The Interaction of Individual Values and Sticky Formal Institutions in Economic Development.

• Author: Kapas, Judit

1. Introduction

   Although Adam Smith (1759) was the first to analyze how norms, beliefs, morality and culture affect economic development, an upsurge of interest in the role of culture has occurred only recently. (1) Nowadays, the majority of research on the impact of culture is empirical, and up to the present time a relatively large and widespread literature has developed (e.g., Putnam et al. 1993, Knack and Keefer 1997, Zak and Knack 2001, Tabellini 2010, Williamson 2009, Gorodnichenko and Roland 2011).

   While the insight that culture matters in economic performance is commonly shared by the scholars in the field, views and arguments differ when it comes to the details: how culture matters, directly or indirectly; which particular component of culture matters; through which channels culture matters; what the causal mechanisms are; etc. If the findings of the literature concerning these details are in some sense inconclusive, and indeed they are, it is because the empirical literature is "struggling" with theoretical/definitional problems (e.g., Guiso et al. 2006, Tabellini 2010): the empirical literature on culture is "trying to do empirical research on a fragile conceptual base" (Moore 1999:75), accordingly, what the econometrics of culture shows is that there is an impact of something on economic performance, but we do not know what it is exactly (Herrmann-Pillath 2014).

   To progress further in "culture-development" research, some prominent scholars in the field suggest stepping back from "grandiose approaches", and focusing instead on an analysis of the impact of a very specific aspect or dimension of culture (2) (Durlauf and Fafchamps 2005); and relying primarily on theoretically-derived hypotheses (Manski 2000).

   This paper intends to follow the route proposed above: based on the theory of institutional stickiness (Boettke et al. 2008). On the one hand, I will unbundle culture and focus on the development-enhancing impact of only one cultural layer (the "deep" one), and on the other, I will derive theory-driven hypotheses for my empirical analyses.

   When it comes to unbundling culture, my argument is that this can be based on the degree of stickiness of the different cultural layers (dimensions), similarly to the way in which different institutions are categorized on the basis of their degree of stickiness in Boettke et al. (2008). Then, my focus will be on the "deep" cultural layer to which other cultural layers and the formal institutions are stuck, which will be proxied by exogenous individual values. For my concern, the main implication of the theory of institutional stickiness is that the "deep" cultural values are embodied and crystallized in the stickiest formal institutions of a society, leading to a special interaction phenomenon ("stuck-togetherness"). The aim of this paper is to investigate this specific effect empirically, and evidence the genuinely unique role of the "deep" cultural layer in economic development.

   My cross-country regression results have provided details as regards the "stuck-togetherness" of values and sticky formal institutions. More specifically, besides establishing that both values and sticky formal institutions are strong determinants of long-run income, I have found that their "stuck-togetherness" acts as a separate factor in development, and in addition, it amplifies the impact of values on development in the good-institution countries. Furthermore, I have evidenced that better formal institutions increase the marginal income-increasing effect of those individual values that are favorable to development. Both findings suggest that there is a strong complementary effect between individual values and sticky institutions in economic development. My findings are very robust for various robustness checks.

   The paper is organized as follows. In section 2 I will give a short review of the empirical literature on the impact of culture on development. In section 3 I will set out my main hypotheses, and the model, and will present the data. Section 4 will discuss the baseline results of the empirical investigations, while section 5 will provide robustness checks. Section 6 will conclude.

2. Review of the literature: how culture affects economic performance

   The quantitative analyses on the impact of culture that have developed up to the present time are diverse in terms of the measure of culture, the empirical strategy, and the samples of countries or regions used in the studies. Because of the focus of my concern, in this review I will briefly summarize only those studies that are interested in analyzing the impact of culture on long-run development, i.e., those in which the dependent variable is a measure of economic performance.

   Undoubtedly, the early "reference studies" are Hofstede (1980), Putnam et al. (1993), and Inglehart (1990), all documenting that cultural differences are the primary source for growth/income differences across countries. "The opening through which culture entered the economic discourse was the concept of trust" (Guiso et al 2006:29). (3) In the literature trust has been seen as the most important dimension of social capital (e.g., Fukuyama 1995). The first study to investigate the economic effect of social capital, which "has opened a Pandora's box of research" (Casey 2004:96), was Putnam et al. (1993). In their book Putnam et al. (1993) analyze Italian regions and argue that the critical factor in explaining differences in the economic performance of various Italian regions can be found in regional differences in social capital: in regions with a horizontal social structure, based on trust, social capital is higher, and economic outcomes are greater.

   This work has been followed up by numerous analyses, of which Knack and Keefer (1997) is the most influential. The two scholars associate social capital with interpersonal trust and civic cooperation. To measure them, they take data from the World Values Survey. In their cross-country regression Knack and Keefer (1997) find that both cultural variables significantly affect economic growth. More importantly, when including an interaction term of trust and GDP per capita in the regression besides trust and civic cooperation, they provide evidence that both trust and civic cooperation are stronger in countries with higher and more equal incomes, with institutions that restrain the predatory actions of chief executives, and with better-educated and ethnically homogeneous populations.

   Zak and Knack (2001), in some respects, is an extension of Knack and Keefer (1997), by confirming its main findings, but at the same time, providing new insights, as well: they prove the existence of the low-trust poverty trap. Another interesting result is that they are able to identify trust as a channel, and not only a factor on its own to induce growth.

   A few studies have examined whether social capital is a prerequisite for prosperity at the sub-national level. Schneider et al. (2000) is one example. This paper analyzes how political culture and social capital affects growth on a wider sample of the regions of Europe. As opposed to Putnam et al.'s (1993) results, these authors argue that the impact of culture on economic growth is marginal, at best.

   Beugelsdijk and Van Schaik (2005a) is another paper looking at regional differences in Europe in the field of social capital-development. In their empirical investigations they come to a very similar conclusion to that drawn by many others: social capital is positively and significantly related to regional economic growth. With the intention of answering the question of whether the findings of Putnam et al. (1993) on Italian regions can be generalized, the two authors refine their investigations in another paper (Beugelsdijk and Van Schaik 2005b). In this study they renew the regression analysis, by modifying the specification, but more importantly, by providing an extensive set of robustness checks. The main finding is that trust is not a significant determinant of regional growth, but another dimension of social capital, namely active group membership is, a result that partly confirms Putnam et al.'s (1993) hypothesis.

   Akgomak and ter Weel (2009) focus on the indirect effects of social capital on economic growth. As a channel, this paper identifies innovation. The authors apply 3SLS strategy, and throughout the empirical investigation, trust is instrumented by historical institutions. For 102 regions in Europe, this paper provides evidence for the fact that innovation has a strong positive impact on growth, the former being significantly affected by social capital, but social capital does not have a significant effect on growth.

   The only paper discovering a negative relationship between trust and growth is Roth (2009). (4) This paper uses panel data, and reveals that when excluding transition countries from the sample, the relationship becomes curvilinear, meaning that in low-trust countries an increase in trust leads to higher growth, but in high-trust countries an increase in trust leads to a decrease in growth. But, interestingly, when analyzing the relationship in a cross section of countries, the positive association of trust with growth detected my many, appears.

   The research question of Ahlerup et al. (2009) is unique because the authors' primary interest lies in understanding whether social capital substitutes or complements institutions in growth. The results obtained from a standard cross-country Barro-type growth regression provide evidence that trust and formal institutions substitute each other in growth. In the interpretation of the authors this indicates that the marginal effect of social capital decreases with better institutions: trust matters the most when formal institutions are weak. On the other hand, the marginal effect of an improvement in institutions depends on the level of trust.

   The basic econometric difficulties with the social capital...