Diary of a Financier

A chat about housing with Bill McBride (Calculated Risk): Is the future so bright?

In Dissertation, Economics on Mon 30 Sep 2013 at 07:50

Throughout September, I’ve had a fantastic exchange with Bill McBride over at Calculated Risk.  Housing’s continued recovery was a central tenet to my bull thesis, and it seems like the notion has been adopted by much of the Street, as well as the Federal Reserve.  While I expect a pause in the housing rally here in the intermediate term (2014), its potential energy could (I’ve downgraded from “should”) help propel the US economy onward and upward over the longer term.

Given that gravity, I have to continually play devil’s advocate and tease the null hypothesis.  To wit, I turned to Mr. McBride, an economist who’s one of the foremost authorities on macroeconomic housing trends.

My unedited conversation with Mr. McBride follows herein. In this dialogue, I raise a few concerns that antagonize the housing recovery meme; in particular, ‘Compounding stagnant postcrisis real incomes and baby boom demographics, why should we expect household formation to resume its prior trend?’

My comments follow herein. Enjoy…



Hi Bill,
Quick question about housing: Do you think the modest real income gains in the US justifies enough new housing starts to make any difference in the economy (GDP)?

Rising rates are going to hamper housing, as we’ve already seen it mortgage applications and builders. I’m trying to figure out if incomes and demographics justify enough gains to offset the headwinds.

For instance, given the baby boom generation, we may never return to old housing trends/levels? In peak spending, baby boomers owned 2 or 3 cars and two houses; won’t they my consolidate down to 1? Considering stagnant real incomes postcrisis, I don’t think the old rules apply…


As I’ve noted several times on my blog, I think housing starts will continue to increase over the next few years. Of course I don’t have a crystal ball, but I think demographics are favorable for housing at this point–not negative.


Thank you for the response. I’m an avid reader, so yes, I’ve caught your references of housing starts & demographics. Just a little more granularly, that’s exactly what I questioned:

You have rising rates for housing to contend with now. You have a baby boom generation falling from peak spending. You have a working class that’s hindered by stagnant post-crisis incomes.

I think the thrust of your argument for housing starts is the same as most analysts’ durable goods argument: pent up demand deferred from 2008-12. I just wanted to drill-down and confirm whether or not I’m reading you correctly…?

Oh, one more question too: what’s the status of the shadow housing inventory overhang? What does nobody talk about that anymore?


Hi Romeo,
There was never a clear definition of “shadow inventory”. I’ve been writing about this for a number of years, and I started arguing several years ago that this was overblown. I think people were double counting (like adding properties in foreclosure to inventory even though many were already listed).

Also there were rumors that many lenders were sitting on a huge number of foreclosed homes (I started tracking it closely, and the widely reported numbers were way too high).

Plus many adjusted rate loans adjusted down as interest rates fell. And many people refinanced into more affordable loans.

The bottom line is “shadow inventory” isn’t a huge problem.

I track the number of seriously delinquent loans (and in-foreclosure) from LPS, MBA, and from the GSEs. All are down significantly (many of these are already listed, so they aren’t “shadow”).


I found Mr. McBride’s comments on the shadow housing inventory very helpful.  He satisfactorally allays concerns about the issue, which now seems to have been a giant, misreported, non-event–partially sidestepped by ZIRP.

In regards to the rest, I can only infer from Calculated Risk and the above discussion that Mr. McBride’s optimism relies on a couple, key inputs.  First, pent-up demand from the slack accumulated between 2007-09.  Second, demographics and household formation trends.

This is not intended to be a takedown of his opinion, but I do disagree with him.  My problem with the first variable is that 2006 is a high base for housing data comps; 2009 is a low base.  We cannot assume that either outlier is an anchor when estimating the magnitude of pent-up demand that’s been displaced and now needs to be replaced.

More appropriately, we should look at trends.  Accordingly, I charted US monthly housing starts (1959-2013) Housing starts (1959-2013)- including linear & exponential regression trendsand incorporated both a linear and exponential regression.  Both trends are negative, which makes sense to me: There’s preexisting housing inventory that should force building volumes to exponentially decay over the long term. If you’re looking for an upside target for housing starts, the linear regression trendline at 1.25M SAAR is an objective base case target (vs. 891k in 9/2013). That sounds pretty bullish, leaving a lot of upside to be realized as the slack gets reeled-in.

That decay should always be impeded by a constant demand for new homes, the depreciation of old, and population growth. Hence, I was careful to recommend “exponential decay” (i.e. never actually reaching zero).  Housing starts/population growth (1959-2013)- including linear regression trendIn another chart, I did my best to control for these factors by deflating housing starts by population growth.¹  Again, the trend there is negative, and more importantly, it looks like housing starts already returned to trend growth earlier this year (~950k-1.0M SAAR).

Now, there’s no science to this. This is just math, and human interactions are not strictly governed by data. In other words, housing starts could blast through their historic trend growth and even return to bubble-era highs again. However, this analysis is valuable–particularly from an investor’s standpoint–for a few reasons. First, it dispels the myth that the housing recovery has only just begun. Second, it exposes housing-related theses as having thin margins of safety; since mean reversion has already occurred, the probability of one outcome or another is insignificant.

