Trust in private and common property experiments.

• Author: Cox, James C.

1. Introduction

   Numerous experimental studies involving private property endowments have demonstrated that individuals' decisions, in a variety of situations, reflect complex and diverse motivations beyond simple own-income maximization (see, for example, Ostrom 1998; Camerer 2003; Ashraf, Bohnet, and Piankov 2006; Camerer and Fehr 2006; Cox, Friedman, and Gjerstad 2007; Cox, Friedman, and Sadiraj 2008). More pertinent to the research presented here, extensive research has been generated showing that subjects in "Trust Game" experiments (also called "Investment Game" experiments) achieve higher levels of efficiency than predicted by the stage game equilibrium for self-regarding (or "economic man") preferences. In the standard form of this game, first movers are assigned private property rights to an endowment that they can either keep or invest in a potentially advantageous way with an anonymous second mover (without the presence of a third-party enforcer of contracts).

   The implication of such fairness behavior for individuals who own assets as common property has not been explored. Although not addressing the specific institutional context examined in this article, the traditional literature suggests a rather negative view of how individuals may cope with common property arrangements (see, for example, Alchian and Demsetz 1973). In addition, there is often confusion in the policy literature regarding open or limited access common-pool resource arrangements versus common property arrangements. Discussion of common-pool resource settings may, in fact, ignore the issue that commons may be utilized under a wide variety of property arrangements. A negative view of expected individual behavior of those who jointly own assets has been compounded by a confusion stimulated by Hardin's (1968) allegorical analysis of the "tragedy of the commons," which was effectively open access and not common property. Government ownership and private property have both been recommended as "panaceas" to correct for the presumed inefficiencies of what have loosely been called common property resources, even though in the extreme case of open access no one has property rights to these resources (Ostrom, Janssen, and Anderies 2007). Considerable field research has now challenged the frequently accepted conclusion that common property systems are more inefficient than private property systems and has illustrated the substantial difference in incentives and behavior between open access resources and resources owned as common property (NRC 1986, 2002; Dietz, Ostrom, and Stern 2003; Dolsak and Ostrom 2003).

   Since contracts are typically incomplete, many of the possible gains from exchange with private property require trust and reciprocity (Fehr, Gachter, and Kirchsteiger 1997). Similarly, efficient outcomes with common property also require trust and reciprocity. In addition to the institutional context in which users of the commons make decisions, in both experimental and field research, trust appears to be a core variable explaining why participants in some settings tend to cooperate while tending not to cooperate in other settings (Ostrom, Gardner, and Walker 1994; Ostrom 1998; Walker and Ostrom 2007).

   Recently, we embarked on an experimental research program designed to explore whether a fundamental difference exists in the trusting behavior of individuals holding private property assets as contrasted to holding common property assets. The experimental design used here focuses on the effects of assigning subjects' endowments as common property rather than private property. This study can be viewed as exploring the robustness of previously reported results for the two-person private property Trust Game. In addition, the sequential common property treatment used in this study complements earlier research on time-independent and time-dependent common-pool resources (Ostrom, Gardner, and Walker 1994). In common-pool resource experiments where no property rights are assigned and communication is not allowed, participants substantially overexploit the common-pool resource. When given an opportunity to communicate and to choose their own property rights, however, participants tend to trust one another to adhere to nonbinding agreements, generating near optimal decisions (Ostrom, Walker, and Gardner 1992).

   The Private Property Trust Game and the Common Property Trust Game investigated here are both based on the standard Two-Person Trust Game (Berg, Dickhaut, and McCabe 1995). In the standard game, the first mover, endowed with a private good, makes a decision that can create a surplus to be shared with a second mover because each $1 sent to the second mover by the first mover is tripled by the experimenter. The first mover faces a social dilemma, however, because the second mover will be given full ownership and full authority over the distribution of the first mover's investment and the resulting surplus. That is, to create the efficiency-enhancing surplus, the first mover must either trust that the second mover will reciprocate the first mover's generous transfer or have altruistic preferences for the second mover's payoff (Cox 2004; Cox and Deck 2005; Ashraf, Bohnet, and Piankov 2006).

