As widely reported, the soft employment data for August and declines in August retail sales and industrial production (manufacturing IP also down) have reduced market odds on a Fed rate hike at its meeting 20-21 September. According to the CME FedWatch Tool, based on trading in federal funds, the probability of a rate hike tomorrow is only 0.12. The same CME tool gives a probability of .46 the Fed will stand pat through December. Now what? I wish I knew. Here is how I think about the question.

First, it now appears that the Fed will go into its December meeting, as it did last year, with forward guidance on the table for a federal funds rate increase. The FOMC might, of course, alter its 2016 forward guidance at its September meeting. If the Committee reduces its guidance to indicate a fed funds range 25 basis points higher than now, but below prior guidance, will that create a strengthened implied “promise” to act in December? That would double down on its current problem with forward guidance. Will the FOMC hike even if employment data through November remain soft? Or, suppose employment growth resumes; will the market take seriously that the FOMC would consider a 50 bps hike in December as implied by current forward guidance?

Second, what are Janet Yellen’s incentives? A year from now, looking back, is the Fed likely to be in a better position and her reputation enhanced if the Fed has raised the federal funds target rate in 2016 and it turns out to be premature or the if Fed has held steady whereas it would have been better to have tightened in 2016? Given the data in hand as I write, it seems to me that waiting makes more sense. Yes, unemployment is below 5 percent and recent employment growth solid, but softening. However, there is little sign of rising inflation. On conventional measures, there is still slack in the labor market; for example, the labor-force participation rate is still well below prior levels. And, don’t forget that in 1999 unemployment fell to almost 4 percent.

Third, if the Fed gets behind by not moving in 2016, how hard will it be to catch up? How much difference can it make if the Fed moves in early 2017 rather than in 2016? Only an old-fashioned fine-tuner can believe it makes much difference.

We can replay this same argument at every future FOMC meeting. What must happen to create a compelling case for the Fed to move? My interpretation of the rate increase last December is that it had less to do with compelling new information than with the fact that the Fed had long promised to move in 2015. That says much more about the wisdom of forward guidance than about sensible monetary policy.

Here is a suggestion for the FOMC, which seems so obvious that I assume the Committee must already be considering it. The FOMC should recast its forward guidance away from the calendar. At its September meeting, the guidance should apply to end of third quarter 2017, 2018 and 2019 rather than end of those calendar years. At each meeting, the guidance would then apply to 4 quarters ahead, 8 quarters ahead and 12 quarters ahead. With this approach, the Committee would never again face an apparent calendar deadline to act.

Seems obvious to me, and very simple. Yes, perhaps guts forward guidance and that would be a good thing. The mantra should be “data dependence, not date dependence.”