A report (PDF) from Max-Planck-Instituts für ausländisches und internationales Strafrecht about data rentention was recently featured in Heise.de and the online edition of Der Spiegel. Below is a summary in English.
Data retention is not worthwhile, says a study conducted by the Max-Planck-Institute for foreign and international Criminal Law.
According to the study, the logging and retention of certain telecomminications traffic data for six months that was made compulsory in Germany in January 2008 will only have mariginal effect and that the loss traffic data due to deletions will only impact around 2 % of the total number of criminal cases (p. 254). This is within the margin of statistical error and the annual variation in criminal cases solved is larger.
This testimony (PDF) to the German Bundesverfassungsgericht by chartered auditor Meinhard Starostik argues that the impact rate is even lower: 0,002 % (footnote 6).
This finding corresponds to estimates from Bundeskriminalamts, who in a separate study from the summer of 2007 says that data retention will incease the percentage of solved crimes "from 55 percent today to, at most, 55.006 percent."
A study of data retention (PDF) from the Dutch Erasmus University from June 2005 looked at 65 police investigations that were provided by the Dutch Ministry of Justice as good examples of the usefulness of traffic data for law enforcement. They concluded that "in virtually all cases" the police could get all the traffic data they needed, based on the average 3 months availability of telephony traffic data. The researchers also warned they could not qualify the usefulness of these data as direct or indirect evidence, or the representativeness of the sample of cases for law enforcement in general.
The Max-Planck study also shows an exponential increase in use of traffic data by law enforcement, from 5000 queries in year 2000 to about 41000 in the year 2005 (see summary and figures on pages 77, 90, and 402 in the report). In Bayern traffic data queries increased by 60 percent from 2006 to 2007 according to this report.
With respect to types of crime, 50 percent of IP-address queries concerns fraud and 25 percent concerns copyright violations. The argument that traffic data are needed to prevent terrorism is not supported by the statistics.
The study also warns about dangers from abuse due to unauthorized access to the stored data by inside or outside agents at well as the potential to use such data for "strategic information analysis" (data mining) of large segments of the population.
(My German is a bit rusty, so corrections are welcome.)