Don’t Confuse Uncertainty With Risk (Or Diamonds With Coal)

There’s been a mini bull market in discussions about uncertainty lately, thanks to the recent drop in the so-called Economic Policy Uncertainty Index. “‘Uncertainty’ isn’t a problem anymore,” declares The Washington Post’s Wonkblog, which Paul Krugman references by noting that “the index of uncertainty that everyone was pointing to has plunged, with no visible boost to the economy.”

The revelation, we’re told, is that the recent habit of equating heightened uncertainty in matters of policy (regulations, fiscal decisions, etc.) with slow economic growth is a dead issue. As Wonkblog advises, the “decline of uncertainty” has yet to unleash a noticeable upturn in hiring, relative to the trend in the recent past. The connection between uncertainty and the economy, in sum, has been resolved: there is none.
Taking the numbers at face value, it appears that uncertainty wasn’t a problem after all, or at least as far as blaming uncertainty for the economy’s troubles. If uncertainty has fallen and we’re still stuck in the land of sluggish growth, the connection between the two looks null and void.
Perhaps, but there’s a problem. The Economic Policy Uncertainty Index (EPU) isn’t really a measure of uncertainty; rather, it’s a measure of the perception of uncertainty, as defined by news stories, reports from the Congressional Budget Office on matters of federal tax code, and economic projections via the Federal Reserve Bank of Philadelphia’s Survey of Professional Forecasters. Yes, EPU is an interesting and innovative index. But its merits in matters of macro analysis are still an open debate at best. Uncertainty about the future, after all, is absolute and constant.
There’s nothing wrong with trying to quantify uncertainty, but measuring unknown unknowns is destined for failure if we look at the challenge with sober eyes and an open mind. By definition, uncertainty is a state of being clueless about the future. We can make guesstimates and crunch data for estimating what could happen, and we can go quite far down this road. But the future is still uncertain, no matter what we do in the here and now. That doesn’t mean we should throw up our hands and let fate wash over us. But it’s important to recognize that deciding what may be probable, or not, doesn’t reduce uncertainty. In other words, we must maintain a healthy respect for distinguishing between risk and uncertainty.
The economist Frank Knight forged the modern distinction for the two concepts in his 1921 book Risk, Uncertainty, and Profit. “There is a fundamental distinction between the reward for taking a known risk and that for assuming a risk whose value itself is not known,” he advised. Risks that can be quantified can be “easily converted into an effective certainty.” By contrast, “true uncertainty” is “not susceptible to measurement.”
As one simple example, imagine you’re about the roll the dice in a high-stakes craps game in Las Vegas. You’ve made a $10,000 bet on 8 as your point number. If 7 comes up on the next roll, you lose; if 8 comes up, you win. The risk here is clear. There are five ways to make 8: 5 and 3, 4 and 4, and so on. But there are six ways to make 7: 6 and 1, 5 and 2, etc. The probabilities work out to a roughly 13.9% chance of rolling an 8 on the next throw vs. a 16.7% chance of rolling a 7. Your risk, in other words, is slightly skewed in favor of losing. But that doesn’t mean that you will lose. Indeed, the outcome is uncertain because the outcome of the next roll is unknown.
A similar analysis applies to looking ahead in matters of the economy. We can analyze history in search of clues about how the macro profile could evolve. But this is an exercise in estimating the risk of a given outcome. But that doesn’t change the uncertainty, which is total in terms of what tomorrow may bring.
Some things just can’t be modeled. We may have a high degree of confidence that a particular outcome is likely, but any confidence must be based on what’s already happened. The critical factor, of course, is the mystery about what has yet to unfold. There are countless possibilities that could trip up our finely crafted projections for reasons that no one can anticipate. Granted, you can imagine any number of extreme outcomes and in that sense you can prepare yourself for almost anything. But rank speculation is hardly a robust way to model the future. In practical terms, we’re still beholden to uncertainty.
Let’s say you crunched the numbers thoroughly and decide that the economy is poised to grow in the near term. But you wake up the next morning and discover that the Strait of Hormuz in the Persian Gulf has been shut down in a new Middle East war. Crude oil prices instantly triple and the economy tanks. Yes, you might have prepared for such an event, but the news still comes as a shock and throws a monkey wrench in your model’s projections.
“Uncertainty must be taken in a sense radically distinct from the familiar notion of Risk,” Knight wrote. Thinking otherwise, no matter how many pretty graphs you create in Excel or R, won’t change that fundamentally sound advice. Keep that in mind the next time someone tells you that uncertainty is higher or lower. In reality, the future is always unclear, in absolute terms. Short of our transition into gods, that’s not going to change.