This chapter commences by arguing that individuals must form collective-action mechanisms to protect themselves from their own protectors. We next consider the erection of a collective-action mechanism and the ability of that mechanism to punish defectors. The discussion then turns to the peculiar nature of the balance between the power of this mechanism and that of the protector and the measures needed to prevent the protector from taking over. We continue with a discussion of the decision-making mechanism, vote allocation, and the need for a voting supermajority. The prevention of capture and the notion of rentseeking follow. Then comes a discussion of controlling the protectors without collective action, followed by considerations of expanding the state and the accompanying collective-action mechanism. We close with conclusions.

THE NEED FOR COLLECTIVE ACTION IN CONSTRAINING PROTECTION SPECIALISTS

Rule-of-law regimes, that is, regimes where clients control the protection organizations, are characterized by the relative ease with which they can be turned into dictatorships and by the difficulty of restoring the rule of law once it has been lost. Both difficulties stem from the problems of collective action. Unless individuals cooperate to prevent dictatorship from emerging and to overturn it once it seizes power, the strongest individual among any unorganized group can subdue the rest, one at a time, and thus become and remain a dictator. Not only can a dictator be expected to enrich himself at his subjects' expense, but also he can be expected to adopt oppressive measures to lower the opportunity for revolt.

Agreements fall into two broad categories: (i) self-enforced and (2) thirdparty-enforced. As discussed in Chapter 3, the second category can be further broken down into two basic subcategories: (2a) those enforced by non-violent means and (2b) those enforced by violence. Individuals, of course, can also act in isolation (grow vegetables at home).

The definition of the state adopted here assigns to the state's enforcement only the agreements that fall into category (2b). Although the focus of this book is on the state, enforcement by the state shares many features with enforcement by non-violent third parties. The distinctions between self-enforced agreements and agreements enforced by any type of third party, then, may shed light on the nature of the state. In the first part of this chapter I discuss factors that affect individuals' choices between the two main categories; the rest of the chapter is devoted to the choice between the two modes of third-party enforcement.

Before commencing the main discussion, I make several points to enhance clarity. The first is that the agreement categories are not hierarchical. For example, owners of idiosyncratic commodities (authors' private diaries) are expected to protect them by themselves. As these become generic (published diaries), we expect that categories (1) and (2a) will be bypassed and that the state will undertake the protection. It is also important to recognize that agreements are routinely subject to multiple enforcement forms. For instance, parts of agreements may be enforced by violence, and other parts of the same agreements may be self-enforced. Still another point is that the existence of third-party enforcement, as argued in Chapter 2, is conditional on the presence of self-enforced relations.

In this and the next three chapters I discuss four topics that will elucidate the character of the state and promote further understanding. The order of these chapters is arbitrary. They concern problems that earlier chapters raised indirectly, but were not integral parts there. I begin with autonomy and merger.

It appears that small states should be able to avoid being captured by, and forced to pay tribute to, larger ones by voluntarily merging in order to mount a unified defense. Yet historical examples of voluntary merger are few. In this chapter, I argue that the prevailing ideas related to the advantages of merger are ill-conceived. I then explore conditions leading to merger, concluding that considerations of trade and its enforcement, not cooperation for military defense, are the factors that drive merger agreements.

WHY POOLING MILITARY STRENGTH IS INSUFFICIENT TO EXPLAIN MERGER

Individually, small states make tempting prey for larger, more powerful states. The combined military powers of small neighboring states, however, may exceed that of a threatening large state. The small states can do well if they can pool their forces. If merger between states were easy to arrange, it would be of great value. Cooperation, however, is always problematic; each party can gain the most from the cooperative enterprise by keeping its own contribution to a minimum. This difficulty is relevant to the pooling of military forces. Agreements among states must be self-enforced, and the temptation for a free ride is bound to be strong. One method of reducing the chance of a free ride is by merger, where the individual states submit to the combined power of the newly merged state.

