Read an Excerpt

Cambridge University Press
0521848296 - Social Traps and the Problem of Trust - by Bo Rothstein
Excerpt

1

Ref lections after a long day in Moscow

All intellectual journeys have a particular beginning. This one commenced one evening after a long day in Moscow five years ago. In November 1997, I was invited by the Swedish Ministry for Foreign Affairs to speak at the Swedish Embassy in the Russian capital. President Boris Yeltsin was soon to make his first (and only) state visit to Sweden and his entourage of civil servants and politicians was eager for more information about Swedish society. There were many issues upon which they sought enlightenment, according to the Foreign Affairs official who contacted me, and what they wanted from me was a lecture that addressed how the Swedish welfare state worked and how we "controlled the Swedish state." It sounded intriguing, and I accepted the invitation with alacrity. I should add that I had no real personal experience of Russia, and my contacts with the extensive Swedish and international research community concentrated on Eastern Europe were even more limited then than they are now.

The lecture, which was interpreted to Russian, was sadly of a somewhat more mundane sort. Much of it dealt with the sometimes esoteric differences between universal and selective welfare programs, the right of Swedish citizens to occasionally appeal the decisions of government agencies, the mysteries of the principle of public access to official records, and peculiar institutions like the Parliamentary Ombudsman. I cribbed a bit from the basic class I taught in public administration policy and some more from a study of Swedish welfare policy that I had published a few years before (Rothstein 1998a). Still, my Russian listeners seemed enthralled, especially when they grasped the economic magnitude of the Swedish public sector.

Following my talk, and those given by other invited guests from Sweden, the Embassy had arranged the kind of refreshments customary in diplomatic contexts to promote more informal discussions between us and the Russians. One of the people I spoke to introduced himself as the third-ranking official in the Russian tax administration. He let me know that he was exceptionally interested in Sweden and Swedish state administration, for a very particular reason. He had been in touch with his colleagues at the National Tax Board in Stockholm and they had told him something that struck him as highly improbable - that the Swedish National Tax Board collected 98.7 percent of what they billed Swedish taxpayers. Could this be possible, he wondered, or were they pulling his leg? He wondered because his and the Russian tax administration's most pressing problem at that time was that they could not collect more than about 24 percent of the total taxes due from Russian citizens according to their tax returns.

"Oh yes," I said after a few seconds' thought, "That sounds about right." The figure did not count the "black" and "gray" market economy, of course, but that the Swedish National Tax Board probably collected such a percentage of the total amount it actually assessed citizens sounded about right to me. To his next question of how that could be possible, I answered that it was probably owing to two things. For some years in the mid-1980s, I had had the privilege of working closely with Urban Laurin at the Department of Political Science in Uppsala, whose penetrating and skillfully crafted doctoral dissertation had been on the inclination (or disinclination) of Swedes to pay their taxes, so I was not entirely at sea on the subject (Laurin 1986). Through long-standing collaboration with political scientist Margaret Levi at the University of Washington in Seattle, I had also been in touch with certain aspects of American research on this intriguing subject (Levi 1988; Scholz 1998; Scholz and Lubell 1998).

Using the research by Laurin and others that I knew something about at the time, I answered that Swedes' willingness to pay taxes was founded on a widespread belief that the tax administration was reasonably competent and compelled most other citizens to pay in one way or another. And since people believed that other people generally paid what they were supposed to, they also paid. Laurin's dissertation in particular supported that hypothesis - i.e. that tax compliance and evasion depended to a great extent on what people believed other people did. This is not unique to Sweden. Two American researchers summarize their findings as follows: "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: 411). A large Danish study based on survey data has also shown such a correlation: "the lower the social trust, the lower moral standards when it comes to paying taxes" (Goul Andersen 1998: 246).

But, I went on, research has provided an explanation. Most Swedish citizens understood that all the money was not stashed in Stockholm by the National Tax Board. At the very least, a substantial portion would conceivably come back in the form of child benefit, old age pensions, health care, public schools, the wages of professors of political science, and other purposes of general interest to the public and the individual. According to that research, acceptance of the need to pay taxes cannot be based solely on compulsion or threats of audits, as such an apparatus of compulsion and control would become far too expansive and costly (Levi 1988). It is also unlikely that most people pay taxes for purely altruistic reasons. Some form of conditional assent must come into the picture. We understand this to mean that citizens are prepared to pay their taxes under certain conditions (albeit somewhat grudgingly). Those conditions are, I said, first that people believe that "most others" probably pay what they are supposed to, and secondly that most of the money is used for purposes people consider legitimate.

