May’s Job Numbers Are No Surprise.
As I checked the notifications on my phone this morning, I found myself scrolling through another fresh litany of hot takes, the sort that have become commonplace these days. Sadly, egregiously dishonest articles that would have previously shocked me are now the norm, not the exception. In fact, the only time I find myself even remotely surprised by the news is when they decide to switch it up, and actually report something positive. Today brought just such good news, in the form of May’s staggering decrease in unemployment claims, and the addition of 2.5 million new jobs. As prone as I am to be stunned by the media’s admission of anything good taking place in the world, the only thing shocking about these particular headlines was the surprise with which they were delivered.
“Shocking” said CNN. A “May Surprise” declared NPR. “Surprising data” proclaimed NBC News. “How did the experts get it so wrong?” they asked. I don’t know, but probably the same way “experts” get things wrong all the time: the use of flawed hypothetical models over real-world experience and observations. Rejecting the ‘Fat Tony’ approach, as Nassim Taleb would call it. I’ll be the first to admit that I’m no expert, so the headlines about how wrong all the experts were is perhaps technically correct. But I found these job numbers, and the continual rise in the Dow to be anything but surprising. On the contrary, this is precisely what I assumed would happen, though admittedly I was off by a month (I originally predicted June as the first month of solid unemployment data, and September as the month when the numbers are closest to accurate). While it’s always fun to be ahead of the curb, my primary goal here isn’t ego stroking, but rather to explain why some rando who studies economics as a hobby and makes YouTube videos in his bathroom/office had this correct from the beginning. I want to help explore the simplicity in how I got here, and elaborate on which noise is worth turning up, and which noise is worth turning down. For those of you already bored, the TL;DR version is that the “experts”-
1) Focused on the forest (their models), not the trees (individuals in the real world). They looked at unemployment numbers, not what (and who) those numbers represent
2) Treated those unemployment numbers as = unemployment numbers from any other month under any other circumstances, and fed those faulty assumptions into their stupid models
3) Downplayed the impact of emergency legislative measures
4) Underestimated American employers and workers
5) Dismissed any notion of “hitting pause” on the economy
6) Perplexingly applied “conventional wisdom” to a situation that was completely and unequivocally unconventional
There is likely another factor to consider, and that is the propensity of virtually every talking head to oppose whatever it is that President Trump says. Since President Trump was endlessly optimistic about the potential for economic bounce-back, many in the intelligentsia and mainstream media took a different route. However, this doesn’t explain why many Conservatives also treated the economy, and more specifically, the unemployment numbers, as accurate. Last night I was discussing the Dow numbers with some friends, and again lamented for the 80th time in as many days “Why is Ben Shapiro treating the unemployment numbers like they’re even remotely accurate? He knows better.” I like Ben, but I think he has one thing in common with many of the other folks who got this so wrong, and it’s the one thing that often prevents them from seeing and understanding many of these things: busy urgency + detachment.
We live in a pretty small house. I don’t make a lot of money, but that’s ok, because we’re fairly low-maintenance. I like to go for walks around my community, and talk to business owners at the places I frequent. I don’t feel pressed to constantly present an opinion or hot take on something every single day, so I’m able to take time and think through my observations and subsequent conclusions. Here are a few of the things I saw and experienced during the beginning/middle of the lockdowns:
-movie theaters selling “curbside popcorn” as a way to keep customers engaged
-restaurants, including fast food, quickly switching over to delivery and online-ordering
-breweries offering beer vouchers with delivery orders, for free beer when things opened back up
-a local butcher shop experiencing the biggest boon in business they’d ever had, according to the owner
-I sold a car I’d had listed for months
-Food trucks with long lines of social-distanced customers, patiently waiting for their food
-The grand opening of not one, but two restaurants. One of them simply didn’t open their dining area, and did drive-through only. On opening day, the line was wrapped all the way around the building, and extended into the parking lot across the street
-I visited this restaurant a few weeks later, again seeing a long line. The woman at the window told me she had lost her job a few weeks prior, and moved here to work at this restaurant. She said the pay was good, and seemed pretty happy.
-A friend who has his own contracting business, and a friend who works at a lumber yard, both telling me they’ve never been busier in their lives. My contractor friend is currently backed up by 6 months on jobs.
-I personally made thousands of dollars selling digital weapons and armor on an Xbox game. No, I’m not kidding. Yes, that’s a thing. Let’s not get too distracted here. The point is, that does not happen during a recession.
These are just a few of the things I noticed during the beginning, and a few weeks into the lock down. Here’s the next piece, the vital “trees” that were missed:
Every person I spoke with who filed for unemployment still had a job. People in California, Washington, Oregon, Illinois, Ohio, Missouri, Wyoming, Florida. I spoke to dozens of friends, family, and acquaintances. Some were encouraged by their employers to file unemployment for March and April, with employers predicting getting back to work in May. Others filed for only a few weeks, while their employers waited on PPP loans. One friend said his employer was paying him as if he was working 40 hours a week, despite only working 18. He said the decrease in actual work hours qualified him to file for unemployment, and his hair stylist wife was also able to collect unemployment. As soon as Ohio opened back up, she made $1200 on her first day back. My own barber reported something virtually identical. Every barber at that shop filed for unemployment for the month of April, and simply got right back to cutting hair when we opened back up. My barber’s wife, who has an entirely different career, did the same. One restaurant owner told me they were doing a “soft hiring” soon, and hoped to beat the other restaurants to the best help, as he knew of several other establishments that would be doing the same. Recently, I had someone tell me that despite never losing or slowing down on her primary job, the fact that she lost her second job (a part-time server gig), she qualified and had filed for unemployment.
Obviously these are all personal anecdotes, but that’s without even digging into the reports of folks around the country who opted to collect unemployment not due to job loss, but because it paid more than their regular job. However, an anecdote is an anecdote. And I do truly detest anecdotes as a substitute for empirical data. However, the opposite is also true. Empirical data (“expert models”) that is found to be in tension with everyday observations is more dubious than an anecdote presented as empirical data. If necessary, please read that last sentence again before continuing.
Think of it this way: you don’t need to do a 6-month double-blind study to empirically determine the effect of throwing a horse off a skyscraper, regardless of whether or not you’ve ever actually seen a horse thrown from a skyscraper. In such a ridiculous and gory hypothetical, who would you trust more: the everyday person who says “Dude, that horse is going to splat everywhere.” Or, an “expert” who confidently informs you that their modeling indicates the horse might actually somersault away from the point of impact, and ride off into the sunset?
Back to the issue of unemployment numbers, it’s important to briefly note what I’m not saying. I’m not saying that every one of the individuals represented in these numbers is someone who remains gainfully employed. Obviously many of those folks have indeed become unemployed in the sense that we’re accustomed to thinking of it. My main point is that the majority of those numbers do not reflect people who are experiencing that kind of traditional unemployment. Jobs temporarily unable to be performed are not the same as jobs lost.
This leads us to the next point, and it’s a bit technical. Fortunately, it’s only technical like using pi to describe a rolling object, otherwise known as a wheel, is technical. The basic assumption of unemployment numbers (and their implication on the overall economic climate) is first crudely based on some approximation of this formula for individual economic stability. Assume:
(-A) + (B) = +/-C
Where
(-A) = Monthly expenses (often mistakenly assumed as static)
(B) = Monthly revenue (often mistakenly assumed as static)
(+/- C) = Leftover capital/purchasing power (individual economic stability)
This obvious oversimplification remains a decent basis for understanding the implications of any given individual’s state of financial stability. Those assumptions are then used to fuel another chain of assumptions:
Lack of employment = lack of sufficient capital/purchasing power
Lack of sufficient capital = inability to pay bills/spend money into the economy
Inability to pay bills/spend money into the economy= falling behind/further into debt/potential homelessness/falling business profits/falling profit motive to start a business/ wage stagnation/ etc.
Like the last formula, this set of assumptions is also a decent method for understanding the relationship between unemployment numbers and economic stability under normal circumstances. But applying it to what took place in our economy over the past several months is nothing short of insane.
Where to start. First, between expansions of unemployment benefits and the PPP, the “Unemployment = lack of sufficient capital” assumption is not applicable, or at least not as applicable as one might be led to believe. Second, all assumptions about the static nature of bills and other expenses also fail to apply as they are understood in the contemporary sense. Temporary suspension of rent and mortgage payments, virtual 100% reduction of gasoline expenses, and inability to spend money on outside entertainment changed the nature of expenses for everyday citizens. Again, the presumed static nature of core inputs and outputs is often what creates wild extremes and general lack of accuracy with many of our “expert models.”
Alright, technical talk is over. Back to the question and explanation at hand. Two final thoughts for you:
1) Anyone pressured to quickly provide explanations to complex and evolving situations will take unconscious, mental shortcuts to save the expenditure of unnecessary intellectual bandwidth. In simple terms, someone much smarter than myself (this is a very long list) but who is also pressed for answers which no one at the time possesses, will often use faulty defaults as inputs into their mental calculus. In this situation, you end up with talking heads from all over the political spectrum looking at unemployment benefit claims and conflating those benefit claims with job loss, in a situation where the two are not at all axiomatically interchangeable.
2) The subconscious falling back on mental defaults in situations where they are not as useful to us is an example of turning the volume up in the wrong area of informational inputs. By leaning on these defaults, there is a proportional turning down of the volume on information that might contradict the originally established assumptions. Since I didn’t personally experience job loss, trying to “understand” the unemployment situation was not something I mentally attempted to do. It just wasn’t pressing for me, and neither was having a canned explanation for some daily podcast or news broadcast, either. I was able to set my default to ‘Fat Tony’ mode, thus avoiding the Ludic Fallacy: in unprecedented situations, use observation to form conclusions, rather than trying to fit observations into a model concocted in a lab. If a coin comes up heads 99 times in a row, what are the odds of the 100th time being heads? “Models” tell you that the odds are 50:50. Five seconds of critical thinking based on your observations shows your models to be utterly worthless, and there’s clearly some information that you’re missing (it’s probably a rigged coin, and you should definitely bet on it being heads a 100th time).
I saw the catastrophizing over the “staggering unemployment numbers” every single day, just like everyone else. But I looked around and saw those numbers reflected absolutely nowhere within my scope of observation. Thus, I trusted what I could see in the absence of any supporting evidence for the apocalyptic vision being fed to me.
History indeed doesn’t repeat itself, but it does often rhyme. I wanted to pull my hair out every time someone tried to make Covid 19 predictions based on what happened during the Spanish Flu outbreak. This is asinine. History can be used to help us understand what we don’t know much better than what we do know, by honestly identifying how situations are different. This is what it means to use “via negativa” as a means for understanding. By describing what something isn’t, we can equip ourselves to discover the gaps in our knowledge or understanding about what that something is. One of the most useful things we can learn about Covid 19 from looking at the Spanish Flu Pandemic is how completely different these situations are from each other, and thus the futility of filtering Covid 19 through the Spanish Flu lens in the first place.
The good news is that most of us don’t experience the detachment and tyranny of the urgent that many of our intellectuals in the model-making and talking head community suffer from. We have the luxury of considering observations much more carefully than they can. This is my final encouragement to you: yes, listen to people who are able to authoritatively explain something occurring, or that has occurred, within their respective field of expertise. Thus it is perfectly reasonable for the announcer at a professional football game to say “I predict the play clock will run for approximately 60 total minutes over the course of this game.”
However, the minute that person begins to make future predictions in a new and dynamic situation based on their understanding of the past, consider turning the TV off. If that same announcer was observing a different but similar game and said “I’m not sure what sport they’re playing, and I know they’re making the rules up as they go, and the clock itself is broken, but I predict the play clock will run for approximately 60 total minutes over the course of the game,” you would have serious reservations about the reliability of that claim. Sometimes “expert analysis” falls into the first category. Sadly, it’s often the second, also. Ingest grain(s) of salt as needed.
Author’s Note: It’s worth mentioning that my discussion of any positive outcomes to the CARES Act is not an endorsement of the act itself. There’s a compelling argument that government is ultimately responsible for the destruction is causes, including economic destruction to the lives of its citizens. That said, I’m far from convinced that MMT is a recipe for the long-term economic stability of a country.