When the MSM says interest rates are going up it's really the federal funds rate, or the rate that banks charge each other to loan funds deposited at the federal reserve.  This is a paper transaction in most cases, but can be stemmed from a cash crunch which would require the cash to be deposited to actual banks.  

What the FED does is they do not "set" the rate, it is determined by Supply & demand, but for money.  They know they can control the SUPPLY of money.  They can print more(more on that in a sec) or they can take money out of circulation.  When they raise rates they are suppose to be taking cash out of the system.  This ine effect reduces the supply of money and therefore the new "price" of money is the new FF rate.  That rate sets many other prime rates etc.  

I think this is BS, because it's far easier to set a rate by telling everyone it costs 25bp more today, than yesterday.  They don't need to reduce anything, but instead charge more.  I say this because you bring up M3 which is the broadest measure of money.

I have no confidence in the Fed's compilation of banking system "net" repurchase agreement positions, and that it generally excludes Wall Street "repos" makes this quite worthwhile number worthless. There are also issues with "eurodollar deposits." Additionally, any broad measure of "money-like" instruments today must at the minimum include CP [commercial paper], some ABS [asset backed securities] and should include some "structured products." [derivatives?]

What all this means is that the Fed can "print" money in ingenous ways that may or may not inflate the monetary base.  Basically M3 includes all the games that the finance community has come up with to make money on something.

Pulled from Wiki: According to the last published data from 16 March, 2005, M3 has been growing at an annual rate of over 8.22%. As of 16th March 2006 M3 was $10'336.3 billion. One year earlier, on 14th March 2005 the M3 was $9'550.5 billion.  

CPI & GDP are easy numbers to alter and they've done that quite well.  M3 isn't nearly as easy to fudge.  One last note about M3 from http://www.capitalspectator.com/archives/2005/11/does_m3_matter.html

Superficially, at least, it looks a bit suspicious when the Fed says it'll soon bury a series of money supply numbers that contradicts the central bank's public relations effort to convince the world that it's squeezing the monetary system and thereby taking a hawkish view on fighting any and all future inflation. Of course, as Cosgrove observed, the Fed has relatively lesser influence on M3 vs. M2, or certainly the monetary base. Nonetheless, the Fed's only explanation for the planned demise for M3 boils down to sparse interest in the series and the allure of saving money by discontinuing its publication. The former is a reasonable, but ultimately unpersuasive point. After all, the Fed publishes a massive amount of data, and much of it is obscure for all but a few economics geeks. As for the bit about saving money, well, that one just rings hollow from the outset. How much could the termination of M3 save the Fed? Measured against its annual budget, it's barely a rounding error, if that.