#26 - Out of Sample
No Y-Axis is Safe
No y-axis is safe in the Covid-19 world. Events that seemed implausible a few months ago are now occurring regularly.
Consider US unemployment claims. From 1967 through 2019, the number of weekly claims never exceeded one million. In contrast, the past six weeks have seen between 3.3 and 6.9 million:
Source: U.S. Employment and Training Administration, Initial Claims [ICSA], retrieved from FRED, Federal Reserve Bank of St. Louis; https://fred.stlouisfed.org/series/ICSA, May 3, 2020
Similar story with oil. Historically, a barrel of oil was worth something. For a brief moment recently, oil was worth less than nothing:
Source: New York Times, Too Much Oil: How a Barrel Came to Be Worth Less Than Nothing, April 20, 2020
In general, it’s not a bad prediction that the near future will look similar to the near past. However, that’s becoming a riskier proposition these days.
What happens when huge swaths of the economy are shut down? What happens when social animals are faced with extended periods of isolation? What happens when both of these occur at the same time? We don’t know because we’re out of sample.
Let’s say that you live in Los Angeles. Every morning you jog from Santa Monica to Marina del Rey and back. You’ve done this for the past decade, 3,650 times in total. Typically it’s between 65 and 75 degrees outside and you complete the eight mile run in 60 to 62 minutes.
Then one morning something is very off. You look outside your window and there’s a hurricane. And a tornado. And a plague of locusts. Not feasible, but bear with me.
Being a creature of habit, you lace up your running shoes and hit the road. How long should your jog take today?
The problem here is that your previous 3,650 runs get thrown out the window. You’ve never run in these conditions before so your past experience won’t be that helpful in determining your outcome. You’re out of sample.
In a sense, this is where we find ourselves today. Some old assumptions no longer apply. History is a less reliable guide. In The Signal and the Noise, Nate Silver notes that:
“We forget - or we willfully ignore - that our models are simplifications of the world. We figure that if we make a mistake, it will be at the margin. In complex systems, however, mistakes are not measured in degrees but in whole orders of magnitude.”
As unemployment claims and the price of oil illustrate, the world isn’t working the way it was a few months ago. Time to draw wider axes.
👉 If you enjoyed reading this post, feel free to share it with friends!
For more like this once every weekend, consider subscribing 👇