"Fascinating, truly fascinating," said my Russian interlocutor. He explained that there were two reasons most Russians did not pay their taxes, which jibed with my reasoning to a certain extent. Russian citizens believed that since most other people did not pay, it was rather pointless to play the honorable taxpayer. Moreover, they believed that most tax bureaucrats and other civil servants were corrupt to the core. Either they took bribes to let people get out of paying taxes, or else they personally confiscated a considerable portion of the taxes that were, despite everything, actually paid. If, contrary to all expectations, some tax revenues reached the proper addressee in the Russian state administration, the general belief was that those civil servants were also corrupt or that the funds were spent for generally illegitimate purposes.

The Russian bureaucrat then wondered whether it was true that most officials in the Swedish state administration could not be bribed. I answered in the affirmative and then inquired, somewhat discreetly, whether the beliefs of Russian citizens about widespread corruption and bribery in his tax administration were founded. "Oh yes," he answered forthrightly, to my surprise. "It is a large bureaucracy with more than 100,000 civil servants, and sure, many are ready, willing, and able to take bribes. But most of them also realize that the current situation is untenable and are fundamentally opposed to the generally rampant corruption." He said that the problem is actually the same as that of the taxpayers. It is rather pointless to be the only civil servant who does not take bribes if one believes that almost everyone else does. My new Russian friend explained that if he could just find some way to convince the majority of civil servants that most others would stop taking bribes and putting tax revenues in their own pockets, he was sure the overwhelming majority would also be prepared to desist from corruption.

At the time, there was a great deal of coverage in Russian and Swedish newspapers about the non-payment of wages and pensions that was engendering widespread nervousness across Russia. With that in mind, I asked my Russian friend again whether most Russian citizens realized that if they did not pay their taxes, the state would never have provide them with schools, health care, and retirement pensions. He replied that most Russians understood that very well but, again, most also believed there was no point in being the only honest actor in such a rotten game. Why should they loyally cooperate with a state they perceived to be genuinely corrupt, and why should they behave honorably when everyone they knew - neighbors, friends, and coworkers - cheated? Who wants to play the part of the village idiot in rose-colored glasses? Or, as put in the English terminology that dominates the social sciences, "who wants to be a sucker?" I could not come up with a reasonable counter argument. Unadulterated altruism is a rare bird, at least when it comes to paying taxes. Another problem is that in situations like these, no good actually came out of altruistic behavior. Those who loyally kept paying their taxes despite knowing about the general disloyalty in the game fed nothing but the corruption.¹

Certainly, this insight into the state of affairs is as logical as it is grim, and we cogitated over the issue as we made further judicious use of the delights the Ministry for Foreign Affairs had laid before us that evening. But my increasingly interested Russian interlocutor continued to probe. He wondered whether I, as a political scientist, had any sound theories that could explain the state of his tax bureaucracy and the Russian society. I perked up, and said that indeed was something for which we in the social sciences actually had remarkably good theories. The Russian situation he had outlined was, I was able to say, a brilliant illustration of a phenomenon given the metaphorical designation of the social trap, among many other names. Especially in the expanding area of non-cooperative game theory, it is one of the central problems -, that is, how to explain the way that cooperation can be established among self-interested utility maximizing actors. Cooperation is based on trust - or, to use another word, social capital. Without trust, I explained, societies, groups, and organizations fall into similar social traps.

The fine art of driving a taxi in Palermo

Diego Gambetta, one of the foremost researchers into the question of how to explain the southern Italian mafia and society, has provided one of the best illustrations of the "social trap." Gambetta's case has to do with taxi drivers in Palermo, Sicily. I am sure they are like all other taxi drivers for the most part, but according to Gambetta they have one rather unique trait: they do not use their two-way radios and have no use for a dispatch center. The reason for this is that when they introduced those new-fangled ideas in the early 1980s, the system degenerated into chaos and universal anarchy (Gambetta 1993: 220ff.).

