About ten years ago, Greg Mankiw devised a formula to model monetary policy in the United States in the 1990s, which ended up saying that the federal funds rate should equal:
8.5 + 1.4 (Core inflation - Unemployment)
That's with "core inflation" defined as CPI inflation rate over the previous 12 months excluding food and energy and "unemployment" defined using the seasonally adjusted rate. Essentially it's a form of backward-looking Taylor Rule. This isn't, I think, the best way to make monetary policy but it serves as a useful illustration anyway especially since Eddy Elfenbein has taken the formula and projected it forward:
The key thing here is that by these standards the Fed should have set interest rates well below zero during the depths of the recession and it continues to be the case that rates should be negative. The Fed could have and should have achieved this goal by attempting to elevate inflation expectations, but they didn't so the recession has been long and painful. But by this standard, at least, we're fast approaching the moment at which our zero-bounded monetary policy becomes appropriate. At that point, we should expect the economy to start soaking up excess capacity rapidly. I like this post because it's a great wonky demonstration of my recovery winter accelerating growth hypothesis. 12 or 18 months ago, money was way too tight. Now it's still tighter than it should be, but we're much closer to the mark and thus poised for better growth. And by this model at least, a further fall of 0.2 percentage points in the unemployment rate would be enough to get us into the liftoff zone.
That's not really what Mankiw's model was build for, so I wouldn't take the 8.3 percent unemployment rate thing to the bank, but the underlying dynamic will hold for any reasonable variant of the model.