EScroogeJr (< 20)

US housing is 50% undervalued.

April 01, 2008 – Comments (8)

I have repeated many times that we are in the early stage of a monster housing bull. But there are many occasions when this point becomes sort of hard to prove because all real estate is local and if someone feels bearish, he can always point to this or that community where the general rule does not apply. The best way to counter these objections is to look at the aggregate numbers and then work with statistical averages. This way you can model the cash flow from an average house so that your model takes into account both Phoenix and New York and you can see what's happening on the national scale and why home prices are set to double again. By the way, I hope that at least some of you were taken in by my previous post which was written to honor the April Fool's Day tradition. No, I am not shorting housing either in real life or in Caps. Not until prices double from today's level. Then I will think whether it's better to short or to stay on the long side.  

I will begin my argument by posting this piece of statistics: http://www.bea.gov/papers/pdf/RIPfactsheet.pdf

From this data (which is obsolete by 1.5 years, so the numbers should be multiplied by approximately 1.05-1.07) we can find the aggregate cash flows produced by all 100+ million houses in America. Turning this number into free cash flow by throwing out ficticious expences, we can arrive at a realistic PE ratio of the average American house. I am going to disregard such exotic types of housing as trailers and farm buildings because they are mostly irrelevant to the discussion.

But I am going to adjust my FCF model a little bit, because my goal is not to produce some number, whether meaningful or not. The correct context is essential here, assuming you don't want to postpone the purchase of your dream home only to find its price going to the stratosphere because you've done sloppy analysis. So I am going to ignore some types of expences, most notably mortgage interest and insurance. Some people will object to that. Well, it's democracy, so if you thing you have a better model, feel free to construct your own number. This is similar to the way we calculate FCF for stocks. Some people prefer to remove the cost of buybacks because they feel that it's just another part of management compensation. Other people argue that it's an investment to incentivize management to grow the business and as such should be considered a part of FCF. This is exactly what we have here. I am neglecting interest and mortgage insurance payments because my model focuses on finding the upper boundary for house values. I am interested in finding the point at which the majority of homeowners will become anxious to sell their houses. This price mark is crucial because this is when the bubble stops growing as the majority of its participants head to the exit door. The key assumption involved is that the government is going to keep inflating property values as long as possible. Again, this postulate is not the subject of discussion. If you feel uncomfprtable with it, by all means use some other key assumption that you think is closer to reality. In my model, the buble pops when the typical average homeowner decides that he should sell his house, put the proceeds in a bank or into S&P 500, and rent. Relocations and trade-downs don't count because they create as many buyers as sellers, so there is no overall accumulation or distribution of housing...only pure sales can make a sizable difference. I must emphasize this point: I am not measuring the well-being of this or that homeowner who is living a frugal lifestyle because his mortgage payments bite so painfully...or because he took a second mortgage and then HELOCed himself to the eyeballs...or because the mortgage came with house insurance and with PMI insurance...or because he built a McMansion in the middle of Death Valley and is now wondering why it's not appreciating. Some of these people may indeed choose to cash out or walk away from their troubles (assuming their lender was an idiot who offered them a HELOC to 120% of appraised value...but let's not shed too many tears, after all, we know the end to that story: the lender is getting bailed out by JPM). But these people are irrelevant to the thesis because after they are taken to cleaners, the growth resumes, prices go up, mortgage-backed securities trade above par value, bankers are eager to lend, equity sets new records, and everybody except renters is happy again. What we need is a mass exodus when every Joe and their brother would want to get rid of properties and start renting. Otherwise we can only have a correction, but corrections pose no threat to the bubble if you look beyond a one-year time frame.  And what does that mean in turn? It means that the homeowner who owns the place free and clear and has no mortgage interest to serve should also stand in line to cash out. This is when further growth becomes impossible, and we can say we've reached the limit. This is why I'm not counting interest payments, mortgage insurance payments and other such crap. The decision to stop the bubble will be made by people who have no mortgage to pay off, and yet the incoming cash flow generated by their house will be equal to the national average. But again, it's a matter of your personal philosophy, just like FCF for stocks. If you believe that share dilution goes a long way toward enhancing job performance of your CEO, you count the matching buybacks as a part of capex, and if you believe that options have little correlation with performance, then you count buybacks as an operating expense. The argument between these two schools is not about whose arithmetics is correct, but about who's got a better model of CEO behavior. So make your adjustments, substituting your preferred model of seller psychology.

A final disclaimer: this has nothing to do with my personal feelings. I am not "supporting" the bubble. I am just looking at the numbers and saying that such and such numbers are pointing toward such and such outcomes. Not that my position matters anyway. The process will run its due course whether we like it or not.

