Trust, social dilemmas, and the strategic construction of collective memories
Department of Political Science
SE 405 30 Göteborg
voice: + 46 31 772 1224
fax: + 46 31 773 45 99
Generalized trust and trust in government
In November 1997, I was invited to Moscow to lecture about the Swedish civil service for academics, politicians and bureaucrats. The topic given to me was, "How To Control the Civil Service". After the talk, I had the opportunity to speak with one of the top officials in the Russian tax authority. He was very interested in what I had to say because he had learned, in Sweden, taxpayers are actually paying their tax bills. That is, of the total amount of taxes that people are suppose to pay, according to tax authorities, ninety-eight percent also reaches the government. The tax official told me that, at that time, the equivalent level in Russia at that time was twenty-eight percent. Again, we are talking about the taxes the government knows its citizens are suppose to pay according to the tax laws, thus, excluding the "black" and "gray" parts of the economy. When we talked about this huge difference and how to comprehend it, he gave me the following explanation for the situation in Russia. He argued that it was not the case that most Russians did not want to pay taxes or were especially dishonest, instead, the problem, he explained, was that of a classical "social dilemma" (Dawes 1980; Ostrom 1998). Most Russian citizens actually wanted to pay what they were supposed to because they cherished the things that would be done with their tax money (health care, education, pensions, defense, etc.). But, they would only be willing to pay their tax bills on two conditions, which were both never fulfilled. First, they rightly did not believe that all "the other" taxpayers where paying their taxes properly, so it was really no point in being "the only one" who acted honestly. The goods (public, semi-public or private) that the government was going to use the money to produce, would simply not be produced because there were too little taxes paid in the first place. Secondly, they believed that the tax authorities where corrupted, so that even if they paid their taxes, a significant part of the money would never reach the hospitals or schools, etc. Instead, the money would fill the pockets of the tax bureaucrats. In both cases, trust in others was in extremely short supply.
I then asked if this was true, and was told that "yes", there is a significant amount of corruption in the huge Russian tax bureaucracy, consisting of more than 100,000 persons. But, again, the problem was explained to me as a standard "social dilemma" (or very large n-persons "prisoners dilemma"). I was told that most of the bureaucrats did not really want this corruption to continue. They were willing to stop taking bribes and stealing tax money, if they could be convinced that (almost) all the other bureaucrats were also willing to comply with the rules. The problem was, so far, top officials in the tax authority had found no way of convincing the lower level bureaucrats that corruption would come to an end. Again, this was a case of lack of trust.
I told him that I could very well believe his argument, that the reason Swedes were willing to pay their (high) taxes was not least that they a) thought that most other citizens paid what they were supposed to pay, b) that the degree of corruption was very low. In fact, there is pretty good survey research on tax-evasion in Sweden that confirms that citizens belief in if other citizens complied with the tax rules had great influence on their own willingness to pay their taxes (Laurin 1986; cf. Rotter 1980). Moreover, the story also seemed to confirm the survey based research about tax paying in the U.S. done by Scholz and Lubell, who concludes that "citizens will meet obligations to the collective despite the temptation to free ride as long as they trust other citizens and political leaders to keep up their side of the social contract" (Scholz and Lubell 1998, p. 411). This was, I told him, variations in political behavior that social scientists had excellent theoretically sophisticated models to explain. In the words of Bardhan: "corruption represents an example of what are called frequently-dependent equilibria, and our expected gain from corruption depends crucially on the number of other people we expect to be corrupt" (Bardhan 1997, p. 1331). In other words, the more people that we think are corrupt, the more people will be corrupt. The sad news is of course that especially a non-efficient equilibrium of the kind existing in the Russian tax system, seems to be very robust (Bendor and Mookherje 1987; Bendor and Swistak 1997).
The question I then got from the Russian tax bureaucrat, and that had to admit I could not answer, was the following: "How do you go from a situation such Russia's to the situation which exists in Sweden?" That is, how do you move a society (or an organization) from a low trust situation with massive tax fraud and corruption, to a high trust situation where these problems, while existing, are much less severe. The question boils down to how do you change peoples' perception of the type of strategic situation in which they are situated. In the words used in game-theory, what does it take to move a society from a stable, but inefficient, equilibria, to a stable efficient equilibria?1
I think this is a very important problem for both theoretical and political reasons. Simply put, I think we know quite a lot about the "statics" in this area. Field research, historical-comparative research, formal game theory and experimental psychological research now seem to have reached a point where very different types of equilibria can be explained when it comes to establishing more or less efficient solutions to social dilemmas (Bates, de Figueiredo Jr, and Weingast 1998; Ostrom 1998). It is not difficult to understand or explain the social mechanisms that make the agents, such as the Russian taxpayers or tax bureaucrats, when stuck in an equilibria , choose to act in ways which reproduce that equilibria, even if it is collectively irrational for them to do so. In popular language, we know why "once the system gets there, it stays there".2 As it has been formulated by Fritz Scharpf:
If actors were able to choose their orientations (…), they would be better of, in purely individualistic terms, if they could opt for solidarity - and could trust others to do likewise. But the ability to trust is of course the crucial problem. If one party acts from a solidaristic orientation while the other is motivated by competitive preferences, then the trusting party would be left with its own worst-case outcome… In other words, being able to trust, and being trusted, is an advantage-but exploiting trust may be even more advantageous (Scharpf 1997, p. 86f).
