Andy Harless comments at Waldman's latest:

OK, the Fed has two instruments, IOR and OMO’s, and two targets, the condition of the macroeconomy (which we can proxy with NGDP, though it is actually some mysterious combination of inflation and employment known only to the nonexistent collective consciousness of the Fed) and the health of the banking system. It sets its two instruments so as optimize the expected outcomes for its two targets. But since the instruments interact heavily, it’s somewhat arbitrary to say which instrument is affecting which target. 

The view in this post seems to be that, under a “floor” system, we should regard IOR as the lever that controls the macroeconomy and OMO’s as the lever that controls the banking system. Thus OMO’s are not inflationary: they are a microeconomic tactic that has little to do with macroeconomic outcomes. I have some sympathy for that view, inasmuch as it comports with what is likely to be the Fed’s own perception: when it wants to change macroeconomic conditions, it will change IOR, and when it wants to affect the health of the banking system, it will change its OMO policy. 

But I also find Scott Sumner fairly persuasive on this point. To hold the condition of the banking system constant, the Fed will have to adjust its OMO’s in response to its own changes in IOR. If it reduces IOR, liquidity will flow out of the banking system, and, to avoid making the condition of the system more precarious, the Fed will have to increase the base money stock. If it raises IOR, liquidity will flow into the banking system, and, to avoid paying for more liquidity than it wants, the Fed will have to reduce the base money stock. Holding the intended condition of the banking system constant, a change in IOR is an implicit promise of a change in OMO’s. Holding the intended condition of the banking system constant, the two changes are inseparable, so why should we associate the macroeconomic outcome with the former change rather than the latter? 

In the long run, unless there is a dramatic shift in the Fed’s preference regarding the condition of the banking system, it’s very clear that NGDP will be strongly correlated with the base money stock and very little correlated, if at all, with IOR. In the short run, it’s not clear which correlation will be stronger. (Even in the short run, as Scott points out, base money, at the very least, puts an upper constraint on expected NGDP, because IOR cannot go below zero.) I think there’s a strong case that, given the long run correlation, the presumption in associating an instrument with a target should be that we associate the base money stock with macroeconomic conditions. Hence OMO’s are inflationary, even if the commitment to do those OMO’s is inherent not in the OMO’s themselves but in a prior change in IOR.

If forced to guess, I would go with Jan Hatzius's predictions. 2.5 percent growth second half of 2013 and 3 percent in 2014. The Fed doesn't expect full employment until 2015. Nate Silver talks to him in "The Signal and the Noise."

Bill McBride interviews him here.

9.Will the Federal Reserve stop buying assets? 

No. Admittedly, the minutes of the December 11-12 FOMC meeting suggest that most Fed officials currently expect QE3 to end by late 2013. But we would not make too much of this. For one thing, it is important to remember that the outlook for monetary policy depends on the outlook for the economy.
...
10.Will interest rates rise? 

Not much. ... At the longer end of the curve, we do expect a small increase in 10-year Treasury yields to 2.2% by the end of 2013.

Abbreviations:

OMO is Open Market Operations

IOR is Interest on Reserves

I guess what I don't full understand is this bit where Greg Ip comments on what Krugman wrote:

...I’ve concluded the economics are more complicated and more benign than appreciated, but the political consequences are graver. 

(This is going to get rather abstruse, so bear with me.) Let’s start with the role of coins in the money supply. Long before there was a Federal Reserve, the United States government minted coins, often of silver and gold (specie). Private banks used these coins to back their own, private bank notes. The Fed took over issuance of bank notes in 1913, but coins remain the purview of the Treasury. Since bank notes are a liability of the Fed, they are, like reserves, part of the monetary base, the building block of the broader money supply. Coins, however, are a liability of the Treasury. In fact, in economic terms, coins are analogous to perpetual, zero coupon Treasury bonds. The Treasury can issue coins to the public and use the proceeds to finance the budget, just as if it issued bonds. 

