Housing starts grew to an annualized rate of 872K in September, the highest level in nearly four years. Meanwhile, building permits rose to 894K, a sign that we should continue to see improvement in the future. This improvement is good news for the homebuilders (XHB), but this doesn't necessarily mean its "good news" for the American economy.
It's important to remember that housing starts merely gives us a small piece of the picture: the supply of new homes on the market. Starts do not tell us much about the level of demand, or the quality of demand. Nor does it say much about the supply of existing homes.
There are several other important questions to ask:
(1) What is overall housing demand, and is supply increasing rapidly enough to meet it?
(2) How much of the increase in demand is a result of more "owner-occupants" entering the market, and how much is result of real estate investors?
(3) Could supply of existing homes be artificially suppressed due to regulatory and foreclosure issues?
(4) Could supply of new homes be lesser than needed due to the destruction of smaller homebuilders and the Dodd-Frank financial reform bill?
(5) Are demand increases merely the result of an increasing rate of money supply growth?
There are a lot of issues to look at, that often get ignored in the mainstream press until it's too late. It's worthwhile to explore a few of these issues, but I particularly want to focus on the money supply growth issue.
Housing Starts and Recovery
For those who have followed me over the past year, you know I've been a housing bull and I remain one. Housing will likely improve regardless of what Congress, the President, and the Federal Reserve do in my view and that's why I've found it to be such an attractive investment area.
The only real question is whether we'll have a healthy recovery or whether the housing recovery will come with some new problems, such as rising inflation. Indeed, my biggest concern is that we could be heading back to a 1970's-like environment of lower real growth and higher inflation that results in a booming housing market.
The chart below shows housing starts since 1970. Based on historical data and population projections, I estimate that we need around 1.5 million starts to be at normalized levels again. You can see even with dramatic improvement, we are still well below that figure, and far below historical levels.
This second chart shows the more recent view of starts, running from 2007 to 2012. You can see more clearly in this chart, how dramatic the rebound has been. Around May of 2011, we were at 550,000 starts. 16 months later, we're at 872,000, or a 59% improvement.
Of course, even if we are at historically low levels of construction, this rapid rise in housing starts still begs the question, why is everything suddenly taking off right now? Part of the answer to that may lie in a housing shortage.
Rental vacancies are at their lowest point in over a decade. The homeowner vacancy rate is at its lowest point since March 2006 at 2.1%. Inventories seem to be low in most places, which is odd given all the talk we heard over the past few years about "shadow inventories."
So we actually have data that seems to suggest that the recent surge is the result of a shortage, which begs the question of whether we're increasing demand rapidly enough. There's another issue here: could supply be artificially constrained by bulky foreclosure regulations? We see some evidence that this may be the case in places like Las Vegas. These somewhat artificial supply constraints might help create more near-term demand for new homes, but it's also a less than ideal situation, and falls into the category of economic dead-weight.
I also want to explore the idea that maybe the housing market is rapidly coming back as a result of money supply growth, which is making real asset investment preferable to other forms of investment.
Inflationary Housing Boom?
Inflation and housing booms can go hand in hand, as we saw in the 1970's. While we're a long way off from a "housing boom" at the moment, I'm not particularly comforted by recent increases in M2 money supply, nor some of the recent Case-Shiller data.
M2 money supply growth has historically been a reasonably good predictor of future inflation. We can see in the chart below that M2 growth has been at elevated levels for most of the past year, hovering in the 8% to 10% range, with a recent dip down to around 6% to 7%.
Case-Shiller housing price numbers are likewise very interesting and I'm not totally sure what to make of them. Year-over-Year [YOY] gains in most major metropolitan areas have been subdued or even negative (with some notable exceptions such as Phoenix). Quarter-over-quarter [QOQ] gains, on the other hand, have been very large. This is true whether we look at seasonally adjusted figures or the unadjusted ones.
Atlanta is probably the most dramatic example of this. In the Atlanta MSA, we've had YOY price declines of -10.1% on a seasonally adjusted basis. Yet, we've had quarter-over-quarter price gains of around 5.7%. If we were to annualize the QOQ gain, we'd end up with 24.7% growth. Obviously, I don't really believe we're seeing that huge of an annual swing in housing prices, but it does show how dramatically prices have been going up in the past quarter.
Here's the Case-Shiller price data for the 10-city and 20-city indices on both a seasonally adjusted and unadjusted basis.
It's also interesting to look at this data for the individual cities in the data. Here is seasonally adjusted data for all 20 metropolitan areas in the Case-Shiller index.
From this data, we can see wild variations in each local market, which might suggest that whatever is happening in the housing market is based on more localized concerns, rather than broad macro trends. But I still can't help but to wonder about the rapid nature of the price increases.
On one hand, we've seen this sort of thing before. From about 2009 to early 2010, we saw housing prices increase, before drifting back down again until late this spring. Yet, it wasn't nearly as dramatic that time. The two are compared below.
This data becomes even more odd when put into a broader context. I decided to look at the historical quarter-over-quarter gains in the Case-Shiller 10-City index, all the way back to 1987. Here are the results.
