It only took seven years, but zero interest rates are about to be a thing of the past. The question, though, is for how long.

That's because any doubt that the Federal Reserve was going to start increasing interest rates later this month has been erased. The economy was already doing well enough that, in the Federal Reserve's estimation, the first rate rise in nine years seemed justified, but it would have been a little less so if jobs growth had stalled out in November. It didn't. Instead, the economy added 211,000 jobs, another 35,000 in revisions to previous months, and the unemployment rate stayed flat at 5.0 percent for the good reason that more people were both looking for and finding work. And that should keep falling as long as the economy is adding anywhere near the 218,000-jobs-a-month it has averaged for the last three.

But how much further will the Fed let it fall? That's really what it's deciding when it decides how fast to raise rates the second and third and fourth times. It's said it will do that slowly—maybe 0.25 percentage points every other meeting—but will that speed that up if the recovery doesn't slow down?

Well, that depends on what the Fed thinks the "natural rate" of unemployment is. That's the economy's Goldilocks zone where unemployment isn't so low that inflation increases, but not so high that unemployment does either. The Fed used to think that was around 5.2 percent unemployment, but has since marked that down to 4.9 percent. But who's to say that it isn't even lower than that? The truth is we don't know. It's all guesswork. The only way to tell is to wait until wages actually start rising more. That's a sign that unemployment has fallen about as far it can, and companies are having to fight over workers instead.

Sounds simple, but there are two problems. The first is that the Fed is about to start raising rates even though wages aren't. Average hourly earnings were up just 0.2 percent in November and 2.3 percent in the past year, well below the 3.5 to 4 percent that be consistent with a healthy economy. The Fed, in other words, is acting as if it's sure the economy is near full employment when it isn't clear that it is. And the second is that the Fed wants unemployment to go below its natural rate, but not too much. The idea is that this would make inflation, which is far, far below its 2 percent target, start rising towards it without going over it. But how can you do the second if you're also doing the first? The only way is if your guess—and that's all it is, really—is right that economy is already at the natural rate of unemployment. But if it isn't, then you'll slow the economy down before you wanted to—and you might have to cut rates back down to zero a lot sooner than you wanted to, either. Maybe just a year or two from now.

The safer course of action would be none at all. It would be waiting until wages pick up for at least a few months. That might mean something as small as 2.5 percent annual wage growth for three months time, just enough to tell us that this isn't a blip, but rather the beginning of a trend.

The tricky thing is that unemployment says the recovery is complete, but wages don't. One might be enough, but why take a risk that you don't have to? The only thing worse than zero interest rates is trying to end zero interest rates too soon—because you end up right back where you started.