Revising Jobs Numbers Hurts Market Confidence: Sullivan
Here we go again. Another big piece of economic data being revised in a big way.
The January jobs report caused nary a ripple in the market when it came in near expectations. But both December and November data were revised much higher. The combined revisions to those two months added 127,000 more jobs than were originally reported. (Read More: Economy Adds 157,000 Jobs; Rate Up to 7.9%.)
Good news, right? Maybe.
We want—and need—more jobs. But we also want—and need—to have confidence in government data when it is released. Big swings in official data down the road can crush that initial confidence.
Let's be clear: This is no conspiracy theory controversy. I am in no way suggesting there is any "funny" data here. Rather it is simply incomplete and perhaps not ripe for consumption at the time of picking.
Market participants acknowledge this is an issue and has been for some time.
Two economists I spoke with on Friday about the revisions in the jobs data offered essentially the same advice: Take the original (January) jobs numbers with a very large grain of salt. They remind us that the monthly employment data merely comes from two Bureau of Labor Statistics surveys. Both of these rely on the agency being able to reach the surveyed individuals and employers about their job status. Thus, the more time the government has to collect information the better the data tends to be.
(Read More: Main Street Remains Pessimistic, Sees Little 2013 Hiring.)
One economist I spoke with went so far as to say these revisions increasingly render the first look at the jobs report "basically worthless."
This kind of thinking is dangerous to market confidence. We need to believe that the data we see each month or quarter is the most accurate it can be at the time of release. The constant changing of important numbers reduces our ability to understand how the U.S. economy is performing at a given time.
Put bluntly, we simply can't have leading economists and investors blowing off key data when it comes out.
I offer two possible solutions:
Either delay big data points, like the payroll and GDP reports, for an additional month or quarter to enable more time to collect data, or only release the data once it is ready.
Yes, the second option would force the government to end its regular calendar of releases for these reports, but alternatives aren't hard to envision. For example, the Bureau of Labor Statistics could give the market a heads up as to when the jobs data will come out (i.e. "we will release the jobs report next Wednesday at 8:30am"). At least a week or two heads up would be needed, but that also provides even more additional time to fine-tune the surveys.
Though these changes may make it more difficult to plan around the data (on the TV side, too!), it seems a small price to pay for making sure we have the most confidence we can in the data as it's released the first time.
Of course, we could always just have the states—who have more complete and accurate jobs data than the BLS—better communicate with the federal government. But let's not get crazy. This is the government we're talking about, after all. Baby steps.
"Street Signs" airs every weekday at 2 p.m. ET
—By CNBC's Brian Sullivan; Follow him on Twitter: @SullyCNBC