The utility to taxi owners of a dispatch center that can call taxis over the radio is obvious. Customers need keep track of only one phone number and can be served by the nearest car and thus save time, while taxi owners get more customers and shorter routes. Customers, taxi owners, and the drivers they employ all profit by such a system, which is why taxi owners in most areas of the world have formed alliances and shared the costs of similar dispatch centers, even though they are actually in competition with each other. This simple example of what it needs to create efficient competition among profit maximizing actors in a market shows that competition is not enough. The actors must also agree to establish institutions that are not ruled by competition and self-interest, but are rather driven by norms such as impartiality and the public good. In this case, the idea of such an institution is that customers can call a dispatch center that inquires which driver is closest to the address and, when that driver responds, requests them to take the fare.

But, according to Gambetta, it turned out to be impossible to get this rather elementary system to work in Palermo. The reason for this was that that in order to get the most fares, taxi drivers in Palermo frequently lied about how close they were to the places in the city where they were ordered by the dispatch center to pick up fares. Soon everyone knew that everyone else was embroidering the truth, and so everyone added a few more stitches . . . and a few more. The dispatch center concept is based on the fundamental but uncertain principle that taxi drivers can be confident that none of the others will say they are closer than they really are in order to get the fare. Such a social norm must be established for the system to work. We can safely say that this is a rational strategy for the collective of taxi drivers as the fares are evenly allocated, for reasons of probability, if all drivers state their locations honestly. But since taxi drivers in Palermo, according to Gambetta, could not trust one another, a snowball of deceit upon deceit started rolling and finally everyone lied, always saying they were "just around the corner" in order to claim the fare. Taxi driver A gives his location, B waits to hear it and then says that he is a little closer, whereupon C calls in and says he is even closer, and so on.

Gambetta says there is no incontrovertible evidence that many taxi drivers cheated this way, but the very belief held by the majority that "most drivers" cheated was enough to break down the system as increasing numbers chose to leave the organization. Gambetta concludes that, without trust, there is no possibility of establishing a cooperative equilibrium (1993: 224). That lack of trust led to the closure of the dispatch center and taxi drivers had instead to wait in line at taxi stands around the city, got substantially fewer fares, and had to drive further every time they picked up a customer. The social trap had snapped shut around them. Suspicion had led them all into a lose-lose situation, despite the fact that they all understood that everyone would have profited if they had trusted one another.

But . . . how do you get from Moscow to Stockholm?

I held forth for some time, giving other examples of this fascinating theory, even though it had until then played a somewhat obscure role in my own consciousness, in part because it was frequently presented in an intricately mathematized - and thus, for me, rather inaccessible - form (cf. Scharpf 1997). However, a number of recently published books following that theoretical line but with a distinctly empirical orientation, including those by the American political scientists Gary Miller (1992), Elinor Ostrom (1990), and Robert Putnam (1993), had increasingly roused my interest in the phenomenon. Why did the extent of interpersonal trust and the capacity to establish what some economists call "efficient" institutions (everything from local taxi dispatch centers to all the institutions of states governed by rule of law) vary so widely among different societies, regions, cities, and individual organizations (cf. Myhrman 1994)?

Anyway, for my Russian interlocutor I rolled out large parts of the theoretical and empirical arsenal that social science could contribute towards explaining the situation in which he and all of Russian society then found themselves - one of widespread corruption, lawlessness, mafia control, and crippled public welfare programs. I must admit that I felt rather pleased with myself, especially because the Russian tax official nodded in agreement at many points during my rather lengthy monologue. But then he asked a question that in one blow stripped me of answers and gave me the basic theme of this book. "Tell me, Professor Rothstein," he said, "now that we know all of this and have all of these marvelous theories and intriguing studies, what should I do to make Moscow like Stockholm?"

