I must admit that it isn't easy to find good statistics on the web.

First, take care with the prices from solarbuzz, as they are based solely on prices for a single module purchase, found on the web. Real purchase prices are clearly lower. Moreover, it was quite astonishing that, although the U.K. market in absolute terms is totally negligible in comparison with the rest of Europe, the recent devaluation of the British Pound had such an effect on the European Module Price Index. From http://www.solarbuzz.com/moduleprices.htm (January 2009):

Price movement started throughout the PV chain during the Fourth Quarter, but it has yet to be fully reflected in retail prices through this sales channel. While there has been a dramatic move in the European index this month, this was actually a consequence of exchange rate movements (mainly Sterling/Euro) within Europe rather than a large number of price changes in the survey result.

Nevertheless, the statistics from Solarbuzz are still valuable, especially because they are updated monthly. Me too, I regularly check them.

For 2000-2005 data, IEA-PVPS could be a source:

http://www.iea-pvps.org/products/download/rep1_17.pdf

Figure 10 on page 28 gives a nice graph with 1997-2007 module and system prices for Germany (country #1), USA (country #2) and Japan (country #3).

For trends since 2006, personally, I have more trust in some other statistics, like these from photonconsulting (module and system prices for 2006 and 2007, Table 1):

http://www.photonconsulting.com/files/sa_2008_exec_summary.pdf

or from BSW-Solar (updated every quarter, in euro):

http://www.solarwirtschaft.de/preisindex

To end with, I would like to point to a rather old (2003), but still interesting presentation, saying something more about the importance of BOS Cost (Balance of System Cost = System Cost - Module Cost):

http://www.ecn.nl/fileadmin/ecn/units/bs/PHOTEX/PV_expert_ws/introductio...

They were also correct expecting price increases because of silicon feedstock problems.