In Part I of this book we developed a framework to study the nature of the state. Part II focuses on legal institutions and puts more flesh on the previously described bones of the state. In this chapter, I discuss the relationship between protection and delineation and how the formation of that relationship amounts to the founding of legal institutions. Protection efforts vary in terms of the assets in need of protection and the types of threats made against them. I demonstrate that legal institutions will vary with the threats and the asset types. In order to proceed, legal rights must be defined.

The “assets” in the definition are all-inclusive. They encompass, among other things, physical assets, as well as individuals (owned by themselves or, in the case of slaves, by others), intellectual creations, brand names, and reputations. The enforcement and protection measures that the state provides include the prevention of any kind of uncompensated use of or damage to one's legal assets by other persons. Such enforcement covers, for instance, damage from rape and from copyright infringement. I argue later that the state's delineation of assets can take two distinct forms. In one, the state explicitly agrees to protect certain assets. In the other, the state agrees to protect any assets that individuals register with it, as well as those that individuals delineate in contracts.

Besides the violence-using enforcer, other third parties can delineate and enforce claims over assets. The rest of this chapter is devoted primarily to the topic of protection and delineation by force-using protectors, touching only briefly on delineation by other third parties.

CONDITIONS FOR LEGAL DELINEATION

In this chapter, probing further into the nature of legal rights, I first ask this: What characterizes assets that are exchanged by contracts, that is, by agreements enforced by the state? To qualify as contracts, agreements must satisfy certain requirements imposed by the state; they must not violate any law, they must involve quid pro quo, and so forth. In addition, in some cases the state may stipulate the registration of contracts and of properties. An agreement that satisfies these requirements is, therefore, a “legal” contract (though not necessarily a written document).

As long as contracts meet these requirements, the state does not take an active part in determining what goes into them. If the state is to enforce what the transactors choose to delineate, the transactors must provide sufficient information to enable the state to adjudicate in case of dispute. In addition, for the third party to be able to adjudicate disputes about the performance regarding an asset that is changing hands, the asset delineated in the contract must be matched to the recipient. The ease of matching will depend on the nature of the asset; matching unique assets is easier than matching generic ones. Among other attributes, the former can be more readily authenticated than the latter. The state may engage in authenticating ownership, but ownership can be authenticated by other means too. In the case of land, an ownership registry that matches individuals to parcels of land is usually maintained by the state. Authentication by a party other than the state also occurs – for instance, where appliance manufacturers serially number the appliances they produce, and then register the buyers. Ownership registration can reduce disputes and help settle them when they occur.

I have argued (Barzel, 2000a) that people create institutions in order to prevent confiscation and use “checks and balances” to enhance the viability of those institutions. Severe shocks tend to upset the checks and balances and increase the probability that one individual or another will be able to usurp power and become a dictator. In that paper I also described some of the features of dictatorships and the conditions under which those regimes can evolve into rule-of-law states.

In this chapter, I first explore the notion of shock and the paths its effects can take, followed by an inquiry into the distinction between mild and severe shocks. I then compare, along two main dimensions, dictatorial regimes and regimes governed by rule of law. First, I ask how an outside threat will affect each of the regimes. Second, I examine how the two regimes differ in terms of (1) the rights held by the head of state, (2) the state's participation in enterprises, (3) the extent to which legal rights are developed in the state, and (4) the distinction the states make between legal rights and economic rights.

THE EFFECTS OF SHOCKS

At any point in time there is a balance of power between protector and clients. Each particular balance will determine what actions the protector can take, as well as the size of his reward. A shock upsets such a balance and, more importantly, increases the level of uncertainty in the assessment of power. As a rule, the damage caused by a shock is not evenly distributed across states, nor is it within a state.

INTRODUCTION

In the discussion that follows, I use the terms “agreement” and “contract” in very specific ways that require definition.

By these definitions, then, the scope of the state will increase with an increase in the extent of contracting.

In this chapter, I claim that, as a rule, the state enforces only the contractual components of individual transactions; transactions, as a rule, have additional attributes, and these are enforced by other means. As is obvious, I predict that a decrease in the cost of contracting will induce transactors to cover more of the commodity attributes in their contracts. I also predict that a decrease in that cost will generate a number of more “macro” effects:

1. Legal delineation and anonymous exchange will become more common.

2. The level of vertical integration (defined later) will decline.

3. The extent of the market and the role of the state will increase.

By the model here, then, the cost of contracting and the precise contents of agreements can significantly affect the scope of the state.

