Having some specific stable quantity of money in the base seems totally unrelated to inflation.
Let me show you how to make base money and inflation seem related. First, you grab the FRED graph for base money:
You divide it by a measure of output like GDP, because prices are proportional to "the quantity of money relative to output."
Then you add a second line to the graph, to show the price trend. I'll use the CPI:
You can see already that the red and blue lines are very similar:) To make this more obvious I will multiply all the blue-line values by a constant, to magnify the line and get a better comparison:
Next, to convert the output numbers to real output, multiply the blue-line numbers by a price index. This is just like what Milton Friedman did in Money Mischief:
There ya go. Now there seems to be some relation between money and inflation! However, it is pretty easy to see that until the crisis of 2008, prices (the red line) were going up faster than the money/output ratio. We can fix this by replacing base money with a faster-growing money. Milton Friedman used M2. I'll do the same:
Look at that! Now there seems to be a striking similarity between prices and the money/output ratio! Wasn't that easy?
Before anybody misinterprets the above post, let me clarify a few things. First of all, my sense of humor is generally inaccessible. But there is the title. And there is my insistence on making the one set of numbers "seem" like the other. And there is the little smiley face. Those are clues.
The big joke occurs between graphs #4 and #5, where I arbitrarily introduce the conversion of output to real output. Yes, I 'defend' it by saying that's what Milton Friedman did. And it is. But that doesn't mean I think it is good arithmetic.
The use of "real" output in this calculation is completely out of place. As I attempt to show, the similarity of money/output to the price-trend arises because of it -- because it factors the price trend into the money/output ratio!
// Update 19 May 2011
For the short form of this critique of Friedman's MRTO graphs, see The Friedman Factor.