Ordinarily, the Fed buys coins from the Treasury in response to demand from banks and pays for them by printing money which goes into Treasury’s account at the Fed. Sound familiar? It is, in economic terms, exactly the same as quantitative easing, when the Fed buys government bonds and pays for them with newly created money, deposited in banks’ reserve accounts. 

There are some differences, however. While the Fed’s balance sheet expands when it buys a coin, the monetary base- the building block of the money supply – does not. That’s because, coins, unlike bank notes, are an asset to the Fed, not a liability. You can see this by examining the Fed’s balance sheet here. The Fed would pay for the $1 trillion platinum coin by creating an equivalent deposit in the Treasury’s account. Treasury deposits aren’t part of the monetary base, either. As the Treasury spent that money, however, it would shift from the Treasury’s accounts to the reserve accounts of commercial banks. For example, suppose Treasury sends a $1000 refund check to a taxpayer, who deposits it in his account at Citibank. The Fed moves $1,000 from the Treasury’s account to Citibank’s reserve account. The Fed's liabilities remain the same, but the monetary base expands, by as much as $1 trillion, eventually. 

Is that inflationary? Paul Krugman says not now, but maybe later: 

      "Printing money isn’t at all inflationary under current conditions — that is, with the economy depressed and interest rates up against the zero lower bound. But eventually these conditions will end. At that point, to prevent a sharp rise in inflation the Fed will want to pull back much of the monetary base it created in response to the crisis, which means selling off the Federal debt it bought. "

I disagree. The Fed does not have to sell its bonds, or the $1 trillion coin, to control inflation (though it may do so anyway). It only needs to retain control of interest rates, and that does not depend on the size of its balance sheet. 

This, by the way, is at odds with the textbook IS-LM model in which the Fed sets a fixed money supply (represented by the upward sloping LM curve)...

Emphasis added. And:

The Fed influences the level of output by changing its interest rate, i.e. shifting the MP curve up or down. At any given interest rate the money supply is infinitely elastic, which simply means it is determined by demand, rather than the other way around, as monetarists think. I wish undergraduate courses used Mr Romer’s IS-MP model instead of the IS-LM model; students would have a much easier time applying their class work to the real world. 

This essentially explains why no major central bank targets the money supply. The money supply does matter, but for narrow, technical reasons. The Fed controls the short-term interest rates by adjusting the supply of reserve balances between banks. Ordinarily, creating trillions of dollars of reserves through QE (or buying a $1 trillion coin) would overwhelm any conceivable demand by banks for interbank funds, forcing the Fed funds rate down to zero. This would rob the Fed of control of interest rates and thus inflation. In this circumstance, Mr Krugman would be right: the Fed would have to sell vast amounts of bonds and other assets, and the $1 trillion platinum coin, to drain those reserves and regain control of interest rates. 

But in 2008, Congress gave the Fed authority to pay interest on reserves. Because banks should not lend reserves to each other for less than they can get from the Fed, this restores the Fed’s control over interest rates regardless of the size of its balance sheet, and thus over inflation.

Emphasis added. Ip doesn't agree with the Monetarists, like Nick Rowe? Is Ip agreeing with Waldman and Kaminska? Tim Duy's original reply:

Ip argues that interest on reserves gives the Fed the power to control interest rates, and consequently the power to control inflation, regardless of the size of the balance sheet. If you follow Ip's analysis through to its logical conclusion, then why should the Treasury issue debt at all? Why not just issue platinum coins? Could cash and government debt combine to serve the same functions together that they serve separately? Consider the disruptiveness of that outcome to the status quo.

Bottom Line: The platinum coin idea was ultimately doomed to failure because neither the Federal Reserve nor the Treasury could allow for even the remote possibility it might be successful. Its success would not just alter the political dynamic by removing the the debt ceiling as a threat. The success of a platinum coin would fundamentally alter the conventional wisdom about the proper separation of fiscal and monetary policy and the need to control the debt immediately.