Monday, June 25, 2018

Housing: Part 306 - The boom was good. The bust was bad.

The Federal Reserve has a new article (HT: Noah Smith) posted that is a great example of the problem of trying to develop new understanding in a context where canonized wisdom is wrong.

The main point of the article is that, "The rise in rent-inclusive PIRs for below-median income households suggests that many of these households are about as vulnerable as near the onset of the financial crisis, as more of their income is committed to rent payments, possibly at the expense of saving for down payments for home purchases."

PIR's are payment to income ratios.  They are saying if we estimate housing expenditures only using mortgage payments, then it looks like housing costs have declined since the housing boom, but if we include rent payments in that measure, then households with lower incomes actually have more precarious housing expenses than they did during the boom.

This is the correct conclusion.  What's fascinating is that the data they use to come to this conclusion also contradicts the idea that the housing boom was associated with unsustainably high housing expenditures.  But, the idea that there was an unsustainable housing bubble is canonical.  So, the fact that this data contradicts that presumption is not important.  The canon does not require empirical confirmation.

I don't blame the authors of the article for this.  We have to have a canon.  We can't require that the canon be reconfirmed every time we try to learn something new.  But, it still fascinates me to see a single data set that is used to confirm some new idea within the canon of accepted facts, even while the fact that the data disputes the canon itself goes unnoticed.  As long as the canon is wrong, this will be inevitable.

We can see what is happening in the graphs from the article.

Here (figure 2 from the article), the portion of households with debt payments (which are mostly mortgage payments) above 40% of income is shown in blue.  The portion of households who spend more than 40% of income on the sum of debt and rent is shown in red.

The point that the authors try to make here is that debt payments are an incomplete measure of distressed housing costs.  But, this graph supports an interesting point about the housing boom.  There was a lot of concern expressed about that rising debt-to-income ratio from 2001 to 2007.  This rise in high levels of debt was completely from increased debt among households with high incomes.  And, notice what we see in the graph here.  If we add rent payments to mortgage payments, there was little change from 2001 to 2007.  In other words, households were not increasing their total housing expenses.  They were simply substituting mortgage payments for rent payments.  The households who were taking on mortgage payments greater than 40% of their incomes had been making rental payments that were more than 40% of their incomes.  These mortgages were rent hedges being taken out by households with high incomes who lived in housing deprived cities.

From 1995 to 2007, there was a housing boom, and that boom was associated with a reduction in rent payments.  Since 2007, mortgage repression has lowered the prices of homes, so that households that own homes spend less, but renters are spending more.

If we could expunge the error-plagued canon from the public consciousness, then the obvious lesson to be learned from this graph is that the housing boom greatly reduced the number of households with distressed levels of housing expenditures.

Here, I will compare Figure 1 and Figure 4 from the article.  Again, the blue line is for debt payments.  The red line is for debt payments plus rent payments.  (These are rough approximations of the measures shown.  See notes in the article for details.  Some other minor payments are included in some of these measures.)  Figure 1, on the left, shows the portion of all incomes going to these payments.  Figure 4, on the right, shows the portion of incomes going to these payments for households with below median incomes:

For all households, we can see that total payments (in red) were fairly stable throughout the boom, and have declined since then.  As described above, the boom reflected mostly a substitution of mortgage payments for rent payments.  But, take a look at payments for households with below median incomes.  Remove the terrible, distorting, wrong canon from your head so that you can actually see this graph in full.  There is one overwhelming point that this graph makes clear: The housing boom was associated with a significant decline in total housing payments for households with low incomes.  The housing boom was reducing economic distress of households with low incomes, and that process was halted in 2007 when the housing boom was interrupted.

The authors do not note this because even though the idea that low income households were especially vulnerable at the cusp of the financial crisis is wrong, it is canonical.  It is not a point that can be disputed with a single data point.  So, the most significant piece of information that this graph makes clear is not noticeable.  I would suggest that it is effectively invisible to the naked eye, even though it lies in plain sight.