Instantly, I was at a loss for words.² I had never thought along those lines and I immediately realized, not without some embarrassment, that social science can offer no answers on this issue that are in the least reliable and even fewer that are useful in practice. We have excellent models for explaining static situations and systematic corruption as well as various forms of trustful cooperation over time, but there are no useful models for explaining what causes a change from one state of affairs to another. How can you get people who have long harbored deeply rooted mutual suspicion to suddenly begin to trust one another and cooperate loyally for the common good? Why should people with long-standing and extensive experience and memories of the untrustworthiness (evil, duplicity, cruelty, etc.) of "other people" suddenly begin to rely on one another? In a game like this, trust is not just an empty gesture or a personal preference. It is a matter of fundamentally changing a worldview to one that says most other people will also act in solidarity and cooperate - for example, by giving up tangible resources (paying taxes and refraining from taking bribes). It is not simply a matter of changing values, either. People who take bribes or evade taxes may simultaneously hold values by which they actually consider what they do to be morally wrong and harmful, not only to society but also to themselves over the long term. The reason they continue to act treacherously or opportunistically is not necessarily that they (or their culture) suffer from some kind of moral defect, but rather that there is no point in being the only honest player in a rotten game at which everyone else cheats (or is perceived to be a cheater). This is a case when rationality fails because one cannot rationally decide to forget treacherous behavior (cf. Elster 1983). The act of trusting people who cannot be trusted can be very risky.

According to the logic of the social trap, even people with clear preferences for "fair play" will continue their disloyal behavior because they believe, and for good reason, that almost all "other people" are going to keep playing dirty. And, again, this is not because most other people are actually evil and fundamentally disloyal, but because they expect that everyone else will cheat. Changing the situation is thus a matter of changing the worldview of large groups of citizens about the kind of society they live in and how people might conceivably act in that society. Therein, we have captured two of the central insights of non-cooperative game theory which will dominate this book. First, that political and economic actions should be understood as "strategic" in the sense that what we do depends on what we expect "other people" are going to do (Schiemann 2000). Secondly, that the end result of individual rationality may very well be collective irrationality (Lichbach 1997). Any group of agents risks being trapped in a non-cooperative equilibrium, even though they all realized that a more cooperative equilibrium would bring welfare gains to all of them. As Per Molander has argued, this is in fact a deathblow to every type of idyllic notion that rational agents without coordination can establish efficient equlibria (Molander 1994: 84).

On the difficulty of seeing what does not exist

The following days were full of reflection and contrition. The power of my Russian friend's dilemma was suddenly clear to me. I also realized that the social trap problem certainly did not apply exclusively to Russia, but also to all of the post-socialist societies, not to mention the developing countries with their persistent poverty and corruption (Kornai, Rose-Ackerman and Rothstein 2004; cf. Rodrik 1999). My lecture at the Embassy about such strange phenomena as universal child benefit, active labor market policy, and the parliamentary ombudsman in a public sector that encompassed more than half the gross domestic product (GDP) must have seemed exceedingly odd to my Russian listeners, considering that I devoted not a single word to issues such as bribery and corruption. It was also entirely clear to me that a great deal of the research I had done and been involved in to that point, and which had to do with Swedish social, labor market, and education policy, was based on two tacitly accepted premises - circumstances that I and my colleagues in this type of research had taken for granted, but which we should have analyzed and problematized. First, the existence of fundamental trust in "most" other citizens in Swedish society. Second, the belief that public administration may certainly be both complex and bureaucratic, but that it is not being eroded by corruption to any significant extent.³ In our defense, it is not easy for the research community to study that which does not exist, but from a comparative perspective it should have been clear to us that these were core issues to be addressed (cf. Blomkvist 1988).

Much of this welfare state research has involved the attempt to explain differences in the scope and direction of welfare and social policy in the Organization for Economic Cooperation and Development (OECD) countries. The problem may be described as follows: How should we explain the great variations in social and welfare policies among these countries that are otherwise rather similar in terms of socioeconomic conditions? When all is said and done, Sweden, the United States, Italy, Denmark and Belgium, Germany and Japan are all western, capitalistic, industrial, patriarchal, democratic, liberal market economies. Given the structural logic of the market (or that of the class struggle, the gender struggle, etc.), these countries should have developed rather similar social insurance and social service systems, but they have not. On the contrary, national public policies in these areas so critical to the civic welfare have evolved very differently. A battalion of international welfare state and social policy researchers has devoted extraordinary effort since the 1970s to describe and attempt to explain these differences. Researchers in the Scandinavian countries have primarily emphasized the symbiosis between strong unions and social democratic parties. Some have added to the mix the existence of certain unusual political institutions that favored the inception of a general welfare policy. But after my long conversation in Moscow, it became clear to me that we who are engaged in this research have failed to see an important piece of the puzzle in the building of the Scandinavian welfare state - i.e. the lack of significant corruption and the high level of interpersonal trust in Scandinavian societies. It seems utterly unreasonable to think that it would have been possible to shape public opinion in favor of transferring such large economic resources to various public welfare administrations if the people had strongly believed that those administrations were basically corrupt and/or engaged in systematic abuse of power. It seems equally unlikely that it would have been possible to create these comprehensive social insurance systems if citizens were convinced that most other citizens abused or cheated the taxation or distribution systems.