We really only have one other housing bust in this data and that was the one that occurred in the late 80s / early 90s. You can see a brief surge around 1991, but it topped out at 0.8%. That compares to the current QOQ surge of 2.3%. After that small uptick in 1991, we saw a slow and gradual recovery, before prices began to take off around 1997 (and didn't reverse again till late 2005).
What's really odd about the two recent bumps in our data is that they look more like the surges in the housing boom from 1997 - 2005, rather than the bottom-of-the-market bounce in 1991 that was followed by gradual price recovery.
Money Growth, GDP, and Housing Prices
Want to make things even more interesting? After looking at this data, I started to wonder about money growth in relation to GDP, and how inflation and housing prices were related to these factors. Sifting through monetary data is tricky, because it's extraordinarily volatile, so that it often breeds completely unreadable charts. But I came up with an interesting metric that yielded some intriguing results.
In the chart below, I decided to look at the difference between M2 money supply growth and nominal GDP growth. Because the data was too volatile to make any sense using smaller time frames, I decided to do a three year rolling average of year-over-year gains. The results are interesting to say the least.
This chart might seem confusing at first, but I'll explain it. The main thing to focus on is the 0% line. When the trend line drifts below that, it suggests that nominal GDP is growing more rapidly than money supply. If the trend line moves above 0%, then it's the reverse and money supply is growing more rapidly than nominal GDP.
Around 1994 and 1995, right before the tech boom really took off, we see a massive 4.6% difference between GDP and money supply growth, with money supply growth being very small comparatively speaking. This trend begins a dramatic reversal around 1996 and 1997, before finally crossing over the 0% line again (more money growth than economic growth) in 1999, and eventually peaking at around 3.8% in 2003.
We once again saw a reversal and started drifting below the line again around November 2005 till June 2008, before a new very dramatic reversal takes place. We have surged back over the line since then, topping out at a record-high 5% (at least in this data set) in early 2010, and we're now around 3.7%.
Of course, this isn't a perfect measure of anything. There are a lot of flaws. For one, even if this seems to foreshadow the tech boom of the 90s and the housing boom that eventually followed, it doesn't seem to explain the inflation of the 1970s very well. This might be because I'm looking at the difference between M2 and GDP growth on an absolute basis, rather than looking at M2 growth as a % of GDP.
If we look at money supply growth as a % of GDP, this seems to predict the 1970's inflation better.
This metric also does a good job showing the money supply stabilization in the mid 90's that preceded the tech boom, but it's unclear whether it's a good predictor of the 00s housing bubble. After all, it peaked out at only about half of the level we saw in the 1970's.
Yet, the overall trend does show a more dramatic shift in that timeframe than it did in the 70s and maybe that is the important thing. After all, investing with 1% money supply growth might make certain investments attractive, while less stable 4% growth might make a completely different set of investments attractive. This may very well explain the shift from tech (XLK), which is more attractive in a low inflation environment, to real assets like housing.
I've posted a lot of data, but what can we conclude from all of it? I'm not totally sure to be honest, but it certainly seems to challenge some conventional wisdom right now.
The mainstream investment community might be somewhat divided on whether we've finally seen a housing bottom, but there's broad agreement amongst both housing bulls and bears that when the market does find its footing again, that housing prices will return to a slow growth rate (i.e. 1% to 3%). There have been very few who have been anticipating a rapid housing price recovery.
The data I've presented here, however, might suggest that the mainstream consensus is wrong. Housing prices do not seem to be following the gradual recovery pattern that emerged in the early 90s. Rather, the price increases we're seeing actually seem more typical of an unstable housing boom market like we saw from 1997 - 2005.
Whether this trend is a short-term aberration is difficult to tell, but money supply data that is painting inflationary signs could suggest that we are seeing the infant signs of stagflation. There may be no disco this time around, but it might begin to look a lot like the 1970's soon.
How to Invest
Regardless of what happens over the next few years, I think housing will continue to recover, but investing on that theme without directly buying houses is becoming trickier. The homebuilders have already taken off. Even though I believe that a builder like Pulte (PHM) still has a significant amount of upside, the risk / reward tradeoff is certainly much less attractive now than it was 15 months ago.
Companies that have a high degree of leverage in their capital structure and will benefit from improvement in the housing market certainly could be winners. This would include many banks (KBE) and insurers (KIE). Fixed-income securities, such as US treasury bonds (SHY) and high-grade corporate bonds (LQD) would seem most vulnerable; alongside of mortgage REITs (MORT)
But I'd add the caveat that a lot of the above depends upon political decisions. If the US Federal government dramatically cuts spending and raises taxes, that could change things significantly. My present view is that we'll continue the current path regardless of how the upcoming elections go; it's merely a question of degree. But I don't pretend to have a crystal ball that allows me to see what the pols are planning to do and if we see a significant change in course, that could also cause a shift in what investments are attractive.
Disclosure: I am long PHM.
Additional disclosure: I am long PHM, as well as a few banks and insurers.