From a political economy point of view, it has been forcefully argued by Mancur Olsen and Douglas North that it is the lack of efficient institutions that explains why so many countries remain poor, despite growth in the world economy. Without efficient institutions, agents can usually not produce public goods and the transactions costs in economic exchange can become detrimental to economic growth. Buying private protection to protect ones property rights is usually both expensive and an uncertain investment (North 1990; Olson 1996). To quote Mancur Olsen's last published article:
the large differences in per capita income across countries cannot be explained by differences in access to the world's stock of productive knowledge or to its capital markets, by differences in the ratio of the population to land or natural resources, or by differences in the quality of marketable human capital or personal culture. Albeit at a high level of aggregation, this eliminates each of the factors of production as possible explanation of most of the international differences in per capita income. The only remaining plausible explanation is that the great differences in the wealth of nations are mainly due to differences in the quality of their institutions and economic policies (Olson 1996, p. 19).
The human costs, from Sierra Leone to Kosovo, from Rwanda to Palermo, for the people living in societies without efficient institutions, are immense (Bardhan 1997). Thus, the problem we face is to explain the dynamics, that is, under what circumstances a system is likely to move from the one situation to the other. How can Palermo become Milan, and how can Moscow become Stockholm? And what does it take for Stockholm to become Moscow? As Elinor Ostrom has argued, "the really big puzzle in the social sciences is the development of a consistent theory to explain why cooperation levels vary so much and why specific configurations of situational conditions increase or decrease cooperation in first- or second level dilemmas" (Ostrom 1998, p. 9). If it is some form of trust that is needed, and there seem to be many good arguments for this, then how can trust be created? It is well known from field and experimental research that placing persons in small groups where they have to talk to one another, is beneficial for creating trust and collaboration. Talk, or personal communication, or engaging in mutual discourse of common problems, can build trust. "Cheap talk", as is the game theoretic expression, turns out to be not so cheap (Sally 1995). But we can not put 100,000 Russian tax bureaucrats in a room for face-to-face communication, not to mention the entire Russian population. Deliberative democracy, for all its sophisticated theoretical elegance, seems to be difficult solution for simple reasons of practicality (cf. Elster 1998; Gutmann and Thompson 1996). Interpersonal trust, and trust in societal institutions and in other citizens whom you do not know, are two very different things (Levi 1998a; Rothstein 1998).
While the "discourse" solution is ruled out in such large-n situations, the simple answer to the Russian bureaucrat would have been that they should introduce "the rule of law" type of institution. Corrupted bureaucrats and tax-cheating citizens would then be caught by the police and punished by the courts. By "fixing the incentives" in this way, standard economic theory tells us that the problem would be solved. It's simple, just increase the negative pay-off for cheating and corruption (including the risk of being caught) to a point where the fear of being caught would be higher than the greed that leads agents to engage in tax fraud and corruption. When society is constructed so that fear is larger than greed, things go well.
But as have been argued by e.g., Pranab Bardhan, Michael Hechter, Mark Lichbach, Gary Miller, Barry Weingast, and others, accomplishing this is not an easy thing to do, because constructing such an institution is in itself a collective action problem (Bardhan 1997; Miller 1992). Presuming standard utility maximizing self-interested rationality, where do you find the uncorrupted judges and policemen in a society where corruption is rampant? Most judges and policemen may reason like the tax bureaucrats above, perfectly willing to act honestly if they trust most other policemen and judges to do the same (Hechter 1992; Lichbach 1995; Weingast 1993). Comparing four different solutions to this problem, (markets, hierarchies, norms and contract enforcing institutions) and writing from within a game-theoretic approach, Lichbach has stated that: "The major difficulty with every solution to the Cooperator's Dilemma is that each presupposes the existence of at least one other solution. All solutions to the Cooperator's Dilemma, in other words, are fundamentally incomplete. This is, of course, quite troubling" (Lichbach 1995, p. ix). It should be added that from a rational choice point of view, it is the efficient "Stockholm" solution to problems of collective action and social dilemmas that are difficult to explain. What is most likely to happen with the "homo economicus" type of players is that they end of in "social traps", i.e., the "worst for all" outcome predicted by the famous prisoners' dilemma game (Platt 1973). As Michael Hecther has formulated the problem, if agents act from the standard self-interested rationality, it is most likely that "the agents will outsmart themselves into supoptimal equilibria" (Hechter 1992). Adam Smith's famous 'invisible hand" can not run an efficient market economy by itself, because the institutions needed to prevent "opportunistic" behavior will not automatically be produced by self-interested utility maximizing agents (Ben-Ner and Putterman 1998b, p. 4). The standard solution presented in game theory to this problem, known as "iterated play", may work in two-persons' games, but does not hold for n-persons' games. It has even become doubtful for two-persons' games because of the "shadow of the last game". If the agents know that there will be an end-game where it will be in the interest for the one who has the last move to "defect", this will make the player with the next to the last move, "defect", and so on (Rapoport 1987).
