Chapter 17

Many government decisions can be defined as strategic, a term we've alluded to in previous chapters, particularly Chapter 16.

The term strategy is one of those loose words which has come to embellish normal business and government activities. We no longer plan; rather we enter a strategic planning process. We no longer have an employment plan; rather we have a human resource strategy.

Unfortunately this loose terminology detracts from the traditional meaning of strategy, which is about decision-making when there is interdependence between, but not necessarily cooperation among, decision makers. In a strategic situation I make my decision while thinking about what other decision makers will be thinking. They are planning their decision thinking what I will be doing. For an excellent description of strategy in a game theory context, see the writings of Tom Schelling.⁽¹⁾

Obvious strategic situations include many recreational games - chess, poker, football (but not sports of individual achievement, such as high jumping). Less obvious are everyday decisions, such as whether we will go out to dinner on Saturday night:

And so on - an infinite regress. Strategic situations are often intrinsically indeterminate. They are amenable to simulation by game theory, and the game most relevant to public finance is prisoners' dilemma, modelled in this chapter in several contexts.

This chapter illustrates strategy first by outlining the final topic in pricing, oligopoly, for it integrates pricing and strategic theory. It then goes on to look at strategy in a broader context, especially in the contexts of market failure and collective action. The exercises in this chapter are for group learning; it is hard to get an understanding of game theory without going through such simulations.

Oligopoly

Oligopoly is a market situation in which there are few sellers and many buyers. It is more true to life than either monopoly or perfect competition. It describes many markets in Australia, such as airlines, beer and telephones.

In general, pricing decisions in oligopoly come neither from the invisible hand of the market nor from some rational profit-maximizing formula as would be the case in a pure monopoly.

Rather, pricing in oligopoly is a strategic process, best understood by game theory simulation..

The oligopolist must think of her competitors when making a pricing decision, and they, in turn, will be wondering what her decision will be. Of course, under trade practices legislation they are not permitted to communicate, so they must guess one another's thoughts. "Will I raise my price and get more profit? That would be great if my competitors followed me - we'd all be better off. But what if they decide not to follow me and to hang out?" "Will I lower my price and get more market share? Or will they simply follow me down, so that we finish ruining one another in a price war?"

Economists tend to analyze oligopoly in terms of a kinked demand curve - inelastic demand in response to lower prices, and elastic demand in response to higher prices. It's a valid model, but we can see the same through a simple game theory simulation.

We can read about game theory in Schelling and Axelrod, and can study oligopoly theory in textbooks, but the easiest way to understand it is to go through a simulation. The simulation outlined below requires at least two teams, out of hearing range, but able to see each other. There should be a game organizer to count down the rounds.

In a medium-sized Australian city, there are only two pathology laboratories, and they each hold around half the market for blood testing. There is no national health insurance scheme.

One laboratory, The Pathology Corporation, is government-owned and has to maximize its profits on commercial pathology to cross-subsidise work it is doing on non-commercial work, such as AIDS testing.

The other laboratory, Blood Business, is owned by a group of private investors, who are keen to maximize their own profits for commercial reasons.

Both laboratories are currently selling diagnostic tests for a unit price of $200. So long as they sell at the same unit price, each laboratory can expect to keep its half share of the market. But if one laboratory sells at a lower price than the other, it will expand its market share and increase its profits at the expense of the other company. Variable costs are low and fixed costs are high; therefore increasing sales through undercutting is very good for profits. Both laboratories are large, and neither can realistically expect to put the other out of business by undercutting the other's price for a few months.

You are the manager of one of these companies. Each month over the next nine months you will be asked to set the price for your laboratory's tests. Your goal is to maximize your own firm's profits. You are entirely indifferent to the profit of the other firm. Factors such as market share are of no concern to you, except in so far as they can help increase your profits.

You will post prices only in multiples of $100, and neither will post a price above $300 - at higher prices it becomes profitable for customers to use a laboratory in another city. (For this purpose you need three cards for each team, marked $100, $200 and $300, and the organizer will have to make sure that there is a countdown for each round, so that the teams can show their cards to each other simultaneously.)

Your accountant has calculated that the monthly profits of your laboratory will depend on the price you set and the price set by your competitor. For example, if you both post low prices, you will both suffer low profits. Profits can be read off the matrix below.