Discussing first anonymous exchange, I argue that when information is costly, only caveat-emptor transactions meet the conditions for anonymity, and only these are enforced fully by the state. The state never enforces entire non-caveat-emptor agreements; other enforcers enforce part or all of such agreements. The nature and the consequences of mixed enforcement, tightly wound with vertical-integration issues, occupy the rest of this chapter.

A BEGINNING

The model developed here attempts to capture the emergence of social interactions. What should be its starting point? Knowing what is cause and what is effect is essential, and if one attempts to start at a relatively advanced stage of human development, disentangling effects and causes is exceedingly difficult. It is difficult to know which of the forces that might affect political institutions are exogenous and which are endogenous. Were it possible to go back in time to observe the sequence of development of such forces, this problem might be overcome. But at its dawn, coming out of the “primordial soup,” society left behind no social record. A social record is a by-product of social order, which did not exist at that early stage. It is not possible, then, to reconstruct faithfully the early conditions of human life. Instead, we must develop a model for that purpose. This is a problem similar to that of the origin of life, and Dennett's solution seems appropriate here also. Dennett (1995, p. 454) theorized about how life originated: “Since the relevant period was prehistoric, and since he [the evolutionist] had no fossil record to consult, his story would have to be rational reconstruction.”

What could constitute the model or at least a model of the beginning? A candidate for one of the starting assumptions is that humans socialized and cooperated with one another from the outset. To proceed, however, we must have answers to questions such as these: Were they then ruled autocratically, democratically, or in some other way? How specialized were individuals within groups? Were the groups small or large? No a priori way to come up with answers seems available.

In the fall of 1995 I was visiting New College at Oxford University as a guest of Michael Hechter. At that time, I was working on a paper on the nature of the state, and I asked Michael to read it. Once he asked “Why don't you make it a book?” it became clear to me that that was where I was heading, and there was no going back. Michael, who is now back at the University of Washington, has been a constant source of advice and encouragement ever since.

A major theme of this book concerns agreements, their enforcement, and control of the force-using enforcer. The enforcement organization the state employs specializes in the use of violence. People, however, can use other third parties, such as religious institutions or private firms, for enforcement, or they may engage in self-enforcement. In fact, they frequently use more than one enforcer to enforce individual agreements. The approach in this book differs from other approaches concerned with the state, such as those of Olson and North (as well as Hobbes), by focusing on how enforcers are chosen and how those force-using enforcers are prevented from becoming dictators. An analysis of the enforcement of agreements and the choice of enforcers is shown to require the tools of conventional price theory and of game theory. For that reason, in spite of its non-conventional subject, this book uses tools from the economist's tool-bag.

The price theory used here is directed primarily toward the analysis of property rights and the cost of transacting. A major distinguishing characteristic of this analysis is the absence of absolutes. Acquisition of information is costly, and knowledge of economic entities is never complete.

RECAPITULATION

In this book I am attempting to model the emergence of the state and follow its development, primarily when it evolves toward the rule of law. In the model I develop, all individuals, including the Hobbesian king, are assumed to be maximizers. As they begin to socialize, they first do not allow members of the society to specialize in the use of violence, in order to avoid dictatorial takeover by such specialists. However, when threatened by outsiders, they will allow such a specialist to emerge to help protect the members of the society, but only after they form a collectiveaction mechanism that will be able to control the specialist. The emergence of a specialized protector who adjudicates disputes and enforces his decisions defines the state.

The enforcement of agreements is a major feature of the model. The reason is that in only a subset of all agreements has the state a comparative advantage in enforcement; other agreements will be self-enforced or enforced by other third parties. The determination of which agreements the state will enforce and what fraction of all agreements it will enforce will determine the scope of the state.

