St. Louis Federal Reserve watchers, rejoice! And yes, I’m talking to both of you. The St. Louis Fed is freshly relevant this week thanks to a paper it published back in 2007 that examined the economic effects of the 1918 Spanish flu. Drawing on old newspaper articles, local surveys, and other studies—national data back then was scarce—the report found that the damage done to businesses by the outbreak was both severe and short-lived. The impact on the next generation, however, was longer-lasting. Those in utero during the pandemic went on to attain less education and lower incomes than had previous generations.

What we wouldn’t give for that kind of glimpse from the future today. The coronavirus has killed hundreds of thousands while sledgehammering the economy, leaving close to a quarter of working-age Americans either unemployed or underemployed. And we still have no idea how it will end. It may be that this recession is similar to the one in 1918, cutting deeply but easing rapidly. Or it may be that we’re in for another lost decade of stubborn unemployment and stagnant growth. It may be that the virus is seen off this summer, remembered as a frightening but ultimately brief ordeal. Or it may be that it lurks into the autumn, whereupon it comes roaring back.

We don’t know, and we hate that we don’t know. Consequently a cottage industry has sprung up around our uncertainty, hawking models, projections, expert opinions. These things have valid scientific purposes, of course, but thrown down the rabbit hole of our popular discourse, they’ve taken on a kind of hysterical clairvoyance, supposedly able to tell us what’s coming and how we should respond. With climate change, we grew accustomed to the idea that scientists could see into the future. Now we’re demanding they do the same with the coronavirus. That’s despite the fact that so far, none of these projections have demonstrated any greater predictive ability than your average call to Miss Cleo.

Take the government’s official death toll projections. Back in January, the White House was largely complacent over the coronavirus, with President Trump comparing it to the seasonal flu and his health secretary saying that Americans need “not worry for their own safety.” Then in late March, the pendulum swung towards apocalypse. Actually, the White House said, 200,000 Americans could die. Two weeks later, the death toll projection fell to a far rosier 60,000, and the country breathed a sigh of relief ahead of Easter weekend. Then the projections ticked upwards yet again. Today, IHME, the White House’s principal modeler, predicts that 147,000 Americans will be killed by August 4.

Some of the issue here may be the choice of models. IHME has been criticized by epidemiologists, as have the Imperial College modelers in Britain (who have lately been distracted by, er, more extracurricular activities). But the bigger problem is best summed up in a quote to Politico by the head of IHME, explaining why his organization’s projections were so wrong. “We had presumed, perhaps naively,” he said, “that given the magnitude of the epidemic, most states would stick to their social distancing until the end of May.” In other words, the models are premised on assumptions that can be scrambled by real-world events, whether political decisions or acts of God or the caprices of the virus itself. They aren’t showing us the future so much as extrapolating off of a snapshot, one that can easily change. Yet we treat them as practically mystic. “200,000 could die!!” scream the headlines, with “could” ever the weasel word.

We don’t just do this with the death toll. On the economy, too, we seem hopelessly confused. Here’s a smattering of headlines from the past two months: “Unemployment rate could exceed 20% by June, top White House adviser says.” “Economists see uneven jobs recovery, high U.S. unemployment through 2021.” “Top JPMorgan investment advisor: It will take ’10 to 12 years’ for U.S. employment levels to return.” “The coronavirus recession will be deeper and faster than the financial crisis.” “Economists say quick rebound from recession is unlikely.” “Trump’s baseless claim that a recession would be deadlier than the coronavirus.” “U.N. warns economic downturn could kill hundreds of thousands of children in 2020.”

Stare into this blurry puddle long enough and you might conclude that no one has any idea what the hell they’re talking about. Or you might fall back on your own biases, choosing to believe stories that buttress your political beliefs and speak to your own personal circumstances. Either way, this kind of confusion can have long-reaching effects. Consider, for example, a new study that was released last week, which found that there could be 75,000 so-called deaths of despair—meaning suicides and drug and alcohol overdoses—as a result of the coronavirus recession. It called to mind another social science finding, one of the most consequential of the last decade: that life expectancy among less educated, middle-aged, white Americans was declining, driven primarily by those deaths of despair.

That claim, courtesy of researchers Anne Case and Angus Deaton, made its way around the internet. It fed into the narrative of the populist right and Donald Trump. It provided an empirical grounding for “American carnage.” But wait: a less noticed study a year later, which took Case’s and Deaton’s data and adjusted for age, found a more mixed picture. According to research from Columbia University, while middle-aged white women had indeed seen increased mortality rates, middle-aged white men had reversed this trend back in 2005. And then came another study, in the American Journal of Public Health, that challenged the very concept of “deaths of despair,” warning that “the gap between deaths of despair as a claim and deaths of despair as a rigorously tested scientific concept is wide.”

There is a Grand Canyon-sized gap between what we think we know and what we actually know. How to navigate this chasm? Two maxims can help.

The first comes from Friedrich Hayek: “The curious task of economics is to demonstrate to men how little they really know about what they imagine they can design.” Hayek was concerned with what he called the “fatal conceit,” which he defined as the belief “that man is able to shape the world around him according to his wishes.” We might add a corollary: that man is able to anticipate the world around him according to his wishes. Because knowledge is complex and dispersed, Hayek argued, no one can ever marshal enough of it to centrally plan an economy. Likewise even a sophisticated model can’t have enough data to foresee how a pandemic will play out. There are simply too many variables, drawing on too many areas of life.

The second maxim comes from a very different source: John Dickinson, perhaps our most conservative founding father. “Experience must be our only guide,” Dickinson said. “Reason may mislead us.” Of course, by reason, he didn’t mean vast computer algorithms struggling to track contagion across seven continents; he was thinking of 18th-century rationalism, which he contrasted with the more reliable yardstick of historical experience. While what seemed philosophically sound in the abstract could be tainted by personal bias or disconnected from real life, precedent was far more settled. How something had worked in the past was a good indication of how it would work in the future.

Unfortunately we have very little precedent when it comes to the coronavirus, though the Spanish flu can perhaps offer some clues. The 1918 influenza, like the current pandemic, began in the spring, only to enter a second wave in the fall that killed more people than the first. A third wave then began that winter and stretched into the summer of 1919. That’s chilling, yet there’s good news too: the recession that followed was short and quickly blossomed into the 1920s, one of the most dizzying economic expansions in our history.

So top hats and flapper dresses all around? Who knows? It’s called the novel coronavirus for a reason. The awful truth is that we have very little idea how long this will go on and how it will ultimately turn out. And the reason for that is that we know so very much less than we think we do.