Inventories were stacking up in America's warehouses and store shelves in the third quarter. (Scott Sady/Associated Press)

The U.S. economy grew at a 3.6 percent annual rate in the third quarter, the government now says, much better than the 2.8 percent previously estimated.

That's the kind of number that, if it reflected the true, underlying growth rate of the economy, would be fantastic news. A couple of years of 3 percent to 4 percent growth would go a long way toward ending the long, post-crisis economic malaise that has afflicted the United States. Too bad that's not at all what the details of the new GDP report suggest is going on.

A run-up in private inventories added a whopping 1.68 percentage points to the overall GDP number, the most since 2010, meaning that "final demand" - -the actual purchases and investments in the second quarter -- rose at a sub-2 percent rate, about where it has been for four years and counting. Inventories can grow for good reasons (businesses are more optimistic about the future and want to stock up), or bad reasons (expected demand for goods doesn't materialize, so more stuff is sitting in warehouses and on store shelves), and it's hard to know in real time which is happening.

But what is consistently true is that inventory shifts are onetime affairs. If anything, a run-up in inventories in one quarter predicts an inventory drawdown in the following quarter (since 1947, the correlation between inventories' contribution to GDP with that same number for the subsequent quarter is negative 20 percent). The steep upward revision is already prompting analysts to downgrade their forecasts for growth in the final months of 2013 and the start of 2014.

But there's a bigger story here than the weird blips to GDP being driven by inventories. It's that we keep getting mixed signals on how robust the U.S. economy really is as 2014 approaches. Some recent data have been good. The Institute for Supply Management survey of manufacturers for November indicated the strongest growth in output since the spring of 2011. The number of people filing new claims for jobless benefits has been hitting rock-bottom levels (including only 298,000 last week, the Labor Department said Thursday, though that was dragged downward by seasonal adjustment quirks tied to the late Thanksgiving).

At the same time, overall growth has remained tame once inventory effects are taken out, and we've seen enough false starts and moments of unjustified optimism during this long, slow recovery that policymakers, particularly those at the Federal Reserve, may want more overwhelming evidence that things are picking up before taking it for granted that above-trend growth has finally arrived.

Which brings us to the November jobs report, due out Friday morning at 8:30. Yes, it's worth mentioning all the usual caveats about not putting too much faith in any one data point, the large margin of error in the survey, the revisions that could ultimately make the initial reading meaningless.

All that's true, but you go to battle (or in this case, set monetary policy) based on the data you have, not on the data you might wish to have. So the Friday jobs numbers, which analysts expect will show 185,000 net jobs added in November, are the single most important data point in determining whether the Fed begins slowing its monthly bond purchases at its Dec. 17-18 policy meeting. And for the rest of us, it will be the best indicator of whether this recovery is starting to take off, or whether the long slog continues apace.