I'll admit it, I was the skeptical college professor mentioned by Eric Edberg. After that talk, I also did some googling. Searching the RIAA website, I found their Research and Data site, with total sales and Consumer Profiles. First, the whole recording industry has gone through a cycle in the last ten years. In 1996 $12.5 billion of recordings were sold*. In 2005 $12.3 billion were sold. The high point was in 1999 ($14.6 billion), during the crazy days of Amazon free shipping. The low was in 2003 ($11.9 billion). Classical recording sales has followed the same trend through this time, maintaining a rather constant 3% of market share. In digital sales, however, the Classical genre does much better. As a Guardian article from last spring points out, classical music accounts for 12% of iTunes sales. And 2005 Year-End Statistics for RIAA show that the recording market started in an upward trend after 2003 when download sales started being recorded. But this is sales, not production.
I couldn't find any industry-wide data, but I looked at Deutsche Grammaphon's catalog to tally their number of releases each year. Here it is:
Year # of releases
A very steady upward climb, with the normal fluctuations expected from a random walk (so I understand from some Econ colleagues). Granted, the huge increase in 2005 and 2006 include re-releases as DG (and all the other record labels) rush to digitize their back catalog. But even so, there are plenty of new releases in these last two years as well. Finally, I point to a relatively old Wired article on the Long Tail. I think too many of my colleagues and their acquaintances are still thinking in the old brick-n-mortor economic mentality, including the speaker from last night. The new digital economy allows for niche markets to thrive, including the various Classical niches. Independent labels such as Canteloupe have taken off, with great access through Amazon, iTunes, eTunes, etc.
I will agree that the major labels did cut many of their artists during the late 90's, probably most from panic when sales temporarily levelled off in 1997. As Greg Sandow (via Eric) points out, Classical recordings did rely upon subsidies, so cutting high-priced artists would shave some expenses, at least from a knee-jerk middle-manager perspective. But I will have to see conclusive data refuting what I have seen, to convince me that the Classical recording industry has indeed died. So consider this a bleg, repeating what Eric asks for: actual statistics along the lines of: "In 1980, there were X new classical releases; in 1990, Y; in 2000 Z; etc."
*Estimated, based on shipping data.