Ok. Maybe I'm misunderstanding the author. Let's take those quotes in context from the conclusion of the article: "There are numerous factors that overstate an

Ok. Maybe I'm misunderstanding the author. Let's take those quotes in context from the conclusion of the article: "There are numerous factors that overstate and understate the actual number of unemployed, and we’re not able to pin them all down. I mentioned seasonal adjustments as one of those factors. But there are also fraudulent claims that overstate claims data. And then there are claims that have been denied for various reasons, often due to mismatched information from the employer that take weeks or months to sort out. And they would understate unemployment. Then there are states that still don’t process all federal claims. And even states that do process them, do so inconsistently. Then there are people who don’t qualify for unemployment insurance, and when they lose their work, they don’t count here at all. But give or take maybe 1 million people, the total number of unemployed still hovers around 32 million. That’s the number to go by, not the fallen-off-the-cliff nonsense produced these days by the Bureau of Labor Statistics in its now ludicrously distorted survey-based jobs report." I take "there are also fraudulent claims that overstate claims data" to mean that some non-zero number of "fraudulent claims" presumably are counted in the set of all unemployment claims, and that this is one of several subtleties in the data that might contribute to marginally both under- and over-reporting. It seems to me that BLS is just one of the reporting entities, and that the survey results they provide are regarded by the author as incomplete an unreliable self-report, but not that all of the various sources are providing unreliable data (although it’s not clear exactly how it is all collated). Is that incorrect?