Monthly Profit Schedule $'000
Their price	Our Price
	100		200		300
	Their profit	Our profit	Their Profit	Our profit	Their profit	Our profit
100	0	0	100	-20	100	-30
200	-20	100	30	30	130	-30
300	-30	100	-30	130	60	60

Thus, if you both post prices of $200, for example, both firms will get a profit of $30 000 for the month. If one posts a price of $100 and the other $200, the undercutting firm will get a profit of $100 000, while its competitor will suffer a loss of $20 000.

Provisions of the Trade Practices Act apply, but they are enforced only where agreement on price-fixing can be demonstrated with hard evidence - that is through communicating before setting price. There is no danger of prosecution for merely anticipating or reacting to the pricing decision of the other firm.

Round	Price		Profit this month		Cumulative profit
Round	PC	BB	PC	BB	PC	BB
1
2
3
4
5
6
7
8
9

Prisoners' Dilemma

The game above is an example of what is, perhaps, the most famous and realistic of all games of strategy - prisoners' dilemma. In that context it teaches about oligopoly, but it is a metaphor for a much wider set of situations.

It is named prisoners' dilemma because of a hypothetical situation in which two people are caught red-handed committing a crime, and after arrest are immediately separated. The prosecution has enough evidence, however, to convict them on only a minor charge, which carries a one year sentence. The prosecutor decides to offer immunity from prosecution if one squeals on the other. The squealing criminal gets off free; the other gets ten years. If they squeal on each other, however, they'll both get a moderately long sentence - five years perhaps.

		The other's decision
		Silence	Squeal
My decision	Silence	-1 -1	-10 0
My decision	Squeal	0 -10	-5 -5

It's obvious that both prisoners will squeal, to their mutual misfortune. That is, unless there's some contract to make the squealing option unattractive. That's built into the Mafia's code of 'Omerata', or silence, basically a severe set of penalties for squealing. Because of this contract a member of the Mafia can trust other members of the Mafia - an example of Roger Fisher's dictum that we should base trust on incentives to be trustworthy rather than on any subjective moral judgement.⁽²⁾

Prisoners' dilemma is not just a two party game. Examples of multi-party prisoners' dilemma (sometimes called social dilemmas) abound.

An example in literature where one party contemplates the decision of many other parties comes in Catch 22, in relation to the dangerous 'milk run' bombing raids. Captain Yossarian is talking to his commanding officer:

A public policy application of the prisoners' dilemma is in international trade. If all countries maintain trade restrictions, then they are not as well of as they would be with open trade. If just some countries liberalize, however, then they run the risk of exploitation. Although a mutually beneficial arrangement is self-evident, the result is stalemate.

This is a case of a multi-party prisoners' dilemma situation, directly relevant to public finance. In Australia the states and territories all suffer the high cost of road wear because of the operation of heavy trucks. States have it in their power to collect registration fees from trucks registered within their borders, or to impose road taxes on diesel fuel. But it is impossible to determine where a truck is domiciled; most transport operators will register a truck where registration is cheapest. And to overcome state diesel taxes, it is easy to fit long range fuel tanks for interstate use.

Each state government would like to raise its registration fees or fuel tax to a level sufficient to cover the cost of road wear. That is economically "rational", and would overcome a major financing problem. But if any one state holds out and does not implement such measures, it will win at the expense of its competitors. If a full cost recovery pricing is achieved with a registration fee of, say, $2000 a truck, then the state which discounts to $500 will get all the nation's registrations and will be ahead of the pack. It is this incentive to defect which ensures no deal is ever struck, and the states all remain in a lose/lose situation, with low taxes and high costs of road wear. Interest groups frame the issue in terms of "state rights", but the issue is one of a lack of any mechanism to enforce a contract of collective rational choice. It is a case of the tyranny of small jurisdictions.

This is a case where all parties can get together, can agree on ends (user charges for roads), means (taxes), and yet there will never be any implementation of the agreement. Means/end agreement is not, in itself, a sufficient basis for agreement when there are strategic advantages from breaking an agreement.

Schelling gives examples of a whole set of cases where there are externalities, which economists would not normally consider to be externalities, where the sum of private choice does not equate to how people would choose collectively. The overwhelming majority of the population may want strict gun control, but if there is not gun control they may wish to carry a gun. Public opinion pollsters often forget, when framing questions, to clarify whether they are asking for views on individual or collective choice.

For example, a "rational" utility maximising person would use leaded gasoline. A car designed for leaded gasoline gets longer engine life, better fuel consumption, and better acceleration. There is a high cost in terms of air pollution, but in a city with 100 000 cars 99.999 percent of that cost is an externality, not borne by the person who decides to use leaded gasoline.

