按：這是筆者在港大政政系research course的一篇短文, 主要就Robert Putnam的名著Making Democracy Work作出批判–該文認為社會資本的水平越高, 該地區政府的施政就會越成功。筆者嘗試對社會資本的概念作出剖析, 指出Putnam的方法論有以偏蓋全之虞, 社會資本既不是民主體制成功的必要條件, 也不是充份條件, 很可能只是成功民主體制下的伴隨現象而已。
Does social capital contribute a significant cause to democratic institutional success?—a critical review on Putnam’s Making Democracy Work
Since the publication of Robert Putnam’s book Making Democracy Work in 1993, the term “social capital” has been made popular in socio-developmental theory. Social capital has been claimed as the “missing link” in development and promoted as an essential element of democratic institutional success. As Putnam (1993) puts it, “Building capital will not be easy, but it is the key to making democracy work”.
We can have two levels of analysis of making “democracy work”. The first one is about the role of social capital in democratization. In the book, Putnam attempts to answer the possible connections between political culture, democracy and economic growth. Social capital is defined as “features of social organization, such as trusts, norms, and networks, that can improve the efficiency of society by facilitating coordinated actions”. Social trust, a key concept in social capital, is comprised of two components, norms of reciprocity and networks of civic engagement. Though not very explicitly stated, Putnam seems to positively correlate the level of social capital with the degree of democratization of institution, at least in Italy. This in turn brings us to the second level of analysis on good governance of democratic institutions. Putnam puts two causal arguments in the book to show how social capital can “make democracy work”: social capital is the significant cause of good governance and economic development, and social capital is the result of path-dependent historical legacies. Throughout the book, Putnam illustrates his claims by identifying different empirical correlates of democracy, through comparing quantitative and qualitative data with historical information in Italy. He presents a historical account of virtuous circle in Northern Italy where good governance was observed and, on the other hand, that of a vicious circle in Southern Italy where bad governance was resulted. Based on the empirical data presented, Putnam makes the claim that good regional governance can be found in areas where there is a higher level of “civicness”. This contributes to the institutional success in Northern Italy, while Southern Italy is doomed to have bad governance.
There are numerous critiques on theoretical as well as empirical grounds on Putnam’s work. It is simply out of the scope of this paper to present a comprehensive review on these stuffs. In this paper, I would focus my analysis on the concept of social capital itself. The term “social capital” is vague as it does not have a clear definition of what it is. It seems to be anything that exists in the community or society network. Such an ambiguous concept fails to explain institutional performance. I would argue that the use of the concept to explain institutional success would result in empty proposition.
The failure of explanation also rests on the selection bias of the empirical evidence to favor certain conclusions that Putnam might want to make. As a result, his arguments fail to universalize to other countries, especially in Southeast Asia. The selection bias, together with the vagueness of the concept of social capital, makes the explanatory power of his theory rather limited.
The vagueness of the concept and the measurement bias
Putnam is not the first one to coin the term social capital. Bourdieu (1986) defines three dimensions of capital—economic, cultural and social. Bourdieu puts emphasis on class conflict in which social capital is a resource in social struggles. Under this context, social capital is defined as “the aggregate of the actual or potential resources which are linked to possession of a durable network of more or less institutionalized relationships of mutual acquaintance and recognition” Thus, social capital is a “resource” that can be mobilize and a “quality” produced by the totality of relationship between actors. Coleman (1988) gives a new relevance to Bourdieu’s concept of social capital. He defines social capital by its function. It is not a single entity, but a variety of different entities, with two elements in common: they all consist in some aspect of social structures, and they facilitate certain actions of actors within structure”. There still many other definitions of social capital. For instance, Becker (1974) defines social capital as a particular kind of intermediate good for the production of assets entering as arguments in agents’ utility functions. Many economists also focus on the ownership argument and claim that the use of the word “capital” is inappropriate as social capital can hardly “change” its ownership.
