THE ECONOMIC ORIGINS OF GOVERNMENT

THE ECONOMIC ORIGINS OF GOVERNMENT
By
Jan-Erik Lane (Geneva and Singapore)
INTRODUCTION
When he was struck down by premature death Mancur Olson worked intensively to
find an economic explanation of the omni existence of government in well-ordered
societies. His last book Power and Prosperity: Outgrowing Communist and
Capitalist Dictatorships (2001) outlines his answer to the question of the economic
reasons of government. We will never know more precisely how Olson would have
modelled the collective action reasons that lead people to set up a state. One answer
towards an endogenous theory of government is to be found with Barzel (2002). Let
us in this Chapter employ some game theoretical tools to restate this economic theory
of the state.
An economic theory of government would explain why people set up and support a
political authority, starting from their self cantered reasons. This way of framing the
question of the origin of government sound much like the contract school thinking
that dominated Western political philosophy from Grotius to Rousseau. It also links
up with the most recent developments in game theory, especially evolutionary game
theory. What, then, would lead rational players to pay for a government in the short
run as well as in the long run? Yoram Barzel answers using classical contract theory
combining Hobbes and Locke in an argument that (1) the state is a super Hobbes’
protector whose advantage is economies of scale in enforcing contracts:
Exchange requires agreements, and agreements must be enforced. Initially,
All agreements must have been self-enforced, relying on long-term relations.
Increases in the extent of social interaction enabled third-party enforcement
without the use of force, by elders for example, to emerge. When the state
emerged, it provided third-party enforcement by using its ability in violence.
(Barzel, 2002: 269)
Moreover, this Hobbesian position is combined with (2) that rule of law also emerges
in order to protect the protected against the protector, i.e. the Lockean position:
The need to contain the specialized protectors is never-ending. When subjects
Control their protectors, we expect them to impose constraints, such as the
length of service and on their budget. In addition, we expect subjects to pay
wages to the protectors, rather than allowing them to retain the spoils of war.
(Barzel, 2002: 268)
Let us try to put more substance into these somewhat vague statements forming parts
of a purely economic theory of government by employing the powerful tools of game
theory. What would lead utility maximizing players to set up a government with a rule
of law mechanism to govern it?
2
TRANSACTION COSTS AND COORDINATION
The argument that I will put forward focuses upon coordination failures. The ultimate
source of government is the transaction costs which human beings incur as they
interact in groups. Social existence is impossible without coordination between
individuals. Social coordination opens up the possibility of civilisation, i.e. joint effort
between individuals to secure common ends. Human interaction runs from conflict
and destruction to close or intimate cooperation. Coordination is the process of
interaction through which individuals learn to cooperate. Coordination failure takes
place when individuals clash or they fail to develop mutually beneficial interaction.
Government would under this economic interpretation basically offer a mechanism
which enhances human coordination.
Coordination is the process through which human beings take each other into account,
explore the alternatives of action which are of concern to them and choose a strategy
that leads to a stable outcome which is in their common interest. It has been rightly
suggested that markets constitute a coordination mechanism of immense
consequences for the economy and society. It is one of the major mechanisms of
coordination, which has evolved over time to the benefit of mankind. Simple markets
like the Saturday grocery market build upon the same preconditions as complex
markets like futures’ markets, namely property rights, trust in exchange and
contractual enforcement.
Exchange is a most powerful tool of coordination. Human beings can interact freely
and negotiate the terms of an agreement, which they stick to enforcing it themselves.
However, exchange has two limitations:
- (a) Coordination failures: Here one finds several situations, including that
individuals fail to strike a deal, that individuals make an agreement but it is not
enforced, that individuals renege, or simply that individuals choose the worst
possible outcome.
- (b) Transaction costs: When it is no longer a matter of two persons interacting,
then
transaction costs start running high as groups of people decide about how
to cooperate negotiating the terms of cooperation endlessly.
Government or the state resolves coordination failures and reduces transaction costs.
These are the basic rationales of government in a society. Let us begin with the
analysis of collective action, or group decision-making. Sometimes a group of people
needs to take action in order to safeguard a common interest. Government may
resolve collective action problems by operating public programmes and holding
public assets.
COOPERATION AND COORDINATION FAILURES
Let us make the analysis of coordination as simple as possible by only considering
two individuals who have an interest in maximising their gain from interaction. Thus,
these players may wish to cooperate. However, they may fail to do so when
3
coordination problems beset the interaction. Various coordination failures may be
identified arising in different kinds of interaction. Cooperation is merely one strategy
in the games that will be analysed below. What is crucial is whether cooperation is a
Nash equilibrium, meaning it is the best response of a player towards another player.