The same "rational" utility maximising person may value clean air more than the extra motoring benefits of leaded gasoline. The more people use unleaded gasoline, the higher is his or her utility.

We can model his or her utility along a dimension which takes into account the collective choice - that is the decision of the population of car owners in the region. In the unregulated market the collective choice will be at position A, even though the position B represents an improved utility for all players. Axelrod points out that in such situations we need some level of coercion, either a tax to offset the utility of the sub-optimal behavior, or a prohibition.⁽⁴⁾ These are forms of social contract, to ensure that collective choice aligns with individual choice. Economists will recognize that what is dealt with here is a treatment of negative externalities. Schelling's contribution has been to match this with game theory, to counter the economic libertarian argument which sees any coercion or restraint on market transactions as an infringement on basic liberty.⁽⁵⁾ Collective choice may be possible only if there is there is the possibility of coercion to make the decision binding.

The region of Id has five cities of around 40 000 each, strung either side of a river.

Id is an old industrial region. Once host to a number of "smokestack" industries, it has seen rapid de-industrialization over recent years, with a consequent rise in unemployment, now running at 15 percent.

The environment is not in great shape, but none of the local city governments have the funds to devote to urgent environment projects - reclamation of disused mine sites, replanting trees, cleaning up abandoned factories, disposing of toxic chemicals, , and stabilizing old dumps.

There is, however, the chance to get investment projects in the area. The region could easily take up to two big projects, with net employment and environmental benefits to the region. Although each project would add to pollution, that addition would be marginal, and revenue from the project would help the city concerned raise taxes to pay for urgent environment work. The benefits would accrue not only to the host city, but also to the region. Past two projects in the region, however, the benefits would start to outweigh the costs.

You are part of a small executive government group in your city, and you must decide whether or not to accept a project in the coming period. The benefits (and costs) depend not only on your decision, but also on what other cities do. Environmental scientists, engineers and economists have developed a schedule of benefits for your city associated with accepting a project, as shown below. These benefits are in cardinal terms; you may wish to think of them in terms of millions of dollars, for they bring all benefits, economic, environmental and social, to a common measure.

Schedule of Benefits for my City
Decision of Other Cities	Decision of my City
Decision of Other Cities	Accept a Project	Reject Project
If no other cities accept	100	0
If one other city accepts	60	-10
If two other cities accept	0	-30
If three other cities accept	-20	-50
If all others accept	-40	-80

You are at a meeting of representatives of each city, where the developers will be offering projects to the cities.

Round	Accept/Reject	No Others Accepting	Score	Cumulative
1
2
3
4
5
6
7
8
9
10

This game should help remind us that although benefit-cost analysis can produce realistic simulations of the results of projects, they often do so in isolation. Costs and benefits may be changed by the decisions of other parties. For example, if we are establishing an industrial park in our city, we need to think of the plans of competing cities, and they aren't going to share them with us. Situations like this get to be truly strategic, in that we must make our plans in ignorance of the plans of other parties, and they must take our likely plans into consideration. Such situations are indeterministic.

The cases we have modelled are examples of repeated round prisoners' dilemma. Unlike the two prisoners, many real life prisoners' dilemma situations are ones we meet again and again.

We can think of situations such as a refrigerator in a student dormitory. If we all simply look after ourselves, and steal a cold bottle of beer whenever we find it, all the students will keep their beer in their rooms, warm. The refrigerator will stay unused.

If most people are prone to stealing cold beer, but some suckers put their beer in the refrigerator, the thieves will get cold beer, and the suckers will lose. They will soon learn not to keep their beer in the refrigerator.

If the whole group agrees not to steal beer, then the situation may reach a cooperative equilibrium, with a social contract among the group. But it could become unstable. Unless there is guidance from higher authority - e.g. a moral code or sanctions - the situation could unravel, say as the end of the year approaches. The cooperative equilibrium, in itself, is unstable. The uncooperative equilibrium is stable. In everyday parlance, once trust is lost it's very hard to re-establish it.

One problem with prisoners' dilemma situations is the condition of displaced objectives. In the oligopoly game we spell out the rules "Your goal is to maximize your own firm's profits. You are entirely indifferent to the profit of the other firm." But very often in a classroom game and in real life, the objective gets displaced, to one of righting past wrongs, or of getting one-up on the competition. (Once we've gone one-down it's very hard to decide to cooperate in future.)

