I Don’t Get It (Housing Starts Edition)

by Dave Schuler on November 20, 2012

In reaction to the sharp increase in housing starts in 2012 Bill McBride says:

And the growth in housing starts should continue over the next few years. Even with the significant increase this year, starts in 2012 will still be the 4th lowest year since the Census Bureau started tracking starts in 1959 (the three lowest years were 2009 through 2011).

My estimate is the US will probably add around 12 million households this decade, and if there was no excess supply, total housing starts would be 1.2 million per year, plus demolitions, plus 2nd home purchases. So housing starts could come close to doubling the 2012 level over the next several years - and that is one of the key reasons I think the US economy will continue to grow.

I honestly don’t get that. I can see that housing starts have increased relative to the very low number of starts in 2011. Either Mr. McBride is saying that there are some other factors at work that he doesn’t talk about in his post or he’s claiming that the economy will continue to expand. According to the NBER, the official recordkeeper, we’ve had about 41 months of expansion. Based on past experience we should expect that expansion to slow and even reverse within the next couple of years.

Consider the graph above and the associated table. What I see is that what we’ve been experiencing over the last dozen years or so is a marked departure from the normal. During most of the period housing starts increased during expansions and declined during recessions. From about 1993 onwards (or, arguably, from about 2000 onwards), that pattern has been thwarted. To me it looks as though we’ve been overbuilding houses for quite a long time.

One data point doth not a trend make. 2013 may see more increases in housing starts or it may see a decline, as there was in 2011 relative to 2010. It’s just too early to tell but, given where we probably are in the business cycle, banking on a new housing boom sounds premature to me.

{ 21 comments… read them below or add one }

PD Shaw November 20, 2012 at 11:29 am

It looks to me like housing starts do not correlate that strongly with the general business cycle.

Housing starts rise consistently from around 1991 to 2008, without hardly a hiccup for the early 2000 recession.

The decline from about 1984 to 1991 appears to run counter-cyclical, does it not? And the first half of that graph looks quite erratic overall, and becomes much smoother and longer after the mid 80s.

My thinking would be that housing starts are increasingly long term enterprises with significant planning and investment before shovel hits the ground. A short contraction like the early 2000s won’t effect housing starts. The reason for the countercyclical contraction might be a lagging impact of the 1980 recession. McBride might be wrong to say usually housing starts rebound sharply.

PD Shaw November 20, 2012 at 11:31 am

I’d be curious about the home values associated with the recent uptick in housing starts. More modest construction for new realities, or bigger homes for those unaffected by the recession?

Steve Verdon November 20, 2012 at 12:24 pm

It looks to me like housing starts do not correlate that strongly with the general business cycle.

Check again.

Recession: 1960-1961; Housing Starts 1960-1961: -18% & 4%
Recession: 1969-1970; Housing Starts 1969-1970: -1% & -3%
Recession: 1973-1975; Housing Starts 1937-1975: -13%, -35% & -13%
Recession(s): 1980-1981; Housing Starts 1980-1910: -24% & -16%

Edward Leamer’s research into business cycles suggests that 9 out of 10 recessions are significantly linked to the housing boom/bust cycle.

Link

C&P:

Housing IS the Business Cycle
Edward E. Leamer

NBER Working Paper No. 13428
Issued in September 2007
NBER Program(s): EFG ME

Of the components of GDP, residential investment offers by far the best early warning sign of an oncoming recession. Since World War II we have had eight recessions preceded by substantial problems in housing and consumer durables. Housing did not give an early warning of the Department of Defense Downturn after the Korean Armistice in 1953 or the Internet Comeuppance in 2001, nor should it have. By virtue of its prominence in our recessions, it makes sense for housing to play a prominent role in the conduct of monetary policy. A modified Taylor Rule would depend on a long-term measure of inflation having little to do with the phase in the cycle, and, in place of Taylor’s output gap, housing starts and the change in housing starts, which together form the best forward-looking indicator of the cycle of which I am aware. This would create pre-emptive anti-inflation policy in the middle of the expansions when housing is not so sensitive to interest rates, making it less likely that anti-inflation policies would be needed near the ends of expansions when housing is very interest rate sensitive, thus making our recessions less frequent and/or less severe.

