No, actually I don’t worship at the altar of the CDC, but there is no other compiler of Excess Deaths for the US that I know of. I once found a UK site quoting them, but probably from the CDC. More important, the UK numbers look a lot like the US.
The Economist has a model for global: https://www.economist.com/graphic-detail/coronavirus-excess-deaths-tracker
Attribution during one stimulus package did indeed reward hospitals that coded Covid, but that item disappeared from PPP in a later stimulus package. There IS a requirement that all patients get a COVID test now, so docs and nurses know how what protection they need treating that patient.
This requirement likely does elevate case count, but not likely death count because if a patient is dying over a period of time, his test will make clear why.
Probably the most compelling issue in the world of attribution is the elderly person living alone who is aware that the vast majority of people recover. So they start coughing at age 73 and try to tough it out. They die. Alone. Smell brings an ambulance a few weeks later, and presto a death occurs that should have been but is not attributed to Covid, because never tested. This is a hugely common scenario. It is why all community police forces now have a “wellness check” process on the books, from a neighbor call.
Overall, the reason to largely believe attribution is how the death curves align to All causes EDs. In March/April last year, before lockdowns or media splash, the northeast spiked deaths on the attributed graph and it looked just like EDs. Nothing else was happening in March/April. No meteor hit NYC. Just old people were dropping like flies, and Covid declarations aligned.
You say, "..there is no other compiler of Excess Deaths for the US that I know of. I once found a UK site quoting them, but probably from the CDC." Correct, as the Economist estimates from the CDC, so gathering data is seen but estimating is also seen. As to data, the taxonomy and coding with precede excess death calculations still are interesting as I had mentioned before. With changing definitions, coding and data acquisition, much remains a question.
Still amusing was from mid 2020 -- "WHAT'S UP, DOC? Dr Deborah Birx ‘doesn't trust the CDC’ as experts fear coronavirus cases have been INFLATED."
https://www.thesun.co.uk/news/11599541/dr-deborah-birx-doesnt-trust-coronavirus-cdc-data-inflated/
Johns Hopkins today reports 5,353,969 global deaths, and US (as 4.25 percent of the world population) deaths at 806,439.
As a small share of the global population and assuming that the virus affects humans somewhat equally, the US' deaths should be about that percentage of the global deaths. It is not. 227,544 deaths would be that number. The number collected from their data set is currently the larger 806,439 as of today. Almost three times the global toll. But....
India with its 1.4 billion is half the death toll of the US, at 477,158 as of today.
Your Economist site states, "In India, for example, our estimates suggest that perhaps 2.3m people had died from covid-19 by the start of May 2021, compared with about 200,000 official deaths." Which data are correct? That collected or that estimated?
Their "estimate" of 2.3 million is far larger than the JHU' tabulation of 477,158 reported today, which is larger than a normative estimate. More than four times as many as that counted and reported. So the data and the Eocnomist-admitted "estimates" vary hugely.
All the excess death calculations do not explain such data discrepancies. This is because they tabulate with their data collection and estimated excess deaths. Before the data is collected is the issue.
"From" and "with" have been conflated. "Including the "assumed" in reports has been the WHO methodology since April of 2020. The Economist "corrects" a value for India, as above, with an estimate wildly outside any data set I have reviewed.
So my skepticism after looking at your sources actually has been reinforced.