If I follow this correctly, Daniel, it seems the chain of events is that
- Speculative developers build lots of new homes
- Economic downturn forces poorer workers to economise on housing
- Economy takes the form of greater concentration (house sharing) rather than moving in search of lower rents
- This deflates demand despite rising population
- Speculative developers scale down housebuilding.
FWIW, (3) is the logical approach in an era when commuting was problematic (more people walked to work so major relocation was inefficient), economic downturn encouraged one to hang onto work rather than relocate, and social networks were far more local.
I also don’t see that spending 30–40% of disposable income on housing is unusually high. In most OECD countries today, the poorest fifth of the population will spend about that on rent, while overall the average rent is 20–30% of disposable income.
I’m not sure that what you have described demonstrates that “there was never a time when the market alone was delivering quality, cheap housing in abundance.” That appears to be exactly what “the market” was doing in the pre-WWI period, and it did so again in the 1930s.
What it *does* suggest is that the market can be cyclical and that the relationship between population and *demand* is not linear. This is certainly true — demand for housing is highly income elastic; at the other end, as people become richer they demand much more space, such that they probably continue to pay 20–30% of income on housing even if their income doubles.
(I have a couple of graphs to support this but there is no way to insert here so will tweet).