PD Shaw November 20, 2012 at 1:53 pm

Thanks Steve, I was eyeballing it. The linked piece seeks to explain the two major “eyeball” discrepancies: (1) 2001 was an Internet caused recessionary event that played out without the traditional housing drama. (2) What appears to be a countercyclical decline in the 80s is the result of tight monetary policy.

OTOH, Table 2 in the link suggests that residential (mal) investment as a source of GDP weakness heading into a recession has diminished since 1980, relative to other factors such as durables, services, equipment & software, and to some extent exports (which were a source of weakness before 1960 and after 1980, but not in between(?))

PD Shaw November 20, 2012 at 2:55 pm

Edward Leamer’s answer to Dave here might be that we don’t have a business cycle, so much as a consumer cycle, of which housing starts are the best measure. That would mean that we bottomed out as recently as this year and we might expect another three years of expansion, or as much as five.

Steve Verdon November 20, 2012 at 4:08 pm

Well, the housing part of the decline can lead to a decline in other consumer durables later on, according to Leamer. According to the linked data, 2009 April is when we hit the bottom 478,000 housing starts. So the protracted sluggishness could be in part due to the secondary phase in other consumer durables. Given the depth of the housing market decline the second phase could also be fairly long.

Of course, if housing starts are going up due to government policies vs. normal market processes then this could all be false. After all, the government could be simply distorting incentives and “moving” housing starts forward in time implying that down the road we’ll have another recession as housing starts fall. Unless of course the government has a hidden room that has an endless supply of money (which we don’t).

PD Shaw November 20, 2012 at 5:06 pm

FWIW, I got the early 2012 as a “bottom” from looking at the linked Calculated Risk graph for “Housing Starts, One Unit Structres, Rolling Twelve Month.” That graph looks to me like sharp decline until late 2009, slight bounce, and then declining further until January 2012, when it starts rising. The similar graph for Five Plus Unit Structures shows a more robust bounce.

What I find most counterintuitive about the Leamer analsysis is that its all about residential units, not residential prices. My gut tells me that the current downturn was problematic in large part because of the price problem.

Steve Verdon November 20, 2012 at 5:20 pm

Housing prices are sticky downward, but volume on the other hand is highly volatile. That is why the focus on units. Yes, it is somewhat counter intuitive since you’d think new home builders might be more flexible on housing prices than people in currently existing homes.

steve November 20, 2012 at 7:23 pm

I think a lot of the optimism comes from data showing increases in the number of households.

Steve

Icepick November 20, 2012 at 10:30 pm

Yeah, all those college kids with the shiny new careers, plus all the people that have decided to move on up from the hood. (THAT’S why we’ve got all these empty houses!)

steve November 21, 2012 at 7:59 am

Unemployment/underemployment for college grads is about 50%. The record low for that group, set at the end of the dot com bubble was 41%.

Steve

Dave Schuler November 21, 2012 at 8:34 am

Note that when they say that education is the key to a prosperous future that they aren’t saying for whom.

Icepick November 21, 2012 at 9:00 am

Yeah, steve, it’s all happy days. That’s why a couple of thousand lay-offs have been announced in the Orlando area in the last week or so, including a few hundred at the largest local hospital operator. It’s all coming up roses because Black Jesus and his Disciples say so.

steve November 21, 2012 at 5:35 pm

We are talking about housing needs. If we are really seeing just a 9% difference, then over the course of three of four years demand should build up. The math is pretty simple.

Steve

Dave Schuler November 21, 2012 at 6:07 pm

What happens when everybody who can qualify for a home loan already has one and speculators aren’t willing to finance as much house building as they used to? Let me say it another way. Head count and marketing universe aren’t the same thing.