In commercial situations firms often pursue goals which are more akin to 'getting on top' than to looking after one's self interest. Much of the self-destruction of business in the eighties was fuelled by the philosophy of market share - getting the largest share of the market became the objective which displaced the self interest objective of profit. (In fact that philosophy was advocated by several reputable business schools.) What happens when two or more players are operating with the same objectives? Market share objectives suffer from the fallacy of composition - it's great for one firm to think that way, but if all competitors are thinking the same way ....

Contract and Convention

In prisoners' dilemma situations parties may often agree among themselves to have a contract to make the defection option in prisoners' dilemma unattractive. That contract may include sanctions for defection.

Sometimes the contract may be implicit, or unspoken, like the British notion 'we don't do that sort of thing around here', with the sanction being withdrawal of the ticket to the Royal enclosure at Ascot. This is an example of a social contract. More broadly the term 'social contract', as developed by philosophers like Plato and Hobbes, refers to a whole set of reciprocal obligations among members of a society.

The value of social contract without sanctions is dependent on the existence of ongoing relationships. Also social contract is more easily applied in a small group than in a larger group. It is important to note, however, that there is no automatic mechanism whereby social contract arises. In the absence of outside authority it requires a conscious act of negotiation within the group.

In commercial and government situations specific regulation is usually necessary.

For example it may be easy to wind back the odometer on a used car. If I'm a used car dealer, though I may be as innocent as a newborn lamb, I think that my competitors are winding back their odometers. Like Yossarian in Catch 22, I'd be a damned fool not to wind back odometers myself. The prisoners' dilemma situation removes my option to act ethically.

Most other dealers may feel the same way - we don't want to cheat, but we know the incentives on everyone else to do so are high. If we want to be free not to cheat we need some combination of regulation and sanctions. In some situations industry self regulation may be possible. Provided expulsion is a sufficiently strong sanction, it may be possible, say, to get a certificate of compliance, issued by an industry association. Trademarks and green labels come into this category. When internal controls are weak, then external regulation by a higher authority is the only feasible option. Economic libertarians argue that all economic regulation removes freedom; they rarely look at the prisoners' dilemma situation in which a lack of regulation reduces the freedom of the ethical supplier.

Establishment of conventions is another way in which governments or other parties may intervene to help markets work.

Conventions differ from contracts in that contracts are established to overcome any incentive to defect in prisoners' dilemma situations. There is a class of situations in which there is no incentive to defect, in which all sides want to cooperate, but they need some guidance. Such situations call for the establishment of conventions.

Schelling describes a situation in which a community builds roads for itself, but, being libertarian, lays down no rules about which side of the road to drive on, or who has precedence at intersections. In order to use the roads safely the community needs road rules.

This is a metaphor for rules, regulations, standards and other instruments to help compliance where compliance is in everyone's interests. Color codes for electrical wiring, sizes of doorways, air traffic rules, communication protocols for computers are all cases in point. Sometimes conventions may arise in an unregulated market, especially where there is industry dominance (e.g. computer operating systems). Some conventions, like the convention that when a telephone call is disconnected the caller re-initiates the call, have arisen in the antiquity of social custom. In some situations it may be necessary for a central authority to make a unilateral decision, especially when it is desirable to change a convention (e.g. metrication)..

In government organizations over the last ten years or so devolution has been very much in fashion. Often there has been little thought to the setting of conventions in this process. For example, in Victoria, the Department of Health allowed, in the name of community responsiveness, each community health center to go its own way in financial management. As a result there is no uniform chart of accounts and therefore no comparative performance data to set benchmarks. Similarly in most states public hospitals have no central purchasing standards in relation to medical supplies like sterile products. The result is considerable inefficiency and waste. Accounting and purchasing standards are useful conventions which help in financial management.

Notes

1. Thomas C Schelling The Strategy of Conflict (Harvard University Press 1980), Choice and Consequence - Perspectives of an Errant Economist (Harvard University Press 1984), Micromotives and Macrobehavior (W W Norton 1978)

2. Roger Fisher and Scott Brown Getting Together - Building a Relationship that Gets to Yes (Houghton Mifflin 1988)

5. For a typology of situations in which collective choice demands restriction on individual choice, within a liberal framework, see Arthur Okun Equality and Efficiency (Brookings 1975)

Chapter 17 - Strategy and Public Management

Introduction

Oligopoly

Prisoners' Dilemma

Contract and Convention

Notes