Anyway, let's get down to arithmetic using the link above (here and afterwards, the numbers are rounded for the sake of conveniance). The total cash flow from housing (rental revenue + owner's equivalent rent) is1.3 trln. Admittedly, the actual number could be somewhat lower because owners' equivalent rent is based by owners' estimates, which might conceivably be a few percent off the mark. I subtract the 270 bln of "intermediate goods and services consumed". This includes maintenance, repairs, property insurance, property management fees, and even brokers' commissions on land and closing costs (sic!). I then subtract 215 bln of property taxes. As I said, I'm not subtracting "net interest payed" (primarily mortgage interest); furthermore, I am not subtracting subsidies for Section 8 housing, because the question who pays the rent - the tenant of the government program - is immaterial to the landlord, and of course I'm not subtracting depreciation, because it is an accounting gimmick (wear and tear is already accounted for - see maintenance and repairs). As a bone to the bears, I will even subtract 15 bln worth of compensation to employees (in case you are terrified of watering the lawn on your own). So we have 500 bln of expenses, and 800 bln of FCF.

Next, what multiple do we put on that FCF? Once again, I don't mean the multiple that Warren Buffett would consider cheap. After all, this is not a stock. A government-sponsored bubble grows until motivated sellers appear in large numbers, and these have to be sellers motivated by considerations of profit because distressed sellers alone will not suffice. That's what distinguishes it from an ordinary bubble which easily pops when first distressed sellers appear, because everybody in the pyramide was a mo-mo investor...nobody is intending to wait it out. For instance, would 20x make sense? No, because if you sell for 20x and rent, then after a mere 20 years you will have ho money and no house. (A conterargument: "But I will invest the proceeds!" Reality check: you will not invest it all because that would be a Jim Cramer proposition, and any conservative 50% stocks, 50% CD investment will only keep you in line with rent inflation, so you will run out of money in 20 years). And there is no downside risk to owing because 20x would be a fair valuation for a defensive stock. How about 30x? No, this is not very attractive either, because the downside risk of about 33% (all the way down to 20x valuation) does not justify losing both money and house over the period of 30 years (Objection: "But I can buy back later!" Reality check: very low probability of determining the top and the botom accurately, plus 7-8% in frictional costs mean that your most likely profit will be zero even if the pullback materializes as expected). How about 40x? Yes, this may produce some sellers, 40 years worth of rent is a large sum of money, especially for people whose actuarial life expectancy is shorter than that. But...1) why sell when you can reverse-mortgage? 2) what about your sentimental attachment to the home? 3) now, at 40x, there's surely going to be a tax if you sell and don't buy again, so to get 40x, your actual sale price should be higher, 4) by then, CDs will be paying 1% at most, and S&P will seem dangerously overvalued (after all, this is 40x, so you can be sure by that moment Bernanke's successors have brought the Fed rate down to zero because how else would they achieve such valuations?), so inflation will be eating your proceeds much faster. While 40x is the conservative estimate, 50x looks like a more realistic valuation. However, this model assumes that  1) rental rates as proportion of incomes will not rise further, 2) there will be no additional perks associated with homeownership, and 3) Fed Funds rate will not go below zero. None of these assumptions looks realistic, and it is extremely unlikely that all of them are valid. It is only reasonable therefore to admit the possibility of 60x multiple, just 20% above the 50x valuation.

Applying these valuations, I put the ultimate value of American housing between 32 trl and 48 trl dollars in 2006 money. In 2006, the total value of houses stood at 20.3 trl, it is likely a couple of points lower today, so let's round it off to 20 trl. This means the market is currently assigning them a super-conservative valuation of 25. My valuation model says that real prices should double from here, but they could grow anywhere from 60% to 140% depending on circumstances.  Once we grow 80%, we'll be approaching the danger zone, and if we pass the 120% mark, you should seriously consider shorting, but there will still be plenty of risk on the short side. That will be the level where house values stood in Japan, and we know that the Japanese government did a very sloppy job supporting them. The American bubble should have no trouble reaching these heights before running out of steam.

Risks: The first risk to that valuation is that the bubble might lose its political support. That would be a crushing blow to the model because as we all know, the fair value of housing in a free market is 100K for a new house and 50K for an old house, so if you believe there's going to be a free market in housing, you will be right to short the Case-Shiller index forever. Personally, I'd be more worried about the invasion of little green men from Mars, but if you believe it, you should definitely stay on the short side. The second risk is that more people will choose to leave the country for the tropics where housing is more affordable (or, alternatively, fewer people will immigrate to the US). England is already showing some symptoms of this response. Polls show that more and more Britons express willingness to leave the country to get rid of mortgage payments (the English bubble seems to be far worse than ours, and London prices make Manhattan look cheap). But Americans are the world's champions in sales promotion, so it is likely that the US will still be able to sell the "American dream" both at home and abroad even as affordability worsens enough to make it just that - a dream. The final risk is that the health care system could grow even more voracious than it is today and gobble up the baby-boomers' housing equity, forcing more people to trade down from a McMansion to a one-week timeshare in a hospital room. That would put a large distressed inventory on the market, severely limiting the upside. Seriously, though, I think the economy no longer has much room for two bubbles. In the long run, either housing or health care will have to give, and the health care bubble looks more vulnerable because it does not have the support of 68% of the population.

Summary: We have found an asset class that is poised to double over the next decade. Nominal returns should be higher due to inflation. A leveraged RE investment should comfortably outperform S&P for the patient owners who are both willing and able to stomach the current turbulence.