The distinction between cooperative and non-cooperative game theory is crucial here.
Cooperation will only be pursued if it is part of a non-cooperative strategy that is
Nash. This assumption makes it possible to analyse how players may bypass Pareto
optimal outcomes and end up in coordination failures.
I) Coordination failures with pure strategy Nash equlibria
Here we have two situations: (a) PD situations, i.e. the two players defect from the
Pareto-optimal solution; (b) Chicken situations. i.e. the two players may push for their
best outcome and thus bring about a Pareto inferior result. The PD game is the most
dismal type of coordination failure, as both players have the dominating strategy of
defection against cooperation:
Table 2.1. The PD Game
Player B
Cooperation Defection
Cooperation 3, 3 0, 6
Player A
Defection 6, 0 1, 1
The value of the game is 6, but the players only achieve 2, which amounts to clear and
expensive strong coordination failure. Two solutions are open:
i) Correlated strategies: Communication between the players could lead to a new
strategy (cooperation, cooperation), but it will not be self-enforced. Actually a number
of correlated strategies could be devised to arrive at the value of the game through
side payments, but they all require a third party to police the agreement. From where
comes this third party and how is he/she to be paid?
Suppose that the PD game is played two times consecutively. Then the value of the
game is 12 but the players only achieve 4 due to the chain store paradox.
Communication and a cooperative solution involving for instance alternation between
cooperation/defection and defection/cooperation would be Pareto superior. But how is
such a correlated strategy to be enforced?
ii) Meta-strategies: If the game is played several times without a fixed time horizon,
then the discount rate of the players will lead them to coordination only if they
discount future cooperation little. The discount rate d in an infinite sequence of PD
games would have to satisfy the following condition in order to favour cooperation
ahead of defection:
3/1- d > 6 + 1/1-d; 3 > 6(1-d) + 1; 2 > 6 – 6d; -4 > -6d; 6d > 4; d > 2/3.
4
If one of the player has a discount rate lower than 2/3, then cooperation is not going to
be the outcome of coordination. The Chicken game is less dismal from the point of
view of coordination. Let us look at its structure:
Table 2.2. The Chicken Game
Player B
Wimp Macho
Wimp 3, 3 0, 6
Player A
Macho 6, 0 -4, -4
The value of the game is 6 but brinkmanship may bring the game to the dismal
solution of – 8. However, the negative outcome is not strictly Nash and thus may be
avoided by the players. Typical of both the PD game and the Chicken game is
communication does not help. In fact, agreements ex ante based upon communication
may trigger the negative solution.
One way to handle the risks in the Chicken Game is to employ a mixed strategy.
However, it entails that sometimes there will occur the coordination failure of macho,
macho. Thus, we have: 3p + 0(1-p) = 6p + -4(1-p); p = 4/7. And we also have: 3q +
0(1-q) = 6q + -4(1-q); q = 4/7. Thus, coordination failure may occur as frequently as
4/7 X 4/7 = 16/49. Consider a version of the same game in Table 2.3 with the Civic
Duty Game, or who is to call the fire services when they smell smoke:
Table 2.3. The Civic Duty Game
Player B
Ignore Telephone
Ignore 0, 0 10, 7
Player A
Telephone 7, 10 7, 7
If the two players choose mixed strategies in the Civic Duty Game, then they cannot
avoid the worst outcome. A correlated strategy such as always one player calling
would be a Pareto improvement but how is such an agreement to be struck between
two players who both would prefer not to call? And who would enforce the agreement?
Communication is a vital strategy in the other two games of cooperation, bargaining
games and assurance. Here, it is in the interest of the players to communicate
truthfully and treat agreements as binding. However, these games are still noncooperative
games. And the players may fail to coordinate upon the Pareto optimal
outcomes, although coordination should be a rather straightforward in these two
games. Let us see how coordination failures may arise.
5
Table 2.4. The Battle of Sexes
Player B
Diner Tennis
Diner 2, 1 0, 0
Player A
Tennis 0, 0 1, 2
In bargaining situations the two players may fail to arrive at the best outcome because
they can not arrive at division of the cake. If bargaining end with no solution, then
there is a strong coordination failure. It is in the interest of all players to extract the
entire value of the game (3), but one player may prefer solutions that are inside the
negotiation set if it provides an acceptable solution from a distributional point of view.