This illustrates one of the difficulties of conducting social scientific research, that of studying what does not exist. In general, this is categorized as counter-factual history, in which questions such as "what would have happened if . . . ?" are asked. These "if . . . so" questions can sometimes be less meaningful (what would have happened if Napoleon had had access to nuclear weapons at Waterloo?) but, properly used, they are an important element of research because they indicate potential lines of development that could have been entirely logical. In particular, counter-factual thinking constitutes one of the cornerstones of comparative policy research. For example, if interpersonal mistrust and widespread popular suspicion of authorities based on corruption or discrimination are much more common around the world than the opposite, it becomes interesting to ask two questions in order to deepen our understanding of Swedish policy. The first is counter-factual: What would have happened if Swedish policy had been characterized by the kind of interpersonal mistrust and corruption illustrated in my conversation in Moscow? Secondly, what is the origin of the relatively high level of trust that Swedes feel in each other and in their public agencies?

A possibly true story set in Rome, some time in the late 1960s

Allow me to give a specific example of this weak spot in our thinking. Along with many other Swedish researchers, I have studied the Swedish active labor market policy for sound reasons, including its large scope from an international perspective. The active labor market policy has also constituted a central component of the Rehn-Meidner model, a unique macroeconomic model that dominated Swedish economic policy from the late 1950s well into the 1980s. It was in many respects a centerpiece in what became known as "the Swedish Model" (Milner and Wadensjö 2001). A great deal has been written about the origin and function of the model, but it can be concisely described as a means of combining the internal need of the Swedish Trade Union Confederation (Landsorganisationen - LO) for uniform wage development with the government's interest in applying strong pressure for structurally transforming industry in order to stimulate economic growth. Instead of allowing the financial strength of individual firms or industries to determine local wage demands, the unions' wage demands were made progressively more uniform. The consequence was that less efficient firms and industries that were unable to pay wages at the centrally determined level were eliminated while expansive firms/industries could earn large profits, most of which they were forced by tax policy to use for further expansion. The policy provided several advantages to the unions, primarily with respect to internal wage policy. It also benefited the social democratic governments because they were able to control inflation by pursuing an austere finance policy while at the same time harvesting the fruits of strongly increased economic growth (Lindvall 2004). However, one main problem was how to manage the labor force eliminated by structural rationalization - i.e. the many workers who lost their jobs in firms and industries that could not match the centrally determined uniform wages. For a party and a union movement strongly committed to "full employment," this was a hard problem because the Rehn-Meidner model would create unemployment for those who happened to work in less efficient firms and industries. The idea of the Rehn-Meidner model was that it would be possible to transfer this part of the labor force to the type of expansive industry favored by the prevailing wage policy (mainly the large export oriented firms/industries).

According to the "inventors" of the model (the trade union economists Gösta Rehn and Rudolf Meidner), this required a comprehensive public labor market apparatus supplied with extensive administrative and financial resources. Through various "active" measures (well-equipped employment offices, generous subsidies to workers who relocated, and a large program for vocational training), redundant labor could be transferred to new employment. In his memoirs, former Prime Minister Tage Erlander describes the crucial meeting between the leaders of the Social Democratic Party and the leaders of the (blue collar) LO in 1955, when he became persuaded that the model was feasible:

For the first time, I declared my support for the Meidner-Rehn labor market policy. Before then, I had been concerned that it would hasten the closure of companies compelled by free market economizing. But if the mobility-creating policy did not only increase mobility in the labor market, but also led to more socially aware use of our capital assets, if, in other words, the labor market policy gained a stronger selective element, then my fears were erased. (Erlander 1976: 41)