Underlying this book is the methodology of the economics of (not well defined) property rights or of (positive) transaction costs, which by my definition is its dual. It models a world that radically departs from that underlying the Coase theorem, where economic property rights are (somehow) everywhere well defined. Institutions, including the state, are superfluous in such a world. Such an idealized world cannot exist, however. Indeed, there is no niche in reality where property rights can be fully and costlessly delineated.

Power plays a curious role in economists' attempts to formulate a theory of the state. North (1981) and Olson (1982), two of the more prominent economists who have written on the state, do not give power much of a role. North, for instance, initially brings up the importance of power. In summarizing his discussion, however, he does not even refer to power, but rather states that the ruler is subject to “a competitive constraint and a transaction cost constraint” (p. 28). Olson (1993) introduces the notion of the power of a “roving bandit” who, like the dictator I have discussed (Barzel, 2000a), turns into a “stationary bandit.” But neither he nor I have explored the role of power itself.

Umbeck (1981) seemingly was the first modern economist to construct a model of the state that starts from a “primitive” beginning. In his model, “might makes right”; individuals' power determines everything. Power is also the central force in the accounts of many historians and political scientists. I concur with Umbeck that individuals' power determines the initial distribution of wealth, but the picture changes as Umbeck's model is extended to accommodate specialization.

Individually, specialized protectors have more power than others do. When power is delegated to specialized agents, others must deal with the problem of preventing the agents from taking over and becoming principals themselves. However, physical power, along with other personal attributes, such as cunning, that individuals possess, will predominate only until cooperation emerges. Given the trade-off opportunities between wealth and power, then as wealth is accumulated, sheer physical power will quickly recede to share or even yield center stage to the power that organizations control. I focus in this chapter on two issues related to power.

To assist the reader, certain basic terms and results often used in the discussion are catalogued here. This is followed with brief notes and references to help the interested reader to further investigate these topics. Rather than providing a complete listing of relevant references, my intent is to suggest places to start. Finally, comments about axiomatics and a technical, geometric proof of Arrow's Theorem are given. This proof is designed to emphasize how binary independence loses information about voters and allows nonexistent irrational voters to influence the procedure's outcome.

Glossary

Anscombe Paradox. (Page 123.) A profile example showing that, by pairwise voting, a majority of the voters can be on the losing side on a majority of the issues.

Arrow's Theorem. (Page 43.) This assertion is responsible for inspiring much of the research activity about voting and decision processes over the last half century. It requires the voters to have strict (i.e., no indifference) transitive preferences; there are no restrictions on how each voter ranks the alternatives. The societal outcomes are to be transitive. The only restrictions imposed on the decision procedure is that it satisfies the Pareto and the Binary Independence Conditions. The conclusion is that with three or more alternatives, the only possible procedure is a dictatorship.

Binary independence. (Page 39.) Procedures satisfy this condition if the societal rankings they produce for each pair depends strictly on each voter's relative ranking of the pair; the procedure does not use any information about the other alternatives.

Some assembly required

What do games of chance, analyzing the spread of AIDS, gambling on football games, hedging on the market, appointment of congressional seats to states, and even strategic behavior have to do with this part-whole conflict?

I already explained why an array of important assertions — Arrow's Theorem, Sen's Theorem, as well as the difficulties experienced by pairwise voting, agendas, tournaments, conflicting comparisons, and even the search for consensus — reflect the loss of central, readily available information. Simply stated, by not using crucial information, the integrity of the outcomes cannot be ensured. Problems must be expected whenever information about the disconnected parts fails to characterize the connected whole.

The situation resembles those three feared words which can accompany a new purchase — some assembly required. Panic rushes in. Where is the instruction sheet? Without instructions explicitly explaining how the parts are related, without a guide to clarify how to put them together, the purchase can become an expensive pile of junk — a useless collection of parts collecting dust rather than meeting an intended goal. Similarly, as described in the last chapter, whenever a voting or decision procedure concentrates on the parts, it can unintentionally ignore the assumption that the voters have rational preferences. Lost is the instruction sheet describing how the parts — how each voter's ranking of the pairs — should be assembled. Since the properties of a procedure reflect its inability to use “connecting” information, we must anticipate results such as those of Arrow, Sen, and the free rider.