Furthermore, regarding the standard "rule of law" solution, Weingast have made a very important comment, namely that "(a) government strong enough to protect property rights is also strong enough to confiscate the wealth of its citizens" (Weingast 1993, p.287). There is no guarantee if that the government has the power to establish the institutions to implement a "rule of law" system, they would not abuse that same power to break the "rule of law" (i.e., infringe on property rights) if it is in their interest to do so. Rulers, if acting according to standard economic models of human behavior would, without a very long time horizon, and lacking sophisticated knowledge of the relation between incentives, investment and economic growth, are not likely to respect "the rule of law" (Levi 1988).
How to understand the existence of the "rule of law' and other similar efficient institutions from within a rational choice perspective is complicated. Weingast's analysis of the role of the U. S. Supreme Court in the New Deal era is an interesting example. The story is well-known – in order to stifle the Court's resistance to his New Deal policies, President Roosevelt wanted to increase the number of judges in the Court. The new judges he would appoint would be expected to vote against propositions which deemed the New Deal legislation unconsitutional. In his analysis of why this strategy to "pack" the Court failed, Weingast falls back on the following, very non-rationalistic, argument: "(b)ecause it constituted a direct assault on the constitutional principle of the separation of powers, large number of citizens, including many of the intended beneficiaries, viewed the plan as illegitimate" (Weingast 1997, p.254). Thus, Weingast seem to argue that even if it was in their instrumental interest to support Roosevelt's strategy to pack the Court, for the majority of Roosevelt's supporters, the norm against tampering with the constitution took precedence. This may very well be true, but then Weingast ambition to explain the "rule of law" from within rational choice theory is incomplete, because if agents act against their instrumental interests, the assumptions of the theory are violated.
Apparently, without strong ethical norms against self-interested behavior, the "rule of law" can not work as a trust enhancing institution. The first uncorrupted judge or civil servant would not have seen the light of day, given the assumptions made within rational choice theory (Miller and Hammond 1994). Thus, the question is: How can the government, or any other powerful actor, establish credibility and a reputation for trustworthiness so that other actors (citizens, firms and organizations) believe that state officials will honor their commitments in the future? In fact, it is not formal institutions, per se, that can solve this problem of "credible commitments, but instead the actors "cognitive maps" about how trustworthy the actors are in operating these institutions. This is, by and large, an effect of the institutions historical record of being trustworthy or not (Rothstein 1998). As Williams et. al. have stated this problem:
Credible commitments are sticky - hard to establish and hard to change - because they are based on a long history of past policies. Jump-starting capitalism, with the importation of institutions, thus looks problematic. The beliefs to support capitalism must evolve with institutions' longetivity, …, For those societies without a policy tradition of respecting property rights, the perceptions of credible commitment may therefore be all but possible to establish. Institutions can be changed almost at will, but political memories are long and hence belief systems relatively entrenched." (Williams, Collins, and Lichbach 1995, p. 18)
It is not the formal institutions as such, but the perceived history of how these institutions have acted, that matters (Bardhan 1997, p. 1333). That is why exporting a formal institution from one setting to another usually fails to have the expected positive effects. Institutions as such do not automatically change values and behavior, because they do not operate in a history or context free environment (Katznelson 1998). In game theoretic parlance, "history of play" matters. As Fritz Sharpf writes in his Games Real Actors Can Play, "(t)he fact that two actors have a memory of past encounters as well as an expectation of future dealings with each other is assumed to have an effect on the individual interactions” (Scharpf 1997, p. 137).
All the authors cited so far (except myself) work within a rational choice and game theoretic approach. The problem stated above, how individual and collective rationality comes into conflict, is the major theoretical problem within this approach. And I agree with Elinor Ostrom when she argues that this is the major problem in political science (Ostrom 1998). However, as I hope to have shown, it can not be solved within this approach. Agents acting solely from self-interested utility maximization, from which the theory starts, are not likely to solve problems of collective action. The temptation to free-ride will, sooner or later, overwhelm such agents. This occurs because as soon as the payoffs are high enough, they will engage in opportunistic behavior and thus destroy what credibility and trust there is (Miller and Hammond 1994). As Ziegler has recently argued: "In a strategic trust game there usually is an incentive for the trustee to misuse trust", (Ziegler 1998, p.445). Obviously, something other than pure self-interested utility maximization is needed to solve the problem of collective action which rational choice theory so forcefully has put forward (Ben-Ner and Putterman 1998b). This is probably the reason why there is an increasing interest in "non-rationalistic" variables such as norms, signs, history, culture. etc. within the rational choice approach in political science (Levi 1998b; Ziegler 1998). Clearly, something more than agents with standard utility self-interested strategic rationality is needed to solve social dilemmas. There are good arguments in favor of this being the ability to trust and the ability to be trustworthy (Braithwaite and Levi 1998).