Thursday, 12 May 2016

Counting Laws in Comparative Law? – A Comment on the ‘Global Regulation’ database

I came across www.global-regulation.com a few weeks ago and its ‘analytics’ function seems quite interesting indeed, despite some problems discussed below. Counting ‘laws’ is also not something entirely new to comparative law - in my book:
Actual attempts to measure the quantity of legal rules have been limited to the codified law. In the 1980s Heinz Schäffer and Attila Racz led a research project on ‘Quantitative Analyses of Law – A Comparative Empirical Study: Sources of Law in Eastern and Western Europe’. This project used questionnaires to estimate the total amount of generally binding normative acts, for instance, in terms of pages and the number of single norms. Schäffer and Racz also tried to identify the amount of legal changes within a period of ten years, as socialist governments were expected to be less hesitant in changing existing legal orders. More recently, Tom Ginsburg and colleagues were interested in whether and how legal systems differ in the ‘specificity’ of codified law. The first paper called this ‘leximetrics’, and examined whether in Europe the length of laws implementing EU directives varies systematically across countries… Ginsburg and colleagues also examined the specificity of constitutions, using data from most countries of the world...
Global Regulation’s analytics enables searching for the frequency of terms based on a large number of laws from (at present) 48 countries; for non-English speaking countries this is based on machine-based translations of laws. For the purpose of comparison the search results are displayed in two bar charts showing the absolute number of laws that mention a particular term and then the percentage of each country’s laws. Naturally, the absolute data usually have the countries with many laws at the top position (notably the US); thus, the percentage data are the more meaningful ones.
  Searching for general substantive terms such as ‘justice’, ‘efficiency’, ‘welfare’ etc and then comparing the results across countries is interesting, but it is unlikely to be reliable due to differences in legislative drafting and problems of translation.
  A bit more specific terms may work better such as ‘socialism’, ‘social’ and ‘god’ – and the following results may indeed be somehow plausible (click to enlarge):



The chart below for Europe is interesting as well: so apparently countries that are not quite in the EU – Switzerland and Turkey – have many laws that need to deal with their relationship to the European Union.

Next: what about the relationship to international laws and other countries more generally? Searching for ‘international’ and ‘foreign’ gives you the following (and I also searched for further terms such as ‘China’, ‘India’, ‘English’, ‘French’, ‘Spanish’ etc):


Those latter results show the main limitation of the database: some countries, eg Estonia, Switzerland and Turkey, are always at the top of the results. How is this possible? Well, the data only present the number of laws, regardless of the length of the laws. Thus, I’d suggest that for those countries the database seems to have mainly covered fairly long laws (such as the main Codes) that are likely mention any term, while for other countries many smaller statutes seem to be included.
  The obvious solution would be not simply to measure and report the number of laws in percentage terms, but also to consider the length of these laws. Apparently, Global Regulation has this information as it also reports the ‘complexity’ of laws which it defines as the length of laws; yet, unfortunately, this is not (yet) included in the comparative bar charts.