That means that it's sufficient for the lower income tiers (relative to the house value) to have a few months of high interest rates (in a time frame of 15-25 years) in order to lose the house.
Buying at the current rates is risky, while the risk of rising rates in 1988 was much lower. Factoring that in should change your conclusion.
http://www.freddiemac.com/perspectives/sean_becketti/2017041...
People got burned bad by ARMs in the housing crash of 2009 and they still aren't willing to take the risk.
("Yes and") There's also very little rate benefit to adjustable mortgages right now, so it's quite sensible to take a fixed rate mortgage. When I bought my first house (in 1996), I recall that fixed rate mortgages were several points higher than adjustable rate mortgages during their fixed portion. In cases like that, it made for an actual decision.
With both fixed rates and ARMs having roughly the same initial rates, there's little reason to choose adjustable.
1. Every idiotic meme about how goods X, Y, and Z have more than doubled in price while [real] wages haven't. I'm torn between thinking the meme creators are completely clueless or epic trolls.
2. There is a substantial linkage between nominal wages and nominal inflation, especially for limited, competitively bid goods (like housing).
Looking at median house prices https://fred.stlouisfed.org/series/CSUSHPINSA , it looks like they went up slightly more than median wages (2.85x vs 2.5x in 30 years), which I think is likely more than explained by today's lower than historical norms interest rates (increasing affordability). https://www.macrotrends.net/2604/30-year-fixed-mortgage-rate...
Mortgage rates over 10% in 1988 and under 5% today means that 1988 houses were less affordable on 1988 wages than 2018 houses are on 2018 wages.