If you meet their demand with new housing, the demand and price pressure on existing housing inventory will decrease, even if you don’t build any new “affordable” housing.
How are those the two options?
Aren't you forgetting about new construction in established neighbourhoods? I.e. buying up two or three low-density lots, or a older/decrepit low-land-value condo building, to replace these lot(s) with a modern, high-density mixed-use condo building, composed mostly of high-sqft condos that cost more than most SFHs?
There are more units on the new lot than there were on the old lot(s), so the people buying these units are definitely "moving into new construction"; but at the same time, the creation of these units displaced previous low-cost rentals.
This is 99% of what people in my city (Vancouver) are complaining about when they talk about gentrification through "luxury construction": being priced out of a neighbourhood they previously rented in, as buildings containing low-price rental units are torn down and replaced by buildings containing only high-price rental units.
You are correct, that it would be possible to build new luxury housing and then lower the cost of old luxury housing until it becomes affordable. However, that seems unlikely to actually happen since the owners of the old luxury housing are likely to resist losing money. In some cases, luxury housing may have features (size, density, amenities, etc.) that make is economically unviable to price it as affordable housing.
Reality is that it's too expensive to stay a "luxury" housing, all the labor costs and major renovations just don't make sense from a financial perspective.
The fundamental issue that people miss is that the cost of building a "luxury" housing and "affordable" is not that different. Throwing quartz countertops and a tiny gym with outdated cheap equipment for a few extra k is nothing when the place costs 500k+. Sizes are also not relevant, since newer apartments are smaller than older ones. We can't just build an old affordable house, that's it. We need to start building as much as we can and impact the demand by evaluating the impact of new businesses coming to the area.
You could evict a bunch of people from existing houses and then re-sell those within the constraints of affordable housing parameters, but I doubt that is a feasible method of significantly scaled up expansion the pool of available affordable housing.
See Oakland.
Gradually? That's roughly what's happening in Denver right now - nobody's been building anything but luxury castles for years, but we saw our average rent drop last quarter for the first time since the recession.
Reason? There were 11,000 more units available than there were this time last year and there are 40k more on the way, so apartment managers are having to think about vacancy rates instead of myopically focusing on profit per unit for the first time since... probably the recession again.
https://www.denverpost.com/2023/01/25/apartments-metro-denve...
The good news is there is basically no place in North America where there is a finite construction infrastructure or land causing housing shortages, except for maybe very low population remote areas and maybe Manhattan. The vast majority of shortage is driven by bad government zoning and approvals policy (looking especially hard at you San Francisco). If you fix the policy in a way that developers will be reasonably sure it won't revert back and be used to punish them for what they did under new policy, your shortage problem would sort itself out in a decade or two as the massive building boom led to a glut and lower prices.
Let's refine this statement, there's plenty of land and housing, but not enough near the jobs or attractions that middle and upper class desire or that the lower class can afford, who provide support for the middle and upper class.