I was thinking about the issue of lost coins last night.  I read in the FAQ the idea that lost coins aren't lost value, as the value of a lost coin will be redistributed to all other coins in the economy.  However, it also states that there is no way of distinguishing a lost coin from a coin in savings.  It follows that whatever impact the coin's "loss" had on the economy (apart from its owner) took effect as soon as it exited active circulation, not when it was lost.  It follows from that that savings have the power to increase the value of all the coins in circulation, which would motivate more savings, and... of course, what I'm describing is the deflationary spiral which has already been talked to death here.

But assuming the simplest economic model of bitcoin possible, let's follow this thought experiment through.

So everybody wants to save their bitcoins instead of spending them, because the value is increasing, and as more people start saving, that increase appears to accelerate.  However, savings as a fraction of the total bitcoin money supply will follow an S-curve; the deflation will start to decelerate as the majority of the money supply winds up in savings.

At this point, bitcoin owners will realize that they are sitting on piles of gold, but that deflation has become so slow that their savings cannot reasonably gain any more value (in this simple model).  Furthermore, if some other saver were to get the notion to start spending, the increased circulation would cut into the value of your savings.  Clearly it is imperative to be the first one to start spending, and thus maximize your buying power before everybody else jumps on the spending bandwagon!  This would start an inflationary "spiral" (though not hyperinflation, obviously, because there is an upper limit on the amount of currency in circulation).

I imagine these forces would not swing back and forth between such extremes, but would rather balance each other out somewhere around the middle.  Once new bitcoins are no longer being produced, I foresee that fairly constant ratios of coins will remain in circulation and in savings.  The inflation/deflation rate should tend to zero, excluding the very important effects of population growth, economic growth, etc.

Do others agree with this assessment?