If people bought their houses anticipating high rental income to make ends meet, that's a separate problem, and less bad than "people are becoming homeless because there are too few housing units".
It's not like the inventory of lots for sale doubles -- it stays exactly the same, but some of those lots increase in price because the ADU adds value.
> When a house with an ADU will reliably fetch 20% more than one without, then eventually all houses on the market will have ADUs, with the consequence that people buying will need to depend on short term rental income in order to make their mortgage.
If you significantly increase the number of units available via ADUs, then there's a lot more options to rent, and the demand to purchase houses goes down, bringing the prices down.
If the number of properties with ADUs doesn't go up enough to have that effect, then STR via ADUs will not significantly impact the housing market.
The reality is that if every house in the city had an ADU, most of them will not be used for STR, as the demand for STR isn't that high. They bring in a lot of money precisely because there aren't many options. If every house in LA suddenly had an ADU, you wouldn't get a proportionate increase in visitors. There will be an equilibrium point.
This is also ignoring the pain involved in building an ADU on an existing property. Depending on the locale, it can more than double your property taxes. Most existing homeowners will not build an ADU. Too much hassle, with uncertain income - most AirBnB hosts don't make much money via AirBnB - it's a fairly active investment and requires a lot of time/coordination.
Why would demand to purchase houses go down? People aren't buying houses because they can't find rentals. People buy houses because they want to be homeowners.
One classic example is the case of a very popular event happening in a small market that is otherwise unremarkable. Say, the Masters golf tournament: for about two weeks a year, Augusta, GA, is full of non-locals. Many residents will take vacation during this time and rent their home out to golf fans from around the world. Very few people want to rent a home in Augusta any other time of year. (It's also small enough that you're not required to fill out any additional tax forms, just report the income.)
Another is the case of a home that, for whatever reason, will not sell in the current market but would reasonably be expected to in relatively short order. A brief, 3- to 6-month rental for, e.g., a family that is constructing a new home that is not yet ready but who have sold their existing home and need somewhere to live is an excellent use.
Like I said, somewhat niche uses, but it's worth keeping in mind that there can be legitimate, minimally-distorting uses.
Why solve a problem when you can just nuke it, right?
I hang out with AirBnB investor folks. There's a simple solution that works: Require the owner live on the property. Every AirBnB investor will skip the city/county if that is the rule, but still leaves room for legitimate residents to make some money (e.g. rent out the basement, etc). The latter isn't taking housing away from anyone - for most owners it's not economical to convert the basement to make it a long term rental.