Earlier quoted context omitted.
That's a cool site, I made a graph that I think illustrates my point more clearly: https://fred.stlouisfed.org/graph/?g=B6ZP It's the change in population vs. the change in new housing with a separate line for each. It's not super easy to understand because the population change is in thousands of people and housing is just in individual units, but I think it's clear from the graph (if I'm reading it correctly) that…
Wouldn't you have to have some sort of adjustment for singles vs families over time? and for average number of people in a household? Number of people in a household has a cultural factor and age factor also (young single people often share households, more young people => higher density). I have no idea impactful those numbers would be, just pointing out that there are lots of confounding factors to consider.
I'd also guess a factor of young single people sharing households is due to limited supply and high cost. If it was affordable to have your own apartment, more people would do so. The roommate situation is partly (though not entirely) a symptom of failure to build.
You can see this when you compare regions, where I grew up (western new york) housing is super cheap even though there's a lot less economic opportunity. Nobody I know from high school that stayed there lives with roommates. Many have bought their own (nice, new) house for ~150k. That house in the bay area would be 3 million easily.
So while I agree there are confounding factors at play, the housing supply issue is so extreme I think the others are largely rounding errors.