I want to look at the graph from yesterday's post some more, because I think that graph highlights the turning points of the housing bust well.
We can roughly divide the period into 6 sections:
1) Normal market, with Closed Access supply problem.
2) GSEs and FHA in decline while private securitizations rise. Since the GSEs and FHA tend to serve first time buyers, this exchange of market share coincided with a decline in homeownership. But, private securitizations, probably helped by both a loosening of regulations regarding marginal lending and by the decline in the public conduits, grew to take their place. The less stringent limits on terms for the private securitizations allowed home prices in credit constrained markets of the Closed Access cities and Contagion cities to rise, though these loans were generally to households with high incomes. The change during this time was in the terms, not in buyer incomes. So, even though homeownership was beginning to decline, price increases during this time were at their strongest. First time buyers were as strong as ever here, so we can presume that tactical selling out of ownership was strong. During this time, there was strong out-migration from the Closed Access cities - weighted toward low income households, but not necessarily weighted toward renters. During this period, there was probably some profit taking among existing homeowners in the Closed Access cities, who, because the private securitization boom was pushing their home prices up, were incentivized to take capital gains, even where they had been enjoying artificially low property taxes.
Researchers like Mian & Sufi write about a transfer from high wealth to low wealth households, but in a lot of ways, wealth is a proxy for age. If we think in terms of income, migration data from this period suggests that there might have been some profit-taking among low income households. I suspect there was a bimodal distribution of migrants - poor renters forced out and low income owners taking profits - both induced by an expansion of housing consumption among young, educated, high income workers in the Closed Access cities, which was made possible by the lenient terms of the private securitization market.
3) In the spring of 2006, the Federal Reserve intentionally pushed the Federal Funds Rate high enough to slow real housing growth. And they succeeded royally. Mortgages held on banks' balance sheets had been strong, helping to counteract the relative decline in GSE and FHA lending, but banks predictably reduce lending in the face of an inverted yield curve. We can see the effects of this policy choice. So, at the margin of the market - new home sales - we see a sharp drop. New home buyers were still relatively strong, however. The continued entry of first time buyers begins to cause leverage to rise during this time, as the wave of tactical selling builds from existing homeowners exiting the market.
4) By 2007, the number of first time buyers begins to drop, but the exit from ownership continues. This causes new home sales to fall even farther, and home ownership rates really begin to fall. But, intrinsic value rules. Rent inflation was rising sharply, and home prices remained near their peaks, because intrinsic value dominates supply and demand in asset markets that are not in disequilibrium.
5) In late 2007, disequilibrium hits. Both buying and selling dry up. Notice that the homeownership rate levelled off somewhat in 2008 despite the drop in first time buyers. Tactical selling was strongest when prices were high. That stopped after prices began to drop.
6) In late 2008, the GSEs were taken into conservatorship and the banks were in disarray. The GSEs severely tightened credit standards, to widespread approval. The Fed finally fully committed to stability. So, prices and sales finally bottomed out and stabilized. But, because the stabilization was targeted at the high end and left the low end out, the apparent stabilization of the housing market was a bit of a mirage. This period was the worst period for the bottom half of the housing market. And the continuing drop in homeownership was concentrated at middle incomes and the bottom.
Keep in mind that homeownership to begin with was already 80% to 90% for the top 40% of households, by income, so the rise in ownership among those groups, as a proportion of existing non-owners, was much stronger than in lower income groups. And, today, compared to 1995, ownership rates are lower for the low income groups and higher for the high income groups. We gave low income households a bust when they never had a bubble.
These last charts are annual changes in median home prices, by zip code (from Zillow Data). Zip codes are arranged by price on the x-axis, on a natural log scale.
Across many cities, the same pattern developed in 2008 and 2009. The aggregate market appeared to have stabilized, but we removed liquidity from the bottom half of the market. The number of defaults that happened during this period is many, many times the number of defaults that had happened before the subprime market collapsed in 2007.