   The Common Property Trust Game is the inverse of the standard game--which we refer to in this paper as the Private Property Trust Game. In the Common Property Trust Game, the initial endowment of wealth (equivalent to the maximum possible tripled amount available to the second mover in the Private Property Trust Game) is assigned jointly to the two players. In this game, the first mover has the option of withdrawing resources from the joint fund, and in the process destroying surplus because every dollar withdrawn reduces the joint fund by $3. (1) Similar to the Private Property Trust Game, the first mover faces a social dilemma in the sense that the second mover has full authority over the final distribution of the fund remaining after the first mover makes the withdrawal decision. In order to refrain from withdrawing from the joint fund, the first mover must either have altruistic preferences or believe the second mover is trustworthy--as the first mover must do in order to invest a positive amount in the Private Property Trust Game.

   The traditional Trust Game has been examined experimentally by many scholars. The first study to explore this game was undertaken by Berg, Dickhaut, and McCabe (1995). Using a double-blind protocol, the first mover and second mover are each individually endowed with $10. The experimenter triples any amounts that are sent by the first mover to the second mover. The second mover then has a decision as to how to divide the amount received. A large proportion of first movers sent money to the second movers, while only two of the first movers sent nothing. Half of the second movers returned either $0 or $1. On average, the second movers returned slightly less than what the first movers sent to them.

   The experiment undertaken by Berg, Dickhaut, and McCabe (1995) has been replicated by a large number of scholars in Bulgaria (Koford 1998), France and Germany (Willinger et al. 2003), Germany (Jacobsen and Sadrieh 1996), and Sweden and Romania (Rothstein and Eek 2006). Camerer (2003) provides an overview of findings from Trust Games. In general, one finds that a large proportion of first movers send a substantial portion of their endowments to second movers, and many second movers return substantial amounts. Findings vary by country and the specific experimental design, but there is more congruence across experiments than disparity. Some experimental designs (for example, Cox 2002, 2004; Cox and Deck 2005) implement control treatments that discriminate between distinct motives for sending positive amounts (such as altruism or trust) and distinct motives for returning positive amounts (such as altruism or reciprocity). Methodological issues concerning such discriminations have recently been explored (see, for example, Cox, Sadiraj, and Sadiraj 2008).

   As discussed above, the focus of this study is to expand the set of property rights environments in which the Trust Game is played. In particular, our focus is on the extent to which decisions by first movers and second movers are affected by whether endowments are assigned as private property or common property. The traditional model of self-regarding (or "economic man") preferences, as well as models of unconditional social preferences (Fehr and Schmidt 1999; Bolton and Ockenfels 2000; Charness and Rabin 2002), all predict that behavior in the Trust Game will be the same, regardless of whether endowments are private property or common property because subjects' feasible choices and payoffs are invariant with the type of endowment. In contrast, recent models of reciprocal preferences (Cox, Friedman, and Gjerstad 2007; Cox, Friedman, and Sadiraj 2008) can be extended to explain why behavior may differ with the type of endowment, as follows.

   The second mover's feasible set, corresponding to the private property endowment, allows the second mover to take for himself or herself an amount not greater than his/her private endowment of $10. The second mover's feasible set, corresponding to a positive amount s sent by the first mover, allows the second mover to take for himself or herself any amount up to $10 + $3s. Therefore the second mover's feasible set, corresponding to the endowment in the Private Property Trust Game, is less generous (to the second mover) than all other feasible sets corresponding to positive amounts that the first mover might send to the second mover, according to the MGT ("more generous than") ordering of feasible sets defined by Cox, Friedman, and Sadiraj (2008, p. 36). The larger the amount of his/her private endowment that the first mover sends to the second mover, the more generous is the feasible set to the second mover, and the more altruistic he/she...