You are confused. The denominator includes the total estimated population. You could at least read the official definition before wasting everyone's time.
Here is the definition, from the first paragraph of the source that you yourself posted[1]:
The United States homeownership rate represents the percentage of occupied housing units where the resident is also the owner. A constantly evolving figure, the United States homeownership rate currently rests at 65.2%, while renter-occupied housing units make up 34.8% of the national stock.
From this definition it trivially follows that a rented property becoming vacant increases the homeownership rate, because the vacant property (and the person who was renting) will both no longer be counted, at all and thus the total percentage of owner occupied units will increase.
At this point I'm not sure if you failed to read your own link, don't understand what "percentage of" means, or are just obstinately wrong, but regardless it doesn't much seem worth continuing does it?
The Census breaks the homeless population into two groups - those whom live in shelters and those who do not. They completely ignore the later. For the former, shelters are not considered housing units.
> "Census Bureau surveys do not collect data on the
population experiencing homelessness and living on the streets"