I’ve decided that I have no interest in investing in or around this theme here today, which happens to be something Wall Street’s sell-side is selling hard right now, while the buy-side is unloading.


There’s a lot more to housing’s contribution to GDP than just housing starts:

“Housing contributes to GDP in two basic ways: through private residential investment and consumption spending on housing services. Historically, residential investment has averaged roughly 5% of GDP while housing services have averaged between 12% and 13%, for a combined 17% to 18% of GDP. These shares tend to vary over the business cycle.

“Residential investment includes construction of new single-family and multifamily structures, residential remodeling, production of manufactured homes and brokers’ fees. Consumption spending on housing services includes gross rents (which include utilities) paid by renters, and owners’ imputed rent² (an estimate of how much it would cost to rent owner-occupied units), and utility payments.”

So, what else might push housing enough to move the needle in GDP growth?

In yet another contradiction to my skepticism, The Washington Post’s “Wonkblog” recently made some positive comments regarding a number of housing dynamics.  This supports Mr. McBride’s sentiments:

“First, household formation is heading up nicely–people who bunked with parents, put their marriage plans Housing starts & Household formation (1980-2020est)on hold or stayed in that ratty apartment with three college friends are apparently spreading their wings. That should boost housing demand, not to mention consumer traffic at Bed, Bath & Beyond ($BBBY).

“Second, households Household liabilities/Disposable income (1980-2025est)have largely tackled their debt problems–or are at least on the path to doing so. A debt profile poised for disaster–at an average of more than 130% of household disposable income [in 2008]–is quickly heading below 100%, good news for consumer spending… [and] fuel for healthy economic expansion in 2014.”

US population by age (1900-2060est)

Click for animation

That narrative about all these kids in their parents’ basements coming out to buy a home sounds great… the aggregate demographic data tell me a different story, however.  Even if that narrative were true in the face of rising mortgage rates, it should be massively offset by baby boomers’ descent from peak spending. The net-net GDP growth effect should be almost nil.

If you look at US population distribution by age over the years, the baby boomers pass through the distribution like an elephant that was swallowed by a snake.  The boomers’ echo is comparatively a mere antelope–formidable and perhaps feisty, but not as massive.   Further, with boomers already descending from peak spending, the echo won’t summit peak spending until 2020-30.  As aforementioned, retirees’ divestiture of their extra “2 or three cars and two houses” is a formidable headwind when we’re dependent on a generation with >$1T in student debt to counterbalance the effect.  There will be some latency before the echo net-net displaces boomers’ carbon footprints.

Plus, the first chart presented by WaPo (above) projects a narrow window (2013-15) for housing formation to surge–back to prior highs nonetheless–before gently tailing-off into the end of this decade.

I agree with WaPo that American households have deleveraged, but I do not find any evidence that we may rely on releveraging/household consumption to tow GDP growth hereafter.  In losing that tailwind, I maintain that GDP growth is dependent on the external sector (i.e. trade balance improving maybe even toward a net export surplus) and corporate spending.

In sum, I remain skeptical of the household formation thesis at this time, and therefore, by extension, I’m skeptical of housing’s potential contribution to GDP until we realize some meaningful combination of (continued) low rates, strong wage growth, and Millennials’ rise. It’s worth revisiting this theme in a few years, at which time capital markets should start discounting these more secular forces — all things else being equal.


¹Formula = Monthly Housing Starts / Monthly Population Growth. It’s obviously an imperfect analysis, since it ignores so many complexities of household formation. For example, population growth is arguably a leading (not coincident) indicator for housing starts, which is why you see the spike in population-adjusted starts in the 70s, when baby boomers joined the workforce. Over time, the trend is still significant, because you can see how even today we build so many more homes (>4x) than in which individual people are available to reside. Unless I’m thinking about this all wrong–a distinct possibility–that’s essentially pulling-forward economic production from the future.

²i.e. Owners Equivalent Rent (OER)

  1. […] Shadow housing inventory reaches 5-year low | World Property Channel Shadow inventory -22% y/y in July to 1.9mm homes, -38% since 2010 high. Completed foreclosures -34% y/y in August to 48k. Foreclosure backlog -33% y/y to 939k homes. [Previously] […]

  2. […] swapped-out of our Ford ($F) position with big gains in September, pluse we’re avoiding US housing and consumers.  A 2014 durables spending boom has become the preeminent null hypothesis to my […]

  3. […] fiscal drag waning 5. Population demographics improving with baby boom’s echo coming-of-age [Counterpoint] […]

  4. […] keeping rates below normal into the early-2020s. [Fed policy is starting to react to my demographic assessments. This has the potential to change my bearish 2014 outlook & form the blasting cap at the end […]

  5. […] policy is starting to react to my demographic assessments [e.g. demographics start aiding organic growth from aggregate demand between 2017-20, when the baby […]

  6. Update: Housing starts end 2013 with weak showing | Calculated Risk

    October & November retroactively revised downward; December data surprises with miss (414k vs 450 expected).

    FY13 numbers showed good sequential growth (+16.4% y/y @ 428k SAAR), but it was still 6th weakest year on record.

    [Housing starts are volatile m/m, but my thesis is coming to fruition so far, with rising mortgage rates already stunting housing activity in 4q13…]


Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Google+ photo

You are commenting using your Google+ account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )


Connecting to %s