In the game Battle of Sexes a mixed strategy has no sense, as the players should
coordinate on a pure strategy pair, but which one? Both players would be tempted to
push for their best outcome at the risk of arriving at a Pareto inferior solution. The
coordination failure (0,0) is, however, not as likely in the Battle of Sexes as in the PD
game or Chicken, because a negotiated settlement is always better than no agreement
at all.
If the Battle of Sexes is played several times, then a mixed strategy involving a
combination of the negotiated solutions (2, 1) and (1, 2) is feasible. The problem of
finding this mixed strategy is in reality the same as the cake splitting problem of how
two players can agree to divide a cake into two pieces that add up to 1 or 100 per cent.
A number of different solutions have been suggested to the bargaining problem.
However, they all require that the solution arrived at is respected and enforced.
Cooperation in human affairs appears difficult when it is modelled with the above
mentioned games. Thus, the PD game entails defection undoing agreements, the
Chicken game promises threat which may destroy agreements and the Battle of Sexes
models bargaining as a struggle that may lead the negotiators wrong. Let us look at
the game Assurance which models human cooperation as basically peaceful and
capable of attaining Pareto optimal outcomes. Yet, even here there is a minor
coordination problem.
Table 2.5. Assurance
Player B
Floppies Hard discs
Floppies 1, 1 0, 0
Player A
Hard discs 0, 0 2, 2
Coordination may appear innocuous in assurance situations. However, there may be a
failure to coordinate, as communication disturbances may lead to one player choosing
6
floppies while the other opts for hard discs. Suppose that one player has a special
preference for floppies and is known to be stubborn. Then perhaps the other player
will settle for the Pareto inferior outcome?
II. Coordination failures with mixed strategy Nash equilibrium
When there is no pure Nash equilibrium strategy, then the players may fail to
coordinate, one chasing the other. Consider the following game:
Table 2.6. The Samaritan’s Problem
Player B
Work Loaf
Aid 3, 2 -1, 3
Player A
No aid -1, 1 0, 0
Here there is coordination failure in the weak sense that no pure strategy exists, each
strategy pair being dominated by another strategy pair. The only way to break the
infinite regress that is contained in this game in Table 2.6 is to play a mixed strategy:
Player A: 2p + 1 (1-p) = 3p; p = p + 1 = 3p; 1 = 2p; p = ½.. Payoff: 1/5 X 3 + -1X4/5
= -1/5.
Player B: 3q + -1(1-q) = -1q; 3q -1 + q = -q; 5q = 1; q = 1/5; Payoff: ½ X 2 + ½ X 1 =
3/2
Again we may establish that a cooperative solution (3, 2) is Pareto superior to the
mixed strategy Nash solution. But how can the players coordinate upon this
cooperative solution?
To Sum Up
The most elementary forms of human interaction besides conflict which as a zero sum
game will not be considered here presents the possibility of several forms of
coordination failures. The loss to the players involved could be so substantial due to
coordination failures that they would seek cooperative solutions to their interaction
problems. But cooperation is only generally feasible when the players accept so-called
correlated strategies, meaning that there must exist a third party. Government is the
best candidate available for constituting this third party mediator and enforcer. We
will call this third party the “enforcer”.
THE ENFORCEMENT MECHANISM
When two persons have a common interest, then the existence of a mechanism that
offers correlated strategies entails that they can make an agreement about what to do
and that the agreement will be enforced. The enforcement mechanism envisaged here
is of a minimum type, as it only provides for contractual verification and enforcement.
7
Such an enforcement mechanism may be called “government” or “political authority”
when it is made valid for a larger group of individuals. Again the reason would be
economical in nature. An enforcement mechanism would develop economies of scale
in providing third party coordination. Thus, it would be efficient for a society to have
one enforcement mechanism that would develop its order to be valid for a group of N
persons, i.e. create a legal order.
What may drive the process of enlarging the application of the enforcement
mechanism is the search for rents on the part of the enforcer. Drawing upon the
economies of scale in enforcement the enforcer may wish to seize upon monopoly
rents. Thus, he/she may try to eliminate any competitor and charge a high “price” for
his/her services. Among a group of potential enforcers one family may emerge
victorious, to be called the “king” or the “royal” family.
It is in the interest of the enforcer to establish his/her position as firmly as possible.
The danger is again economical in nature, namely free riding.
GROUP SIZE: The Temptation of Free Riding
The enforcement mechanism will have to pay for its cost. With economies of scale the
unit cost will be lower if many individuals contribute to its financing. However,
enforcement being a public good it is not in the interest of any single individual to
contribute. The incentives of the individual are modelled according to:
Cost of enforcement Cost of enforcement
------------------------ < -------------------------- ; Condition of group rationality (1)
Group Size N Group Size N - 1
But for any individual i it holds that
Individual Cost i > Free riding by individual i ; Condition of individual rationality
(2)
Thus, the group will not accept (2) but insist upon (1). It will accept to make the
enforcement mechanism obligatory, i.e. provide it with the characteristics of public
law. Individuals would have to contribute whatever their willingness to pay could be.
This is all a static analysis of the economic reasons of government. Let us add a few
words about the dynamics of the enforcement mechanism.
GROUP SIZE: Transaction Costs
How can a group of a large size decide upon and implement a common action like an
enforcement mechanism? When groups take action in order to provide for some
common interest, then transaction costs surface at once, as group consent requires
negotiations among all persons concerned. The group will have to decide about how
to decide about collective action.
Two-to-two person bargaining may be too cumbersome: If a group involving more
than two persons wishes to take a decision about a common project, then striking a
8
series of deals between two persons runs up lots of transaction costs. It may be
feasible when the group is truly small, but with a group size of N > 5, the process
quickly becomes unwieldy. Other decision mechanisms than exchange become
necessary, such as group meetings where all participate, discuss and take a decision at
one point in time. Thus, bargaining may fail to resolve an issue: If a group considers a
set of alternative projects about what to do, it may not be able to resolve the issue of
which one to select by two-to-two person bargaining. Negotiating all projects among
all the participants again runs up transaction costs. The only viable alternative would
be to introduce a quicker decision technique when conflicts emerge within a group: If
a group of people has in the past been able to coordinate upon common interests, then
there is no guarantee that conflicts among the people in the group may not surface.
The group may not wish to dissolve itself, but could employ a decision technique,
which permits it to resolve the issues, such a voting.
Thus, a group may decide to set up a permanent enforcement mechanism - a
government, which handles collective decision-making. This would be a mechanism
that may be employed for continuous collective action. A government would take
action when it has identified a common objective on the basis of some procedure for
taking decisions such as monarchy, aristocracy or democracy. The classical Wicksell
analysis of group decision-making posits a trade-off between external and internal
decision costs (Mueller, 1989). Here we will add legitimacy to the economic analysis
of government.
Why would an individual accept the decisions of a government, if he/she had not
participated in the making of them or dissents in relation to them after having
participated? A person would like to benefit from the enforcement mechanism but
he/she may wish that the group engages in other forms of collective action. How is
this paradox of collective action to be resolved? Let us look at how this question was
resolved by the major theoreticians in the contractarian school of government.
THE OLD CONTRACTARIAN SCHOOL
Although political authority has existed since the dawn of civilisation, one may still
ask why men and women accept it (Weber, 1978). Contractarian theories focus upon
an economical quid pro quo between people and government whereas utilitarian
theories refer to expediency or total utility (Binmore and Binmore, 1994, 1998). If the
hypothesis suggested above is correct, namely that transaction costs constitute the
economic source of government, then minimising transaction costs lies behind both
the contractarian and the utilitarian frameworks. It is because men and women do not
find time and opportunity to make a long series of agreements between them that it is
necessary to hand over power to a government (Hobbes, Locke, Rousseau) or it is
convenient for society that a few rule the many (Hume). Both the contractarian and
the utilitarian approach look upon government and society as a quid pro quo
relationship under which society accepts to be ruled, i.e. submission or political
obligation, against receiving something back: safety, human rights or the
maximisation of average utility. How did the masters perceive this Wicksellian quid
pro quo?
9
a) HOBBES: Safety traded against obligation or even submission
The agreement between government and society for Hobbes concerned the trade-off
between life and safety on the one hand and obligation, if not total submission on the
other hand. Without government the life of man would be “poor, brutish, nasty and
short”. Locke argued that the list could be made longer and thus more binding upon
government once it had been put in place, viz. life, liberty and property. In his
criticism of the contractarian framework, Hume replaced the agreement with
considerations of utility, which to him led men and women to favour government
ahead of anarchy. The distinction between a contractarian and a utilitarian approach to
government may also be found in modern political theory, as e.g. with Rawls against
Harsanyi. However, the common core of both approaches is the implicit hypothesis
that people cannot make and enforce agreements with each and every one all the time
for the things they value the most such as safety, liberty and property. Transaction
costs would be overwhelming.
Hobbes’ predicament – the protection of life as the foundation of government – is
only a necessary condition for the state and not a sufficient one, as government could
itself fail in protecting the lives of citizens, thus itself reneging upon the contract. To
Hobbes society needs government so badly that the contract entails the submission of
citizens to a Leviathan sovereign. But why would people accept an arbitrary
government, which commits genocide against them? Even if government or the state
is merely a transaction cost saving mechanism, a government that fails to respect the
social contract would violate peoples’ rights and increase transaction costs in society.
Thus, the power of government must be restrained through some legitimation device
like an institutional mechanism like e.g. a constitution with Kant or participation with
Rousseau.
b) LOCKE: Anticipating principal-agent interaction
With Locke the quid pro quo basis of government and political obligation could not
be modelled as one-sided as with Hobbes. Thus, the interaction between rulers and
ruled is conceptualised as a trust. I Anglo-American law trust denotes a relationship
of quid pro quo between one person who has the power to manage something, e.g.
property, and another person who has the privilege of receiving the benefits from
that property. There is no exact equivalent to the concept of a trust in civil-law
systems, but the synonyms include: CUSTODY, care, guardianship, keeping,
safekeeping and ward This reciprocity between rulers and ruled leads Locke to his
theory of limited government, revolution and human rights. However, we still have no
answer to the question when political obligation is legitimate as seen from all the
members of the polity and not only the gentry.
c) ROUSSEAU: Obligation comes from participation
It is true that Rousseau did not target transaction costs as the basis of government, but
he was not really far off the crux of the matter, namely minimising transaction costs.
Man and woman may live as peaceful savages before the advent of civilisation using
their natural force as their liberty in society. However, when these primitive men and
women realise the implications of private property, then institutions must be created
in order to eliminate or minimise chaos. What was natural before the great contract
10
becomes institutional after the contract. What people want – rights – are now much
more safely protected, as they cannot be violated so easily. Freedom has been turned
into law.
In reality, Rousseau is looking at the transaction cost argument from a different angle
than the making of agreement. He underlines the implications of the enforcement of
contracts, or the difficulty of enforcing agreements without consent. Enforcement is
of course a transaction type cost. When each man and woman is going to enforce their
rights, then the whole of society becomes a mess. If a sovereign is erected to enforce
the rights of citizens, then all must obey, or be forced to be free, as Rousseau stated. If
enforcement could at each time and place be called into question, then how could
transaction costs be minimised? Thus, a government once created by the social
contract must be obeyed, but there is only an obligation to obey when this contract
meets with the consent of all.
d) Kant: Rule of law restrains
Kant looked upon the contractarian idea of one or two great public contracts as a
fiction that allowed people to solve the legitimacy question of why a few people
would be entrusted with the task of governing the many. Speaking of political power
as based upon a contract between rulers and the ruled allowed one, argued Kant, to
introduce restrictions upon government in order that the state would be a
“Rechtsstaat”. The legitimacy perspective complements the transaction cost approach
but does not replace it.
CONCLUSION
The so-called Nash programme in game theory implies that only non-cooperative
solutions to human interaction can be accepted, as one is not allowed to assume that
agreements are self-enforced. However, non-cooperative game theory entails that
there will occur coordination failures with attending losses to the players. These
economic costs may be minimised by the creating of a third party institution that
enforces agreements. Government is the third party enforcer par preference since it
may capture the economies of scale in contract protection.
The resort to an enforcement mechanism in society leads further to the classical
problem in all contract theory of government: Sed quis custodet ipsos custodies? It
may be argued that a Kantian institutional mechanism of rule of law is the most
transaction cost saving device for solving this problem of controlling the enforcers.
LITERATURE
Barzel, Y. (2002) A Theory of the State. Cambridge: Cambridge U.P.
Binmore, Ken G. and Binmore, Ken (1994) Game Theory and the Social Contract,
Vol.1. Cambridge: MIT Press.
11
Binmore, Ken and K. G. Binmore (1998) Game Theory and the Social Contract. Vol
II. Cambridge: MIT Press.
Mueller, D. (1989) Public Choice II. Cambridge: Cambridge U.P
Olson, M. (2001) Power and Prosperity: Outgrowing Communist and Capitalist
Dictatorships . New York: Basic Books.