Despite the (partially) fierce criticism of a quantitative approach to comparative law, the use of statistics to answer comparative law questions seems to have inspired some new comparative law research. A working paper by Mathias Siems with the title A Network-Based Taxonomy of the World’s Legal Systems uses statistics in order to “develop a difference-based taxonomy in accordance with the idea of legal families.” (p. 10)
After discussing pros and cons of classifications in comparative law, Siems comes up with three groups of four variables ranging from “English as official language” to “Paid annual leave.” He then applies these variables to a dataset of 157 countries in order to determine the difference of each single variable with the same variable for each of the remaining countries. This allows him to plot a histogram for the distribution of the country differences which in turn allows him to create and graphically represent cluster for all the 157 countries studied. Siems subsequently applies his variables to the legal origins literature concluding – somehow surprisingly – that his “network data does not support the legal origins taxonomy.” (p. 17) The last part of the paper is dedicated to developing an alternative classification of the legal systems of the world. By playing with numbers, Siems comes up with four groups: “The Global Anglosphere”, “The Modern European Legal Culture”, “The Rule by Law or Religion” and the “Weak Law in Transition”.
With his paper, Siems opens an interesting avenue for comparative law research. Statistics in general and network analysis in particular may play a role in helping to answer sometimes longstanding questions in comparative law. Yet, statistics are not inherently objective and they need to be properly applied in order to make use of their full potential. In this regard, Siems’ paper raises several concerns. First, the choice of variables by Siems – while well elaborated though probably extendable from a legal standpoint – is problematic from a statistical viewpoint as he mixes categorical variables (here questions with yes/no answers, e.g. “Islam state religion”) with ordinal (or numerical) ones (where the answer lies on a spectrum, e.g. “literacy rate”). Also, the exact translation of the variables into numbers is not really explained. Second, Siems does not really elaborate on the reasoning behind the thresholds (“cut-offs”) for his clusters. Some representations are based on a 0.11 threshold, others on 0.13. Why these thresholds were chosen remains unclear. Third, the conclusion that the classification of the legal systems according to their legal origins is wrong rests on shaky statistics. Little details are given as to the numbers and the reasoning seems somehow circular. Fourth, Siems selects the number of groups into which legal systems can be divided randomly by playing with different numbers and looking which number of groups makes more sense. This, of course, takes away a lot of objectivity from the study.
Overall, the general idea behind A Network-Based Taxonomy of the World’s Legal Systems by Mathias Siems, i.e. to examine traditional issues of comparative law is very innovative and may potentially shape future comparative law research. As any pioneering work it inevitably faces some challenges – mostly due to the fact that the statistical analysis of networks is complex and at the same time so foreign to the conventional way to think about law.
Mathias Siems, A Network-Based Taxonomy of the World’s Legal Systems, 2014 Working Paper Durham Law School, available here.