What Determines Protection of Property Rights? An Analysis of Direct and Indirect Effects
| Profession | School of Business, George Washington University |
| Pages | WPS3940 |
Page 2
Protection of property rights is a key determinant of the efficient operation of contracts and the development of financial institutions. The critical question is why some countries have managed to develop strong protection of property rights, while others have not. A substantial body of theoretical work tries to explain the historical determinants of these differences. There is also a growing body of empirical work that assesses the relative contribution of different historical determinants in cross-country variation of property rights protection.
However, attempts at empirical validation of institutional theories face challenges stemming from severe data limitations. There is only one realization of the data with relatively few observations, which have by now been well explored in the literature. Given the overlapping nature of the theories of property rights, it is difficult to fashion empirical tests which are consistent in their treatment of the competing theories. Different investigators focusing on different variables may find different specifications persuasive to test and to report. Moreover, it is possible to quickly develop a heuristic about which variables are jointly significant in the regressions. This is a matter of concern because out-of-sample tests are not feasible. Similar concerns exist in the asset pricing literature, where Foster, Smith and Whaley (1997) show that there is a significant bias "when a researcher has had access to many potential regressors (or, equivalently, has read past research that suggested which regressors to choose)".1
Page 3
In this paper we address these issues by adopting an empirical approach that relies on the data rather than investigator discretion to specify a model linking property rights and a set of potential explanatory variables advanced in the literature. Our approach treats the potential explanatory variables together and evenhandedly, and allows us to explore relations between them. Empirically, we use cross country data on 158 countries and evaluate four theories concerning historical determinants of property rights protection. While there are overlaps, the four theories focus on different and distinct causal mechanisms in shaping institutions, as captured by legal origin, endowments, ethnic diversity and religion. We begin with a set of variables suggested by the institutional theories and then allow the data to reject potential causal relations between these variables and property rights protection. At the end of the process, we are left with a set of potential causal relations between our measure of property rights protection and the set of proposed explanatory variables. The procedure also suggests possible relations among the set of proposed explanatory variables.
Specifically, we employ Directed Acyclic Graph (DAG) methodology developed in computer science that allows us to consistently evaluate the four theories concerning historical determinants of property rights protection (Spirtes, Glymour, and Scheines, 2000). This algorithm uses the correlation matrix of a set of variables to determine whether a variable meets certain criteria, derived from probability and graph theory, for it to be classified as a direct or indirect cause of another variable. Using this methodology, whose purpose is to discover causal patterns in the data, we are able to identify which historical factors are direct determinants of property rights protection and which are not. We subject the results to a battery of robustness tests and compare our methodology to Page 4regression analysis to illustrate how a regression-based analysis of the question can lead to misleading results.
Our results show that at the 5% and 10% significance level, Common Law, Latitude and Ethnic Fractionalization are all significant predictors of property rights protection where as Catholic Religion is not. However at the 1% significance level, only ethnic fractionalization is a causal determinant of property rights protection. Further, Ethnic Fractionalization is the only variable which is robust to different sample compositions where as the effect of Latitude and Common Law is strongly dependent on the sample of countries under study. The data offers only limited support for the proposition that Common Law origin or Latitude is a determinant of property rights protection and the support is not robust to different definitions of the variable or sample composition
Our paper is related to Beck, Demirguc-Kunt and Levine (2003) and Ayyagari, Demirguc-Kunt and Maksimovic (2005a). Using firm-level survey data, Ayyagari, Demirguc-Kunt and Maksimovic (2005a) evaluate determinants of firm level perceptions of property rights protection and find the ethnic fractionalization and endowments view to explain a greater proportion of variation in the data compared to other historical factors However, they consider one variable at a time and do not rule out any variable. Beck, Demirguc-Kunt and Levine (2003) find more evidence supporting the law and finance view. However, in this paper we are able to illustrate how the regression-based methodologies, as normally applied, can be misleading in identifying causal factors and that legal origin is not a robust determinant of property rights protection.
Page 5
Our paper is most closely related in spirit to Kormendi and Meguire (1985), Barro (1991), Levine and Renelt (1992) and Sala-i-Martin (1997). These authors examine a parallel problem: determining which of many possible proposed macroeconomic variables could reliably be classified as predictors of economic growth. Kormendi and Meguire (1985), Barro (1991), and Levine and Renelt (1992) use Extreme Bounds Analysis (EBA) and Sala-i-Martin (1997) uses a similar technique. DAG analysis has several advantages over these methods. Whereas these methods start from an equation that is specified by the researcher that embodies a causal ordering that is then tested, DAG can endogenously discover the causal ordering. Moreover, whereas EBA treats one relation at a time, the graphs produced by DAG show robust relations between all the variables being analyzed, taking into account the implications of robust relations elsewhere in the system, on the causal ordering in a specific relation. The DAG analysis also allows the researcher to explore the implications of imposing a causal restriction in one relation on robust relations throughout the system.
The remainder of the paper is organized as follows. Section 2 discusses the hypotheses we investigate. Section 3 outlines the empirical methodology and the data. Section 4 applies the methodology to identify historical determinants of property rights and presents the main results, comparing them to regression analysis. Section 5 provides additional robustness results, with different samples of countries and alternative variable definitions and compares DAG methodology to Extreme Bounds Analysis and a methodology for examining robustness due to Sala-i-Martin (1997). Section 6 concludes.
Page 6
We evaluate four potential historical determinants of property rights protection. First, the law and finance view predicts that historically determined differences in legal traditions help explain differences in protection of property rights today (La Porta, Lopez-de-Silanes, Shleifer and Vishny, hereafter LLSV, 1998). Focusing on the differences between the two most influential legal traditions, the British Common law and the French Civil law, this theory holds that legal traditions differ in terms of the priority they attach to protecting the rights of private investors against the state (Hayek, 1960). The reasons for the differences can be found in the way different legal traditions evolved. While the British Common law evolved to protect private property owners against the crown (Merryman, 1985), the French Civil law evolved to eliminate the role of a corrupt judiciary by restraining courts from interfering with state policy and to solidify the power of the state. Over time, these trends led French Civil law to focus on the rights of the state and less on the rights of the individual investors when compared to British Common law (Mahoney, 2001). Thus, the law and finance theory predicts legal origin to be an important determinant of property rights protection, with countries that have adopted the British Common law tradition placing much more emphasis on such protections than countries with the French Civil law tradition. As these legal origins spread around the world through colonization, British colonizers brought with them a legal tradition that stressed private property rights protection, while French colonizers spread a legal tradition that is less conducive to such protection.
Second, the endowment view emphasizes the role of geography and the disease environment in shaping the institutional environment and the property rights that Page 7underline such development. Acemoglu, Johnson and Robinson (2001) argue that it is not the identity of the colonizer but the colonization strategy that determined the extent of property rights protection. In settler colonies such as the United States, Australia and New Zealand, Europeans settled themselves and created institutions to protect private property rights and check the power of the state. On the other hand, in colonies where the colonization strategy was to extract resources from the indigenous population rather...
To continue reading
Request your trial