The Data Will Return, But They'll Be Messy

Markets may be jumpy too.

Photo by Ian Talmacs on Unsplash

Reopening the U.S. government gives economists and investors reasons to cheer, too, as economic data streams return to their regular schedules. But like your kitchen faucet after the plumber has interrupted your water supply, the best advice is to let the water run for a while before you trust it.

The first releases on unemployment are expected to come early next week, along with retail sales, trade, and personal consumption expenditures —the Fed’s preferred inflation gauge. If this sounds too good to be true, it’s because this will be the September data that was almost ready for publication before the government shut down on October 1.

October will present a variation on the classic metaphysical question: “If a bunch of people get thrown out of work and no one is there to count them, did the economy tip into recession?” There may be efforts to recreate some data sets, but survey data is tricky when there were no surveys.

Even this month’s numbers will be challenging, because this week is the typical reference period for key surveys. These releases, when they come, will have to rely on much more inference and statistical modeling than normal. Some will be misleading, some will be wildly wrong, and some won’t even get released before the Fed’s next rate decision on December 10.

Then there’s the even greater challenge of deciding which of these measures may have been distorted by the shutdown itself. Are weak readings a temporary dip as furloughed government workers delayed big purchases? Or did the shutdown aggravate the downturn in the jobs market that has been building through the fall? Is inflation losing steam, or was there just a brief cooling because of weak demand?

Investors should brace for some surprising numbers that both bulls and bears will use to support opposing arguments. Even more than usual, they will embrace what seems to reinforce their instincts and reject contradictory data that must surely be erroneous.

Data analysis is more art than science in the best of times. But it may be at least until January before the best forecasters can feel they are once again on reliable ground.