as less effort; so employer prefers more effort/care • But no 1 to 1 relationship between effort/care and Q; accidents can happen even if very careful • Low effort in good state of the world leads to a relatively high Q – Employer observes neither state of the world nor the effort/care • Can only calculate expected output based on probability of different states of the world – for given level of effort

Game theoretic model of moral hazard More effort (less utility) P contract A Less effort (more utility) Uncertainty for the principle State of world Good or bad? Output? Agent/worker prefers less effort Principle/employer prefers more effort – output generally, but not always higher Principle never knows what the state of the world was/is or the agent’s effort as there is no 1 -2 -1 relationship between action/effort and output.

Moral hazard Output State of world Agent/worker prefers less effort Principle/employer prefers more effort: 50 More effort (less utility) P contract A Less effort (more utility) Uncertainty for the principle Good Expected output = pgood 50 + p. Bad 40 Good 40 Expected output = pgood 40 + p. Bad 20 Bad With wage constant what effort level do you think the agent/worker will choose? What would you do? 20

Game theoretic model of moral hazard Output State of world 50 More effort P contract A Less effort Uncertainty for the principle Good Bad Good 40 40 Bad Effort doesn’t map cleanly to output so agent has an incentive to make less effort/ be careless but lie: say made an effort but the state of the world was ‘bad’ e. g. agent was unlucky that accident/mistake happened – especially if p. Bad is low 20

Possible solutions to moral hazard • – Motivation/incentives for the agent to perform the action that benefits the principle e. g. – – – • But it may not be possible to contract for compliance e. g. so that a person who takes out health insurance takes care of their health, or a mechanic works hard all day etc. because of : 1. 2. • Material incentives such as promotion, higher pay, a better job, payment by result, piecework Social norms Reputation so that the transaction is repeated Cost sharing and exclusions in insurance measurement, monitoring issues and; uncertainty means that even if the desired actions are taken outcome is not certain Maybe the agent won’t take the job?

Summary • 2 main problems associated with asymmetric information – Adverse selection • Detailed examples: second hand car markets, hiring in labour markets, markets for health care – Moral hazard • Detailed examples: management employment contracts, wage contracts, markets for health care