When it came to excess and waste, the Soviet Union was hard to beat - its defense industry produced military equipment in inordinate numbers. For some reason sea-launched ballistic missiles were particularly popular. Each SLBM launcher had more than three missiles produced for it - "for tests and storage".
A note in the Kataev archive has numbers for projected production volumes for some of the systems: each R-27 launcher (SS-N-6, Project 667A/Yankee submarine) had 3.64 missiles produced, R-29 launcher (SS-N-8, Project 667B/Delta I and Project 667BD/Delta II) had 4.7, R-29R (SS-N-18, Project 667BDR/Delta III) - 4.3, R-29RM (SS-N-23, Project 667BDRM/Delta IV) - 4.0, R-39 (SS-N-20, Project 941/Typhoon) - 3.
Land based ICBMs were produced in smaller numbers - each launcher of R-36, UR-100N, and MR UR-100 was supposed to have respectively 1.5, 1.6, and 1.7 missiles produced for it. As the Soviet data indicated, that was roughly on par with the numbers in the United States.
It's hard to tell why the difference is so large. Somewhere else in his notes Kataev mentioned that the Soviet Navy had a plans for their submarines to return to port after they fire their missiles for a few reloads. I would say the Rocket Forces had a more realistic view of the nuclear war business they were in.
While it may seem no more than a curiosity today, these numbers have some significance for missile proliferation. A number of recent accounts of the North Korean and Iranian missile programs indicate quite convincingly that those programs used some components of the R-27/SS-N-6 missile. The East-West Institute study and most strongly Ted Postol's technical addendum (PDF) to it indicate that Iran used the R-27 vernier engine in its Safir space launcher. North Korea also appears to have access to a number of R-27/SS-N-6 missiles and used them in its missiles (again, I rely on Ted Postol's analysis).
As far as missile proliferation is concerned, the good news is that neither North Korea nor Iran seem to have the technology to produce powerful rocket engines, so they have to rely on whatever the Soviet components they managed to get hold of. The bad news is that there are a lot of those components out there, enough to build a not-so-small missile force.
At the peak of the R-27 deployment, in mid-1970s, the Soviet Union had 544 launchers for missiles of that type. This means that almost 2000 missiles were manufactured. Of those, at least 653 missiles (492 R-27 and 161 R-27U) were used in flight tests during 1968-1988. The 544 deployed missiles were fueled, so they most likely have been liquidated after being withdrawn from service. This leaves about 800 missiles that have been produced, but never fueled. Unfortunately, it looks like some of them ended up outside of Russia. We don't know how many, of course, but given the number of missiles available, it can easily be a few dozen and could be as high as a few hundred.
UPDATE 01/17/10: As some of my colleagues pointed out, the number of "dry" missiles is probably significantly lower than 800 - most produced missiles were deployed on submarines at some point of their relatively short lives and therefore were fueled. Main engine, which was submerged in the fuel tank, would also come into contact with fuel - this would would definitely limit its lifetime. Vernier engines, however, were outside of the tank and therefore could be used again, if necessary,