Despite the enormous definitions of social capital in terms of resources, social structures or utility functions, Putnam makes his own definition in the book. Recall the definition that listed in the beginning of the article: Social capital is defined as “features of social organization, such as trusts, norms, and networks, that can improve the efficiency of society by facilitating coordinated actions”. It is the dense network of social reciprocity and civic engagement that build and maintain democracy, and therefore make institutional success. From the definition, it seems that social capital includes anything, such as public goods, social networks, civic culture, etc. Civil associations, one of the “indicators” of social capital, provide the “networks of civil engagement” within which reciprocity is learned and enforced, trust is generated, and communication and pattern of collective action are facilitated. In the book, Putnam also emphasizes the role of horizontal networks in building an efficient democratic institution. Those networks, such as sport clubs, cooperatives, mutual aid societies, cultural associations and so on, are important in building networks of civil engagement, as opposed to the vertical networks of patron-client arrangements or of traditional hierarchical organizations such as the Catholic Church.
Despite the “impossibility” to provide a single, universal definition of social capital, Putnam uses four indicators to measure social capital. They are the number of voluntary organization, the number of local newspapers’ readers, voter turn-out at referenda and the relevance of preference votes expressed by voters in elections. It seems that Putnam has ignored the important role of civil association (especially the political one) in counteracting the oppressive states. During the “third wave” of democratization, the increase in the number of political associations is very important in the transitional period of post-communist countries. In Eastern Europe, the highly politicalized NGOs, which fight for common wealth, nation-cultural activity, or particular group interest and those which advocate human rights and civil rights were central to the oppositional activity in the democratization process in those highly oppressive states. It is unclear why Putnam only selects those “apolitical” voluntary associations as the index of measuring “civicness” and the degree of democratization. A possible response from Putnam would be this: he tries to reconcile the need for an engaged political community with the need for political order and stability. Only those associations that can win the social-political as well as cultural struggles in the transitional process can qualify as “civic” associations to contribute to democracy. However, certain selection biases have to be made in order to construct the goodness of the existence of civil society to democratic institutional success. It is a highly controversial empirical claim that one is not hard to find many counter examples to argue with. The idea of “dense networks of civic engagement” is very much alike the concept of “GuanXi” in China. It is easy to find enormous examples of how the Chinese businessmen in Southeast Asia try hard to reduce transactional costs by creating a network of reciprocity and trust amongst the participants, i.e. to establish good “GuanXi” among business partners. Nevertheless, such an existence of network cannot make these participants an advocate of democracy. In many cases, elitism is a one of the commonly observed phenomenon in the Chinese culture. In addition, such a network can neither guarantee any efficiency nor responsiveness of any institutions that embedded in a vast network of “GuanXi” culture. These networks need not to be vertical in nature. There exist horizontal “GuanXi” networks in Chinese culture. But even if such networks exist, there is no guarantee that the inter-communal “GuanXi” must override the “GuanXi” within the community. If the network within a community becomes dominate, it seems to have a negative impact on democratic institutional success due to the conflict of different interest groups. The linkage between strong civil society and strong democratic state might not be as strong as Putnam might think.
Besides selection bias, another problem concerns with the flawed conception of democratic institutional success. First, the civic community index cannot truly represent civic virtues. There are simply many controversies on the selections of suitable indicators, I just mention some here. For instance, isn’t the educational level of people and crime rate also count for “civicness” of citizens? Second, why Putnam just measures institutional success in terms of efficiency and efficacy, but not the level of fairness instead? There are simply more for the requirement of “political equality” than Putnam portrays in the book. It seems to me that there is not enough “democratic elements” in Putnam’s index for institutional performance. There is just one element for “responsiveness”. It seems that such an index is also suitable for an autocratic or an authoritarian government. One of the possible responses from Putnam would be that, democracy is defined (or characterized, in a weaker sense) by the high level of social capital. This move, though very commonly found in many literatures, is problematic. I would come to this problem in the next section.