Icepick November 21, 2012 at 7:27 pm

steve, we’ve got lots of empty houses in Orlando. But they’ve started building again. Signs stating things like “Starting from the $400′s!” are springing up in well developed parts of town. Go to some of the areas on the fringe and you can see stuff a lot more expensive than that. Are those houses being filled by new college grads? Or are they being bought up by foreign speculators trying to get what money they can out of the various parts of the world economy that are collapsing? Given that just in the last couple of weeks we’ve had many hundreds of layoffs announced in this area by large employers, I don’t really see an employment boom on the horizon.

Hmm.

Also, as I pointed out elsewhere tonight, we’re still millions of jobs short* of where we were in December of 2007. That doesn’t exactly scream for the need of more housing. Or rather, doesn’t scream for the demand for more housing.

* Depends on which sets of data one looks at, but the numbers run from a little under 3,000,000 short to about 4,000,000 short. That does not account for population growth. All Employees: Total nonfarm (PAYEMS) indicates that we’re 4,227,000 jobs short.

NOTE: Employment for people in the 25-54 age range in down over 6,000,000 people from the start of the recession. THAT does not say HOUSING BOOM to me.

steve November 21, 2012 at 9:10 pm

I dont expect a boom, just a move away from the bottom. It has started in that direction. Barring trouble in Europe, I expect it to up a bit more. I dont expect new grads to buy 400k houses, but they will want to rent or buy small places. There are certainly anecdotal reports of things picking up. The kids we know who graduated from my son’s school are doing ok, so it is not all doom and gloom. We still have lots o f debt to work through, so Iexpect we will jsut muddle along a bit longer.

Steve

Icepick November 21, 2012 at 10:42 pm

steve, millions of us have had our working careers destroyed. I’m out well over a quarter million dollars so far, I’m going to be out a few million over the course of what would have been my working career. I’m not alone in that. That will never, ever be recovered. Our skills are lost even if someone did want to hire us. Our working lives are wasted.

The people coming through college now are coming out loaded down with greater amounts of debt. That debt is going to eat into their ability to start households and families. That is a long-term drag on the economy going forward.

Muddling through is the same as failing at this point. This is a complete and total disaster, and we’re not going to get a WWII to jump start us out of it either. Debt levels are high and they’re not coming down save for bankruptcy and foreclosure. And need I go through the food stamp and SSDI numbers again? If things are getting better why are more and more people on food stamps? If things are getting better and more and more people are on food stamps, why are the friggin’ food banks experiencing record utilization? Every story I’ve heard on the food banks so far this holiday season has them complaining that they’re getting more demand than ever. I’m sure some food bank somewhere has seen lesser demand, but I haven’t heard of it. How can that be three and a half years into this supposed economic recovery? Recovery for who?

And the American public want four more years of this. Outstanding. Just simply outstanding.

Icepick November 21, 2012 at 11:01 pm

We still have lots o f debt to work through, so Iexpect we will jsut muddle along a bit longer.

And what’s this “WE” shit you’re talking? You suffering from the economy? What, couldn’t buy the missus the new car you wanted? Hunh. Did you lose your job? Have your professional prospects been ruined? No? Of course not.

So spare me the “WE” talk. You haven’t earned the right, doctor. There ain’t no “WE”.

jan November 21, 2012 at 11:06 pm

Icepick

we’ve got lots of empty houses in Orlando. But they’ve started building again.

Sounds a lot like what China has attempted to do — stimulate the economy by building homes/cities they don’t need, just to keep people employed. Consequently the appearance of ‘ghost’ communities — built up and ready to go, except for people…

jan November 21, 2012 at 11:10 pm

There are certainly anecdotal reports of things picking up.

I would wait for 2013 to unfold before talking about ‘things picking up.’ From what I’m seeing and reading, employers are pulling in, deciding not to expand because of the election results, and letting more people go.

Leave a Comment

Previous post:

Next post: