We are now in year 6 of the housing downturn and it is difficult to find too many signs of optimism amongst the investing masses. Housing busts have historically lasted about six to eight years, but there is a perception that it’s “different” this time around and massive shadow inventories can help keep the housing market suppressed for at least another half decade. A recent estimate from CoreLogic puts the shadow inventory at around 1.7 million homes. Some even suggest that this number has potential to rise more over the next few years.
Given this, a lot of people view the housing market as a lost cause right now. Until the excess inventories are worked out of the system, it is argued, the housing market cannot recover. This is where I find myself largely in disagreement.
We don’t have a supply problem. We have a demand problem.
The demand problem comes from two major sources:
- High unemployment / low job security, and
- Overpriced housing
The latter is important because attempts to prop up the market by keeping prices artificially high tend to exacerbate the problem, by not allowing prices to push down towards a new equilibrium. I’d also throw in the Federal government’s 20% down payment rule proposal as another hurdle to a healthy recovery, and it’s not totally clear what will happen on that front, yet. It’s something to keep an eye on, but at some level, prices will get low enough to compensate even for that.
Before I get into why things will improve more than people expect, let’s take a look at just how bad things are right now.
The ‘00s Housing Boom
The ‘00s housing boom was not completely unprecedented in terms of volume. When you look at the historical data, we had much bigger boom years than we did in the ‘00s. However, the ‘00s boom was significantly lengthier than previous booms, lasting nearly nine years, compared to prior booms that tended to last roughly three to six years.
The following table shows US housing starts and population from 1959 to the present. Note that the last column is “Housing Starts per Person”, which was my attempt to create a normalized measure for housing.
click to enlarge images
Over that 52-year period, the average number of housing starts was 1.48 million annually. The chart below shows housing starts each year (red line) versus that average figure (green line).
While the boom was somewhat typical, if lengthier than average, the housing bust is completely atypical and unprecedented over the past five decades. Before 2008, the previous low level of housing starts in our sample period was 1.01 million during the early ‘90s recession. Right now, we’re on course to have our third consecutive year with less than 60% of that total.
I created the “starts per person” stat to try to gauge a normalized level of housing production given our increasing population. However, the metric has some flaws. If you look back at the table, it would appear that “starts per person” is decreasing over time. Perhaps as American cities become more developed and the quality of the existing housing stock becomes higher, we simply do not need as much new housing per person as we did before.
With the following chart, I try to piece together what a “normalized” housing production level might look like in the 2010s.
My estimate from the above info is that a normalized level of production would imply around 1.4 to 1.6 million housing starts, or about a 150% increase from current levels. But even if we veer towards the extreme conservative estimate and assume we should have .004 housing starts per person; that would still put us at 1.25 million starts, a 110% increase from the current levels.
This puts my estimation very close to one generated by Macroeconomic Advisers, which is estimating that based on demographic trends, household formation rates, and housing stock depletion; the US would need to add, on average, 1.6 million new homes each year for the next decade.
While all of this is based on somewhat abstract population, housing, and demographic projections, it is somewhat telling that all these figures suggest that housing production will need to increase dramatically over current levels of production. Let’s shift away from this discussion, however, and look at pricing and supply/demand dynamics.
Prices and Scarcity
It may seem preposterous to suggest that we currently have a condition of housing scarcity in many places, but there’s an increasing amount of evidence supporting this view. One of the first signs of this is rising rental rates. While it is anecdotal, I had noticed that rental rates were increasing at a dramatic pace in urban Atlanta over the past year. Rents in my area had increased about 10% to 15% year-over-year on comparable units of nearly every apartment complex I had monitored; and the prices continue to go up now.
This suggests that the rental market isn’t very soft at all. With rapid rental price increases, coupled with falling housing prices, suddenly buying is looking a lot more attractive in Atlanta. A recent New York Times article affirms my anecdotal analysis. The Times article looks at price to rent ratios. Not unlike the P/E ratio for stocks, a ratio of 15 seems to be near the long-term average for Price/Rent.
Under this line of reasoning, housing in Atlanta, Los Angeles, Miami, Minneapolis, St. Louis, Las Vegas, Cleveland, Detroit, Phoenix, Pittsburg, and Tampa look relatively inexpensive compared to the alternative of renting. Whereas, housing still looks comparatively unattractive in San Diego, New York, Washington DC, Silicon Valley, and the Pacific Northwest. While this would suggest that buying a home is only attractive in part of the country, this is nonetheless, dramatically different than the environment a few years ago, where buying a home still looked extraordinarily expensive compared to renting in nearly every major US metropolitan area.
The Wall Street Journal offers another interesting metric: the price to income ratio. WSJ examines the historical norms for the price/income ratio in several major US markets. Once again, we get a divergence, with many cities now below their historical averages; but several are still above historical norms.
In the aggregate, WSJ finds the current Price/Income ratio to be 14% above the historical average for the entire US. The US historical average for P/I is at 2.9 and the current ratio is 3.3. Still, 14% is not outlandishly high, and WSJ’s data only goes back to 1985, meaning it contains two major housing busts, but only one housing boom.
Using WSJ’s analysis, Detroit (-35%), Las Vegas (-25%), Phoenix (-13%), Cleveland (-12%), Dallas (-11%), Orlando (-9%), and Atlanta (-8%) are amongst the most beaten down major markets. Silicon Valley (+38%), Los Angeles (+38%), San Diego (+32%), New York (+32%), San Francisco (+26%), Washington DC (+26%), Denver (+23%), Portland (+22%), and Seattle (+14%) are amongst the more historically expensive markets.
In the aggregate, all this data would seem to suggest we should at least be prepared for another 10% to 15% drop on the national level, but some major markets might have bottomed or be very close to bottoming right now.
Of course, Price/Income has a few major flaws; particularly in markets with major development restrictions. In those sorts of markets, housing prices always stay artificially high due to scarcity. Therefore, you have to compare housing prices with rental prices to know the full story.
One thing that is clear is that rental vacancy rates appear to be declining rapidly in many markets. Using Census Bureau data, I pieced together the rental vacancy rates over the past four years for many major US metropolitan areas and examined the trends over the past two years.
You can see all the data below; the last column shows the decline in vacancies between the most recent quarter (Q2, 2011) and the quarter two years prior (Q2, 2009). Highlighted in yellow are all the cities where the vacancy rate declined by more than 25%. Notice there are an awful lot of them!
In the two columns prior, you can see the YOY decline over the past four quarters and the four quarters prior to that. Highlighted in blue are the cities where the declines were greater than 12%. As you can see, nearly every city in the sample has a decline in vacancy rate over the past two years, with many have major declines.
If you’re wondering what all the green is in the above chart, I used that color for any quarter where the rental vacancy rate was below 8% in a given city. The data here suggests that the rental market is booming, which seems to be supported by what we’ve witnessed with residential REITs over the past few years.
Another issue being overlooked is that there might be a certain dichotomy developing. The housing boom was more an exurban phenomenon than an urban one. Subdivisions filled with “McMansions” popping up 40 miles outside any major city is still one of the preserving images of excess. Yet, there may be a shortage in housing in urban and inner suburban markets, and we seem to be experiencing a shift in consumer preference, with more people favoring these submarkets. This could suggest that we may have another boom that is coming over the next decade; but that the nature of it could be significantly different from the last boom.
Employment and Housing
This all brings us to the issue of employment. Employment is one of the major factors holding back a fuller economic recovery. It’s also holding back housing. However, employment and construction are intricately related.
Construction workers, along with people in production and transportation related professions, tend to be less educated than their counterparts in management, office, and service occupations. Using the BLS data, we can see that those with lower levels of educational attainment have suffered the most during the housing bust. The unemployment rate for those without a high school diploma has gone from a low of 7.1% in 2006 to the current level of 15%.
Taking a closer look at occupational unemployment data, we find the same sort of trend.
Notice that construction unemployment hovered around 6% during the housing boom, but is now around 13.7%. Even this is a dramatic decrease from two years ago when the figure was 18.4%. However, we should add an asterisk to that thought, because part of that drop in construction related unemployment is the result of a lower labor force participation rate; suggesting the real unemployment for construction workers might actually be a bit higher than the current figures suggest.
Even many of the other occupations listed will experience lower unemployment levels with a boost in construction. After all, more construction activity will mean more accountants, more office workers, and more support staff. So the takeaway here is that unemployment remains stubbornly high precisely because the lack of a rebound in housing. If housing rebounds, we’ll probably finally start to see some progress on lowering the stubbornly high unemployment rate; which, in turns, means we’ll see more demand for housing. It’s a virtuous cycle of sorts and we could see things improve more rapidly at an unexpected future time.
What's Holding Things Back?
I've already mentioned two high-level issues holding housing back: unemployment and high prices. There are several other factors that I believe are also contributing to the poor environment right now.
- Consumer confidence
- Lending activity
- Underwater mortgages
- Regulatory environment
Consumer confidence is a simple one. Due to the falling prices over the past few years, we're seeing the opposite dynamic as we did during the boom, with people more fearful of buying. This seems like the most minor issue to me, because once things bottom and prices start moving upwards for a few years, this fear will likely disappear.
Lending activity has been a more important factor holding back housing. As the banks repaired their balance sheets, they were much less likely to make mortgage loans. However, there are some signs of stabilization in the banking sector. The rate of FDIC closures has certainly slowed in 2011. Moreover, banks hold significantly more capital now than they did in 2008 or 2009. Declining yields on US treasury bonds could also provide further incentive for the banks to start making more mortgage loans. Overall, this issue hasn't totally disappeared, but I believe we're starting to turn the corner on it.
Underwater mortgages remain a big issue in the housing market, as many homeowners are essentially stuck in their current house as a result of it. It's not uncommon for a homeowners to be employed and be able to service their mortgage; but if they are underwater, they have to take a huge hit if they sell the property. This has led to a situation where many people who would otherwise buy a new home have decided to stay put and not take the huge one-time hit. This situation, too, will improve with time.
That leaves the regulatory environment, which remains my biggest concern. During the housing boom, Congress and the Federal Reserve underreacted to what was going on. The Fed even had the audacity to lower interest rates in spite of major inflation in the housing market.
Now, we've moved to the opposite end of the spectrum, where government entities are overreacting. The 20% down payment is one example, but there are several more. Dodd-Frank has also put a lot of arbitrary requirement onto the banks, which has made them less profitable, and hence, less willing to lend out. Many bankers remain annoyed by Federal micromanagement of lending, as well. Main Street Bank of Texas even went so far as to surrender their banking charter, to get away from the FDIC's arbitrary enforcement actions against them. They decided to pursue a deal with a private equity firm and become a non-banking lender. From all the signs I see, the regulatory environment right now is the issue most likely to hold back the housing sector in the long-run.
In spite of these issues, I think the supply/demand dynamics will eventually trump all, and we will start to see a recovery. However, these issues could help determine how long it will take for housing to recover and how large the recovery will be.
Why I’m Buying the Homebuilders
At some point, the realities of urban population growth coupled with an increased consumer preference for rental housing are going to force rental prices upwards enough, so that a significant amount of new construction is needed to alleviate price pressures. There’s already evidence that we’re reaching that stage in some markets. Given the fact that very little construction has taken place over the past few years, relative to the prior decades, it stands to reason that we could see evidence of significant undersupply in the housing market.
While it’s difficult to predict when things will start to rebound and at what pace the rebound will take place, suffice it to say that the current level of housing production is not sustainable. My prediction is that we will see a gradual recovery of the housing market over the next few years and we’ll slowly move back towards 1.5 million starts per year; this seems to be much stronger than most people are expecting right now.
All of this makes the home builders look like a very attractive investment opportunity. Even if, after three years, we only see an improvement to 900,000 starts, many of the builders should see a large increase in profitability and cash flows that could drive up stock prices significantly.
Buying into the builders is not for the faint of heart. Given all the carnage we’ve seen already, it wouldn’t surprise me to see another 50% drop in prices sometime in the next 12 months. It could take several years for the builders to rebound, so this is a poor strategy for anyone who might need to pull their cash out of the market any time soon. But for those willing to ride it out over the long-term, I think this will prove to be one of the better investment classes over the next 5-8 years.
As far as builders go, I favor Pulte (PHM), Toll Brothers (TOL), and DR Horton (DHI). I may add some others or add to my current stakes in these three builders if I see attractive pricing and fundamentals.