It might be unfair to Putnam using an external perspective for criticism, given one aim of the book is to explain the institutional success in Northern Italy. However, from an internal critical point of view, Putnam’s argument would remain more likely to be a circular one if he only selects evidence that favor the explanation of institutional success. The explanatory power is only limited in Italy case. What is more, due to a highly historical deterministic nature of his arguments on civic roots in Italy, the explanation on institutional success would be likely a tautology.
Methodological weakness that impedes the explanatory power of the model
The measurement of social capital remains highly problematic without a clear articulation of what social capital is. If social capital means anything in social networks, trust and social norms, can we really measure it in terms of indirect indicators such as membership in mutual aid societies or that in cooperatives? It seems that such use has confused about what social capital is, as distinct from its outcomes, and what the relationship between social capital and its outcomes may be. If the definition simply includes everything, social capital would become tautologically present whenever an outcome is observed. In Putnam’s studies, social networks, voluntary organizations and civic participation are used together with variables like the efficiency of justice administration and responsiveness in performing a multivariate analysis aiming to build synthetic indicators of social capital. Such a construct puts higher levels of life satisfaction and social well-being altogether tautologically present with anywhere social capital indicators are observed. It makes the concept of social capital too board to include the entire institutional network of a society, which makes the concept useless to explain anything.
In addition to the vagueness of the concept, Putnam tries to use a highly historically-determined model to explain institutional success in Northern Italy and institutional failure in Southern counterparts. He traces back the civic roots to 11th century. Despite the problem of subjective historical interpretation, tracing back explanation to historical facts cannot help to explain institutional success. It seems an infinite regress that exists in an extremely complex and un-traceable web of historical events and explanation. I try to illustrate the problem with the following hypothetical dialogue:
A: Why are there democratic institutional success observed in Northern Italy?
B: Because there is a higher level of social capital observed.
A: Why is there a higher level of social capital?
B: Because there are denser networks of civic engagement as well as higher level of trustworthiness between citizens. These are the important indicators of “civicness” exists in a society.
A: Where does the “civicness” culture come from?
B: It can be traced back to the Italian historical events happened several centuries ago.
Under such a picture, it seems that for Putnam, a democratic intuitional success is defined by a high level of social capital. Social capital is also defined and determined from history. When someone disagrees with Putnam’s observation with the argument that a high level of social capital does not always contribute to democratization and institutional success, a possible response to save his theory is to re-define social capital as everything (such as trust or networks or norms) that can make agents cooperate or market works or institution performs better. Nonetheless, these factors are the results (or epi-phenomenon that have nothing to do with causal relationship) of successful democratic institutions, not the causes. If we include these factors in the definition of social capital, then surely any empirical analysis will find that social capital does increase cooperation among agents and hence improves the efficiency of institutions and makes institution more democratic. It is an empty proposition that cannot explain anything. We are just using a tautology to support an empirical statement. The argument runs like this:
1. Social capital is either a sufficient or necessary condition (or at least, less logically valid, an important condition) for democratic institutional success.
2. Statement 1 is a generalized and empirical claim from induction.
3. Therefore, any empirical evidence that shows democratic institutions with high level of social capital but without good governance can sufficiently reject statement 1.
So far, statement 1 is not a tautology. When does this statement become an empty proposition?
4. Successful democratic institution is defined (or characterized) by having a high level of social capital.
It is only with the addition of statement 4 that makes the whole argument an empty proposition. If anyone rejects the claim that social capital can cause democratic institutional success, then simply either his definition of social capital is not comprehensive enough, or one can respond by adding more elements in the definition of social capital so as to “fit” those characteristics that appears in any successful democratic intuitions observed so far. Putnam’s move resembles to the second strategy. The bias selection of evidence and the board, vague definition of social capital make his stand very close to statement 4. To explain the origin of the high social capital, he adds some historical deterministic element in the definition of social capital in which the appearance of the networks and trust are all culturally and historically determined. But this cannot help to increase the explanatory power of the model. It can only make the model more “localized” which can only explain the situation of a specific region with that particular historical context. It simply adds one more argument in the closed, self-explanatory model.
As we can see from the above analysis, the vagueness of the concept and the methodological weakness greatly reduce the explanatory role of Putnam’s model. Social capital is neither a necessary nor sufficient (nor even an important) condition for democratic institutional success. With the construction of a closed-systemic explanation of the civic tradition in Italy, the application of the model towards other places in the world is highly limited, if not virtually impossible.
Arrow, K., (1999), “Observations on Social Capital”, in: Dasgupta, P. Serageldin, I. (Eds), Social Capital. A Multifaceted perspective. The World Bank, Washington D.C.
Becker, G. (1974), “A Theory of Social Interactions”, Journal of Political Economy, 82, n. 6, pp.1063-1093.
Bourdieu, P., (1986), “The forms of capital”, in: Richardson, J.G. (Ed.), Handbook of Theory and Research for the Sociology of Education. New York, Greenwood Press, pp. 241-258.
Bowles, S. and H. Gintis (2002), “Social Capital and Community Governance”, The Economic Journal, Vol. 112, I. 483, 419-436.
Coleman, J., (1988), “Social Capital in the Creation of Human Capital”, American Journal of Sociology 94, 95-120.
Letki, N. (2006), “Investigating the Roots of Civic Morality: Trust, Social Capital, and. Institutional Performance,” Political Behavior, 28, 305-326.
Putnam, R.D., Leonardi, R. and Nanetti, R.Y. (1993), Making Democracy Work, Princeton: Princeton University Press.
Chase, R., Woolcock, M. (December 12-15, 2005) “Social Capital and the Micro-Institutional Foundations of CDD Approaches in East Asia: Evidence, Theory, and Policy Implications”, Arusha Conference, New Frontiers of Social Policy. Retrieved April 14, 2008, from http://siteresources.worldbank.org/INTRANETSOCIALDEVELOPMENT/Resources/Chase&Woolcock.rev.1.pdf
Sabatini, F. (2006), “The Empirics of Social Capital and Economic Development: A Critical Perspective”, FEEM Working Paper 15.06, January 2006, Eni Enrico Mattei Foundation, Milan. Retrieved April 14, 2008, from http://www.socialcapitalgateway.org/Sabatini%20(2006)%20-%20The%20empirics%20of%20social%20capital%20and%20economic%20development%20-%20January%202006%20-%20Published%20FEEM.pdf
 Putnam (1993), p.185.
 Ibid, p.167.
 Ibid, p.171.
 Bourdieu (1986), p.119.
 Coleman (1988), p.98.
 For more arguments please see Arrow (1999) and Bowles and Gintis (2002), also see Becker (1974, 1996).
 Putnam (1993), pp.167-175.
 Ibid, p.96, the civic community index.
 This phenomenon resembles to the idea of epiphenomenalism in philosophy of mind.
 We cannot simply assume that these factors are the causes for democratic institutional success. It only begs the question as we want to find what contributes to such a success.
 I should here give what “empty proposition” means. A tautology is an analytic sentence that is always true. An analytic sentence does not give information about the empirical world. For instance, 1+1=2 do not give any information about the empirical world, and it is true by definition, so it is a tautology.
If we try to use a tautology to support an empirical statement, empty proposition is resulted. Suppose I am going to justify this argument: Humans are rational by nature. Normally we would try to justify it by empirical observation, using the method of induction. But it is always hard to do so, as it is impossible to examine all “possible” humans. Then I try to give an analytical sentence: Humans are defined as a rational being. If anyone tries to use any empirical observation to argue that some people are irrational, then I simply reply that they are not human at all. This kind of justification is referred as empty proposition.
 It would be even better if we separate the arguments into two: a)social capital is either a sufficient or necessary condition for democracy; and b)social capital is either a sufficient or necessary condition for (any) institutional success. For simplicity, I just combine a) and b) together.