U.S. added just 73,000 jobs in July and numbers for prior months were revised much lower

[progree]

97. Checking against private sources sounds smart, but warning: they normally vary a great deal from the BLS numbers

for example private payrolls from ADP vs. BLS

The first column is the BLS private payrolls number month-over-month increases https://data.bls.gov/timeseries/CES0500000001?output_view=net_1mth

The second column is the ADP private payrolls month-over-month increases https://fred.stlouisfed.org/data/ADPMNUSNERSA (I had to do spreadsheet work to come up with the month to month differences)

The 3rd column is the difference between the BLS and ADP numbers

The first 2 columns are month-over-month increases beginning with February 2023. In thousands.
Positive ones in the 3rd column mean that the BLS number was higher than the ADP number.
BLS ADP DIFF
250 157 93
48 -53 101
167 132 35
166 52 114
170 146 24
110 138 -28
108 119 -11
89 95 -6
117 130 -13
97 100 -3
213 95 118
73 129 -56
151 91 60
169 82 87
129 113 16
160 164 -4
66 136 -70
40 42 -2
33 180 -147
208 194 14
-1 221 -222
244 204 40
287 176 111
79 186 -107
107 84 23
114 147 -33
133 60 73
137 29 108
74 -23 97 - ADP revised from -33 to -23
83 104 -21 - The latest, for July, reported 7/30 by ADP and 8/1 by BLS

===============================================================

I'm also reminded of the wild difference between two BLS employment measures:

https://www.democraticunderground.com/10143505265#post17

The headline payroll job numbers (+73,000 in July) come from the Establishment Survey
https://data.bls.gov/timeseries/CES0000000001
Monthly changes (in thousands): https://data.bls.gov/timeseries/CES0000000001?output_view=net_1mth
YEAR: JAN FEB MAR etc.
2022: 225 869 471 305 241 461 696 237 227 400 297 126
2023: 444 306 85 216 227 257 148 157 158 186 141 269
2024: 119 222 246 118 193 87 88 71 240 44 261 323
2025: 111 102 120 158 19 14 73
The last 2 months (June and July) are preliminary, subject to revisions
Last 12 months: 128k/month average

# Employed in thousands (down 260,000 in July) come from the separate Household Survey, http://data.bls.gov/timeseries/LNS12000000
Monthly changes (in thousands): http://data.bls.gov/timeseries/LNS12000000?output_view=net_1mth
If one adjusts the date range from 2021 to 2025, the graph is much more meaningful because it leaves out the huge swings of 2020 that greatly enlarges the Y axis and makes what follows look like tiny almost undiscernible squiggles around the zero axis
YEAR: JAN FEB MAR etc.
2022 1016 483 608 --315 487 --284 164 477 75 --126 --177 752
2023 958 178 417 162 --178 183 204 292 --33 --231 675 --762
2024 66 --177 412 70 --331 --9 64 206 377 --346 --273 478
2025 2234 --588 201 461 --696 93 --260
Last 12 months: 157k/month average
January and February of each year are affected by changes in population controls.
A very volatile data series from month to month. I used a double minus to make the negative ones stand out a little better
This Household Survey also produces the unemployment rate and labor force participation rate among many other stats

Assuming these are all honest numbers up to now, it just shows how different surveys get different results.

So I hope that a Nobel economics prize winner can tell us, in the future, when surveys vary wildly, which part of that is manipulation, and which part is, well statistical noise and methodology differences.

This posting stimulated by "Caracas on the Potomac (Krugman)"
https://www.democraticunderground.com/100220530941

From here on, I’m going to be paying a lot more attention to private surveys. And when they tell a different story from the official numbers, there will no longer be a reason to take the official data more seriously.