This post is an excerpt from Take the Threat of this Triple Whammy Seriously Now.
Tax data is reality. BLS jobs data is artistic impressionism. It is initially drawn from an infinitesimal survey sample as of the 12th of the previous month. Then the BLS statistically manipulates, massages, and fudges the number, primarily with two statistical crap machines. One is seasonal adjustment, and the other is the business birth/death adjustment. Both are crap, and the result, as often as not, is crap.
For a change, Friday’s jobs report was reasonably close to reality. But it’s irrelevant, and bond traders showed as much by not buying enough to keep yields from rising.
I noted the following about this in our last tax revenue update:
8/4/21 This weakness [in withholding tax collections] occurred in the last two weeks of the month, after the July 12 jobs survey date. But actual wages are paid in arrears, so the collections at the end of the month should reflect the number of people working as of the survey date. On a month to month basis, the change is negative.
Whether the BLS data reflects the facts in a given month is anybody’s guess. Their statistical massage process presents a mirage that only gets corrected toward reality in ensuing months and years.
Over time, the BLS fits their stats to what actually happened, as shown by unemployment claims and tax data. We just look at the tax data. In other words, why go for the impressionism when you can look at the reality?
After posting crap for a month or two, the BLS number eventually catches up to reality. If you’ve been around these [Liquidity Trader Federal Revenue]reports for a while, you know that they revise the previous two months, and then they re-benchmark the data every year for the next 5 years. That’s because the seasonal adjustment is based on a ten year average.
Just one problem. It’s based on the last 5 years, and a projection of what the next 5 years are expected to be. So they use past data and future data. Future data. Right. Data that doesn’t exist yet. As the real data comes in on the state and federal unemployment programs, and is cumulated at year end on tax data, they benchmark what they previously published for the last 5 years to the new actual data. It’s crazy.
The other way they get caught up with past reality, is to simply add or subtract the over or under estimate from previous months in the current month. We saw that in July, because they undercounted a couple of big gains in previous months.
Ironically, the economists who participate in the stupid “beat or miss” surveys, sometimes get essentially the right number. But the BLS misses big. And sometimes the economists get the wrong number because they overreact to missing the previous month.
That happened last month when there was a blowout BLS number the economic priesthood didn’t expect. So what did they do this month? They went too far the other way, boosting their guesses significantly, when they had been on the right track the month before. Those boosts left them slack-jawed when Friday’s lousy number came out.
The whole game is dizzying, and not worth your time from a swing trading or long term trend perspective. True, maybe you can use these very short term knee jerk reactions to the false news for entering a position, but that’s also a very tricky game.
I do not understand why the BLS doesn’t just use the tax data to build their reports. It’s real data, and it covers everybody. Sure, you need to do some smoothing, but that can be done with 7th grade math. Which is about my level.
The point is, let’s try not to get whipsawed by this BLS nonsense. Taxes are real. BLS jobs estimates, and whether economists guess right or wrong, is completely irrelevant. They have nothing to do with the direction of stock and bond prices.
8/4/21 By following the tax data, we knew that the US economy was undergoing an explosive rebound early in the spring, well before it was the mainstream view. The withholding data in recent months had called into question the BLS jobs survey accuracy. So we weren’t surprised by the strength in recent jobs data [July]. In March and April the jobs numbers were too weak. We knew that the BLS massage process had missed the strength for a couple of months. It had to start catching up.
So we shouldn’t care at all what the jobs number is. We should care about how much tax revenue the Federal government is taking in because that has a direct impact on Treasury supply. Since we know what that number is, we don’t need to do all kinds of extrapolation and interpolation to guess what economic data means for the market. That’s second and third order stuff. It’s useless.
We have the direct data we need to help us know what we need to know. What will the current and near term future be for Treasury supply, and how much will the Fed need to do to keep the rigged game going.
Now the question is why do I write so much about things that don’t matter?
Redeeming social value. In other words, we need to be reminded to focus on what matters, and either ignore, or laugh at the published garbage that everybody else follows religiously.