It is OK to say a recession is about to hit the U.S., and it is OK to predict one won’t. What probably isn’t OK is to put too much confidence in either forecast these days.

There is plenty of evidence with which to argue the U.S. is on the cusp of a downturn. The yield on the three-month Treasury bill, for example, is well above that on the 10-year Treasury note. In the past such yield-curve inversions have almost always augured a recession. Indeed, a yield-curve-based model from the Federal Reserve Bank of New York puts the odds of a recession occurring over the next year at 71%.

Meanwhile, in another precursor to past recessions, the Federal Reserve has raised its target on interest rates a lot in a hurry, and it might not be done. Borrowing costs are rising, and, with the recent cracks in the banking system, lenders are tightening standards. Consumer-confidence measures are in the gutter, and corporate profits are falling.

Then there is the evidence that a recession isn’t coming. The labor market remains extremely strong with 339,000 jobs added last month. And, considering the very large number of unfilled job openings still out there, it doesn’t seem as if the economy is about to start shedding workers. Some industries, such as healthcare, are still dealing with shortages.

The automobile sector, finding it easier to obtain needed semiconductor chips and other materials, is finally able to deliver more of the cars, SUVs and pickup trucks people had been waiting for. Light-vehicle sales and production are rising as a result. The housing sector also looks as if it has moved past its nadir, with home sales picking up from the end of last year, and home builders becoming more optimistic. This matters because declines in the automotive and housing sectors typically have typically been features of, and major contributors, to past recessions. When they are on an upswing it is hard to see the economy going into a downswing.

Some recession signals started flashing more than a year ago and now look like false alarms. Take the so-called Anxious Index: In each quarter since the late 1960s the Federal Reserve Bank of Philadelphia has asked forecasters what they think the chances are that gross domestic product will decline in the following quarter. Before the pandemic, if it got to 20%, the chances a recession was coming soon were high. If it got to 40%, a recession was probably there. It hit 20% in the second quarter last year and cleared 40% in the fourth quarter. So far, no recession.

The rising Anxious Index was one reason Sean Snaith, the director of the University of Central Florida’s Institute for Economic Forecasting, was convinced a year ago that a recession was coming. When The Wall Street Journal polled economists last June on the chances of the economy slipping into a recession within the next year, he said 99%, and others were nearly as confident.

He notes that gross domestic product contracted in the first two quarters of last year, which usually happens during recessions. But the National Bureau of Economic Research, which has been the arbiter of U.S. recessions since before the U.S. government reported GDP, looks at an array of indicators, including employment, income and industrial production. GDP data are also subject to revisions: Of the 29 quarters since 1965 where GDP was initially reported to have declined, 7 are now in the positive column while several initially reported gains are now marked negative. Snaith is less sure about his recession call: “I’m down to a coin flip now.”

Overconfidence among forecasters isn’t a new phenomenon. Don Moore, a professor at the University of California, Berkeley’s Haas School of Business, and Sandy Campbell, a graduate student there, have found that the economists surveyed by the Philadelphia Fed have consistently been too convinced that their forecasts are accurate. It isn’t just economists. “As human beings, we want to be certain,” says Moore. “The truth is we live in a highly uncertain, probabilistic world.”

If ever there were a moment to put bigger error bars on economic forecasts, though, now would be it. The ease with which one can tell a story about how a recession is coming, and with which one can tell the opposite story, is a reflection of how out of whack the economy is after all the distortions the pandemic introduced.

The problem is those stories can be so convincing, told with the confidence of a student essay generated by ChatGPT. It can seem wishy-washy to not express certainty about what will happen. But for those who need to balance the risks of recession with the opportunities of continued expansion, a willingness to utter a hard-to-say three-word sentence might still be in order: “I don’t know.”