I will point out a couple of other details from Figure 4.  As with the broader measure I described above, the rise in debt payments in 2004 is completely countered by a decline in rent payments.  There was no net increase in housing expenditures.  But, note that this measure peaked in 2004 (which is when homeownership rates peaked).  The so-called subprime bubble happened from 2004 to 2007.  Mortgage payments for households with low incomes peaked before the subprime and Alt-A markets exploded.  That is because the rise in private securitizations was mostly associated with the households with high incomes buying homes in housing deprived cities while households with lower incomes migrated to less expensive parts of the country.

The housing bubble was the acceleration of the national segregation by income that is set in motion by localized housing deprivation.  Households with high incomes were taking on the expense of claiming their piece of limited residential space in order to claim high incomes in the Closed Access cities.  Households with low incomes were fleeing from those cities and making compromises in order to reduce their expenses.  The public obsession with predatory lending has blinded us to the primary social development of our time.  And, for a decade we have been poisoning our economy as a result.

Here is Figure 6 from the article.  The general reaction I see to this graph from most audiences is that (1) the period since 2007 is just another example of the common American getting screwed - that things keep getting harder for those with less power and easier for those with more power - and also (2) the most important thing we must do is to avoid going back to the policies that were in place from 1995 to 2007.

What is happening here is that we "solved the problem" of high home prices by kneecapping the mortgage market - especially mortgages for households with lower incomes, FICO scores, etc.  This has reduced home prices across the board, and it has especially reduced home prices in entry level markets.  Since entry level prices are so low, homebuilders can't compete with existing homes on price, and new entry level supply has been curtailed.

So, for households with high incomes who can qualify for a mortgage, housing is cheap.  For households with high incomes who can't afford a mortgage, rents are high, but they can make adjustments in their real housing consumption to bring payments down (smaller unit, longer commute, etc.)  Households with lower incomes have less elastic demand for housing because they already tend to spend more of their incomes on housing that is smaller or less convenient than they would prefer.  Few of them can qualify for a mortgage, so they don't get to take part in the housing sale.  And, as renters, they must eat much of the rising costs.

The data is all there.  There is no need to argue about the details.  I take no issue with the facts in this article.  That is the good news.  The facts can be stipulated.  The bad news is that the canon needs to be erased and rewritten, and as the old saying goes, you can't easily reason people out of something they weren't reasoned into.  Yet, it seems like there is room for optimism when there are so many pieces of highly informative and accurate research out there about this topic, such as this article, which simply need to believe their own findings.


  1. IMHO more great blogging.

    Orthodox macro economics appears caught in a rut, often defined by conventional wisdom, ideology or agenda rather than dispassionate inquiry.

  2. "A 28-year-old Democratic Socialist just ousted a powerful, 10-term congressman in New York"

    Let's see: Decades of rising housing costs, heavier taxes and stagnant wages.

    America's chattering classes obsessed with "free trade."

    The employee class may be making the "wrong" decision, but I cannot blame the ordinary employee who votes for socialism.

    When, for example, will America's leaders really tackle rising housing costs? There is a solution---wipe out property zoning and go to free markets.

    Conveniently, many believe in free markets when it come to global trade--not in property development.

    Bernie Sanders may the D-Party nominee in 2020.

  3. I wish you would do a post on the government regulations that prevent mortgage lending to would-be home-buyers with relatively low FICO scores. Especially significant, in my view, are regulations that apply to *non-bank* mortgage originators. Obviously federal regulators exercise control over *banks*, primarily because of federal deposit insurance; but what sort of control do they exercise over non-banks? Without strong regulation of non-bank mortgage originators, many would spring up to serve the low-FICO market, alleviating the difficulty you point to.

    1. I wish I had more details for you.

      Here is the CFPB rule for non-banks:

      GSE standards also apply to banks and non-banks. I don't think they even use FICO scores explicitly in their models. Yet, the data is extreme regarding the shift in lending to various FICO score ranges.

    2. It would be good to know just what is the form of the CFPB's threat against a non-bank mortgage lender who lends to buyers with lower credit ratings. I am also curious about what powers the CFPB has to demand that non-banks supply it with information about the loans they are making.